MEETING ABSTRACTS

Driving Knowledge to Action:

Building a Better Future
 61st
Annual Meeting of the
American Society of Clinical Oncology

May 30-June 3, 2025

2025 Annual Meeting Abstracts
(a supplement to Journal of Clinical Oncology)

Copyright 2025 American Society of Clinical Oncology

Editor: Ryan Gentzler, MD, MS

Associate Editor: Yara Abdou, MD

Managing Editor: Krystal Ingram

Production Manager: Julie Crose

The 2025 ASCO Annual Meeting Proceedings are published by Wolters Kluwer on behalf of the American Society of Clinical Oncology.

Requests for permission to reprint abstracts should be directed to healthpermissions@[Link]. Editorial correspondence should
be directed to abstracts@[Link].

Copyright © 2025 American Society of Clinical Oncology. All rights reserved. No part of this publication may be reproduced or transmitted in
any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without
written permission by the Society.

The abstracts contained within are copyrighted to Journal of Clinical Oncology and are reprinted within the Proceedings with Permissions.

The American Society of Clinical Oncology assumes no responsibility for errors or omissions in this publication. The reader is advised to check
the appropriate medical literature and the product information currently provided by the manufacturer of each drug to be administered to verify
the dosage, the method and duration or administration, or contraindications. It is the responsibility of the treating physician or other health care
professional, relying on independent experience and knowledge of the patient, to determine drug, disease, and the best treatment for the
patient.

Abstract management provided by CONFEX, Cumberland, RI. Composition services provided by Wolters Kluwer Health.
2025 ASCO Annual Meeting Abstracts
Guide to Abstracts

Plenary Session .............................................................................................................................................................................. 1s

Special Clinical Science Symposia ................................................................................................................................................ 3s

Breast Cancer—Local/Regional/Adjuvant .................................................................................................................................... 5s

Breast Cancer—Metastatic ............................................................................................................................................................ 37s

Care Delivery/Models of Care ........................................................................................................................................................ 72s

Central Nervous System Tumors .................................................................................................................................................. 112s

Developmental Therapeutics—Immunotherapy ......................................................................................................................... 138s

Developmental Therapeutics—Molecularly Targeted Agents and Tumor Biology ................................................................. 188s

Gastrointestinal Cancer—Colorectal and Anal ............................................................................................................................ 235s

Gastrointestinal Cancer—Gastroesophageal, Pancreatic, and Hepatobiliary .......................................................................... 272s

Genitourinary Cancer—Kidney and Bladder ................................................................................................................................ 328s

Genitourinary Cancer—Prostate, Testicular, and Penile ........................................................................................................... 361s

Gynecologic Cancer ........................................................................................................................................................................ 394s

Head and Neck Cancer ................................................................................................................................................................... 427s

Hematologic Malignancies—Leukemia, Myelodysplastic Syndromes, and Allotransplant .................................................... 458s

Hematologic Malignancies—Lymphoma and Chronic Lymphocytic Leukemia ...................................................................... 480s

Hematologic Malignancies—Plasma Cell Dyscrasia ................................................................................................................... 504s

Lung Cancer—Non–Small Cell Local-Regional/Small Cell/Other Thoracic Cancers .............................................................. 521s

Lung Cancer—Non–Small Cell Metastatic .................................................................................................................................. 553s

Medical Education and Professional Development ..................................................................................................................... 594s

Melanoma/Skin Cancers ................................................................................................................................................................ 606s

Pediatric Oncology ......................................................................................................................................................................... 634s

Prevention, Risk Reduction, and Genetics ................................................................................................................................... 653s

Quality Care/Health Services Research ........................................................................................................................................ 685s

Sarcoma ........................................................................................................................................................................................... 732s

Symptom Science and Palliative Care .......................................................................................................................................... 754s

Letter From the Editor

The 2025 ASCO Annual Meeting Abstracts (a supplement to Journal of Clinical Oncology) is an enduring record of the
more than 3,000 abstracts selected by the ASCO Scientific Program Committee for presentation as part of the ASCO
Annual Meeting.

Publication-only abstracts are included in the online supplement to the June 1 issue of Journal of Clinical Oncology at
[Link]. Abstracts can be also accessed online through ASCO Meeting Experience ([Link]). Online
abstracts include the full list of abstract authors and their disclosure information. All abstracts carry Journal of Clinical
Oncology citations, for example:

J Clin Oncol 43:16s, 2025 (suppl; abstr 500).

Should you have any questions or comments about this publication, I encourage you to provide feedback by contacting
abstracts@[Link].

Sincerely,

Ryan D. Gentzler, MD, MS

ASCO Meeting Abstracts Editor, Journal of Clinical Oncology

Yara Abdou, MD
ASCO Meeting Abstracts Associate Editor, Journal of Clinical Oncology
ASCO Abstracts Policy

Public Release of Abstracts

The 2025 ASCO Annual Meeting Abstracts were publicly released by ASCO at 5:00 PM EDT on Thursday, May 22,
2025. Abstracts are publicly available online at [Link]/abstracts. Late-Breaking Abstracts, which include all
Plenary Abstracts, will be publicly released according to the following schedule:

• Late-Breaking Abstracts presented in a scientific presentation on Friday, May 30, will be publicly released Friday,
May 30, at 8:00 AM EDT.
• Late-Breaking Abstracts presented in a scientific presentation on Saturday, May 31, will be publicly released
Saturday, May 31, at 8:00 AM EDT.
• Late-Breaking Abstracts presented in a scientific presentation on Sunday, June 1, will be publicly released Sunday,
June 1, at 8:00 AM EDT.
• Late-Breaking Abstracts presented in a scientific presentation on Monday, June 2, will be publicly released Monday,
June 2, at 8:00 AM EDT.
• Late-Breaking Abstracts presented in a scientific presentation on Tuesday, June 3, will be publicly released Tuesday,
June 3, at 8:00 AM EDT.

In the unlikely event that ASCO publicly releases an abstract in advance of the scheduled time, the release will be publicly
announced on [Link]
exceptions/embargo-abstract-release.

Conflict of Interest Disclosure

All financial relationships reported by contributors to this activity are provided to learners prior to the start of the
activity. During planning and development of the activity, relevant financial relationships were mitigated for all
contributors. Relationships are considered self-held and compensated unless otherwise noted (I 5 Immediate family
member; Inst 5 My Institution). Disclosures are submitted per the ASCO Policy for Relationships with Companies.
 PLENARY SESSION 1s

LBA1 Plenary Session LBA2 Plenary Session

Randomized trial of standard chemotherapy alone or combined with ate- NIVOPOSTOP (GORTEC 2018-01): A phase III randomized trial of adjuvant
zolizumab as adjuvant therapy for patients with stage III deficient DNA nivolumab added to radio-chemotherapy in patients with resected head and
mismatch repair (dMMR) colon cancer (Alliance A021502; ATOMIC). First neck squamous cell carcinoma at high risk of relapse. First Author: Jean
Author: Frank A. Sinicrope, Mayo Clinic Rochester, Rochester, MN Bourhis, CHUV, Bâtiment Hospitalier, Lausanne, Switzerland

The full, final text of this abstract will be available at The full, final text of this abstract will be available at
[Link] on the day of presentation and in the [Link] on the day of presentation and in the
online supplement to the June 10, 2025, issue of the Journal online supplement to the June 10, 2025, issue of the Journal
of Clinical Oncology. of Clinical Oncology.

LBA3 Plenary Session LBA4 Plenary Session

Results from VERIFY, a phase 3, double-blind, placebo (PBO)-controlled Camizestrant + CDK4/6 inhibitor (CDK4/6i) for the treatment of emergent
study of rusfertide for treatment of polycythemia vera (PV). First Author: ESR1 mutations during first-line (1L) endocrine-based therapy (ET) and
Andrew Tucker Kuykendall, Moffitt Cancer Center, Tampa, FL ahead of disease progression in patients (pts) with HR+/HER2– advanced
breast cancer (ABC): Phase 3, double-blind ctDNA-guided SERENA-6 trial.
First Author: Nicholas C. Turner, Royal Marsden Hospital, London, United Kingdom

The full, final text of this abstract will be available at The full, final text of this abstract will be available at
[Link] on the day of presentation and in the [Link] on the day of presentation and in the
online supplement to the June 10, 2025, issue of the Journal online supplement to the June 10, 2025, issue of the Journal
of Clinical Oncology. of Clinical Oncology.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
2s PLENARY SESSION

LBA5 Plenary Session

Event-free survival (EFS) in MATTERHORN: A randomized, phase 3 study of
durvalumab plus 5-fluorouracil, leucovorin, oxaliplatin and docetaxel che-
motherapy (FLOT) in resectable gastric/gastroesophageal junction cancer
(GC/GEJC). First Author: Yelena Y. Janjigian, Gastrointestinal Oncology Service,
Memorial Sloan Kettering Cancer Center, New York, NY

The full, final text of this abstract will be available at

[Link] on the day of presentation and in the
online supplement to the June 10, 2025, issue of the Journal
of Clinical Oncology.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 SPECIAL CLINICAL SCIENCE SYMPOSIA 3s

100 Clinical Science Symposium 101 Clinical Science Symposium

Performance evaluation of a reflex blood-based methylated ctDNA multi- Prognostic significance of blood-based multi-cancer detection in circulating
cancer early detection test in individuals with obesity. First Author: Dax Kur- tumor DNA (ctDNA): 5-year outcomes analysis. First Author: Robert Charles
begov, Sarah Cannon Research Institute, Nashville, TN Swanton, Cancer Evolution and Genome Instability Laboratory, The Francis Crick In-
Background: Obesity increases risk of 13 distinct cancers, most without screening stitute, London, United Kingdom
programs, and collectively representing 40% of all annual US cancer [Link] Background: In the case-control Circulating Cell-free Genome Atlas (CCGA) study
recent CDC data (2023) showing that obesity rates exceed 35% in 23 U.S. states, multi- (NCT02889978), a multi-cancer early detection (MCED) test was developed that uses next-
cancer early detection (MCED) testing represents a unique opportunity to address a generation sequencing to detect a cancer signal shared across . 50 cancer types using
critical screening gap in this population. In this case-control study, we examined the ctDNA in blood. The concentration of ctDNA in blood is associated with cancer ag-
performance of a reflex blood-based ctDNA methylation MCED test in individuals with gressiveness and prognosis. Previous analysis of participant outcomes in the second
obesity (BMI $ 30 kg/m2), evaluating cancer-specific intrinsic accuracy (sensitivity) and (cross-validation) CCGA substudy evaluated the prognostic value of cancer signal de-
positive predictive value (PPV). Methods: We analyzed peripheral blood samples tection by an early version of the MCED test with 3-year follow-up. Participants with
(NCT05435066) from individuals with obesity, including 424 treatment-naı̈ve cancer confirmed cancer and no cancer signal detected (NCSD) MCED test result had better 3-year
patients and 458 non-cancer controls, using a reflex test by Harbinger Health. The test survival than those with a cancer signal detected (CSD). In the present analysis, we
system consists of a primary methylome profiling test optimized for high sensitivity evaluated the prognostic value of cancer signal detection by a refined version of the MCED
(rule-out), followed by a confirmatory reflex analysis utilizing an expanded methylation test in the third (validation) CCGA substudy (CCGA3) using an updated statistical
panel designed to achieve high PPV (rule-in) and tissue of origin (TOO) classification. methodology with 5-year follow-up, a typical timeframe for cancer-survivor status.
Test performance metrics were derived using 10-fold cross-validation with patient-level Methods: Participant blood samples collected during CCGA3 were analyzed using the
stratification. Overall cancer incidence in the obese population was estimated at 1.6%. MCED test. Participants with confirmed cancer were followed for up to 5 years and their
Prospective PPV for each TOO readout was calculated using cancer-specific intrinsic overall survival stratified by cancer signal detection (CSD/NCSD). Observed survival was
accuracy estimated from the case-control study in combination with incidence values. compared to the expected survival of a reference population calculated using Surveillance,
Results: In this cohort (mean age 57.1 6 13.4 years; 63.3% female; 67.8% White; 22.4%, Epidemiology, and End Results (SEER) data matched to the distribution of age, sex, cancer
Black or African American), the test achieved an overall 29.7% (95CI, 25.3-34.6) correct- type, and stage in each signal detection group. A one-sample proportional hazard model
was used to assess differences between observed and expected survival based on cancer
TOO sensitivity (66.1% false negative, 4.2% incorrect TOO) at 98.9% (95CI, 97.6-99.6)
signal detection status. Results: Follow-up data were available for 2475/2513 (99%) of
specificity and 98.9% negative predictive value (95CI, 98.85,99.0). Among a subset of
participants with stageable, invasive cancer. Of these, 792 (32%) died during follow-up,
cancers associated with obesity, TOO-specific PPVs were: hepatobiliary (HB: liver, biliary
673/792 (85%) of whom had a CSD; of the 1683 (67%) participants alive at follow-up, 579/
duct; 100%; 95CI, ND), CRC (87.5; 95CI, 61.1-96.9), upper gastrointestinal (UGI; 81.8;
1683 (35%) had a CSD. Overall observed survival rates of both groups were higher than the
CI95, 48.6-95.5), uterine (66.7%; 95CI, 26.5-91.7), pancreaticobiliary (PB: pancreas, expected survival rates based on SEER data matched for known clinical factors (43% vs
gallbladder; 17.6%; 95CI 3.1-59.0). Corresponding cancer group sensitivities (early 40% [CSD]; 88% vs 81% [NCSD]). Observed vs expected survival rates for participants with:
stage, I-II) were: HB 50.0 (35.1), CRC 51.9 (28.6), UGI 40.9 (9.1), uterine 8.5 (3.8), PB 56.5 stage I cancer were 66% vs 71% (CSD) and 90% vs 85% (NCSD); stage II cancer were 64% vs
(25.0). In a modeled cohort of 100,000 tested individuals, the expected TOO readout 67% (CSD) and 92% vs 83% (NCSD); stage III cancer were 48% vs 42% (CSD) and 79% vs
counts (correct case-type odds) were: 23 for HB (23:0), 60 for CRC (7:1), 42 for UGI (9:2), 66% (NCSD); stage IV cancer were 22% vs 16% (CSD) and 56% vs 32% (NCSD). Overall HR
23 for uterine (2:1), 279 for PB (1:5). Conclusions: Inindividuals with obesity at in- for NCSD vs CSD across all stages was 0.60 (95% CI: 0.50-0.72; P = 6.18e-09). HRs for
creased risk for multiple cancers, especially those lacking established screening signal detection group vs SEER reference populations were , 1 at all stages with NCSD;
guidelines, the reflex ctDNA methylation MCED test demonstrated clinically meaningful with CSD, HRs were , 1 at stages III and IV and $1 at stages I and II. Conclusions: In
PPV and early-stage sensitivity for each cancer type. These results warrant prospective CCGA3, 5-year follow-up confirmed that while CSD was associated with hazard of death,
validation to assess the test’s clinical validity and utility in early-stage cancer detection early-stage cases had long-term survival close to expected rates. These results suggest
in this high-risk population. Research Sponsor: None. that a CSD MCED test result may inform prognosis and urgency of treatment. Clinical trial
information: NCT02889978. Research Sponsor: GRAIL, Inc.

102 Clinical Science Symposium 103 Clinical Science Symposium

A phase 1 study of intracerebroventricular (ICV) delivery of bivalent chimeric Neoadjuvant biomarker trial of pepinemab to enhance nivolumab or ipili-
antigen receptor (CAR) T-cells targeting EGFR and IL13Ra2 in patients with mumab activity in resectable head and neck cancer. First Author: Conor Ernst
recurrent glioblastoma (rGBM). First Author: Stephen Joseph Bagley, University of Steuer, Emory University, Atlanta, GA
Pennsylvania, Philadelphia, PA Background: Neoadjuvant treatment with immune checkpoint blockade (ICB) improves
Background: Outcomes in patients with rGBM are poor, with historical median overall clinical benefit in patients with multiple types of cancers. Window of opportunity studies
survival (OS) of 6-9 months. Here we report the results from the dose exploration phase of a permit integrated assessments of safety and efficacy, including biomarker assessments
phase 1 trial investigating ICV-delivered, bivalent CAR T-cells targeting EGFR epitope 806 of treatment effects and mechanisms of resistance in the tumor microenvironment.
and IL13Ra2 in rGBM. Methods: Patients with EGFR-amplified GBM that was recurrent/ SEMA4D blocking antibody, pepinemab (pepi), has been reported to overcome resistance
progressive following front-line radiotherapy were enrolled using a 3+3 design (dose levels: mechanisms including immune exclusion and myeloid suppression in preclinical and
5.0 x 106, 1.0 x 107, and 2.5 x 107 cells). Patients underwent surgery for (1) maximal safe clinical studies. We conducted a neoadjuvant integrated biomarker study (NCT03690986)
resection and confirmation of viable tumor and (2) Ommaya reservoir placement. Post- to evaluate the effect of pepi alone and in combination with ICB on the immune profile in
operatively, patients received a single ICV dose of CART-EGFR-IL13Ra2 cells without the tumor and blood of patients with resectable head and neck squamous cell carcinoma
lymphodepleting chemotherapy. Primary endpoints included dose-limiting toxicity (DLT) (HNSCC). Methods: Patients were randomized to receive one dose of either pepi alone,
and determination of the maximum tolerated dose (MTD). Secondary endpoints included pepi/ipilimumab (ipi), pepi/nivolumab (nivo), ipi, nivo (n = 6 patients/group), or no
objective radiographic response, progression-free survival (PFS), and OS. Serial CSF treatment (n = 4); followed by surgery within days 17-36. The primary objective is
samples were analyzed for CAR T-cell pharmacokinetics and single-cell RNA sequencing biomarker assessments; clinical endpoints include pathologic response (pMR), safety,
(scRNAseq). Results: Eighteen patientsreceived CART-EGFR-IL13Ra2 cells (n=6 per dose surgical delays, RFS, and OS. Analysis of pretreatment and surgically resected tissue and
level). Median age was 57, 15 (83%) were male, 13 (72%) had MGMT unmethylated tumors, blood to evaluate spatial distribution of tumor and immune populations employed high
and 7 (39%) had .1 prior relapse. One DLT (grade 3 lethargy/fatigue) was observed at the dimensional 36+ multiplex IHC and 32-color flow cytometry. Biomarker results were
MTD (2.5 x 107 cells). Acute neurotoxicity related to CAR T-cells occurred in all patients. stratified by demographic and clinical outcome measures. Results: Thirty-four patients
Using immune effector cell-associated neurotoxicity syndrome (ICANS) grading, 10 of 18 were enrolled (median age 63 (58-69); 70.6% male; 79.4% OC, 20.6% OP; 82.4% HPV/p16
patients (56%) experienced grade 3 neurotoxicity; none had grade 4-5 neurotoxicity. Using Neg). All patients proceeded to surgery without delay; no additional or unexpected TRAEs
tumor-inflammation associated neurotoxicity (TIAN) grading, 2 of 18 patients (11%) had
were observed in pepi combinations; 9/10 patients who experienced TRAEs were Grade 1-
grade 3 and 1 patient (6%) had grade 4 neurotoxicity. Grade 1-2 fever occurred in all patients.
2. Biomarker analysis was stratified by HPV status due to vast difference between highly
Eleven of 13 patients (85%) with measurable disease at time of CAR T-cell infusion ex-
infiltrated HPV+ compared to immunologically cold HPV- TME. Among 24 available HPV-
perienced tumor shrinkage, ranging from 1-62% reductions (median 35%, IQR 12 – 39%) in
resected tumors,an increase in the number of intratumoral tertiary lymphoid structures
target lesions and with one confirmed partial response by modified RANO criteria. PFS and
(TLS) was observed in pepinemab containing cohorts, whereas B cells lined the tumor
OS continue to mature and will be presented. CAR T-cell expansion in CSF was robust with a
dose-response relationship observed. The CAR transgene remained detectable in CSF and edge and were generally excluded from tumor bed in cohorts lacking pepinemab. A
blood 12 months post-CART infusion in a patient experiencing durable stable disease lasting significant increase in density of mature TLS (including CD21+ follicular DC and CD23+
for 17 months (ongoing at data cut-off). In patients undergoing repeat resection following germinal center B cells) was observed in patients treated with pepi+nivo compared with
treatment, CART-EGFR-IL13Ra2 cell infusion markedly increased the number of tumor- pepi or nivo alone and untreated patients. This finding was unexpected, as mature TLS are
infiltrating lymphocytes. scRNAseq in post-treatment CSF revealed higher cytotoxicity and generally rare in poorly immunogenic HPV-negative HNSCC. Clinical assessments of
exhaustion scores in CD8+ CAR T-cells as compared to the infusion product, indicative of pathologic response and RFS are being analyzed. Conclusions: Neoadjuvant treatment
target cell engagement. Conclusions: ICV delivery of CART-EGFR-IL13Ra2 is feasible and with pepi enhanced the density and maturity of TLS deep within the tumor which was
appears safe. CART-EGFR-IL13Ra2 cells are bioactive and exhibit an encouraging early most prominent in combination with nivo notably in HPV-negative disease. Pepi
efficacy signal in rGBM. Clinical trial information: NCT05168423. Research Sponsor: Kite represents a novel strategy to boost tumor immunity and organization of functional TLS
Pharma (a Gilead company). to overcome limitations of ICB in HPV- HNSCC. Clinical trial information: NCT03690986.
Research Sponsor: Winship Cancer Institute of Emory University, Atlanta, GA.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
4s SPECIAL CLINICAL SCIENCE SYMPOSIA

104 Clinical Science Symposium 105 Clinical Science Symposium

Efficacy and safety of IBI363 monotherapy or in combination with bevaci- Safety and efficacy of ABBV-706, a seizure-related homolog protein 6
zumab in patients with advanced colorectal cancer. First Author: Zhenyu Lin, (SEZ6)–targeting antibody-drug conjugate, in high-grade neuroendocrine
Cancer Center, Union Hospital, Tongji Medical College, Huazhong University of Science neoplasms. First Author: Alissa Jamie Cooper, Department of Thoracic Oncology,
and Technology, Wuhan, China Memorial Sloan Kettering Cancer Center, New York, NY
Background: The prognosis of patients (pts) with microsatellite stable/proficient Background: SEZ6 is a potential neuroendocrine lineage marker that is expressed in small
mismatch repair (MSS/pMMR) colorectal cancer (CRC) who failed standard chemo- cell lung cancer (SCLC) and high-grade neuroendocrine neoplasms (NENs). NENs have a
therapy is poor, highlighting a significant unmet need. No immune-oncology therapy has significant unmet need for novel effective targeted therapies. ABBV-706, a unique
succeeded in this indication due to the "cold" tumor nature. IBI363 is a PD-1/IL-2a-bias antibody-drug conjugate comprising a SEZ6-directed antibody conjugated to a potent
bispecific antibody fusion protein that blocks PD-1 and stimulates tumor-specific T cells topoisomerase 1 inhibitor payload, is being evaluated in a phase 1 study (NCT05599984) in
that could potentially turn "cold" tumors into "hot" tumors. Methods: The analysis of patients (pts) with advanced solid tumors. Preliminary results from ABBV-706 mono-
efficacy and safety data were from 68 pts treated with IBI363 monotherapy and 73 pts therapy dose escalation showed a manageable safety profile with promising efficacy in
treated with IBI363 plus bevacizumab (beva), respectively. Eligible pts were locally SCLC and NENs (J Clin Oncol 2024;42[suppl 16]: abs 3001). Herein, updated safety and
advanced unresectable or metastatic CRC who failed or were intolerant to the standard efficacy of ABBV-706 monotherapy in NENs are presented. Methods: Pts ($18 yr) with
treatment. Data cutoff date was Dec 6, 2024. Results: A total of 68 pts and 73 pts (None relapsed/refractory high-grade NENs (well-differentiated grade [G] 3 neuroendocrine tu-
were confirmed as microsatellite instability-high/deficient mismatch repair [MSI-H/ mors [G3 NETs] and poorly differentiated neuroendocrine carcinomas [NECs]), atypical
dMMR]; MSS/pMMR: 86.8% and 90.4%; unknown microsatellite/MMR status: 13.2% and lung carcinoid, and medullary thyroid cancer (MTC) were enrolled in dose-escalation and
9.6%; liver metastases: 61.8% and 54.8%; KRAS/NRAS mutations: 42.6% and 41.1%; -expansion cohorts of a phase 1, open-label study. Pts received ABBV-706 monotherapy IV
previous treatment lines $ 3: 63.2% and 53.4%; previous immunotherapy: 27.9% and at 1.3–3.5 mg/kg once every 3 weeks. Primary study objectives are assessment of safety,
16.4%) were treated with IBI363 monotherapy (0.1 mg/kg to 3 mg/kg every week [QW], PK, and efficacy. SEZ6 expression is evaluated retrospectively. Results: As of Aug 27,
every 2 weeks [Q2W] or every 3 weeks [Q3W]) and IBI363 plus beva (0.6 or 1 mg/kg Q2W, 2024, 191 pts were enrolled overall, including 64 with NENs. In the NEN cohort, median age
1.5, 2 or 3 mg/kg Q3W, plus beva 5 mg/kg Q2W or 7.5 mg/kg Q3W), respectively. Median was 63 yr (range 33–86) and pts had received a median of 3 (range 1–8) prior therapies.
follow-up time was 11.8 months (range: 0.4–22.5) for monotherapy and 5.1 months NEN histologies were large cell NEC (LCNEC; 22%, n=14), gastro-enteropancreatic NEC
(GEPNEC; 19%, n=12), MTC (9%, n=6), neuroendocrine prostate carcinoma (NEPC; 8%, n=5),
(range: 1.2–14.9) for combination. In efficacy-evaluable pts (n = 63 for monotherapy and
G3 NETs (8%, n=5), and other NECs (34%, n=22). The safety profile for ABBV-706 was
n = 68 for combination), the objective response rate (ORR) was 12.7% (95% confidence
similar across NEN subtypes and aligned with the entire study population. For the overall
interval [CI]: 5.6–23.5) and 23.5% (95% CI: 14.1–35.4). The median duration of response
study population, TEAEs occurred in 184 (96%) pts and G$3 in 134 (70%). Most frequent
was 7.5 months (95% CI: 1.2–19.6) for monotherapy and not mature for combination.
hematologic TEAEs were anemia (58%; G$3: 45%), neutropenia (44%; G$3: 33%), and
The median OS was 16.1 months (95% CI: 10.1–not reached) for monotherapy and not thrombocytopenia (35%; G$3: 21%). Most frequent nonhematologic TEAEs were fatigue
mature for combination. Especially, in pts without liver metastasis who received the (45%; G$3: 3%) and nausea (38%; G$3: 2%). Unadjudicated pneumonitis/interstitial lung
combination therapy (n = 31), the ORR was 38.7% (95% CI: 21.9–57.8), the DCR was disease rate was 4% (G$3 in 2 pts). For the entire NEN cohort, the objective response rate
83.9% (95% CI: 66.3–94.6), and median PFS was 9.6 months (95% CI: 4.1–12.2). (ORR) was 31.3% (20/64) and the clinical benefit rate was 92.2% (59/64). ORR by NEN type
Grade $ 3 treatment-related adverse events were reported in 16 (23.5%) pts with was: LCNEC, 28.6% (4/14); GEPNEC, 16.7% (2/12); NEPC, 60.0% (3/5); G3 NET, 60% (3/5);
monotherapy and 22 (30.1%) pts with combination. Arthralgia, rash, and thyroid dis- MTC, 16.7% (1/6); other NEC, 31.8% (7/22). The median duration of response was 5.59 mo
orders were commonly reported immune-related adverse events. Conclusions: IBI363 (95% CI: 4.24, not estimable) and median progression-free survival was 6.80 mo (95% CI:
monotherapy demonstrated prolonged overall survival in pts with advanced CRC 5.45, 7.75). Correlation analysis of efficacy with SEZ6 expression is ongoing.
compared to historic data of standard of care. IBI363 plus beva showed even more Conclusions: ABBV-706 showed preliminary efficacy in several high-grade NENs with a
encouraging efficacy with acceptable safety and warrants further development. Clinical high unmet need, supporting its further development in specific subtypes. Clinical trial
trial information: NCT05460767. Research Sponsor: None. information: NCT05599984. Research Sponsor: AbbVie Inc.; n/a

106 Clinical Science Symposium 107 Clinical Science Symposium

Phase 1 study of SHR-1826, a c-MET–directed antibody-drug-conjugate Phase 1 trial of SHR-A2102, a nectin-4–directed antibody drug conjugate
(ADC), in advanced solid tumors. First Author: Yang Zhang, Sun Yat-sen University (ADC), in advanced solid tumors. First Author: Runbo Zhong, Shanghai Chest
Cancer Center, Guangzhou, China Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
Background: MET alterations are key drivers of diverse oncogenic processes, including tumor in- Background: Nectin-4 is a cell adhesion molecule that is highly expressed in a wide variety of
vasion, growth, and metastasis, and are associated with poor prognosis. SHR-1826 is a novel ADC of a cancers and is associated with poor prognosis. SHR-A2102 is a novel ADC consisting of a fully
humanized c-MET-directed IgG2 monoclonal antibody attached to a topoisomerase I inhibitor payload humanized Nectin-4–directed monoclonal antibody, bound to topoisomerase I inhibitor payload
via a tetrapeptide-based cleavable linker. We conducted a multi-center, first-in-human, phase 1 trial of via a cleavable linker. We conducted a multi-center, phase 1 trial to evaluate SHR-A2102 in
SHR-1826 in advanced solid tumors, and here report preliminary results from the dose-escalation and advanced solid tumors. Methods: Patients (pts) with Nectin-4 positive, locally advanced unre-
expansion portions. Methods: Patients (pts) with advanced solid tumors harboring MET alterations sectable or metastatic solid tumors were enrolled. The study included dose-escalation (D-ESC),
(overexpression, amplification, or activating mutation) who had failed standard therapy or no available dose-expansion (D-EXP) and efficacy-expansion (E-EXP) phases. SHR-A2102 was given IV at 2–10
standard therapy, were enrolled. The study consisted of dose-escalation (i3+3 design), dose-expansion mg/kg, Q3W during D-ESC, and at 6 mg/kg and 8 mg/kg Q3W during D-EXP and E-EXP. The primary
and efficacy-expansion phases, during which pts received SHR-1826 at 2.2–6.0 mg/kg, Q3W, iv. objectives were to assess safety and tolerability. Results: As of Dec. 20, 2024,369 pts were
Primary objectives were to assess safety and tolerability. Results: As of Dec.5, 2024, 116 pts were enrolled and treated (median age, 59 y; ECOG PS 1, 85.6%; $2 lines of prior therapy, 64.0%). During
enrolled and treated (NSCLC/CRC/GC/PC, n=72/32/10/2; median age, 59.2 yrs; ECOG PS 1, 87.9%; $3 D-ESC, DLT occurred in 1 pt (10 mg/kg; grade 4 decreased platelet count). Overall, grade $3 TRAEs
lines of prior therapy, 44.0%; median c-MET H-score, 163 [range 9-300]). During dose-escalation, 1 DLT occurred in 167 (45.3%) pts, with the most common ($3%) being decreased neutrophil count
was observed at 6.0 mg/kg (grade 3 febrile neutropenia). Grade $3 TRAEs were reported in 56 (48.3%) (25.5%), decreased white blood cell count (16.3%), anaemia (11.7%), decreased lymphocyte count
pts, with the most common being decreased neutrophil count (32.8%), decreased white blood cell (8.7%), decreased platelet count (4.9%), asthenia (3.5%) and nausea (3.3%). 2 (0.5%) pts dis-
count (22.4%), anaemia (13.8%), and decreased platelet count (11.2%). Interstitial lung disease continued treatment due to TRAE. ILD occurred in 1 (0.3%; grade 3) pt. In 304 evaluable pts for
occurred in 3 (2.6%; grade 1-2, n=2; grade 3, n=1) pts. 2 (1.7%) pts discontinued treatment due to response, ORR was 35.2% (95% CI 29.8-40.9) and DCR was 84.2% (95% CI 79.6-88.1). As of data
TRAE. There were no treatment-related deaths. Among 58 evaluable pts with NSCLC, ORR was 39.7%
cutoff, 146 (39.6%) pts had disease progression or died; median PFS was 4.7 mo (95% CI 4.3-5.6).
(95% CI 27.0–53.4) and DCR was 94.8% (95% CI 85.6–98.9; Table 1); response was observed across a
Efficacy in selected tumor types is shown in Table. Conclusions: SHR-A2102 demonstrated a
range of c-MET expression levels, and in both EGFR-mutated and wild-type tumors. Median duration of
manageable safety profile and promising activity across a variety of pretreated advanced solid
response was not reached, with 21 of 23 responses ongoing. In all 72 NSCLC pts, median progression-
tumors. Multiple trials are ongoing to further assess SHR-A2102 both as monotherapy and in
free survival was 6.8 mo (95% CI 4.5–7.2). Conclusions: SHR-1826 demonstrated a manageable
safety profile in pts with heavily pretreated advanced solid tumors. Promising anti-cancer activity was
combination therapy for solid tumors. Clinical trial information: NCT05701709. Research Sponsor:
observed in MET-altered NSCLC, warranting further investigation in this population. Clinical trial Jiangsu Hengrui Pharmaceuticals.
information: NCT06094556. Research Sponsor: Jiangsu Hengrui Pharmaceuticals Co., Ltd. Efficacy in selected tumor types (efficacy evaluable set).
EGFR-mut Nsq EGFR-wt Nsq Sq NSCLC HR+/HER2-BC TNBC HNSCC All patients†
Efficacy in pts with NSCLC. NSCLC (N=69) NSCLC (N=44) (N=44) (N=20) (N=32) (N=12) (N=304)
2.2 mg/kg 4 mg/kg 5 mg/kg 6 mg/kg All patients
Best overall
(n=2) (n=24) (n=31) (n=1) (n=58)
response, n (%)
Best overall response, CR* 0 1 (2.3) 0 0 0 0 1 (0.3)
n (%) PR* 30 (43.5) 10 (22.7) 11 (25.0) 13 (65.0) 18 (56.3) 6 (50.0) 106 (34.9)
SD 28 (40.6) 21 (47.7) 33 (75.0) 4 (20.0) 9 (28.1) 5 (41.7) 149 (49.0)
Complete response 0 0 0 0 0
PD 11 (15.9) 11 (25.0) 0 3 (15.0) 5 (15.6) 1 (8.3) 47 (15.5)
Partial response* 0 9 (37.5) 13 (41.9) 1 (100.0) 23 (39.7) Not evaluable 0 1 (2.3) 0 0 0 0 1 (0.3)
Stable disease 2 (100.0) 14 (60.9) 16 (51.6) 0 32 (55.2) ORR*, % (95% CI) 43.5 25.0 25.0 65.0 56.3 50.0 35.2
Progressive disease 0 0 2 (6.5) 0 2 (3.4) (31.6-56.0) (13.2-40.3) (13.2-40.3) (40.8-84.6) (37.7-73.6) (21.1-78.9) (29.8-40.9)
Not evaluable 0 1 (4.2) 0 0 1 (1.7) DCR, % (95% CI) 84.1 72.7 100.0 85.0 84.4 91.7 84.2
ORR*, % (95% CI) 0.0 (0.0–84.2) 37.5 (18.8–59.4) 41.9 (24.5–60.9) 100.0 (2.5–100.0) 39.7 (27.0–53.4) (73.3-91.8) (57.2- 85.0) (92.0-100.0) (62.1-96.8) (67.2-94.7) (61.5-99.8) (79.6-88.1)
DCR, % (95% CI) 100.0 (15.8–100.0) 95.8 (78.9–99.9) 93.5 (78.6–99.2) 100.0 (2.5–100.0) 94.8 (85.6–98.9) PFS‡, mo (95% CI) 5.7 4.3 4.5 5.6 5.6 6.8 4.7
(5.1-NR) (2.0-7.3) (4.1-6.8) (4.3-NR) (4.3-7.1) (2.4-6.8) (4.3-5.6)
Data are shown for pts with $1 post-baseline assessment.
*Including 6 unconfirmed responses across groups. *Include responses to be confirmed.
†
Other tumor types include ESCC, PAAD, CRC, CC and UC.
‡
Evaluated in full analysis set (n=369).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 5s

LBA500 Oral Abstract Session 501 Oral Abstract Session

De-escalated neoadjuvant taxane plus trastuzumab and pertuzumab with or Predicting pathologic complete response (pCR) from clinicopathologic var-
without carboplatin in HER2-positive early breast cancer (neoCARHP): A iables and HER2DX genomic test in stage II/III HER2+ breast cancer treated
multicentre, open-label, randomised, phase 3 trial. First Author: Hong-Fei Gao, with taxane, trastuzumab, and pertuzumab (THP): Secondary results from
Department of Breast Cancer, Cancer Center, Guangdong Provincial People’s Hospital the EA1181/CompassHER2 pCR trial. First Author: Nadine M. Tung, Beth Israel
(Guangdong Academy of Medical Sciences), Southern Medical University, Guangzhou, Deaconess Medical Center, Boston, MA
China Background: EA1181 (NCT04266249) is a single-arm trial of neoadjuvant THP for patients with clinical
anatomic stage II/III HER2+ breast cancer; patients with cT4 or cN3 disease were excluded . Assessing the
primary endpoint, 3-year recurrence-free survival in patients with a pCR (ypT0/Tis, ypN0), requires longer
follow-up. Here, we present results for the secondary objective of pCR rate and its relation to clinico-
pathologic factors and the HER2DX pCR likelihood score (Reveal Genomics) derived from gene expression
and clinical features. Methods: Patients received 4 cycles of trastuzumab and pertuzumab (HP) with
weekly paclitaxel (12 weeks) or docetaxel (q3w x 4), followed by surgery. Clinicopathologic features were
assessed for all patients and HER2DX pCR score (stratified by ER status) was determined using the
diagnostic biopsy in a representative subset of study participants. Results: 2175 patients were enrolled.
Median age was 55 years (range 22-88 years); 58% had clinical stage IIA, 33% stage IIB, and 9% stage III.
45% had nodal involvement (mostly cN1). 781 tumors were HER2+/ER- and 1394 HER2+/ER+ (locally tested
). 2141 patients started THP, for whom the pCR rate was 44% overall, 63.7% in HER2+/ER- and 32.4% in
HER2+/ER+ tumors. Disease progressed during THP in 16 patients (0.7%). The pCR rate varied inversely to
The full, final text of this abstract will be available at the proportion of cells staining for ER among patients with HER2+/ER+ breast cancer: 1-10%+, 62.5%; 11-
69%, 51.6%; $70%, 22.5% (p ,0.001). The pCR rate was significantly associated with higher grade,
[Link] on the day of presentation and in the especially in HER2+/ER+ disease. T and N stage did not significantly affect pCR rate. Among 569 patients
online supplement to the June 10, 2025, issue of the Journal assessed for the HER2DX pCR score, the pCR rate was significantly greater for patients with a higher vs
lower score, regardless of ER status (Table). Further correlations and interactions will be presented.
of Clinical Oncology. Conclusions: Neoadjuvant THP resulted in pCR in nearly two-thirds of pts with clinical stage II/III HER2+/
ER- and in one-third with HER2+/ER+ breast cancer. There was no association with clinical stage. Lower ER
expression and higher grade were associated with higher pCR rates. The HER2DX pCR score was a
significant predictor of pCR, regardless of ER status. Clinical trial information: NCT04266249. Research
Sponsor: National Cancer Institute; Breast Cancer Research Foundation; Susan G. Komen for the Cure.
All Participants HER2+/ER- HER2+/ER+
# patients pCR rate # patients pCR rate # patients pCR rate
enrolled (95% CI) enrolled (95% CI) enrolled (95% CI)
Evaluable Cohort 2141 44% 774 64% 1367 32%
(42-46%) (60-67%) (30-35%)
HER2DX Cohort 569 48% 230 69% 339 34%
(44- 52%) (63-75%) (29-39%)
HER2DX pCR-high 182 (32%) 68% 147 (64%) 70% 36 (11%) 58%
score (60-74%) (62-77%) (41-74%)
HER2DX pCR-medium 161 (28%) 67% 70 (30%) 74% 89 (26%) 61%
score (59- 74%) (62- 84%) (50-71%)
HER2DX pCR-low 226 (40%) 19% 13 (6%) 31% 214 (63%) 18%
score (14- 25%) (9-61%) (13-24%)
P-value ,0.001 0.010 ,0.001
HER2DX score

502 Oral Abstract Session 503 Oral Abstract Session

Prediction of survival after de-escalated neoadjuvant therapy in HER2+ early Comparison of marking techniques for target lymph nodes in 2,596 patients
breast cancer: A pooled analysis of three WSG trials. First Author: Monika Karla with node-positive breast cancer treated with neoadjuvant chemotherapy:
Graeser, West German Study Group and Ev. Hospital Bethesda, Breast Center Nie- Results from the prospective multicenter AXSANA/EUBREAST-03/AGO-B-
derrhein, Moenchengladbach, Germany and Department of Gynecology, University 053 study (NCT04373655). First Author: Maggie Banys-Paluchowski, Department of
Medical Center Hamburg, Hamburg, Germany Obstetrics and Gynecology, University Hospital of Schleswig Holstein, Campus Lübeck,
Background: Current treatment de-escalation strategies in HER2+ early breast cancer (eBC) Lübeck, Germany
aim to mitigate acute and late toxicities by reducing or entirely omitting systemic chemotherapy Background: Surgical axillary staging in patients with node-positive (cN+) breast cancer
(sCTx). We analyzed the outcomes and investigated predictors of survival in three randomized scheduled for neoadjuvant chemotherapy (NACT) varies significantly, and includes axillary lymph
de-escalation trials investigating short (12-week) neoadjuvant treatments (NAT) with and node dissection (ALND), sentinel lymph node biopsy (SLNB), targeted axillary dissection (TAD),
without sCTx (paclitaxel, pac) in HER2+ eBC. Methods: In total, 713 patients (pts) were an- and target lymph node biopsy (TLNB). SLNB/TAD/TLNB aim at reducing surgical morbidity
alyzed. WSG-ADAPT-HR-/HER2+ (NCT01817452) compared trastuzumab and pertuzumab (T + without loosing staging accuracy. Comparative data on marking techniques for TAD/TLNB are
P, n=92) vs. T +P + pac (n=42); WSG-ADAPT-HR+/HER2+ (NCT01779206) compared trastu- limited. Here, different marking techniques from the largest available international prospective
zumab emtansine (T-DM1, n=118) vs. T-DM1 + standard endocrine therapy (ET, n=125) vs. T + ET cohort are critically evaluated. Methods: AXSANA is an ongoing cohort study investigating
(n=129); WSG-TP-II (NCT03272477) compared neoadjuvant/adjuvant T + P + ET (n=100) vs. T + oncological and patient-reported outcomes after different axillary procedures in cN+ breast
P + pac (n=107). Omission of further sCTx was allowed in pts with pathological complete cancer treated with NACT. In the present analysis, the subgroup of patients receiving marking of
response (pCR, ypT0/is ypN0); sCTx was mandatory for non-pCR pts. pCR was the primary their TLN is selected, and detection and removal rates are analyzed. The entire dataset is
endpoint of each trial; survival was a secondary endpoint. Kaplan-Meier method and Cox re- continuously and systematically monitored for data quality assurance. Results: 6,129 patients
gression were applied for survival analysis. Results: Median follow-up of 60.7 months was from 291 sites in 26 countries were included between June 2020 and January 6th, 2025. Of these,
available for 713 pts (sCTx: n=149; sCTx-free NAT: n=564). 395 tumors (55%) were cT2-4, 414 2,596 had $ 1 TLN marked before NACT and had completed surgery at time of analysis. The
(58%) were grade 3, and 223 pts (31%) were clinically node-positive. Ten (7%) and 74 (13%) pts mean number of suspicious nodes at diagnosis was 1.9 ($ 4 in 13.4%). 2,484 patients (95.7%)
had iDFS events, 8 (5%) and 51 (9%) had dDFS events, and 6 (4%) and 34 (6%) pts died in the received a minimally invasive biopsy of $ 1 node. TLN marking was performed using a clip in
sCTx and sCTx-free NAT groups, respectively. In the sCTx and sCTx-free NAT groups, the 2,003 patients (77.2%), a magnetic seed in 287 (11.1%), carbon ink in 192 (7.4%), radar marker in
respective 5-year survival rates were 98% (95%CI 93, 99) and 97% (95%CI 95, 98) for OS (HR 0.88; 119 (4.6%), radioactive seed in 18 (0.7%), radiofrequency identification device (RFID) in 12 (0.5%)
95%CI 0.36, 2.11; p=0.775) and 96% (95%CI 91, 98) and 88% (95%CI 85, 91) for iDFS (HR 0.56; or other methods in 2 (0.1%). . 1 type of marker was placed in 36 patients (1.4%). 1 TLN was
95%CI 0.29, 1.08; p=0.083). 95 (66%) and 171 (31%) pts had a pCR after sCTx and sCTx-free NAT, marked in 2,427 patients (93.5%), followed by 2 TLNs in 138 (5.3%) and $ 3 in 27 patients (1%).
respectively. iDFS events occurred in 5 (5%) pts with pCR and 5 (10%) without pCR after sCTx The mean number of marked TLNs was highest if carbon ink was used (mean 1.21), followed by
and in 14 (8%) with pCR and 59 (16%) pts without pCR after sCTx-free NAT. 5-year iDFS rates in clip (1.07), magnetic seed (1.06) and radar marker (1.04); no patient received . 1 radioactive
pts with pCR were 98% (95%CI 91, 99) after sCTx and 94% (95%CI 89, 97) after sCTx-free NAT seed/RFID. 1,895 patients (73.0%) achieved ycN0 status. Targeted removal of the TLN was
(HR 0.76; 95%CI 0.27, 2.14; p=0.609). In univariate analysis, iDFS was associated with pCR (HR planned in 2,100 patients (80.9%): 2,076 (80.0%) were scheduled for a TAD and 24 (0.9%) for a
0.18; 95%CI 0.04, 0.77) in the sCTx group and with cT (3-4 vs 1: HR 2.54; 95%CI 1.22, 5.28) and TLNB. TLN was detected and removed during TAD/TLNB in 1,915 patients (91.2%). TLN detection
cN stage (cN+ vs cN-: HR 2.27; 95%CI 1.44, 3.58), grade (3 vs 1-2: HR 1.79; 95%CI 0.86, 3.74) and rate was highest in patients whose TLNs were marked with probe-guided techniques (96.6%;
pCR (HR 0.47; 95%CI 0.26; 0.84) in the sCTx-free NAT group. Detailed subgroup analyses radioactive seed: 100%, magnetic seed: 96.9%, radar marker: 96.1%, RFID: 90%), followed by
including the impact of standard chemotherapy on outcome will be presented at the meeting. carbon (94.9%) and clip (89.6%; p , 0.001). TAD/TLNB removed a median number of 3 nodes
Conclusions: This pooled analysis demonstrates that de-escalation trials in HER2+ eBC are (mean 4.1, SD 2.77; carbon: median 4, mean 4.29, SD 3.52, probe-guided: median 3, mean 3.82,
feasible and safe for patients. 123 weekly paclitaxel + HER2 blockade is an effective and well- SD 2.63, clip: median 3, mean 4.15, SD 2.75). Conclusions: This large prospective analysis of
tolerated regimen with excellent 5-year survival. The favorable survival after pCR to sCTx-free patients with initially cN+ breast cancer receiving NACT demonstrates that probe-guided markers
NAT lays the groundwork for further de-escalation strategies, such as the currently ongoing provide superior TLN detection rates. Clinical trial information: NCT04373655. Research
WSG-ADAPT-HER2-IV evaluating T-DXd as NAT. Clinical trial information: NCT01817452, Sponsor: AGO-B; AWOgyn; Claudia von Schilling Foundation for Breast Cancer Research; Ehmann
NCT01779206, NCT03272477. Research Sponsor: None. Foundation Savognin; EndoMag; Merit Medical; Mammotome.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
6s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

504 Oral Abstract Session 505 Oral Abstract Session

Predicting nodal burden after neoadjuvant chemotherapy (NAC) with cir- 15-year outcomes for women with premenopausal hormone receptor-
culating tumor (ct)DNA for surgical planning: Results from the I-SPY2 trial. positive early breast cancer (BC) in the SOFT and TEXT trials assessing
First Author: Rita Mukhtar, Division of Surgical Oncology, Department of Surgery, benefits from adjuvant exemestane (E) + ovarian function suppression
University of California, San Francisco, San Francisco, CA (OFS) or tamoxifen (T)+OFS. First Author: Prudence A. Francis, Peter MacCal-
Background: Axillary surgery in breast cancer is used for staging and therapeutic purposes, lum Cancer Centre, Melbourne, Australia
but axillary lymph node dissection (ALND) confers high risk of complications including Background: Long-term follow-up of the SOFT and TEXT randomized trials has shown persistent
lymphedema. Accordingly, trials have focused on right-sizing axillary surgery, with options reduction of recurrence from inclusion of OFS in adjuvant endocrine therapy, and clinically meaningful
ranging from complete omission of surgery, sentinel lymph node (SLN) surgery, targeted improvement in overall survival (OS) among patients at higher baseline risk of recurrence. We report a
axillary dissection (TAD), or ALND. Identifying a biomarker to reliably predict nodal burden final update after a median follow-up of 15y in SOFT and 16.6y in TEXT. Methods: SOFT and TEXT
would facilitate accurate selection of surgical management. We evaluated whether the enrolled premenopausal women with HR+ early BC from November 2003 to April 2011 (2660 in TEXT,
presence or absence of ctDNA in the blood pre and post NAC is associated with residual nodal 3047 in SOFT intention-to-treat populations). TEXT randomized women within 12 weeks of surgery to 5y
E+OFS vs T+OFS; chemotherapy (CT) was optional and concurrent with OFS. SOFT randomized women
burden. Methods: I-SPY2 is a prospective, multicenter NAC trial for patients with clinical stage
to 5y E+OFS vs T+OFS vs T alone, within 12 weeks of surgery if no CT planned, or within 8 months of
II-III high-risk breast cancer. Patients are randomized to novel NAC agents, with pathologic completing (neo)adjuvant CT. Both trials were stratified by CT use. The primary endpoint was disease
complete response being the primary endpoint. As part of the trial, serial ctDNA is assessed free survival (DFS) which included invasive local, regional, distant and contralateral breast events,
with a highly sensitive tumor-informed assay using up to 16 patient-specific tumor mutations second non-breast malignancies and deaths. Secondary endpoints included invasive breast cancer-free
(Signatera) at baseline, 3 weeks, 12 weeks, and post-NAC. We determined whether ctDNA interval (BCFI), distant recurrence free interval (DRFI) and OS. 20y data collection was completed in
positivity or negativity post-NAC, and the change in ctDNA status baseline/post-NAC (-/-, -/+, Q4’2024: 80% of surviving patients had final follow-up during or subsequent to 2020, for 70% it was
+/+, +/-) are associated with ypN category (N0, N1, N2). Results: ctDNA status was available during 2023-2024. 15y Kaplan-Meier estimates and hazard ratios (HR) with 95% CIs are reported.
post-NAC in 495 patients and change in ctDNA status from baseline was available in 493. At Results: There were 815 DFS events and 388 deaths reported in SOFT; and 669 DFS events and 325
baseline, ctDNA was detected in 160/220 (72.3%) cN0 patients and 227/273 (83.2%) cN+ deaths in TEXT. In SOFT, a moderate DFS benefit of T+OFS vs T (HR 0.85; 0.72-1.00) persisted, however
patients (p=0.006). Post-NAC, ctDNA was detected in 11/220 (5%) cN0 patients and 34/273 1/6 DFS events were not BC related; BCFI benefit was HR 0.82 (0.69-0.98). E+OFS vs T further reduced
(12.5%) cN+ patients (p=0.004). While baseline ctDNA status was not associated with ypN DFS events (HR 0.73; 0.61-0.86). The 15y DFS in SOFT was 67.0% for T, 70.5% for T+OFS and 73.5% for
category, there was a significant association between post-NAC ctDNA status and ypN E+OFS. There were consistent but non-significant decreased hazards of death for T+OFS vs T (HR 0.87;
0.68-1.10) and E+OFS vs T (HR 0.85; 0.67-1.08). 15y OS was 85.3%, 86.7%, 86.9% respectively. For the
category. For ctDNA + patients post-NAC, 33.3% were ypN0, 31.1% were ypN1, and 35.6% were
TEXT+SOFT combined analysis of E+OFS vs T+OFS (n=2346 vs 2344) DFS, BCFI and DRFI continued as
ypN2 at surgery; in contrast, for ctDNA - patients post-NAC, 67.1% were ypN0, 23.1% were significantly improved for E+OFS over T+OFS. 15y DFS was 74.9% vs 71.3% (HR 0.82; 0.73-0.92). 15y OS
ypN1, and 9.8% were ypN2 (p,0.0001). Dynamic ctDNA changes were also associated with was 87.8% vs 87.0% (HR 0.94; 0.80-1.11) respectively. 15y estimates by CT use are tabulated.
ypN category, with significantly more ypN0 patients among ctDNA -/- or ctDNA +/- cases, and Conclusions: The high level 15y final results of the SOFT and TEXT confirm a role for OFS- and
more ypN2 patients in those who did not clear ctDNA (+/+) (p=0.0001, Table). Conclusions: To aromatase inhibitor-containing adjuvant endocrine therapy for premenopausal women. Analysis is
our knowledge this is the first study to demonstrate a significant relationship between ctDNA ongoing. Clinical trial information: NCT00066690 (SOFT) and NCT00066703 (TEXT). Research Sponsor:
and ypN category, which has important surgical implications. CtDNA negativity was associated ETOP IBCSG Partners Foundation, BCRF, US NCI, Pfizer, Ipsen, et al have supported long-termfollow-up
with low likelihood of ypN2 disease, making these patients excellent candidates for SLN of the trials.
surgery or TAD. Ongoing analyses will incorporate receptor subtype and timing of clearance of
15y (%) Events SOFT Prior CT (n=1628) SOFT no CT (n=1419)
ctDNA. CtDNA in breast cancer may help tailor surgical management of the axilla, potentially
reducing patient morbidity without compromising prognostic information. Research Sponsor: CT+noCT T / T+OFS / E+OFS T / T+OFS / E+OFS
None. DFS 536+279 60.9 / 63.0 / 66.3 73.9 / 79.1 / 82.1
DRFI 367+56 73.5 / 73.8 / 77.6 94.7 / 94.7 / 96.8
ctDNA baseline/post-NAC ypN0 (n=317) ypN1 (n=116) ypN2 (n=60) OS 318+70 77.4 / 79.4 / 79.8 94.4 / 95.1 / 95.2
TEXT CT (n=1607) TEXT no CT (n=1053)
ctDNA -/- (n=105) 72.4% 19.1% 8.6% T+OFS / E+OFS T+OFS / E+OFS
ctDNA -/+ (n=1) 0.0% 100.0% 0.0% DFS 456+213 68.5 / 72.1 76.8 / 81.8
ctDNA +/+ (n=44) 34.1% 29.6% 36.4% DRFI 286+69 79.0 / 81.3 91.6 / 94.6
ctDNA +/- (n=343) 65.9% 23.9% 10.2% OS 266+59 82.7 / 84.3 94.1 / 94.8

506 Oral Abstract Session 507 Oral Abstract Session

Updated survival outcomes and predictors of benefit from ovarian function The impact of ovarian function suppression with adjuvant endocrine therapy
suppression in premenopausal women with hormone-receptor–positive on survival outcomes in young germline BRCA mutation carriers with breast
breast cancer: Results from the ASTRRA trial. First Author: Jai Min Ryu, cancer: Secondary analysis of an international cohort study. First Author:
Division of Breast Surgery, Department of Surgery, Samsung Medical Center, Sung- Paola Zagami, UNC Lineberger Comprehensive Cancer Center/University of Milan/
kyunkwan University School of Medicine, Seoul, South Korea European institute of Oncology IRCSS (IEO), Chapel Hill, NC
Background: The ASTRRA trial previously demonstrated that adding ovarian function Background: In young women with hormone receptor-positive (HR+) breast cancer (BC),
suppression (OFS) to tamoxifen (TAM) showed consistent disease free survival (DFS) benefit ovarian function suppression (OFS) has been shown to improve outcomes when combined
at 8-yr follow-up analysis in premenopausal women with hormone-receptor (HR)–positive with adjuvant endocrine therapy (ET). However, limited evidence exists on its efficacy in
breast cancer who remain premenopausal or resume menstruation after chemotherapy. Here, germline BRCA (gBRCA) carriers. Here we investigated the association between OFS plus
we aimed to update the survival outcomes and identify patients most likely to benefit from ET and outcomes in the largest global cohort of young gBRCA carriers with BC.
OFS to tailor clinical decision-making. Methods: A total of 1,282 premenopausal women were Methods: The BRCA BCY Collaboration (NCT03673306) is an international, multicenter,
randomized 1:1 to receive either 5 years of TAM alone (TAM-only) or 5 years of TAM with OFS hospital-based, retrospective cohort study of women harboring germline BRCA1/2
for 2 years (TAM + OFS). The primary endpoint was DFS, and the secondary endpoint was pathogenic/likely pathogenic variants, diagnosed between 2000 and 2020 with stage I-III
overall survival (OS). For the HER2-negative cohort, a composite risk score (range: 0–5) for invasive BC at age of # 40 years. The analysis included patients with HR+ BC and available
breast cancer-free interval (BCFI) was calculated based on tumor size, nodal status, and data on ET and OFS. The OFS group included patients treated with luteinizing hormone-
tumor grade using a Cox regression model. The impact of OFS was analyzed by composite risk releasing hormone agonists (LHRHa) and/or bilateral risk-reducing salpingo-oophorectomy
score and stratified by age. The events for BCFI were defined as local, regional, or distant (RRSO) within 1 year of BC diagnosis. Outcome analyses included disease-free survival
recurrence; invasive contralateral breast cancer; or death resulting from breast cancer as the (DFS), BC-free interval (BCFI) and overall survival (OS). Cox proportional hazard models,
first event. Results: With a median follow-up of 117.6 months, the 10-year DFS rate was stratified for country, year of diagnosis, nodal status, and surgery type and adjusted for
83.7% in the TAM + OFS group compared to 75.9% in the TAM-only group (hazard ratio [HR], RRSO and bilateral risk-reducing mastectomy (time-dependent), were used to explore the
0.68; 95% CI, 0.53-0.87). Meanwhile, there were no significant differences in 10-year OS: association between OFS use (vs non-use) and outcomes. Sensitivity analysis explored OFS
94.6% in the TAM + OFS vs. 93.2% in the TAM-only group (HR, 0.79; 95% CI, 0.50-1.27). In the as time-dependent covariate. To address immortal time bias, an additional Cox model
776 patients with HER2-negative breast cancer, there were no significant differences in the accounted for left truncation, considering differences in time to BRCA testing.
distribution of age group (P = .320), tumor size (P = .572), lymph node status (P = .577), or Results: Among 5,660 patients from 109 centers, 1,865 patients with HR+ BC were in-
histologic grade (P = .249) between TAM + OFS and TAM-only groups. Worse 10-year BCFI cluded, of whom 1,071 (57%) received OFS plus ET (35% with an aromatase inhibitor [AI],
was significantly associated with younger age ( , 40 vs. 40-45 years, P = .026), larger tumor 65% with tamoxifen [tam])and 794 (43%) received tam alone. Patients receiving OFS were
size ($ 2cm vs. , 2cm, P , .001), lymph node positivity (positive vs. negative, P , .001), and more likely to have node-positive disease (56% vs 47%), receive treatment in recent years
aggressive histologic grade (III vs. II vs. I, P = .006), respectively. Among patients with a high (36% vs 17%), undergo mastectomy (70% vs 57%) and be tested for gBRCA at diagnosis
composite risk score (4–5, n = 282, 36.3% of the HER2-negative cohort), the 10-year BCFI was
(46% vs 30%). With a median follow-up of 7.8 years (IQR 4.6-12.1), OFS combined with ET
significantly improved with OFS: 76.6% in the TAM + OFS group vs. 65.7% in the TAM-only
was associated with significantly improved DFS (adjusted HR [aHR] 0.79, 95% CI 0.66-0.94),
group (HR, 0.62; 95% CI, 0.40–0.98). This benefit was particularly pronounced in patients
BCFI (aHR 0.74, 95% CI 0.61-0.89) and OS (aHR 0.66, 95% CI 0.50-0.88) over tam alone.
aged 40–45 years. Conclusions: We demonstrated the consistent benefit of adding OFS for 2
Sensitivity analysis using OFS as a time-dependent factor yielded consistent results. No
years to TAM in improving 10-year DFS. In patients with HR-positive/HER2-negative breast
significant interactions were observed between OFS use and specific gBRCA mutations or
cancer and a high composite risk score, the addition of TAM plus OFS resulted in a 10.9%
HER2 status. Sub-analyses by type of ET (OFS + AI vs. OFS + tam vs. tam alone) will be
improvement in the 10-year BCFI, suggesting this approach may be beneficial, especially for
presented at the conference. Conclusions: In this global cohort of young BRCA mutation
those aged 40-45 years. Clinical trial information: NCT00912548. Research Sponsor: National
Research & Development Program for Cancer Control through the National Cancer Center carriers, OFS combined with ET was associated with improved DFS, BCFI and OS versus tam
funded by the Ministry of Health & Welfare, Republic of Korea; No. HA23C014400. without OFS. These findings support the consideration of OFS as a key component of
adjuvant therapy in this population. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 7s

508 Oral Abstract Session LBA509 Rapid Oral Abstract Session

Efficacy and safety of elinzanetant for vasomotor symptoms associated NRG-BR003: A randomized phase III trial comparing doxorubicin plus cy-
with adjuvant endocrine therapy: Phase 3 OASIS 4 trial. First Author: Fatima clophosphamide followed by weekly paclitaxel with or without carboplatin
Cardoso, ABC Global Alliance, Lisbon, Portugal for node-positive or high-risk node-negative TNBC. First Author: Vicente Valero,
Background: Vasomotor symptoms (VMS) associated with adjuvant endocrine therapy (AET) The University of Texas MD Anderson Cancer Center, Houston, TX
impact quality of life and decrease treatment adherence, worsening breast cancer outcomes. There
are few effective treatment options and none approved for this indication. Methods: The 52-week
randomized phase 3 trial OASIS 4 (NCT05587296) evaluated the safety and efficacy of elinzanetant
(EZN), a dual neurokinin-1 and -3 receptor antagonist, in women aged 18–70 years being treated for,
or at high risk of developing, hormone receptor-positive (HR+) breast cancer and experiencing $35
moderate-to-severe VMS/week associated with AET. Women were randomized 2:1 to receive once-
daily EZN 120 mg for 52 weeks or placebo (P) for 12 weeks followed by EZN for 40 weeks. Primary
endpoints were mean change in moderate-to-severe VMS frequency from baseline to weeks 4 and 12
analyzed using mixed model with repeated measures (one-sided p-values). Secondary endpoints
were mean changes from baseline in moderate-to-severe VMS frequency to week 1 and moderate-to-
severe VMS severity to weeks 4 and 12. Treatment-emergent adverse events (TEAEs) were reported
throughout the study. Results: Mean (standard deviation [SD]) baseline daily VMS frequency was
11.4 (6.9) in the EZN group (n=316) and 11.5 (6.4) in the P group (n=157). Reductions from baseline
in VMS frequency were observed from week 1 (EZN: -4.0 [5.1]; P: -1.8 [3.8]). At week 4, mean (SD)
VMS frequency reduced by -6.5 (6.1) with EZN and -3.0 (5.0) with P, with statistical significance The full, final text of this abstract will be available at
between EZN and P (least squares [LS] mean difference [95% confidence interval (CI)]: -3.5 [-4.4,
-2.6]; p,0.0001). At week 12, reductions in VMS frequency were -7.8 (6.2) with EZN and -4.2 (6.1)
[Link] on the day of presentation and in the
with P, with statistical significance between EZN and P (LS mean difference [95% CI]: -3.4 [-4.2, -2.5]; online supplement to the June 10, 2025, issue of the Journal
p,0.0001). Reductions in VMS severity were greater with EZN vs. P (week 4: -0.7 [0.6]; -0.4 [0.4], of Clinical Oncology.
week 12: -1.0 [0.7]; -0.5 [0.6]). During the placebo-controlled period, 220 (69.8%) and 98 (62.0%)
patients reported TEAEs in the EZN and P groups, respectively. Somnolence, fatigue, and diarrhea
were more frequently reported with EZN (Table). Fewer TEAEs were reported in both groups during
weeks 13–52. Conclusions: EZN was efficacious with a fast onset and well tolerated for the
treatment of VMS associated with AET. TEAE frequency was as expected for this type of trial.
Adequate VMS management can improve adherence to AET and, therefore, improve cancer out-
comes and quality of life. Clinical trial information: NCT05587296. Research Sponsor: Bayer.
EZN P Total EZN
Week 1–12 Week 1–12 Week 1–52
n (%) n=315 n=158 N=465
Any TEAE 220 (69.8%) 98 (62.0%) 368 (79.1%)
Headache 30 (9.5%) 20 (12.7%) 56 (12.0%)
Arthralgia 20 (6.3%) 10 (6.3%) 52 (11.2%)
Fatigue 30 (9.5%) 8 (5.1%) 43 (9.2%)
Somnolence 34 (10.8%) 6 (3.8%) 42 (9.0%)
Diarrhea 16 (5.1%) 3 (1.9%) 32 (6.9%)
Back pain 10 (3.2%) 7 (4.4%) 29 (6.2%)
Nausea 19 (6.0%) 10 (6.3%) 29 (6.2%)
Any serious TEAE 8 (2.5%) 1 (0.6%) 33 (7.1%)

510 Rapid Oral Abstract Session 511 Rapid Oral Abstract Session
Prospective randomized phase II trial to assess the efficacy and safety of A phase 2 study of response-guided neoadjuvant sacituzumab govitecan
neo-adjuvant olaparib/carboplatin (OC) in comparison to docetaxel/ and pembrolizumab (SG/P) in patients with early-stage triple-negative
epirubicin/cyclophosphamide (TAC) in patients with early triple-negative breast cancer: Results from the NeoSTAR trial. First Author:
breast cancer (TNBC) with homologous recombination deficiency (HRD): Rachel Occhiogrosso Abelman, Massachusetts General Hospital Cancer Center, Harvard
Primary results from the ABCSG 45 trial. First Author: Christian F. Singer, Medical School, Boston, MA
Department of Obstetrics and Gynecology and Center for Breast Health, Comprehensive Background: Sacituzumab govitecan (SG) is a TROP-2 directed antibody-drug conjugate
Cancer Center, Medical University of Vienna and Austrian Breast and Colorectal Cancer (ADC) approved for metastatic triple negative breast cancer (TNBC). Pembrolizumab (P), an
Study Group, Vienna, Austria anti-programmed death 1 monoclonal antibody, is approved for early-stage TNBC and
Background: Carboplatin-based regimen are effective in patients (pts) with eTNBC, and olaparib metastatic PD-L1 positive TNBC. However, safety and efficacy of SG+P in early TNBC is not
improves the outcome of pts with BRCA1/2 pathogenic variants (PV), but the safety / efficacy of known. We published results of Arm A1 investigating neoadjuvant SG monotherapy in early
OC co-treatment in HRD-positive TNBC is unknown. ABCSG 45 (EU CT 2024-512821-10) is a TNBC (Spring et al. Annals of Onc 2024). Here we present results from Arm A2 of the NeoSTAR
prospective multicenter phase II study investigating the efficacy and tolerability of OC compared study investigating the combination of neoadjuvant SG + P in early-stage TNBC
to conventional chemotherapy in HRD-positive eTNBC. Methods: Pts with HRD (Myriad (NCT04230109). Methods: Patients (pts) with early TNBC (tumor size $2 cm or node
genetics)-positive eTNBC were randomized to 6 cycles of olaparib (100 mg bid, days 4-16) / positive) with no prior treatment were eligible. Pts received SG at starting dose of 10mg/kg on
carboplatin (AUC 5) q3w, or 6 cycles of docetaxel/epirubicin/cyclophosphamide (75/50/500) days 1,8 of a 21-day cycle for 4 cycles with P 200 mg given on day 1 of each cycle. After trial
q3w (TAC). In an initial dose-finding phase, 100 mg bid was identified as olaparib combination regimen, pts underwent imaging to determine residual radiographic disease per RECIST v1.1.
dose. Stratification factors were tumoral BRCA1/2 and menopausal status. Primary endpoint A biopsy was performed if residual disease (RD) was suspected. Additional neoadjuvant
was centrally assessed residual cancer burden (RCB), pCR and QoL were secondary endpoints. chemotherapy (ANACT) was at discretion of the treating physician prior to definitive surgery.
Planned sample size was 90 pts, randomized 1:1 to achieve 80% power (two-sided alpha=0.05) The primary objective was rate of pathologic complete response (pCR) with neoadjuvant SG/
to detect a RCB 0/I difference of 31%. Differences between treatment arms were assessed with a P. Secondary objectives included need for ANACT, radiographic response (RR), safety and
two-sided Cochran Mantel-Haenszel test using stratification factors. Pre-defined subgroup tolerability (adverse events [AEs] per CTCAE v5.0) and event-free survival. A Simon two-stage
analysis was performed with logistic regression. Results: A total of 90 pts (OC: n=46; TAC: design and standard descriptive statistics were utilized, including 95% binomial confidence
n=44), of whom 42 (47%) were BRCA1/2 PV carriers, were randomized between November 2019 intervals for all rates estimated. Results: From 5/19/23-8/13/24, 50 pts were enrolled
and December 2023. Median age was 50.5 years (range 27.0-80.0). 40% had cT1, 55.6% cT2, and (median age: 57 years, range 23-77). Clinical anatomic stage was II in 48 pts (96%) and III in 2
4.4% cT3/4 tumors, and 60% of pts were clinically N0. 94.4% of tumors were G3, and Ki67 pts (4%). 64% of pts were node negative at diagnosis. 44 pts (88%) completed the trial regimen
was .60% in 71.1%. Overall, the RCB0/I rate with OC was 52.2% vs. 70.5% with TAC (stratified (5 pts had toxicity, 1 pt progressed on treatment). In interim analysis, 5/15 had pCR and so the
risk difference = -18.8% (95%CI: -39.6% to 2.0%); p=0.068). In pts with BRCA1/2 PV, RCB0/I rates remaining 35 were enrolled. The pCR rate per protocol (pts with pCR at surgery directly after
were comparable: 77.3% (OC) vs. 65.0% (TAC), while in 47 pts with BRCA1/2 wild type (WT), OC SG/P without ANCT) was 16/50 (34%, 95% CI 19.5-46.7). The RR rate (complete CR or partial
was significantly less effective: RCB0/I of 29.2% vs 73.9% in TAC (interaction p=0.008). pCR was response PR) was 66% (95% CI 50-78%), 30% CR and 36% PR. Of 26 pts who received ANACT, 9
achieved in 47.8% (OC) vs 59.1% (TAC; p=0.231). In pts with a BRCA1/2 PV, OC resulted in 77.3% experienced pCR (2 biopsy-confirmed RD, 6 negative or non-diagnostic RD biopsy, 1 no
pCR rate, vs. TAC 65.0%, in BRCA1/2 WT pts pCR was achieved by 20.8% (OC) vs. 56.5% (TAC) biopsy). Overall, 25 (50%, 95% CI 35.5-64.5) pts had pCR at surgery. Of 5 pts with pathogenic
(interaction p=0.021). OC treatment resulted in more $ grade 3 hematologic toxicities with 30% BRCA mutations, 3 (60%) had pCR after SG/P, and 1 pt had pCR after ANACT. 20 pts (40%) had
vs 3% thrombocytopenia and 43% vs 18% neutropenia but caused fewer non-hematological grade 3 or higher AEs. The most common AEs were nausea (28, 56%), alopecia (26, 52%),
toxicities. Conclusions: In this prospective randomized study in HRD-positive TNBC, 6 cycles of fatigue (23, 46%), and diarrhea (22, 44%). Dose reductions of SG occurred in 4 pts (8%).
TAC resulted in strikingly high RCB0/I and pCR rates, independent of BRCA1/2 status. 6 cycles of Updated survival and biomarker data will be presented at the meeting. Conclusions: In the
OC achieved a pCR rate of .77% in BRCA1/2 PV but were less effective in pts with BRCA1/2 WT first trial to investigate the SG/P combination in early TNBC, 34% of pts had pCR. Additional
disease. These results may help to optimize neoadjuvant treatment strategies in TNBC. This research is needed to determine the optimal duration and sequence of neoadjuvant SG/P and
research was conducted with support from AstraZeneca Austria GmbH. Clinical trial infor- chemotherapy for pts with TNBC. Clinical trial information: NCT04230109. Research Sponsor:
mation: 2024-512821-10. Research Sponsor: None. None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
8s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

512 Rapid Oral Abstract Session 513 Rapid Oral Abstract Session
The Promise study: A presurgical randomized clinical trial of CE/BZA vs The WinPro trial: A window of opportunity study of endocrine therapy with
placebo in postmenopausal women with ductal carcinoma in situ. First and without prometrium in postmenopausal women with early stage hor-
Author: Swati Kulkarni, Northwestern University Feinberg School of Medicine, Chicago, mone receptor-positive breast cancer. First Author: Lucy Haggstrom, St Vincent’s
IL Hospital Sydney and University of New South Wales, Sydney, NSW, Australia
Background: Conjugated estrogen/bazedoxifene (CE/BZA), the first tissue selective es- Background: Preclinical evidence has shown that progesterone is a tumour suppressor in estrogen
trogen complex, was developed as an alternative to combination estrogen and progesterone receptor positive (ER+) breast cancer. Prometrium is bioidentical to human progesterone, currently
therapy to treat hot flashes and osteoporosis. Preclinical studies found that CE/BZA reduced used for treating menopausal symptoms. Methods: The WinPro trial is a randomised, multi-centre,
mammary ductal proliferation and increased expression of anti-tumorigenic markers in phase 2, window of opportunity trial of preoperative endocrine therapy in post-menopausal women
with early-stage, ER+, progesterone receptor (PR) +, HER2- breast cancer. Patients (pts) were
breast stroma. Our study aimed to determine if a pre-surgical window-of-opportunity in-
randomised [Link] to letrozole (let) 2.5mg daily, letrozole 2.5mg daily + prometrium (pro) 300mg daily,
tervention with CE/BZA in women with ductal carcinoma in situ (DCIS) had a protective or tamoxifen (tam) 20mg daily + prometrium 300mg daily for 11-17 days before surgery. The primary
effect on the duct epithelium and stroma of DCIS lesions without impacting quality of life. endpoint was the percent proportional reduction in Ki67 between biopsy and surgery (‘Ki67 sup-
Differences between CE/BZA and placebo arms for the primary endpoint, change in Ki-67 pression’) in let vs let + pro in the per protocol population. Other endpoints were safety, changes in ER,
protein expression, and quality-of-life endpoints are reported here. Methods: This multi- PR and androgen receptor (AR) via immunohistochemistry (IHC) and H-scores, spatial transcriptomics
center, randomized, double-blind placebo-controlled Phase 2 trial was conducted between and RNA sequencing. Results: From Feb 2018 to June 2024,244 pts were enrolled across 6 Australian
9/19/17 and 8/21/24. Postmenopausal women with estrogen receptor positive (ER+) DCIS sites. 239 pts were randomised to let (n=78, 32.6%), let + pro (n=79, 33.1%), and tam + pro (n=82,
undergoing surgery were randomized to CE 0.45 mg /BZA 20 mg or placebo for 28 +/-7 days. 34.3%). 189 pts completed per protocol: let (n=66, 34.9%, let + pro (n=64, 33.9%), and tam + pro (n=59,
Percentage of nuclei staining for Ki-67 was evaluated on slides from the baseline core 31.2%). Baseline characteristics were well balanced across arms. There was no significant difference
biopsy and surgical specimen. Changes were compared between arms using the two-sample in Ki67 suppression between let (88.2%) vs let + pro (89.2%) (p = 0.4). Ki67 suppression appeared
t-test, while changes within arms were analyzed using paired t-test. Analyses were done on inferior with tam + pro (61.5%). Treatments were well tolerated, with hot flushes less frequent in let +
pro (13.3%) vs let (22.4%) or tam + pro (20.5%). IHC analyses showed no change in ER% after
log2 scale to satisfy the normality assumption. The Breast Cancer Prevention Trial Eight
treatment, a decrease in PR% after treatment with let and let + pro, and a decrease in AR% after
Symptom Scale (BESS) and Menopause-Specific Quality of Life (MENQOL) surveys were treatment in all groups. H-score, spatial transcriptomics, RNA sequencing and PAM50 analyses are
self-administered by patients before and after the intervention. Wilcoxon’s signed-rank and underway. Conclusions: The WinPro trial showed that the addition of pro to let in post-menopausal
rank-sum tests were used to perform within- and between-arm comparisons, respectively. women with ER+, PR+, HER2- breast cancer was safe, reduced hot flushes and led to similar reduction
Results: Of the 171 patients consented,141 enrolled, and 117 completed the study. Ninety- in Ki67 as let alone. Ongoing translational analyses are underway which will examine the changes in
four patients (46= CE/BZA, 48=placebo) took .80% of the medication and had Ki-67 gene expression in malignant, immune and stromal cells at sub-cellular resolution and provide further
evaluated at baseline and post-intervention. The BESS and MENQOL surveys were com- insight into the mechanisms of response and resistance to endocrine therapy. Clinical trial infor-
pleted by 100 and 108 patients, and 125 patients were evaluated for toxicity. The mean mation: ACTRN12618000928213. Research Sponsor: Cancer Council New South Wales; National
absolute change in Ki-67 was -5.62 (SD=10.2; p=0.003) in the CE/BZA arm and -1.07 Health and Medical Research Council, Australian Government; Besins Healthcare.
(SD=10.8; p=0.6) in the placebo arm, with a greater reduction in CE/BZA arm (p=0.016). Baseline Surgery
There was no difference between arms in the BESS score across all 8 domains or in the
Let Let + pro Tam + pro Let Let + pro Tam + pro
MENQOL score. However, within each arm, vasomotor symptoms decreased in the CE/BZA N = 67 N = 64 N = 60 N = 67 N = 64 N = 60
arm (p=0.002) but not in the placebo arm (p=0.4). No grade . 3 treatment related adverse Ki67% suppression, - - - 88.2 89.2 61.5
events were reported. Conclusions: In this prospective randomized clinical trial, CE/BZA median (Q1, Q3) (74.6, 93.0) (78.3, 93.9) (32.3, 75.9)
significantly reduced epithelial proliferation in ER+ DCIS with no impact on quality of life Ki67%, median (Q1, Q3) 8.5 10.5 10.8 1.0 1.0 3.5
(4.5, 16.0) (5.0, 16.3) (6.0, 17.0) (0.5, 2.5) (0.5, 2.0) (2.0, 7.5)
compared to placebo. These results support consideration that CE/BZA is a safe option to ER%, median (Q1, Q3) 95 95 95 95 95 95
manage menopausal symptoms for women concerned about their risk of developing breast (95, 95) (95, 95) (95, 95) (95, 95) (95, 95) (95, 95)
cancer, and provide supportive evidence that CE/BZA may reduce the risk of developing PR%, median (Q1, Q3) 90 90 90 20 20 90
(50, 95) (35, 95) (50, 95) (4, 70) (2, 70) (33, 95)
invasive breast cancer. Clinical trial information: NCT02694809. Research Sponsor: U.S. AR%, median (Q1, Q3) 40 28 70 15 10 10
National Institutes of Health; 1R01CA218436-01. (10, 90) (8, 80) (20, 90) (2, 60) (2, 75) (1.3, 45)

514 Rapid Oral Abstract Session 515 Rapid Oral Abstract Session
Early results of the French multicenter, randomized SHARE trial comparing Dalpiciclib (Dalp) plus endocrine therapy (ET) as adjuvant treatment for HR+/
whole breast irradiation versus accelerated partial breast irradiation in HER2– early breast cancer (BC): The randomized, phase 3, DAWNA-A trial.
postmenopausal women with early-stage low risk breast cancer: Analysis First Author: Zhi-Ming Shao, Fudan University Shanghai Cancer Center, Shanghai, China
of toxicity and cosmetic outcomes. First Author: Yazid Belkacemi, AP-HP, De- Background: Dalp, a potent CDK4/6 inhibitor, has demonstrated significant improvements in PFS
partment of Radiation Oncology and Henri Mondor Breast Center, Henri Mondor Uni- when combined with ET in both first-line and later-line settings for HR+/HER2– advanced BC. We
versity Hospital, University of Paris Est Creteil (UPEC), Paris, France conducted a randomized, double-blind, phase 3 trial (DAWNA-A) to further evaluate Dalp with ET
Background: Locoregional management of early stage breast cancer (BC) has evolved from as adjuvant therapy for high-risk, early HR+/HER2– BC and here present findings from the pre-
maximal tolerable to minimal effective therapies. Significant advancements in radiation specified first interim analysis (IA1). Methods: Women aged 18-75 y, with stage II-III, HR+/HER2–
therapy (RT), such as limited target volumes and hypofractionation, have led to accelerated BC, who had completed definitive local therapy (surgery and/or radiotherapy) and had patho-
partial breast irradiation (APBI). This study reports on toxicity and cosmetic outcomes of APBI logically confirmed ipsilateral axillary lymph node involvement, were enrolled. Patients (pts) were
in post-menopausal women with unifocal pT1-N0-M0 invasive BC. Methods: The SHARE trial randomized (1:1) to receive oral Dalp (125 mg QD; 3-wk on/1-wk off, for 2 y) + ET (letrozole 2.5
(NCT01247233) is a non-inferiority, multicenter, randomized trial comparing local control of mg/anastrozole 1 mg/tamoxifen 10 mg/toremifene 60 mg QD, for 5 y) or placebo + ET. Pre/peri-
menopausal pts received LHRH agonists (perimenopausal use at investigator’s discretion).
APBI versus Whole Breast Irradiation (WBI). Eligible patients were postmenopausal women
Stratification factors were menopausal status (pre/peri vs post), clinical stage (II vs III), number of
over 50 years who had lumpectomy with surgical margins . 2mm. Only patients who had at
involved nodes (,4 vs $4), and adjuvant chemo (y vs n). The primary endpoint was invasive
least 4-5 clips placed in the tumor bed during surgery were eligible. Patients were randomized
disease-free survival (iDFS). IA1 was pre-planned at ~254 iDFS events (~50% of total expected).
to receive either WBI (50Gy in 25 fractions (fr) with optional 16Gy-boost or 40Gy in 15 fr, or
As of Oct 25, 2024, 268 iDFS events occurred and IA1 was performed; the actual superiority
42.5Gy in 16fr) or APBI (38.5Gy or 40Gy in 10fr twice daily). Primary endpoint was local
boundary was a 1-sided p ,0.00205 (Lan-DeMets [O’Brien-Fleming] boundary). Results: Between
recurrence. Secondary endpoints included grade .2 toxicity (NCI-CTCAE-v4) and cosmetic Apr. 30, 2021 and Jul. 19, 2024, 5274 pts were randomized (Dalp, n=2640; placebo, n=2634). As of
outcomes (good/excellent versus intermediate/poor) evaluated by both patients and doctors, data cutoff, median follow-up was 20.3 mo (range 0.0–41.9). Dalp + ET significantly prolonged
over the follow-up time. We estimated the cumulative incidences (CI) using the Kalbfleisch- iDFS vs placebo + ET (HR 0.56, 95% CI 0.43–0.71; 1-sided p ,0.0001); iDFS benefits with Dalp
Prentice method due to competing events, and cause-specific Hazard Ratios (cs-HRAPBI/WBI) were generally consistent across stratification factors and other baseline subgroups. DFS and
from Cox models adjusted on stratification factors. Results: Among 1006 patients (503 per distant DFS (DDFS) also favored Dalp + ET over placebo + ET (Table). TRAEs led to discontinuation
arm) enrolled between December-2010 and July-2015, with a median follow-up of 5.8 years, 28 of Dalp in 2.1% of treated pts in Dalp arm and of placebo in 0.8% in placebo arm. Tr-SAE occurred
deaths and 11 local recurrences were reported. The risk of severe toxicity appeared signif- in 3.7% and 1.5%, respectively. There was no death due to TRAEs. Conclusions: Addition of Dalp
icantly reduced in the APBI-arm when considering all type of toxicity (cs-HRAPBI/WBI=0.74, to ET as adjuvant treatment significantly improved iDFS, with a tolerable safety profile. These data
[95%-CI: 0.61-0.89], p=0.001; 3-year CI=45% [41-49] in WBI vs 36% [32-40] in APBI), or only support the use of Dalp for treating HR+/HER2- early BC. Clinical trial information: NCT04842617.
breast skin toxicity (cs-HR=0.55 [0.44-0.70], p,0.001; 3-year CI=36% [32-40] vs 21% [18-25], Research Sponsor: Jiangsu Hengrui Pharmaceuticals Co., Ltd.
respectively). Conversely, for non skin breast toxicities, WBI was less toxic: cs-HRAPBI/WBI=2.06
Efficacy outcomes.
(1.49-2.86), p,0.001). We observed no significant difference of patient-reported cosmetic
results: cs-HRAPBI/WBI=1.08 (0.85-1.37), p=0.54. Findings were similar for doctor-evaluated Dalp + ET (n=2640) Placebo + ET (n=2634)
results. Rib fractures incidence was nearly double in APBI compared to WBI. Conclusions: The iDFS Event, n (%) 98 (3.7) 170 (6.5)
SHARE trial showed that APBI is associated with reduced severe and skin-related toxicities 3-y rate*, % (95% CI) 89.1 (85.8–91.7) 86.2 (83.3–88.6)
HR† (95% CI); p value‡ 0.56 (0.43–0.71); p ,0.0001
compared to WBI, with no significant difference in cosmetic outcomes. Conversely, WBI was DFS Event, n (%) 108 (4.1) 195 (7.4)
less toxic concerning non-skin breast toxicity, mainly breast fibrosis. The question that 3-y rate*, % (95% CI) 88.0 (84.5–90.7) 83.8 (80.5–86.6)
currently remains open on a practical level is how to consider APBI in the context of the HR† (95% CI); p value‡ 0.53 (0.42–0.67); p ,0.0001
widespread adoption of the “Fast Forward” regimen for patients at low risk of recurrence. DDFS Event, n (%) 93 (3.5) 149 (5.7)
3-y rate*, % (95% CI) 90.2 (87.2–92.6) 88.7 (86.2–90.8)
Clinical trial information: 2010-A00243-36. Research Sponsor: French Ministry of Health HR† (95% CI); p value‡ 0.60 (0.46–0.78); p,0.0001
PHRC-2010 Cancérologie; La Ligue contre le cancer.
*Kaplan–Meier method. †Stratified Cox proportional hazards model. ‡Stratified Log-rank test (1-sided).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 9s

516 Rapid Oral Abstract Session 517 Rapid Oral Abstract Session
Efficacy and safety of ribociclib (RIB) + nonsteroidal aromatase inhibitor The TRADE study: A phase 2 trial to assess the tolerability of abemaciclib
(NSAI) in NATALEE: Analysis across menopausal status and age. First Author: dose escalation in early-stage HR+/HER2- breast cancer. First Author: Erica L.
Kevin Kalinsky, Winship Cancer Institute at Emory University, Atlanta, GA Mayer, Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston,
Background: The NATALEE trial demonstrated significant invasive disease–free survival MA
benefit with RIB + NSAI vs NSAI alone in patients (pts) with stage II/III HR+/HER22 early Background: The CDK4/6 inhibitor abemaciclib (abema) is approved with adjuvant en-
breast cancer (EBC) at high risk of recurrence. Here, we report outcomes by menopausal status docrine therapy (ET) for high-risk node positive hormone receptor positive (HR+) HER2-
and age. Methods: Pts were treated with RIB + NSAI or NSAI alone in NATALEE; premen- breast cancer. This regimen reduces cancer recurrence, yet therapy may be complicated by
opausal (PreM) women also received goserelin. Men were excluded from this analysis. Ef- toxicity, limiting patient (pts) ability to maintain dose or continue medication. In the phase
ficacy, safety, and quality of life were analyzed by menopausal status (assessed at III monarchE study, 25.8% of pts discontinued abema early for reasons other than re-
randomization or start of adjuvant endocrine therapy, whichever was first) and age (PreM [,40 currence, 18.5% for adverse events (AEs), and 43.6% required dose reduction. Experiences
y, $40 y]; postmenopausal [PostM; ,60 y, $60 y]). Data cutoff was April 29, 2024. Results: A with other targeted therapies suggest initial dose escalation may reduce toxicity and
greater portion of PreM vs PostM pts had ECOG performance status 0 (86.8% vs 80.1%), Ki- discontinuation. TRADE is a prospective, single-arm, phase 2 study evaluating whether a
67 .20% (39.9% vs 34.4%), N1-N3 nodal stage (63.4% vs 56.9%), and T3/T4 tumors (28.7% vs dose-escalation strategy of adjuvant abema improves drug tolerability. Methods: Eligible
24.0%) at diagnosis. There was consistent treatment benefit with RIB + NSAI vs NSAI alone
pts had node-positive HR+/HER2- breast cancer and were candidates for adjuvant abema
across groups and ages (median follow-up, 44.2 months) (Table). Fewer PreM pts discontinued
with ET. All pts started abema at 50 mg BID for 2 weeks (wks), escalated to 100 mg BID for
RIB due to AEs vs PostM (16.1% vs 22.9%); reductions due to AEs were similar (22.4% vs
2 wks, then escalated to final dose of 150 mg BID onwards. Escalation required absence of
23.6%). Of pts who discontinued due to AEs, more PreM pts did so without a dose reduction vs
PostM (75.4% vs 67.5%). Within menopausal groups, fewer pts in the younger cohorts dis-
ongoing grade 3/4 or persistent grade 2 toxicity; anti-diarrheal medication was used as
continued RIB due to AEs (Table). Alanine aminotransferase elevation was the most common needed. The primary endpoint, measured at 12 wks, was a composite rate of discon-
AE leading to discontinuation in the PreM (6.2%) and PostM groups (8.0%). Time to dete- tinuation of abema for any reason or inability to reach or maintain the 150 mg dose. Based
rioration in global and physical functioning scales of the EORTC QLQ-C30 was similar between on assumptions from monarchE, the experimental hypothesis was that a dose-escalation
treatment arms for all subgroups. Conclusions: RIB + NSAIprovides treatment benefit to a schedule of abema would significantly reduce rate of the composite primary endpoint at 12
broad range of pts with stage II/III HR+/HER22 EBC across menopausal status and age. In wks from 40%. Results: 90 pts enrolled, 89 evaluable for the primary endpoint (1 pro-
younger PreM pts, who typically have more aggressive disease characteristics, treatment gression before 12 wks). Median age was 58 [range 24-78], 4% were Black, 3% were
favored RIB + NSAI, and these pts were least likely to discontinue RIB due to AEs. Clinical trial Hispanic. 48% had stage II disease, 52% had stage III, all received AI, 14% concurrent OFS.
information: NCT03701334. Research Sponsor: None. The study achieved the predefined primary endpoint with 26 pts (29.2%; 90% CI [21.3-38.2];
p=0.046) meeting the composite endpoint at 12 wks: 6 (6.7%) for early discontinuation (3
PreM (n = 2238) PostM (n = 2844)
[3.4%] for toxicity), 8 (9.0%) for inability to reach 150 mg, and 12 (13.5%) for dose reduction
All <40 y ‡40 y All <60 y ‡60 y from 150 mg. The most frequent .grade 2 treatment related AEs by 12 wks were
a
Hazard ratio RIB = 1115 RIB = 237 RIB = 878 RIB = 1424 RIB = 703 RIB = 721
(95% CI) NSAI = 1123 NSAI = 276 NSAI = 847 NSAI = 1420 NSAI = 735 NSAI = 685 neutropenia (23.3%), diarrhea (22.2%), and fatigue (20%). Rates of clinically significant
Invasive disease– 0.671 0.690 0.662 0.746 0.835 0.673 diarrhea (. grade 2) within 0-4, 4-8, and 8-12 wks were 5.6%, 14.6%, 15.3%, in contrast to
free survival (0.518-0.870) (0.419-1.137) (0.488-0.897) (0.607-0.917) (0.619-1.128) (0.506-0.896) rates from monarchE of 20.5%, 12.1%, 7.3% in the same periods. Conclusions: The TRADE
Distant disease–free survival 0.655 0.647 0.659 0.759 0.854 0.681
(0.498-0.861) (0.383-1.091) (0.478-0.908) (0.612-0.941) (0.625-1.168) (0.506-0.916) study is a positive trial meeting its primary endpoint. Use of an adjuvant abema dose
Recurrence–free survival 0.641 0.723 0.610 0.735 0.811 0.668 escalation strategy allowed a greater number of pts (70.8%) to reach and maintain the
(0.486-0.845) (0.429-1.220) (0.439-0.846) (0.588-0.919) (0.590-1.114) (0.487-0.915)
Disposition in RIB
150 mg dose at 12 wks than in monarchE. Early discontinuation was infrequent, and 93.3%
arm, n (%) were continuing therapy at 12 wks. Reduced incidence and severity of clinically important
RIB discontinuation due to AE 179 (16.1) 25 (10.5) 154 (17.5) 326 (22.9) 125 (17.8) 201 (27.9) toxicity such as diarrhea was observed. This dosing strategy could be considered when
RIB reduction due to AE 248 (22.4) 64 (27.0) 184 (21.1) 332 (23.6) 169 (24.2) 163 (22.9)
a
initiating adjuvant abemaciclib. Further follow-up will assess long-term tolerability, dosing
Hazard ratios between treatment arms (RIB + NSAI; NSAI alone), stratified by stage, prior chemotherapy, and geographic maintenance beyond 12 wks, and correlative analyses. Clinical trial information:
region.
NCT06001762. Research Sponsor: Lilly.

518 Poster Session 519 Poster Session

Impact of chemotherapy on financial toxicity in African-American breast Out-of-pocket cost modeling of EUROPA trial arms for adjuvant breast
cancer patients: Early findings from the navigator-assisted hypofractiona- cancer therapy: Five days of radiation versus five years of antiestrogen
tion (NAVAH) phase I clinical trial. First Author: Maya J. Stephens, Medical College therapy. First Author: Ena Chinedum Oboh, The George Washington University School
of Georgia, Augusta, GA of Medicine and Health Sciences, Washington, DC
Background: With the rising cost of chemotherapy, the financial toxicity (FT) of systemic Background: Optimal adjuvant therapy following breast-conserving surgery in adults with
therapy can substantially impair patient quality of life. FT is also associated with various low-risk, early-stage breast cancer remains uncertain. Cost and financial toxicity remain
socioeconomic factors, one being race. Patients of African American race often bear the significant concerns for breast cancer patients. There is limited granular analysis of the
worst burden of cancer treatment-related FT, with a 40% increased mortality from breast role of insurance in the aggregate cost of five fractions of adjuvant radiotherapy (RT)
cancer. The degree to which chemotherapy prior to radiation therapy (RT) impacts FT has versus adjuvant endocrine therapy (ET), as is being explored in the EUROPA Phase III trial.
yet to be formally quantified. We report early FT findings among African American breast This study aims to disaggregate costs, estimate out-of-pocket (OOP) expenses by in-
cancer patients prior to receipt of adjuvant RT on the ongoing Navigator-Assisted Hypo- surance plan, and increase transparency to better inform treatment decisions.
fractionation (NAVAH) Phase I clinical trial ([Link] ID: NCT05978232) to assess Methods: Treatment protocols were aligned with the EUROPA trial arms. For our financial
the impact of chemotherapy on FT. Methods: African-American breast cancer patients model, we used five-fraction RT and ET prescribed over 5 years, with follow-up after two
undergoing RT were eligible if age 18+ with pathologically confirmed breast cancer following years (consistent with EUROPA trial results) and five years (full duration of prescribed
resection. As part of the trial, patients were assisted by a patient navigator during and after antiestrogen therapy). OOP costs, deductibles, and copays/coinsurance were calculated
treatment. FT was measured using the validated 12-item COmprehensive Score for financial using Medicaid, Original Medicare, Medigap Plan G, and Medicare Part D plans. Data
Toxicity-Functional Assessment of Chronic Illness Therapy (COST-FACIT) survey instru- sources included [Link], [Link], [Link], and the CMS
ment. COST-FACIT scoring was used to find FT in patients before receipt of RT. Values from physician fee schedule. Price estimates reflect actual insurance plan costs rather than
26-44 represent Grade 0 FT (none), values from 14-25 represent mild Grade 1 FT (mild), claims data. The model assumes a Medicare- and/or Medicaid-eligible patient aged $ 70
values from 1-13 represent Grade 2 FT (moderate), and values of 0 represent Grade 3 FT years with early-stage estrogen-receptor positive breast cancer after breast-conserving
(severe). The chi-square test was used to identify statistically significant differences surgery. Results: Original Medicare beneficiaries face estimated OOP treatment costs of
(p ,0.05) between patients who received chemotherapy versus no chemotherapy prior to $1,049.06 for adjuvant ET at 24 months and $2,130.25 at five years. For five-fraction RT,
receipt of RT. Results: The first 32 enrolled patients completing the pre-RT COST-FACIT the estimated OOP costs are $1,490.93 at 24 months and $2,320.12 at five years. Medigap
survey were evaluated. 53% of patients underwent chemotherapy before RT. Mild to Plan G beneficiaries incur lower OOP expenses: $682 for adjuvant ET at 24 months and
moderate FT was apparent in 56% of patients. The mean and median COST-FACIT score $1,705 at five years, and $514 at 24 months and $1,285 at five years for five-fraction RT. In
(range 4.4-39) was 25 (+/- 10.4). 78% of patients who experienced some level of FT un- contrast, Medicaid beneficiaries have no OOP expenses for either treatment option, as
derwent chemotherapy and 22% of patients experiencing FT did not receive chemotherapy Medicaid fully covers all treatments. Conclusions: The EUROPA trial highlights HRQOL
(p = 0.0015). Of patients who did not experience FT, 21% received chemotherapy and 79% of outcomes at 24 months for endocrine therapy versus radiotherapy in women aged 70+ with
patients did not. In total, 82% of patients who underwent chemotherapy before RT reported luminal A-like early breast cancer. This cost analysis, based on actual cost estimates,
mild to moderate FT. Grade 3 FT was not observed. Conclusions: The NAVAH study is the provides a clear comparison of OOP expenses across Medicaid and Medicare plans. While
first to objectively compare FT among patients receiving chemotherapy before RT for early- RT has higher upfront OOP costs at 24 months under Original Medicare, the cost difference
stage breast cancer. Our findings indicate that more than 80% of patients who underwent narrows over five years, with RT incurring slightly higher cumulative costs. This reflects
chemotherapy experienced FT. Approximately 1 in 5 patients not experiencing FT received RT’s one-time nature versus the ongoing costs of ET, which more than doubles over the
chemotherapy. The findings indicate that chemotherapy plays a significant role in patient same period. Medigap Plan G beneficiaries experience significantly reduced OOP costs,
quality of life, highlighting a subsection of patients who may benefit from proactive financial favoring RT further at five years, while Medicaid eliminates OOP costs entirely for either
assistance to reduce the detrimental effect of FT on their livelihood. Clinical trial infor- treatment, ensuring equitable access to both options. These findings support the initial
mation: NCT05978232. Research Sponsor: Susan Komen Foundation; National Medical EUROPA trial results favoring RT over ET for HRQOL and treatment-related toxicity.
Fellowships. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
10s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

520 Poster Session 521 Poster Session

Impact of body mass index (BMI) on efficacy and safety of abemaciclib in The effect of endocrine therapy omission on survival in ER-negative PR-low
breast cancer patients (pts) treated in the monarchE trial. First Author: Christine (1–10%) early-stage breast cancer treated with chemotherapy. First Author:
Desmedt, Laboratory for Translational Breast Cancer Research, KU Leuven, Leuven, Shawn Michael Doss, Medical College of Georgia, Augusta, GA
Belgium Background: The effect of endocrine therapy omission on outcomes in breast cancer (BC)
Background: Two years (yrs) of adjuvant abemaciclib + endocrine therapy (ET) resulted in sustained im- with low progesterone receptor (PR) expression (1–10%) is unknown. Previously, omitting
provement in invasive disease-free survival (IDFS HR=0.68, 5 yrs rates: 84% abemaciclib + ET vs 76% ET, 8% adjuvant endocrine therapy in estrogen receptor (ER)-low (1–10%) BC was associated with
absolute benefit) in pts with hormone receptor positive, human epidermal growth factor receptor 2 negative,
node-positive, high-risk early breast cancer (EBC). Obesity is an established factor influencing the biology and
worse survival (1), but whether this is also true for PR 1–10% (PR-low) BC has not been
prognosis of breast cancer; however, the specific impact on treatment (tx) outcomes remains uncertain. Here studied. We analyzed outcomes for omission of endocrine therapy in patients with ER-
we report efficacy and safety by BMI in monarchE. Methods: Pts were randomized 1:1 to receive ET for at least negative PR-low stage I–III BC who received chemotherapy. Methods: We identified
5 yrs +/- abemaciclib for 2 yrs. Groups were defined by baseline BMI (kg/m2): as obese ($30), overweight 46,704 patients from the National Cancer Database (diagnosed 2018–2020) with PR
(25,30), and non-overweight (,25). IDFS/DRFS in each group was assessed using Kaplan-Meier method and 1–10% stage I–III BC, of whom 3651 (7.8%) were ER-negative. Of these, 2,915 (79.8%)
unstratified Cox model. Safety was summarized by subgroup. Results: 1507 pts (27%) were obese, 1762 (32%)
received chemotherapy. After excluding incomplete data for covariates, 2,796 remained for
were overweight, and 2227 (41%) were non-overweight. Most obese pts were postmenopausal (67%), received
aromatase inhibitor as first ET (75%) and a substantial proportion (47%) had $4 comorbidities, vs 60%, 69%, analysis. Cox proportional hazards models were used to analyze overall survival (OS).
34% among overweight pts and 47%, 63%, 30% among non-overweight, respectively. Pts $65 yrs constituted Multivariate Cox regression and propensity score matching were performed to account for
18%, 17% and 12% of these groups. Disease characteristics were balanced across BMI groups. A consistent tx confounding by age, stage, comorbidity score, HER2 status, year of diagnosis, and grade.
benefit in IDFS was observed with the addition of abemaciclib to ET across all 3 BMI groups: obese (HR = 0.67, This study protocol was developed and reviewed by oncology faculty prior to imple-
95% CI: 0.53, 0.85), overweight (HR = 0.73, 95% CI: 0.58, 0.91), and non-overweight (HR = 0.68, 95% CI: 0.55, mentation. Results: Of the final cohort of 2,796 ER-negative PR-low BC patients, 73.6%
0.83), with an interaction p-value of 0.858. 5yr IDFS rates in the ET arm were lowest in obese pts (74%) and
similar for overweight and non-overweight pts (77% in each). In the abemaciclib + ET arm, absolute im- were HER2-negative and 85.0% were high grade. Stage distribution was 34.8% stage I,
provements in IDFS were 8.8%, 5.8% and 7.9% across each respective BMI group. Similar findings were 43.6% stage II, and 21.6% stage III. Endocrine therapy was omitted in 2,051 (73.4%). OS
observed for DRFS. Obese pts had fewer grade$3 (G$3) neutropenia, and related dose hold/reductions. was 93.9% (95% CI 93.0–94.9%) at 2 years and 86.1% (95% CI 84.3–87.9%) at 4 years, with
Despite higher G$3 diarrhea in obese pts, related dose reductions and discontinuations were similar in all 3 267 total deaths. In the univariate (unadjusted) analysis, omission of endocrine therapy
groups. G$3 ALT elevations were low in all groups. Serious AEs (SAEs) were more common in obese pts, was associated with worse 4-year OS (hazard ratio [HR] 1.84, 95% CI 1.34–2.53, p,0.001).
across tx arms (Table). Conclusions: In pts with high-risk EBC, adjuvant abemaciclib + ET showed consistent
and clinically meaningful tx benefit across BMI subgroups, along with a manageable safety profile. Additional
In the multivariate (adjusted) analysis, the HR was 1.72 (95% CI 1.25–2.37, p,0.001).
analyses are planned to adjust for the impact of confounding factors such as comorbidities. Clinical trial Interaction testing for endocrine therapy and HER2 status was not significant. To account
information: NCT03155997. Research Sponsor: Eli Lilly and Company. for possible pandemic impacts on OS, a sensitivity analysis of 2,639 patients (after ex-
cluding those who did not survive six months beyond definitive surgery) was performed
Abemaciclib + ET ET
and yielded a HR of 1.50 (95% CI 1.08–2.10, p=0.016). After propensity score matching of
Non- Non-
Obese Overweight overweight Obese Overweight overweight 1,490 ER-negative PR-low BC patients (matched by age, stage, comorbidity score, HER2
% n=723 n=886 n=1118 n=784 n=876 n=1109 status, year of diagnosis, and grade), omission of endocrine therapy was still associated
‡1AE Any grade 98 99 98 91 89 88 with worse 4-year OS (HR 1.63, 95% CI: 1.12–2.36, p=0.010). In contrast, omission of
G ‡3 50 48 52 21 15 15 endocrine therapy in matched ER-negative PR-negative (instead of PR-low) BC patients
Neutropenia 14 19 24 0.1 0.5 2
Diarrhea 10 8 6 0.3 0.2 0.2 was not associated with worse OS (HR 1.12, 95% CI: 0.92–1.37, p=0.27). Conclusions: In
ALT 2 3 3 1 0.7 0.5 ER-negative PR-low (1–10%) early-stage BC treated with chemotherapy, omission of
All tx 7 6 5 2 0.3 0.5
discontinuation endocrine therapy may be associated with worse overall survival, suggesting that en-
due to AE docrine therapy could improve survival in ER-negative BC patients even with only low
Dose reductions 41/18/ 43/18/6 45/17/11 - - -
related to any 6 (1–10%) PR expression. Further investigation is recommended as this retrospective study
AE/diarrhea/
neutropenia
design cannot establish causality. 1. Choong, 2024, JCO. Research Sponsor: None.
‡1 SAE 20 16 13 11 8 9
Fatal 3 2 1 2 1 1

522 Poster Session 523 Poster Session

Practice patterns related to ovarian function suppression in the SWOG Beyond the 5-year mark: Adherence to and continuation of extended adju-
S2010 clinical trial of young women with breast cancer. First Author: vant endocrine therapy in non-metastatic breast cancer patients. First Author:
Norah Lynn Henry, University of Michigan Medical School, Ann Arbor, MI Jenny Wei, Department of Hematology-Oncology, Kaiser Permanente San Francisco, San
Background: Addition of ovarian function suppression (OFS) to adjuvant endocrine Francisco, CA
therapy (ET) improves disease free survival for young patients at high risk of breast Background: Extending adjuvant endocrine therapy (AET) beyond five years has been
cancer recurrence. SWOG S2010, an ongoing clinical trial that has completed enrollment, shown to be beneficial for women with non-metastatic, hormone-responsive breast
is examining the benefit of active symptom monitoring for improving 18-month per- cancer. While many studies have examined adherence to (taking medication according
sistence with ET plus OFS. Because data are currently limited on the use of OFS and to prescribed regimen) and continuation of (taking medication for prescribed duration)
endocrine therapy in routine practice, this report of S2010 enrollees describes practice AET, most have focused on the first five years of use and data beyond that period are
patterns related to the prescription of ET for young patients starting OFS plus ET for lacking. To address this knowledge gap, we conducted a large retrospective cohort study
treatment of stage 1-3 hormone receptor-positive breast cancer. Methods: We char- to determine adherence to and continuation of AET beyond the first 5 years among a
acterized the baseline demographic, clinical, and pathologic data from the pre- and cohort of non-metastatic breast cancer patients treated in a community setting.
perimenopausal patients starting OFS plus ET for treatment of breast cancer enrolled in Methods: We estimated adherence to and continuation of extended AET (tamoxifen or
S2010 using descriptive statistics. Results: Of the 557 participants enrolled on S2010, aromatase inhibitor) among 13,675 women at Kaiser Permanente Northern California
median age was 44.5 years (range 23.5-57.1), 25% were Hispanic, 6% Black, 6% Asian, (KPNC), an integrated healthcare system caring for approximately 4.5 million members.
and 76% White. In addition, 39% were not college graduates, 9% were uninsured, and 58% The cohort consisted of women diagnosed with Stage I-III, estrogen receptor positive
worked full time. Most patients were overweight (34%) or obese (39%). Planned ET was (ER+) breast cancer and were treated at KPNC from 2008 to 2017. Adherence was
aromatase inhibitor (AI) therapy for 423 (76%) and tamoxifen for 134 (26%). For OFS, 425 defined as medication possession ratio .80% for each 12-month period during follow-
(80%) received GnRHa injections, 33 (6%) underwent bilateral salpingo-oophorectomy, up. Continuation was defined as time to last pill possession date before a 180-day gap of
and 74 (14%) had chemotherapy-induced ovarian failure (CIOF). Of the 32 (43%) par- AET and was based on the Kaplan-Meier estimator. Both adherence and continuation
ticipants with CIOF who were under the age of 45, 12 (38%) were planning to receive AI were examined for the 10-year period after initiation. Results: Of the 13,675 eligible
therapy without concomitant GnRHa therapy. There were 294 participants (55%) with women who initiated AET, 81% were 50 years or older, 61% were white and 76% had an
anatomic stage 1 ER-positive breast cancer, of whom 127 (43%) did not receive Elixhauser Comorbidity Score of 2 or less. Most women had Stage I or II breast cancer
chemotherapy. Of the 174 participants (33%) with anatomic stage 2 disease, 32 (18%) (90.9%) that was progesterone receptor positive (79%) and HER2 negative (84%). We
did not receive chemotherapy. Conclusions: Most participants on this clinical trial observed a gradual decline in adherence to AET each year from 79% in year 1 to 60% in
received GnRHa therapy as their method for OFS, and an AI for their endocrine therapy. year 5, followed by a dramatic drop to 23% in year 6. After year 6, annual adherence again
We found that a small number of patients under the age of 45 with CIOF did not receive dropped gradually and was 10% in year 10. Median AET continuation time was 5.1 years.
GnRHa therapy and planned to start AI therapy, even though such patients are at high Similarly, there was also a striking decline in AET continuation between year 5 and 6, with
risk of recovery of ovarian function. Additionally, approximately one-quarter of trial 52% continuing AET to the end of year 5 and 20% continuing to the end of year 6. Only
participants appeared to have low risk disease based on non-receipt of chemotherapy 4.5% of the cohort continued to the end of year 10. Conclusions: We observed a
and anatomic Stage 1, despite NCCN guidance recommending that OFS be reserved for substantial drop in both adherence to and continuation of AET following year 5 into year
patients with higher risk features. These findings suggest that additional education 6 with only 4.5% of the cohort continuing to the end of year 10. This period may represent
about the optimal use of OFS in young women with hormone receptor-positive breast an opportunity for intervention to improve use of extended AET. Future studies are
cancer may be warranted. Clinical trial information: NCT05568472. Research Sponsor: needed to identify factors affecting adherence to and continuation with extended AET.
National Cancer Institute; R01CA266012; National Cancer Institute; UG1CA189974; Research Sponsor: None.
Hope Foundation for Cancer Research.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 11s

524 Poster Session 525 Poster Session

Impact of initial chemotherapy dosing on subsequent dosing patterns and Chemotherapy (CT) declination among patients with early-stage hormone
treatment completion in early-stage breast cancer. First Author: Maria Jose receptor positive breast cancer (BC) and high Oncotype DX recurrence
Monroy Iglesias, Memorial Sloan Kettering Cancer Center, New York, NY scores (RS). First Author: Inimfon Jackson, The University of Texas MD Anderson
Background: While breast cancer treatment guidelines provide dosing recommendations, Cancer Center, Houston, TX
some patients do not receive the full expected dose at outset. Preemptive dose reduction, Background: Among patients with hormone receptor positive, HER2-negative (HR+/
often due to toxicity concerns, may influence subsequent reductions or early discontinuation. HER2-) BC, the 21-gene Oncotype DX assay is both prognostic of recurrence risk and
The impact of initial dose reductions on subsequent care delivery remains unclear. predictive of CT benefit. However, some patients decline CT despite their physician’s
Methods: The Optimal Breast Cancer Chemotherapy Dosing (OBCD) study is a cohort of 34,109 recommendations. We investigated the factors associated with CT declination and its
women diagnosed with stage I-IIIA breast cancer at two U.S. integrated healthcare systems impact on overall survival (OS) among patients with early-stage HR+/HER2- BC and high
between 2004-2019. We examined associations between first-cycle dose proportion (FCDP), RS. Methods: Patients ($18 years) diagnosed with HR+/HER2- BC from 2018-2021, with
categorized as $95%, 90-95%, 85-90%, ,85%, with average relative dose intensity (ARDI) pathologic (p) T1-T3, pN0-N1 disease and RS .25, were identified in the National Cancer
categories and early discontinuation. Adjusted prevalence ratios (aPRs) were estimated using Database. Multivariable logistic regression was used to examine the factors associated
generalized linear models of the Poisson family to evaluate associations between FCDP with CT declination. Furthermore, multivariable Cox proportional hazards regression was
($95% vs. ,95%) and the likelihood of further dose reductions (based on ARDI) and early
used to evaluate the association between CT declination and OS based on a propensity
discontinuation. Analyses were performed overall and stratified by age, body mass index (BMI),
score matched 1:5 cohort using year of diagnosis, age, race/ethnicity, pT and pN.
and Charlson comorbidity index (CCI), as older age and comorbidities, including obesity, are
Results: Among 23,416 patients with early-stage HR+/HER2- BC and RS .25, 74.3% were
linked to higher risk of dose reductions. Results: Among 9,724 women receiving adjuvant
chemotherapy, 66% of those with FCDP $95% remained fully dosed throughout treatment. In
non-Hispanic White (NHW) and 12.1% were Black. Overall, 2601 (11.1%) patients declined
contrast, 46% of patients in both the 90–95% and 85–90% FCDP groups stayed in the same CT despite physician recommendation (median RS of 30). Among those declining CT,
category. The highest likelihood of early discontinuation was seen in patients with FCDP ,85% 15.8% also declined endocrine therapy. On univariate analysis, CT declination was as-
(19%) compared to 13% in the FCDP $95% group (p ,0.01). Multivariable analyses showed sociated with older age, Black race and lobular histology. After adjustment, each unit
that women with FCDP ,95% were more likely to experience further dose reductions (aPR increase in RS was associated with lower odds of CT declination (aOR=0.97; 95%CI
1.34; 95%CI 1.21-1.49), than those with FCDP $95%; but interactions with BMI, CCI, or age 0.96–0.97). A more recent year of diagnosis was associated with lower odds of CT
were not statistically significant. However, the associations between FCDP and early dis- declination while older age and Black race (aOR=1.33; 95%CI 1.17–1.51) were associated
continuation (aPR 1.29; 95%CI 1.11-1.50) varied by BMI (p-interaction 0.02) and age (p- with higher odds. Additionally, patients on Medicaid (aOR=1.66; 95%CI 1.40–1.97) and
interaction 0.03). A lower FCDP was positively associated with early discontinuation in women Medicare (aOR=1.29; 95%CI 1.12–1.48) had higher odds of declination compared to those
with BMI 25-30kg/m2 (aPR 1.62; 95%CI 1.27-2.07) and BMI 30-35kg/m2 (aPR 1.50; 95%CI 1.16- on private insurance. Having pN1 disease was associated with lower odds of declination
1.94), but no association was observed for BMI ,25 or .35 kg/m2. Patients aged #49 years than pN0 disease (aOR=0.74; 95% 0.66–0.83). There was no association between
(aPR 1.84; 95% CI 1.39–2.45) and 50–64 years (aPR 1.25; 95% CI 1.07–1.46) with FCDP ,95% comorbidity and declination. Notably, CT declination was associated with an increased risk
were more likely to discontinue early, with no association observed in older adults. of death after a median follow-up of 3 years (aHR=1.28; 95%CI 1.02–1.61) among 10,909
Conclusions: Women with breast cancer who have FCDP ,95% are more likely to experience matched patients. Sensitivity analyses among patients with RS .30 showed similar
further dose reductions, regardless of age, BMI, or CCI, suggesting that initiating treatment results. Conclusions: Though prospective studies have demonstrated the benefit of CT
with reduced doses may not prevent subsequent reductions. Early discontinuation was more among patients with high RS, 11% of patients declined CT. We observed a decrease in CT
likely among patients with FCDP ,95%, particularly those with BMI 25–30 or 30–35 kg/m² or declination over time, as well as with increasing RS. Of note, Black patients, and those on
younger age. These findings may reflect frailty, treatment intolerance, or patient preferences. Medicaid or Medicare were more likely to decline chemotherapy. CT declination was
Understanding the drivers of these decisions is key to guide strategies that balance toxicity associated with worse OS. While the reasons for treatment declination are multifactorial,
concerns with maintaining adequate dosing to reduce treatment cessation in at-risk groups. research is needed to understand the underlying disparities and work toward improving
Research Sponsor: National Cancer Institute; R37CA222793, U24CA171524, U01CA195565, cancer care delivery. Research Sponsor: Susan G. Komen Foundation; SAC220221; The
P30CA008748, and P01CA154292; Geoffrey Beene Cancer Research Center at Memorial Sloan Breast Cancer Research Foundation; 23-190.
Kettering Cancer Center; National Cancer Institute; R50CA211115.

526 Poster Session 527 Poster Session

Treatment-related neutropenia as a predictor of response to adjuvant Real-world (RW) analysis of characteristics and risk of recurrence (ROR) in
palbociclib in the PALLAS trial (ABCSG-42/AFT-05/BIG-14-13/PrE0109). Black patients (pts) with HR+/HER22 early breast cancer (EBC) eligible for
First Author: Kristina Fanucci, Dana-Farber Cancer Institute, Boston, MA NATALEE. First Author: Yara Abdou, Division of Oncology, Lineberger Comprehensive
Background: The PALLAS trial (NCT02513394) investigated the efficacy of the addition of pal- Cancer Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC
bociclib (palbo) to standard adjuvant endocrine therapy (ET) to reduce breast cancer recurrence. Background: The NATALEE trial showed statistically significant and clinically meaningful
Previous analyses of this trial have not shown significant benefit of combination palbo+ET over ET invasive disease–free survival benefit with ribociclib + nonsteroidal aromatase inhibitor
alone. Given prior data showing that extent of neutropenia is associated with response to palbo and (NSAI) vs NSAI alone in pts with stage II/III HR+/HER22 EBC at high ROR that deepened
other cell cycle-specific therapies, we evaluated whether extent of neutropenia could identify
even after all pts were off ribociclib. Prior RW analyses have shown clinically meaningful
responders to palbo in the adjuvant setting. Methods: PALLAS is a global, open-label, phase III trial
that randomized patients (pts) with stage II-III hormone-receptor positive, HER2-negative breast
ROR at 5 y despite endocrine therapy (ET) among NATALEE-eligible pts. TAILORx and
cancer to receive ET for $5 years with or without standard-dose palbo for 2 years in 28-day cycles. RxPONDER reported racial disparities in recurrence and outcomes; however, further
The primary endpoint is invasive disease-free survival (iDFS). For this exploratory analysis, the understanding of characteristics and unmet needs in Black pts is needed. This RW
palbo population was classified into pts with treatment-emergent high-grade neutropenia (HGN) analysis describes pt characteristics and outcomes among Black pts eligible for NATALEE
with maximum grade $3 (absolute neutrophil count ,1000), or low-grade/no neutropenia (LGN) and receiving ET. Methods: Data from the US Flatiron Health EHR-derived deidentified
with maximum grade ,2; these groups were compared to each other and to the ET alone group for database (2011-2023) were used. Selection criteria included pts aged $18 y with an-
5-year iDFS outcomes. Logistic regression examined individual baseline characteristics associated atomical stage I-III (AJCC) HR+/HER22 EBC who had primary tumor surgical resection
with HGN during the first 3 cycles within the palbo group. Impact of HGN during the first 3, 6, and 12 and started adjuvant ET. Pt characteristics and treatment patterns were analyzed in Black
cycles on iDFS was tested using univariate and multivariable landmark Cox regression. vs White pts eligible for NATALEE. ET (any) adherence was defined as proportion of days
Results: The safety population included 5736 pts, 2840 allocated to palbo+ET, 2896 to ET alone. covered $80%. Recurrence-free survival (RFS), distant recurrence–free survival (DRFS),
Prior publications reported no new safety signals, low rates of serious infection, and no grade 5
and overall survival (OS) were assessed descriptively and using multivariate Cox re-
treatment-related events. The palbo+ET group consisted of 1006 (35.4%) LGN and 1834 (64.6%)
HGN. 5-year iDFS results are shown in the table. Pts who received palbo+ET and developed HGN by
gression analysis. Results: A total of7481 pts met selection criteria. Overall, 41.2% (242/
the end of cycle 6 had significantly improved 5-year iDFS compared to those who received ET alone 588) of Black and 33.0% (1697/5142) of White pts met NATALEE eligibility criteria.
(p=0.04), which remained statistically significant when adjusting for body mass index (BMI), prior Compared with White pts, Black pts were younger (median age, 59 y vs 62 y) and more
chemotherapy, and race. Multivariable logistic regression showed lower BMI, prior chemotherapy, likely to be premenopausal (29.3% vs 23.8%), have anatomical stage III disease (35.1% vs
Asian race, and prior mastectomy were significantly associated with HGN (all p,0.05). 25.7%), more extensive nodal involvement (19.8% N0, 54.1% N1, 14.0% N2, and 7.0% N3
Conclusions: In this exploratory analysis of the phase III PALLAS adjuvant trial, addition of palbo vs 24.0% N0, 53.3% N1, 11.3% N2, and 6.2% N3), higher use of Medicaid (13.2% vs 3.6%),
to ET appeared to be superior to ET alone in pts who developed HGN in the first 6 cycles of lower socioeconomic status (lowest socioeconomic status quintile, 33.9% vs 11.4%),
treatment but not in those who had LGN. These findings are consistent with observations in the higher obesity (body mass index $30, 52.9% vs 38.4%), and lower ET adherence (3 y ET
metastatic setting suggesting that neutropenia could be a useful biomarker for palbo concen- adherence: 57.0% vs 65.2%; 5 y ET adherence: 48.3% vs 56.2%). Five-y RFS, DRFS, and OS
tration and efficacy. Clinical trial information: NCT02513394. Research Sponsor: Pfizer. rates were 74.3%, 77.6%, and 85.0%, respectively, in Black pts and 83.2%, 84.5%, and
Maximum grade 5-year iDFS 90.9% in White pts. Compared to White pts, Black pts had worse RFS (hazard ratio, 1.5;
neutropenia P=0.0045), DRFS (hazard ratio, 1.4; P=0.0272), and OS (hazard ratio, 1.7; P=0.0023) after
measured Palbo+ET Palbo+ET ET alone Hazard p-
at end of cycle: HGN (%) LGN (%) (%) Ratio value
adjusting for age, menopausal status, body mass index, tumor size and grade, nodal
status, chemotherapy use, and socioeconomic status index. Conclusions: In a RW
3 84.9 84.4 1.06 0.54
84.9 82.9 1.17 0.06 dataset of pts eligible for adjuvant ribociclib based on NATALEE inclusion criteria, Black
6 85.6 84.9 1.08 0.44 pts had a higher recurrence risk and worse survival compared with White pts, under-
85.6 83.4 1.19 0.04 scoring the opportunity to improve outcomes and address racial disparities in Black pts
12 86.3 85.9 1.06 0.57
86.3 85.0 1.12 0.20 with HR+/HER22 EBC. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
12s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

528 Poster Session 529 Poster Session

Real-world patterns of Oncotype DX (O-Dx) testing and chemotherapy (CT) Estrogen receptor expression in residual breast cancer following neoadju-
use among patients with early-stage, hormone receptor–positive (HR+) vant chemotherapy. First Author: Sarah K. Premji, Mayo Clinic, Rochester, MN
breast cancer (BC). First Author: Marija Sullivan, The University of Texas MD Background: Neoadjuvant chemotherapy (NAC) is commonly administered to patients
Anderson Cancer Center, Houston, TX (pts) with estrogen receptor alpha (ER) negative (immunohistochemistry [IHC] 0%) and ER-
Background: O-Dx, a 21-gene assay, has transformed the management of patients with low (IHC 1-10%) breast cancer (BC). For pts without pathological complete response (pCR),
early-stage, HR+, HER2-negative (HER2-) BC by guiding personalized treatment decisions ER expression may differ between baseline and residual BC following NAC. We previously
regarding adjuvant CT. However, real-world patterns of CT use based on recurrence score demonstrated adjuvant endocrine therapy (ET) omission in ER-low early-stage BC was
(RS) remain a topic of interest. Using a large, hospital-based database, we evaluated testing associated with significantly worse overall survival (OS); however, this effect was re-
patterns, factors associated with O-Dx testing, and examined predictors of CT use among stricted to pts with residual BC following NAC and those with higher baseline ER levels (IHC
patients by RS. Methods: Patients $18 years with early-stage, HR+/HER2- BC diagnosed 6-10%) (Choong et al., ASCO 2024). Here, we assessed how often ER was expressed in
from 2018-2021 who underwent surgery and had pT1-T3, pN0-N1 disease were identified in residual invasive BC for pts treated with NAC for ER-negative and ER-low BC.
the National Cancer Database. Descriptive analyses were used to compare baseline Methods: Weidentified pts with pre-treatment (tx) ER-negative and ER-low (regardless of
sociodemographic, clinical and treatment characteristics by receipt of O-Dx testing. HER2 expression) stage I-III BC treated with NAC who underwent BC surgery at Mayo Clinic
Multivariable logistic regression was used to examine factors associated with receipt of O- Rochester between 2009 and 2023. ER IHC was performed in a CAP/CLIA laboratory. We
Dx testing and predictors of CT use, stratified by RS into 0-10 (low risk), 11-25 (intermediate evaluated ER expression in residual invasive BC for pts without pCR. The percent of pts
risk) and . 25 (high risk). Results: Of 319,771 patients with early-stage, HR+/HER2- BC, with a post-tx change in ER status was estimated and reported with 95% Wilson confidence
54% (172,491) received O-Dx testing. Median age was 61 among those who received testing intervals. Results: 955 pts (838 [88%] pre-tx ER-negative; 117 [12%] pre-tx ER-low) met
and 65 among those who did not. Black (aOR = 0.93; 95%CI 0.90–0.95) and Hispanic inclusion criteria, of whom 69% had HER2-negative and 31% HER2-positive BC. The median
patients (aOR = 0.89; 95%CI 0.86–0.92) were less likely to receive testing than White age at diagnosis was 52 (range: 24-86). 496 (52%) had residual BC. Residual BC was more
patients. Compared to private insurance, those on Medicare (aOR = 0.93; 95%CI 0.91–0.95), common in HER2-negative versus (vs) positive tumors (56% vs 42%, p , 0.001) but did not
Medicaid (aOR = 0.89; 95%CI 0.86–0.92), or no insurance (aOR = 0.91; 95%CI 0.85–0.98) had differ significantly by ER status (ER-negative 51% vs ER-low 57%, p = 0.22). Of those with
lower odds of O-Dx testing. While older age, higher comorbidity scores and pN1 disease were residual BC, 277/496 (56%) had ER re-testing. Rates of ER re-testing did not vary for pre-tx
associated with lower odds of testing, recent year of diagnosis, pT2 disease, lobular ER-negative vs -low (57% vs 51%, p = 0.37) but were significantly lower for HER2-positive
histology, higher grade and treatment in academic facilities were linked to higher odds of vs HER2-negative tumors (39% vs 62%, p , 0.001). Among those with post-tx testing, 31/
testing. Overall, 16% (51,213) of patients received CT. Of those, 17.3% did not have an O-Dx 277 (11%, 95% CI: 8-15%) had an increase in ER expression from pretreatment levels
test, while 76% of those who received CT and testing had RS . 25. 4.2% of patients aged 18- (defined as ER IHC , 1% to either 1-10% or . 10%; or ER IHC 1-10% to . 10%). In these 31
49 years received CT despite RS 0-10. Among patients without a test, being Black (aOR = pts, the original NAC was for either TNBC (21/31; 68%) or HER2+ BC (10/31; 32%). In pts
1.19; 95%CI 1.12–1.26) or Hispanic (aOR = 1.08; 95%CI 1.01–1.16) was associated with with pre-tx ER-negative BC, 27/243 (11%, 95% CI: 8-16%) had ER expression in the residual
higher odds of receiving CT. Black patients with RS . 25 (aOR = 0.84; 95%CI 0.76–0.93) BC including 14/243 (6%, 95% CI: 3-9%) with ER . 10% and 13/243 (5%, 95% CI: 3-9%) with
were less likely to receive CT than White patients. Larger tumors, pN1 disease and higher- ER 1-10%. For pts with baseline ER-low BC, 4/34 (12%, 95% CI: 5-27%) had ER . 10% in the
grade tumors were associated with greater odds of CT receipt while older age at diagnosis residual BC. Among the 31 pts where ER increased following NAC, 21/31 (68%) received
and lobular histology were associated with lower odds regardless of RS. Conclusions: O-Dx adjuvant ET, including 17/18 if ER was . 10% in the residual disease and 4/13 in those with
testing has been increasingly incorporated into clinical practice. Our findings highlight ER 1-10%. Conclusions: In pts treated with NAC for ER-negative or ER-low BC not
disparities in the receipt of O-Dx testing and CT use, particularly according to RS. Black achieving pCR, we identified higher ER expression in the residual breast cancer in . 10% of
patients who did not undergo O-Dx testing were more likely to receive CT, while those with pts. Given that omission of ET in ER-low BC with residual cancer following NAC is as-
RS . 25 were less likely to receive CT. Further research is needed to explore physician and sociated with worse survival, repeat biomarker testing should be considered in those
patient decision-making regarding O-Dx testing and adjuvant CT. Research Sponsor: Susan without pCR, and ET individualized according to ER expression. Research Sponsor: None.
G. Komen; SAC220221; Breast Cancer Research Foundation (BCRF); 23-190.

530 Poster Session 531 Poster Session

Molecular and immune profiling of breast cancer from pregnancy to post- Prospective decision impact study of the Breast Cancer Index: Results from
partum: Insights into the tumour-immune landscape during breastfeeding the BCI Registry study. First Author: Tara B. Sanft, Yale University, New Haven, CT
from GEICAM EMBARCAM study. First Author: Regina Pe~ na-Enriquez, Instituto Background: The Breast Cancer Index (BCI) is a validated gene expression assay that
Maimónides de Investigación Biomédica de Cordoba (IMIBIC)-Hospital Universitario provides an individualized risk of late distant recurrence and predicts the likelihood of
Reina Sofia, Universidad de Cordoba, Córdoba, Spain benefit from extended endocrine therapy (EET) in HR+ early-stage breast cancer. The
Background: Breast cancer (BC) is the most common malignancy among young women of objective of this analysis was to assess the influence of BCI on clinical decision-making
childbearing age, with a rising incidence in this population. Pregnancy-associated breast regarding EET. Methods: The BCI Registry study is a prospective, multi-institutional
cancer (PABC) is an aggressive entity linked to poor prognosis and elevated metastatic risk. study investigating the long-term clinical outcome, decision impact, and quality of life in
Despite advances in BC research, the impact of pregnancy, breastfeeding, and postpartum HR+ breast cancer patients receiving BCI testing as part of routine clinical care.
mammary gland remodeling on tumor biology remains unclear, highlighting the need to explore Physicians and patients completed pre- and post-BCI test questionnaires to assess
their interaction with the tumor microenvironment for novel therapies. Methods: A gene physician decision-making; physician confidence; and patient preferences and concerns
expression and immune cell profiling analysis was conducted using the nCounter Breast
for EET. Pre- and post-BCI responses were compared using McNemar’s test and the
Cancer 360 panel (NanoString) and CIBERSORTx (Newman 2019, Nat Biotechnol) on FFPE
Wilcoxon signed rank test. The BCI Registry Study is registered on [Link]
tumor samples from GEICAM/2017-07 EMBARCAM study (NCT04603820) PABC patients
during the gestation (PABC_GS, n=21), breastfeeding (PABC_BF, n=21), and first-year post- under NCT04875351. Results: In the current analysis,pre- and post-BCI testing
partum (not during lactation period, PABC_FY, n=15) vs non-PABC tumours (n=49). Differential questionnaires were completed for 2850 physicians and 2832 patients. 88.6% of pa-
expression analysis per gene and biological signature was performed using the limma package. tients were postmenopausal, 76.5% N0, 73.0% T1, 53.5% G2, and 13.0% HER2-positive.
P-values were adjusted with the Benjamini-Yekutieli false discovery rate (FDR) method. Following BCI testing, physicians changed treatment recommendations for EET in 41.2%
Additionally, the LM22 matrix from CIBERSORTx was employed to quantify 22 immune cell (1175/2850) of patients (p,0.001). In cases where physicians recommended EET prior
types from normalized NanoString data. Statistical significance was set at 5%. Results: PABC to BCI testing, 49.8% (775/1555) changed their recommendation to not treat with EET,
clinical subtypes were 46% HR-positive/HER2-negative, 21% HER2-positive, and 33% triple- while 31.2% (400/1280) of those who did not recommend EET prior to BCI testing
negative. Differential gene expression analysis identified similar significant enrichment changed their recommendation in favor of EET. Following BCI testing, 43.9% (1250/
pathways across all PABC groups compared to non-PABC: DNA repair-related signatures (HRD, 2850) of physicians felt more confident in their recommendation (p,0.001) and 43.2%
BRCAness, BC p53) and BC proliferation (adj p,0.05), along with higher CDK4 expression and (1223/2832) of patients felt more comfortable with their EET decision (p,0.001). The
genomic risk (p,0.05). Conversely, key regulatory pathways such as apoptosis, TGF-Beta, and percentage of physicians having high confidence levels (confident or strongly confident)
PD-L1 were downregulated (adj p,0.05). Nonetheless, it is noteworthy that the PABC_BF increased from 63.6% (N=1813) pre-BCI testing to 88.2% (N=2515) post-BCI testing. The
group showed a unique profile marked by increased immune activity and cell abundance percentage of physicians having low confidence levels (not at all confident, not con-
(cytotoxic cells, CD8 T cells, T-reg, cytotoxicity), elevated SOX2 expression (adj p,0.05) and fident, or ambivalent) decreased from 33.1% (N=943) pre-BCI testing to 11.0%(N=313)
inflammatory chemokines levels (p,0.01) compared to non-PABC. The CIBERSORTx analysis
post-BCI testing. In BCI (H/I)-Low patients, 48.9% (868/1776) showed a decreased
supported these findings, demonstrating a significantly higher abundance of several immune
preference for EET (p,0.001). In BCI (H/I)-High patients, 34.6% (365/1056) showed an
cells in PABC_BF, remarkably CD8+ T-cells and T-regs, compared to all other groups (p,0.05).
Conclusions: This GEICAM EMBARCAM sub-study reveals that PABC tumors display ag-
increased EET preference (p,0.001). After BCI testing, significantly more patients were
gressive molecular features across all subtypes, contributing to poor prognosis. Notably, the less concerned about cost (23.9%, p,0.001), drug safety (25.7%, p,0.001), and EET
breastfeeding-associated subset (PABC_BF) exhibits a highly active tumor-immune micro- benefit (30.9%, p,0.001). No significant change in concern regarding side-effects was
environment with robust immune cell infiltration and inflammatory signalling, highlighting observed (p=0.58). Conclusions: Incorporating BCI into clinical practice resulted in
potential for targeted immunotherapy. These findings underscore the need for further clinical significant changes in physician recommendations for EET, while at the same time
research to optimize immune-based strategies in PABC patients’ management. Clinical trial increasing physician confidence. Knowledge of the BCI result improved patient pref-
information: NCT04603820. Research Sponsor: Instituto de Salud Carlos III (PI18/00817 and erence, satisfaction and reduced patient concerns regarding cost, drug safety and
PI22/00969).; Fundacion Le Cado. benefit of EET. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 13s

532 Poster Session 533 Poster Session

HER2DX prognostic value in older patients with HER2-positive early breast The difference of clinical and molecular characteristics between HR-posi-
cancer: A correlative analysis from the RESPECT phase III trial. First Author: tive/HER2-positive and HR-negative/HER2-positive early breast cancer: A
Kazuki Nozawa, Department of Advanced Clinical Research and Development, Nagoya secondary analysis of 11 clinical trials. First Author: Hongmei Zheng, Department
City University, Graduate School of Medical Sciences, Nagoya, Japan of Breast Surgery, Hubei Cancer Hospital, Tongji Medical College, Huazhong University of
Background: HER2DX, the first multigene assay specifically designed for HER2+ breast Science and Technology; Hubei Provincial Clinical Research Center for Breast Cancer,
cancer, has demonstrated potential to guide treatment decisions. However, its validation in and Wuhan Clinical Research Center for Breast Cancer, Wuhan, China
the context of de-escalated chemotherapy regimens, including trastuzumab monotherapy, Background: HER2+ early breast cancer (EBC) are treated as an identical cluster of patients in clinical
in older patients remains limited. In the RESPECT trial (NCT01104935, JCO 2020), 1-year of practice. However, the various status of hormone receptor (HR) leads to different outcomes. We aimed to
trastuzumab monotherapy was shown to be a clinically meaningful adjuvant option clarify the difference of clinical and molecular characteristics between HR+/HER2+ and HR-/HER2+ EBC.
Methods: This secondary analysis assessed pathologically complete response (PCR) and disease free
compared to de-escalated chemotherapy and trastuzumab in older patients with HER2+
survial (DFS) among HR+/HER2+ and HR-/HER2+ EBC patients enrolled in the 11 clinical trials. These
early breast cancer. This exploratory analysis of HER2DX within the RESPECT trial (Trans- studies included 16866 HER2+ EBC patients, with available immunohistochemistry and/or in situ hy-
RESPECT study) aimed to evaluate the assay’s prognostic value. Methods: The RESPECT bridization results. There are five neoadjuvant trials (TRYPHAENA, Kristine, Neosphere, BERENICE and
Phase III trial enrolled patients aged 70–80 years with stage I–IIIA HER2+ early breast Peony) and six adjuvant trials (HERA, HannaH, Aphinity, Katherine, PrefHer and SafeHer). The primary
cancer. Participants were randomized to receive trastuzumab monotherapy (H group) or endpoints of the trials were PCR, DFS, overall preference proportions and adverse event rates.
trastuzumab plus chemotherapy (H+CT group). Chemotherapy regimens included paclitaxel Kaplan–Meier approach and Cox proportional hazards model were applied to estimate the association of
monotherapy (35.1%), anthracycline/cyclophosphamide alone (22.9%), CMF (19.8%), treatment strategies with PCR and DFS among HR+ and HR- populations. The 11 trials were all registered
with [Link], number NCT00976989, NCT02131064, NCT00545688, NCT02132949,
docetaxel monotherapy (14.5%), or docetaxel-carboplatin (3.1%). Risk stratification into
NCT02586025, NCT00045032, NCT00950300, NCT01358877, NCT01772472, NCT01401166 and
HER2DX low- or high-risk groups used both the standard 50 cutoff (scale 1-99) and an NCT01566721. Results: In the16866 HER2+ EBC patients, except for PrefHer and SafeHer trials, which has
exploratory 32 cutoff, previously reported in the APT trial (Tolaney et al., Lancet Oncol, no information about HR status, there are 13801 patients with HR details, of which, HR+ 8004 (58.0%) and
2023). The primary endpoint was relapse-free survival (RFS), with secondary endpoints HR- 5713 (41.4%). In HER2+ EBC, the various status of HR leads to different outcomes. Our study revealed
including overall survival (OS). Results: Among 275 patients in the RESPECT trial, 154 an interesting phenomenon that compared to HR-/HER2+ EBC, HR+/HER2+ EBC has a lower pCR rate.
tumors (56.0%) were profiled using HER2DX (H group: 74; H+CT group: 80). Baseline However, trials from adjuvant therapy studies suggested HR+/HER2+ EBC has a longer DFS, which is not
characteristics of the profiled cohort mirrored those of the overall trial population. Most consistent with other subtype patterns. The pCR rate is a near-term indicator that visualizes the response
patients (92.9%) had a performance status of 0, 28.6% were older than 75 years, 53.2% were rate to a particular treatment, and the DFS is a distant indicator that reflects the overall biological behavior
of the individual. The HR+/HER2+ EBC, despite its low pCR rate, potentially improves its long-term DFS due
HR-negative, 80.5% had node-negative disease, and the majority had pT1c (41.6%) or pT2 to the addition of adjuvant endocrine therapy postoperatively. Conclusions: Compared to HR-/HER2+ EBC,
(44.8%) tumors. Using the HER2DX 50 cutoff, 40 patients (26.0%) were classified as high HR+/HER2+ patients has a lower pCR rate, but has a longer DFS, which deserve further exploration.
risk, and 114 (74.0%) as low risk. RFS was higher in the HER2DX low-risk group compared to Research Sponsor: None.
the high-risk group (hazard ratio [HR] = 2.02, 95% CI: 0.97–4.19), with 5-year RFS rates of The stratification of the HER2+ early breast cancer patients according to HR status in the 11 clinical trials.
92% and 77%, respectively. OS was also superior in the HER2DX low-risk group (HR = 2.74,
Trial Total, n HR+, n(%) HR-, n(%) Unknown, n
95% CI: 1.18–6.36), with 5-year OS rates of 97% and 84%. Using the HER2DX 32 cutoff, 81
patients (52.6%) were classified as high risk, and 73 (47.4%) as low risk. In the H group, 3- TRYPHAENA 225 114(50.6) 111(49.4) 0
Kristine 444 276(62.2) 168(37.8) 0
and 5-year RFS were 97% and 94% in the low-risk group, compared to 87% and 81% in the Neosphere 417 197(47.2) 219(52.5) 1
high-risk group. In the H+CT group, 3- and 5-year RFS were 95% and 95% in the low-risk BERENICE 401 252(62.8) 140(34.9) 9
Peony 329 173(52.6) 156(47.4) 0
group, compared to 93% and 83% in the high-risk group. Conclusions: The HER2DX ge- HERA 5099 2571(50.4) 2528(49.6) 0
nomic risk score demonstrates prognostic value in older patients with HER2+ early breast HannaH 596 265(44.5) 257(43.1) 74
cancer, including those treated with trastuzumab monotherapy. This assay may aid in Aphinity 4804 3082(64.2) 1722(35.8) 0
Katherine 1486 1074(72.3) 412(27.7) 0
identifying patients suitable for treatment de-escalation strategies. Additional analyses will PrefHer 488 NA NA NA
be presented at the conference. Clinical trial information: NCT01104935. Research Sponsor: SafeHer 2577 NA NA NA
Chugai Pharmaceutical Co., Ltd. HR, hormone receptor.

534 Poster Session 535 Poster Session

Patterns of early and late recurrence across breast cancer subtypes in the Real-world evidence of PARPi-related MDS/AML risk in breast cancer pa-
CCTGMA.32 trial. First Author: Ana Elisa Lohmann, University of Western Ontario, tients: An international collaborative network analysis. First Author: Michela
London, ON, Canada Palleschi, IRCCS Istituto Romagnolo per lo Studio dei Tumori (IRST) "Dino Amadori",
Background: An ongoing constant risk of recurrence out to 20 years is well established in hormone receptor Meldola, Italy
positive breast cancer (BC) less so in other BC subtypes. This study aims to describe patters of early (# 5 years
Background: Poly (ADP-ribose) polymerase (PARP) inhibitors have emerged as a
of BC diagnosis) and late recurrence (. 5 years of BC diagnosis) across immunohistochemically defined BC
subtypes - luminal (ER/PgR+HER2-), triple negative (TN: ER/PgR/HER2-) and HER2+ (any ER/PgR) in significant therapeutic advance in breast cancer treatment. However, concerns about
CCTGMA.32 (NCT01101438) which investigated metformin vs placebo in patients enrolled 2010-2013. therapy-related myeloid neoplasms, specifically myelodysplastic syndrome (MDS) and
Methods: 3649 patients with high-risk non-metastatic BC were enrolled and followed for first locoregional and acute myeloid leukemia (AML), necessitate thorough investigation of their safety profile
distant recurrence, new primary cancers and death. Annual rates of these events were calculated in each BC in real-world settings. Methods: Using the TriNetX Global Collaborative Network, we
subtype and averaged for early (years 0-5) and late (after 5 years) post randomization. Results: In luminal (n =
2104), TN (n = 925), HER2+ (n = 620) BC the median follow-ups were 96.2 (range 0.2 to 120.7), 94.5 (0.03 to
conducted a retrospective analysis comparing breast cancer patients treated with PARP
120.5), 95.2 months (0.03 to 119.8), respectively. Patterns of events varied across subtypes and early vs late. In inhibitors (PARPi) versus conventional chemotherapy only (anthracyclines and taxanes).
luminal BC, the early vs late annual invasive cancer event rates (ICERs) was 3.04 vs. 2.31 % (late rate 0.76 of Our primary analysis utilized propensity score matching (1,702 patients per group)
early rate). The annual early vs late rates of distant recurrence (DR) were 2.33 vs 1.72% (late rate 0.74 of early accounting for age and race, to evaluate the risk of developing MDS/AML. Hazard ratio
rate). Bone was the most common site of DR both early and late. In the TN BC, the early vs late annual ICERs (HR) was used to compare the incidence of MDS/AML between the matched cohorts.
were 4.6 and 1.21% (late rate 0.35 of early rate). Annual early vs late DR rates were 3.09 vs. 0.20 % (late rate
0.28 of early rate). Visceral metastases (lung, liver, CNS) were most common early. In HER2+, early vs late
Secondary analyses included an unmatched Cox proportional hazards model in a larger
annual ICERs were 2.93 vs 1.47% (late rate 0.50 of early rate). Annual early vs late DR rates were 2.25 vs 0.71% cohort (1,826 vs 36,257 patients), comparison between different PARP inhibitors, and
(late rate 0.32 of early rate). Bone and visceral metastases were common early. CNS was rare after 5 years in all assessment of mortality risk factors. Results: In the propensity score-matched anal-
BC subtypes. Second primary cancers (new BC and non-primary BC) were frequent across BC subtypes, with no ysis, PARPi treated patients demonstrated a statistically significant higher risk of
fall-off over time; they were responsible for the majority of late events in TN and HER2+ BC. Conclusions: In developing MDS/AML versus chemotherapy only cohort (16 cases versus 10, HR=5.25;
luminal BC, risk of late ICER remains high (annual rate about three-quarters of early rate), while risk of late
events was lower in TN and HER2+BC (late rates one quarter to one-third of early rates). Risk of second primary 95% CI: 1.96-13.92; p,0.0001). Treatment patterns differed notably, with PARPi-treated
cancers did not decrease over time, and second primaries were the most frequent late events in TN and patients receiving more carboplatin (HR=1.73; 95% CI: 1.42-2.10) but less anthracycline
HER2+BC. Clinical trial information: NCT01101438. Research Sponsor: London Health Sciences Foundation - therapy (HR=0.25; 95% CI: 0.20-0.31). The unmatched Cox regression analysis con-
Ontario, Canada; Canadian Cancer Society Research Institute; National Cancer Institute; The Breast Cancer firmed these findings with a higher risk of developing AML/MSD in the PARPi cohort
Foundation - New York; Canadian Breast Cancer Foundation - Ontario, Canada; Ontario Institute for Cancer
Research; Apotex Canada; Hold’em for Life Charity.
(HR=3.47; 95% CI: 1.87-6.27) and identified age (HR=1.03; 95% CI: 1.02-1.05) and
platinum therapy (HR=2.12; 95% CI: 1.26-3.59) as independent risk factors. Within the
Luminal TN HER2+ triple negative group, the data remains statistically significant (HR=3.14; 95% CI: 1.459-
Annual event Annual event Annual event 6.757 ). No significant differences in MDS/AML risk were observed between olaparib and
rate (%) rate (%) rate (%)
talazoparib. While mortality was comparable between groups, prior platinum exposure
Year 0-5 Year 5+ Year 0-5 Year 5+ Year 0-5 Year 5+
emerged as a significant mortality risk factor (HR=2.39; 95% CI: 1.09-5.09).
Any Invasive Cancer Event
Locoregional Event
3.04
0.50
2.31
0.29
4.60
1.15
1.21
0.26
2.93
0.64
1.47
0.15
Conclusions: Our findings indicate a significantly increased risk of therapy-related
Distant Recurrence* 2.08 1.29 3.09 0.20 2.03 0.57 myeloid neoplasms with PARP inhibitor treatment compared to conventional chemo-
Sites of First Metastasis: therapy, particularly in the context of previous platinum exposure in breast cancer
Bone 1.40 0.88 1.13 0.10 0.65 0.42
Lung 0.59 0.56 1.73 0.20 0.83 0.07 patients. These results underscore the importance of careful patient selection and
Liver 0.71 0.56 0.73 0.00 0.50 0.21 monitoring during PARP inhibitor therapy, while highlighting the need for extended
CNS 0.18 0.06 0.65 0.00 0.61 0.00
Second Primary Cancer** 0.66 0.92 0.96 0.90 0.60 0.74 follow-up studies to fully characterize long-term safety profiles. Research Sponsor:
*Including distant recurrence after a local regional events. None.
**Non-breast cancer and new breast cancer events.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
14s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

537 Poster Session 538 Poster Session

Extended endocrine therapy after 5 years of adjuvant LHRH-agonist in Patterns in male and female breast cancer care: A comparative analysis of
premenopausal patients with node-positive hormone receptor (HR)- stage at presentation, treatment, and survival in the Veterans Health
positive early breast cancer. First Author: Carmine Valenza, Division of Early Administration. First Author: Ariana Naaseh, Washington University in St. Louis
Drug Development, European Institute of Oncology IRCCS, University of Milan, Milan, School of Medicine, St. Louis, MO
Italy Background: Male breast cancer (MBC) is a rare disease that is often managed using treatment
Background: There is no evidence regarding the benefit of extended endocrine therapy (eET) beyond protocols derived from female breast cancer (FBC). Given its rarity, limited large-scale national
5 years of adjuvant treatment with LHRH agonists (LHRHa) in premenopausal women with node- cohort data exists to inform clinical treatment decisions for MBC. This study aims to compare
positive, HR-positive early breast cancer (eBC). Methods: We conducted a retrospective study on contemporary treatment trends and survival outcomes for MBC and FBC within the Veterans Health
two prospectively maintained datasets (Young Women Study and IEO Breast Cancer Dataset) to Administration (VHA). Methods: We conducted a retrospective cohort study of patients diagnosed
evaluate the clinical benefit of eET in women who had completed 5 years of adjuvant LHRHa, with MBC and FBC between 2000 and 2022 using national data from the VHA Cancer Cube registry.
remained premenopausal, and had no evidence of distant or locoregional recurrence. This study Demographics, tumor characteristics, treatment (surgery, chemotherapy, hormone therapy, and
included ,40y women at diagnosis (between 2006 and 2016) with node-positive HR+ eBC, with radiation), and survival were compared between MBC and FBC patients. Descriptive statistics and
ductal, lobular, or mixed histological subtypes, receiving or not eET (tamoxifen monotherapy or chi-square tests compared the cohorts. Kaplan-Meier methods and Cox proportional hazards
LHRHa+tamoxifen/aromatase inhibitor [AI]). The primary endpoint was the invasive breast cancer- regression model were utilized to examine overall survival (OS). Results: Of the 14,018 total
free survival (IBCFS), calculated from the 5th year of endocrine therapy (ET) and adjusted for dataset, patients who met inclusion criteria, only 13.9% (n=1952) were males. MBC patients were sig-
age at diagnosis, histotype, stage, disease subtype, type of adjuvant chemotherapy and ET received. nificantly more likely to get diagnosed at an older age (MBC 68.8 vs. FBC 60.0 years; p#0.001) and
Results: 503 patients were included (see Table): 287 received eET for a median duration of 3.6 years present with stage III or IV disease (25.8% vs. 12.9%; p,0.001). Compared to FBC patients, MBC
(Interquartile Range: 2.1–5.0). At a median follow-up of 7.05 years (calculated from the 5th year of patients had significantly higher rates of receiving hormone therapy (56.8% vs. 51.6%; p,0.001)
ET), 50 and 72 IBCFS events occurred in the eET and non-eET groups, respectively. The adjusted and lower rates of chemotherapy (34.3% vs. 37.4%; p=0.007), radiation (21.2% vs. 47.3%;
Hazard Ratio (HR) for IBCFS comparing the eET to the non-eET group was 0.60 (95% CI,0.41-0.88; p ,0.001), and surgery (92.1% vs 93.0%; p=0.01). MBC patients were significantly less likely to
p,0.001). For distant recurrence or death, 28 and 46 events occurred, respectively, and the adjusted undergo breast-conserving surgery (BCS) (11.2% vs. 52.0%; p ,0.001) than FBC patients. In a Cox
HR for distant disease free-survival was 0.43 (95% CI, 0.27-0.71). Among patients receiving eET, the proportional hazard model including age and stratified by stage, MBC patients had reduced OS (6.9
adjusted HR for IBCFS comparing tamoxifen monotherapy (n=137) with LHRHa+tamoxifen/AI vs. 19.0 years; p,0.001) and higher risk-adjusted hazard of all-cause mortality (adjusted hazard
(n=150) was 0.75 (95% CI, 0.41-1.38). Conclusions: Extending endocrine therapy beyond five ratio 1.40, 95% CI 1.30-1.49). OS for MBC was lower than FBC across all stages (Table).
years of LHRHa treatment resulted in significantly higher IBCFS and distant metastasis free-survival. Conclusions: Our national cohort study is the largest series of patients with MBC and FBC in the
Larger prospective studies are required to confirm this finding and determine the most effective eET VHA population to date. We demonstrate that MBC patients present with advanced-stage cancer,
strategy. Research Sponsor: None. are less likely to receive aggressive treatments, are less likely to undergo BCS, and have reduced OS
Patients’ characteristics. across all stages, compared to FBC patients. Our findings highlight the need for further research to
Extended endocrine No extended endocrine
optimize outcomes of MBC patients. Research Sponsor: National Cancer Institute; T32CA009621.
therapy therapy Stage and survival of the cohort.
Characteristic (N=287) (N=216)
Median Survival Median Survival
Age at diagnosis, median (IQR) 37 (35-39) 37 (33-39) Clinical All Patients Males for Males Females for Females
Dataset: IEO | YWS, n 273 | 14 212 | 4 Stage (n=12066) (n =1952) (years) (n=12066) (years) P-value
Histotype: ductal | lobular | mixed, % 91 | 5 | 4 95 | 3 | 2 0 2518 141 (7.2%) 12.3 2377 (19.7%) Not reached ,0.001
pT: pT1 | pT2 | pT3-4, % 37 | 48 | 15 41 | 52 | 7 I 5877 601 9.8 5276 (43.7%) 20.2
pN: pN1 | pN2 | pN3, % 64 | 22 | 14 73 | 17 | 10 (30.8%)
Luminal A-like | B-like (G3 or HER2+), % 47 | 53 49 | 51 II 3589 709 7.1 2880 (23.9%) 17.5
LHRHa combination during years 1-5: tamoxifen | 66 | 34 76 | 22 (36.3%)
aromatase inhibitor, % III 1324 314 5.4 1010 (8.4%) 10.6
Previous chemotherapy, % 77 70 (16.1%)
Previous radiotherapy, % 64 62 IV 730 189 (9.7%) 1.9 541 (4.4%) 2.8
G3, grade 3; IEO, European Institute of Oncology; IQR, interquartile range; YWS, Young Women Study.

539 Poster Session 540 Poster Session

Racial differences in the prognostic value of Oncotype (RS) and MammaP- The impact of racial and socioeconomic disparities on radiation therapy
rint (MP) in postmenopausal, estrogen receptor (ER)–positive, node- delays in breast cancer patients: A National Cancer Database analysis. First
negative (N0) breast cancer (BC) patients with low genomic risk: A National Author: Mehmet Murat Zerey, Miami Cancer Institute, Baptist Health South Florida,
Cancer Database (NCDB) study. First Author: Prashanth Ashok Kumar, George Miami, FL
Washington University Medical Faculty Associates, Washington, DC Background: Despite advancements in breast cancer (BC) treatment, significant dis-
Background: Postmenopausal patients with ER+, HER2 negative, N0 BC often rely on parities in outcomes persist in the United States. Timely initiation of adjuvant radiation
genomic testing, such as RS or MP, to determine the benefit of chemotherapy. However, therapy (RT) is crucial, as delays are associated with increased recurrence. Using the
the prognostic value of RS, particularly among ethnic minority patients, remains un- National Cancer Database (NCDB), this study provides a larger-scale analysis of racial
certain. Methods: We utilized the 2021 NCDB to include postmenopausal female BC and ethnic (R/E) disparities, examining the role of socioeconomic (SE) factors, including
patients aged .50 years. Inclusion criteria were ER+, HER2-negative patients with, education and income, on RT delays and overall survival (OS) differences. Methods: A
stage T1-4, N0, RS ,26, or MP low risk. Patients were stratified by race, Caucasian (W) retrospective analysis was conducted using data from the NCDB. The study included
and African American (AA). Univariate analysis was performed to evaluate patient and female patients (pts) $18 years with stage I-III BC diagnosed between 2004 and 2020.
tumor characteristics. Five-year overall survival (OS) rates were constructed using Pts who received chemotherapy were excluded. Pts were categorized by R/E into four
Kaplan-Meier (KM) product limit estimation. Unadjusted and covariate adjusted as- groups: White (W), Black (B), Hispanic (H), and Other (O). SE variables, including income
sociations between race and OS were evaluated using Cox proportional hazard (PH) and education, were stratified into quartiles (Q1–Q4) as defined by the NCDB. A delay in
regression. Associations between race and OS among subgroups were similarly eval- RT initiation was defined as starting treatment more than 3 months (mos) after surgery.
uated to highlight disparities and trends within the population. Hazard ratios (HRs; AA: Kaplan-Meier (KM) analysis with the log-rank test was used to compare OS across
W), 95% confidence intervals (CIs) and p-values are reported. Results: 96,411 patients groups. Cox proportional hazards regression was performed to estimate hazard ratios
had an RS,26 [W- 89,105 (92.42%) AA-7,306(7.58%)] and 3,146 had MP low [W- (HR) and assess the impact of income and education on delays in RT and survival
2,929(93.1%) AA-217(6.9%)]. Over 93% of patients in both groups received endocrine outcomes. Results: A total of 395,328 female BC pts were included in the analysis. The
therapy, ~75% underwent partial mastectomy and radiation, and 7% received che- median age of the cohort was 65. Most were W (85.2%), followed by B (7.4%), H (4.0%),
motherapy. Significant racial differences in tumor size and grade were observed in the and O (3.5%). Almost all pts had stage I (81.6%) BC. The majority of patients were in the
RS group: T1 tumors (AA: 75% vs. W: 78%), T2 tumors (AA: 23% vs. W: 20%, p,0.001), G1 highest (Q4) income and education quartiles (40.2% and 45.6% respectively). Higher RT
(AA: 29% vs. W: 34%), and G3 (AA: 13% vs. W: 9.3%, p,0.001), but these differences delays ( . 3 mos) were observed in B (11.07%) and H (11.38%) compared to W (5.31%),
were not seen in the MP group. At 5 years, survival for RS ,26 was 95.5% (AA) vs. 97.1% (p , 0.001). Pts in the lowest income and education quartiles (Q1) experienced delays
(W), and for MP low, 96% (AA) vs. 96.7% (W). The adjusted HR for RS ,26 showed worse more frequently (8.02% and 9.29%, respectively) compared to those in Q4 (5.72% and
outcomes for AA (HR: 1.2, 95% CI 1.1–1.31, p,0.001), especially in younger patients, 5.27%, respectively p , 0.001). KM survival analysis revealed significant differences in
grades 1–2 tumors, T1 stage, partial mastectomy and low income. For MP low, the HR OS for delayed RT, with median survival of 218 mos for 0–3 mos, 211 mos for 3–6 mos,
was not significant (HR: 1.08, 95% CI 0.59–1.97, p=0.8). Conclusions: Our analysis and 209 mos for . 6 mos (p , 0.001). KM survival analysis also demonstrated worse
shows that among ER+, HER2- N0 cohort with RS ,26, AA patients have a 20% increase survival for pts in Q1 income and education compared to Q4 (p , 0.001). Cox pro-
in the risk of death compared to W patients, after adjusting for patient- and tumor- portional hazards model, when adjusted for clinical, R/E and SE factors revealed that pts
related factors. In contrast, no survival disparities were observed in the MP low-risk in Q1 for income (HR = 1.411, p , 0.001) and education (HR: 1.036, p , 0.001) had a
group. These findings suggest that RS may not be fully prognostic for AA patients even significantly higher risk of mortality compared to Q4. Conclusions: This study highlights
when accounting for clinic-pathologic risk factors. Study limitations include its ret- significant R/E and SE disparities in timely RT initiation and survival outcomes among
rospective design, potential biases, and incomplete consideration of biological and BC pts. B, H, and pts with lower S/E status experienced greater RT delays, emphasizing
socioeconomic factors. Research Sponsor: None. the critical need for targeted interventions to address delays in care and improve equity
in cancer treatment. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 15s

541 Poster Session 542 Poster Session

Association between chemotherapy use and clinical outcomes in young Long-term outcomes of patients with HER2-positive invasive lobular car-
BRCA carriers with T1N0 breast cancer: Results from an international cohort cinoma in the ALTTO trial (BIG 2-06/NCCTG N063D [Alliance]). First Author:
study. First Author: Filipa Lynce, Dana-Farber Cancer Institute, Boston, MA Guilherme Nader Marta, Medical Oncology, Dana-Farber Cancer Institute, Harvard
Background: Systemic treatment decisions for young BRCA carriers with small node- Medical School, Boston, MA
negative breast cancers present unique challenges due to limited evidence on the benefits Background: Invasive lobular carcinoma (ILC) is the second most common histologic
of chemotherapy in this setting. This study evaluated chemotherapy use and survival subtype of breast cancer (BC), representing 10–15% of cases. HER2 overexpression is rare
outcomes among these patients. Methods: The BRCA BCY collaboration (NCT03673306) in ILC, and there is limited data on the clinical characteristics and outcomes of patients
is an international, multicenter, hospital-based, retrospective cohort study that included (pts) with HER2-positive (HER2+) ILC treated with adjuvant trastuzumab. This study aims
carriers of germline pathogenic variants in BRCA1/2 diagnosed with invasive breast cancer to investigate the prognostic value of ILC histology in this setting. Methods: ALTTO was a
at the age of #40 years between January 2000 and December 2020. Among patients multicenter, randomized phase III trial evaluating the efficacy of trastuzumab, lapatinib,
diagnosed with T1N0 disease, survival outcomes - disease-free survival (DFS) and overall their sequence, or combination as adjuvant therapy in pts with HER2+ early BC. Pts with
survival (OS) defined from breast cancer diagnosis - were compared between patients who pure ILC or invasive BC of no special type (NST) who were enrolled in trastuzumab-
received chemotherapy and those who did not, using multivariate Cox models adjusted for containing arms of the ALTTO trial were included in this analysis. Central pathology review
propensity score (including age at diagnosis, histology, grade, country and year of di- confirmed histologic subtype and was used for classification, while local pathology was
agnosis) and risk-reducing surgeries (bilateral risk-reducing mastectomy (RRM) and/or used when centralized review was unavailable (USA and China). Survival outcomes, in-
risk-reducing salpingo-oophorectomy (RRSO)), and accounting for the delayed entry at the cluding disease-free survival (DFS), and overall survival (OS) were evaluated using Kaplan-
time of BRCA testing (i.e. left truncation). Subgroup analyses were performed according to Meier method and multivariate Cox regression adjusted for prognostic factors. Patterns of
tumor subtype (HR+/HER2- vs triple negative breast cancer (TNBC)). Results: Out of 5660 relapse were analyzed and compared across histological subtypes. Time to distant re-
from 109 centers, 1280 patients had T1N0 breast cancer: T1mic (n = 14, 1.1%), T1a (n = 92, currence (TTDR) and time to CNS recurrence were summarized using cumulative incidence
7.2%), T1b (n = 303, 23.7%), T1c (n = 778, 60.8%), and T1 size unknown (n = 93, 7.3%). Most functions. Results: Among pts in the trastuzumab-containing arms (N = 6281), 84.4%
patients received chemotherapy (80%), although use was less frequent over time. Among underwent central pathology review, with a concordance rate of 67.4% for ILC diagnosis. A
patients who received chemotherapy, the majority were treated with an anthracycline- total of 61 pts with pure ILC (1.0% of the cohort) and 5981 pts with NST were included in the
containing regimen (83.6%) and in the adjuvant setting (85.7%). Patients who received analysis. Pts with ILC were older (mean 54.8 vs. 50.9 years; p=0.002), more likely White
chemotherapy were younger and more likely to have high grade, TNBC or larger tumors (95.1% vs. 68.8%; p,0.001), and postmenopausal (72.1% vs. 56.3%; p=0.01). The pro-
compared to those who did not. The median follow-up was 8.7 years (IQR 5.0-13.4 years), portion of pts with ILC (vs NST) was higher in Europe (67.2% vs 53.7%) and lower in Asia-
during which 428 had DFS events including second primary breast cancer (n = 174), Pacific (8.2% vs 30.6%) (p,0.001). A significantly higher proportion of ILC (vs NST) were
locoregional (n = 130), distant relapse (n = 65), second primary non breast cancer (n = 53), hormone receptor-positive (80.3% vs. 57.4%; p,0.001), Grade 1-2 (51.7% vs. 39.3%;
and 88 had died. Overall, 8-year DFS was 69.4%. In multivariate analysis, no significant p=0.05). At a median follow-up of 9.8 years (IQR 6.9-10.0), no significant differences in DFS
differences in DFS or OS were observed between patients who received chemotherapy and (hazard ratio [HR] 1.14, 95% CI 0.66-1.97; adjusted HR [aHR] 1.33,0.77–2.31), OS (HR 0.96,
those who did not (DFS HR = 0.92, 95% CI 0.65-1.31; OS HR = 0.68, 95% CI 0.37-1.24). No 0.43-2.15; aHR 1.09, 0.48–2.44), or TTDR (HR 1.67, 0.91–3.05) were observed between ILC
significant difference in 8-year DFS was observed between patients with HR+/HER2- and NST. Central nervous system (CNS) relapses were more frequent in ILC (13.6% at 10y,
breast cancer (n = 474) treated with or without chemotherapy (HR 0.91: 95% CI 0.59-1.43), 95% CI 7.1–26.1%) than in NST (5.0%, 4.5–5.7%), with an HR of 3.14 (1.52–6.48) for CNS
or between patients with TNBC (n = 637) treated with or without chemotherapy (HR 0.84: recurrences in ILC when compared to NST. Conclusions: Long-term outcomes were
95% CI 0.46-1.53). Conclusions: In this global study of young BRCA carriers with T1N0 comparable between ILC and NST in HER2+ early BC treated with trastuzumab-containing
breast cancer, chemotherapy was not associated with better DFS or OS overall. However, regimens. The higher incidence of CNS metastases in ILC highlights its unique relapse
these patients remain at high risk of events and warrant investigation of additional risk- pattern, necessitating further investigation to optimize treatment. High discordance
reduction strategies. Research Sponsor: None. between central and local pathology emphasizes the need for standardized histological
review in trials and treatment decisions. Clinical trial information: NCT00490139. Research
Sponsor: Novartis.

543 Poster Session 544 Poster Session

Cost-effectiveness of HER2/neu 655 genotyping in managing trastuzumab- Neuromorphological effects of simultaneous exercise during neo-/adjuvant
induced cardiotoxicity risk in HER2-positive breast cancer patients. First chemotherapy in breast cancer patients: The Exercise Cancer and Cognition
Author: Isabel Blancas, Hospital Universitario Clı́nico San Cecilio / Medicine Department, (ECCO) study. First Author: David Kiesl, Ordensklinikum Linz - Elisabethinen, Linz,
Granada University / Instituto de Investigación Biosanitaria de Granada (ibs Granada), Austria
Granada, Spain Background: Cancer-Related Cognitive Impairment (CRCI) is commonly experienced by
Background: Trastuzumab has significantly improved survival in HER2-positive breast breast cancer patients, often linked to chemotherapy. It manifests as deficits in learning,
cancer patients. However, around 20% of patients experience cardiotoxicity. Car- memory, reaction time, and concentration, with the hippocampus—a critical region for
diotoxicity has been defined as a $10% drop in left ventricular ejection fraction (LVEF) cognitive and affective functions—playing a central role. The hippocampus retains neuro-
or LVEF ,50%, or the appearance of clinical cardiac insufficiency. The HER2/neu 655 genesis potential throughout life, with physical exercise shown to enhance hippocampal
A.G polymorphism has been linked to cardiotoxicity risk. This study evaluates the cost- volume and memory function. High-intensity exercise appears to have superior cognitive
effectiveness of HER2/neu 655 genotyping. Methods: Eighty-eight HER2-positive benefits compared to moderate aerobic activity. The ECCO study aimed to assess the
breast cancer patients treated for early disease with trastuzumab were retrospec- neuropsychological and brain morphological effects of high-intensity interval training (HIIT)
in breast cancer patients, focusing here on imaging data subanalysis. Methods: The ECCO-
tively analyzed. All were genotyped for HER2/neu 655 A.G (AA: n=53, AG: n=32, GG:
study was designed as monocentric two-armed 1:1 randomized controlled trial (RCT),
n=3). LVEF was monitored by echocardiography or isotopic ventriculography at baseline
including a 12-month intervention group and a control group. Patients in the intervention
and regular intervals. Cardiotoxicity was defined as above. Logistic regression adjusting group were instructed to perform strength- and low-intensity endurance training at least
for hormonal status and anthracycline use estimated the association between genotype once a week at home according to the American College of Sports Medicine (ACSM) rec-
and cardiotoxicity. Cost data from the Andalusian Regional Health Service included ommendations and to complete a supervised high-intensity interval training (HIIT) once a
diagnostic tests, cardiology visits, pharmacologic therapy, and hospitalizations. week at our health center. Patients in the control group were advised physical activity
Results: Among the 53 patients with the AA genotype, 3.7% experienced a decrease in recommendations according to the World Health Organsiation (WHO). Participants ran-
LVEF, while 9.4% developed clinical symptoms. For the AG genotype (32 patients), 9.3% domized to the control arm received usual care and physical activity recommendations have
showed an LVEF reduction, and 28.1% presented clinical symptoms. In the GG genotype been given according to usual standards. The volume measurements of the brain were
group (3 patients), 1 patient (33.3%) developed clinical symptoms. AG carriers had a performed with an automatic segmentation tool FreeSurfer version 7.4. The FreeSurfer
significantly higher risk of cardiotoxicity than AA patients (OR adjusted for hormonal algorithm performs an automatic segmentation of subcortical volumes and reconstructs the
status and anthracycline treatment =4.42; p=0.037). HER2/neu 655 A.G genotyping cortex after the elimination of non-brain tissue. Results: MRI data were available at baseline
costs V38. Asymptomatic LVEF reductions usually required 3 cardiology visits including and after 12 months for a total of 67 patients diagnosed with breast cancer. Randomized into
echocardiography (V121.05 each), and one year of pharmacological treatment (car- an intervention arm (n = 35) and control arm (n = 32). The groups were comparable in all
vedilol and/or enalapril therapy; V84.72 total). Cardiac insufficiency costs range from demographic factors such as age, BMI and VO2max. There were no statistically significant
V2,992.61 (grade 1) to V9,363.56 (grade 4). Conclusions: HER2/neu 655 genotyping is changes (p , 0.05) in the volumes between the baseline measurement and the 12-months-
cost-effective for identifying patients at higher risk of trastuzumab-induced car- follow-up measurement. Other brain areas also showed no significant (p , 0.05) alterations
diotoxicity. The low cost of genotyping is outweighed by the potential savings in between baseline and follow-up. No differences were found in the volume change between the
preventing severe cardiac events. Genotype-driven monitoring and proactive cardiac and exercise group and the control group. Conclusions: This analysis of the ECCO study revealed
targeted cardiovascular risk management in AG carriers could reduce both the incidence that there were no significant changes in hippocampal volume or in the more precise
and severity of cardiotoxicity. Research Sponsor: None. categorization into its subfields in either the intervention or control group of breast cancer
patients. Contrary to hypotheses suggested in the literature, it was shown that breast cancer
therapy does not lead to morphological changes in the hippocampus, indicating that CRCI is
not based on morphological damage. Clinical trial information: NCT04789187. Research
Sponsor: Verein zur Forschungsförderung der Krebshilfe OÖ.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
16s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

545 Poster Session 546 Poster Session

Effect of a multidisciplinary intervention based on a supervised training Survival impact of adjuvant chemotherapy regimens for small (T1mi/a/b),
program on cardiovascular risk and quality of life in early stage breast cancer node-negative (pN0), triple-negative breast cancer (TNBC). First Author:
patients. First Author: Maria Torrente, Hospital Universitario Puerta de Hierro-Maja- Kai Conrad Cecil Johnson, The Ohio State University - James Comprehensive Cancer
dahonda, Madrid, Spain Center, Columbus, OH
Background: There is growing evidence that physical activity can enhance cancer care. Background: While triple negative breast cancer (TNBC) represents an aggressive
Exercise programs have been shown to help manage treatment side effects, improve subtype of breast cancer worldwide, clinically low-risk cases are still encountered. When
functional outcomes, enhance overall quality of life, and reduce fatigue. Additionally, managing such cases, questions remain regarding the true added benefit of adjuvant
obesity and cardiovascular disease are among the most common comorbidities in breast chemotherapy on survival outcomes given the absence of dedicated prospective trials
cancer survivors. However, despite these benefits, exercise is still not widely prescribed for this population. Additionally, in the event that systemic therapy is pursued for these
to oncology patients. A multidisciplinary approach involving various healthcare pro- patients, it remains unclear whether anthracycline-containing treatments lead to im-
fessionals is crucial to ensuring that exercise interventions are tailored to individual proved long-term outcomes. Methods: Given the rarity of stage 1 triple negative breast
needs. Our study aims to determine whether a 12-week exercise intervention can cancer, we extracted recurrence & survival data from within the ASCO-developed
improve physical fitness and reduce cardiovascular risk in patients with early-stage CancerLinQ database to perform an in-depth retrospective analysis involving women
breast cancer after completing oncologic treatment. For the first time, impact is with small (pT1mi/a/b), node-negative (pN0), TNBC who underwent curative breast
assessed by measuring cardiorespiratory fitness (VO₂ max) and muscle strength, in- surgery and were diagnosed between the years 2002-2023. For the adjuvant chemo-
cluding range of motion, speed, and power. Methods: A total of 75 women with his- therapy recipients, only those who received a regimen of either taxane chemotherapy
tologically confirmed Stage I-III primary breast cancer who have recently completed all plus cyclophosphamide (TC) or an anthracycline & cyclophosphamide combined fol-
cancer-related treatments were included. Through computer-generated simple ran- lowed by taxane chemotherapy (AC-T) were included in this study. Our co-primary
domization, participants were assigned to resistance training (RTG; two sessions/week objectives were to compare the invasive recurrence-free survival (iRFS) and overall
for 12 weeks) or control (CG; general physical activity guidelines recommendations). survival (OS) of patients who received adjuvant TC or AC-T versus those receiving
Outcomes were evaluated at baseline and week 12. Muscular strength ( including range locoregional therapy alone. Our secondary outcome included an iRFS comparison
of motion, speed, and power) was the primary outcome. Secondary outcomes included between AC-T, TC, & locoregional therapy. Clinicopathologic variables were compared
cardiorespiratory fitness (measured by VO₂ max-maximum rate of oxygen consumption with appropriate tests for the categorical and continuous variables. Results: Among the
attainable during exercise), cancer-related fatigue and HRQoL. All participants had 159 patients identified with T1mi/a/b N0 TNBC who met inclusion criteria, 42 had
Performance status 0 or 1 and completed the EUROQOL-5D 5L and EORTC-QLQ-C30 QoL undergone locoregional therapy alone, 77 had received TC chemotherapy, and 40 re-
online survey. Results: The expected number of 75 patients was enroled in the study ceived AC-T. Baseline demographics found that the locoregional group had a higher
(mean age 55.9 6 7.4 years, all female). Patients assigned to the intervention had a proportion of T1mi/a vs T1b patients (p , 0.001) & a higher average age (p , 0.002). No
significant positive change in HRQoL total score [mean difference 3.8; 95% confidence differences were seen between groups in terms of germline mutations (BRCA1, BRCA1,
interval (CI) 0.2; 7.3; P = .038], body mass index [mean difference 20.7 kg/m2 (95% PALB2, CHEK2, & ATM), tumor grade, lymphovascular invasion, surgery type, race,
CI 21.3; 20.1); P = .022], muscle strength [mean difference 2.5 (95% CI 0.1; 5); P = .044; ethnicity, or average body-mass index. After a median follow up period of 57.2 months
effect size 0.39], and cardiorespiratory fitness (VO2 max) [mean difference 2.7 (95% CI overall, we found there was a significant benefit in both iRFS (HR 2.52, 95% CI 1.1-5.83, p
0.8; 4.6); P = .007]. No significant changes were observed in the control group between = 0.025) & OS (HR 6.95, 95% CI 1.62-29.79, p = 0.0027) for those who received adjuvant
week 0 and 12. Conclusions: This 12-week supervised exercise-based programme chemotherapy (TC or AC-T) compared to locoregional therapy alone. The 5-year iRFS
improved HRQoL, body mass index, muscle strength, range of motion, and power in loads was 89.9% with AC-T, 77.1% with TC, & 69.1% with locoregional therapy, whereas the 5-
while notably enhancing cardiorespiratory fitness. Integrating exercise into standard year OS was 96.9%, 96.3%, 85.8%, respectively. Conclusions: These findings suggest
healthcare practice can significantly improve patient’s quality of life and reduce car- that a recurrence & survival benefit is seen with the application of adjuvant chemo-
diovascular risk. Research Sponsor: SPANISH SOCIETY OF MEDICAL ONCOLOGY. therapy, even among this clinically low-risk population. However, whether it needs to be
AC-T or TC appears less significant. Research Sponsor: None.

547 Poster Session 548 Poster Session

Obesity, chemotherapy dosing, and toxicity: Results from the Optimal Clinical outcomes of patients with stage I triple-negative breast cancer
Breast Cancer Chemotherapy Dosing study. First Author: Elizabeth Kantor, (TNBC) treated with or without chemotherapy: The Mayo Clinic experience.
Memorial Sloan Kettering Cancer Center, New York, NY First Author: Sumeet Kumar Yadav, Mayo Clinic Health System, Mankato, MN
Background: ASCO guidelines state that most cytotoxic drugs should be dosed according to Background: TNBC is associated with higher risk of recurrence and poorer survival rates
full body surface area (BSA) without limiting the dose on the basis of obesity. This approach is than other breast cancer subtypes. For node-positive TNBC or tumors larger than 0.5 cm,
supported by a lack of evidence to suggest that patients with obesity, when fully-dosed, chemotherapy is generally recommended. For stage I TNBC, the benefit from chemotherapy,
experience higher risk of toxicity. Indeed, historical evidence suggests that obese, fully-dosed the optimal regimen, and whether to administer it in the neoadjuvant versus adjuvant setting
patients may experience lower risk of neutropenia than normal weight patients. However, remain controversial. Here, we report the treatment patterns and clinical outcomes of
questions remain regarding the representativeness of historical trial data on which this ev- patients with stage I TNBC treated at Mayo Clinic. Methods: Using the Mayo Clinic Tumor
idence is based. We examined this issue in the Optimal Breast Cancer Chemotherapy Dosing Registry data, we identified patients with stage I TNBC treated between 2010 and 2021. We
(OBCD) Study. Methods: The OBCD Study is a real-world cohort of 34,109 women diagnosed used the pathologic tumor (T) category for patients treated with upfront surgery and the
with stage I-IIIA breast cancer at Kaiser Permanente Northern California and Kaiser Per- clinical T category for patients receiving neoadjuvant chemotherapy. We used the Kaplan-
manente Washington between 2004-2019. Among women receiving the full dose of che- Meier method and log-rank test to compare recurrence-free survival (RFS) according to
motherapy at treatment initiation ($90% of intended dose, n = 7,644), we examined risk of treatment groups, measured from the day of surgery. RFS was reported at a median follow-
toxicities in women with obesity (BMI $30 kg/m2) compared to non-obese women (BMI 18.5- up of 3.9, 6.2, and 7.3 years for patients who received chemotherapy neoadjuvantly,
, 30 kg/m2). We examined hematologic (neutropenia, anemia, thrombocytopenia) and non- adjuvantly, and surgery alone with no chemotherapy, respectively. Results: A total of 602
hematologic (nephrotoxicity, hepatotoxicity, neuropathy, and cardiotoxicity) toxicities. Hazard patients with Stage I TNBC were included, with a median age of 62 years. 290 (48%)
ratios (HR) and 95% confidence intervals (CI) were calculated using Cox proportional hazards underwent upfront surgery followed by adjuvant chemotherapy (pT1a: 2%, pT1b: 25%, pT1c:
regression, adjusted for covariates including prevalent comorbid conditions. Secondary an- 74%), 127 (21%) received neoadjuvant chemotherapy followed by surgery (cT1a: 2%, cT1b:
alyses examined associations pertaining to specific BMI groups and also stratified by ad- 11%, cT1c: 87%), and 185 (31%) underwent primary surgery without any chemotherapy
ministration schedule (standard vs dose-dense). Results: Fully-dosed patients with obesity
(pT1mi/T1a: 33%, pT1b: 32%, pT1c: 35%). Most patients treated with neoadjuvant therapy
experienced lower risk of neutropenia (HR: 0.80; 95% CI: 0.73-0.88) and any hematologic
received anthracycline/cyclophosphamide + taxane (70%), while most patients treated in
toxicity (HR: 0.83; 95% CI: 0.75-0.91) but increased risk of neuropathy (HR: 1.34; 95% CI: 1.18-
the adjuvant setting received a non-anthracycline containing regimen (60%). The 5-year RFS
1.52), cardiotoxicity (HR: 2.30; 95% CI: 1.13-4.67), and non-hematologic toxicities overall (HR:
was 96% (95% CI: 93-100%) for patients who received chemotherapy neoadjuvantly, 95%
1.31; 95% CI: 1.15-1.48). The strength of these associations increased with increasing BMI
category. The inverse association between obesity and hematologic toxicity was evident for
(95% CI: 92-98%) for those who received chemotherapy adjuvantly, and 85% (95% CI: 80-
standard administration schedules (HR: 0.54; 95% CI: 0.45-0.66) but not dose-dense 91%) for those treated with surgery alone and no chemotherapy (P , 0.0001). 71 (56%) of
schedules. However, the positive association between obesity and non-hematologic toxic- the 127 patients who received neoadjuvant therapy achieved a pCR, with only 1 RFS event
ities persisted regardless of administration schedule. Conclusions: Women with obesity given (at 14 months) in that group, compared to 4 RFS events among those not achieving pCR (at
the full BSA-determined chemotherapy dose are less likely to experience neutropenia than 12, 20, 28 and 75 months, respectively). Among patients not receiving chemotherapy (either
fully-dosed non-obese women. Importantly, this holds among patients with more severe adjuvantly or neoadjuvantly), the 5-year RFS rates were 94% for pT1mi/T1a, 92% for pT1b
obesity, but not when restricted to newer dose-dense administration schedules. Findings also and 72% for pT1c (T1mi/a/b vs T1c, P = 0.009). Conclusions: In this large cohort of stage I
suggest that fully-dosed patients with obesity may experience higher risks for neuropathy and TNBC, patients who received chemotherapy (adjuvantly or neoadjuvantly) had better 5-year
cardiotoxicity. These findings highlight the importance of better understanding the risks and RFS compared with those treated with locoregional therapy only. Among patients receiving
benefits of dosing strategies as treatments and patient populations continue to evolve. chemotherapy, 5-year RFS was nearly identical regardless of whether chemotherapy was
Research Sponsor: National Cancer Institute; R37CA222793; National Cancer Institute; administered before or after surgery. Interestingly, patients undergoing upfront surgery
U24CA171524; National Cancer Institute; U01CA195565; National Cancer Institute; were more likely to receive an anthracycline-sparing chemotherapy regimen. Research
P30CA008748; National Cancer Institute; P01CA154292; Geoffrey Beene Cancer Research Sponsor: Mayo Clinic Breast Cancer SPORE P50CA 116201; ASCO, the Conquer Cancer -
Center at Memorial Sloan Kettering Cancer Center; National Cancer Institute; R50CA211115. Breast Cancer Research Foundation Advanced Clinical Research Award.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 17s

549 Poster Session 550 Poster Session

Clinical validation of a multi-modal Ataraxis AI platform for recurrence Pathologic complete response to neoadjuvant chemotherapy in early-stage
prediction in early-stage breast cancer across multiple patient cohorts. male breast cancer across molecular subtypes and racial/ethnic groups. First
First Author: Jan Witowski, Ataraxis AI, New York, NY Author: Jincong Q. Freeman, Department of Public Health Sciences, The University of
Background: Breast cancer (BC) treatment selection is traditionally guided by clinical Chicago, Chicago, IL
characteristics. However, as clinical characteristics cannot capture the complexity of a Background: Male breast cancer (mBC) accounts for ~1.0% of all breast cancers in the U.S.
disease, genomic tools have been developed. Recent advances in artificial intelligence Neoadjuvant chemotherapy (NACT) is often used to downsize locally advanced tumors and/
(AI) have allowed pathology imaging to be used to build more accurate and compre- or allow for lumpectomy in early-stage breast cancer. However, data on pathologic complete
hensive prognostic/predictive models. In this study, we validated an AI test, powered response (pCR) after NACT in mBC is scarce. This study aimed to explore how pCR in male
by a pan-cancer histopathology foundation model, that integrates digital pathology patients with early-stage breast cancer differed by molecular subtype and by race/ethnicity.
images with clinical variables to predict breast cancer recurrence. Methods: The Methods: This retrospective study analyzed data from the 2004-2021 U.S. National Cancer
Ataraxis AI prognostic model (ATX) was developed using 4,659 stage I-III BC patients Database registry. Patients were eligible if they were male sex, aged $18 years, diagnosed
from 10 distinct cohorts. Ataraxis AI platform first extracts novel morphological features with stage I-III disease, and underwent NACT. pCR (achieved/did not achieve) was defined as
from digitized H&E slides using a pre-trained AI foundation model. These morphological ypT0/TisypN0. Molecular subtypes included HR+/HER2–, HR+/HER2+, HR–/HER2+, and
features are then integrated with common clinical characteristics, such as TNM staging, TNBC. Racial/ethnic groups included Asian or Pacific Islander, Black, Hispanic, White, and
ER/PR/HER2 status, age at diagnosis, or lobular or ductal histology to generate a risk Other. We performed multivariable logistic regression, adjusting for age, race/ethnicity,
molecular subtype, clinical T/N, and tumor grade. Results: Of 1428 patients, the mean age
score between 0 and 1. We evaluated ATX on 3,502 patients from 5 external cohorts,
was 58.5 years (SD=12.8); 69.3% identified as White, followed by 18.8% as Black, 6.2% as
including 858 patients with available Oncotype DX (ODX) scores. The primary endpoint of
Hispanic, 3.4% as Asian or Pacific Islander, and 2.4% as Other. Most (87.2%) patients had
this study was disease-free interval (DFI), defined as the time until first recurrence, with
invasive ductal carcinoma. 51.3% were HR+/HER2–, 31.9% were HR+/HER2+, 11.7% were
deaths prior to recurrence censored. Results: Across 3,502 patients spanning five TNBC, and 5.1% were HR–/HER2+. Overall, the rate of pCR was 10.9%. Patients with HR–/
validation cohorts, ATX accurately predicted DFI with a C-index of 0.71 [0.68-0.75] and HER2+ tumors achieved the highest pCR rate (37.3%) compared to 33.8% with TNBC, 14.9%
hazard ratio (HR) of 3.63 [3.02-4.37, p , 0.01], computed for every 0.2 unit increase in with HR+/HER+, and only 3.7% with HR+/HER2– tumors (p,.001). The pCR rate trended
the test score. Compared to ODX (n = 858), the ATX was more accurate, achieving a C- higher in Asian or Pacific Islander (14.6%) or Hispanic (13.6%) patients than in Black
index of 0.67 [0.61-0.74] versus 0.61 [0.49-0.73]. Additionally, ATX added independent (11.2%), White (10.4%), or Other (8.8%) patients, though not statistically significant
prognostic information to ODX in a multivariate analysis (HR: 3.11 [1.91-5.09, p , 0.01]). (p=.728). On multivariable regression analysis, patients with HR+/HER2+ (adjusted odds
ATX demonstrated robust accuracy in TNBC (n = 230, C-index: 0.71 [0.62-0.81], HR: 3.81 ratio [aOR] 3.89, 95% CI: 2.24-6.76; p,.001), TNBC (aOR 8.80, 95% CI: 4.76-16.28; p,.001),
[2.35-6.17, p = 0.02]) and HER2+ (n = 353, C-index: 0.67 [0.55-0.80], HR: 2.22 [0.99-5.01, or HR–/HER2+ (aOR 13.45, 95% CI: 6.40-28.28; p,.001) tumors had greater odds of having
p = 0.05]) groups. Conclusions: (1) ATX is predictive of breast cancer recurrence, (2) achieved pCR than those with HR+/HER2– tumors. No significant differences in odds of pCR
ATX improves upon the accuracy of ODX, (3) ATX demonstrates robust performance in by race/ethnicity were found. Additionally, older age (aOR 0.84 [per 10-year increase], 95%
all main BC subtypes. Research Sponsor: Ataraxis AI. CI: 0.72-0.98; p=.026) and grade 1/2 (vs grade 3) tumors (aOR 0.39, 95% CI: 0.25-0.60;
ATX evaluated across 5 cohorts individually and pooled, for both Harrell’s C-index p,.001) were associated with lower odds of pCR. Conclusions: In early-stage mBC, the
and hazard ratio. post-NACT pCR rate varied significantly across molecular subtypes, with the lowest rate in
HR+/HER2– tumors, mirroring patterns observed in female breast cancer in the neoadjuvant
Cohort N C-index HR
setting. pCR rates were similar by race/ethnicity but lower among patients who were older or
Karmanos 168 0.62 [0.49-0.75] 3.82 [1.33-10.98, p=0.01] had low-grade tumors. These data suggest pCR dependence on tumor biology and could
Basel 269 0.67 [0.58-0.77] 3.98 [1.92-8.25, p,0.01] help neoadjuvant treatment selection to achieve optimal outcomes for early-stage mBC.
TCGA 911 0.70 [0.63-0.77] 3.0 [2.1-4.28, p,0.01] Future research could investigate survival outcomes by pCR in this mBC population.
Providence 1733 0.74 [0.7-0.79] 4.02 [3.09-5.23, p,0.01] Research Sponsor: National Institute on Aging; T32AG000243; Susan G. Komen Breast
Chicago 421 0.70 [0.60-0.80] 3.25 [1.45-7.31, p,0.01]
Cancer Foundation; TREND21675016.
Pooled 3502 0.71 [0.68-0.75] 3.63 [3.02-4.37, p<0.01]

551 Poster Session 552 Poster Session

Artificial intelligence (AI) based spatial assessment of tumor- Ultrasensitive circulating tumor DNA (ctDNA) detection for prognostication
infiltrating lymphocytes (TIL) and pathologic complete response in early in triple-negative breast cancer (TNBC) post-neoadjuvant chemotherapy
HER2+ breast cancer (BC): Secondary analysis of NSABP B-41. First Author: (NAC). First Author: Luc Cabel, Institut Curie, Paris and Saint-Cloud, France
Ilana Schlam, Dana-Farber Cancer Institute, Boston, MA Background: NAC +/- immune checkpoint inhibitors is the standard of care for most
Background: Manual quantitative assessment of stromal TILs has shown promise as a biomarker in early-stage TNBC patients. After NAC and breast surgery, adjuvant treatment decisions
HER2+ BC. We present an AI-powered single-cell TIL assessment. Methods: Manual TIL assessment rely on pathological complete response (pCR) status, which does not inform on the
was completed per guidelines. Zero-shot, AI-powered pipeline (Case45) was used to analyze tumor presence of distant micrometastases. Blood-based assays for ctDNA enable non-invasive
microenvironment (TME) from H&E slides, focusing on TILs and their spatial interplay with cancer
monitoring of residual disease levels at sensitivities down to 1 part-per-million (PPM). We
cells. The algorithm identified all cells, deriving three metrics: pct_lymphocyte (lymphocytes/total
cells), AI_TIL (adjacent-tumor lymphocyte to stromal cell ratio), hotspot_immune (normalized fraction report the prognostic value of ultrasensitive ctDNA detection during NAC in patients with
of immune cell aggregates in relation to cancer/tissue). Spearman correlation coefficients evaluated TNBC. Methods: Early-stage TNBC patients treated with NAC in the SCANDARE pro-
correlations; logistic regression models assessed the relationship between TIL measurements and spective study were evaluated for ctDNA before, during, after NAC/pre-surgery and during
pCR, with and without gene expression adjustments. AUC assessed predictive performance, and post-surgical follow-up. Plasma ctDNA was profiled using NeXT Personal (Personalis), an
univariate Cox models examined TILs’ association with event-free survival (EFS). Results: Our an- ultrasensitive tumor-informed MRD assay that tracks up to 1,800 patient tumor-specific
alyses included tumors of 262 patients with early-stage HER2+ BC, 67% estrogen receptor (ER) variants based on whole genome sequencing to attain sensitivity down to 1-3 PPM
positive, 51% positive lymph nodes. Poor histologic grade (p,0.001), non-luminal (p=0.006), and ER- with .99.9% specificity. Results: Plasma ctDNA was analyzed for 279 samples from 84
tumors (p=0.003) were associated with higher manual TILs. Manual TILs were moderately associated TNBC patients (Stage I: 2%, II: 77%, III: 20%), of whom 35 (42%) achieved pCR after NAC.
with pct_lymphocyte (r= 0.34) and AI_TIL (r= 0.43). Perthe table, manual TILs were positively as-
sociated with pCR, the association was numerically stronger in ER- tumors (Interaction p=0.38).
After a median follow-up of 53 months, 16 patients (19%) developed distant metastases,
pct_lymphocyte and AI_TIL were positively associated with pCR, regardless of ER status. hotspo- and 18 patients (21%) died. ctDNA was detected before NAC in all 82 patients with an
t_immune was strongly associated with pCR (OR=1.26 for all, 1.29 in ER-, 1.22 in ER+, p=,0.001). TILs available sample (median=3461 PPM, IQR: 1168-22078), with pretreatment levels in the
and ESR1 and ERBB2 provided complementary prognostic utility in pCR in trastuzumab-treated ultrasensitive range (,100 PPM) in 12%. Most ctDNA detections during (51%) and post-
patients (AUC: 0.699-0.757). Among all subjects, there was no association between manual TILs and NAC (55%) were ,100 PPM. Patients with rates of early on-treatment ctDNA reduction
EFS (p=0.2); there was a marginal association between AI_TIL and EFS (p=0.06). Conclusions: The faster than the median, or clearance, had significantly improved distant relapse-free
spatial characterization of TILs using an AI-powered tool shows promise as a prognostic biomarker in interval (DRFI, log-rank P=7.7x10-3) and overall survival (OS, log-rank P=7.9x10-3). Pa-
both HER2+/ER+ and HER2+/ER- BC, manual TIL assessment is prognostic in HER2+/ER- BC. The tients with post-NAC, pre-surgery ctDNA clearance had significantly improved DRFI (log-
assessment of immune aggregates appears to be highly predictive of pCR. Further validation through
rank P=3.5x10-7) and OS (log-rank P=1.4x10-11), and were enriched for pCR (59% vs 9%,
prospective-retrospective studies, focused on the spatial immune heterogeneity in the TME, is required
before integrating these biomarkers into routine clinical practice. Clinical trial information: OR=13.9, 95% CI [2.8,137.2], Fisher’s exact P=1.4x10-4). Multivariable Cox models in-
NCT00486668. Research Sponsor: BCRF; BCRF 21-156. cluding pCR and ctDNA detection post-NAC, pre-surgery performed significantly better
than models including only pCR for predicting DRFI (LRT P=8.0x10-4) and OS (LRT
TILs measurements and pCR.
P=6.6x10-4). ctDNA detection significantly stratified survival outcomes among patients
Variable (continuous) Cohort OR (95% CI) p-value
without pCR (DRFI HR=8.2, 95% CI [1.8,37.0], Cox P=6x10-3), with 60% of ctDNA-positive
Manual TILs % All 1.13 (1.04, 1.23) 0.004 patients developing distant metastases vs 10% of ctDNA-negative patients. During follow-
(10-unit inc.) ER- 1.16 (1.02, 1.31) 0.02
ER+ 1.07 (0.95, 1.21) 0.27
up, all plasma samples from non-relapsing patients were ctDNA-negative. In relapsed
Percentage of Lymphocyte All 2.00 (1.30, 3.07) 0.002 patients, ctDNA was detected in 95% of samples collected at relapse or tumor pro-
(10-unit inc.) ER- 1.75 (0.95, 3.21) 0.07 gression. Conclusions: Ultrasensitive ctDNA detection informs on the outcome of early
ER+ 1.93 (1.02, 3.62) 0.04 TNBC treated by NAC, independently of pCR status. These results, obtained with samples
AI TILs All 1.22 (1.06, 1.40) 0.005
(one-tenth inc.) ER- 1.19 (0.98, 1.44) 0.09 taken post-NAC but before-surgery warrant investigating the benefit of implementing
ER+ 1.19 (0.97, 1.46) 0.10 ctDNA detection in an interventional setting. Research Sponsor: Agence National de la
Recherche; Site de recherche intégré contre le cancer (SiRIC).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
18s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

553 Poster Session 554 Poster Session

Association of immune, proliferation gene signatures and stromal tumor Association of ImPrintTN signature with survival outcomes by race in basal-
infiltrating lymphocytes (sTILs) with outcomes in patients with stage I triple- type triple negative breast cancer (TNBC): FLEX registry analysis. First Author:
negative breast cancer (TNBC). First Author: Paolo Tarantino, Dana-Farber Cancer Priyanka Sharma, University of Kansas Medical Center, Westwood, KS
Institute, Boston, MA Background: ImPrintTN is a triple negative breast cancer (TNBC) immune classifier
Background: Approximately one third of all TNBC diagnoses are stage 1. No validated biomarker is signature that has been associated with pathological complete response (pCR) to im-
routinely utilized to guide treatment at this early stage. Methods: Samples from patients with stage I TNBC munotherapy (IO) plus chemotherapy in ISPY2. Real Word data (RWD) from FLEX was
or ER-low (1-10%) breast cancer undergoing surgery at Dana-Farber/Brigham Cancer Center between 2016 utilized to assess the association of ImPrintTN with self-reported race and its impact on
and 2021 were identified. The 10-gene Core Immune Gene (CIG) signature and the 4-gene proliferation
signature (both part of the TNBC-DX tool) were derived from extracted RNA. Central evaluation of sTILs was clinical outcomes in early stage TNBC. Methods: Patients (pts) enrolled in the FLEX
conducted at Dana-Farber, with 5% and 20% used as thresholds. All markers were tested for prediction of (NCT03053193) trial diagnosed with early-stage TNBC and BluePrint Basal molecular
recurrence free survival (RFS) using the Kaplan-Meier method. Results: We identified 253 patients with subtype with available survival data, who self-identified as Black or White, were eligible for
stage I TNBC (n=218) or ER-low tumors (n=35) treated at Dana-Farber. Median age was 61 (31 – 85), most this analysis. ImPrintTN results (+/-) were acquired through whole transcriptome profiling.
tumors were ductal (89%), high-grade (73%), 48% were .1 cm and 65% received chemotherapy. 5-year RFS Chi-squared and Fisher’s exact tests assessed differences in clinical characteristics.
in the overall cohort was 86.8%, with numerical variation by tumor size (T1a 100%, T1b 93.8%, T1c 81.7%, Association of pCR outcomes and ImPrintTN+/- were tested by binary logistic regression.
p=0.26). Gene signatures and sTILs were obtained for 117 and 123 patients, respectively (both for 110
patients), with their association with outcomes described in Table 1. Median follow-up was 3 years. A total
Recurrence-free survival (RFS) was compared between race and ImPrintTN+/- using
of 20/117 patients (17.1%) had medium-high CIG score, with none experiencing RFS events prior to year 5. Kaplan-Meier estimates and log rank tests. Cox proportional hazards model was used to
Similarly, no recurrence was observed prior to year 5 in 29 patients (24.8%) at the upper CIG quartile (vs 83- analyze the association of ImPrintTN, race, and clinical features with RFS.
88% 5-year RFS in other quartiles). Worse outcomes were seen among patients in the upper quartile of Results: Among 279 eligible patients with early stage Basal TNBC, 23.7% were Black,
proliferation (5-year RFS 83%, vs 88-100% in other quartiles). Overall, 33/123 patients (26.8%) had high 76.3% were White, 27.7% had node positive disease, 49.8% received neoadjuvant therapy,
sTILs (.20%) and experienced the highest 5-year RFS (97%, vs 78% if low sTILs). OS data will be presented. 47.3% adjuvant therapy, 2.5% IO, and median follow-up was 3 years. 56.6% of pts were
Conclusions: High expression of the 10-gene CIG immune signature or high sTILs are associated with ImPrintTN+, similarly distributed by race (Black: 60.6%, White: 55.4%, p=0.761). Among pts
numerically improved outcomes in patients with stage I TNBC that did not reach statistical significance,
warranting further study as prognostic tools. Research Sponsor: None.
treated with neoadjuvant therapy (n=139) no significant differences in pCR rates were
observed by race (Black: 26.5%; White: 35.2%; p=0.46). However, a higher pCR rate was
3- and 5-year recurrence free survival (RFS) according to gene signatures and sTILs. P values were
obtained via Cox proportional hazard models.
achieved in ImPrintTN+ vs ImPrintTN- cancers (39.3% vs 22.0%; OR=2.29, 95% CI [1.04-
N 3-year RFS 5-year RFS
5.08]; p=0.039). The 3-year RFS was similar for Black (82.5%) and White (83.5%; p=0.91)
pts. Significantly improved 3-year RFS was associated with ImPrintTN+ (87.9%) compared
CIG score
- Low 97 91% (85%, 98%) 89% (80%, 98%) to ImPrintTN- (77.5%; p=0.01). Among ImPrintTN+, RFS was similar for Black (89.7%) and
- Med-High 20 100% (100%, 100%) 100% (100%, 100%) White (87.3%; p=0.30) pts. However, ImPrintTN- observed a trend towards lower 3-year
CIG score (quartiles) RFS in Black (71.1%) compared with White (79.2%; p=0.24) pts. In a multivariate model,
- £25% 49 94% (87%, 100%) 88% (74%, 100%)
- 25-50% 10 83% (58%, 100%) 83% (58%, 100%) RFS probability was significantly associated with ImPrintTN (ImPrintTN+ vs ImPrintTN-:
- 50-76% 29 87% (74%, 100%) 87% (74%, 100%) HR= 0.41, 95% CI [0.22-0.75]; p=0.004) and nodal status (LN+ vs LN-: HR= 2.98, 95% CI
- >75% 29 100% (100%, 100%) 100% (100%, 100%)
Proliferation score [1.66-5.37,]; p,0.001), while race and neo/adjuvant therapy were not. Conclusions: The
- Low 58 96% (89%, 100%) 90% (78%, 100%) analysis found that 56.6% of Basal TNBCs in the FLEX trial were ImPrintTN+, with similar
- Med-High 59 90% (82%, 99%) 90% (82%, 99%)
Proliferation score (quartiles)
proportions observed among Black and White pts. ImPrintTN status was prognostic for
- £25% 30 100% (100%, 100%) 88% (67%, 100%) both pCR and RFS in TNBC and was associated with significantly improved RFS across
- 25-50% 29 90% (77%, 100%) 90% (77%, 100%) racial groups. However, the negative prognostic impact of ImPrintTN- appeared more
- 50-76% 29 100% (100%, 100%) 100% (100%, 100%)
- >75% 29 83% (68%, 100%) 83% (68%, 100%) pronounced among Black compared with White pts. Ongoing research is focused on
sTILs exploring biological differences within the ImPrintTN- subgroup by race. Clinical trial
- 1-5% 62 94% (87%, 100%) 78% (63%, 98%)
- >5-20% 28 91% (81%, 100%) 91% (81%, 100%)
information: NCT03053193. Research Sponsor: Agendia, Inc.
- >20% 33 97% (90%, 100%) 97% (90%, 100%)

555 Poster Session 556 Poster Session

Real-world (rw) ctDNA testing trends and associated outcomes in patients Association of genetic predisposition to low-grade systemic inflammation
(pts) with early stage breast cancer (EBC). First Author: Erin Fidyk, Flatiron Health, with cancer-related fatigue in women receiving chemotherapy for non-
New York, NY metastatic breast cancer in URCC07012 and URCC10055. First Author: Ayo
Background: Emerging evidence from prospective studies underscores the prognostic potential of Olowofela, Department of Radiation Oncology, Medical College of Wisconsin, Milwaukee,
ctDNA to inform risk stratification and clinical decision-making in EBC (stage I-III). However, its use WI
and correlation with outcomes in routine clinical practice remains less understood. We describe Background: Cancer-related fatigue (CRF) is reported by ~75% of patients receiving che-
tumor-informed ctDNA testing trends and the association of test results with recurrence risk among
motherapy for breast cancer. CRF has been linked to inflammation. Chronic, low grade
pts with EBC in the rw setting. Methods: Pts with an EBC diagnosis (dx) after 1/1/2018 who had
documented HR/HER2 status at initial dx, surgery, and $1 commercial ctDNA test in the early stage
systemic inflammation is a polygenic trait, and a polygenic risk score for inflammation (iPRS)
setting were selected using machine learning models from the Flatiron Health US-nationwide EHR- might be associated with risk of CRF. Methods: Using data from the UK Biobank,we de-
derived, deidentified, longitudinal database of .750 000 pts with BC (data cutoff 8/31/2024). ctDNA veloped an iPRS using the INFLA-score, a composite measure of serum C-reactive protein,
positivity (ctDNA+) was defined as having $1 positive test in EBC. Baseline characteristics were white-cell count, platelet count, and neutrophil-lymphocyte ratio. The iPRS was evaluated for
stratified by EBC subtype and ctDNA status. To examine the association of ctDNA status with re- association with CRF among women with non-metastatic breast cancer enrolled in one of
currence, unadjusted Kaplan-Meier (KM) plots and adjusted Cox proportional hazards models were two completed multi-site clinical trials of the University of Rochester Cancer Center NCI
performed to assess recurrence-free survival among ctDNA-tested pts, controlling for age, race/ Community Oncology Research Program (NCORP) Research Base. CRF was measured before
ethnicity, stage, ECOG status, dx year, insurance status, practice type, and neoadjuvant/adjuvant and after standard-of-care chemotherapy using the Multidimensional Fatigue Symptom
treatment. Recurrence was indexed to surgery date and defined as locoregional or metastatic re- Inventory-Short Form (MFSI-SF). Linear regression evaluated the change in MFSI-SF score
currence or death. Results: In a cohort of 195 279 pts with EBC, 14 496 ctDNA tests were performed in from pre- to post-chemotherapy; logistic regression evaluated a binary outcome of any vs no
4639 pts (median 2 per pt) with most in stage I (43.3%) and II (37.1%). Testing prevalence was highest worsening of scores. Analyses were adjusted for patient and treatment factors. Results: The
in HR-/HER2- (4.9%), followed by HR-/HER2+ (3.5%), HR+/HER2+ (2.9%), and HR+/HER2- (1.9%).
NCORP cohort included 802 women who received chemotherapy (anthracycline-based =
Testing increased from 1.6% (n = 450) of EBC pts dx in 2020 to 4.25% (n = 1278) in 2023 with a
decrease in median time to first test pre- and post-2022 (35 vs 8 months respectively). Among tested 51.8%; previous surgery = 85.0%) at a median age of 55 years (range = 22 to 81). There was
pts, 921 (19.9%) had $1 positive test and were more likely to be younger (58 vs 64 years) and have an increase in MFSI in 55% of the women, with a mean increase of 8 (range=-64 to 71) from
stage III disease compared to non-tested pts. ctDNA+ patients had a worse 3-year overall survival (OS) prechemotherapy (mean=8, range = -24 to 83) to post-chemotherapy (mean=15.3, range =
as well as a strong association with recurrence (Table). Conclusions: In the largest rw study of ctDNA -24 - 88), indicating an overall increase in CRF. The iPRS was associated with a significant
testing in EBC to date, pts with ctDNA+ disease across all subtypes were more likely to recur, decrease in MFSI-SF (b=-3.29; 95%CI=-6.25 to -0.34; P=0.029; covariate-adjusted b=-2.71;
highlighting the potential prognostic value of ctDNA testing to inform pt counseling, monitoring and 95%CI=-5.50 to 0.08; P=0.057) and lower odds of worsening CRF (OR=0.66; 95%CI=0.47-
treatment strategies. These rw results, coupled with findings from prospective randomized trials, 0.93; P=0.016; covariate-adjusted OR=0.67; 95%CI=0.47 to 0.96; P=0.029). The negative
support the case for ctDNA+ as a distinct risk category in the management of EBC. Research Sponsor: relationship between the iPRS and change in CRF was partially explained by the finding that
Flatiron Health, Inc. women with an iPRS in the highest quartile have worse pre-chemotherapy MFSI-SF scores
Unadjusted 3-year OS probability (95% CI) (b=4.33; 95%CI=0.23 to 8.43; P=0.038). Conclusions: Women with genetic predisposition to
low-grade systemic inflammation, indicated by a higher iPRS, have worse CRF pre-
Adjusted HR (95% CI)
EBC Subtype ctDNA- pts ctDNA+ pts All with P <0.01 chemotherapy that does not worsen, and may improve, over the course of treatment
while women with a lower iPRS have less CRF pre-chemotherapy and are at greatest risk of
HR+/HER2- 0.98 (0.97-0.99) 0.76 (0.7-0.82) 10.7 (7.08-16.1) developing new or worsening CRF during treatment. If validated, the iPRS could identify
N = 2786
HR-/HER2- 0.96 (0.95-0.98) 0.61 (0.52-0.72) 10.7 (6.34-18.1) patients in need of supportive care interventions to reduce CRF. This work was supported
N = 1002 by the National Institutes of Health National Cancer Institute Contract No.
HR+/HER2+ 0.97 (0.95-0.99) 0.85 (0.77-0.94) 11.8 (4.54-30.8) HHSN261201500003I, Task Order No. HHSN261000039 and by UG1CA189961, URCC NCORP
N = 592
HR-/HER2+ 0.97 (0.94-1) 0.78 (0.62-0.97) 8.94 (1.72-46.4)
Research Base. Research Sponsor: National Cancer Institute/U.S. National Institutes of
N = 259 Health; Contract No. HHSN261201500003I, Task Order No. HHSN261000039.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 19s

557 Poster Session 558 Poster Session

Assessment of ovarian function suppression (OFS)-containing adjuvant Dynamic changes in circulating tumor DNA among Taiwanese early breast
endocrine therapy in premenopausal women by Breast Cancer Index. First cancer patients undergoing upfront surgery: Results from the VGH-TAYLOR
Author: Ruth O’Regan, University of Rochester, Rochester, NY study. First Author: Chi-Cheng Huang, Taipei Veterans Genreal Hospital, Taipei City,
Background: Breast Cancer Index (BCI) previously identified that premenopausal pa- Taiwan
tients with HOXB13/IL17BR ratio (H/I)-Low tumors derived greater benefit than BCI(H/I)- Background: Circulating tumor DNA (ctDNA) has emerged as a promising prognostic marker
High tumors from OFS-containing adjuvant endocrine therapy vs tamoxifen alone in the in breast cancer. Post-treatment ctDNA detection is associated with increased recurrence risk
Suppression of Ovarian Function Trial (SOFT). This translational study of the Tamoxifen and reduced long-term survival. This study evaluated dynamic ctDNA changes in early breast
and Exemestane Trial (TEXT) was conducted to assess the predictive benefit of BCI (H/I) cancer patients with distinct immunohistochemical (IHC) subtypes. Methods: Liquid biopsies
from exemestane (E) plus OFS over tamoxifen (T) plus OFS and validate the prognostic were performed using the Oncomine Breast cfDNA Assay v2. Samples were collected at
performance of BCI. Methods: Blinded BCI testing was performed in all available tumor baseline (pre-surgery, visit 1), after adjuvant therapy (visit 2), and every six months thereafter
(visit 3 and subsequent visits) for patients in the upfront surgery (Group 1A) arm of the VGH-
samples from patients enrolled in TEXT, of which 1782 of 2660 had BCI successfully
TAYLOR study. Pre-operative and follow-up ctDNA detectability and its impact on recurrence-
assessed and BCI categories assigned per established clinical cutpoints. Primary
free survival (RFS) were evaluated. Results: A total of 577 early breast cancer patients with at
endpoints were breast cancer–free interval (BCFI) for predictive and distant least one ctDNA test were analyzed; the majority (74%, n=425) were hormone receptor (HR)-
recurrence–free interval (DRFI) for prognostic analyses. Per pre-specified SAP, a positive/human epidermal growth factor receptor 2 (HER2)-negative. Among 500 pre-operative
secondary analysis of predictive benefit combined the two OFS arms common to TEXT samples, ctDNA was detected in 24% (n=121) of patients, with TP53 (21%, n=106) and PIK3CA
and SOFT (2896 of 4690 patients); clinicopathologic subgroup analyses were conducted (6%, n=28) being the most prevalent mutations. During follow-up (visit 2 and later), ctDNA was
in the combined TEXT+SOFT cohort. Cox proportional hazards models, stratified by detected in only 3% (n=13) of patients; all harbored TP53 mutations, with one case also
chemotherapy use and nodal status, that included treatment assignment, BCI(H/I) exhibiting an ERBB3 mutation. All patients with detectable follow-up ctDNA had also tested
status, and interaction term were used to assess BCI predictive performance by testing positive pre-operatively (Table 1). Five-year RFS was 94% in the ctDNA-positive group (n=121)
for treatment-by-BCI(H/I) interaction. The median follow-up was 13 years. and 95% in the ctDNA-negative group (n=319). Among HR-negative/HER2-positive and HR-
Results: Among TEXT patients, 58% had BCI(H/I)-Low [Link] with BCI (H/I)- negative/HER2-negative subtypes, ctDNA positivity was associated with numerically worse
Low tumors exhibited a 6.6% absolute benefit in 12-year BCFI (HR=0.61 [95% CI, 0.44- RFS (90% vs. 94% and 88% vs. 89%, respectively). Conclusions: Following surgery and
0.85]) for E+OFS versus T+OFS while those with BCI(H/I)-High tumors had an 6.3% adjuvant therapy, most pre-operative ctDNA-positive cases became undetectable (89%, 108/
absolute benefit (HR=0.78 [95% CI, 0.57-1.07]) (P-interaction = 0.29). Results were 121). Although ctDNA positivity showed a trend toward compromised RFS, particularly in HR-
consistent in the combined TEXT+SOFT cohort and adjusting for clinicopathological negative/HER2-positive and HR-negative/HER2-negative subtypes with TP53 mutations, the
variables. Clinical subgroup analyses consistently showed benefit of E+OFS vs T+OFS high clearance rate of pre-surgery ctDNA positivity warrants longer follow-up to fully evaluate
its prognostic value and the impact of liquid biopsy in early breast cancer. Clinical trial in-
for BCI(H/I)-Low tumors, and more variable relative treatment effects among BCI(H/I)-
formation: NCT04626440. Research Sponsor: Yonglin HealthCare Foundation.
High tumors, including by age. Post-hoc exploratory time-varying estimates suggested
the treatment-by-BCI relationships may differ in years 0-5 vs .5 years. BCI and BCIN+ as Dynamic changes of circulating tumor DNA before surgery and after treatment across
immunohistochemical subtypes among early breast cancer.
continuous indices were prognostic for distant recurrence in N0 (P = 0.0004) and N1
(P , 0.0001) cancers. The 12-year DRFI was 96.3%, 90.3% and 84.9% for BCI-low, Subtype Pre-surgery ctDNA positive Post-treatment/follow-up ctDNA positive
intermediate and high-risk N0 cancers, respectively. Conclusions: BCI was confirmed as HR+/HER2+ 26%(9/35) 0%(0/35)
HR+/HER2- 22.8%(84/369) 1.9%(7/369)
prognostic in premenopausal women with HR+ early breast cancer enrolled in TEXT. HR-/HER2+ 25%(7/28) 7.1%(2/28)
BCI(H/I) status did not clearly predict differential benefit from E+OFS vs T+OFS. The HR-/HER2- 30.8%(20/65) 6.2%(4/65)
TEXT results complement the prior results from SOFT, indicating premenopausal pa-
HR: hormone receptor, HER2: human epidermal growth factor receptor II, ctDNA: circulating tumor DNA.
tients with BCI(H/I)-Low tumors benefit from more intensive endocrine therapy.
Research Sponsor: Biotheranostics, Inc., a Hologic company.

559 Poster Session 560 Poster Session

Impact of germline BRCA status on clinical outcomes of patients with HR+/ Cadence of circulating tumor DNA (ctDNA) testing for molecular surveillance
HER2- early breast cancer. First Author: Antonio Marra, Division of Early Drug in early-stage breast cancer (eBC). First Author: Marla Lipsyc-Sharf, UCLA Health
Development for Innovative Therapies, European Institute of Oncology IRCCS, Milan, Jonsson Comprehensive Cancer Center, Los Angeles, CA
Italy Background: Post-surgical detection of molecular residual disease (MRD) via plasma
Background: Germline pathogenic variants (PVs) in the BRCA1 and BRCA2 (gBRCA1/2) ctDNA testing is strongly associated with recurrence of eBC. Our prior real-world data
genes increase the risk for breast cancer (BC) [Link] prognostic significance of suggests that adjuvant MRD testing impacts clinical care for most patients (pts) with
gBRCA1/2 in patients with hormone receptor-positive/HER2-negative (HR+/HER2) early BC positive ctDNA (ctDNA+) results. Ongoing trials are studying whether adjuvant ctDNA
is still controversial. Methods: This cohort study derived from a prospectively-maintained testing improves eBC outcomes. While serial testing is known to improve MRD detection
institutional database of all consecutive patients with BC who underwent germline testing, rates, the optimal cadence of testing is unknown. Here, we investigated the role of timing,
including BRCA1, BRCA2 and PALB2, at the European Institute of Oncology (May 2002-Jan cadence, and quantitative results of real-world ctDNA testing on detection of clinical
2024). The study population comprised patients with stage I-III HR+/HER2- (estrogen recurrence of eBC. Methods: We identified eBC pts with available recurrence-free survival
receptor expression .1%) invasive BC who underwent surgery and (neo)adjuvant treatment, data who had adjuvant plasma MRD testing via a clinically validated, personalized, tumor-
as endocrine therapy (ET) +/- chemotherapy (CT) (Jan 2000-Dec 2022). Primary endpoints informed mPCR-NGS ctDNA test (Signatera, Natera, Inc.). All tests were ordered in the
were distant relapse-free interval (DRFI) and invasive disease-free survival (iDFS) by STEEP United States in the real-world clinical setting between 2019-2024; clinical records were
2.0. Univariate and multivariate Cox proportional-hazard models were employed for survival reviewed. The cumulative incidence of clinical recurrence after each ctDNA test was used
analyses, with left-truncated models to account for the time from BC diagnosis to germline to calculate negative predictive value (NPV) for recurrence within 3, 6, 12, 18, 24, and
testing. Results: A total of 1,730 patients were included in the analyses, with 52 (3%) 30 months (mo) post-test. ctDNA levels at different timepoints prior to recurrence were
BRCA1, 180 (10%) BRCA2, and 9 (0.5%) PALB2 PV carriers. Compared to non-carriers, analyzed separately by ER and HER2 status. Results: For 819 pts with stage I-III eBC (ER+/
patients with gBRCA1/2 and gPALB2 PVs were younger (median age: 39 vs 42 yrs, p,.001), HER2-: 249, HER2+: 68, triple-negative [TNBC]: 502), there were 4689 total plasma samples
had advanced disease stage (stage II-III: 71% vs 58%, p,.001), higher tumor grade (G3: 54% obtained in the adjuvant setting (median 5.7 time points per patient). Median time of first
vs 26%, p,.001) and Ki-67 expression (median: 26% vs 20%, p,.001). Patients with adjuvant ctDNA testing was 7 mo after surgery (range: 0.1-214.9). Median follow-up was
gBRCA1/2 and gPALB2 PVs were also more likely to receive neoadjuvant (13% vs 6%, 18.1 mo (range: 0.7-239.4). For pts with ER+/HER2- tumors, median time to first test was
p,.001) and/or adjuvant CT (56% vs 36%, p,.001) and mastectomy (56% vs 45%, p=.002). 14.6 mo (range: 0.3-214.9) versus 16.1 mo (range: 0.3-151.8) in HER2+ and 7.1 mo (range:
All patients received adjuvant ET, as tamoxifen or aromatase inhibitor +/- GnRH analogue. 0.1-178.8) in TNBC. Among pts with multiple adjuvant ctDNA tests, median interval
No patient received adjuvant olaparib or CDK4/6 inhibitor. At a median follow-up of 9.7 (IQR between tests was 2.8 mo (consistent across subtypes). For ER+ and TNBC tumors NPV
6-13.9) years, 335 (19%) patients experienced local relapse, 316 (18%) distant metastasis, (95%CI) gradually decreased over time from 99.5% (98.8-99.8) and 99.7% (99.2-99.9) at 3
and 124 (7.2%) died due to BC. At multivariate analyses, gBRCA2 P/LPVs were independently mo to 97.7% (96.7-98.4) and 97.8% (96.6-98.5) at 30 mo, respectively. Among ctDNA+
associated with shorter DRFI (HR 1.46, 95%CI 1.04–2.06, p=.028) and iDFS (HR 1.34, 95 CI TNBC pts, quantitative ctDNA levels were higher among those who recurred , 3 mo after a
1.01–1.78, p=.045), regardless of stage, nodal status, (neo)adjuvant CT, type of surgery and +ctDNA test than those with recurrence in 3-6 mo; median ctDNA levels were lowest in pts
adjuvant ET, whereas gBRCA1 were not. Exploratory analyses showed that among 232 who recurred . 6 mos after a ctDNA+ test (median [range], 2.7 [0.03-1089.6] vs 2.0 [0.1-
gBRCA1/2 carriers, 47 (20%) and 96 (41%) were eligible for adjuvant olaparib or abemaciclib 15.3] vs 0.3 [0.3-1.7] MTM/mL, p = 0.0041 and p = 0.0071, respectively). This trend was less
therapy per OlympiA and monarchE criteria, respectively, with 37 (16%) eligible for both pronounced in ER+HER2- disease between pts who relapsed , 6 mo, 6-9 mo, and . 9 mo
therapies. Additional analyses to unravel interaction of gBRCA status with adjuvant after a ctDNA+ test (median [range], 11.6 [0.26-299.4] vs 4.5 [0.2-27.0] vs 3.8 [2.3-124.5]
treatment are underway. Conclusions: Patients with HR+/HER2- early BC harboring MTM/mL, p = 0.5, p = 0.8). Conclusions: The data from this real-world analysis of tumor-
gBRCA2 PVs had a significantly increased risk of recurrence, with a potentially distinct informed ctDNA testing in pts with eBC during surveillance demonstrate a high NPV for
impact of BRCA2 vs BRCA1. Only a small proportion of this population currently qualify to both ER+ and TNBC disease. These data guide future prospective ctDNA-guided studies
adjuvant treatment escalation with targeted therapies, underscoring the need of expanding aimed at therapeutic interception to improve clinical outcomes. Research Sponsor:
the therapeutic options in this setting. Research Sponsor: None. Conquer Cancer, the ASCO Foundation; Natera, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
20s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

561 Poster Session 562 Poster Session

Association of lifestyle factors and pathological characteristics in patients Association of infiltration of hematopoietic stem cells (HSC) with cell pro-
with early breast cancer and overweight/obesity: Results from the Breast liferation and patient survival in breast cancer. First Author: Masanori Oshi,
Cancer Weight Loss (BWEL) trial. First Author: Caterina Sposetti, Dana-Farber Roswell Park Comprehensive Cancer Center, Buffalo, NY
Cancer Institute, Boston, MA Background: HSCs, also known as blood stem cells, are self-renewable cells that can
Background: Obesity and other lifestyle factors are associated with breast cancer (BC) risk and outcomes. These develop into all types of blood cells. They are found in bone marrow and peripheral blood.
relationships appear to vary in pre- vs postmenopausal women. Here we explore associations between pathological
features, diet quality and physical activity (PA) in patients (pts) with BC enrolled in the BWEL (Alliance for Clinical Trials in However, the clinical relevance of HSCs in the BC TME remains [Link] elucidate
Oncology A011401) trial, stratifying for menopausal status. Methods: BWEL is a phase III trial evaluating the impact of a the clinical significance of HSC infiltration in the tumor microenvironment (TME) of
weight loss intervention on disease outcomes in 3180 pts with stage II-III HER2– BC and body mass index $27 kg/m2. breast cancer (BC). Methods: In silico analyses were conducted on 5,176 BC patients,
The first 542 BWEL pts underwent assessment of PA (7-Day PA Recall) and diet (24-hour Diet Recall) at enrollment.
Healthy Eating Index 2020 (HEI) score (0-100, higher value = healthier diet) and minutes/week (min/wk) of moderate/
including large independent cohorts; The Sweden Cancerome Analysis Network-Breast
vigorous PA (MVPA) were calculated. Estrogen receptor (ER) and progesterone receptor (PR) expression level (%) were (SCAN-B) and the Molecular Taxonomy of Breast Cancer International Consortium
abstracted from pathology reports. Analyses of links between HEI score, MVPA min/wk, % ER/PR, stratified by (METABRIC), as well as multiple single-cell sequenced cohorts. HSC were identified
menopausal status, were conducted using t-tests; differential associations by menopausal status used linear regression
models with interaction term. Results: In 523 pts with available pathology data, 76.5% had ER/PR+ BC, 56.2% were
through the xCell algorithm, and patients with high HSC levels were defined as those with
postmenopausal. Median HEI score was 50.1 (range 18.1-96.4), median MVPA min/wk was 0 (range 0-630), median time HSC expression above the median in each cohort. Results: Fraction of HSCs ranged
from diagnosis to enrollment was 10.0 months (range 2.4-13.1). In postmenopausal pts, higher diet quality was linked to from 0.04-0.50% of all cells in BC TME by single cell transcriptome analyses. HSC
increased % ER and PR (HEI above vs below median, mean % ER 78.0 vs 66.1, p=0.012; mean % PR 57.7 vs 43.1, p=0.004). infiltration was not correlated with its lineage cells, common myeloid progenitor cells
The relationship between HEI score and % ER/PR differed significantly by menopausal status (interaction term of
menopausal status and % ER: p=0.006; and % PR: p=0.012); there were no significant associations between HEI scores and common lymphoid progenitor cells, but was associated with high infiltration of
and % ER/PR in premenopausal pts. There were no significant associations between PA and % ER/PR. dendritic cells and stromal-related cells and low infiltration of Myeloid-related cells; M1-
Conclusions: Healthier diet was associated with higher % ER and PR in postmenopausal pts with BC, but no rela- macrophages, and eosinophils, and lymphoid-related cells; Th1 cells, Tregs, NK T cells,
tionship was seen in premenopausal pts. These data suggest that lifestyle factors may influence BC pathologic features
related to outcomes in older women. Support: U10CA180821, U10CA180882, UG1CA189823, U24CA196171, https:// and memory B cells. HSC high BC enriched TGF-b signaling, myogenesis, coagulation,
[Link]. Clinical trial information: NCT02750826. Research Sponsor: National Cancer and angiogenesis gene sets. On the other hand, all the cell proliferation-related gene sets
Institute/U.S. National Institutes of Health; U10CA180821; National Cancer Institute/U.S. National Institutes of Health; in Hallmark collection; E2F targets, G2M checkpoint, MYC targets-v2, and mitotic
U10CA180882; National Cancer Institute/U.S. National Institutes of Health; UG1CA189823; National Cancer Institute/
U.S. National Institutes of Health; U24CA196171; National Cancer Institute/U.S. National Institutes of Health;
spindle, enriched to low HSC BC, and HSC infiltration was significantly lower in high
U10CA180820; National Cancer Institute/U.S. National Institutes of Health; U10CA180868; National Cancer Institute/U.S. histological grade BC and in Ki67 high expression BC. HSC high patients were sig-
National Institutes of Health; U10CA180863; National Cancer Institute/U.S. National Institutes of Health; UG1CA189974; nificantly associated with better overall survival compared to low patients in ER+/HER2-
National Cancer Institute/U.S. National Institutes of Health; CCS 707213; Italian Association for Cancer Research (AIRC) /
Gianni Bonadonna.
(both p,0.02), but not in TNBC in both cohorts. Interestingly, there was no survival
difference by HSC infiltration in ER+/HER2- when neoadjuvant chemotherapy (NAC) was
% ER mean (SD) p value % PR mean (SD) p value used. Together with our finding that HSC in the TME markedly reduced by NAC, we
Post-Menopausal cannot help but speculate that the loss of HSCs by NAC may have contributed to lose
HEI* score
£50.1 [n=134] 66.1 (43.3) 0.012 43.1 (41.8) 0.004 their benefit in patient prognosis. Lastly, high levels of HSC were associated with
>50.1 [n=156] 78.0 (36.8) 57.7 (41.0)
MVPA† min/wk significantly lower risk of lung metastasis and better survival, but not with brain and
0 [n=147] 74.5 (39.3) 0.37 50.0 (41.8) 0.41 bone metastases. Conclusions: This is the first report that quantified HSCs using
‡1 [n=147] 70.3 (41.2) 53.0 (42.0)
Pre-Menopausal transcriptome in TME and demonstrated that they are rare, associated inversely with cell
HEI* score
£50.1 [n=124] 70.5 (38.5) 0.15 53.5 (41.1) 0.43
proliferation and with better survival in ER+/HER2- BC patients. Survival benefit of HSC
>50.1 [n=103] 65.5 (43.0) 49.1 (43.0) infiltration was lost with NAC that reduce its infiltration. HSC high BC was associated
MVPA† min/wk
0 [n=112] 66.3 (41.0) 0.99 49.5 (41.4) 0.59 with lower risk of lung metastasis and with better survival, but not with brain nor bone
‡1 [n=113] 66.3 (41.2) 52.6 (42.7) metastasis. Research Sponsor: National institutes of healthes.
*Healthy Eating Index.
†
Moderate/Vigorous Physical Activity.

563 Poster Session 564 Poster Session

MRI-based radiomic signature and its association with genomic complexity Pregnancy-associated breast cancer: Tumor infiltrating lymphocytes TILt-
in breast tumor heterogeneity. First Author: Joonoh Lim, Inocras Inc., San Diego, CA ing the balance? First Author: Carsten F.J. Bakhuis, University Medical Center Utrecht,
Background: Breast cancer is inherently heterogeneous, posing challenges for effective Utrecht, Netherlands
treatment. Uncovering the relationship between imaging features and genomic profiles Background: Pregnancy-Associated Breast Cancer (PABC), which includes breast
could improve patient stratification. In this study, we evaluated whether radiomic cancer diagnosed during pregnancy (PrBC) or postpartum (PPBC), is often more ag-
features can capture the underlying genomic complexity of breast tumors, potentially gressive and diagnosed at more advanced stage compared to breast cancer in non-
offering a non-invasive means to better characterize tumor heterogeneity. Methods: We pregnant young women. The aggressive tumor growth in PABC may be influenced by the
analyzed 284 breast cancer patients using an integrated radiogenomic approach. MRI- maternal shift towards immunotolerance during pregnancy, aimed at safely harboring
derived radiomics features were extracted and clustered using unsupervised learning the semi-allogenic fetus. Potentially, this shift allows cancer cells to evade immune
methods, resulting in 12 distinct clusters. We then analyzed these clusters against detection. In recent years, the importance of stromal tumor-infiltrating lymphocytes
matched whole-genome sequencing and transcriptome data, focusing on heterogeneity- (sTILs) as a marker of the anti-cancer immune response has become evident. Given the
related radiomics features. Clustering was performed using dynamic tree cutting after aforementioned relevance of immunotolerance in PABC development, we have evalu-
hierarchical clustering of 10 principal components derived from 214 radiomic features. ated the presence and prognostic importance of sTILs in PrBC and PPBC patients in this
Results: We identified distinct patterns of tumor heterogeneity among the 12 identified most extensive study to date. Methods: We assessed tumor tissues from the Dutch
clusters, named according to descending cluster size (range: 10-46). Clusters 9, 4, and 3 Pregnancy-Associated Breast Cancer Cohort, which includes PrBC and PPBC (#1 year
exhibited the highest homogeneity (in that order), with cluster 9 being the most ho- postpartum) patients diagnosed between 1988 and 2022. Whole slide images (H&E) of
mogeneous overall. Cluster 12, 11, 8, 7, and 5 showed varying degrees of heterogeneity, tumors from 200 patients were uniformly reviewed, and sTILs were scored according to
while clusters 1 and 2 were moderately heterogeneous. Cluster 1-3 were HER2-enriched international guidelines. Given the lack of previous sTIL cutoffs for PABC, a data-driven
(PAM50). Clusters 1 and 2 together had ERBB2 amplifications (33%; Fisher’s exact test, approach was chosen. Results: Our initial analysis revealed a clear survival benefit for a
P = 0.056), whereas cluster 3 was HER2-positive (IHC) without amplifications. Cluster 1 sTIL score of at least 20%. Therefore, our PrBC/PPBC patient group was divided into a
leaned toward the basal-like subtype, while 3 leaned toward luminal A. Cluster 2 was “Low sTILs” ( , 20%, n = 153) and a “High sTILs” ($20%, n = 47) group. High sTIL scores
enriched in luminal B (50%; P = 0.012). Cluster 1-3 were distinguishable by their degree were associated with a higher histologic grade (89% grade III versus 73% in the low sTIL
of radiomics-quantified heterogeneity. Cluster 4 was enriched in high Myc expression group, p = 0.043) and more frequent triple negativity (68% versus 43%, p = 0.021) Despite
(17%; P = 0.059). Cluster 5 was enriched in whole-genome-based HRD (40%; P = 0.01) this more aggressive histopathology, higher sTILs were associated with a significantly
and basal-like (52%; P = 0.001). Cluster 6 was deprived of TP53 mutations (37%; P = better 5-year overall survival (OS) probability (94% versus 69%, p , 0.001). In a
0.04), had low tumor mutational burden, and was characterized by small volume but multivariable analysis, correcting for disease stage, intrinsic subtype and tumor grade,
higher surface-volume ratio, suggesting irregular shape. Cluster 8 was enriched in high sTIL scores remained a strong prognostic indicator in PABC (HR 7.3, 95% CI 2.24 –
PIK3CA mutations (60%, P = 0.046), cluster 10 was enriched in CHEK2 mutations (9%; 23.9). In a subgroup analysis for triple negative disease only (n = 93), patients with a high
P = 0.039), cluster 11 showed high TERT (40%, P = 0.005) and CDKN2A (40%; P = 0.048) sTIL score (n = 30) showed a better 5-year OS probability (93% versus 60%, p = 0.002),
expression, cluster 12 was predominantly post-menopausal (80%; P = 0.47), and both which also persisted in a multivariable analysis (HR 4.7, 95% CI 1.4 – 15.7).
clusters 10 and 12 exhibited low ESR1 expression (20%; P = 0.035). Conclusions: This Conclusions: Despite the immunotolerance in pregnancy, this study demonstrates the
comprehensive radiogenomic analysis demonstrates that MRI-based radiomics features presence and prognostic importance of sTILs in PABC. Importantly, patients with high
can effectively capture tumor heterogeneity patterns that correlate with specific ge- sTILs ($20%) had a markedly better prognosis despite having more aggressive disease
nomic alterations in breast cancer. The identification of 12 distinct clusters, each with characteristics, regardless of subtype. Therefore, sTILs may be an important prognostic
characteristic genomic features, provides new insights into the biological basis of tumor indicator and may aid in selecting patients to forgo adjuvant systemic therapy, especially
heterogeneity, potentially opening new avenues for breast cancer subtyping. Research during pregnancy. Research Sponsor: A Sister’s Hope for Breast Cancer Research;
Sponsor: None. Private philantropist (unrestricted funding for this project).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 21s

565 Poster Session 566 Poster Session

Effect of allostatic load and measures of segregation on cancer detection Prognostic value of systemic inflammation in early-stage breast cancer in
and false positive rate after screening mammography. First Author: Niam the CANTO cohort (SIM-CANTO). First Author: Julia Dixon-Douglas, Peter Mac-
Abeysiriwardena, Massachusetts General Hospital, Boston, MA Callum Cancer Centre, Hobart, Australia
Background: Allostatic load (AL), a cumulative stress measure and residential seg- Background: The importance of host anti-tumor immunity in early-stage breast cancer
regation (MRSs) have been associated with breast cancer (BC) outcomes. We assess the (eBC) is now well recognised. Neutrophil / lymphocyte ratio (NLR) is a peripheral blood-
association of AL and MRSs on breast screening outcomes. Methods: From the Mass based measure of systemic inflammation and immune status that has been associated
General Brigham Biobank, we retrospectively identify women aged $40, who underwent with prognosis in other tumour [Link] aimed to evaluate the prognostic value of NLR
screening mammography from Jan 1, 2021 to Dec 31, 2021. We collected age, self- in a large prospective cohort of eBC. Methods: Patients with eBC (stage I-III) in the
reported race/ethnicity, and zip code. Five MRSs were computed: Dissimilarity, Isolation national French CANTO (NCT01993498) cohort with baseline peripheral blood counts
(BI), Delta Index, Absolute Centralization, and Spatial Proximity. To compute AL, we (obtained after diagnosis and before any eBC treatment, including surgery) were in-
assessed cardiovascular, metabolic, immune and renal lab values. AL was assigned one cluded, regardless of systemic (neo)adjuvant therapy. The independent variable of
point for each lab value in the worst quartile and summed. We assessed any false interest was baseline NLR assessed as a continuous variable. Outcomes included in-
positive (FP) and cancer diagnoses within 12 months after the index screen. Multiple vasive- and distant-disease-free survival (iDFS, DDFS) and overall survival (OS). We
imputation was performed for missing covariates. Multivariable logistic regression performed univariable analysis followed by multivariable Cox regression models se-
models were constructed to assess age, race, AL, and each MRSs association with quentially adjusting for age, biologic subtype, TNM stage and treatment. Main analyses
cancer detection and FP rates. Rubin’s rules were applied to estimate overall odds ratios were conducted in the overall cohort, while additional analyses explored the role of NLR
(OR), confidence intervals (CI), and p-values for all covariates. Results: Of 13,754 in subtypes (ER+/HER2-, HER2+, TNBC). For a cohort of patients receiving neoadjuvant
women assessed, 1.2% (n=169) women received a cancer diagnosis. Most of the women therapy we tested the impact of NLR or pCR using Wilcoxon test. Sensitivity analyses
were White (87.6%), 2.6% Asian, 5.3% Black; 1.5% self-identified as Hispanic; mean age, used NLR as a categorical variable (using median NLR as cutoff to define high vs low
64.4611.3SD. Each point increase in AL increased the risk of cancer detection after NLR). Results: Overall, 10 470 patients were included. Median follow-up was 6.7 years
screening mammography by 15% (OR=1.15,95% CI[1.06, 1.25],p=0.001). No association (5.1 – 8.5). The median age at diagnosis was 56.4 years. Most (78%) of patients had
was detected between each MRS and cancer detection (all p.0.22). BI, the expected stage I/II eBC, 77% ER+/HER2-, 13% HER2+ and 9% TNBC. The median NLR was 2.03. In
proportion of neighbors belonging to the same group, was associated with FP rate, with the univariate analysis, there was a significant association between increasing NLR and
BI . 0.6 increasing the odds of FPs (OR=2.80[1.13-6.92],p=0.026). After adjusting for AL worse DDFS in the overall cohort (HR: 1.1, p = 0.004; 95% CI:1.1 – 1.16) and in the ER+/
and MRSs, age and race were not significantly associated with cancer detection, but 5- HER2- cohort (HR: 1.1; p = 0.03; 95%CI:1.1 – 1.2 ). In a model adjusted by age and
year changes in age were associated with lower FP rate (OR=0.82[0.80, 0.84],p,0.001). biologic subtype, NLR showed significant associations with DDFS (HR 1.07; p = 0.04) in
Conclusions: AL was associated with an increased risk of cancer detection after the global cohort, but these associations did not maintain significance after further
screening mammography after adjusting for age, race, and MRSs. The MRS BI above 0.6 adjustment for TNM stage and treatment. Similarly, in the ER+/HER2- cohort, NLR was
was associated with an increased false positive rate. Further work is needed to confirm significantly associated with DDFS when adjusted for age (HR 1.08; p = 0.02), which was
these observations. Clinical Relevance Statement: AL and MRS (i.e., BIS) may represent no longer significant after adjusting for TNM stage. No statistically significant dif-
potential biomarkers for personalized mammographic screening for BC. Research ferences were observed across other subtypes for DDFS, iDFS or OS, nor for pCR in the
Sponsor: None. neoadjuvant cohort. Sensitivity analysis showed consistent results, in particular low NLR
was significantly associated with improved DDFS (HR: 0.8, p = 0.03; 95%CI: 0.7 – 0.9) in
the ER+/HER2- subgroup. Conclusions: Systemic inflammation, as measured by
baseline NLR, was associated with significantly shorter DDFS in the overall CANTO
cohort and in the ER+/HER2- subgroup in univariable and age- adjusted analysis.
However, this association disappeared after adjustment for known clinicopathologic
prognostic characteristics. Research Sponsor: None.

567 Poster Session 568 Poster Session

Exploring tumor genomics and clinical outcomes in in adolescent and young Global trends and regional disparities in young-onset breast cancer: Age-
adult (AYA) breast cancer. First Author: Faris Tamimi, King Hussein Cancer Center, specific patterns from 1990 to 2021. First Author: Muzamil Khan, The George
Amman, Jordan Washington University School of Medicine and Health Sciences, Washington, DC
Background: Breast cancer in adolescents and young adults (AYAs, defined as individuals aged 39 or Background: Young-onset breast cancer (YOBC), defined as breast cancer in women
younger) often exhibits aggressive biological behavior and distinct clinical patterns compared to older under 40, is a growing global health concern with unique biological and clinical im-
patients. This research investigates the somatic mutations and clinical features that distinguish AYA plications. Analyzing its patterns and distribution is essential for developing effective
breast cancer, aiming to uncover unique genomic and prognostic characteristics. Methods: We
analyzed data from the METABRIC cohort using cBioPortal, focusing on 173 genes sequenced from
healthcare strategies and ensuring efficient resource allocation. Methods: Data on
2,509 breast cancer cases. The dataset included information on copy number variations, gene ex- incidence, prevalence, mortality, and morbidity indicators (DALYs, YLD, and YLL rates per
pression, and long-term clinical outcomes. Tumor characteristics, mutation frequencies, and relapse- 100,000) were obtained from the Global Burden of Disease (GBD) 2021 database for
free survival (RFS) outcomes were compared between AYAs and older patients. Statistical methods women under 40 from 1990 to 2021. The dataset, covering 204 countries, was stratified
employed included Chi-square tests for categorical data, Wilcoxon Rank-Sum tests for medians, and into five age groups: 35–39, 30–34, 25–29, 20–24, and , 20 years. Linear regression
Cox regression for survival analysis. The 10 most commonly altered genes across the dataset were was used to calculate average annual percent change (AAPC) and 95% confidence
examined. Results: Of 2,498 patients included in the analysis, 143 were AYAs (5.7%) and 2,355 were intervals (CI) for each indicator and age group. Statistical significance was set at p ,
older patients (94.3%). Compared to older patients, AYAs demonstrated: Lower ER-positive rates
0.05. Results: Between 1990 and 2021, YOBC accounted for 16,783,674 deaths globally.
(37.8% vs. 74.9%), Higher ER-/HER2- subtype prevalence (25.2% vs. 11.6%), Greater lymph node
positivity (52.4% vs. 41.3%), and Worse Nottingham Prognostic Index scores (median 5.04 vs. 4.04). Incidence increased across all age groups, with the largest increases in women under 30.
Genomic analysis revealed significantly higher TP53 mutation frequencies in AYAs (58.9% vs. 33.2%, AAPC for incidence was 2.27 (95% CI: 2.21–2.33) in those under 20, 2.18 (95% CI:
p , 0.01) and lower PIK3CA alterations (28.1% vs. 42.1%, p , 0.01). Other significantly altered genes 2.11–2.24) for ages 20–24, and 1.60 (95% CI: 1.51–1.70) for ages 25–29 (all p ,
in AYAs included TG, TRPS1, CASC8, POU5F1B, MYC, CASC11, NDRG1, and LINC02912. However, 0.0001). Prevalence showed similar increases: 2.35 (95% CI: 2.29–2.41), 2.24 (95% CI:
stratification by receptor subtypes (ER-/HER2-, ER+/HER2-, HER2+) showed no significant differences 2.18–2.30), and 1.64 (95% CI: 1.54–1.73), respectively (all p , 0.0001). Mortality AAPCs
in TP53 or PIK3CA alterations between AYAs and older patients. AYA status ( , 40 years) was were 1.54 (95% CI: 1.46–1.61) for under 20, 1.35 (95% CI: 1.23–1.46) for ages 20–24,
associated with worse recurrence-free survival (RFS) in univariable analysis and remained a significant and 0.68 (95% CI: 0.59–0.76) for ages 25–29, while older cohorts saw declines, including
predictor in multivariable analysis, adjusting for ER status, HER2 status, and nodal involvement (HR:
1.33, 95% CI: 1.00–1.77, p = 0.048). Conclusions: AYA breast cancer is marked by distinct genomic
-0.38 (95% CI: -0.53 to -0.23, p , 0.0001) for ages 35–39. DALYs also increased in
and clinical features, including higher TP53 mutation rates, lower PIK3CA mutation rates, more younger groups, with AAPCs of 1.56 (95% CI: 1.49–1.63) for under 20 and 1.37 (95% CI:
aggressive subtypes such as ER-/HER2-. While some genomic differences are less pronounced within 1.26–1.48) for ages 20–24 but declined in older cohorts. Regionally, Turkey had the
biomarker-matched subgroups, AYAs remain at higher risk for recurrence. These findings highlight the highest AAPCs across metrics, Malawi had the highest incidence and YLL for those
urgent need for age-specific therapeutic strategies to improve outcomes in this population. Research under 20, and Zimbabwe, Yemen, and Lesotho reported the highest incidence, preva-
Sponsor: None. lence, YLL, and DALY. In contrast, Ukraine, Mariana Island and Armenia experienced the
Gene alteration event frequency by age group (AYA versus >39 years). greatest declines in incidence, prevalence, YLL, and DALY. Conclusions: The global
Gene AYA >39 p-Value q-Value burden of YOBC has risen sharply, especially in the youngest cohorts (, 20 and 20–24
TP53 58.90% (86/146) 33.16% (780/2352) ,0.001 ,0.001 years), with significant increases in low- and middle-income regions like Turkey, Sub-
PIK3CA 28.08% (41/146) 42.05% (989/2352) ,0.001 0.0442 Saharan Africa, and the Middle East. Age-specific interventions, early detection, and
TG 35.62% (52/146) 23.72% (558/2352) ,0.01 0.0645
NDRG1 29.45% (43/146) 19.56% (460/2352) ,0.01 0.126
preventive measures are needed to address the growing YOBC rates. Research Sponsor:
TRPS1 30.82% (45/146) 23.04% (542/2352) 0.0347 0.340 None.
CASC8 30.14% (44/146) 21.51% (506/2352) 0.0178 0.231
LINC02912 29.45% (43/146) 20.88% (491/2352) 0.0167 0.229
CASC11 30.14% (44/146) 21.73% (511/2352) 0.0235 0.268
MYC 30.14% (44/146) 21.68% (510/2352) 0.0234 0.268

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
22s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

569 Poster Session 570 Poster Session

Interferon signaling and outcomes in triple-negative breast cancer (TNBC) in MHC class I expression and outcomes in breast cancer in the real-world
FinXX, CALGB 40603 (Alliance) and real-world clinico-genomic data. First clinico-genomic data and the FinXX trial. First Author: Yi Liu, Department of
Author: Saranya Chumsri, Mayo Clinic Florida, Jacksonville, FL Quantitative Health Sciences, Mayo Clinic, Phoenix, AZ
Background: Several studies established the prognostic role of both the amount and locations Background: Major histocompatibility complex class I (MHC I) plays a critical role in
of tumor-infiltrating lymphocytes (TILs) in TNBC. Three distinct immunotypes were described immune surveillance by binding peptides derived from intracellular proteins and presenting
based on the amount and locations of TILs: immune enriched (IN), immune excluded, and them on the cell surface for recognition by CD8+ T cells. Loss or downregulation of MHC I
immune desert. Using single-cell spatial transcriptomic analysis in the Mayo Clinic TNBC cohort, expression has been identified as a key mechanism of immune evasion in cancers. Here,
our previous studies showed the central role of interferon (IFN) signaling in IN phenotype. we evaluated MHC I expression and outcomes in all subtypes of breast cancer (BC).
Herein, we evaluated the association between IFN and outcomes in TNBC in 3 independent Methods: 9,038 BC samples were analyzed via NGS (592-gene panel, NextSeq; WES/WTS,
datasets. Methods: NanoString IO360 was performed in 114 samples from FinXX NovaSeq; Caris Life Sciences, Phoenix, AZ), including triple-negative BC (TNBC) 3,038,
(NCT00114816) to generate 22-gene IFNɑ and 33-gene IFNg signatures. RNA sequencing was HER2-positive (HER2+) 1,082, and hormone receptor-positive (HR+HER2-) 4,918. Immune
performed in 388 samples from CALGB 40603 (NCT00861705). 3038 TNBC samples were tested
cell fractions were estimated using WTS deconvolution (Quantiseq). MHC I (HLA-A/HLA-B/
by WTS (NovaSeq; Caris Life Sciences, Phoenix, AZ). Median values were used as cut-offs for
HLA-C)-high (H) and -low (L) were classified by RNA expression above or below the 25th
high vs low IFNg RNA expression and 18-gene IFNg signature scores. Caris Life Science CODEai
was used to evaluate real-world overall survival (OS) obtained from insurance claims and
percentile. Real-world overall survival (OS) was derived from insurance claims and cal-
calculated from tissue collection to last contact using Kaplan-Meier estimates. Chi-square, culated from tissue collection to last contact using Kaplan-Meier. NanoString IO360 was
Mann-Whitney U, ANOVA, and Cox regression were used. Results: A high 22-gene IFNɑ sig- performed in 114 samples from the FinXX trial (NCT00114816). Statistical significance
nature score was associated with significantly improved recurrence-free survival (RFS) in FinXX was assessed using chi-square, Mann-Whitney U, ANOVA, and Cox regression with multiple
(hazard ratio [HR] 0.32, 95% confidence interval [CI] 0.14-0.74, p 0.007) and OS (HR 0.28, 95%CI comparison adjustments (q,.05). Results: TNBC had higher expression of HLA-A and
0.12-0.66, p 0.003). Similar findings were observed with 33-gene IFNg signature with significant HLA-B (median TPM: 169 and 191) compared to HER2+ (146.6 and 170, q,0.05) and
improvement in RFS (HR 0.21, 95%CI 0.09-0.51, p , 0.001) and OS (HR 0.18, 95%CI 0.08-0.44, HR+HER2- (141.7 and 157.5, q,0.05). However, there was no significant difference in
p , 0.001). Furthermore, in CALGB 40603, both IFNɑ and IFNg scores were positively as- HLA-C expression across 3 BC subtypes. In TNBC, MHC I-H tumors had higher frequencies
sociated with pathologic complete response (pCR: IFNɑ p 0.019 and IFNg p 0.007) and residual of PD-L1 positivity (66.2% vs. 13.1%) as well as higher infiltration of B cells (4.5% vs. 3.2%),
cancer burden (RCB: IFNɑ p 0.044 and IFNg p 0.013). Using the Caris data platform to further M1 macrophages (5% vs. 1.5%), M2 macrophages (4% vs. 2.1%), Tregs (2.8% vs. 0.8%),
validate, we identified 2899 TNBC patients (pts) with genomic and clinical outcome data. High CD8+ T cells (1.8% vs. 0%), dendritic cells (3.2% vs. 2.8%), higher T-cell inflamed score (137
IFNg expression was associated with significant improvement in OS (25.95 vs 17.43 months; HR vs. -144), and IFNg score (0.02 vs. -0.49) compared to MHC I-L TNBC (all q,.05). MHC I-H
0.65, 95% CI 0.59 – 0.72, p , 0.001). Similarly, pts with high IFNg signature scores had TNBC was associated with significant improvement in median OS (30.1 vs. 15.2 months,
significant improvement in median OS (25.79 vs 16.22 months; HR 0.66, 95% CI 0.6 – 0.73, p , HR 0.55, 95% CI 0.46-0.65, p,0.0001). However, this survival difference was not observed
0.001). Conclusions: This study underscores the pivotal role of IFN signaling in TNBC. High in patients with MHC I-H vs. MHC I-L in HER2+ (HR 1.04, 95% CI 0.74-1.47, p = 0.81) and
IFNa and IFNg signatures were consistently associated with improved RFS, OS, higher pCR HR+HER2- (HR 0.87, 95% CI 0.75-1.02, p= 0.09) BC subtypes. We further validated the MHC
rates, and lower RCB across clinical trial cohorts and real-world data. These findings signify IFN I expression in the FinXX trial. Similarly, patients with MHC I-H had significant im-
signaling as a potential key biomarker and therapeutic target in TNBC. Support: U10CA180821, provement in recurrence-free survival (HR 0.27, 95%CI 0.11-0.66, p = 0.002) and OS (HR
U10CA180882, U24CA196171; Breast Cancer Research Foundation, Mayo Clinic Breast
0.23, 95% CI 0.09-0.57, p = 0.0005) compared to MHC I-L. Conclusions: Our findings
Cancer SPORE (P50CA116201-17), Bankhead Coley, W81XWH-15-1-0292, P50CA015083,
demonstrate that higher MHC I expression is associated with higher immune infiltration
R35CA253187; [Link] Genentech. Clinical trial informa-
tion: NCT00114816 and NCT00861705. Research Sponsor: U.S. National Institutes of Health; and improved outcomes in TNBC but not in HER2+ or HR+HER2- BC subtypes. These
U10CA180821, U10CA180882, U24CA196171; Mayo Clinic Breast Cancer SPORE results suggest that MHC I expression plays a critical role in the tumor microenvironment
(P50CA116201-17), P50CA015083, R35CA253187; Bankhead Coley; Breast Cancer Research of TNBC. Future studies are needed to evaluate the prognostic value and potential
Foundation; U.S. Department of Defense; W81XWH-15-1-0292; Genentech. therapeutic target of MHC I in TNBC. Support: Breast Cancer Research Foundation,
Bankhead Coley, W81XWH-18-1-0564. Clinical trial information: NCT00114816. Research
Sponsor: None.

571 Poster Session 572 Poster Session

Impact of HER2 low status on pathologic response after neoadjuvant che- Reframing hormone-positive DCIS management: Effects of adjuvant ther-
motherapy in TNBC: A large scale retrospective cohort study. First Author: apies and surgical extent on any invasive recurrence. First Author: Thomas
Julia Tchou, University of Pennsylvania, Philadelphia, PA O’Keefe, University of California, San Francisco, San Francisco, CA
Background: HER2 low status, defined as HER2 1+ or HER2 2+ and nonamplified by in-situ Background: The treatment of DCIS is still primarily informed by trials that are now decades old.
hybridization, has demonstrated the ability to identify a population of patients with triple- We have learned since then that only invasive subsequent events, not in situ ones, increase a
negative disease who benefit from trastuzumab deruxtecan in the metastatic setting. The woman’s risk of eventual metastasis and breast cancer mortality, and may necessitate more
implications of HER2 low status in the early-stage setting is unclear. This study evaluated aggressive systemic treatment. This suggests a need for the reassessment of the impact of
whether HER2 low status impacts rates of pathologic complete response in response to adjuvant therapies. Methods: Women diagnosed with a first breast cancer of unilateral hormone
neoadjuvant chemotherapy (NAC). Methods: Using the national cancer database (NCDB), pa- positive DCIS undergoing breast conserving surgery were identified in the Surveillance, Epide-
tients with clinically invasive non-metastatic breast cancer between 2010-2021 were retro- miology and End Results Program registry. Propensity-matching was performed between treat-
spectively identified (n=1,926,979). Patients with unknown receptor status or who did not have ment groups using age, race, lesion size and grade. Competing risks methods were used to
triple negative breast cancer (n=1,755,897) and those who did not have surgery (n=9,849) were estimate the cumulative incidence of any invasive event at 10 years and subdistribution hazard
excluded. Clinicopathologic, treatment, and outcome variables were compared using chi-square ratios (sHRs) were calculated from multivariate models adjusting for the same covariates used for
and anova tests. Subgroup analysis was performed among patients who received NAC and had matching. Results: A total of 14,189 patients diagnosed from 2007 to 2011 were eligible for
available pathologic response results. Multivariable binary logistic regression was performed to matching, among whom 900 (6.3%) suffered a subsequent invasive event.A cohort was developed
assess clinical variables associated with pCR after NAC. Cox proportional hazards model was by matching from the smallest treatment group, lumpectomy with endocrine therapy. There were
performed to assess clinical variables associated with overall survival (OS) in TNBC receiving 2,996 matched patients, among whom 511 (17.1%) had an invasive subsequent event. Adjuvant
NAC. Results: Of the 161,233 individuals eligible for analysis, 79,268 (49.4%) had HER2 low endocrine therapy was associated with reduced risk of any invasive event (sHR=0.38, p,0.001)
disease. The proportion of HER2 0, HER2 1+ and HER2 2+/ish negative (HER2 2+) were 50.6, 34.8, compared to patients undergoing lumpectomy without adjuvant therapy. Radiotherapy alone was
and 14.6% respectively. Overall, 49,994 (31%) individuals received NAC with an overall pCR rate of not associated with reduced risk (sHR=1.03, p=0.81): it was associated with reduced risk of an
42.9%. The pCR rate was significantly lower in those with HER2 2+/ish negative compared to ipsilateral invasive event (sHR=0.65, p=0.006) but an increased risk of a contralateral invasive
those with HER2 0 TNBC at 39.9 vs. 44.2%, p ,0.001. On multivariable analysis, HER2 1+ status event (sHR=1.50, p=0.01). In subgroup analyses, lumpectomy with radiation therapy was noted to
trended towards a lower likelihood of pCR R 0.95, 95% CI 0.91-1.00, p=0.06) while HER2 2+/ish be non-significantly associated with increased risk of any invasive disease in patients younger than
negative status was significantly associated with a lower likelihood of pCR (OR 0.87, 95% CI 0.82- 60 years (sHR=1.38, p=0.053) and with decreased risk in patients 60 years or older (sHR=0.79,
0.94, p,0.001) compared to HER2 0 disease. Cox proportional hazards model analysis dem- p=0.12). Conclusions: Our results suggest that endocrine therapy may confer the greatest risk
reduction to the development of any subsequent invasive recurrences, and whole-breast radio-
onstrated that the strongest clinical factor for worse OS was non-pCR with HR 3.76, 95% CI 3.48 –
therapy may be associated with increased risk for younger women. The risk posed by DCIS as a
4.05, p,0.001 while neither HER2 1+ or HER2 2+/ish negative was associated with worse OS.
high-risk marker may outweigh its risk as a premalignant lesion. Research Sponsor: None.
Conclusions: In this analysis, tumors that were HER2 2+/ish negative were associated with a
significantly lower pCR rate after NAC compared with HER2 0 tumors. This is the first large-scale Any invasive subsequent breast cancer, and subgroup analyses by age. Only treatment effect shown, but
results are adjusted for race, grade, age, and lesion size.
study to demonstrate that HER2 low status may be prognostically unfavorable in early-stage
TNBC. Examination of novel neoadjuvant therapeutic approaches tailored based on HER2 status Any Invasive, Any Age Any Invasive, < 60 years old Any Invasive, ‡ 60 years old
including trastuzumab deruxtecan may be warranted to improve pCR rates and outcomes in sHR (95% CI) p sHR (95% CI) p sHR (95% CI) p
patients with HER2-low early-stage breast cancer. Research Sponsor: None. Treatment
BCS Ref - Ref - Ref -
pCR rate stratified by HER2 low status. BCS + ET 0.38 (0.29-0.51) ,0.001 0.34 (0.21-0.54) ,0.001 0.40 (0.28-0.58) ,0.001
HER2 0 HER2 1+ HER2 2+ BCS + RT 1.03 (0.82-1.28) 0.81 1.38 (1.00-1.91) 0.053 0.79 (0.59-1.07) 0.12
n 81605 56109 23519 sHR = Subdistribution hazard ratio.
Neoadjuvant 26458 32.4% 16841 30.0% 6695 28.5% BCS=Breast conservation surgery.
pCR 11690 44.2% 7064 41.9% 2672 39.9% ET=Endocrine therapy.
RT=Radiation therapy.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 23s

574 Poster Session 575 Poster Session

Impact of postmastectomy radiotherapy on health-related quality of life and Development and validation of prediction models for 5-year and 10-year
safety in breast cancer patients undergoing breast reconstruction: A multi- ipsilateral breast tumor recurrence after breast-conserving surgery. First
center cross-sectional study (Reborn-02). First Author: Hirohito Seki, Department Author: Yasuaki Sagara, Department of Breast Surgical Oncology, Social Medical
of Breast Surgery, Kyorin University School of Medicine, Tokyo, Japan Corporation Hakuaikai Sagara Hospital, Kagoshima, Japan
Background: Postmastectomy radiotherapy (PMRT) is the standard treatment for im- Background: Ipsilateral breast tumor recurrence (IBTR) remains a critical concern for
proving the prognosis of patients with high-risk breast cancer. Expanded indications for patients undergoing breast-conserving surgery (BCS). Reliable prediction tools for IBTR
breast reconstruction (BR) will likely increase the number of patients undergoing BR who risk can support personalized surgical strategies and adjuvant treatment decisions,
require PMRT, but combining BR with PMRT raises concerns about complications and especially in the era of evolving systemic therapies. This study aimed to develop and
aesthetics. Evaluating the impact of PMRT on the health-related quality of life (HR-QOL) in validate prediction models for 5-year and 10-year IBTR. Methods: This multi-center
Japanese patients is essential for shared decision-making (SDM). This study assessed retrospective cohort study included 10,089 women who underwent partial mastectomy
differences in postoperative HR-QOL and complications between patients who underwent for invasive breast cancer between 2008 and 2017. Cases involving conversion to
BR with or without PMRT. Methods: We conducted a multicenter cross-sectional study mastectomy, use of neoadjuvant chemotherapy, bilateral/multiple cancers, or missing
using a questionnaire survey of patients with primary breast cancer who underwent BR key data were excluded. Prediction models were developed using Cox proportional
between January 2008 and December 2022 at participating institutions, which was ap- hazards regression and validated via bootstrap resampling. Model performance was
proved by the respective institutional review boards. We used the Japanese version of the assessed using Harrell’s C-index, Brier scores, calibration plots, and goodness-of-fit
BREAST-Q questionnaire and questions on patient backgrounds. Results: We included tests. The cumulative incidence of IBTR, which served as the baseline for the prediction
1078 patients with primary breast cancer. The questionnaire response rate was 77.0% model, was calculated using the Fine and Gray model, treating death as a competing risk.
(830/1078). The non-PMRT and PMRT groups comprised 616 and 214 patients, re- Results: The median age of patients was 55 years [interquartile range (IQR): 46–65].
spectively. The PMRT group had higher rates of axillary lymph node dissection (11.7% vs. During a median follow-up of 8.9 years (IQR: 6.4–10.8), IBTR occurred in 292 patients
52.4%; P,0.001), adjuvant hormonal therapy (67.2% vs. 90.6%; P,0.001), and peri- (3.1%). The initial model, based on variables from Sanghani et al. (JCO 2010), achieved a
operative chemotherapy (31.8% vs. 84.0%; P,0.001) than the non-PMRT group. Moreover, Harrell’s C-index of 0.70. Incorporating hormonal receptor status, HER2 status, ra-
the overall complication rate (45.3% vs. 76.2%, P,0.001) and the rates of dermatitis (9.9 diotherapy, and targeted therapy as predictors reduced the C-index to 0.60, despite their
vs. 48.1%; P , 0.001), skin necrosis (2.9 vs. 9.8%; P,0.001), breast asymmetry (17.0% vs.
clinical relevance. Importantly, the inclusion of these factors improved calibration,
24.3%; P=0.002), capsular contracture (4.9 vs. 16.4%; P,0.001), and lymphedema of the
demonstrating better alignment between predicted and observed IBTR probabilities. The
upper limb (1.1% vs. 7.5%; P,0.001) were higher in the PMRT group. Multivariate analysis
final Cox model exhibited strong clinical and statistical robustness (p , 0.001), pro-
revealed PMRT as an independent risk factor for dermatitis, skin necrosis, and capsular
viding individualized IBTR risk estimates. Cox-Snell residual analysis confirmed
contracture. In the BREAST-Q assessment, the PMRT group showed lower satisfaction
with the breast (55 vs. 49; P,0.001), and with physical (85 vs. 76; P,0.001), psychosocial
goodness-of-fit, with the cumulative hazard closely following the 45-degree line up to
(55 vs. 49; P,0.001), and sexual well-being (36 vs. 34; P=0.021) than the non-PMRT group. 0.3, indicating reliable model performance for observed events. While hazard ratios
Multiple regression analysis revealed PMRT as an independent factor associated with low (HRs) for chemotherapy and radiotherapy were consistent with results of EBCTCG meta-
BREAST-Q scores for breast satisfaction and physical and psychosocial well-being in analyses (MA), HR for endocrine therapy was lower than reported in MA. Consequently,
patients with BR. Conclusions: This is the first large-scale, multi-institutional study to use HRs from MA were adopted to account for treatment effects in our prediction model.
patient-reported outcomes to assess the effects of PMRT on HR-QOL in Japanese patients Conclusions: We have developed and validated a new prediction model for 5-year and
with breast cancer who underwent BR. PMRT was associated with an increased risk of 10-year IBTR using Cox regression and bootstrap methods. A web-based tool is under
complications and decreased HR-QOL in patients with BR. Of note, these findings do not development to enable individualized risk assessment and treatment planning. Future
negate the role of PMRT in patients undergoing BR, but highlight the importance of SDM research will focus on external validation and the integration of genetic and novel
based on realistic HR-QOL expectations after breast reconstruction surgery with PMRT. therapeutic data to enhance model robustness and clinical utility. Research Sponsor:
Research Sponsor: None. Japanese Breast Cancer Society.

576 Poster Session 577 Poster Session

Regional node recurrence after sentinel lymph node biopsy for ycN0 patients Preoperative axillary US and MRI in early-stage breast cancer: Potential to
treated with primary chemotherapy in cT1-3N1M0 breast cancer (second prevent unnecessary axillary surgery. First Author: Xin Wang, West China Hospital
report from SHARE study). First Author: Shigeru Imoto, Kyorin University, Mitaka, (China), Chengdu, China
Japan Background: Recent studies, including the INSEMA and SOUND trials, have demonstrated
Background: Axillary management after sentinel lymph node biopsy (SLNB) is still that omitting sentinel lymph node (SLN) biopsy does not adversely affect survival outcomes.
debated for clinically node-positive breast cancer (BC) patients treated with neoadjuvant Notably, patients with positive axillary findings on ultrasound (US) tended to show more
chemotherapy (NAC). The Japanese Society for Sentinel Node Navigation Surgery extensive axillary nodal metastasis on final histopathological examination. Thus, we aimed to
conducted a prospective non-randomized phase 2 study (SHARE study, focused on the advanced utility of preoperative axillary US and Magnetic Resonance Imaging
UMIN000030558). False-negative rate (FNR) as the primary endpoint was reported at (MRI) in preventing unnecessary axillary surgery in a large series of patients with early-stage
ASCO2023. In brief, it was 11.5% (90% confidence interval, 5.5% and 20.5%). When breast cancer treated with both breast surgery and SLN biopsy. Methods: Between January
multimodal imaging, 1 nodal metastasis, multiple tracers, multiple sentinel lymph nodes 2012 to December 2023, a total of 7879 patients who underwent breast surgery for clinical
were considered for subset analysis, the FNR was 12.1%, 9.1%, 11.1%, and 10.6%, T1–2/N0 cancers and SLN biopsy with or without axillary lymph node dissection were
respectively. Here we report a short-term prognosis of ycN0 BC patients. included. Preoperative axillary US and MRI findings were compared with clinical-pathologic
variables, considering the presence of SLN metastasis. Patients with positive results
Methods: Clinical T1-3N1M0 BC was eligible. Nodal metastasis was histologically
routinely underwent US-guided fine needle aspiration biopsy (FNAB), and negative FNAB
confirmed. In case of ycN0 BC, SLNB was planned and lymphatic mapping depended on
results were also considered. Multivariate logistic regression analysis was used to identify
each institutional practice. In cases of pN0(sn), pN0(i+)(sn) and pN1mi(sn), SLNB factors associated with SLN metastasis. Results: A total of 7879 eligible patients (47.48%
followed by lymph node sampling (LNS) had been allowed instead of axillary lymph node with clinical T1 cancer and 52.52% with T2 cancer) were included in our study. Among them,
dissection (ALND). Results: Between February 2018 and May 2021, 185 patients from 2048 (25.99%) had positive SLN biopsy, and 1971 (25.02%) underwent axillary lymph node
19 institutes were registered. Twenty-seven cases were exculded by protocol deviation, dissection due to positive SLN biopsy. Patients with SLN metastasis were younger and had a
non-ycN0 or withdrawal of SLNB. Among the 158 ycN0 cases, sentinel lymph nodes were higher frequency of positive axillary findings on US and MRI, as well as clinical T2 stage
detected in 153 cases. The median age was 52 years old. Clinical stage was IIA in 40 (P,0.05). At multivariate analysis, positive axilla at US (P =0.001), positive axilla at MRI (P
cases, IIB in 105 and IIIA in 8. Luminal subtype classified by ER, PR and HER2 expression =0.02), clinical T2 stage (P =0.004), grade 3 (P =0.004) and lymphovascular invasion
was found in 60 cases, HER2 in 34, Luminal-HER2 in 35 and triple-negative in 24. SLNB (P=0.001) were significantly associated with SLN metastasis. The number of nodal me-
alone was performed in 37 cases, SLNB and LNS in 72 and SLNB followed by ALND in 44. tastases increase with increasing tumor size. Among 3595 (45.06%) patients with negative
Regional nodal irradiation was planned in 63 cases. At the median follow-up of axilla at US and MRI and clinical T1 stage cancer, 698 (8.75%) had SLN metastasis. For these
24 months, 13 cases relapsed and 5 died of BC. Regional node recurrence-free survival people, the median follow-up was 73.6 months. The estimated 5-year invasive disease–free
rate at 2 years after surgery, disease-free survival rate and overall survival rate were survival rate was 96.8%, while the 5-year overall survival rate was 98.6%. The analysis of the
98.6%, 92.6% and 97.3%, respectively. Conclusions: SLNB-guided axillary management first primary-outcome events (occurrence or recurrence of invasive disease or death from
is feasible and safe in ycN0 BC after NAC. Clinical trial information: 000030558. any cause), showed apparent differences between the positive axilla at US and MRI and
Research Sponsor: The Japanese Society for Sentinel Node Navigation Surgery. clinical T1 stage cancer group and negative axilla at US and MRI and clinical T1 stage cancer
group in the incidence of recurrence (4.7% vs. 3.2%) and death (2.0% vs. 1.4%).
Conclusions: Preoperative axillary US and MRI results, along with clinical T stage are
significant predictors of SLN metastasis in patients with early-stage breast cancer. The
results of this study suggest that preoperative axillary US and MRI can help select patients at
minimal risk of SLN metastasis, for whom axillary surgery could be omitted. Research
Sponsor: National Natural Science Foundation of China; National Clinical Research Center for
Geriatrics, West China Hospital, Sichuan University; Natural Science Foundation of Sichuan
Province.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
24s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

578 Poster Session 579 Poster Session

Use of local estrogen therapy among breast cancer patients in SEER-MHOS Efficacy and safety of post-mastectomy radiation therapy for patients with
database. First Author: Olivia Mitchel, University of Arizona- College of Medicine breast cancer after breast reconstruction: A retrospective multicenter cohort
Tucson, Tucson, AZ study (Reborn-03). First Author: Wakako Tsuji, Department of Breast Surgery, Shiga
Background: Anti-hormonal therapy with tamoxifen or aromatase inhibitors is a key General Hospital, Moriyama Shiga, Japan
component in the treatment of hormone receptor-positive breast cancers. One of the Background: A cornerstone of multidisciplinary treatment for breast cancer, breast recon-
adverse effects of this type of therapy is genitourinary syndrome of menopause, which may struction has recently been extended to high-risk patients with breast cancer who require post-
be treated with vaginal estrogen. It remains unclear whether local vaginal estrogen use mastectomy radiation therapy (PMRT). PMRT reduces the 10-year risk of locoregional re-
carries any increased risk of recurrence or mortality in patients with a history of breast currence; however, its association with postoperative complications remains unclear. This
cancer. This has led to conflicting advice in the clinical setting and potentially unnecessary study aimed to evaluate the efficacy and safety of PMRT in patients with high-risk breast
avoidance of hormone-based products that could possibly provide women symptomatic cancer who underwent breast reconstruction. Methods: This retrospective cohort study in-
relief. Further exploration of the relationship between local estrogen therapy and breast cluded patients with high-risk breast cancer who underwent immediate or delayed breast
cancer outcomes is warranted. Methods: A retrospective cohort study of 18,620 female reconstruction after mastectomy between 2008 and 2018. High-risk patients were defined as
breast cancer patients $ 65 years of age diagnosed between 2010-2017 in the SEER-MHOS those with positive axillary lymph nodes, clinical tumor size of .5 cm, chest wall invasion, or
registry was performed, comparing the breast cancer patients who used local vaginal skin invasion. Patient data were collected from participating institutions and analyzed ret-
estrogen (n=800) to those who did not (n=17,820), to assess whether local vaginal estrogen rospectively. Results: Of the 1,138 patients, 427 (37.5%) underwent PMRT and 711 (62.5%)
exposure was associated with any difference in overall survival as a primary outcome. did not. The median age at surgery was 46 years (range, 23–76years), and the median follow-
Breast cancer specific survival was analyzed as a secondary outcome. Missing data was up period was 8 years. The number of patients meeting more than two high-risk criteria was
excluded by complete case analysis. Wilcoxon rank-sum tests were performed to compare 149 (23.9%) and 81 (11.4%) in the PMRT and non-PMRT cohorts, respectively. Breast re-
continuous variables, Chi-square tests to compare categorical variables, Kaplan-Meier construction using silicone breast implants (SBI) was performed in 70% and 71.4% of patients
estimation to summarize overall survival, and sub-distribution hazard regression was in the PMRT and non-PMRT cohorts, respectively. The locoregional recurrence rates were 7.7%
performed to evaluate breast cancer specific survival by group. Multivariate regression and 12.7% in the PMRT and non-PMRT cohorts, respectively (P=0.034). Multivariate analysis
models controlled for age, race, cancer stage, treatment (i.e. surgery, radiation, anti- revealed that PMRT was an independent predictive factor for reducing locoregional recurrence
(P,0.001) (Table 1). Complications occurred in 130 patients (30.4%) in the PMRT cohort and
hormonal therapy), and year of diagnosis. The research protocol was approved by our
168 (23.6%) in the non-PMRT cohort (P=0.016). The reoperation rates due to complications
Scientific Review Committee and submitted for IRB approval. Results: There was a sta-
were 46 (35.4%) in the PMRT cohort and 61 (36.3%) in the non-PMRT cohort (P = 0.100). In the
tistically significant increase in overall survival (HR=0.56, p,0.0001) as well as breast
PMRT cohort, 62 patients (14.5%) experienced grade 2 radiation-induced dermatitis, and three
cancer-specific survival (HR=0.53, p=0.014) among breast cancer patients who used vaginal
patients (0.7%) developed grade 2 radiation-induced pneumonia. Conclusions: Although the
estrogen compared to those who did not. Among those who used vaginal estrogen, there
incidence of adverse events was slightly higher in the PMRT cohort, the reoperation rates due
was a statistically significant increase in overall survival for those with a duration of use .7 to complications were comparable between the cohorts. PMRT is a safe and effective modality
years (median duration of use) compared to those with a duration of use ,7 years (HR=0.01, that provides substantial benefits for locoregional recurrence reduction in patients with high-
p,0.0001). Subset analysis restricted to patients with hormone positive breast cancer risk breast cancer undergoing breast reconstruction. Research Sponsor: None.
showed a statistically significant increase in overall survival for those who used vaginal
estrogen compared to those who did not (HR=0.62, p=0.0007), and a nonsignificant increase Multivariate analysis for locoregional recurrence in patients with high-risk breast cancer who
underwent breast reconstruction.
in breast cancer specific survival (HR=0.62, p=0.08). Conclusions: The use of vaginal
Factors Hazard ratio 95% Confidence interval P-value
estrogen among this SEER-MHOS cohort of breast cancer patients showed improved
survival outcomes. These findings add to a rising contemporary paradigm shift that local PMRT 0.40 0.251-0.623 ,0.001
hormone therapy is not associated with increased risk to overall or breast cancer specific Neoadjuvant therapy 3.43 2.29-5.12 ,0.001
pT (>5 cm) 1.95 1.25-3.03 0.003
survival, which has important clinical implications. Research Sponsor: Departmental Surgical margin 1.96 1.01-3.81 0.047
funding from the University of Arizona College of Medicine - Tucson Department of Surgery.

580 Poster Session 581 Poster Session

Implications for sentinel lymph node biopsy (SLNB) omission in patients Serial circulating tumor DNA (ctDNA) monitoring in early-stage, HR+/HER2-,
with early-stage node-negative HR+/HER2- breast cancer undergoing invasive lobular carcinoma (ILC) of the breast and impact on clinical
mastectomy. First Author: Jennifer Hechao Chen, The University of Texas MD outcomes. First Author: Julia Foldi, University of Pittsburgh Medical Center
Anderson Cancer Center, Houston, TX (UPMC), Pittsburgh, PA
Background: The SOUND and INSEMA trials demonstrated non-inferiority of SLNB omission in HR+/ Background: ILC accounts for ~10-15% of breast cancer (BC) cases in the US with
HER2- early-stage breast cancer (BC) patients with negative axillary ultrasound (AUS) undergoing difficult-to-diagnose patterns of metastasis. Loss of E-cadherin (CDH1) is the patho-
breast conserving therapy (BCT). However, it remains unclear if this approach may be applied to genomic feature of ILC. Tumor-informed ctDNA assays detect molecular residual disease
patients receiving mastectomy. We evaluated outcomes in a similar patient population receiving BCT
in patients who completed definitive therapy for early-stage BC. However, little is known
versus mastectomy. Methods: Using an institutional database (2010-2023), we identified patients
with cT1N0M0 HR+/HER2- unifocal invasive ductal carcinoma with negative AUS who underwent about the ctDNA dynamics and its prognostic relevance in ILC specifically. Here, we
upfront BCT (breast conserving surgery + Radiation therapy [RT]) or mastectomy +/- RT. All patients assessed actionable genomic alterations and correlated ctDNA dynamics with clinical
underwent SLNB. Clinicopathologic and treatment characteristics were compared by surgery type as outcomes in patients with HR+/HER2- ILC. Methods: We utilized Natera’s proprietary
well as clinical outcomes. Results: Among 1,506 patients, median age was 59 years (IQR 51-67), and real-world database linked to commercially available claims data and commercial ctDNA
80% were postmenopausal. BCT was performed in 78.2% (n=1,178) while 21.8% (n=328) underwent testing via a clinically validated, personalized, tumor-informed mPCR-NGS ctDNA assay
mastectomy. Patients who underwent mastectomy were more likely to be younger, premenopausal, (Signatera, Natera, Inc.) to identify patients with early-stage, HR+/HER2- ILC. Tumors with
with larger and higher grade tumors (Table). SLN positivity rate was 7.7% with no significant dif- CDH1 truncated alterations from tumor whole exome sequencing were categorized as ILC.
ferences by surgery type (7.6% vs 8.2%, p=0.684). Only 4 (1.2%) and 2 (0.2%) patients had pN2 disease
HR+/HER2- tumors were categorized using treatment regimens from insurance claim
in the mastectomy and BCT cohorts, respectively. More patients undergoing mastectomy had axillary
lymph node dissection compared to patients undergoing BCT (6.96% vs 0.3%, p,0.001). Among
codes. Targetable alterations in PIK3CA, AKT, PTEN, BRCA1, BRCA2, ESR1, NF1, and
mastectomy patients, 10.1% received RT with higher rates in node-positive patients (52.9% vs 5.1%, ERBB2 were assessed. ctDNA positivity rates and distant recurrence-free survival (DRFS)
p,0.0001). Similarly, chemotherapy use was higher among node-positive mastectomy patients were evaluated using longitudinal ctDNA status (ctDNA + or -). The timing of ctDNA+ after
(76.5% vs 22.5%, p=0.0001). Among SLN-positive postmenopausal patients with 21-gene recurrence primary BC surgery was divided into two categories: ,2 years (y) and .2y. Results: 430
score (RS) ,25, chemotherapy use was 16.6% (2/12) after BCT and 0% (0/4) after mastectomy. When patients with early-stage HR+/HER2- ILC and ctDNA testing were identified. The most
RS was .26, chemotherapy use was 60% (27/45) after BCT and 85.7% (12/14) after mastectomy. common targetable alterations co-mutated with CDH1 were in PIK3CA (54.4%), AKT1
Median follow up was 25.3 months (IQR 13.2-58.9). There were 3 (0.2%) axillary recurrences with 2 (5.1%), NF1 (4.9%), and ERBB2 (4.6%). The first ctDNA test was performed on 258 (59.3%)
recurrences after mastectomy (1 of whom received RT). There were 7 (0.46%) distant recurrences, with and 172 (39.5%) patients within or after 2y surgery, respectively. Of 430 patients, 88
2 following mastectomy. Conclusions: Among patients with cT1N0M0 HR+/HER2- invasive ductal
carcinoma with negative AUS, positive SLNs were identified in 7.7% with only 0.46% having pN2
(20.2%) had .1 ctDNA+ test. A ctDNA+ result was reported in 31/258 (12%) and in 57/172
disease. Low axillary and distant recurrence rates were observed regardless of surgery type. However, (33.1%) patients with testing within or after 2y surgery, respectively. Longitudinal ctDNA+
given that nodal status impacts RT use among mastectomy patients, further research investigating was associated with DRFS based on distant secondary malignant neoplasm claim codes.
omission of SLNB in this cohort is warranted. Research Sponsor: None. Among patients with a ctDNA- test within 2y, only 2.83% (6/212) presented a DRFS event
compared to 48% (12/25) of ctDNA-+ cases (Odds ratio 31, p-value=1.42E-9). In patients
BCT (N, %) TM (N, %) p-value
with a ctDNA- test after 2y, only 2.06% (2/95) presented a DRFS event compared to 37.5%
Age (median, IQR) 60.8 (53.2-67.8) 55.7 (46.9-64.9) ,0.00001 (9/24) of ctDNA+ cases (Odds ratio 27, p-value=4.92E-6). Conclusions: To our knowl-
Premenopausal 198 (16.8) 101 (30.8) ,0.00001
Grade 3 146 (16.7) 66 (22.8) 0.035
edge, this is the first large-scale analysis of the landscape of targetable genomic al-
cT stage 0.002 terations and matching longitudinal ctDNA and outcomes in patients with early-stage,
T1mic 14 (1.2) 12 (3.7) HR+/HER2- ILC. We find that PIK3CA is most frequently co-mutated with CDH1, and a
T1a 56 (4.8) 19 (5.8) comparison with the CDH1 wildtype cohort will be presented. These data provide insights
T1b 376 (31.9) 80 (24.4)
T1c 732 (62.1) 217 (66.1) into the ability of ctDNA to detect recurrence earlier, including sites associated with
pT size (median, IQR) 1.4 (0.9-4) 1.35 (0.8-2.1) 0.022 challenges in the interpretation of imaging results in the early-stage ILC setting. Research
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 25s

582 Poster Session 583 Poster Session

TROP2 overexpression as predictor of outcome in patients with early triple- Disparities in demographics and outcomes of breast cancer in females
negative breast cancer. Exploratory analysis from the GEICAM_CIBOMA undergoing mastectomy in rural vs. urban teaching centers. First Author:
trial. First Author: Federico Gustavo Rojo Todo, Hospital Universitario Fundación Harkaran Shergill, Maulana Azad Medical College, New Delhi, India
Jiménez Dı́az; CIBERONC-ISCCIII; GEICAM Spanish Breast Cancer Group, Barcelona, Background: Breast cancer (BC) sensitization, awareness, and increased screening
Spain have led to significant improvements in prevention, management, and survival rates over
Background: Antibody-drug conjugates (ADCs) have emerged as a promising therapeutic the last three decades. Recent reports have shown that the use of prophylactic and
strategy for triple-negative breast cancer (TNBC), a subtype with limited treatment options therapeutic mastectomy is on the rise. Previous studies have revealed healthcare
and poor prognosis. Understanding the biological and prognostic/predictive implications of disparities between rural (RC) and urban teaching centers (UTC), which may influence
biomarkers for these ADCs could help refine appropriate adjuvant treatment for high-risk outcomes. In this retrospective study, we propose to explore the presence of similar
TNBC. TROP2-targeted ADC sacituzumab govitecan has demonstrated significant im- disparities among BC patients undergoing mastectomy in rural hospitals vs. UTC.
provements in progression-free and overall survival in clinical trials involving the advanced Methods: For this study, we extracted procedures of mastectomy among BC females
TNBC setting. We sought to explore the prognostic/predictive value of TROP2 expression by through the hospitalization records of the National Inpatient Sample (NIS). We stratified
immunohistochemistry (IHC) in early TNBC patients (pts) from the GEICAM_CIBOMA trial our sample into procedures performed at RC and UTC. Procedural and postprocedural
(NCT00130533). Methods: We evaluated TROP2 expression by IHC using an anti-TROP2
complications were compared through multivariable regression models. Results: We
monoclonal antibody (clone ERP20043, Abcam) in tumors from a subset of 70 TNBC pts
studied 75915 BC patients who underwent mastectomy between 2016 and 2022. Around
included in the trial. These patients received either 6 months therapy with capecitabine (n=35)
or observation (n=35) after receiving standard (neo)adjuvant chemotherapy (trial recruitment
93.2% of our sample involved procedures performed at UTC, with 6.8% conducted at RC.
between 2006 and 2011). Semi-quantitative Histoscore (H-score) was estimated to consider Procedures at RC involved an older group(mean age 66.18 years vs. 57.19 years, p ,
the following TROP2 expression categories: H-score 0 to ,100: TROP2 low; H-score 100-200: 0.01), with a higher Charlson Comorbidity Index (CCI) score(mean score 3.79 vs. 3.53,
TROP2 medium; H-score $200: TROP2 high. Median TROP2 H-score was also explored for p , 0.01). An estimated 91.6% of all procedures were performed on an elective ba-
pts categorization. Cox regression models were assessed to predict DRFS (primary endpoint), sis(88.8% of rural and 91.9% of UTC, p , 0.01). Metastasis was present in 25.5% of all
DFS and OS (secondary endpoints). Multivariate models were adjusted for confounding cases (25.4% of UTC and 25.8% of rural cases, p = 0.591). Mastectomies performed at RC
factors, including histological grade, stage, chemotherapy regimen, and treatment. had higher odds of bleeding (aOR 1.170, 95% CI 1.075-1.272, p , 0.01), sepsis (aOR
Results: The TROP2 H-score median value was 165. Medium/high TROP2 expression (H- 2.163, 95% CI 1.557-3.004, p , 0.01), postprocedural respiratory failure (PPRF) (aOR
score $100) was observed in 27 (77%) pts treated with capecitabine, and 25 (71%) in the 2.667, 95% CI 1.479-4.808, p , 0.01), need for mechanical ventilation (MV) (aOR 2.732,
observation group. Higher TROP2 expression was associated with higher histological grade 95% CI 1.845-4.046, p , 0.01), ischemic stroke (aOR 5.073, 95% CI 2.884-8.922, p ,
(p=0.032). Medium/high TROP2 expression significantly associated with better DRFS 0.01), and cardiac arrest (aOR 8.443, 95% CI 4.861-14.666, p , 0.01). Acute kidney injury
(univariate analysis, HR=0.41; 95%CI 0.19-0.90; p=0.026; multivariate analysis, HR=0.24; 95% events were similar (aOR 1.160, 95% CI 0.945-1.423, p = 0.155). The odds of all-cause
CI 0.10-0.60; p=0.002). This prognostic value was confirmed at DFS (multivariate analysis, death were higher among RC procedures (aOR 5.401, 95% CI 3.456-8.442, p , 0.01).
HR=0.33; 95%CI 0.14-0.77; p=0.010) and OS (multivariate analysis, HR=0.29; 95%CI 0.11- Conclusions: The findings of this study have significant implications for healthcare
0.76; p=0.011). High TROP2 expression by median value (H-score $165) was also associated policies. BC patients undergoing mastectomy in rural areas were significantly older and
with better clinical outcome (DRFS univariate analysis, HR=0.43; 95%CI 0.19-0.97; p=0.041; had a higher CCI score. Moreover, rural procedures reported higher risks of bleeding,
multivariate analysis, HR=0.44; 95%CI 0.19-1.02; p=0.057). TROP2 expression categorization sepsis, stroke, cardiac arrests, PPRF, MV use, and death. These results highlight the
did not demonstrate predictive value of capecitabine benefit (DRFS interaction with treat- need for additional studies to establish the causes of these disparities, which may reflect
ment, p=0.852). Conclusions: In this GEICAM_CIBOMA trial sub-study, TROP2 over- the need for improving healthcare services in rural areas. Furthermore, encouraging RC
expression by IHC was observed in 74% of the analyzed cases and significantly associated
to set up review protocols on their adverse events through Ishikawa diagrams may help
with better clinical outcome in early TNBC pts, in terms of DRFS, DFS, and OS. Clinical trial
identify probable healthcare inequities compared to UTC, which can then be remedied.
information: NCT00130533. Research Sponsor: Sociedad Espa~ nola de Oncologia Medica
SEOM (SEOM/FECMA Grant 2022). This research has the potential to influence policy changes that could improve outcomes
for rural BC patients. Research Sponsor: None.

584 Poster Session 585 Poster Session

Changes in the degree of satisfaction and quality of life in breast cancer Bilateral mastectomy and breast cancer morality for invasive lobular car-
patients who are candidates for breast conservation but opted for mastec- cinoma: A SEER-based study. First Author: Richa Jaiswal, Women’s College
tomy: A single-center prospective study. First Author: Changhoon Lee, Hospital, University of Toronto, Toronto, ON, Canada
Department of Surgery, Seoul National University Hospital, Seoul National University Background: Many women with unilateral breast cancer opt for bilateral mastectomy.
College of Medicine, Seoul, South Korea While removing the unaffected contralateral breast lowers the risk of second primary
Background: Randomized trials have shown that breast-conserving surgery (BCS) and cancers, there is no benefit on breast cancer mortality. Studies have not investigated
total mastectomy (TM) provide comparable survival outcomes for early breast cancer whether this holds true for invasive lobular carcinoma (ILC). To estimate the 20-year risk
patients. Consequently, the choice of surgery is now considered a preference-based of breast cancer mortality in women with stage I-III unilateral ILC and compare survival
shared decision-making process. While studies have examined quality of life (QoL) in TM outcomes between unilateral lumpectomy, unilateral mastectomy and bilateral mas-
and BCS patients, it remains unclear how choosing TM over BCS affects postoperative tectomy. Methods: This retrospective cohort study used the Surveillance, Epidemiology,
QoL and satisfaction. This single-center prospective trial investigated the longitudinal and End Results (SEER) database to identify women diagnosed with unilateral invasive
impact of TM on QoL and satisfaction in the BCS-eligible patients who voluntarily opted lobular carcinoma (ILC) between 2000 and 2020. The cohort was followed for up to 20
for TM. Methods: We prospectively screened newly diagnosed breast cancer patients years to assess contralateral breast cancer and breast cancer-specific survival. We
who are eligible for BCS based on the preoperative clinical assessment between Nov estimated crude mortality rates, 20-year cumulative breast cancer mortality, and hazard
2021 and June 2024 at Seoul National University Hospital. Patients were allowed to ratios by surgical treatment group. Kaplan-Meier was used for cumulative risk, and Cox
choose between TM and BCS through the shared decision-making process, and were proportional hazards models for unadjusted and adjusted hazard ratios with 95%
enrolled according to the selected surgical method. A series of surveys were conducted confidence intervals. P-values , 0.05 were considered significant. Results: We
using the BREAST-Q questionnaire and Decision Regret Scale (DRS) before and between identified 58,861 women with unilateral ILC. Of which, 34,561 (59%) had lumpectomy,
6-24months following surgery and then compared the results between two groups. 18,894 (32%) had unilateral mastectomy, and 5406 (9.1%) had bilateral mastectomy. The
Results: A total of 68 patients were enrolled; 28 opted for TM, and 40 underwent BCS, mean age (in years) was 64 611 for unilateral lumpectomy, 62 613 for unilateral
with two patients lost to follow-up in the TM group. The TM group was significantly older mastectomy, and 57611 for bilateral mastectomy (p , 0.0001). The mean tumour size
than the BCS group (57.92 vs. 51.43 years, p=0.015), and had larger preoperative tumor was smallest in the lumpectomy group (1.961.5 cm) compared with 3.66 3 cm in both
sizes (25.77mm vs. 17.88mm, p=0.001). Postoperatively, satisfaction with the breast unilateral and bilateral mastectomy groups (p , 0.0001). The 20-year cumulative in-
(62.0 vs. 53.3, p=0.013) and physical well-being (chest) (83.9 vs. 77.5, p=0.004) de- vasive contralateral breast cancer risk was 7.3% for lumpectomy, 7.5% for unilateral
creased significantly. Satisfaction with the breast was significantly lower in the TM mastectomy, and 0.3% for bilateral mastectomy. The 20-year cumulative breast cancer
group compared to BCS (62.7 vs. 38.7, p,0.001). Psychosocial well-being (80.3 vs. 58.1, mortality was 13.4% in the lumpectomy group, 30.2% in unilateral mastectomy group
p,0.001), sexual well-being (55.3 vs. 24.2, p,0.001), and physical well-being (80.9 vs. and 24.3% in bilateral mastectomy group. However, after adjusting for demographic,
72.9, p=0.043) were all lower in the TM group. When comparing pre- and postoperative clinical, and treatment variables, we observed no difference in breast cancer mortality
scores, the TM group experienced greater declines in breast satisfaction (-1.38 vs. rates among unilateral mastectomy patients versus lumpectomy patients (adjusted
-21.34, p=0.007), psychosocial well-being (5.71 vs. -8.40, p=0.004), and sexual well- hazard ratio [aHR], 1.01; 95% CI, 0.94–1.08), and a statistically significant reduction in
being (1.55 vs. -17.83, p=0.004). The TM group also had significantly higher DRS scores breast cancer mortality rates among bilateral mastectomy patients compared to
(21.92 vs. 13.52, p=0.036), indicating greater regret. Conclusions: Our data demon- lumpectomy patients (aHR, 0.90; 95% CI, 0.82–1.00; p value 0.04). Conclusions: In this
strates that patients who opted for TM over BCS for their breast cancer surgery may cohort of invasive lobular breast cancer, bilateral mastectomy patients had a signifi-
experience more profound decline in QoL and higher degree of regret after their surgery. cantly lower risk of contralateral breast cancer and, after adjusting for differences in the
These findings should be incorporated into the shared decision-making process when surgical treatment groups, had a 10% lower rate of breast cancer mortality as compared
choosing optimal surgical treatment in early breast cancer patients who are eligible for to lumpectomy patients. Research Sponsor: None.
BCS. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
26s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

586 Poster Session 587 Poster Session

Evaluating the outcome of cardiovascular risk factors in breast cancer Gene expression signatures (GES) derived from digital histology to predict
patients. First Author: Michelle Koifman, The Brooklyn Hospital Center, Brooklyn, NY pathologic complete response (pCR) to neoadjuvant chemotherapy (CT) in
Background: Breast cancer patients are at increased risk of cardiovascular disease due ISPY2 and other trial/real world cohorts. First Author: Frederick Matthew Howard,
to factors such as treatment-related toxicity and shared comorbidities. This study University of Chicago, Chicago, IL
analyzes demographics and outcomes of cardiovascular risk factors in hospitalized Background: GES predictive of response to therapy across multiple breast cancer subtypes are
breast cancer patients using nationwide data to guide improved care strategies. commercially available or in development. Deep learning models can predict GES from digital his-
Methods: We used the National Inpatient Sample Database (2016 to 2020) to perform tology, and may serve as a lower-cost alternative immediately available at the time of biopsy.
Methods: Transformer-based models trained to predict 38 distinct breast cancer signatures from
the analysis. ICD-10 codes were used to make the primary diagnosis of patients with pathology (all with Pearson correlation . 0.5 versus true GES) were previously developed using cases
breast cancer admitted to hospitals, and secondary diagnosis of various cardiac risk from The Cancer Genome Atlas. These models were applied to digital H&E from pre-treatment biopsies
factors. We used the chi-square test and the student’s T-test to analyze categorical and from HER2- cases treated with CT or CT + immunotherapy (IO) from the ISPY2 trial. The histology-
continuous variables. Multivariable regression analysis was used to adjust for con- derived GES most predictive of pCR in ISPY2 (as per area under the ROC curve [AUROC]) was tested in
founders and calculate adjusted odds ratio (aOR). Statistical significance was set at two external neoadjuvant cohorts - a subset of a trial from Yale of durvalumab + CT (NCT02489448)
p ,0.05. Results: A total of 855394 patients were detected with primary diagnosis of with TIL annotations, and patients receiving standard of care CT at University of Chicago. AUROC
breast cancer. Out of them, 11% had coronary artery disease (CAD), 13% had congestive significance was assessed with 1000x bootstrapping, with Benjamini Hochberg correction applied in
ISPY2 to account for testing multiple GES models. Tertiles of predicted expression calculated in ISPY2
heart failure (CHF), 6% had atrial fibrillation, 1% had ventricular arrhythmias and 0.06%
defined groups with low, medium, and high likelihood of pCR; these cutoffs were tested in the external
had endocarditis. The prevalence of cardiac risk factors was mostly in older adults. The cohorts. Results: Accuracy for pCR prediction was tested in 578 patients from seven arms of ISPY2 –
mean age of patients with CAD and CHF was 73 years, atrial fibrillation was 75 years, with breakdown by treatment and hormone receptor (HR) status shown in Table. A histology model
ventricular arrhythmias was 69 years and endocarditis was 71 years. The mortality was for a GES defined by estrogen regulated genes (Oh et al, JCO 2006) – including proliferation, apoptosis,
found to be higher in patients with cardiac risk factors as well. The patients with and interferon-response genes – predicted pCR with the highest AUROC (0.794) in ISPY2, and
coronary artery disease had aOR of 1.9 (p-value 0.02), congestive heart failure had 1.3 outperformed a logistic regression fit on grade, HR status, and tumor / nodal stage (AUROC 0.705, p for
(p-value ,0.001), atrial fibrillation had aOR 1.1 (p-value 0.001), for ventricular ar- comparison 0.0001). Tertiles of predicted expression for this GES (computed in ISPY2) identified
rhythmias was 3 (p-value ,0.001) and for endocarditis was 1.4 (p-value 0.3). groups with low / high pCR rates which were robust to treatment, HR status, and consistent in
validation cohorts (Table). This digital signature (AUROC 0.737) compared favorably to pathologist TIL
Conclusions: Breast cancer patients with cardiovascular risk factors experience sig- annotation (AUROC 0.664) from the external Yale cohort. An explainability tool demonstrated that
nificantly higher mortality rates compared to those without. These findings emphasize patterns of lymphocytic infiltrate and poor differentiation contributed to high signature predictions
the need for early cardiovascular risk assessment and proactive management in breast from histology. Conclusions: A digital histology-derived GES consistently identifies patients at low /
cancer patients to improve outcomes and reduce complications. Research Sponsor: high likelihood of pCR with neoadjuvant CT or CT + IO, and may improve treatment personalization.
None. Research Sponsor: National Cancer Institute/U.S. National Institutes of Health.
Outcomes of cardiovascular disease in breast cancer patients. % pCR (low % pCR (mid % pCR (high
Subgroup n AUROC p expression) expression) expression)
With Cardiac Risk Without Cardiac Risk Adjusted Odds Adjusted Odds
Mortality Factors Factors Ratio Ratio ISPY2 579 0.794 2 x 10-28 7.6 26.7 58.6
ISPY2, CT + IO 459 0.810 3 x 10-26 8.0 28.2 64.1
Coronary Artery 4.7% 4.5% 1.9 0.02 ISPY2, CT only 120 0.726 0.001 6.4 20.0 36.8
Disease ISPY2, HR- 239 0.704 4 x 10-7 14.8 29.2 58.4
Congestive Heart 6.6% 4.3% 1.3 ,0.001 ISPY2, HR+ 340 0.817 1 x 10-15 6.5 24.8 59.0
Failure UChicago HR- 151 0.746 3 x 10-7 11.1 27.3 55.1
Atrial Fibrillation 5.5% 4.5% 1.1 0.01 UChicago HR+ 63 0.847 1 x 10-5 5.5 12.0 70.0
Yale HR- 41 0.737 0.005 0.0 50.0 61.5
Ventricular 12% 4.5% 3.0 ,0.001
arrhythmias
Endocarditis 6.6% 4.6% 1.4 0.3

588 Poster Session 589 Poster Session

Neoadjuvant penpulimab combined with taxanes and carboplatin in triple- Neoadjuvant low-dose carboplatin and docetaxel in combination with tor-
negative breast cancer: A single-arm, open-label, multi-center phase II ipalimab for early or locally advanced triple-negative breast cancer (Neo-
clinical study (neoTAPPL). First Author: Wenting Yan, Department of Breast and TOP): A single-arm phase 2 trial. First Author: Jun Tang, Sun Yat-sen University
Thyroid Surgery, Southwest Hospital, Army Medical University; Key Laboratory of Cancer Center, Guangzhou, China
Minimally Invasive Surgery and Precision Treatment for Breast Cancer of Chongqing Background: Neoadjuvant immunotherapy combined with chemotherapy improves the
Municipal Health Commission, Chongqing, China rate of pathological complete response (pCR) and prognosis of triple-negative breast
Background: The integration of immunotherapy with neoadjuvant chemotherapy has been cancer (TNBC). However, chemo-immunotherapy treatment is frequently associated with
shown to enhance pathologic complete response (pCR) and survival outcomes in patients with adverse events, resulting in dose reduction and delays. Approaches to deescalate chemo-
triple-negative breast cancer (TNBC). Nonetheless, additional research is required to ascertain immunotherapy without compromising outcomes for TNBC patients are needed. Evidence
the optimal neoadjuvant regimen. Here we present a prospective phase II NeoTAPPL trial in suggests that low-dose carboplatin potentiates the anti-tumor effect of PD-1 inhibitors
which evaluated the efficacy and safety of penpulimab (anti-PD-1 antibody) in combination with through many mechanisms. The NeoTOP trial aimed to assess the efficacy and safety of
taxanes and carboplatin for TNBC patients. Methods: In this open-label, multi-center phase II low-dose carboplatin and docetaxel in combination with toripalimab as neoadjuvant
study, patients with untreated, histologically confirmed TNBC in stage II-III were enrolled. therapy for early or locally advanced TNBC. Methods: This is a single-arm, open-label,
Patients received 6 cycles of neoadjuvant therapy with penpulimab (200 mg, d1, q3w) plus phase 2 trial. Patients with untreated stage IIA-IIIC TNBC were enrolled. Eligible patients
taxanes (docetaxel 75 mg/m2 or nab-paclitaxel 260 mg/m2, d1, q3w) and carboplatin (AUC=6, d1, received carboplatin (AUC=4), docetaxel (75 mg/m2) and toripalimab (240 mg), every
q3w). Patients who either completed or discontinued the neoadjuvant treatment would undergo
21 days for a total of 6 cycles. Toripalimab (240mg, every 3 weeks) continued post-
breast surgery. Adjuvant chemotherapy and immunotherapy were at the discretion of the
operatively for a further 11 cycles. The primary endpoint was the rate of pCR (ypT0/Tis
treating physician, and radiation therapy was per standard of care. The primary endpoint was the
rate of pCR based on the definition of ypT0/Tis ypN0. Secondary endpoints included residual
ypN0). The secondary endpoints included safety, objective response rate (ORR), residual
cancer burden (RCB), event free survival (EFS), overall survival (OS), adverse events (AE), and cancer burden (RCB) rate, event-free survival and overall survival. This study used Simon’s
immune response biomarkers. Results: 50 patients were enrolled, among which 37 patients two-stage design. Results: From January 2022 to September 2024, 51 patients were
received neoadjuvant treatment and underwent breast surgery. The median age was 51 years enrolled. Among them, 29 patients (56.9%) had stage II, and 22 (43.1%) had stage III. Four
(range, 32-72). 33 (89.2%) patients had stage II breast cancer at diagnosis. 21 of the 37 patients patients prematurely discontinued study treatment due to adverse events (three patients)
achieved pCR (56.7%; 95% CI, 40.9%-71.3%), and 29 patients achieved RCB 0-1 (78.4%; 95% CI, or tumor progression (one patient), including two discontinued toripalimab only and two
62.8%-88.6%). The ORR and DCR were 86.5% (95% CI, 72.0%-94.1%) and 91.9% (95% CI, 78.7%- discontinued both toripalimab and chemotherapy. One of the four patients who dis-
97.2%), respectively. Subgroup analysis showed that 60.6% (20/33) patients with stage II had continued treatment achieved a pCR when proceeding to surgery. One patient achieved
achieved pCR, 25% (1/4) patients with stage III reached this outcome. The pCR rate was 56.5% clinical CR after neoadjuvant therapy and refused to receive surgery. After surgery, pCR in
(13/23) in patients with negative lymph nodes, and 57.1% (8/14) in those with positive lymph both breast and lymph nodes was achieved in 29 of 50 patients, resulting in a pCR rate of
nodes. Treatment-emergent adverse events (TEAEs) of any grade occurred in all 37 pts, in which 58.0%. For all the 51 patients who received at least one dose of neoadjuvant therapy, the
20 (54.1%) were grade $3. The most common grade $3 TEAEs were neutropenia (43.2%), ORR was 90.2%. The proportions of RCB-0, RCB-1, RCB-2, and RCB-3 were 58%, 12%, 22%,
leukopenia (24.3%), anemia (21.6%), and thrombocytopenia (18.9%). 15 patients (40.5%) ex- and 8%. All 51 (100%) patients reported any grade of treatment-related adverse events
perienced immune related adverse events (irAEs), all of which were hypothyroidism. (TRAEs). Grade $3 TRAEs occurred in 23 (45.1%) patients. Serious adverse events were
Conclusions: In this trial, we demonstrated that an anthracycline-free neoadjuvant regimen reported in 5 (9.8%) patients. The regimen was well tolerated, and no new toxicity signals
consisting of penpulimab, carboplatin and taxanes in TNBC showed promising antitumor ef- were noted. Conclusions: The combination of low-dose carboplatin, docetaxel, and tor-
ficacy and manageable safety profile. The study is still ongoing. Clinical trial information: ipalimab showed promising efficacy and manageable safety, suggesting the feasibility of
ChiCTR2300071925. Research Sponsor: Development Center for Medical Science & Technology this regimen in the neoadjuvant setting for TNBC. Further randomized phase III trials are
National Health Commission of the People’s Republic of China WKZX2023CX010002.
warranted. Clinical trial information: NCT06618014. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 27s

590 Poster Session 591 Poster Session

Post-treatment MRI to predict pathological complete response in triple- Efficacy and safety of neoadjuvant TQB2102 in women with locally ad-
negative breast cancer following neoadjuvant chemoimmunotherapy. First vanced or early HER2-positive breast cancer: A randomized, open-label,
Author: Toulsie Ramtohul, Institut Curie, Paris, Ile de France, France multi-centre phase 2 trial. First Author: Junjie Li, Shanghai Cancer Center, Shanghai,
Background: Neoadjuvant chemoimmunotherapy (NACI) has significantly improved China
pathological complete response (pCR) rates in early-stage triple-negative breast cancer Background: Standard neoadjuvant regimens for HER2-positive breast cancer include
(TNBC). However, the predictive accuracy of post-treatment MRI for pCR remains unex- trastuzumab and pertuzumab combined with chemotherapy, and the efficacy and safety
plored. Our objective was to assess the performance of post-treatment MRI in predicting of third-generation HER2-direted antibody-drug conjugate (ADC) is unknown. TQB2102
pCR in TNBC patients treated with NACI. Methods: In this prospective, multicenter study is an anti-HER2 antibody-drug conjugate that targets two non-overlapping epitopes of
(August 2021–June 2024), women with early-stage TNBC were recruited from three centers. HER2 (ECD2 and ECD4). It consists of a humanized HER2 IgG1 bispecific antibody
Post-treatment dynamic contrast-enhanced (DCE) MRI data were analyzed across multiple conjugated to a topoisomerase I inhibitor via a cleavable linker, and the DAR value is 6.
vendors. The predictive performance of radiological complete response (rCR) on MRI for Methods: This open-label, randomized, multi-centre phase 2 study enrolled HER2-
pCR was evaluated using the area under the curve (AUC) of receiver operating charac- positive patients aged 18-75 years with stage II–III disease. Patients were randomly
teristics. A multivariable logistic regression model incorporating rCR, nodal involvement, assigned to receive neoadjuvant TQB2102 6mg/kg every 3 weeks for 6 cycles or for 8
and Ki-67 levels was developed and validated. For patients with residual enhancement, a cycles. The primary endpoint was pathological complete response (pCR). Safety was
radiomics score was generated using first-order and shape-based features. Results: The analysed in patients who received at least one dose of study medication.
study included 175 women in a training set from centers #1 (mean age 49 6 11 years) and
Results: Between 05 February 2024 and 24 Sep 2024, we randomly assigned 52 patients
84 in an external test set from centers #2 and #3 (mean age 52 6 12 years). MRI rCR
to neoadjuvant TQB2102 6 cycles (Arm A, n=26), 8 cycles (Arm B, n=26). The baseline
achieved an AUC of 0.83 (95% CI: 0.75–0.92) for pCR prediction. A combined model with rCR,
nodal status, and Ki-67 levels yielded an AUC of 0.88 (95% CI: 0.81–0.96) in the test set.
characteristics were well balanced; approximately 50% of the patients were hormone
Among patients with rCR, no nodal involvement, and Ki-67 .30%, the false-positive rate was receptor (HR)-positive, and 63% of the patients were stage III. The pCR rate was 57.7% in
3.6% and 3.5% in the training and test sets, respectively, with all cases limited to Residual Arm A (95%CI 36.9%-76.7%), 76.9% in TQB2102 Arm B (95%CI 56.3%-91%). In patients
Cancer Burden-I. For patients with residual enhancement, a model incorporating a radiomics with HR positive disease, the pCR rate was 53.8% in Arm A and 58.3% in Arm B; in
score and lesion count achieved an AUC of 0.80 (95% CI: 0.69–0.90). Conclusions: Post- patients with HR negative disease, the pCR rate was 61.5% and 92.9%. Grade 3 or higher
treatment MRI effectively predicts pCR in early-stage TNBC after NACI, suggesting its adverse events occurred 23.1% in Arm A, and 30.8% in 8 Arm B, 7.69% with increased
potential role in identifying candidates for breast cancer surgery omission trials. Research alanine aminotransferase and aspartate transferase. Dose reduction rate and discon-
Sponsor: Institut Curie. tinuation was 3.8% and 19.2% in Arm A, 3.8% and 23.1% in Arm B, and no treatment-
related deaths occurred. Conclusions: This is the first study to report the efficacy and
Accuracy of rCR for predicting pCR.
safety of third-generation duel-HER2-directed ADC in the neoadjuvant setting for HER2-
Set Subgroup Sensitivity (%) Specificity (%) PPV (%) NPV (%) F1 score (%)
positive breast cancer. TQB2102 is highly efficient and well tolerated. Clinical trial
Training N0 and 81 (54/67) 86 (12/14) [67, 99] 96 (54/56) 48 (12/25) 88 information: NCT06198751. Research Sponsor: None.
Ki- [71,90] [92, 99] [28, 68]
67.30% pCR in All patients and by HR status.
External N0 and 82 (28/34) 92 (11/12) [76, 99] 97 (28/29) 65 (11/17) 89 6 cycles 8 cycles
test Ki- [70, 95] [90, 99] [42, 87]
67.30% All 57.7% 76.9%
HR positive 53.8% 58.3%
Note-Unless otherwise specified, the data are presented as percentages, with the numbers of participants in HR negative 61.5% 92.9%
parentheses and 95% CIs in brackets; pCR = pathological complete response; PPV = positive predictive
value; NPV = negative predictive value.

592 Poster Session 593 Poster Session

Early on-treatment (on-Rx) tumor volume reduction (TVR) to predict re- Use of artificial intelligence (AI)–powered spatial analysis to predict path-
sponse to the KEYNOTE-522 (KN-522) regimen in early stage triple negative ologic complete response (pCR) in HR+ HER2- breast cancer (BC) patients
breast cancer (TNBC). First Author: Clinton Yam, The University of Texas MD treated with neoadjuvant chemotherapy (NAC). First Author: Dae-Won Lee, Seoul
Anderson Cancer Center, Houston, TX National University Hospital, Seoul, South Korea
Background: Monitoring clinical response by breast ultrasound (US) during neoadjuvant therapy is considered standard Background: In the KEYNOTE(KN)-756 study, adding pembrolizumab to NAC increased pCR in
of care. We previously demonstrated that suboptimal on-Rx TVR after neoadjuvant doxorubicin and cyclophosphamide high-risk HR+ HER2- BC. There is an unmet need to discover who will benefit from adding immune
(AC) predicts non-pCR after sequential taxane-based chemo. However, it is unknown if on-Rx TVR has the similar
predictive value in pts receiving the KN-522 chemo-immunotherapy regimen. Methods: Pts with early stage TNBC
checkpoint inhibitors (ICIs). This study aims to investigate the role of pCR in HR+ HER2- BC and to
planned to receive the KN-522 regimen were enrolled on the prospective ARTEMIS trial (NCT02276443). Breast US was identify whether AI-powered TIL analysis could predict pCR in patients treated with NAC without
performed at baseline and after 6 weeks of paclitaxel + carboplatin + pembrolizumab. TVR was defined as the percent ICIs. Methods: This is a single-center study conducted in Seoul National University Hospital,
reduction of tumor volumes calculated using 3 perpendicular measurements of the index breast lesion. Pathological Korea. H&E whole-slide images (WSIs) of archival breast tumor tissues at diagnosis were analyzed
complete response (pCR) was defined as ypT0/isN0. Logistic regression was used to examine associations between by Lunit SCOPE IO, an AI-powered spatial TIL analyzer. Tumors with a high proportion of area with
covariates and pCR. Receiver operating characteristic (ROC) analyses were utilized to assess the predictive value of TVR
high intratumoral TIL were classified as immune-inflamed, those with a high proportion of area with
and determine an optimal TVR threshold. Results: 150 pts were included. Clinicopathological characteristics are
described in Table 1. The pCR rate was 63%. In uni- and multi-variable analyses, TVR was the only covariate to low intratumoral but high stromal TIL as immune-excluded, and the remaining as immune-desert.
demonstrate statistically significant association with pCR (aOR:1.9 per 10% TVR, p,0.001). In ROC analyses, the area Recurrence risk prediction AI model was trained and validated on independent 1,552 H&E WSIs to
under the ROC curve (AUC-ROC) was 0.74 (95% CI: 0.66-0.82). TVR.50%, selected based on the Youden index, predicted predict OncotypeDx score. Patients with histologic grade 3 and tumor size $2 cm with node-
pCR with the following performance characteristics: positive predictive value: 73%; negative predictive value: 79%; positive status, or tumor size $5 cm were classified as those eligible for KN-756 study. Results: A
sensitivity: 94%; specificity: 41%. Conclusions: Early on-Rx TVR by breast US outperforms clinicopathological covariates total of 425 BC patients who were treated with NAC without ICIs between January 2015 and
in the prediction of pCR in pts with TNBC receiving the KN-522 regimen and should be leveraged for risk stratification and
design of response-adapted neoadjuvant clinical trials for pts with TNBC. Clinical trial information: NCT022766443. October 2018 were included. The median age was 47 (range 24-80), 67.1% had stage III disease,
Research Sponsor: Conquer Cancer, the ASCO Foundation; 12238. 93.6% had node-positive disease, and 23.5% were eligible for KN-756 study. pCR was achieved in
57 (13.4%) patients and was higher in patients with whom were eligible for KN-756 study (20% vs
pCR Non-pCR p value Adjusted OR p value
(n=95) (n=55) Odds ratio (OR) (univariable) (aOR) (multivariable) 11.4%, p= 0.041). Patients who achieved pCR had better 5-year event free survival (EFS, 92.9% vs
Median 6w TVR – % (interquartile range [IQR]) 84 (72-90) 69 (38-83) 1.5 <0.001 1.9 <0.001
77.7%, p= 0.010). AI-powered spatial analysis was performed in 340 patients. There were 125
Median age – years (IQR) 51 (40-61) 51 (43-64)
N (%)
0.98 0.25 1.0 0.73 (36.8%) with immune-desert tumor, 138 (40.6%) with immune-excluded tumor, and 77 (22.6%) with
Ethnicity immune-inflamed tumor. Patients with inflamed tumor had higher pCR rate compared to those with
White 50 (53) 33 (60) 1 1
Black 15 (6) 10 (18) 1.1 0.84 0.9 0.91 immune-excluded or immune-desert in the whole population (Table) and in patients not eligible for
Hispanic/Latino
Asian
25 (26)
5 (5)
6 (11)
6 (11)
2.4
0.6
0.08
0.36
1.9
0.2
0.32
0.12
KN-756 study (25.6% vs 11.9% vs 5.7%, p = 0.013). Patients with inflamed or excluded tumor had
T stage better EFS compared to desert group (80.9% vs 71.2%, p = 0.0275). In addition, recurrence
T1/2 79 (83) 46 (84) 1 1
T3/4 16 (17) 9 (16) 1.0 0.94 1.2 0.77 prediction model showed that those who were predicted to have OncotypeDx score $ 26 had
Nodal status
Positive 31 (33) 24 (44) 1 1
higher pCR rate (18.2% vs 4.3%, p= 0.002). Conclusions: Achieving pCR was associated with
Negative 64 (67) 31 (56) 1.60 0.18 1.4 0.55 favorable EFS in HR+ HER2- patients treated with NAC. Patients with immunogenic tumor mi-
Germline BRCA status
Mutant 8 (8) 2 (3) 1 1 croenvironment had higher pCR rate and better EFS. Investigating the role of immune checkpoint
Wild Type 84 (88) 51 (93) 0.41 0.27 0.3 0.31
Unknown 3 (3) 2 (4) inhibitors according to tumor microenvironment may be promising. Research Sponsor: None.
Histology
Ductal
Metaplastic
87
3
(92)
(3)
48 (87)
5 (9)
1
0.33 0.14
1
0.3 0.26
Desert Excluded Inflamed
Other 4 (4) 2 (4) 1.1 0.91 2.0 0.59 (N=125) (N=138) (N=77) p-value
Unknown 1 (1) 0
Histologic grade pCR 8 (6.4%) 18 (13.0%) 23 (29.9%) ,0.001
2 13 (14) 14 (25) 1 1
3 81 (85) 41 (75) 2.1 0.08 2.0 0.59 Stage III 85 (68.0%) 88 (63.8%) 51 (66.2%) 0.32
Unknown 1 (1) 0 LN + 114 (91.2%) 132 (95.7%) 72 (93.5%) 0.34
Ki67
£35% 5 (5) 8 (15) 1 1 HG 3 21 (16.8%) 31 (22.5%) 38 (49.4%) ,0.001
>35% 67 (71) 38 (69) 2.82 0.09 4.6 0.09 KN-756 eligible 20 (16.0%) 29 (21.0%) 34 (44.2%) ,0.001
Unknown 23 (24) 9 (16)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
28s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

594 Poster Session 595 Poster Session

SHR-A1811 as neoadjuvant therapy for HR-positive, HER2-low breast can- Computational pathology-informed immune biomarker for trastuzumab
cer: A single-arm, phase II clinical study. First Author: Zhenzhen Liu, Department benefit in HER2+ breast cancer: Validation in NSABP B-41 clinical trial.
of Breast Disease, Henan Breast Cancer Center, The Affiliated Cancer Hospital of First Author: Satvika Bharadwaj, Wallace H. Coulter Department of Biomedical Engi-
Zhengzhou University & Henan Cancer Hospital, Zhengzhou, China neering, Emory University and Georgia Institute of Technology, Atlanta, GA
Background: Human Epidermal Growth Factor Receptor (HER2)-low breast cancer ac- Background: Trastuzumab, a targeted therapy, considerably improves survival outcomes in
counts for approximately 55% of all breast cancer cases, and it is more prevalent in hormone HER2+ breast cancer patients (pts). However, identifying pts most likely to benefit from tras-
receptor-positive (HR+) breast cancer. Recent breakthroughs in anti-HER2 antibody-drug tuzumab and those who might avoid it, remains a challenge. Data from the NSABP B-41 ran-
conjugates (ADCs) have revolutionized the therapeutic landscape for breast cancer domized clinical trial which evaluated lapatinib-containing regimens against trastuzumab, each
treatment, particularly for HER2-low breast cancer. SHR-A1811 is an anti-HER2 ADC and has given concurrently with chemotherapy in the neoadjuvant setting, holds potential for compu-
demonstrated acceptable tolerability and encouraging antitumor activity in HER2-low tational pathology to uncover treatment response patterns. In this study, we present DeSTIL
advanced breast cancer. Therefore, we conducted the phase 2 trial to investigate the (Density and Spatial Architecture of Tumor Infiltrating Lymphocytes [TILs]), a predictive bio-
efficacy and safety of SHR-A1811 as a neoadjuvant treatment in patients with HR+/HER2- marker derived from Hematoxylin and Eosin (H&E) images, to evaluate the immune microen-
vironment and identify HER2+ pts that benefit from trastuzumab. Methods: Digitized and quality-
low breast cancer. Methods: This is an open-label, two-stage, phase II clinical trial. In the
controlled H&E slides from HER2+ pts in The Cancer Genome Atlas (TCGA; n=175) were used to
first stage, 35 participants will be enrolled, and if at least 18 participants achieve an
develop DeSTIL and was independently validated on NSABP B-41 (n=221) pts. Following nuclei
objective response rate (ORR), the trial will proceed to the second stage, enrolling an segmentation, TILs were identified, and their density and spatial architecture features were
additional 31 subjects. Eligible patients are women aged 18-70 years; treatment-naı̈ve; quantified. A lasso-regularized Cox proportional hazards model was used to select features and
histologically confirmed invasive breast cancer stage cT2-3/N0-2M0; HR+/HER2-low, and compute a continuous DeSTIL risk score, which was then dichotomized at the median into
the expression of Ki-67 exceeds 14%. Eligible patients receive SHR-A1811 as neoadjuvant DeSTIL-positive and DeSTIL-negative groups in the training set. The locked model was validated
therapy. SHR-A1811 is administered intravenously at a dose of 6.4 mg/kg once every three on NSABP B-41 to predict event-free survival (EFS) in pts receiving neoadjuvant chemotherapy
weeks for a total of eight cycles. The primary endpoint is ORR. Secondary endpoints include with trastuzumab, lapatinib, or a combination of trastuzumab and lapatinib. Within the DeSTIL-
safety, residual cancer burden (RCB) 0-1, and pathologic complete response (pCR). stratified groups, treatment-specific progression was analyzed to evaluate the potential benefit of
Results: A total of 66 patients enrolled in this study. The median age was 49 years, with trastuzumab-based therapies. Results: Among 221 HER2+ pts from the NSABP B-41 trial, 61 pts
84.8% (56/66) aged $ 40 years. Of all patients, 66.7% (44/66) were premenopausal, 66.7% (28%) were classified as DeSTIL-positive and 160 (72%) as DeSTIL-negative, based on the median
(44/66) had node-positive disease, 86.4% (57/66) had stage II, 13.6% (9/66) had stage IIIA, training threshold. DeSTIL-positive pts demonstrated a significant benefit with the
and 74.2% (49/66) had HER2 expression of IHC 2+/FISH-, while 25.8% (17/66) were IHC 1+. trastuzumab-alone arm compared to the combination regimen (HR=0.09, 95% CI=0.01-0.77,
In the modified intention-to-treat population (mITT, patients who received at least one cycle p=0.0061) (interaction term p=0.024) and against lapatinib plus the combination regimen
of study treatment and at least one post-baseline MRI assessment), 81.5% (53/65) of (HR = 0.11, 95% CI = 0.01-0.9, p = 0.01) (interaction term p = 0.05). In contrast, no sig-
patients achieved an ORR, and two patients achieved a pathological complete response, nificant benefit was observed in DeSTIL-negative pts when comparing trastuzumab to the
resulting in a pCR rate of 3.1% (2/65). Additionally, the proportion of patients with RCB 0 or combination regimen (HR=1.33, 95% CI=0.47-3.75, p=0.5840) or to lapatinib plus the
RCB I was 9.3% (6/65). 97% (64/66) of patients experienced at least one treatment-related combination regimen (HR=1.01, 95% CI=0.45-2.30, p=0.9701). Conclusions: DeSTIL, a
adverse event (TRAE). Grade 3 or higher TRAEs occurred in 39.4% (26/66) of patients, with biomarker based on the density and spatial architecture of TILs, may help identify HER2+
pts more likely to benefit from trastuzumab. Further validation through prospective trials is
the most prevalent being neutropenia (27.3%, 18/66), leukopenia (16.7%, 11/66), and anemia
warranted. Additionally, this biomarker offers a practical framework for comparing HER2-
(13.6%, 9/66). No interstitial lung disease (ILD) and treatment-related deaths were reported.
targeted therapies, with the potential to minimize unnecessary treatments, reducing as-
Conclusions: As a neoadjuvant treatment, SHR-A1811 achieves a significant improvement sociated cardiotoxicity and financial burden. Research Sponsor: U.S. National Institutes of
in ORR and brings both pCR and RCB 0-I benefits in patients with HR+/HER2-low breast Health; R01CA268287A1, R01CA26820701A1, R01CA249992-01A1, R01CA202752-01A, 1R43
cancer. These outcomes support further exploration of SHR-A1811 in this patient pop- EB028736-011, R01CA208236-01A1, R01CA216579-01A1, R01CA220581-01A1, R01CA25
ulation. Clinical trial information: NCT05911958. Research Sponsor: None. 7612-01A1, 1U01CA239055-01, 1U01CA248226-01, 1U54CA254566-01, 1R01HL15127701A1,
R01HL15807101A1, W81XWH-19-1-0668.; sponsored research agreements from Bristol
Myers-Squibb, Boehringer-Ingelheim, Eli-Lilly and Astrazeneca.

596 Poster Session 597 Poster Session

Exploration of the new classification of hormone receptor-positive breast Analysis of event-free survival (EFS) of stromal tumor
cancer: Based on the genomic landscape of 2111 early to mid-stage Asian infiltrating lymphocytes (sTILs), PD-L1 expression, and their early dynamics
patients. First Author: Cuiyun Zhang, The Affiliated Cancer Hospital of Zhengzhou in the NeoTRIP trial. First Author: Giampaolo Bianchini, Università Vita-Salute San
University and Henan Cancer Hospital, Zhengzhou, China Raffaele, Milan, Italy
Background: The molecular classification of breast cancer using omics data has led to Background: We demonstrated that pre-treatment sTILs and PD-L1, and on treatment sTILs
varied prognoses; however, with the increase of targeted therapy options, current but not PD-L1 were associated with pathological Complete Response (pCR) to neoadjuvant
classifications may not adequately guide treatment strategies. Methods: This study therapy in patients (pts) with triple-negative breast cancer (TNBC) enrolled in NeoTRIP
analyzed 2,111 HR+ Asian breast cancer patients using next-generation sequencing of (NCT02620280) trial (Bianchini G ESMO 2020). Here we assess the association between the
1,021 genes for targeted therapy classification. Results: The majority of patients had same biomarkers and EFS. Methods: NeoTRIP randomized 280 pts to nab-paclitaxel/carbo for
invasive ductal carcinoma (67.0%) and were at stages 0 to III (89.6%), with 29.7% being 8 cycles (CT) or with atezolizumab (CT/A). As Per-Protocol Population, 258 pts were evaluable
HER2 positive. Variations were widespread, with SNV/Indel mutations found in 99.4% and for EFS, the primary endpoint of the [Link] collected samples at baseline (n=258/258; 100%)
and on Day 1 Cycle 2 (D1C2) (n=230/258; 89.2%]. We centrally assessed sTILs (cut-off $30% to
CNV detected in 44.6% of the samples, respectively. The key signaling pathways, in-
be considered high) and PD-L1 expression (SP142) on immune cells (IC) (IC$1% considered
cluding PAM, RTK/RAS, DDR, and TP53, each showed a prevalence of 50% or higher and
positive). Association with EFS was investigated. Imaging mass cytometry (IMC) (Wang Nature
co-existed. Interestingly, 76.1% of patients had actionable mutations, predominantly in 2023) was used to investigate the biology linked to PD-L1 dynamic over-expression.
the PAM (54.8%), RTK/RAS (22.2%), and DDR (11.7%) pathways, which were found to be Results: Median follow-up was 54 months. Pre-treatment high sTILs and PD-L1+ were as-
mutually exclusive (p , 0.01). Subsequently, Patients were categorized into two groups, C sociated with lower risk of recurrence in CT arm (HR 0.35 [0.12-0.99], p=0.049 and HR 0.31
and M, based on CNV presence, revealing significant differences in the prevalence of [0.15-0.64], p=0.001, respectively). In CT/A arm neither pre-treatment marker was significantly
actionable genes and pathways. The C group stood out for its high prevalence of PAM associate with EFS. At D1C2 in CT arm, high sTILs but not PD-L1+ was significantly associated
(43.7%) and HER2 (38.4%), while the M group exhibited prominence in PAM (64.1%) and with lower risk of recurrence (HR 0.34 [0.15-0.81], p=0.015 and HR 0.65 [0.28-1.48], p=0.30,
DDR (15.0%). Notably, responses to neoadjuvant treatment (NAT) varied as well, with respectively). In CT/A arm both high sTILs and PD-L1+ were strongly associated with lower risk
higher pCR rates (51.9% vs. 22.1%, p , 0.001) for the combination of chemotherapy and of recurrence (HR 0.23 [0.07-0.78], p=0.019 and HR 0.25 [0.11-0.57], p=0.001, respectively).
HER2-targeted therapy (Chemo + HP) and lower pCR rates (6.0% vs. 15.2%, p = 0.017) for Only in CT/A arm, on-treatment PD-L1 provided prognostic information independent of baseline
chemotherapy alone (Chemo) observed in the C group compared to those in the M group. biomarkers, on-treatment sTILs and pCR (adjHR = 0.25 [0.11-0.58], p=0.001). In patients with
Additionally, subsequent subgroup analyses based on actionable signaling pathways baseline PD-L1- tumors and paired D1C2 samples (n=87), conversion from PD-L1- to on-
between the two groups showed similar patterns, with the C-HER2+ subgroup having the treatment PD-L1 positivity occurred in 64.3% and 17.8% in CT/A and CT arms, respectively (p=
highest pCR rate (64.1%) for Chemo + HP and M-PAM--DDR- subgroup showing the 7.39x10-6). The tumors converted to PD-L1+ status were significantly associated with better
highest pCR rate (21.5%) for Chemo. Besides, clinical characteristics among the various outcome in CT/A arm (HR 0.23 [0.08-0.68], p=0.008) but not in CT arm (HR 0.82 [0.24-2.85],
groups and subgroups showed differences. For instance, the M-PAM--DDR+ subgroup had p=0.76). Notably, in CT/A arm, patients with baseline PD-L1- tumors had similarly low pCR rate
an earlier onset (p , 0.001) and included more premenopausal individuals (p=0.034), regardless of PD-L1 status on D1C2 (20% and 25.9% in PD-L1- and PD-L1+ D1C2 groups,
respectively). The IMC results for the biological characterization of tumors with induction of PD-
along with higher ER and PR expression (p , 0.001) and lower Ki-67 levels (p , 0.001).
L1+ will be presented. Conclusions: In NeoTRIP on-treatment high sTILs (both CT and CT/A
Multivariate logistic regression analysis identified HER2 amplification, PR negativity, and
arms) and PD-L1 positivity (only in CT/A arm) were significantly associated with lower risk of
TP53 mutation as independent risk factors for the efficacy of Chemo + HP NAT. recurrence independently of pCR and baseline biomarkers. The findings suggest up-regulation
Conclusions: The study suggested that classifying breast cancer patients based on CNVs of PD-L1 as new candidate pharmacodynamic biomarker of benefit from atezolizumab.
and actionable pathways may enhance neoadjuvant pCR rates. This underscored the Whether this observation hold also for pembrolizumab remain to be defined. Research Sponsor:
potential implications for expanding neoadjuvant treatment strategies and facilitating Fondazione AIRC per la ricerca sul cancro (AIRC); IG 2018 - ID. 21787, Grant to GB; Breast
tailored precision treatment options based on distinct molecular profiles and therapy Cancer Research Foundation (BCRF); 20-181, Grant to LG; Breast Cancer Research Foundation
responses. Research Sponsor: None. (BCRF); 18-181, Grant to LG; Fondazione AIRC per la ricerca sul cancro (AIRC); IG 2024 - ID.
30919, Grant to GB.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 29s

598 Poster Session 599 Poster Session

Long-term outcomes in triple-negative breast cancers (TNBC) treated with Neoadjuvant cadonilimab (anti-PD-1/CTLA-4 bispecific antibody) plus che-
talimogene laherparepvec (TVEC) in combination with neoadjuvant chemo- motherapy in early or locally advanced triple-negative breast cancer: A
therapy (NACT). First Author: Hatem Hussein Soliman, H. Lee Moffitt Cancer Center single-arm phase II trial (CABIN study). First Author: Tao Wu, Changde Hospi-
and Research Institute, Tampa, FL tal, Xiangya School of Medicine, Central South University, Changde, China
Background: TVEC is an engineered herpes simplex oncolytic virus (HSV OV) approved Background: Neoadjuvant chemotherapy for triple-negative breast cancer (TNBC) can
for the treatment of melanoma. We published a phase 1/2 trial combining TVEC with improve surgical outcomes; however, many patients fail to achieve pathological complete
NACT in early stage TNBC demonstrating increased pathologic complete response (pCR) response (pCR). Immune checkpoint inhibitors (ICIs) targeting PD-1/PD-L1 show promise,
compared to expected rates with NACT. We are presenting updated long term follow up but responses are limited. This phase II trial evaluates cadonilimab (AK104), a bispecific
data on this cohort of both phase 1 and 2 evaluable patients. Methods: Stage II-III TNBC PD-1/CTLA-4 antibody, combined with nab-paclitaxel and carboplatin in neoadjuvant TNBC
pts were enrolled into a single arm, optimal phase 1/2 trial with TVEC (10^6 PFU 1st dose to improve pCR rates and assess safety. Methods: This phase II, single-arm trial enrolled
then 10^8 PFU x 4 doses) weeks 1,4,6,8,10 + weekly paclitaxel (80mg/m2) IV x 12, patients aged 18-75 with treatment-naı̈ve, stage IA-IIIC triple-negative breast cancer
followed by dose dense AC (doxorubicin/cyclophosphamide 60/600 mg/m2) IV q2weeks (TNBC). Patients received cadonilimab (10 mg/kg) plus nab-paclitaxel (260 mg/m²) and
x 4 alone (wT-AC) given preoperatively. Primary endpoint was pCR rate. Secondary carboplatin (AUC 4) every 3 weeks for six cycles, followed by surgery within 4 weeks.
endpoints included 5 year disease free survival/overall survival rates (DFSR/OSR), Biopsy specimens were collected before treatment for whole-exome sequencing. The
safety, immune correlates. Results: Forty six patients were enrolled at Moffitt (5/2018 – primary endpoint was total pathological complete response (tpCR) rate (ypT0/is ypN0).
4/2020) and evaluable for response and outcomes. Study demographics: median age 49 Secondary endpoints included objective response rate (ORR), disease control rate (DCR),
(27-66), 69.5% White, 13% Black, 13% Hispanic, clinical stage II 80% and III 20%, node + disease-free survival (DFS), event-free survival (EFS), overall survival (OS), and safety.
45%. The pCR rate for the phase 1/2 cohort was 45.6% (95% CI 30.9-60.9). Additionally, Results: A total of 29 patients were enrolled from January 31, 2023, to January 20, 2025.
10 patients had residual cancer burden (RCB) 1 responses (associated w/ favorable The median age was 53 years, and 11 patients (37.9%) were premenopausal. The majority
outcomes) 21.7% (95% CI 10.9-36.3%). At median follow up of 70 months (range 17-98), had positive nodal involvement (75.9%) and stage III (51.7%) disease. of the 29 patients, 19
six patients have had a breast cancer recurrence DFSR=86.9% (95% CI 73.7-95.0) and (65.5%, 95% CI: 45.7-82.1) achieved total pathological complete response (tpCR), and 21
four patients died OSR=91.3% (95% CI 79.2-97.6). DFSR in pCR group = 95.2% (95% CI (72.4%, 95% CI: 52.8-87.3) achieved breast pCR (bpCR). Tumor downstaging occurred in
89.7% (T) and 62.1% (N) of patients. Radiological assessments showed an objective
76.1 – 99.9) and non-pCR group = 80% (95% CI 59.3 – 93.1%). All but one of the
response rate (ORR) of 93.1% (95% CI: 77.2-99.2) and a disease control rate (DCR) of 100%.
recurrences occurred in patients with non-PCR responses (RCB 2-3) to TVEC+NACT.
All patients experienced at least one treatment-related adverse event (TRAEs), with 51.7%
Clinical stages at presentation for patients with recurrences were 5 stage II and 1 stage
having grade 3/4 events, most commonly neutropenia (37.9%), platelet count decrease
III. No patients had any HSV reactivation or autoimmunity events during the post study
(19.2%), and leukopenia (17.2%). Serious TRAEs occurred in 6.9%. Immune-related adverse
surveillance period. Greater immune enrichment of B and T cell subsets in pCR vs. non- events (irAEs) of any grade occurred in 14 (48.0%) patients, while those of grade $ 3
pCR tumors was observed during TVEC treatment. Conclusions: To our knowledge, this occurred in only one patient (3.4%). The most common irAE was hypothyroidism (44.8%,
is the first report on longer term outcomes for early TNBC treated with OV. TVEC plus wT- 13/29), all were grade 1 or 2. Notably, patients with TASOR2 or MST1R mutation
AC demonstrates promising long term outcomes when compared to the more intensive demonstrated a significantly lower pCR rate (both P=0.041). Additionally, CCND1 am-
KEYNOTE 522 checkpoint regimen. Additional investigation of oncolytic viruses ad- plification showed a non-significant negative trend with the pCR rate (P=0.051).
ministered during NACT for TNBC is warranted to confirm this benefit. Clinical trial Conclusions: The combination of cadonilimab, nab-paclitaxel, and carboplatin in neo-
information: NCT02779855. Research Sponsor: Amgen. adjuvant treatment for stage IA-IIIC TNBC showed promising efficacy, with high response
rates and significant tumor downstaging, particularly in stage III patients. The regimen was
well tolerated with manageable adverse events, supporting its further investigation as a
potential neoadjuvant treatment for TNBC. Clinical trial information: ChiCTR2200067005.
Research Sponsor: None.

600 Poster Session 601 Poster Session

Biomarkers of neoadjuvant dalpiciclib in patients with operable HER2- Patient outcomes in WSG-ADAPT according to NATALEE and MonarchE risk
negative luminal B breast cancer in the DANCER trial. First Author: Yunxiang criteria. First Author: Oleg Gluz, West German Study Group and Ev. Hospital Bethesda,
Zhou, Department of Breast Surgery and Oncology, The Second Affiliated Hospital of Breast Center Niederrhein, Moenchengladbach, Germany and University Hospital Co-
Zhejiang University School of Medicine, Hangzhou, China logne, Cologne, Germany
Background: DANCER (NCT05640778) was a circulating tumor DNA (ctDNA)-directed, Background: In HR+/HER2- high-risk early breast cancer (eBC), abemaciclib and ribociclib
single-arm, phase II trial investigating the clinical activity of dalpiciclib combined with improve the efficacy of standard endocrine treatment (ET), as shown in MonarchE (abema-
aromatase inhibitors as a neoadjuvant regimen for operable human epidermal growth ciclib) and NATALEE (ribociclib). However, the absolute benefit varies according to prognostic
factor receptor 2 (HER2)-negative luminal B breast cancer. Although a high complete cell factors. We analyzed the outcome of prognostic groups based on MonarchE and/or NATALEE
cycle arrest (CCCA) rate (primary endpoint) of 86.7% (26/30) was achieved at 2 weeks inclusion criteria in the WSG-ADAPT trial considering Recurrence Score (RS, OncotypeDx) and
(T1), some patients showed suboptimal clinical responses after neoadjuvant therapy. Ki-67 response after preoperative ET. Methods: In WSG-ADAPT (NCT01779206), patients (pts)
This underscores the importance of identifying biomarkers predictive of response to with clinically high-risk HR+/HER2- eBC (cT2-4 or cN+ or G3 or Ki67 $ 15%) initially received a
3-week standard ET before surgery or sequential biopsy. Pts with c/pN2-3 or G3 with Ki67 .
CDK4/6 inhibitors. Methods: Plasma samples collected at baseline (T0), T1, mid-therapy
40% were randomized directly to chemotherapy (CT trial) evaluating (neo)adjuvant 4 3
(T2), surgery (S), and postoperatively (PO) underwent ctDNA and Olink proteomic an- paclitaxel q2w vs. 8 3 nab-paclitaxel q1w, followed by epirubicin + cyclophosphamide q2w,
alyses. Tumor tissues obtained at T0, T1, and S were assessed for somatic variation followed by ET. pN0-1 pts with RS 0-11 or RS 12-25 with ET-response (central Ki67postET #
profiling, immunohistochemical markers and MammaPrint index. Patients achieving 10%) received ET alone (ET trial); RS 12-25 pts without ET-response entered CT trial. Pts with
CCCA at both T1 and S with a concurrent objective response by MRI at S were classified N1-3 or T3 or T2, N0 with either Ki-67 $ 20% or G3 or RS . 25 were classified as NATALEE
as Good Responders (GRs, n = 15); others were Moderate Responders (MRs, n = 15). high-risk, pts with N2-3 or N1 with either T3-4 or G3 or Ki-67 $ 20% were classified as
Results: The baseline clinicopathological features of the patients were balanced be- MonarchE high-risk. Results: In the WSG-ADAPT ET trial, 303 (14.2%) and 784 (36.7%) of 2135
tween the GR and MR groups. Compared to MRs, the GRs at S exhibited significantly pts were classified as MonarchE and NATALEE high-risk, respectively. In the CT trial, 963
lower residual cancer burden (RCB) scores, preoperative endocrine prognostic index (43.2%) and 1572 (70.5%) of 2230 pts were classified as MonarchE and NATALEE high-risk.
(PEPI) scores, histological tumor grades, Ki67 expressions, and CA153 levels. Addi- After 60 months of median follow-up, both high-risk vs. low-risk classifications were highly
tionally, GRs demonstrated significantly higher Miller-Payne grades, tumor- prognostic for iDFS and dDFS in ET and CT trials. However, low-risk pts (by both classifications)
infiltrating lymphocyte levels, and tumor shrinkage rates. In terms of biomarkers, in the ET trial had 5-y iDFS and dDFS of 94.7% and 96.4%, respectively, vs. 90.1% and 93.6% for
GRs had a higher rate of ctDNA clearance at and prior to T2 (100.0% vs 54.5%; p = 0.045), high-risk by just NATALEE but not by MonarchE criteria (p = n.s.) and vs. 88.3% and 88.9% in
as well as higher levels of plasma CCL4 (p = 0.029), plasma CCL19 (p = 0.020), im- both NATALEE and MonarchE high-risk pts (p , 0.001). In the ET-only cohort, survival
munohistochemical pRb (p = 0.044), and immunohistochemical CDK4 (p = 0.034). outcomes were similar between pN0 and pN1 pts at high-risk by NATALEE but not by MonarchE
Furthermore, GSTM1 demonstrated a significant shift in its copy number pattern after criteria (5-y iDFS of 87.7% vs. 90.2%; p = n.s. and 5-y dDFS of 91.7% vs. 91.9%; p = n.s.). In the
treatment at S, with five previously detected baseline deletions no longer being identified CT trial, 5-y iDFS and dDFS rates were 93.9% and 94.9%, respectively, for NATALEE low-risk pts
vs. 84.7% and 87.0% for high-risk by NATALEE but not by MonarchE criteria and vs. 77.7% and
and five de novo amplifications emerging (p = 0.007). Lack of early ctDNA clearance was
79.6% for MonarchE high-risk pts. Conclusions: Among 4365 pts in the WSG-ADAPT trial,
also significantly associated with RCB class of III and PEPI score of $ 4. Besides, subgroups classified as high-risk by NATALEE and MonarchE criteria had poor outcomes.
MammaPrint high-risk patients showed a significant increase in RCB and PEPI scores vs. However, N0-1 pts who were high-risk by NATALEE but not MonarchE criteria and pts with RS #
low-risk patients. Conclusions: Patients with operable HER2-negative luminal B breast 25 and/or ET response had only slightly inferior outcomes compared to low-risk pts with ET
cancer who exhibit early ctDNA clearance, MammaPrint low-risk status, GSTM1 deletion, therapy alone. Assuming a hazard ratio of 0.7 for a ribociclib effect, as shown in NATALEE, an
increased pRb/CDK4 expression, and higher plasma CCL4/CCL19 levels may derive absolute benefit of approx. 2% fewer dDFS events after 5 years can be assumed in this group
substantial benefit from neoadjuvant dalpiciclib therapy. Clinical trial information: based on the WSG-ADAPT experience. Shared decision-making will be key in this intermediate-
NCT05640778. Research Sponsor: National Natural Science Foundation of China; the Key risk group. Clinical trial information: NCT01779206. Research Sponsor: Genomic Health (Exact
Research and Development Program of Zhejiang Province. Science); Celgene; Amgen; Allgemeine Ortskrankenkasse.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
30s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

602 Poster Session 603 Poster Session

Racial disparities in clinical outcomes of early-stage triple-negative breast Outcomes of neoadjuvant endocrine therapy (NAET) versus neoadjuvant
cancer treated with neoadjuvant chemoimmunotherapy: Insights from the chemotherapy (NAC) in stage II-III invasive lobular carcinoma (ILC). First
NCDB. First Author: Zunairah Shah, Roswell Park Comprehensive Cancer Center, Author: Jason A. Mouabbi, The University of Texas MD Anderson Cancer Center,
Buffalo, NY Houston, TX
Background: Triple-negative breast cancer (TNBC) is an aggressive breast cancer (BC) Background: ILC is a distinct subtype of breast cancer, accounting for up to 15% of
subtype, and Black patients (pts) with TNBC have worse survival outcomes after neoadjuvant cases, and is characterized by unique clinical and pathological features. ILC often yield
chemotherapy, likely due to biological and socioeconomic factors. Neoadjuvant immune poor responses to NAC. Case reports suggest that NAET yields favorable outcomes in
checkpoint inhibitors combined with chemotherapy, i.e. neoadjuvant chemoimmunotherapy ILC. However, no large-scale study has evaluated the impact of NAET in ILC patients.
(NACI) have improved pCR rates and overall survival (OS), but its efficacy by race is unclear. This is the first large study that aims to compare the outcomes of NAC versus NAET in
This study evaluates racial disparities in clinical outcomes for pts with early-stage TNBC stage II/III ILC. Methods: A retrospective analysis was performed on patients treated at
treated with NACI, aiming to address this critical gap. Methods: We analyzed the National MD Anderson Cancer Center with a diagnosis of anatomical stage II-III estrogen
Cancer Database (NCDB) for pts with stage II/III TNBC treated with NACI from 2019 to 2022. receptor-positive (ER+) HER2-ve ILC in our prospectively collected and curated elec-
Primary outcomes included pCR and OS, which were analyzed with race using univariate and
tronic database. Data collected included demographics, receptor status (ER, PR, HER2),
multivariable logistic regression and Cox proportional hazards models, while adjusting for
treatments received, clinical anatomic and pathologic stage, type of surgery, surgical
clinicopathologic variables (age, stage, grade, comorbidities (Charlson-Deyo Comorbidity
Classification) and socioeconomic factors (residence (rural/urban area), insurance, income).
pathology outcomes, and recurrence. Endpoints included pathologic complete response
P value #0.05 was considered statistically significant. Results: A total of 5,137 pts were (pCR), modified Preoperative Endocrine Prognostic Index (mPEPI) score 0, endocrine
included. Median age was 51 years (range:39-63); 69.9% were White, 20.5% Black, 9.6% Other, therapy responsiveness (ETR; Ki67 #10%), rates of axillary downstaging (from node
49.3% had stage II and 50.7% had stage III TNBC. Median follow up was 26.6 months (3.3- positive to negative), rate of lumpectomy, rate of axillary lymph node dissection (ALND),
61.9), pCR was achieved in 76.5% pts, (White: 77%, Black: 74%, Other: 76%; p = 0.113). Pts and 10-year distant recurrence-free survival (10y DRFS). For DFRS, NAC patients were
achieving pCR had significantly higher 3-year OS (92% vs 72%, p,0.001) and 5-year OS (84% compared to NAET patients who did not receive adjuvant chemotherapy (ACT). Uni-
vs 56%, p,0.001) compared to those without pCR. Racial disparities in survival were ob- variate analysis identified variables associated with outcomes, and multivariate logistic
served, with 3-year OS of 88%, 84%, and 85% (p ,0.05) and 5-year OS of 83%, 77%, and 85% regression was planned for significant factors (p , 0.05). Results: We analyzed 611
for White, Black, and Others, respectively (p,0.05). After adjusting for covariates, Black pts patients among who the median age was 55 (range 30–91), with 77% White, 8.5%
had a trend toward lower likelihood of pCR compared to White pts although not statistically Hispanic, and 8.1% Black. 65% were postmenopausal and all patients received adjuvant
significant (odds ratio (OR) 0.76 [95% CI: 0.54–1.07]. The factors independently associated ET. 509 (80%) received NAC, and 102 (20%) received NAET. Among NAET patients, 83%
with worse OS were residence in rural areas (HR 1.79 [95% CI: 1.00–3.19], p = 0.05), received a non-steroidal AI, while 98% of NAC patients received anthracycline-containing
tumors $10 cm (HR 1.92 [95% CI: 1.21–3.06], p = 0.006), stage III disease (HR 1.91[95% CI: regimens. pCR was achieved in 13 NAC patients (2.5%) and 2 NAET patients (1.9%).
1.47–2.49], p , 0.001) and Black vs. White group (HR 1.42 [95% CI: 1.10–1.84], p = 0.007). mPEPI score 0 was attained by 10 NAET patients (9.8%), and the ETR rate was 83%.
Conclusions: Black pts with TNBC receiving NACI have worse OS than White pts, possibly due Axillary downstaging occurred in 11% of NAC patients and 5% of NAET patients. Rates of
in part to social, structural, or biological determinants of health. Further research is needed to lumpectomy (17.1% in NAC vs. 17.6% in NAET) and axillary lymph node dissection
investigate personalized treatment strategies that address the unique challenges Black pts (ALND) (68.6% in NAC vs. 62.6% in NAET) were comparable between the groups. With a
face in achieving long-term survival and improving overall prognosis. Research Sponsor: median follow-up of 85 months, 10y DRFS was 55% for NAC and 65% for NAET (HR 0.57,
None. 95% CI 0.33–0.98, P = 0.02). Univariate analysis for DRFS revealed no significant
Survival rates and hazard ratios by race. associations with age, race, or initial stage (all p . 0.05), precluding multivariate
Race 3-year survival rate (%) 5-year survival rate (%) Hazard ratio analysis. Conclusions: This large-scale analysis highlights the limited efficacy of NAC in
White 88% 83% 1.00 ILC, with minimal pCR rates. NAET yields promising results, including superior 10y DRFS.
Black 84% 77% 1.42 [95% CI 1.10–1.84] These findings underscore the potential of NAET as a viable neoadjuvant option for
Other race 85% 85% 1.19 [95% CI 0.82–1.74] patients with stage II/III ER+ ILC. Research Sponsor: None.

604 Poster Session 605 Poster Session

Serum estradiol (sE2) levels in premenopausal (PreM) women receiving Molecular insights into HR+/HER2+ early-stage breast cancer: Neoadjuvant
neoadjuvant ovarian function suppression (OFS) with the oral SERD amce- therapy responses by MammaPrint and BluePrint genomic subtypes. First
nestrant, alone, or in combination with letrozole or abemaciclib in the I-SPY2 Author: Laila Samiian, Baptist MD Anderson Cancer Center, Jacksonville, FL
Endocrine Optimization Pilot (EOP). First Author: Jo Chien, University of California, Background: Clinical HER2+ (cHER2+) early breast cancer (EBC) represents 15-20% of
San Francisco, San Francisco, CA invasive EBC and is typically treated with Neoadjuvant HER2-targeted therapy (NHT)
Background: The addition of OFS to standard endocrine therapy (ET) improves iDFS for preM combined with chemotherapy, regardless of ER status. NBRST and I-SPY2 trials showed
women with early-stage hormone receptor positive (HR+) breast cancer (BC). Incomplete OFS varied NHT responses in cHER2+ tumors based on genomically-defined molecular
occurs in a subset of women, the long-term clinical significance of which is unclear. This subtypes, emphasizing the importance of understanding tumor biology. Genomic assays
study evaluates local sE2 levels in preM women receiving OFS with an oral SERD in the MammaPrint (MP) and BluePrint (BP) predict therapy response and inform treatment
neoadjuvant (NA) setting. Methods: The EOP is a I-SPY2 sub-study investigating NA decisions. Here, we explored the biological pathways underlying differential NHT re-
endocrine-based strategies in pts with HR+ HER2- Stage 2/3 BC predicted to have lower sponse in triple positive (HR+HER2+) tumors using whole transcriptome analysis (WTA).
benefit from chemotherapy. Pts were randomized to oral amcenestrant 200 mg/d, alone or in Methods: Patients with HR+/HER2+ early-stage breast cancer treated with NHT (N = 720)
combination with either letrozole 2.5 mg/d or abemaciclib 150 mg bid. Additional pts were were included from FLEX (NCT03053193). MP classified tumors as UltraLow (UL), Low,
randomized to a standard of care (SOC) arm consisting of an aromatase inhibitor. All PreM pts High 1, or High 2 Risk, while BP categorized them as Luminal A, Luminal B, HER2, or Basal.
received OFS with monthly GnRHa up to 2 wks prior to C1D1 of study therapy. Pts were Differences in clinical characteristics and pathological complete response (pCR) rates
treated for 6 months prior to surgery. sE2 levels were collected locally prior to each GnRHa were assessed by Chi-Square or Fisher’s exact tests and proportional Z-test, respectively.
injection. Tumor Ki67 was assessed at 3 weeks and surgery, and associated with sE2 level. Differential gene expression (DGE) analysis of WT profiles was performed between tumors
Statistical significance was determined using a two-sided Student’s t-test and a significance
with and without pCR, using limma package in R, followed by pathway enrichment
level 0.05. Results: Between 5/2021-8/2022, 74 pts were enrolled to an amcenestrant
analysis in Metascape. Results: Among 720 HR+/HER2+ EBCs, MP classified 19 (2.6%)
containing arm, 40 of whom were preM. 38/40 pts had at least one local sE2 level measured in
as UL, 107 (14.9%) as Low, 385 (53.5%) as High 1, and 209 (29.0%) as High 2. BP classified
follow up. Of these 38 pts (median age 45 years), 37 suppressed sE2 into the postmenopausal
(postM) range at one or more follow-up timepoints. Of the 37 pts that completely suppressed 120 (16.7%) as Luminal A, 307 (42.6%) as Luminal B, 278 (38.6%) as HER2, and 15 (2.1%)
into the postM range, 6 pts had sE2 levels rebound into the preM range (mean 319 pg/mL, as Basal. Compared to other BP subtypes (Luminal A/B), BP HER2 tumors were asso-
range 20-848 pg/mL) with peak sE2 levels occurring at a median of 12 weeks from C1D1. One ciated with younger age (54 vs 60, p , 0.001), premenopausal status (p = 0.002), higher
pt never suppressed to the postM range (peak sE2 1227 pg/mL) and was found to have an grade (G3: 54.7%, p , 0.001), and T3 tumors (10.7% vs 3-4%, p , 0.001). pCR rates with
ovarian cyst after 3 months on amcenestrant, requiring surgery. One additional pt had NHT were higher in BP HER2 tumors compared to Luminal A/B tumors (n = 41, 61.2% vs n
multiple ovarian cysts (sE2 84 pg/mL). Median age of the 7 pts who had sE2 rebound was 43 = 18, 26.5%, respectively, p , 0.001). WTA of BP HER2 tumors with pCR showed 23 genes
years. There was no significant difference in Ki67 at 3 weeks and surgery between pts whose with 2-fold change (not statistically significant after correction), 20 of which were
sE2 levels remained suppressed (median Ki67 2.0%) compared to those whose estradiol upregulated and associated with regulation of bone morphogenic protein encoding genes
rebounded into the preM range or never suppressed (median Ki67 1.5%). In the 5 pts whose and increased cell-substrate/cell matrix adhesion, compared to tumors that had residual
sE2 rebounded to . 200 pg/mL, all pts had tumor Ki67 , 10% at 3 weeks and surgery. disease. Conclusions: These data show heterogeneity within HR+/HER2+ tumors, with
Between 12/2022 and 8/2023, 20 pts were enrolled to AI +/- OFS and 11 pts were preM. Of the approximately 60% genomically reclassified as non-HER2-type by BP. Consistent with I-
11 preM pts, all pts suppressed sE2 to the postM range. No SOC patients had sE2 levels SPY2, BP HER2 cancers showed higher pCR rates than Luminal A/B, suggesting that
rebound into the preM range. Conclusions: In this study of neoadjuvant oral SERD with additional therapeutic strategies are needed to increase the pCR rates in these cancers.
monthly OFS, 7/38 (18.4%) preM pts had sE2 levels remain or rebound into the preM range. Although WTA in BP HER2 tumors with and without pCR identified DGE, the findings were
sE2 levels did not appear to impact Ki67 suppression on NA ET. Work up for ovarian cysts not statistically significant. Future analyses of WTA in larger numbers of BP subtypes
should be considered in pts with symptoms, or significantly elevated or persistent sE2 levels. within the HR+ HER2+ EBC patients who are being enrolled on the FLEX trial may elucidate
Clinical trial information: NCT01042379. Research Sponsor: Quantum Leap Healthcare the biology of the cancers with pCR vs non-pCR. Clinical trial information: NCT03053193.
Collaborative. Research Sponsor: Agendia, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 31s

606 Poster Session 607 Poster Session

Results of the prospective randomized controlled trial VOG-01: Neoadjuvant HER2DX genomic test in HER2-positive breast cancer treated with 15 weeks
endocrine therapy ribociclib + fulvestrant + GnRH-a versus chemotherapy 4 of neoadjuvant paclitaxel, trastuzumab, and pertuzumab (THP): Final anal-
AC + 4 T for early HR+/HER2-negative breast cancer in premenopausal ysis from the BiOnHER clinical trial. First Author: Bartomeu Fullana Grimalt,
patients. First Author: Rashida Orlova, Saint Petersburg State University, Saint Avinguda de la Granvia de l’Hospitalet, 199-203, L’hospitalet De Llobregat, Barcelona,
Petersburg, Russian Federation Spain
Background: CDK4/6 inhibitors are used to treat HR+/HER2- metastatic breast cancer Background: HER2DX is a 27-gene genomic assay providing prognostic and predictive
in combination with endocrine therapy. However, there is still a lack of evidence about insights in early-stage HER2-positive (HER2+) breast cancer. Although widely validated,
combined endocrine therapy in neoadjuvant setting. VOG-01 is a phase II randomized less data exists for its performance in THP beyond the DAPHNe trial (JAMA Oncol 2023),
trial that evaluated the effects of combination ribociclib plus fulvestrant and GnRH-a as which included 80 patients (pts). This study aimed to validate HER2DX for predicting
neoadjuvant therapy in premenopausal patients. Methods: Premenopausal women with pathological complete response (pCR) with THP and compare its performance to
HR+/HER2-negative stage II-III were randomly assigned to Fulvestrant (500 mg on the hormone receptor (HR) status. Methods: HER2DX (Reveal Genomics) was centrally
1st, 15th, 28th days of the first cycle, then once every 28 days), Triptorelin (3.75 mg every evaluated on all tumor biopsies from the BiOnHER trial (NCT05912062), where pts with
28 days) and Ribociclib (600 mg daily, 3 weeks/4) during 16-24 weeks (NET), or stage I-III HER2-positive breast cancer received 15 weeks of neoadjuvant THP at the
Doxorubicin 60 mg/m2 + Cyclophosphamide 600 mg/m2 x4 21-day courses followed by Catalan Institute of Oncology. Biopsies were collected at pre-treatment baseline (D1)
Docetaxel 75 mg/m2 x4 21-day courses (NCT). Primary endpoint was objective response and day 8 (D8) after an HP loading dose but before initiating paclitaxel. HER2DX pCR
rate. Secondary endpoints were pathological response rate (RCB), frequency of breast- group cutoffs were based on predefined thresholds for HER2DX low-, medium-, and high
conserving surgery, severity of adverse events and the quality of life (EORTC QLQ-C30). groups. Logistic regression and receiver-operating characteristic (ROC) curve analyses
Results: Eighty two patients were recruited. The objective response rate was 83% in the were used for statistical evaluation. The primary objective was to assess HER2DX pCR
NET and 71% in the NCT (p = 0.2). Complete pathological response (RCB 0) was in 2 score for predicting pCR (ypT0/isN0). Secondary objectives included evaluating HER2DX
cases (5%) in the NET and in 4 cases (10%) in the NCT; RCB I was in 1 case (2.6%) in the performance by HR status, baseline TILs/Ki67, and additional insights from D8 data.
NCT and did not occur in the NET; RCB II - 55% in the NET and 62% in the NCT, RCB III was Results: HER2DX was successfully evaluated in all 83 enrolled patients. The cohort
in 40% and 26% respectively. Breast-conserving surgery were not so common in both included 65.1% T2 tumors, 62.7% cN0 status, 69.9% stage II disease, and 67.5% HR-
groups: 37% in the NET and 32% in the NCT (p=0.5). Severe adverse events (CTCAE ver 5 positive tumors. The overall pCR rate was 45.8% (95% CI, 34.9–57.0%), and the
G3-4) were 66% in the NET and 87% in NCT (p , 0.019). Quality of life significantly ypT1miN0 rate was 54.2%. HER2DX pCR score was significantly associated with pCR
decreased during NCT compared with NET: the total score was 75.7622.2, 42.0619.7, (odds ratio [OR] = 5.26, P , 0.001), with an AUC of 0.835. Patients were categorized into
66.3612.9 in NCT at visits on 1-12-24- weeks and 76.4620, 70.7625.4, 76.1620.5 in 35.0% low, 37.5% medium, and 27.5% high HER2DX pCR score groups, with pCR rates of
NET at the same visits (p , 0.05). Conclusions: For the first time randomised trial 13.3% (95% CI, 4.4–31.6%), 51.6% (95% CI, 33.4–69.4%), and 81.8% (95% CI,
comparing NET and NCT was conducted in premenopausal women with HR+/HER2- 59.0–94.0%), respectively. Among HR-negative tumors, pCR rates were 78.6% for high-
early breast cancer. NET was not inferior to standard NCT in terms of objective response pCR and 0.0% for low-pCR groups, while in HR-positive tumors, pCR rates were 87.5%
rate, complete or pronounced pathological response rate and breast-conserving surgery. and 13.8%, respectively. HR status alone was associated with pCR (OR = 0.125, P =
At the same time it was associated with a lower severity of adverse events and increased 0.006) but lost significance in multivariable analysis including HER2DX. Baseline Ki67
quality of life. Nevertheless new treatment approach requires confirmation of its ef- (median: 35.0%) and TILs (median: 8.0%) were not associated with pCR. While D8 data
fectiveness in large studies. Clinical trial information: NCT04753177. Research Sponsor: offered biological insights, it did not improve predictive performance beyond baseline
None. HER2DX. Conclusions: HER2DX is a robust predictor of pCR following neoadjuvant THP
in stage I-III HER2-positive breast cancer, outperforming HR status. Baseline TILs and
Ki67 were not predictive of pCR, and HER2DX D8 data did not improve predictive
performance. Clinical trial information: NCT05912062. Research Sponsor: Instituto de
Salud Carlos III (ISCIII); FIS PI20/00544.

609 Poster Session 610 Poster Session

Outcomes of elderly patients with early-stage triple-negative breast cancer Early adverse symptoms to predict response to treatment among patients in
treated with the KEYNOTE-522 regimen. First Author: Renata Colombo Bonadio, the I-SPY trial. First Author: Amrita Basu, University of California, San Francisco, San
Instituto D’Or de Pesquisa e Ensino (IDOR), S~
ao Paulo, Brazil Francisco, CA
Background: The KEYNOTE-522 regimen (neoadjuvant pembrolizumab combined with a four- Background: The adverse event (AE) landscape in oncology is changing due to the in-
drug chemotherapy backbone, followed by adjuvant pembrolizumab) is a standard of care for troduction of immunotherapy and antibody drug conjugates. These AEs come with both short
stage II-III triple-negative breast cancer (TNBC). However, the median age of TNBC diagnosis is and long-term symptoms that significantly impact patient quality of life. Monitoring for early
40–50 years, and elderly patients (pts) are underrepresented in clinical trials. Methods: The onset of symptoms could optimize therapy for a particular patient, maximizing potential
effectiveness and safety of the KEYNOTE-522 regimen were evaluated in patients aged $65 efficacy while mitigating toxicity. It is also possible that some toxicities are directly as-
years (y) in the Neo-Real/GBECAM-0123 trial, a real-world, multicenter study conducted across sociated with drug sensitivity. We sought to identify symptoms associated with pathologic
ten institutions since July 2020. Pathological complete response (pCR) was assessed as the complete response (pCR) using patient-reported outcomes (PROs) in early-stage high-risk
primary endpoint. Patients , 65y served as the control group. Results: Of the 413 pts included breast cancer patients. Methods: Our study population included 288 stage II/III high-risk
in the study, 45 (10.9%) were aged $65y. Compared to younger patients, elderly pts exhibited a breast cancer patients enrolled on the I-SPY2 trial from 2021-2024, who received novel
higher proportion of special histological types (15.6% vs 6.8%, P = 0.055), lower-grade tumors neoadjuvant therapies 6 standard paclitaxel. pCR was assigned if tumor was absent in
(grade 1–2: 35.5% vs 13.6%, P = 0.001), lower Ki-67 index ( , 50%: 46.7% vs 16.6%, P , 0.001), breast and nodes at surgery following neoadjuvant treatment. Patients (n = 288, pCR rate =
and fewer germline BRCA1/2 mutations (2.2% vs 13%, P = 0.039). The majority of patients in 29%, 89% administered immunotherapy) were sent electronic PROs. 33 patient-reported AEs
both groups had stage II disease (77.8% vs 69.3%, P = 0.574). Elderly pts were less likely to were measured using NCI’s Patient Reported Outcomes - Common Terminology Criteria for
receive dose-dense anthracycline and cyclophosphamide (AC) (44.4% vs 55.4%, P = 0.027). Adverse Events (PRO-CTCAE). Each symptom was evaluated using severity, frequency, and
Patients aged $ 65y had lower pCR rates compared to those , 65y (46.3% vs 64%; univariate interference on a Likert Scale. Presence of early PRO symptoms (cycles 1-3 of treatment)
logistic regression: OR 0.48, 95%CI 0.25–0.93, P = 0.030). However, this difference was not were binarized (at least one of moderate or greater), and odds ratios were computed with pCR
significant in multivariable analysis adjusted for histological type, tumor grade, Ki-67 index, as outcome. To assess whether higher grade AEs were enriched in patients that achieved a
BRCA status, and AC schedule (OR 1.80, 95% CI 0.74–4.37, P = 0.188). Elderly pts experienced
pCR, we also performed the Wilcoxan rank sum test using maximum (worst) symptom
significantly higher rates of safety concerns (Table 1). Conclusions: TNBC in elderly pts
severity. Results: Of 288 patients included in our analysis (median age = 48 years, range =
appears to have distinct biological characteristics, which may contribute to lower pCR rates
20-78), 203 (70.5%) were White, 17 (5.9%) were Asian, 33 (11.5%) were Black or African
with the KEYNOTE-522 regimen. Additionally, the higher incidence of safety issues in this
American, and 35 (12.2%) were Hispanic. PRO analysis revealed that patients that had
population underscores the importance of personalized treatment strategies and careful
moderate to severe muscle pain (27% vs 10% OR = 3.15, p , 0.05), joint pain (22% vs 8% OR =
patient selection. Further studies focused on elderly pts with TNBC are needed. Research
Sponsor: None. 3.23, p , 0.05), headache (27% vs 12.5% OR = 2.59, p , 0.05), or mouth/throat sores (16% vs
5% OR = 3.56, p , 0.05) within weeks 1-3 had higher odds of achieving a pCR. When we
Safety outcomes of elderly patients treated with KN522 regimen. looked at maximum severity between weeks 1-3, patients that achieved a PCR had higher
‡ 65y < 65y P grade muscle pain (p = 0.04), heart palpitations (p = 0.035), and significantly lower grade
Drug discontinuation due to AE 33.3% 21.7% 0.140 numbness and tingling (p = 0.002). Beyond 6 weeks, associations were weaker or insig-
AC discontinuation 17.8% 3.8% 0.001 nificant. Conclusions: Our study utilizes an analysis framework that was able to determine
Dose reduction 27.2% 12.2% 0.010 sentinel symptoms such as muscle and joint pain, mouth/throat sores and palpitations as
Delay for neoadjuvant treatment conclusion 42.2% 22.3% 0.006
Hospitalization due to AE 33.3% 15.2% 0.011 early as weeks 1-3 associated with increased efficacy. This may suggest an early immune
Antibiotics use 44.4% 25.5% 0.019 reaction in patients that eventually respond favorably to treatment. Our work can help
Grade ‡ 3 AE 48.9% 34.2% 0.137 provide earlier proactive monitoring to mitigate toxicities, treatment redirection if needed,
Anemia 8.9% 2.4% 0.042 and a potential symptom-based early understanding to personalize treatment efficacy. A
Neutropenia 35.5% 19.6% 0.020
Febrile neutropenia 22.2% 10.3% 0.026 similar analysis is underway to predict immune related AEs. Clinical trial information:
Fatigue 6.7% 1.4% 0.046 NCT01042379. Research Sponsor: National Cancer Institute.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
32s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

611 Poster Session 612 Poster Session

Automated prediction of response to neoadjuvant chemotherapy from dig- I-SPY2 endocrine optimization pilot (EOP): Neoadjuvant lasofoxifene (Laso)
itized H&E slides of pre-treatment core needle biopsies in INFORM (TBCRC in molecularly selected patients with hormone receptor positive (HR+)/HER2
031) patients with low stromal TILs. First Author: Stuart J. Schnitt, Brigham and negative (HER2-) stage 2/3 breast cancer (BC). First Author: Mei Wei, Huntsman
Women’s Hospital, Boston, MA Cancer Institute, University of Utah, Salt Lake City, UT
Background: We previously performed anautomated analysis of whole slide images (WSI) Background: EOP, an I-SPY2 sub-study, evaluates the tolerability and activity of novel endocrine
of H&E-stained pre-treatment core needle biopsies (CNBs) from patients (pts) enrolled in the strategies in stage 2/3 breast cancer (BC) patients (pts) predicted to have lower chemotherapy benefit.
INFORM phase II trial of neoadjuvant cisplatin vs doxorubicin-cyclophosphamide in HER2- Laso, a selective estrogen receptor modulator (SERM), has shown favorable toxicity profile and activity
negative germline BRCA carriers. That analysis demonstrated that a digital biomarker of in HR+/HER2- endocrine-resistant metastatic BC. Methods: Pts with Stage 2/3 HR+/HER2-, Mam-
maPrint (MP) low risk BC were enrolled. Pts with MP High1 BC were included if clinically node-
complex immune response (CmbI) which combines immune heterogeneity index (IHI),
negative. Pts received oral laso 5 mg daily for six 28-day cycles. Laso was continued until the day prior
proliferative, and cell cycle G1/S deregulation signatures, was significantly predictive of to surgery. Premenopausal pts received ovarian function suppression (OFS) starting C2D1. The
response (RCB 0,1) to neoadjuvant chemotherapy (NAC) in all pts, in sub-cohorts including primary endpoint was feasibility (.75% of patients completing .75% study therapy). Baseline (T0), 3-
TNBC, and in both therapy arms. Further, a lower IHI, indicating less heterogeneity of stromal wk (T1) biopsies, and the surgical specimen (T3) was assessed centrally for Ki-67. Breast MRI
tumor infiltrating lymphocytes (sTIL), was predictive of a better response to NAC (RCB 0,1), functional tumor volume (FTV) was performed at T0, T1, 12 weeks (T2), and pre-operatively (T3). Blood
whereas a higher IHI, indicating greater heterogeneity of sTIL was associated with a worse was collected for tumor informed ctDNA at T0, T1, T2, T3. Advance event (AE) was assessed using
response (RCB 2,3). The predictive performance of IHI alone was modest compared to CmbI CTCAE V5. Results: From 3/2023 to 5/2024, 20 pts were enrolled. Median age 50.5 years, 50%
but superior to sTIL. High sTIL is associated with favorable prognosis for NAC, especially in premenopausal, and 1 male pt. 60% cN0, 80% MP low-risk signature. 18 (90%) pts completed .75%
TNBC. However, the impact of heterogeneity of immune cell distribution on NAC response, study therapy. Two pts discontinued treatment due to pt preference. Median Ki67 at T0 was 14.7%. At
particularly in those with low sTIL is unknown. The current study evaluated if IHI could T1, 87.5% of pts remained or suppressed Ki67 to ,10% and 37.5% suppressed to ,2.7%. Ki67 at T1
augment sTIL assessment by identifying NAC responders in patients with tumors dem- was similar between pre-and postmenopausal pts despite OFS (Table). The median MRI FTV was 8.4 cc
at T0, and 3.4 cc at T3. Median % FTV reduction from T0 to T3 was -47.5%. 2/20 pts (10%) achieved a
onstrating low sTIL. Methods: CNBs scanned at 40x on a Hamamatsu Nanozoomer scanner
modified PEPI score of 0. No patients achieved completed pathological response. Of the 16 pts with
were evaluated using the 4D QPOR platform to generate IHI as a continuous index. Among RCB results, 2 (12.5%) RCB-1, 6 (37.5%) RCB-2, 8 (50%) RCB-3 disease. 14 pts had ctDNA available at
88 QPOR analyzable pts, 85 had sTIL scores available from prior visual pathologic review. T0. 4/14 were ctDNA+ at T0, 2 of whom became ctDNA negative, and 2 remained ctDNA+. 10/14 pts
Tumors with low sTIL ( , 30%, a previously documented clinically significant cut-off),were were ctDNA negative at T0, 8 of whom remained negative, 2 became positive at T1 then cleared. All AEs
further stratified into low vs high IHI using median IHI for the population as the cut-off. We were grade(G) 1 except 1 pt with G2 hot flashes. Most common AEs include hot flashes (85%),
then determined the relationship between IHI and likelihood of response to NAC (RCB 0,1) in constipation (50%), fatigue (50%), and nausea (35%). One pt had G3 hypersensitivity and hypertension,
the overall cohort, and in the TNBC and ER low ( , 10%) sub cohorts. The analysis was also both unrelated to therapy. Conclusions: Neoadjuvant laso demonstrates a favorable AE profile and
performed using the median sTIL cut-off , 20% previously used in INFORM. Results: Low promising anti-tumor activity in suppressing 3-wk Ki67 and MRI FTV change in pts with HR+ HER2-
IHI was significantly predictive of NAC response (RCB 0,1) in low sTIL pts in the overall negative early BC. Ki67 suppression in premenopausal pts was seen in the absence of OFS. Clinical
cohort (N = 64, OR = 4.75; 95% CI1.50, 16.21), p = 0.005, PPV = 70%) and in the ER low sub trial information: NCT01042379. Research Sponsor: NIH PO1-CA210961.
cohort (N = 46, OR = 4.04;95% CI 1.04-17.38, p = 0.04, PPV = 72%). IHI was modestly Ki67 expression at pre-treatment, and 3-wk time point.
predictive in the low sTIL pts in TNBC sub cohort (N = 41, OR = 4.28 (0.99, 20.77), p = 0.03, All Patients Premenopausal Postmenopausal
PPV = 71%) but not predictive in high sTIL pts ( , 1/3 pts in each subgroup). For sTIL , 20%, (n=20) (n=10) (n=9)
IHI had stronger predictive ability in all pts (N = 64, OR = 7.63 (1.83, 40.09), p = 0.0016, PPV = Median Ki67 expression
80%). Conclusions: Heterogeneity of immune cell distribution determined by computational Baseline 10.0% 12.5% 10.0%
analysis of WSI of pre-treatment CNB of patients with germline BRCA mutations and HER2- 3-wk 5.1% 3.0% 6.0%
Number of pts with Ki67 expression <10% at 3-wk 87.5%1 87.5% 85.7%
negative cancers in the INFORM trial improves response prediction to NAC in patients with Number of pts with Ki67 expression <2.7% at 3-wk 37.5% 50% 28.6%
low baseline sTIL as determined by visual analysis. Research Sponsor: None. 1
Include the one male pt.

613 Poster Session TPS615 Poster Session

Verification of the BREASTEST assay in an Australian population: A novel A phase III trial evaluating addition of adjuvant chemotherapy to ovarian
liquid biopsy assay measuring lipid biomarkers for early-stage breast cancer function suppression + endocrine therapy in premenopausal women with
screening. First Author: David Speakman, BreastScreen Victoria, Melbourne, VIC, pN0-1, HR+/HER2- breast cancer (BC) and oncotype recurrence score
Australia (RS) £25 (OFSET): NRG-BR009. First Author: Eleftherios P. Mamounas, NSABP
Background: Women with dense breasts are at higher risk of developing breast cancer and AdventHealth Cancer Institute, Orlando, FL
however there is no agreement on how to screen these women, despite mandatory density Background: The TAILORx and RxPONDER trials demonstrated that RS identifies many
reporting becoming more prevalent. Currently available image-based, population level postmenopausal pts with node-neg and node-pos BC and RS #25, who do not benefit from
breast screening modalities have poor sensitivity in these women. Likewise, the positive addition of ACT to endocrine therapy (ET). Both trials also showed that certain subsets of
predictive value of imaging is reduced due to a higher false positive rate/recall rate premenopausal pts (node-neg/high clinical risk/RS 16-20, node-neg/RS 21-25, and node-pos/
compared to that of women with less dense breasts. Liquid biopsy approaches do not RS #25) benefited from adding ACT to ET. Most premenopausal pts in these trials did not
suffer from imaging-dependent challenges with density and provide a promising option in receive ovarian function suppression (OFS) as part of their ET regimen. Given the observed
this population. We have discovered a novel liquid biopsy assay using a lipidomic discovery benefit from OFS in high-risk premenopausal pts with HR+/HER2- BC in the SOFT/TEXT trials,
platform by combining liquid chromatography tandem mass spectrometry (LC-MS/MS) many questioned whether all or part of the observed ACT benefit in the TAILORx/RxPONDER
and machine learning. The BREASTEST assay is intended to complement standard of care trials may have been the result of chemotherapy-induced OFS. To address this question, we
screening and address gaps in the current screening paradigm. Methods: This study was developed OFSET, a phase III, multicenter clinical trial comparing OFS+ET v ACT+OFS+ET.
Methods: We hypothesize that addition of ACT to OFS+ET is superior to OFS+ET in improving
conducted to verify the performance of the BREASTEST assay in an Australian population
invasive breast cancer-free survival (IBCFS) among premenopausal, early-stage BC pts with
(n = 720). This verification study was an observational case-control study that pro-
HR+/HER2- tumors, and a 21-gene RS between 16-25 (for pN0 pts) and 0-25 (for pN1 pts).
spectively recruited women with breast cancer (n = 275) or without (n = 446) across 10 Secondary objectives include invasive disease-free survival, overall survival, distant recurrence-
clinical sites over a 34-month period. A primary imaging modality was identified for each free interval, breast cancer-free interval, and health-related quality of life (HRQOL). Pts must be
subject and a binary classification was assigned to the outcome of this imaging (normal/ node-neg with RS 16-20 (plus high clinical risk), or RS 21-25, or have 1-3 positive nodes with
suspicious) to enable comparison to, and combination with, the BREASTEST assay. An RS #25. Stratification is by nodal status/RS status (pN0 RS 16-25 v pN1 RS 0-15 and pN1 RS
assay with a high sensitivity has utility as a rule-out test and would complement 16-25), intent to receive CDK4/6 inhibitor (yes; no), and age (18-39 v $40). Pts are randomized
population-based imaging (high specificity). Therefore, the assay was designed to after surgery to either OFS+ET or ACT+OFS+ETv ET is an aromatase inhibitor (AI). Choice is per
achieve a sensitivity of 0.90. The combined specificity of imaging with the assay was investigator discretion; tamoxifen is allowed if AI is not tolerated or if OFS is incomplete.
calculated to estimate the clinical benefit BREASTEST could bring in ruling-out women Radiotherapy will be administered per investigator discretion per protocol guidelines. The
without breast cancer. A safe de-escalation rate was also calculated in this study to assess HRQOL sub-study will assess differences in severe menopausal symptoms, measured by the
the potential reduction in unnecessary further assessment if this assay was added to FACT ESS-19 score between arms, as well as increased pain severity (PROMIS). Blood and
standard imaging. Results: The utility of the BREASTEST assay was observed when tumor specimens will be collected for future research. Accrual of 3,960 pts is anticipated to be
results were combined with primary imaging data in the study cohort. Across all imaging completed in 7 yrs, 7 mos. Per NSABP B-28 and RxPONDER data, 5yr IBCFS of pN1 pts on the
modalities and breast densities, the assay improved the combined specificity in 45–75- ACT+OFS+ET arm is estimated at 92.3%. Based on TAILORx data, 5yr IBCFS of pN0 pts on the
year-old women by +6.1% (0.712, 0.773) and had a safe de-escalation rate of 21.0%. ACT arm is ~95%. Assuming 56% of pts to be pN0 and 44% pN1, and a 0.5% annual loss-to-
Highlighting the potential benefit to women with dense breasts, in women aged 30-49 follow-up rate, the definitive analyses to detect a hazard ratio: 0.75 with ACT+OFS+ET v
years with breast density category D the combined specificity was +14.7% (0.585, 0.732) OFS+ET, with one-sided a of 0.025 and 80% power, will require 380 IBCFS events, expected to
and safe de-escalation rate of 37.5%. The BREASTEST assay alone obtained a sensitivity of occur~11 yrs after study initiation. OFSETwas activated Aug 2023. As of 1-6-25, accrual is:
0.90, specificity of 0.369 with an AUC of 0.743. Conclusions: The performance and utility 188/3,960. NCT #: NCT05879926. Support: U10CA180868, -80822, UG1CA189867,
of the BREASTEST assay was verified in this study in an Australian population. It has U24CA196067. Clinical trial information: NCT05879926. Research Sponsor: National Cancer
highlighted the potential of this assay in the workup of women with breast conditions, in Institute; U10CA180868; National Cancer Institute; U10CA180822; National Cancer Institute;
particular women with dense breasts. Research Sponsor: BCAL Diagnostics. UG1CA189867; National Cancer Institute; U24CA196067.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 33s

TPS616 Poster Session TPS617 Poster Session

A phase 3, randomized study of adjuvant sacituzumab tirumotecan plus Adjuvant WIDER: A phase 3b trial of ribociclib (RIB) + endocrine therapy (ET)
pembrolizumab vs treatment of physician’s choice in participants with as adjuvant treatment (tx) in a close-to-clinical-practice patient (pt) pop-
triple-negative breast cancer who received neoadjuvant therapy and did ulation with HR+/HER22 early breast cancer (EBC). First Author: Stephanie L.
not achieve a pathologic complete response at surgery. First Author: Heather L. Graff, Brown University Health Cancer Institute, The Warren Alpert Medical School of
McArthur, UT Southwestern Medical Center, Dallas, TX Brown University, Providence, RI
Background: Trophoblast cell surface antigen 2 (TROP2) expression is higher in triple- Background: The phase 3 NATALEE trial met its primary endpoint with significant invasive
negative breast cancer (TNBC) vs other breast cancer subtypes, and high expression is disease–free survival benefit with RIB + ET vs ET alone in a broad pt population with stage II/
associated with poor prognosis. Sacituzumab tirumotecan (sac-TMT; also known as MK- III HR+/HER22 EBC, sustained with additional follow-up at 44.2 months (hazard ratio,
2870/SKB264) is a novel antibody-drug conjugate composed of anti-TROP2 antibody 0.715). The Adjuvant WIDER trial will evaluate RIB + ET in an HR+/HER22 EBC pt population
coupled to a cytotoxic belotecan derivative via a novel linker (average drug/antibody that reflects pts seen in clinical practice as it has wider eligibility criteria, including an
ratio, 7.4). In a prior phase 3 study (OptiTROP-Breast01), sac-TMT alone improved PFS additional focus on enrolling minority pts underrepresented in NATALEE. Given the unmet
(HR, 0.31; 95% CI, 0.22-0.45; P , 0.00001) and OS (HR, 0.53; 95% CI, 0.36-0.78; need in pts with Stage II/III EBC, the results of this trial will complement existing data on
P = 0.0005) vs chemotherapy in participants with heavily pretreated advanced TNBC. benefits of RIB + ET. Methods: This phase 3b, multicenter, open-label, single-arm trial will
The current standard of care (SOC) for patients with newly diagnosed, high-risk, early- evaluate, with early involvement of key pt advocacy groups, the efficacy and safety of
stage TNBC is neoadjuvant pembrolizumab + chemotherapy followed by adjuvant adjuvant RIB + ET in a close-to-clinical-practice pt population with HR+/HER22 EBC.
pembrolizumab after surgery. Patients who do not achieve a pathologic complete Eligible women and men aged $18 years with an ECOG PS of 0 to 2 and anatomic stage II/III
response (pCR) with the current SOC have higher rates of recurrence and mortality vs disease (AJCC 8th ed), with additional criteria for stage IIA disease (N1 or N0 with grade 3, or
grade 2 with Ki-67 $20% or high genomic risk), will be included. Pts will receive RIB (400
patients who achieve pCR. This study (NCT06393374) evaluates adjuvant sac-TMT +
mg/d; 3 wk on/1 wk off) + ET (letrozole 2.5 mg/d, anastrozole 1 mg/d, or exemestane 25 mg/
pembrolizumab vs treatment of physician’s choice (TPC; pembrolizumab 6 capeci-
d) for 36 months, followed by ET alone as SOC per investigator’s clinical judgment. Pre/peri-
tabine) in participants with TNBC who received neoadjuvant therapy and did not achieve menopausal women and men will receive goserelin 3.6 mg or leuprolide 3.75 mg/4 wk.
pCR at surgery. Methods: This phase 3, multicenter, open-label study is enrolling Switching between ETs during study tx will be allowed in cases of unmanageable toxicity.
participants $18 years old with centrally confirmed TNBC per most recent American Pts may have received (neo)adjuvant ET if initiated #36 months prior to enrollment. The
Society of Clinical Oncology/College of American Pathologists guidelines. Participants number of pts with prior ET between 12 and 36 months will be capped at »30%; this cap will
have non-pCR after $5 cycles of neoadjuvant pembrolizumab + chemotherapy, not be applicable to Black or African American pts. For pts with prior ET .12 months,
including $1 cycle of anthracycline-based neoadjuvant therapy. Participants must restaging is recommended. Pts with prior CDK4/6i tx (except RIB) in the adjuvant setting
provide tissue from the surgical specimen for central TROP2 assessment and be able to for #6 months who discontinued due to toxicity can be included. Study tx may be
continue on adjuvant pembrolizumab. Randomization must be conducted #12 weeks held #28 days (or longer on agreement) to recover from RIB-related toxicity before
from surgical resection (window may be extended in consult with sponsor). Participants restarting. If indicated, pts must have completed radiotherapy or chemotherapy before
are randomized 1:1 to pembrolizumab 400 mg Q6W for 5 doses + sac-TMT 4 mg/kg Q2W screening. Key exclusion criteria are distant metastases and/or recurrence and clinically
for 12 doses or TPC with pembrolizumab 400 mg Q6W for 5 doses 6 capecitabine 1000- significant, uncontrolled heart disease at screening. The primary endpoint is investigator-
1250 mg/m2 BID on days 1-14 and days 22-35 every 42 days for 4 cycles until completion assessed invasive breast cancer–free survival rate at 3 years per STEEP v2.0 criteria.
of therapy or disease recurrence, unacceptable toxicity, or withdrawal. Randomization is Secondary endpoints include invasive disease–free survival, distant disease–free survival,
stratified by residual tumor and lymph node status, TROP2 expression, and intention to distant relapse–free survival, recurrence–free interval, overall survival, quality of life, and
use capecitabine. Primary endpoint is invasive disease-free survival. Secondary end- safety. Exploratory endpoints will assess subsequent antineoplastic tx, potential mecha-
points are OS, distant recurrence-free survival, patient-reported outcomes, and safety. nisms of RIB benefit/resistance to RIB + ET, and RIB efficacy/safety in Black pts. Estimated
Enrollment began Q2 2024. Clinical trial information: NCT06393374. Research Sponsor: enrollment is 1400 pts globally. Recruitment is ongoing. Clinical trial information:
Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. NCT05827081. Research Sponsor: Novartis Pharmaceuticals Corporation.

TPS618 Poster Session TPS619 Poster Session

Phase III study to evaluate the efficacy and safety of GLSI-100 (GP2 + GM- ELEGANT: Elacestrant versus standard endocrine therapy (ET) in women and
CSF) in breast cancer patients with residual disease or high-risk PCR after men with node-positive, estrogen receptor-positive (ER+), HER2-negative
both neo-adjuvant and postoperative adjuvant anti-HER2 therapy: Fla- (HER2-), early breast cancer (eBC) with high risk of recurrence in a global,
mingo-01. First Author: Snehal Patel, Greenwich LifeSciences, Stafford, TX multicenter, randomized, open-label phase 3 study. First Author: Aditya Bardia,
Background: GP2 is a biologic nine amino acid peptide of the HER2/neu protein de- UCLA Health Jonsson Comprehensive Cancer Center, Los Angeles, CA
livered in combination with Granulocyte-Macrophage Colony Stimulating Factor (GM- Background: Adjuvant ET is the standard of care (SOC) for treating ER+/HER2- eBC.
CSF) that stimulates an immune response targeting HER2/neu expressing cancers, the Despite advances to optimize adjuvant treatment in high-risk ER+/HER2- eBC, there
combination known as GLSI-100. Of the 146 patients that have been treated with GLSI- continues to be a risk of local and metastatic (incurable) recurrence that persists, and new
100 over 4 clinical trials, GLSI-100 was well-tolerated and no serious adverse events therapies with desirable safety profiles are warranted. Elacestrant is a next-generation oral
observed were considered related to the immunotherapy. Methods: This Phase III trial SERD that provides a novel mechanism of action that has shown both SERD (degradative)
is a prospective, randomized, double-blinded, multi-center study. After 1 year of and SERM (partial agonist) activity that differs from currently available adjuvant ET
trastuzumab-based therapy, 6 intradermal injections of GLSI-100 or placebo will be (Wardell, ERC 2015). In the EMERALD trial, elacestrant significantly prolonged PFS vs SOC
administered over the first 6 months and 5 subsequent boosters will be administered ET in the overall population (HR 0.70; 95% CI 0.55-0.88; P=0.0018) and in patients with
over the next 2.5 years. The participant duration of the trial will be 3 years treatment plus ESR1-mut tumors (HR 0.55; 95% CI 0.39-0.77; P=0.0005) (Bidard, JCO 2022). In patients
1 additional year follow-up. Study Size – Interim Analysis: Approximately 498 patients with ESR1-mut tumors who received prior ET+CDK4/6i $12 mo, mPFS with elacestrant
will be enrolled. To detect a hazard ratio of 0.3 in invasive breast cancer free survival was 8.6 vs 1.9 mo with SOC ET (Bardia, CCR 2024). In a preoperative, window of opportunity
(IBCFS), 28 events will be required. An interim analysis for superiority and futility will be ER+/HER2- eBC trial (SOLTI-1905-ELIPSE), elacestrant was associated with complete cell
conducted when at least 14 events have occurred. This sample size provides 80% power cycle arrest (defined as Ki67,2.7%) rate of 27% and a statistically significant mean
if the annual rate of events in placebo patients is 2.4% or greater. Up to 250 non-HLA- change from baseline, shifting tumor biology toward a more endocrine-sensitive and less
A*02 subjects will be enrolled in an open-label arm. Eligibility Criteria: The patient proliferative tumor phenotype (Vidal, CCR 2025). Given that elacestrant demonstrated
population is defined by these key eligibility criteria: 1) HER2/neu positive and HLA- efficacy in mBC regardless of ESR1-mut status relative to SOC ET and has shown biologic
A*02, 2) Residual disease or High risk pCR (Stage III at presentation) post neo-adjuvant activity in eBC, it is hypothesized that elacestrant can prolong invasive breast cancer-free
therapy, 3) Exclude Stage IV, and 4) Completed at least 90% of planned trastuzumab- survival (IBCFS) in patients with high-risk eBC who received prior adjuvant ET6CDK4/6i.
Methods: ELEGANT (NCT06492616) is a global, multicenter, open-label phase 3 study
based therapy. Trial Objectives: The trial objectives are to: 1) Determine if GP2 therapy
designed to evaluate elacestrant vs SOC ET (AI or tamoxifen) in patients with eBC and a
increases IBCFS, 2) Assess the safety profile of GP2, and 3) Monitor immunologic
high risk of recurrence. Patients will be randomized 1:1 to continue SOC ET or to ela-
responses to treatment and assess relationship to efficacy and safety. Study Status: The
cestrant for a duration of 5 yrs. Eligible patients are women or men with ER+/HER22 node-
study is actively recruiting and enrolling patients in the US and Europe at up to 150 sites.
positive eBC who have completed 24-60 mo of adjuvant ET6CDK4/6i and have ECOG
Contact Information: Greenwich LifeSciences, Inc., Stafford, TX; Email: Flamingo-01@ PS #1. Patients who received a prior CDK4/6i or a PARP inhibitor must have already
[Link]; Website: [Link] Funding: This trial is completed or discontinued these treatments. Pre/perimenopausal women and men will be
supported by Greenwich LifeSciences. Clinical trial information: NCT05232916. administered a LHRH agonist. Exclusion criteria include inflammatory breast cancer,
Research Sponsor: None. history of prior invasive breast cancer, and .6 mo continuous interruption of prior SOC
adjuvant ET or discontinuation of adjuvant ET .6 mo prior to randomization. The primary
endpoint is IBCFS. Key secondary endpoints include distant relapse-free survival, overall
survival, invasive disease-free survival, safety, patient-reported outcomes-quality of life,
and pharmacokinetics. Status: Planned enrollment is 4,220 patients; recruitment is on-
going. Clinical trial information: NCT06492616. Research Sponsor: Menarini Group.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
34s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

TPS620 Poster Session TPS621 Poster Session

EORTC-2129-BCG: Elacestrant for treating ER+/HER2- breast cancer pa- The SURVIVE study: Standard surveillance vs. intensified liquid biopsy-
tients with ctDNA relapse (TREAT ctDNA). First Author: Michail Ignatiadis, Breast based surveillance in early breast cancer survivors. First Author: Henning
Medical Oncology Clinic, Institut Jules Bordet, Université Libre de Bruxelles (ULB), Schäffler, Department of Gynecology and Obstetrics, University Hospital Ulm, Ulm,
Hôpital Universitaire de Bruxelles (HUB), Brussels, Belgium Germany
Background: (Neo)adjuvant systemic treatment, with chemotherapy and/or endocrine Background: Current breast cancer (BC) follow-up relies on clinical examinations and
therapy (ET), substantially reduces the recurrence rates of estrogen receptor positive breast imaging, as studies from the 1980s demonstrated no survival benefit from distant
(ER+) human epidermal growth factor receptor 2 negative (HER2-) early-stage breast metastasis screening. However, with advancements in treatment strategies and the di-
cancer (BC). However, recurrences still occur up to 20 years after diagnosis. Circulating agnostic potential of liquid biopsies, this approach warrants re-evaluation. To enable pre-
tumor DNA (ctDNA) has emerged as a useful biomarker for surveillance in several solid symptomatic detection of distant relapse, we propose assessing a liquid biopsy-guided
tumors. ctDNA-based detection of molecular recurrence could allow the start of ef- surveillance strategy incorporating tumor markers (CA 27.29, CA 125, CEA), circulating
fective therapies before the clinical evidence of metastases. Elacestrant, a selective ER tumor cells (CTCs), and circulating tumor DNA (ctDNA). Methods: The SURVIVE study
degrader, approved in the advanced setting of ER+/HER2- ESR1-mutated BC following (NCT05658172) is the first large-scale, multicenter, partially double-blinded randomized
progression on a CDK4/6-inhibitor, could be used at the time of ctDNA-based molecular controlled trial comparing intensified and standard surveillance in 3,500 survivors of
relapse to delay or prevent the clinical manifestation of distant metastasis. medium- to high-risk early breast cancer (eBC). All subtypes are eligible. High risk includes
Methods: TREAT ctDNA is an European Organisation for Research and Treatment of (neo-)adjuvant chemotherapy, tumor size .50 mm, positive lymph nodes ($pN1mi), or high
grade ($G3). Patients are randomized 1:1 to standard or liquid biopsy-guided intensified
Cancer (EORTC)-led intergroup international, multicentre, randomised, open label, su-
follow-up. Primary therapy (surgery, adjuvant chemo- or radiotherapy) completion is re-
periority phase III trial to evaluate adjuvant elacestrant vs standard ET in patients with
quired, while adjuvant endocrine, antibody, or targeted therapy is permitted. Enrollment is
ER+/HER2- BC. The study comprises a screening and a randomised phase based on
allowed up to 24 months post-primary therapy for TNBC/HER2+ eBC and 60 months for
ctDNA status using a clinically-validated, tumor-informed molecular residual disease HR+/HER2- eBC. In both arms, guideline-based follow-up is performed, with additional blood
ctDNA assay (Signatera). Screening phase: 1960 patients with intermediate to high-risk samples collected longitudinally (years 1–3 every 3 months; years 4–5 every 6 months). In
stage II or III ER+/HER2- BC on medium to long duration ET will be screened for a ctDNA- the intervention arm, these samples are analyzed for tumor markers, CTCs, and ctDNA
based molecular relapse every 6 months. Randomised phase: 220 ctDNA-positive (RaDaR assay). Abnormal findings indicating minimal residual disease (MRD) trigger full
patients without imaging evidence of recurrence will be randomised 1:1 between staging. Recurrence is treated per national guidelines. In the case of M0 status, liquid biopsy
continuing current ET versus switching to elacestrant for a duration of at least 7 years of testing and staging continue, with the option for inclusion in interventional trials, if ap-
ET in total. Participants will undergo intensive follow-up for 3 years with computed plicable. The study is recruiting, with the first patient enrolled in December 2022. By January
tomography and bone scans, in addition to the standard annual breast imaging. The 2025, 812 patients were randomized across 78 centers. Final enrollment is scheduled for
primary endpoint of the study is distant metastasis free survival and secondary end- 2026 but may occur earlier due to accelerated recruitment. Statistics: The two primary
points are invasive disease-free survival, relapse-free survival, overall survival, safety objectives are to evaluate the lead time effect obtained by liquid biopsy marker testing in the
and quality of life. Recruitment started in December 2023 in Belgium, is open in 12 intensified follow-up arm and to test whether intensified, liquid biopsy-guided surveillance
countries at 74 sites and anticipates up to 120 enrolling sites in 2025. Overall study improves overall survival (OS) compared to standard follow-up. OS will be analyzed in the ITT
status and databases status will be periodically reviewed by the IDMC. Clinical trial population using Kaplan-Meier and Cox regression, while the lead-time effect is assessed
identification: EU trial number 2022-501453-36-00. NCT05512364. Study conducted descriptively. Secondary endpoints include IDFS, DDFS, DRFS, BCSS, and QoL as well as
under the Breast International Group (BIG) umbrella. Collaborative groups: GIM, CTI, biomarker sensitivities and specificities obtained in the intensified follow-up arm. Aims: We
SUCCESS, SOLTI, HeCOG, HORG, BOOG, SweBCG and ETOP-IBCSG. Clinical trial in- aim to determine whether liquid biopsy-guided follow-up enables earlier, sensitive, and
formation: 2022-501453-36-00. Research Sponsor: BERLIN-CHEMIEAG MENARINI from specific detection of distant (oligo-)metastases, facilitating timely intervention and im-
Germany. proving OS. Clinical trial information: NCT05658172. Research Sponsor: German Federal
Ministry of Education and Research.

TPS622 Poster Session TPS623 Poster Session

The SURVIVE HERoes study: Targeting molecular relapse in breast cancer— Short-term pre-operative durvalumab (MEDI 4736) in early small triple-
A secondary adjuvant intervention study of trastuzumab deruxtecan versus negative breast cancer patients (POP-Durva). First Author: Joana M. Ribeiro,
SOC treatment in patients with HER2-positive or HER2-low early breast Gustave Roussy, Department of Medical Oncology, IHU-National PRecISion Medicine
cancer and ctDNA positivity after primary therapy. First Author: Kerstin Pfister, Center in Oncology, France, Villejuif, France
Department of Gynecology and Obstetrics, University Hospital Ulm, Ulm, Germany Background: Pathological response to neoadjuvant immune checkpoint inhibitors (ICI)
Background: Current research on circulating tumor DNA (ctDNA) in the adjuvant setting is associated with excellent survival in several tumor types. In Stage II-III triple negative
of early breast cancer (eBC) underscores its strong prognostic significance. Patients breast cancer (TNBC), neoadjuvant anti-PD-(L)1 with chemotherapy improves patho-
who are ctDNA-positive but show no radiological signs of relapse (i.e., molecular re- logical complete response (pCR) and reduces recurrence. In Stage I TNBC, (neo)adjuvant
lapse) exhibit reduced disease-free and overall survival. Secondary adjuvant intervention chemotherapy remains standard of care. Exceptional responses to ICI in TNBC have been
studies represent an innovative and promising therapeutic approach. Methods: We observed, suggesting a subgroup of Stage I TNBC could be treated with ICI alone;
introduce SURVIVE HERoes (NCT06643585), a phase III randomized clinical trial however, biomarkers to select patients are lacking. Methods: Trial Design: POP-Durva
comparing the potent antibody-drug conjugate trastuzumab deruxtecan to standard of (NCT05215106) is a prospective, single-arm phase II trial evaluating pCR after two doses
care (SoC) in patients with HER2-positive or HER2-low eBC and molecular residual or of durvalumab in Stage I TNBC. Patients with untreated clinical stage I (#2cm, N0) TNBC
recurrent disease (ctDNA-positive, cM0) following primary therapy. Patients tested (ER , 10%, PR , 10%, HER-2 non-amplified) with sTIL of $ 5% will be included. Study
positive for ctDNA in a tumor-informed approach are eligible, if staging examinations do treatment consists of two doses of durvalumab 10mg/kg IV, on D1 and D15. On
not show any residual or recurrent cancer. Participants are randomized in a 2:1 ratio to completion of study treatment, patients will undergo breast US and will proceed to
receive trastuzumab deruxtecan (+ endocrine therapy for HR+ patients) or SoC therapy. surgery, or standard neoadjuvant treatment, per physician preference. Fresh tissue
Stratification factors include hormonal receptor status (positive versus negative) and biopsy and Formalin-Fixed Paraffin-Embedded (FFPE) will be collected at screening, on
HER2 status (positive versus low). The primary endpoint is ctDNA clearance rate after D22 or at surgery; blood will be collected for PBMC and ctDNA at screening, D1, D15 and
12 months of therapeutic intervention. Secondary endpoints include relapse-free on D22; faecal specimen collection will occur at baseline and at end of treatment (for
survival, overall survival, safety, and quality of life (QoL). The trial will enroll a total microbiota analysis). Trial Endpoints: The primary endpoint is pCR (ypT0/is ypN0). In
of 180 participants across 50 centers in Germany. Staging examinations and ctDNA patients who proceed directly to surgery following durvalumab, pCR will be assessed at
assessments will be performed at regular intervals. The study is accompanied by a surgery. Patients with residual invasive disease at the D22 biopsy who receive further
comprehensive translational research program. Recruitment began in January 2025 and neoadjuvant therapy will be considered non-pCR for the primary endpoint. With an
is anticipated to continue until 2030. Discussion: Treating ctDNA-positive patients expected pCR rate of 20%, a sample size of 195 patients provides a 95% confidence
without radiographic evidence of recurrence is a novel therapeutic strategy. If SURVIVE interval of a precision of 6.2%. The secondary objectives are ORR and safety. The key
HERoes and similar studies targeting minimal residual disease (MRD) yield positive exploratory objective is to identify biomarkers of response to ICI. Spectral cytometry,
results, they could pave the way for a new molecularly driven individualized treatment single-cell RNA and TCR sequencing will be performed to describe on-treatment immune
approach aimed at achieving cure by liquid biopsy triggered early intervention in this new cell dynamics and to identify mechanisms of response to ICI monotherapy. Imaging-
therapeutic window of pre-symptomatic MRD. Clinical trial information: NCT06643585. mass cytometry will characterise tumour-immune cell spatial interactions. Microbiome
Research Sponsor: Astra Zeneca. profiles will be correlated with response. 4 sites in France are actively recruiting; as of
27/01/2025, 35 patients have been treated. Clinical trial information: NCT05215106.
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—LOCAL/REGIONAL/ADJUVANT 35s

TPS624 Poster Session TPS625 Poster Session

RECAST (Re-Evaluating Conditions for Active Surveillance Suitability as NRG-BR007: A phase III trial evaluating de-escalation of breast radiation
Treatment) for DCIS: Clinical trial in progress. First Author: Ruolin Lorraine Jiang, (DEBRA) following breast-conserving surgery of stage 1, HR+, HER2-,
University of California, San Francisco, San Francisco, CA RS £18 breast cancer. First Author: Julia R. White, University of Kansas Medical
Background: Ductal Carcinoma In Situ (DCIS) is a condition where cancerous cells are Center Comprehensive Cancer Center, Kansas City, KS
confined to the breast ducts. The standard of care is surgery, either breast conservation Background: Approximately 50% of newly diagnosed invasive breast cancers are stage 1,
and radiation or mastectomy. Data from the COMET study in hormone receptor-positive with the majority being ER/PR-positive, HER2-negative. Genomic assays such as the
(HR+) DCIS demonstrates that active surveillance (AS) is a safe alternative for initial Oncotype DX have identified patients (pts) with reduced risk of distant metastasis and
treatment. Starting with endocrine risk-reducing therapy first may assist in identifying without benefit from chemotherapy added to endocrine therapy (ET), freeing them from
candidates for risk reduction vs. focal surgical removal. The RECAST (Re-Evaluating excess toxicity. Genomic assays are also recognized as prognostic for in-breast recurrence
Conditions for Active Surveillance Suitability as Treatment) study re-orders the treat- (IBR) after breast-conserving surgery (BCS) and could similarly allow de-escalation of
ment, starting with endocrine risk reduction, and uses serial imaging to assess treatment adjuvant radiotherapy (RT). Reducing overtreatment is of interest to pts, providers, and
response to predict who can safely proceed with AS and endocrine therapy. Imaging payers. Methods: We hypothesize that BCS alone is non-inferior to BCS plus RT for IBR
response markers are tested to predict the success of endocrine therapy. Several novel and breast preservation in women intending ET for stage 1 invasive breast cancer (ER and/
endocrine treatments are tested. The trial gives patients a window of opportunity to or PR-positive, HER2-negative with an Oncotype DX Recurrence Score [RS] of #18).
evaluate the impact of endocrine therapy based on their imaging characteristics to Stratification is by age (,60; $60), tumor size (#1 cm; .1-2cm), and RS (#11, .11-18/
explore alternatives to surgery. Rather than being randomly assigned to surgery or AS, all MammaPrint Low). Pts are randomized post-BCS to Arm 1 with breast RT using standard
patients start with AS and serve as their own control. Methods: Women are screened for methods (moderate or ultra hypo- or conventional-fractionated whole breast RT with/
and randomized to 1 of 4 endocrine treatments, one of which is the standard of care without boost, or APBI) with $5 yrs of ET (tamoxifen or AI) or Arm 2 with $5 yrs of ET
endocrine therapy (choice of tamoxifen, baby tamoxifen, or an aromatase inhibitor is left (tamoxifen or AI) alone. The specific regimen of ET in both arms is at the treating
to patient and physician discretion); MRIs are conducted at baseline, 3 and 6 months and physician’s discretion. Eligible pts are stage 1: pT1 (#2 cm), pN0, age $50 to ,70 yrs, s/p
semiquantitative imaging determines suitability for AS. Patients on AS are eligible to BCS with negative margins (no ink on tumor), s/p axillary nodal staging (SNB or ALND), ER
continue treatment for 3 years. Follow-up consists of an MRI alternating with a and/or PR-positive (ASCO/CAP), HER2-negative (ASCO/CAP), and Oncotype DX RS #18
(diagnostic core biopsy or resected specimen). A “low risk” MammaPrint is permissible if
mammogram every 6 months. Quality of life (QOL) is measured using PROMIS and the
completed as part of usual care prior to screening. Primary endpoint is IBR (invasive breast
FACT-ES composite scores. Eligibility: All patients with a diagnosis of HR+ DCIS (any
cancer or DCIS). Secondary endpoints are breast conservation rate, invasive in-breast
grade), defined as . 50% ER+ or PR+ on immunohistochemistry Exclusion: Presence of
recurrence, relapse-free interval, distant disease-free survival, overall survival, patient-
invasive disease, pregnancy or active breastfeeding, history of deep vein thrombosis.
reported breast pain, patient-reported worry about recurrence, and adherence to ET. We
Patients with a solid mass on MRI must have a repeated biopsy Primary objectives: To assume a clinically acceptable difference in IBR of 4% at 10 yrs to judge omission of RT as
determine whether novel endocrine therapy increases the fraction of patients who are non-inferior (10-yr event-free survival for RT group is 95.6% v 91.6% for the omission-of-RT
suitable for long-term AS and how medications are tolerated compared to standard group). BR007 is powered to detect non-inferiority with 80% power and a one-sided
endocrine treatment Primary endpoints: QOL and fraction of patients who meet criteria a=0.025, assuming that there would be a ramp-up in accrual in the first two years (leveling
for remaining on AS after 6 months based on MRI Secondary endpoints: Biomarkers of off in Yrs 3-5); 1,670 pts (835 per arm) are required for randomization. Conservative loss to
response and resistance Progress to date: RECAST activated on 01/22/2024. Cur- follow-up is 1%/yr. Some T1a pts screened may have Oncotype DX scores .18, making
rently,12 sites in the US are open to enrollment. 28 are in the process of activation. There them ineligible for the study. In the accrual process, 1,714 pts will be required to register to
are 22 patients accrued with 6 in screening. RECAST is an important next step in ensure that our final randomized cohort is 1,670 pts. As of 1-6-2025: 1,168/1,670 pts have
elucidating the factors that predict the success of AS and provide a framework for been randomly assigned, and 1,294 screened. Support: U10 CA180868, -180822, UG1
understanding endocrine resistance in the HR+ DCIS population. Clinical trial infor- CA189867. Clinical trial information: NCT04852887. Research Sponsor: National Cancer
mation: NCT06075953. Research Sponsor: Quantum Leap Healthcare Collaborative. Institute; U10 CA180868; National Cancer Institute; UG1CA189867; National Cancer In-
stitute; UG1 CA189867.

TPS627 Poster Session TPS628 Poster Session

MELODY: A prospective non-interventional multicenter cohort study to Radiation omission in patients with clinically node-negative breast cancer
evaluate different imaging-guided methods for localization of malignant undergoing lumpectomy (ROSALIE). First Author: Elena Parvez, McMaster Uni-
breast lesions (EUBREAST-4/iBRA-NET/AGO-B-062, NCT 05559411). First versity, Hamilton, ON, Canada
Author: Maggie Banys-Paluchowski, Department of Obstetrics and Gynecology, Uni- Background: Currently, the standard of care for patients undergoing neoadjuvant
versity Hospital of Schleswig Holstein, Campus Lübeck, Lübeck, Germany chemotherapy (NAC) and breast conserving surgery (BCS) is adjuvant radiation (RT).
Background: In the last decades, the proportion of breast cancer patients receiving breast- However, high rates of pathologic complete response (pCR) after NAC have raised
conserving surgery has increased, reaching 70-80% in developed countries. In case of non- questions regarding the necessity of WBRT in these cases. A meta-analysis of 9 German
palpable lesions, surgical excision requires some form of breast localization. While wire- NAC trials demonstrated a 5-year locoregional recurrence (LRR) of only 4% in patients
guided localization has long been considered gold standard, it carries several limitations, with pCR who underwent BCS with radiation therapy. Data from two large National
including logistical difficulties, the potential for displacement and patient discomfort, and Surgical Adjuvant Breast and Bowel Project (NSABP) neoadjuvant trials (B.18 and B.27)
re-excision rates reaching 21% (in DCIS up to 30%). Other techniques (radioactive seed or demonstrated a local recurrence risk of 5.1% at 10 years (2.5% at 5 years) in patients .50
radio-occult lesion localization, intraoperative ultrasound, magnetic, radiofrequency, and years with node negative breast cancer who had a pCR and were treated with BCS and RT.
radar localization) have been developed with the aim of overcoming these disadvantages. With such low rates of recurrence, we postulated that the absolute benefit that RT can
However, comparative data on the rates of successful lesion removal, negative margins, and offer is limited. Radiation therapy is not without side effects, which include both short-
re-operations are limited. In most studies, the patient perspective, addressing e.g. dis- term and long-term toxicity. As such, a trial of de-escalation of RT is warranted.
comfort and pain, has not been evaluated. The aim of MELODY (MEthods for LOcalization of Methods: This study is a prospective, multi-center, single arm cohort study of omission
Different types of breast lesions) is to evaluate different imaging-guided localization of WBRT following BCS in patients with a pCR following NAC. Eligible and consenting
methods with regard to oncological safety, patient-reported outcomes, surgeon and ra- female patients with newly diagnosed T1-3 node negative breast cancer age .50 years
diologist satisfaction and economic impact. Methods: The EUBREAST and the iBRA-NET with no clinical evidence of distant metastatic disease, who have been treated with NAC,
have initiated the MELODY study to assess breast localization techniques and devices from BCS and axillary staging surgery with final pathology demonstrating a pCR (ypT0N0) will
several perspectives (NCT05559411, [Link] MELODY is a prospective be enrolled to the study and followed. Negative lymph node involvement at initial
intergroup cohort study which enrolls female and male patients. planned for breast- presentation must be documented by imaging (US or MRI), fine needle aspiration (FNA) or
conserving surgery with imaging-guided localization for invasive breast cancer or DCIS. core needle biopsy. Marker clip must have been placed in the tumour bed prior to or during
Multiple or bilateral lesions and neoadjuvant chemotherapy are allowed. Primary outcomes neoadjuvant chemotherapy when the tumour can still be identified. Study participants will
are: 1) Intended target lesion and/or marker removal, independent of margin status on final
not receive adjuvant RT, the current standard of care. Study participants will be followed
histopathology, and 2) Negative resection margin rates at first surgery. Secondary outcomes
and assessed for local recurrence, regional recurrence, distant recurrence, DFS and OS.
are, among others: rates of second surgery and secondary mastectomy, Resection Ratio
Any additional breast cancer treatments received by the participant for the first re-
(defined as actual resection volume divided by the calculated optimum specimen volume),
currence event including repeat BCS, mastectomy, additional systemic therapy and
duration of surgery, marker dislocation rates, rates of marker placement or localization
failure, patient-reported outcomes, rates of “lost markers”, radiologist and surgeon sat-
radiation therapy (RT) will be documented. The primary outcome is ipsilateral breast
isfaction, and health economic evaluation of the different techniques. Target accrual is tumour recurrence (IBTR) at median 5-year follow-up. A local recurrence of 5% without RT
7,416 patients. Enrollment started in January 2023. Until 24 January 2025, 3938 patients was felt to be acceptable. Based on a postulated 5-year IBTR risk of 3.0%, 4 years of
from 20 countries were enrolled in the study. The study is expected to complete patient accrual plus an additional 3 years of follow-up, a 90% two-sided CI for a postulated LR rate
enrollment in year 2026. The study will be conducted in 30 countries and is supported by the of 3.0% at 5 years would have an upper bound of ,5% with 300 patients. To account for a
Oncoplastic Breast Consortium (OPBC), AWOgyn, AGO-B, SENATURK, the American Society 5% potential loss to follow-up and 10% receiving RT contrary to protocol, a sample size of
of Breast Surgeons (ASBS) and the Korean Breast Cancer Study Group (KBCSG). Clinical trial 352 patients will be required. The trial opened in March 2024. Clinical trial information:
information: NCT05559411. Research Sponsor: None. NCT05866458. Research Sponsor: Canadian Institutes of Health Research (CIHR).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
36s BREAST CANCER—LOCAL/REGIONAL/ADJUVANT

TPS629 Poster Session TPS630 Poster Session

HERTHENA-Breast03: A phase 2, randomized, open-label study evaluating Eliminating breast surgery for triple negative or HR-/HER2+ breast cancer
neoadjuvant patritumab deruxtecan + pembrolizumab before or after pem- patients with clinical complete response to combined neoadjuvant chemo-
brolizumab + chemotherapy for early-stage TNBC or HR-low+/HER22 therapy and neoadjuvant radiotherapy: A multicenter, phase 2 trial (EBCS).
breast cancer. First Author: Joyce O’Shaughnessy, Baylor University Medical Cen- First Author: Zhengjun Yang, Tianjin Medical University Cancer Institute and Hospital,
ter, Texas Oncology, Dallas, TX and Sarah Cannon Research Institute, Dallas, TX Tianjin, China
Background: The standard of care for patients with high-risk, early-stage TNBC is neoadjuvant Background: Recent advancements in immunotherapy and targeted therapies have sig-
pembrolizumab (pembro) + chemotherapy followed by adjuvant pembro. Patients with HR-low+/ nificantly improved pathological complete response (pCR) rates in patients with triple-
HER22 breast cancer may also be treated per recommendations for TNBC. There is a need for negative breast cancer (TNBC) and HER2-positive breast cancer undergoing neoadjuvant
improved neoadjuvant therapy to increase the rate of pCR, as patients who do not achieve pCR have a chemotherapy. The combination of neoadjuvant chemotherapy with radiotherapy may
high risk of recurrence, and to reduce risk of long-term toxicities associated with cyclophosphamide
further enhance pCR rates through synergistic effects, prompting a reevaluation of tra-
and anthracyclines. HER3 is frequently expressed in breast cancer and implicated in drug resistance.
Patritumab deruxtecan (HER3-DXd) is an antibody-drug conjugate comprising a fully human anti-
ditional surgical approaches. The Sound trial demonstrated that omitting sentinel lymph
HER3 IgG1 monoclonal antibody linked to a topoisomerase I inhibitor (DXd) via a stable tetrapeptide- node biopsy in node-negative patients is safe and feasible, supporting further de-escalation
based linker that is selectively cleaved within tumor cells. This phase 2 study (NCT06797635) will of surgical interventions. For patients achieving pCR, the necessity of breast and axillary
evaluate neoadjuvant HER3-DXd + pembro before or after carboplatin + paclitaxel + pembro for early- surgery is increasingly questioned, given the potential to reduce surgical morbidity without
stage TNBC or HR-low+/HER22 breast cancer. Methods: Eligible participants (pts) are adults ($18 y) compromising outcomes. Our study investigates whether omitting surgery in patients with
with untreated, locally advanced nonmetastatic (AJCC stage cT1c, N1–N2 or cT2–cT4, N0–N2) TNBC pCR confirmed by vacuum-assisted core biopsy (VACB) yields non-inferior 5-year event-free
or HR-low+/HER22 breast cancer. Pts (N $10 and #30) in part 1 of the study (safety lead-in) will survival (EFS) compared to standard surgery. Methods: This multicenter, phase 2 trial
receive neoadjuvant HER3-DXd + pembro followed by carboplatin + paclitaxel + pembro (Table) then enrolls patients aged $18 years with untreated cT1-2 N0 M0 TNBC or HER2-positive breast
surgery. DLT evaluation and dose finding for HER3-DXd (three dose levels of 5.6 mg/kg Q3W, 4.8 mg/ cancer and ECOG 0-1. Patients receive four cycles of TCb (HP)* neoadjuvant chemotherapy,
kg Q3W and 3.2 mg/kg Q3W) during cycle 1 of neoadjuvant HER3-DXd + pembro will be performed in followed by neoadjuvantradiotherapy starting from the fifth cycle (50 Gy in 25 fractions +
part 1 to determine an acceptable dose of HER3-DXd for part 2. Pts (N ~342) in part 2 will be randomly
14 Gy boost in 7 fractions). The TCb (HP)* regimen is tailored based on tumor subtype: triple-
assigned [Link] to arm A, B or C (Table) for neoadjuvant treatment. Randomization will be stratified by
cancer type (TNBC vs HR-low+/HER22) and, in the TNBC subgroup, PD-L1 status (combined positive negative patients receive TCb (nab-paclitaxel + carboplatin) with or without immunotherapy
score $10 vs ,10), overall stage (II vs III) and HER3 expression (low vs high). After neoadjuvant (pembrolizumab), while HER2-positive patients receive TCbHP (nab-paclitaxel + carboplatin
treatment, pts will undergo surgery (with postoperative radiotherapy if clinically indicated) and receive + trastuzumab and pertuzumab) regimens. After six cycles, patients undergo MRI. If MRI
adjuvant pembro 400 mg Q6W for 5 cycles. Pts with residual disease may receive additional adjuvant suggests complete clinical response (cCR), VACB of the primary lesion is performed under
treatment of physician’s choice. Primary endpoints are safety (part 1 and 2) and pCR (ypT0/Tis ypN0) ultrasound/stereotactic guidance (6 cores, 7-10 G needle). If no residual tumor or atypical
(part 2). Enrollment is ongoing. Clinical trial information: NCT06797635. Research Sponsor: Merck cells are found, breast and axillary surgery are omitted. Patients receive indicated
Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA (MSD). This study is part of a immunotherapy/targeted therapy and are followed every 6 months for 5 years. The primary
collaboration between MSD and Daiichi Sankyo, Inc. endpoint is 5-year EFS. Secondary endpoints include breast pCR rate (bpCR: ypT0), overall
Neoadjuvant cycles 1-4 Neoadjuvant cycles 5-8 survival (OS), patient-reported outcomes (PROs), and safety. TheThis trial is designed to
determine whether the 5-year EFS of patients who avoid breast surgery after pCR confirmed
Part 1 HER3-DXd 5.6 or 4.8 or 3.2 mg/kg Q3W + Carboplatinb + paclitaxelc + pembroa
Arm A pembroa
by VACB is non-inferior to that of patients who undergo standard breast surgery with
Part 2 HER3-DXd (selected dose from Part 1) + Carboplatinb+ paclitaxelc + pembroa confirmed pCR. Based on a 90.3% 5-year EFS in pCR patients (cT1-2N0 TNBC/HER2+), the
Arm A
b
pembroa
c a
trial uses a one-sided test (non-inferiority margin: 5%; power: 80%; a: 0.1) to determine if
Part 2 Arm Carboplatin + paclitaxel + pembro HER3-DXd (selected dose from Part 1) + pembroa omitting surgery is non-inferior. 185 patients are needed to omit surgery. Assuming 80%
B
Part 2 Arm Carboplatinb + paclitaxelc + pembroa Doxorubicind OR epirubicine + cyclophosphamidef + pCR and 10% dropout, 256 participants will be enrolled. The trail is actively recruiting.
C pembroa Clinical trial information: NCT06498154. Research Sponsor: None.
a
200 mg Q3W; bAUC 1.5 mg/mL/min QW; c80 mg/m2 QW; d60 mg/m2 Q3W; e90 mg/m2 Q3W;
f
600 mg/m2 Q3W.

TPS631 Poster Session TPS632 Poster Session

A randomized trial of trastuzumab deruxtecan and biology-driven selection OPERETTA: A phase II study evaluating neoadjuvant and adjuvant olaparib
of neoadjuvant treatment for HER2-positive breast cancer (ARIADNE). First plus pembrolizumab following platinum-based chemotherapy plus pembro-
Author: Alexios Matikas, Karolinska Institutet and Karolinska University Hospital, lizumab for germline BRCA mutated triple negative breast cancer. First
Stockholm, Sweden Author: Yuko Takahashi, Department of Breast and Endocrine Surgery, Okayama Uni-
Background: Neoadjuvant therapy is the standard of care for the treatment of non- versity Hospital, Okayama, Japan
metastatic HER2-positive breast cancer. Studies on first generation antibody-drug Background: Triple negative breast cancer (TNBC) remains the most challenging phenotype
conjugates (ADC) such as trastuzumab emtansine (T-DM1) showed equal or slightly of breast cancer. There is still an unmet clinical need for improving the fine-tuning of
lesser efficacy than chemotherapy combined with dual HER2 blockade. Trastuzumab indications for targeted treatments in this population. In TNBC, the frequency of germline
deruxtecan (T-DXd) is a next generation ADC approved for the treatment of metastatic BRCA (gBRCA) 1/2 mutations was reported to be up to 19.5%. This has led to promising
HER2-positive breast cancer, with greatly improved efficacy compared with T-DM1. clinical strategies based on poly adenosine diphosphate (ADP)-ribose polymerase inhibitors
Methods: ARIADNE is an academic, international, open label, randomized, comparative that inhibit single-stranded DNA damage repair and/or modified chemotherapy approaches
phase IIB trial, actively enrolling in Sweden (ten sites) and in Norway (seven sites), with targeting the DNA damage response, using platinum-based regimens. Based on the results of
sites in Belgium (three), Netherlands (one) and Italy (three) activating during Q2 2025. A the OlympiA and KEYNOTE522 study, the adjuvant treatment with olaparib for gBRCAm and
total of 370 patients with non-metastatic HER2-positive primary breast cancer and an neoadjuvant and adjuvant pembrolizumab for patients with a high risk of recurrence TNBC
indication for neoadjuvant therapy will be offered inclusion and randomized 1:1 to has been treatment options as the standard of care. We hypothesize that neoadjuvant and
adjuvant combination treatment with olaparib and pembrolizumab following combination
receive either i) a taxane, carboplatin, trastuzumab and pertuzumab for three cycles or ii)
treatment with platinum-based chemotherapy and pembrolizumab would synergistically
T-DXd for three cycles. Further treatment is based on the PAM50-defined intrinsic
increase the anti-tumor effect through the enhancement of immunogenicity and DNA
molecular subtype from a pretreatment biopsy: HER2-enriched (approximately 65%) damage in patients with gBRCA mutated breast cancer. Methods: OPERETTA is a multi-
patients continue with the same treatment for three more cycles. Estrogen receptor (ER) centered, prospective single-arm phase II feasibility study of patients treated with neo-
positive and luminal (approximately 25%) patients receive trastuzumab and pertuzumab adjuvant olaparib plus pembrolizumab following platinum-based chemotherapy plus pem-
for three cycles, combined with letrozole and ribociclib for two cycles. Finally, ER- brolizumab in gBRCA 1/2 mutated TNBC. The patients with stage IIA-IIIB TNBC known as
negative and luminal or basal-like (approximately 10%) patients either continue with the gBRCA 1/2 mutated will be registered. The primary objective is the pCR rate defined as the
same treatment for three additional cycles in case of radiologic complete response, or absence of residual invasive disease in the breast and axilla. The secondary objectives
they receive four cycles of dose-dense epirubicin and cyclophosphamide in case of lack include additional efficacy measures (i.e., Residual Cancer Burden [RCB] 0/1rate, 3 years
of complete response. The primary endpoint of ARIADNE is locally assessed rate of overall survival [3y-OS], 3 years event-free survivals [3y-EFS]), and safety. The estimated
pathologic complete response (pCR) in patients with molecularly HER2-enriched tumors, sample size using Simon’s two-stage design, with a null hypothesis of a 45% pCR rate and an
defined as ypT0/Tis, ypN0, as determined by a pathologist blinded to treatment as- alternative hypothesis of 70%, was calculated. Given a significance level of 0.1 and 80%
signment (intention-to-treat analysis). Key secondary endpoints are rates of complete power, the design allows a maximum of 23 patients to be included. Eligible patients will be
radiologic response at three cycles; rates of pCR in the other two molecular groups and received combination treatment with paclitaxel (80 mg/m2 qw), carboplatin (AUC 1.5 qw or
in the two groups of the initial randomization; event-free survival, defined as the time AUC 5 q3w), and pembrolizumab (200mg q3w) for first 12 weeks followed by olaparib (300mg
from randomization to disease progression, locoregional or distant recurrence, con- BID) with pembrolizumab (200mg q3w) for another 12 weeks as neoadjuvant treatment.
tralateral breast cancer, other cancer, or death due to any cause. Tissue and plasma Breast/axillary surgery and radiotherapy are recommended per standard of care. After
samples are collected at baseline, during treatment and surgery, as well as during follow- surgery, the combination of olaparib plus pembrolizumab will be continued for another
up. The first patient was randomized on 26th October 2023; 46 patients had been enrolled 27 weeks as adjuvant treatment. This study is recruiting in Japan, and 2 patients are enrolled
to the study until January 2025. Clinical trial information: NCT05900206. Research as of January 2025. This study is part of the West Japan Oncology Group (WJOG) breast
Sponsor: None. cancer study group: WJOG14020B. Clinical trial information: NCT05485766. Research
Sponsor: Merck; AMED.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 37s

LBA1000 Oral Abstract Session 1001 Oral Abstract Session

Vepdegestrant, a PROTAC estrogen receptor (ER) degrader, vs fulvestrant in Patient-reported outcomes (PROs) in patients with ER+, HER2- advanced
ER-positive/human epidermal growth factor receptor 2 (HER2)–negative breast cancer (ABC) treated with imlunestrant, investigator’s choice stan-
advanced breast cancer: Results of the global, randomized, phase 3 VER- dard endocrine therapy, or imlunestrant + abemaciclib: Results from the
ITAC-2 study. First Author: Erika P. Hamilton, Breast Cancer Research Program, Sarah phase III EMBER-3 trial. First Author: Giuseppe Curigliano, Istituto Europeo di
Cannon Research Institute, Nashville, TN Oncologia, IRCCS, University of Milano, Milano, Italy
Background: Imlunestrant (imlu) is a next-generation, brain-penetrant, oral selective es-
trogen receptor degrader. The EMBER-3 trial, in patients (pts) with ER+, HER2- ABC who had
disease progression on or after aromatase inhibitor-based therapy, showed significant
progression free survival (PFS) improvement with imlu vs standard therapy (SOC, fulvestrant
or exemestane) in pts with ESR1 mutations (ESR1m), and with imlunestrant+abemaciclib
(imlu+abema) vs imlu in all pts, regardless of ESR1m. Exploratory PRO analyses are
presented here. Methods: EORTC QLQ-C30 was administered at baseline (BL) and every 8
weeks until treatment discontinuation. Prespecified QLQ-C30 analysis used a longitudinal
mixed model for repeated measures to calculate mean change from BL in pts with BL and
$1 post-BL score. PRO-CTCAE (diarrhea frequency) was administered weekly, reporting 0
(never) to 4 (almost constantly). PRO-CTCAE (injection site reaction [ISR]) was administered
to fulvestrant recipients weekly for 2 weeks post-injection, reporting yes/no (pain, swelling,
The full, final text of this abstract will be available at redness). Descriptive analysis was used for PRO-CTCAE. Results: In pts with ESR1m, imlu
[Link] on the day of presentation and in the monotherapy was associated with many improved or maintained EORTC QLQ-C30 scores,
online supplement to the June 10, 2025, issue of the Journal whereas scores with SOC were declined or maintained. Specifically, pts with ESR1m on imlu
had improved global health status (GHS)/quality of life (QOL) and physical function (PF)
of Clinical Oncology. scores, while scores with SOC declined (mean change differences between treatments: 9.9
[0.1, 19.7] and 6.2 [-0.8, 13.1], respectively). These PRO findings mirror the PFS findings in
this group. In the overall population, GHS/QOL scores declined similarly with imlu vs SOC
(mean change differences: 0.5 [-4.7, 5.7]), while PF scores were maintained with imlu vs a
slight decline with SOC (mean change difference: 2.5 [-1.1, 6.1]). Most fulvestrant recipients
(72%) reported ISR at any time while on treatment, with a mean of 31% during the first week
of the first 6 cycles. Imlu+abema vs imlu showed broadly similar declines in all pts, with
minimal mean change differences in GHS/QOL and PF scores (0.8 [-7.4, 5.9]; -2.2 [-6.6, 2.2],
respectively). Pts reported similarly low rates of “frequent” or “almost constant” diarrhea
with imlu (3%) and SOC (2%) and higher rates with imlu+abema (22%). Conclusions: PROs
from EMBER-3 demonstrated that patients with ESR1m had better GHS/QOL and PF with
imlu vs SOC, mirroring efficacy results. While the frequency of CTCAE defined ISRs was low,
the high rate of PRO-CTCAE ISR demonstrates that this clinically relevant adverse event is
underappreciated by physicians. Additionally, all pts had generally comparable GHS/QOL
and PF with imlu+abema vs imlu. Overall, these results support the efficacy and safety of
imlu compared to existing SOC. Clinical trial information: NCT04975308. Research Sponsor:
Eli Lilly and Company, Indianapolis, IN, USA.

1003 Oral Abstract Session 1004 Oral Abstract Session

INAVO120: Phase III trial final overall survival (OS) analysis of first-line Phase I/Ib study of inavolisib (INAVO) alone and in combination with
inavolisib (INAVO)/placebo (PBO) + palbociclib (PALBO) + fulvestrant endocrine therapy 6 palbociclib (PALBO) in patients (pts) with PIK3CA-
(FULV) in patients (pts) with PIK3CA-mutated, hormone receptor-positive mutated, hormone receptor–positive, HER2-negative locally advanced/
(HR+), HER2-negative (HER2–), endocrine-resistant advanced breast can- metastatic breast cancer (HR+, HER2– LA/mBC): Analysis of hyperglycemia
cer (aBC). First Author: Nicholas C. Turner, Royal Marsden Hospital and Institute of (HG) in prediabetic/obese pts. First Author: Mafalda Oliveira, Vall d’Hebron Uni-
Cancer, London, United Kingdom versity Hospital and Vall d’Hebron Institute of Oncology (VHIO), Barcelona, Spain
Background: INAVO, a highly potent and selective PI3Ka inhibitor that also promotes mutated Background: INAVO, a highly potent and selective PI3Ka inhibitor that also promotes
p110a degradation, is FDA-approved in combination with PALBO + FULV for PIK3CA-mutated, degradation of mutated p110a, is approved by the FDA in combination with PALBO + ful-
HR+, HER2–, endocrine-resistant aBC, based on the primary analysis of INAVO120 vestrant (FULV) for PIK3CA-mutated, HR+, HER2–, endocrine-resistant advanced BC. HG is a
(NCT04191499), which showed a statistically significant and clinically meaningful investigator- common on-target side effect of PI3K inhibitors. There are limited data for PI3K inhibitors in
assessed progression-free survival (INV-PFS) benefit in the INAVO arm vs. the PBO arm (hazard prediabetic/obese pts. Data from prediabetic/obese pts with HR+, HER2– LA/mBC treated with
ratio 0.43; 95% confidence interval [CI] = 0.32–0.59; p , 0.0001). At that analysis, interim OS INAVO from a Phase I/Ib study (GO39374; NCT03006172) are reported here. Methods: Adults
results were immature. Here we report the final OS analysis, including updated efficacy and $ 18 years of age received INAVO alone (Arm A), + letrozole (LET) + PALBO (Arm B), + LET (Arm
safety. Methods: Pts received INAVO (9 mg orally once daily [PO QD]; Days 1–28 of each 28-day C), + FULV (Arm D), + FULV + PALBO (Arm E), or + FULV + PALBO + primary prophylactic
cycle)/PBO + PALBO (125 mg PO QD; Days 1–21 of each cycle) + FULV (500 mg intramuscularly; metformin (Arm F). Data are reported across all arms unless indicated. Pts with baseline risk
Cycle 1 Days 1 and 15 then every ~4 weeks). OS and objective response rate (ORR) were formally factors for HG were defined by HbA1c $5.7%, fasting blood glucose $ 100 mg/dL, or body
tested; updated INV-PFS and safety analyses are descriptive. Results: Data cut-off was Nov 15, mass index $ 30 kg/m2. Adverse events (AEs) were reported using NCI-CTCAE v4, which
2024, at 34.2 months (mo) of median follow-up. Median OS was 34.0 mo (95% CI = 28.4–44.8) in utilizes fasting laboratory glucose values for HG severity grading, rather than clinical inter-
the INAVO arm and 27.0 mo (95% CI = 22.8–38.7) in the PBO arm (stratified hazard ratio 0.67; ventions used in v5. Results: Clinical cut-off was Jan 1, 2024. From190 pts treated, 110
95% CI = 0.48–0.94; p = 0.0190 [boundary = 0.0469]). The OS benefit was consistent across key (57.9%) were prediabetic/obese; their median time on INAVO was 222 days (range, 7 to 2,152)
subgroups. The survival probability at 6, 12, 18, 24, and 30 mo was 96.8%, 87.0%, 74.3%, 65.8%, and mean cumulative dose intensity was 91.8%. Most prediabetic/obese pts discontinued
and 56.5% in the INAVO arm and 90.1%, 76.7%, 67.2%, 56.3%, and 46.3% in the PBO arm. ORR INAVO due to progressive disease (82 [74.5%]); six (5.5%) discontinued INAVO due to an AE
was 62.7% (95% CI = 54.8–70.2) and 28.0% (95% CI = 21.3–35.6), respectively (p , 0.0001). (one due to HG). HG was reported in 80.9% of prediabetic/obese pts (grade 3–4: 34.5%). In pts
Median time to chemotherapy (TTC) was 35.6 mo (95% CI = 25.4–not reached) in the INAVO arm with two risk factors, 87.9% reported HG (grade 3–4: 39.4%). Among pts with HG, median time
and 12.6 mo (95% CI = 10.4–16.1) in the PBO arm (stratified hazard ratio 0.43; 95% CI = to onset was 14 days (range, 1 to 1,674) and 86.0% of events resolved by clinical cut-off.
0.30–0.60). Updated median INV-PFS was 17.2 mo (95% CI = 11.6–22.2) in the INAVO arm and Median time to improvement or resolution of first worst grade $ 2 event was 8 days (range, 1
7.3 mo (95% CI = 5.9–9.2) in the PBO arm (stratified hazard ratio 0.42; 95% CI = 0.32–0.55), with to 64). INAVO dose interruptions, reductions, and discontinuations due to HG were reported in
landmark analyses supporting durable benefit. 90.7% of pts in the INAVO arm and 84.7% in the 41.8%, 13.6%, and 0.9% of pts, respectively. The most common anti-HG medications were
PBO arm had grade 3/4 adverse events (AEs); there were no new grade 5 AEs; 63.4% and 13.5% metformin (52.7%; biguanide; concomitant use in Arm F excluded), empagliflozin (25.5%;
experienced any-grade hyperglycemia (grouped term); and AEs led to INAVO and PBO dis- SGLT-2 inhibitor), sitagliptin (22.7%; DPP-4 inhibitor), and pioglitazone (13.6%; thiazolidi-
continuation in 6.8% and 0.6% of pts, respectively. Conclusions: INAVO + PALBO + FULV nedione); insulin was used in 8.2% of pts. Median time to metformin start (excluding Arm F)
demonstrated a statistically significant and clinically meaningful OS benefit compared with PBO was 14 days (range, 1 to 1,710); the median start dose was 1,000 mg total daily; and the highest
+ PALBO + FULV. Improvement in INV-PFS was maintained during longer follow-up, along with a daily start dose was 2,000 mg. More than one anti-HG medication was often needed. Con-
substantial and statistically significant improvement in ORR. TTC was also substantially delayed clusions: A high proportion of prediabetic/obese pts were included in GO39374. In most of
(by ~2 years) by the addition of INAVO to PALBO + FULV. With longer exposure to INAVO, no new these pts, HG was manageable with dose interruptions and oral anti-HG medications, most
safety signals, nor changes in the safety profile, were noted, supporting good tolerability commonly metformin. Data support the use of INAVO in prediabetic/obese pts; further in-
(reflected in low discontinuation due to AEs). Clinical trial information: NCT04191499. Research vestigation of INAVO in pts with diabetes is warranted. Clinical trial information:
Sponsor: F. Hoffmann-La Roche Ltd; The authors acknowledge the Memorial Sloan Kettering NCT03006172. Research Sponsor: Genentech, Inc.; The authors acknowledge the Memorial
Cancer Center support grant (P30 CA008748). Sloan Kettering Cancer Center support grant (P30 CA008748).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
38s BREAST CANCER—METASTATIC

LBA1005 Oral Abstract Session 1007 Oral Abstract Session

A double-blind placebo controlled randomized phase III trial of fulvestrant Phase III of oral paclitaxel (DHP107) vs intravenous paclitaxel in HER2-
and ipatasertib as treatment for advanced HER2-negative and estrogen negative recurrent or metastatic breast cancer (mBC): Primary analysis of a
receptor positive (ER+) breast cancer following progression on first line CDK multinational optimal trial (NCT03315364). First Author: Sung-Bae Kim, Asan
4/6 inhibitor and aromatase inhibitor: The CCTG/BCT MA.40/FINER study Medical Center, University of Ulsan College of Medicine, Seoul, South Korea
(NCT04650581). First Author: Stephen K. L. Chia, BC Cancer Agency, Vancouver, BC, Background: DHP107 is a novel oral formulation of paclitaxel that is approved in South
Canada Korea and China for the treatment of gastric cancer. DHP107 had encouraging mono-
therapy anti-tumor activity with objective response rate (ORR) of 55% and median pro-
gression free survival (PFS) of 8.9 months (Mo) as first-line therapy in 31 patients with
HER2 negative metastatic breast cancer (mBC) in the OPTIMAL phase II study (Kim Ther
Adv Med Oncol 2021). The first primary analysis is reported herein. Methods: This phase
III, open-label, randomized, controlled trial evaluated the non-inferiority of DHP107 to
intravenous (IV) paclitaxel in mBC, with non-inferiority margin of 1.33. Patients (Pts) had
received one or more lines of endocrine-based therapy and no chemotherapy for mBC. Pts
from Korea, China, and Europe were randomized 1:1 to receive either DHP107 (200 mg/m²
orally, twice daily) or IV paclitaxel (80 mg/m² weekly). The primary endpoint was in-
vestigator-assessed PFS. Secondary endpoints included overall survival (OS), ORR, disease
control rate (DCR), quality of life (QoL), and safety. Results: With the median follow-up of
The full, final text of this abstract will be available at 38.8 Mo, the median age of the pts was 56 years. Of the 549 pts who underwent ran-
[Link] on the day of presentation and in the domization, 481 pts had hormone receptor positive (HR+) disease and 68 pts had triple
online supplement to the June 10, 2025, issue of the Journal negative disease. Among all pts, DHP107 demonstrated non-inferiority to IV paclitaxel in
PFS (mPFS: 10.02 vs. 8.54 Mo; HR 0.869, 95% CI 0.707–1.068). OS was comparable
of Clinical Oncology. between groups (mOS: 32.95 vs. 32.46 Mo; HR 0.979, 95% CI 0.769–1.246). Among
HR+HER2-pts, the mPFS was 10.74 Mo in the DHP107 arm, and 9.07 Mo in IV paclitaxel
arm (HR 0.869, 95% CI 0.700-1.080). QoL outcomes showed no significant differences.
ORR (45.8% vs. 39.7%) ad DCR (93.5% vs. 86.4%) were higher in the DHP107 group.
DHP107 was associated with lower incidences of peripheral neuropathy (37.91% vs.
48.29%), hypersensitivity reactions, musculoskeletal and connective tissue disorders, and
injection/infusion related reactions compared to IV paclitaxel. Neutropenia was the most
common toxicity in both groups, occurring more frequently in the DHP107 group (81.6% vs.
59.3%) with higher rates of Gr$3,4 neutropenia (67.15% vs. 29.66%), and febrile neu-
tropenia (6.14% vs. 0.76%), but no grade 5 events were reported. Gastrointestinal toxicities
were more frequent in the DHP107 group but were predominately Gr1. In this study,
discontinuation rate due to AEs were comparable (12.27% vs. 8.75%, p=0.2081) and AEs
leading to death occurred rarely in both groups (1.08% vs. 1.90%). Conclusions: DHP107
demonstrated comparable efficacy to IV paclitaxel with tolerable and manageable toxicity.
These results establish DHP107 as an effective, convenient alternative to IV paclitaxel for
patients with HER2-negative mBC, supporting its potential role in routine clinical practice.
Clinical trial information: NCT03315364. Research Sponsor: DAEWHA PHARM. CO., LTD.

1009 Clinical Science Symposium 1010 Clinical Science Symposium

Circulating tumor DNA, pathologic response after neoadjuvant therapy, and Circulating tumor (ct)DNA monitoring of ER+/HER2- high-risk breast cancer
survival: First results from TBCRC 040 (the PREDICT-DNA trial). First Author: during adjuvant endocrine therapy. First Author: Lajos Pusztai, Yale Cancer Center,
Natasha Hunter, University of Washington, Seattle, WA Yale School of Medicine, New Haven, CT
Background: Patients with Stage II/III breast cancer that overexpresses the human epidermal growth factor- Background: ctDNA monitoring during adjuvant endocrine therapy is an opportunity to
2 (HER2+) or is triple-negative (TNBC) generally receive upfront neoadjuvant therapy (NAT) before definitive detect molecular relapse before clinically apparent recurrence. ctDNA positivity rates,
surgery. Pathologic complete response (pCR) after NAT is associated with improved survival but a small
proportion of patients remain at risk for recurrence. Circulating tumor DNA (ctDNA) in patients whose
dynamics and the frequency of asymptomatic imaging-detectable metastatic disease at
primary tumors have detectable mutations could improve the identification of patients who remain at risk the time of ctDNA detection remain unknown in high-risk ER+/HER2- BCs. We present
after NAT. Methods: The Pathologic Response Evaluation and Detection In Circulating Tumor-DNA ctDNA results from a prospective, multicenter, randomized ctDNA interventional trial,
(PREDICT-DNA) trial was a prospective, multi-center study aimed at validating ctDNA as a biomarker for DARE (NCT04567420). Methods: Patients receiving adjuvant endocrine therapy for
treatment response in Stage II/III HER2+ or TNBC. The primary aim was to determine the negative predictive .6 months but ,7 years, with either recurrence risk .15% (PREDICT, RSPC, CTS5),
value (NPV) of ctDNA for residual disease following NAT; secondary aims included five-year invasive disease- .4 positive axillary lymph nodes, (primary tumor .5 cm, or 1-3 positive nodes with grade 3
free survival (IDFS), which were fit to a Cox proportional hazards model. Mutations were identified from
tumor tissue; ctDNA was then analyzed in pre- and post-NAT blood and compared with surgical pathology.
histology, or .3 cm tumor, or high molecular risk (Oncotype Dx RS .26, MammaPrint high
Proposed sample size was 229 patients based on simulation to control expected half-width of a confidence risk, EndoPredict .4, Prosigna score .60) were eligible for ctDNA surveillance with the
interval on NPV to be #15% when NPV=90%. The Personalis NeXT Personal ctDNA assay was centrally Signatera assay (Natera, Inc.) every 6 months. ctDNA+ patients had systemic staging with
performed. Results: 228 participants were enrolled in 24 sites between 2016 and 2018. 53% had TNBC, and imaging and if there was no evidence of metastatic disease patients were randomized to
47% had HER2+ disease. 92.2% (n=166/180) had detectable ctDNA at baseline, and 46% of patients had pCR switching to fulvestrant + palbociclib (Arm A) or to continuation of adjuvant therapy (Arm
(42% TNBC, 50% HER2+). 54% of all post-NAT ctDNA detections were in the ultrasensitive range below 100 B). Negative predictive value (NPV) was calculated for recurrence in the screening group
PPM. Among 112 subjects with undetectable ctDNA prior to surgery, 45 were found to have residual disease
resulting in an NPV of 60% (CI 0.51-0.69). Patients with TNBC and detectable ctDNA prior to surgery were
after each ctDNA- test. In randomized patients, early ctDNA dynamics were correlated with
approximately 12 times more likely to experience a recurrence regardless of pCR (HR 12.8 [95% CI: 2.3-71.5]). recurrence-free survival (RFS) and ctDNA clearance rates were calculated by trial arm.
See Table describing landmark IDFS analyses after surgery. Conclusions: While lack of ctDNA detection Results: 552 patients had tissue sent for assay design; 494 had ctDNA results; 52 failed
after NAT and before surgery did not predict pCR, initial analysis of predefined secondary objectives suggest WES and/or had incomplete tumor/normal/blood sets; 6 had pending reports. Among
that ctDNA-negative patients before surgery have excellent prognosis regardless of pCR, particularly if TNBC. patients not randomized, 432 were ctDNA-, of these N=43 had one time point and 389 had
This suggests that ctDNA may be a better biomarker for long term clinical outcomes than pCR. Further .2 ctDNA- result, overall median screening time 27.4 months (0-45.5), 4 ctDNA- patients
correlations and interactions will be presented. Clinical trial information: NCT02743910. Research Sponsor:
Susan B. Komen Breast Cancer Foundation; Breast Cancer Research Foundation; Johns Hopkins Clinical
had recurrence (NPV 100% at 6 months and 99% at 12 months post-testing). Forty patients
Research Network Research Accelerator and Mentorship Program (RAMP); Commonwealth Foundation; were randomized, 34 had post-randomization ctDNA result. Randomization rates were 53%
NIH/NCI grant; R01CA194024; NIH/NCI grant; R01CA214494; NIH/NCI grant; R01CA289528; NIH/NCI grant; and 76% for patients who tested ctDNA-positive on the first screening (N=19) versus those
P50CA098131; NIH/NCI grant; P30CA06485; The Helen Golde Fund; NIH/NCI grant; P30CA006973; The who turned positive in follow up testing (N=15). At any time post-randomization, ctDNA
Translational Breast Cancer Research Consortium (TBCRC). clearance rates were 63% (10/16) in Arm A and 22% (4/18) in Arm B. Among randomized
Invasive disease-free survival (IDFS) by breast cancer subtype, according to ctDNA after NAT and pathologic patients, 6 of 9 patients with increased ctDNA levels from the pre-randomization to the 3-
response.
month on-treatment recurred (median time to recurrence 4.8 months, range: 3.3-24.3),
3y IDFS 4y IDFS 5y IDFS
TNBC (total n=64) (n=40) (n=32) (n=18)
among those with a decrease in ctDNA post-treatment only 1 of 6 experienced recurrence
at 10.3 months (HR: 5.3, 95% CI: 1.1-53, p=0.04). Conclusions: This study demonstrates
ctDNA- & pCR (n=20) 94.1% 94.1% 94.1%
ctDNA- & RD (n=25) 95.8% 89.8% 89.8% the ability of ctDNA to identify breast cancer patients at high risk of relapse for ran-
ctDNA+ & RD (n=19) 48.9% 48.9% 48.9% domization in a prospective, multicenter, randomized clinical trial. Patients with serially
HER2+ (total n=58) 3y IDFS 4y IDFS 5y IDFS
(n=41) (n=31) (n=17) ctDNA- results during surveillance had 99% RFS after a median f/u of 27.4 months. Interim
ctDNA- & pCR (n=19) 94.1% 94.1% 94.1% analysis revealed higher clearance rates in Arm A compared to patients randomized to Arm
ctDNA- & RD (n=31) 92.6% 87.5% 87.5%
ctDNA+ & RD (n=8) 60.0% 60.0% 60.0% B. Early on treatment ctDNA dynamics is prognostic of patient outcomes. Clinical trial
information: NCT04567420. Research Sponsor: Pfizer; Natera Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 39s

1011 Clinical Science Symposium 1012 Clinical Science Symposium

Circulating tumor DNA (ctDNA) dynamics as a predictor of treatment re- Assessment of ctDNA somatic homologous recombination deficiency (HRD)
sponse in metastatic breast cancer (mBC). First Author: Pedram Razavi, Memorial in triple-negative breast cancer (TNBC) from SWOG S1416 trial. First Author:
Sloan Kettering Cancer Center, New York, NY Shane R. Stecklein, University of Kansas Medical Center, Kansas City, KS
Background: ctDNA testing has emerged as a prognostic and predictive biomarker in Background: HRD is observed in up to two-thirds of gBRCA-wildtype TNBC. S1416
the management of mBC. However, the relationship between ctDNA trends and real- (NCT02595905) showed that addition of a PARP inhibitor (veliparib) to cisplatin improved
world treatment outcomes has yet to be fully characterized. Here, we utilized a claims progression-free survival (PFS) in gBRCA-wildtype metastatic TNBC (mTNBC) with HRD
database to evaluate the association between ctDNA trends and the time to next phenotype (“BRCA-like”). In this study, we sought to evaluate concordance between
treatment (TTNT) in patients with mBC. Methods: We utilized Natera’s proprietary real- circulating tumor DNA (ctDNA)-based detection of somatic homologous recombination
world database linked to commercially available claims data to identify patients who repair (sHRR) alterations and tumor-based HRD and to assess if sHRR deficiency (sHRR+)
received treatment for mBC and had ctDNA testing performed commercially using a was associated with benefit from veliparib in gBRCA-wild type mTNBC in SWOG S1416.
clinically validated, personalized, tumor-informed mPCR-NGS ctDNA assay (Signater- Methods: S1416 enrolled patients with mTNBC who had received # 1 line of prior therapy
aTM, Natera, Inc.). Insurance claim codes for treatment regimens were used to determine and randomized them to cisplatin plus veliparib or placebo. Central gBRCA1/2 testing
BC receptor subtype and therapy dates. Treatment lines were included in the analysis if a classified patients as gBRCA-mutated or -wildtype. An a priori defined biomarker panel
ctDNA test result was available within 4 weeks before treatment initiation (T1) and a classified gBRCA-wildtype patients into BRCA-like (HRD+) and non-BRCA-like (HRD-)
subsequent ctDNA test result was available 2-6 weeks after treatment initiation (T2). groups. A third group with gBRCA-wildtype, but without tissue BRCA classification was
TTNT was calculated as the time from initiation of the first treatment to the subsequent also included. Pre-treatment and progression plasma samples were utilized for assess-
treatment. T1 to T2 ctDNA dynamics were analyzed using the Student’s t-test and were ment of sHRR status. ctDNA was analyzed using the Guardant OMNI next-generation
categorized as favorable (persistently negative, ctDNA-clearance, ctDNA-decrease) or sequencing platform. sHRR+ was defined by detectable somatic alterations (SNVs,
unfavorable (ctDNA-negative to positive, ctDNA-increase). Results: A total of 7,222 INDELs, fusions with a functional impact notation of deleterious, and/or CNVs with a
treatment lines were assessed for duration of treatment and corresponding ctDNA functional characterization of homozygous deletion) in a 24 gene panel. Results: Among
dynamics, including3,117 lines (N=2,362 patients) for HR+/HER2- mBC, 3,717 lines N=213 gBRCA-wildtype patients with evaluable pre-treatment blood samples, 25% were
sHRR+. Among sHRR+ patients, alterations in CHEK2 (18%), BRCA1 (17%), BARD1 (8%),
(N=1,943 patients) for HER2+ mBC, and 888 lines (N=605 patients) for TNBC. Of these,
ATM (7%), BAP1 (7%), CDK12 (7%), NBN (7%), BRCA2 (5%), and FANCA (5%) accounted for
448 treatment lines met the inclusion criteria for ctDNA analysis. In HER2+ breast
80% of sHRR alterations. Most sHRR+ patients (91%) had alterations in only one of 24
cancer, TTNT across 226 treatment lines was significantly longer in patients with fa-
genes, suggesting mutual exclusivity of homologous recombination pathway alterations in
vorable ctDNA dynamics (6.7 [3.18–10.3] months) relative to unfavorable dynamics (2.7
ctDNA. sHRR+ status was numerically higher in BRCA-like compared to non-BRCA-like
[1.4–5.1] months; p,0.0001). Among patients with HR+/HER2- mBC, TTNT across 156 tumors or unclassified tumors (32% vs. 20% vs. 20%, respectively; P=0.12). Among n=98
treatment lines was longer in those with favorable ctDNA dynamics (median [Q1, Q3]: patients with availability of evaluable pre-treatment and progression samples, 31% were
7.51 [3.72–11.57] months) compared to unfavorable dynamics (5.02 [1.8–9.84] months; sHRR+ at baseline and 28% were sHRR+ at progression. Numerically, conversion from
p=0.052). A similar trend was observed in TNBC, where TTNT across 66 treatment lines sHRR+ to sHRR- was more common than conversion from sHRR- to sHRR+ (30% vs. 9%,
was longer with favorable ctDNA dynamics (6.03 [2.89–10.07] months) compared to respectively). sHRR was not prognostic for PFS (median 4.3 (sHRR+) vs. 4.1 (sHRR-)
unfavorable dynamics (2.7 [1.16–5.89] months; p=0.381), though this was not sta- months, respectively; P=0.30) nor predictive of benefit from veliparib (P=0.40). Conclu-
tistically significant. Conclusions: Early on-treatment ctDNA dynamics, assessed within sions: One-fourth of gBRCA-wildtype mTNBC patients have ctDNA sHRR alterations, and
the first 6 weeks of therapy, was associated with TTNT in a real-world ctDNA monitoring there is incomplete overlap between tumor- and ctDNA-assessed HRD. ctDNA sHRR
setting across different mBC subtypes and therapeutic regimens. An early rise in ctDNA alterations were mostly mutually exclusive. Approximately one-third of patients with
levels correlated with the shortest TTNT, whereas ctDNA clearance was associated with baseline sHRR+ converted to sHRR- at time of progression while receiving DNA damaging
the longest TTNT intervals. These findings highlight the potential of serial ctDNA testing chemotherapy. sHRR was not prognostic and did not predict benefit from veliparib in
in mBC for monitoring treatment response and informing clinical decisions. Research S1416. Research Sponsor: NIH/NCI/NCTN; U10CA180888; NIH/NCI/NCTN; U10CA180819.
Sponsor: None.

1013 Oral Abstract Session 1014 Rapid Oral Abstract Session

Exploratory biomarker analysis of trastuzumab deruxtecan (T-DXd) vs phy- Use of artificial intelligence–assistance software for HER2-low and HER2-
sician’s choice of chemotherapy (TPC) in HER2-low/ultralow, hormone ultralow IHC interpretation training to improve diagnostic accuracy of pa-
receptor–positive (HR+) metastatic breast cancer (mBC) in DESTINY- thologists and expand patients’ eligibility for HER2-targeted treatment. First
Breast06 (DB-06). First Author: Rebecca Alexandra Dent, National Cancer Centre Author: David Mulder, Mindpeak GmbH, Hamburg, Germany
Singapore, Singapore, Singapore Background: The advent of HER2-targeted antibody-drug conjugates and the intro-
Background: DB-06 (NCT04494425), a Phase 3, randomized, open-label study, demonstrated duction of HER2-low and HER2-ultralow diagnostic categories have made precise HER2
a clinically meaningful progression-free survival (PFS; 13.2 vs 8.1 months [hazard ratio: 0.64]) IHC assessment crucial for optimal breast cancer treatment. However, reproducible and
benefit with T-DXd vs TPC (capecitabine, nab-paclitaxel, or paclitaxel) in patients with HR+, accurate HER2 IHC scoring, particularly in cases with low level of HER2 expression,
HER2-low (immunohistochemistry [IHC] 1+ or IHC 2+ / in situ hybridization–negative) or remains challenging. Many patients with HER2-low or HER2-ultralow expression risk
-ultralow (IHC 0 with membrane staining) mBC after $1 endocrine-based therapy (primary being misclassified as HER2 null, potentially missing access to effective HER2 targeted
data cutoff: March 18, 2024). Here, we report an exploratory circulating tumor DNA (ctDNA) therapies. Artificial intelligence (AI) assisted HER2 assessment may improve pathol-
analysis based on baseline genomic status. Methods: Baseline ctDNA profiling in blood ogists’ diagnostic accuracy and concordance during interpretation training, especially in
samples was assessed via Guardant OMNI 500-gene liquid biopsy assay. In total, 625 patients challenging cases with minimal membrane staining. Methods: A training platform for
had evaluable ctDNA samples and putative tumor content, and comprised the biomarker
AI-supported digital HER2 IHC assessment of breast cancer samples was developed for
evaluable population (BEP) presented herein. Baseline characteristics and efficacy outcomes
pathologists. A total of 105 pathologists from 10 countries participated in masterclass
were evaluated in key genomic subgroups (PI3K pathway, ESR1m, BRCA1/2m), including
confirmed objective response rate (cORR) and PFS, both by blinded independent central
sessions, assessing 20 digital HER2 IHC-stained breast cancer cases both without and
review. Results: Genomic alterations were observed in 45.0% (PI3K pathway, n=281), 51.5% with AI assistance. Cases assigned ground-truth IHC scores by a central reference
(ESR1m, n=322), and 7.7% (BRCA1/2m, n=48) of patients. The median PFS (mPFS) for each center, were divided into three exams: A (n = 5), B (n = 7), and C (n = 8). The mas-
mutational subgroup was 13.2 (T-DXd) and 7.1 (TPC) months (PI3K pathway), 11.3 (T-DXd) terclasses consisted of: (1) Exam A, (2) a lecture on HER2 IHC scoring, (3) Exam B, (4)
and 7.0 (TPC) months (ESR1m), and 21.4 (T-DXd) and 5.6 (TPC) months (BRCA1/2m). T-DXd discussion of results from Exams A and B, and (5) AI-assisted Exam C. The AI software
improved PFS and cORR outcomes compared with TPC across all mutational subgroups was used for decision support only for Exam C. The HER2 IHC scoring followed ASCO/
reported (Table). Conclusions: In this exploratory ctDNA analysis, T-DXd demonstrated a CAP 2023 guidelines, adapted to include the HER2-ultralow (IHC 0 with membrane
greater clinical benefit vs TPC regardless of PI3K pathway, ESR1, or BRCA1/2 mutation. staining) and HER2 null (IHC 0 with no membrane staining), and provided individual
Clinical trial information: NCT04494425. Research Sponsor: AstraZeneca; Daiichi Sankyo. tumor cell classifications for explainability. Results: Across 1,940 readings, patholo-
gists achieved an average agreement of 76.3% with reference scores without AI (Exams
T-DXd TPC T-DXd TPC
BEP (N=625) subgroup cORR, cORR, mPFS, mPFS, PFS A+B), compared to 89.6% with AI-assistance (Exam C). For HER2 clinical categories (null,
(n=T-DXd/TPC) % % mo* mo* hazard ratio ultralow, low, positive) accuracy improved from 66.7% without AI to 88.5% with AI.
PI3K pathway† 57.6 41.5 13.2 7.1 0.65 Misclassification of HER2-ultralow cases as HER2 null occurred in 29.5% of readings
(139/142) [48.9, 65.9] [33.3, 50.1] [9.9, 15.5] [6.0, 9.5] [0.48, 0.87] without AI but decreased to 4.0% with AI assistance. Conclusions: AI-assisted training
ESR1m 60.2 32.1 11.3 7.0 0.64 improved pathologists’ accuracy in HER2 IHC scoring by 13.3%, compared to central
(166/156) [52.4, 67.7] [24.8, 40.0] [9.8, 13.5] [5.6, 9.3] [0.49, 0.83]
BRCA1/2m 80.0 39.3 21.4 5.6 0.14
reference scores. Furthermore, AI reduced the misclassification of HER2-low and HER2-
(20/28) [56.3, 94.3] [21.5, 59.4] [15.2, NE] [4.1, 6.9] [0.05, 0.33] ultralow cases as HER2 null by 25.5%, potentially enabling more patients to access
HER2-targeted therapies. These findings highlight the value of AI systems in biomarker
Square brackets = 95% CIs (based on the Clopper-Pearson [cORR] or Brookmeyer-Crowley method [PFS]). PFS
hazard ratios and CIs based on Cox proportional hazards model with no stratification factors, and ties handled interpretation training, providing pathologists with enhanced decision-making tools at
by Efron approach. A hazard ratio ,1 favors T-DXd vs TPC. No formal testing of significance was performed; the individual cell level and improving diagnostic precision in HER2 IHC interpretation.
*Number of PFS events: 89 (T-DXd) and 92 (TPC) in the PI3K pathway group, 115 (T-DXd) and 107 (TPC) in the Research Sponsor: AstraZeneca.
ESR1m group, and 7 (T-DXd) and 23 (TPC) in the BRCA1/2m group; †includes AKTm, PIK3CAm, and PTENm;
CI, confidence interval; m, mutation; mo, months; NE, non-evaluable.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
40s BREAST CANCER—METASTATIC

1015 Rapid Oral Abstract Session 1016 Rapid Oral Abstract Session
Treatment rechallenge after trastuzumab-deruxtecan–related interstitial Phase IB and II study of ribociclib with trastuzumab plus endocrine therapy
lung disease: A multi-institution cohort study. First Author: Kelsey H. Natsu- in HR+/HER2+ advanced breast cancer patients: Korean Cancer Study Group
hara, University of California, San Francisco Helen Diller Family Comprehensive Cancer BR 18-2 MINI trial. First Author: Joohyuk Sohn, Division of Medical Oncology, De-
Center, San Francisco, CA partment of Internal Medicine, Yonsei Cancer Center, Yonsei University College of
Background: T-DXd is an antibody-drug conjugate approved for advanced HER2+/low/ultra-low Medicine, Seoul, South Korea
breast cancer and multiple other solid tumors. T-DXd carries a rare but serious risk of ILD (incidence Background: In HER2+ advanced breast cancer (ABC), standard treatment has been
12-15%), requiring frequent imaging and symptom evaluation. For . grade (G) 2 ILD, guidelines anti-HER2 therapy with chemotherapy, regardless of hormone receptor status. While
recommend permanent drug discontinuation. For asymptomatic G1 ILD, drug is held with the option
for rechallenge (RC) if imaging findings resolve. Limited data exist on outcomes of RC after ILD in
prior studies support the use of CDK4/6 inhibitors with anti-HER2 and endocrine therapy
diverse real-world patients (pts). Methods: In this multi-center retrospective study, we analyzed pts in pretreated HR+/HER2+ ABC, data on their first line use without chemotherapy are
with T-DXd related ILD treated from 2017-2024. Pts with ILD were identified via chart/ICD code review. limited. This study investigates ribociclib, trastuzumab, and letrozole as a first-line
Adjudication of T-DXd related ILD was based on treating providers’ assessment and graded via CTCAE combination in HR+/HER2+ ABC. Methods: This multicenter, single-arm, prospective
v5. We collected pt demographics, T-DXd and steroid dosing, imaging results, and outcomes after RC. trial was conducted across 17 academic institutions in South Korea (NCT03913234).
Statistical analysis was performed using Wilcoxon rank sum and Fisher’s exact tests. Results: Four Eligible patients were HR+/HER2+ ABC with no prior systemic therapy for metastatic
centers treated 712 pts with T-DXd, with a 9.1% rate of any grade ILD (n=65). One other center reported disease. The Phase IB study used a 3+3 design to determine the recommended Phase II
only RC data in 18 pts with ILD. In total, 47 pts were RC; 38 after G1 ILD (81%), 9 after G2. Median (med)
dose (RPIID) of ribociclib with fixed doses of letrozole (2.5 mg QD) and trastuzumab (8
time to initial ILD was 145 days (d) after 1st dose (interquartile range [IQR] 78-205). Demographics for
pts RC are shown in the table. Among 50 pts with G1 ILD, including pts not RC, 28/50 (56%) received mg/kg loading, then 6 mg/kg every 3 weeks). The Phase II trial evaluated efficacy and
steroids for a med of 36d (IQR 27-79). Radiographic improvement was seen at a med of 24d (IQR 19- safety at the RPIID. The primary endpoint was progression-free survival (PFS), targeting
63) for pts treated with steroids vs 82d (IQR 48-94) without (p,0.01); and a med of 35d (IQR 22-82) for an improvement from 8 to 12 months. Secondary endpoints included overall survival
pts RC vs 81d (IQR 68-105) for pts not RC (p=0.01). Among pts with G1 ILD, 38/50 (76%) were RC at a (OS), objective response rate (ORR), duration of response (DOR), and safety. PAM50
med of 42d (IQR 36-57) from last dose; 23/38 (61%) were dose reduced. After RC, pts remained on T- testing assessed correlations between intrinsic subtype and treatment efficacy. Re-
DXd for a med of 215d (IQR 60-334); 10/38 (26%) developed recurrent ILD (7-G1, 2-G2, 1-G3) at a med sults: Phase IB (n = 13) identified the RPIID as ribociclib 600 mg QD, with one dose-
of 211d (IQR 47-273) from RC. No statistically significant differences were seen between ILD onset, limiting toxicity (Grade 3 ALT elevation) at 400 mg. In Phase II, 77 patients were enrolled,
time to RC, or demographics for pts with recurrent ILD vs not. Of the 9 pts RC after G2 ILD, T-DXd was
continued for a med of 129d (IQR 49-171); 2/9 (22%) developed recurrent ILD (1-G2, 1-G3). No G5
with a median age of 61 years (range 31–85), 18.2% (14/77) premenopausal, 66.2% (51/
toxicity was seen with RC. Conclusions: In this multi-center study, high RC rates were seen after G1 77) HER2 IHC 3+ and recurrent disease in 64.9% (50/77). 66.2% (51/77) had visceral
ILD with long duration of clinical benefit. Pts treated with steroids had faster radiographic ILD metastases. At a median follow-up of 15.8 months (95% CI: 12.9-19.1) months, the
improvement, highlighting the importance of early steroid use. Among pts RC after G1 ILD, recurrent median PFS was 30.4 months (95% CI: 19.6–NA), meeting the primary endpoint. The
ILD rates were low, with the majority G1 and no G5 events. Notably, 9 pts with G2 ILD were RC, with a median OS was not reached. The ORR was 61.1%, including 3 complete and 41 partial
similar rate of recurrent ILD; this must be interpreted cautiously. Our large cohort data further supports responses, with a DOR of 11.8 months (95% CI: 7.6-13.4). Common adverse events
the safety of T-DXd RC in diverse real-world settings. Research Sponsor: None. included neutropenia (66.7%), pruritus (24.4%), and nausea (22.2%). There was a death
RC Pt Characteristics (n=47) n (%) or Median (IQR) reported due to aortic aneurysm. PAM50 analysis in 75 patients (phase IB/II) showed no
Cancer type
significant correlation between intrinsic subtype and efficacy. Conclusions: Ribociclib,
Breast 43 (91) trastuzumab, and letrozole as first-line therapy in HR+/HER2+ ABC demonstrated a
GI 3 (6) median PFS of 30.4 months with a manageable safety profile, supporting its potential as
Gyn 1 (2) a chemotherapy-free option. Clinical trial information: NCT03913234. Research
Age (yrs) 57 (52-68)
Prior # therapy lines in the advanced/metastatic setting 3 (1-5) Sponsor: None.
Renal impairment (CrCl < 60 mL/min) 8 (17)

1017 Rapid Oral Abstract Session 1018 Rapid Oral Abstract Session
HER2-ADC trastuzumab rezetecan (SHR-A1811) in HER2-positive breast A phase II clinical study of adebrelimab and bevacizumab combined with
cancer with brain metastases: Update results from REIN trial. First Author: Min cisplatin/carboplatin in triple-negative breast cancer patients with brain
Yan, Department of Breast Disease, Henan Breast Cancer Center, The Affiliated Cancer metastases. First Author: Ting Li, Department of Breast and Urologic Medical On-
Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, China cology, Fudan University Shanghai Cancer Center, Shanghai, China
Background: HER2-directed antibody-drug conjugates (ADCs) have been demonstrated Background: Brain metastases (BMs) of triple-negative breast cancer (TNBC) is a lethal
to be of intracranial activity in patients with HER2+ breast cancer (BC) with brain disease often associated with a limited life span of approximately 6 months and local
metastases (BM). Our prospective, non-randomized phase 2 trial (NCT05769010) aimed therapy is usually the first treatment choice due to lack of effective anti-tumor agents.
to assess the feasibility of SHR-A1811, a novel HER2-target ADC, with or without other Here reported a triplet, anti-PD-L1 (Adebrelimab, SHR-1316), bevacizumab plus cisplatin/
anti-tumor agents in HER2-expressing BCBM. Here we report the data of SHR-A1811 carboplatin in BMs of triple negative breast cancer. Methods: This is a single center,
combined with bevacizumab in HER2+ BCBM, and update the results of SHR-A1811 in single-arm, phase II clinical trial involving triple-negative breast cancer patients with
HER2+ BCBM (preliminary ORR data of the first 25 patients in Arm 1 has been published active brain metastases. A total of 35 participants were administered a triplet treatment
at 2024 ASCO), presenting the efficacy and safety of SHR-A1811 alone or in combination consisting of Adebrelimab, bevacizumab and cisplatin/carboplatin. Prior use of bev-
in the treatment of HER2+ BCBM. Methods: Patients with HER2-positive or -low BC with acizumab or anti-PD-1/PD-L1 was not allowed. Prior use of platinum was allowed only in
at least one radiotherapy-naı̈ve measurable intracranial lesion were eligible for our trial. cases with platinum-sensitive disease. The primary endpoint was the objective response
The patients with HER2+ disease enrolled in Arm 1 received SHR-A1811 6.4 mg/kg every rate in the central nervous system (CNS-ORR), and the secondary endpoints included the
3 weeks, while those in Arm 3 were assigned to SHR-A1811 4.8 mg/kg and bevacizumab clinical benefit rate in CNS (CNS-CBR), progression-free survival (PFS), overall survival
15 mg/kg every 3 weeks until disease progression, unaccepted toxicity, or no further (OS), the first progression site and safety. Results: The data cutoff for this analysis was
benefit. The primary endpoint was the intracranial overall response rate (ORR-IC) per on December 20, 2024. A total of 35 patients enrolled in this study from August 2020 to
RANO-BM. Results: Between March 30, 2023, and June 3, 2024, 58 patients were October 2024. Among all patients, 42.9% (15/35) had neurological symptoms at baseline,
enrolled in Arm 1 (n = 33) and Arm 3 (n = 25). Among these, 56 patients (96.6%) had and 80% (28/35) had not received any local treatment for their brain metastases. The
received anti-HER2 therapy previously, and the median number of prior systemic median number of previous lines of therapy for metastatic disease was 2 (range 0-4), with
therapies in advanced setting was 2 (range: 0-9). 54 patients received at least one 40% (14/35) patients having received a prior platinum agent. In the intention-to-treat
efficacy assessment and the confirmed ORR-IC in Arm 1 and Arm 3 were 84.4% (27/32) population,which comprised patients who received at least one cycle of study treatment,
and 72.7% (16/22) respectively, which were numerically identical to the overall ORR in the CNS-ORR was 77.1% (27/35), with 5 complete responses (CR), 22 partial responses
each arm, and all patients achieved intracranial disease control. As of December 31, (PR) and the confirmed CNS-ORR was 71.4%(25/35). Among the 23 patients who
2024, the median PFS of Arm 1 was 13.2 (95% CI: 10.0-15.4) months, while the median progressed, the brain was the site of first progression in 69.6% (16/23) of patients. The
PFS of Arm 3 was not mature. 78.8% (26/33) of patients in Arm 1 and 48.0% (12/25) in median PFS was 7.6 months (95%CI, 5.7-11.5), while CNS-PFS was 10 months (95%CI,
Arm 3 experienced treatment-related adverse events (TRAEs) of grade 3 or 4, and the 7.4-12.6), and median OS was 16 months (95%CI, 11.7 to not reached). Treatment-related
frequencies of grade 4 TRAEs were 36.4% and 4% respectively. The grade 3/4 TRAEs that adverse events (TRAEs) were reported in 100% (35/35) of patients, with the incidence of
occurred in more than one patient included decreased neutrophil counts (Arm 1 / Arm 3: grade$3 TRAEs being 48.6% (17/35), including neutropenia (8.6%, 3/35) and platelet
69.7% / 36.0%), decreased leucocyte counts (51.5% / 16.0%), decreased platelet counts count decreased (8.6%, 3/35). Adebrelimab related serious adverse events (SAEs) oc-
(30.3% / 0%), anemia (21.2% / 8.0%), decreased lymphocyte counts (21.2% / 0%), and curred in one patient (facial nerve disorder), no treatment-related deaths were reported.
nausea (6.1% / 0%). Conclusions: Our findings showed that SHR-A1811 6.4 mg/kg Conclusions: The triplet treatment of anti-PD-L1 Adebrelimab, bevacizumab and cis-
solely or SHR-A1811 4.8 mg/kg combined with bevacizumab both can attain high platin/carboplatin, was the first regimen demonstrating a high intracranial anti-tumor
intracranial remission rates, while the lower-dose combination regimen might exhibit a activity, a prolonged CNS-PFS and OS with a good safety profile. Warranting further
better safety profile. The long-term outcomes will continue to be followed up. Clinical investigation in this highly aggressive disease. Clinical trial information: NCT04303988.
trial information: NCT05769010. Research Sponsor: None. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 41s

1019 Rapid Oral Abstract Session 1020 Rapid Oral Abstract Session
Sacituzumab tirumotecan (sac-TMT) as first-line treatment for unresectable Dose optimization of PF-07248144, a first-in-class KAT6 inhibitor, in pa-
locally advanced/metastatic triple-negative breast cancer (a/mTNBC): Ini- tients (pts) with ER+/HER22 metastatic breast cancer (mBC): Results from
tial results from the phase II OptiTROP-Breast05 study. First Author: Yongmei phase 1 study to support the recommended phase 3 dose (RP3D). First
Yin, Department of Oncology, the First Affiliated Hospital of Nanjing Medical University, Author: Patricia LoRusso, Yale School of Medicine, New Haven, CT
Nanjing, China Background: PF-07248144 is a selective catalytic inhibitor of KAT6, a histone lysine
Background: TROP2 (trophoblast cell surface antigen 2) is highly expressed in TNBC and acetyltransferase. To inform the RP3D, we evaluated two pharmacokinetically distin-
associated with poor survival. Sac-TMT (MK-2870/SKB264) is a TROP2 ADC developed guishable doses of PF-07248144 in combination with fulvestrant (FUL) from a phase 1
with a novel linker to conjugate the payload, a belotecan-derivative topoisomerase I study in ER+/HER22 mBC in a dose expansion phase. Methods: Pts with ER+/HER22
inhibitor. It is approved in China for pts with a/mTNBC who have received at least two prior mBC after prior CDK4/6i and endocrine therapy (ET) received PF-07248144 at recom-
chemotherapies, including one for metastatic disease. The Phase II OptiTROP-Breast05 mended doses for expansion (RDEs) of 5 mg QD alone, 5 mg QD plus FUL, or 1 mg QD plus
study (NCT05445908) evaluated sac-TMT as first-line treatment for pts with a/mTNBC. FUL (N = 107) and were followed up (at least 6 months across all cohorts) to assess for
The study also explored the impact of PD-L1 combined positive score (CPS) status. Pts safety and efficacy. Primary objective wassafety/tolerability per CTCAE 5.0 and RDE
with CPS , 10 (PD-L1-negative, IHC 22C3 pharmDx) have limited treatment options, selection. Other objectives included antitumor activity per RECIST 1.1, PK, PD, and
representing a critical unmet need. Methods: Pts with a/mTNBC who had not received predictive biomarkers. Results: 5 mg QD was identified as the RDE for both PF-07248144
prior treatment for advanced disease were enrolled, regardless of PD-L1 or TROP2 status, monotherapy (35 pts treated) and FUL combination (43 pts treated) based on safety, PK,
to receive sac-TMT at 5 mg/kg Q2W until disease progression or unacceptable toxicity. PD, and antitumor activity. 1 mg PF-07248144 plus FUL (29 pts treated) was selected as
For pts with recurrent TNBC, a disease-free interval (DFI) of at least 6 months was the lower RDE based on a distinguishable PK and safety profile while achieving maximal
required for eligibility. Tumor assessment was performed every 6 weeks per RECIST v1.1 blood and tumor PD marker reduction and efficacious concentrations supported by
as assessed by investigator. Results: As of 18 Nov 2024, a total of 41 pts (median age 55 preclinical models. As of Oct 11, 2024, a total of 107 pts were treated at RDEs. Baseline pt
yrs; 43.9% ECOG PS 1; 78.0% PD-L1 CPS , 10) were enrolled; 61.0% of pts had visceral characteristics from the two RDEs plus FUL were comparable. All pts received prior CDK4/
metastases at baseline, 29.3% of pts had de novo metastasis, 19.5% of pts had a DFI of 6- 6i and ET in the metastatic setting. Positive dose-response relationships were identified for
12 months (mos), and 51.2% of pts had a DFI . 12 mos. The median follow-up was 18.6 both safety (neutropenia) and efficacy (objective response rate [ORR]) endpoints. At 5 mg
mo. The objective response rate (ORR) was 70.7% (29/41, 3 unconfirmed PR) and the and 1 mg doses plus FUL, the most common treatment-related adverse event (TRAE) was
disease control rate (DCR) was 92.7%. Median duration of response (mDoR) was 12.2 mo, dysgeusia (G1+G2: 83.7% vs 89.7%). The most common G$3 TRAE was neutropenia (G3:
while the median progression-free survival (mPFS) was 13.4 mo, and the 12-mo PFS rate 39.5% vs 20.7%; G4: 7.0% vs 0.0%). The neutropenia was reversible and manageable with
was 64.6% (95% CI: 45.0%, 78.7%). Among the 32 pts with PD-L1 CPS , 10, the ORR was dose modifications. No febrile neutropenia was observed. The safety profile of 5 mg PF-
71.9% (23/32, 3 unconfirmed PR) and the DCR was 93.8%. The mPFS in this subgroup was 07248144 monotherapy was consistent with 5 mg RDE plus FUL. No events of pneumonitis
13.1 mo, with a 12-mo PFS rate 59.1% (95% CI: 37.1%, 75.7%). Treatment-related adverse were reported in the 107 pts treated. For FUL plus 5 mg and 1 mg PF-07248144, ORR was
events (TRAEs) of grade 3 or higher occurred in 63.4% of pts. The most common $ grade 37.2% (95% CI: 23.0–53.3) vs 24.1% (10.3–43.5); median duration of response was 15.8
3 TRAEs (occurred in $5% of pts) were neutrophil count decreased (46.3%), WBC count mos (9.2–not estimable [NE]) vs 4.6 mos (3.4–NE); clinical benefit rate was 55.8%
decreased (34.1%), anemia (12.2%), stomatitis (9.8%), lymphocyte count decreased (39.9270.9) vs 37.9% (20.7–57.7). With median duration of follow-up 21.9 mos and 11.0
(7.3%) and fatigue (7.3%). No treatment-related deaths occurred, and there were no mos for pts receiving FUL plus 5 mg and 1 mg PF-07248144, the median progression-free
reports of neuropathy or interstitial lung disease/pneumonitis. Conclusions: Sac-TMT survival was 10.7 mos (95% CI: 5.3213.8) vs 3.6 mos (1.8–5.6), respectively. Conclu-
demonstrated promising anti-tumor activity with a manageable safety profile as a first- sions: Based on a thorough benefit–risk assessment of two pharmacokinetically dis-
line treatment for pts with a/mTNBC, independent of the PD-L1 status. A Phase 3 study tinguishable doses with sufficient number of pts and follow up, 5 mg QD PF-07248144 was
comparing sac-TMT vs investigator’s choice of chemotherapy in first-line PD-L1-negative identified as the optimal dose in combination with FUL with acceptable safety and en-
(CPS , 10) a/mTNBC is currently underway (NCT06279364). Clinical trial information: couraging activity. A pivotal phase 3 trial is planned to address the high unmet medical
NCT05445908. Research Sponsor: Sichuan Kelun-Biotech Biopharmaceutical Co., Ltd. need in ER+/HER22 mBC after progression on CDK4/6i plus ET. Clinical trial information:
NCT04606446. Research Sponsor: Pfizer Inc.

1021 Rapid Oral Abstract Session 1022 Poster Session

Phase II study of trastuzumab-pkrb plus gedatolisib in patients with HER2- Longitudinal tissue analysis and correlation of microenvironmental changes
positive metastatic breast cancer who progressed after 2 or more HER2- with combined immunotherapy and targeted therapy response in metastatic
directed chemotherapies (KM-10A/KCSG BR18-13). First Author: Ju Won Kim, breast cancer. First Author: Jieqiong Liu, Yat-sen Memorial Hospital of Sun Yat-sen
Korea University Anam Hospital, Seoul, South Korea University, Guangzhou, China
Background: The prognosis of patients with HER2 positive metastatic breast cancer Background: The ability to interrogate changes within the tumor microenvironment before,
(MBC) has dramatically improved with the advent of HER2-targeted therapy. However, during and following therapeutic intervention could yield important understanding of
resistance to anti-HER2 therapies remains inevitable. Aberrations in the PI3K-AKT-mTOR treatment response and causes for disease [Link] we conducted a multicenter
pathway are recognized as a key mechanism of resistance to HER2 directed therapies. phase II clinical trial (NCT04521179) examining the effect of a novel CTLA-4/PD-L1 bis-
This study is a multicenter, prospective, single-arm, phase II study to evaluate the an- pecific (KN046) antibody in combination with a a novel anti-HER2 bispecific (KN026)
titumor activity and safety of trastuzumab-pkrb plus gedatolisib in patients with HER2 antibody in treatment resistant metastatic breast cancer. Our on-going trial demonstrated
positive MBC who progressed after 2 or more HER2 directed chemotherapy. that in advanced HER2-positive breast cancer (HER2+ BC) patients, who have progressed
Methods: The primary endpoint was the overall response rate (ORR), assumed to be 25% after prior anti-HER2 combinational therapies, the objective response rate (ORR) of this
with a type I error rate of 0.05 and a power of 0.9. Although the target enrollment was 62 chemo-free therapy of KN026 in combination of KN046 was about 47.2% ( 95% CI: 30.4-
patients, the study was prematurely terminated after 44 patients due to the drug supply 64.5). To explore the underlying mechanism of this regimen, we collected tumor specimens
issue of gedatolisib. Patients with HER2-positive MBC and PI3K pathway genomic ab- from patients before and after receiving this combinational treatment for emerging multi-
errations, identified via tumor-targeted sequencing or cfDNA analysis, were enrolled after modal molecular analyses ("Multi-omics") to provide an in-depth description of the tumor
disease progression on at least two HER2-directed therapies. The treatment regimen immune microenvironment and its correlation with treatment response. Methods: We
included trastuzumab-pkrb and gedatolisib. Safety and efficacy outcomes were evalu- performed matched pre- and on-treatment investigative biopsies on index tumors and
performed single-cell RNA sequencing (scRNA-seq) and single-cell T cell receptor se-
ated, with a data cutoff of December 31, 2024. Results: Primary efficacy and safety data
quencing (scTCR-seq) analysis. Results: We performed comprehensive scRNA-seq on
were evaluable in 44 patients. The median age was 59 years (range: 28-72), and the
tumor biopsies obtained from a total of 17 patients that evaluable for overall response.
median number of prior palliative treatment lines was 4 (range: 2-10). Genomic aber-
Among them, 13 patients had two biopsies taken, and four patients had one biopsy collected
rations included mutations kinase domain (26 patients), helical domain (11), amplification
either before or during treatment. Single-cell RNA and T cell receptor sequencing from
(1) of PIK3CA, deletion of PTEN (2), and other mutations (4). Among the 44 evaluable 334,183 cells from site-matched tumors reveal significant temporal shift of various immune
patients, the best overall responses were complete response (CR) in 2 patients (4.5%), cell populations and phenotypes within the tumor microenvironment associated with
partial response (PR) in 17 (38.6%), stable disease (SD) in 19 (43.2%), progressive disease treatment responses. In-depth analysis of subpopulations revealed that CD8+ T cells are
(PD) in 5 (11.4%), and non-evaluable (NE) in 1 (2.3%), resulting in an objective response activated in responsive patients during treatment, and TREG cells, one of CD4+ T cell
rate (ORR) of 43.2% and a disease control rate (DCR) of 86.4%. The median progression- subtypes, are activated in non-responsive patients after therapy. Moreover, we also found
free survival (mPFS) was 5.8 months. After a median follow-up of 32.5 months, 22 deaths that combined-therapy activates cDCs and induces an inflammation shift in Mfs of re-
were recorded, and 19 patients were alive. The median overall survival (mOS) after study sponsive patients. Conclusions: We identified that regulatory T cells are activated while
enrollment was 18.4 months. Common treatment-related adverse events (TRAEs) in- effector T cells, natural killer cells, and dendritic cells were significantly depleted in non-
cluded oral mucositis (32.3%; 4.1% $ grade 3) and skin reactions (14.1%; 1.8% $ grade responding tumors. The immune response in responsive patients was effectively enhanced,
3). Hyperglycemia was reported in 6.8% (0.5% $ grade 3). No fatal adverse event related whereas in nonresponsive patients, it was significantly diminished. And higher baseline
to trial medications were reported. Conclusions: In this phase II study, the combination of levels of Mfs were associated with therapeutic resistance. These results support that
trastuzumab-pkrb and gedatolisib demonstrated a 43.2% response rate with manageable longitudinal analysis of tumor microenvironment to generate multi-omics data that can lead
toxicity in patients with HER2 positive MBC and PIK3CA mutations. A translational to rich insight disease process and to provide clinical value in evaluating treatment re-
research study focused on the analysis of cfDNA and PBMC is currently being planned. sponses. Clinical trial information: NCT04521179. Research Sponsor: Guangdong Science
Clinical trial information: NCT03698383. Research Sponsor: Korea Health Industry De- and Technology Department (2023B1212060013, 2022B1515020100, 2022A1515012238);
velopment Institute; HI17C2206. the Natural Science Foundation of China (82273033, 82072924, 82072906).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
42s BREAST CANCER—METASTATIC

1023 Poster Session 1024 Poster Session

Zongertinib in HER2-altered breast cancer: Preclinical activity and prelim- Transcriptomic analysis of HER2 expression in metastatic breast cancer:
inary results from a phase Ia dose-escalation study. First Author: David Berz, Insights from a UAE patient cohort. First Author: Rachel Su Jen Wong, Department
Valkyrie Clinical Trials, Inc., Los Angeles, CA of Haematology-Oncology, National University Cancer Institute, Singapore, Singapore
Background: Preclinical studies have demonstrated that zongertinib, an irreversible TKI, Background: HER2-low metastatic breast cancer (mBC) has emerged as a clinically
selectively and potently inhibits oncogenic HER2 in a variety of cancer models. An significant subgroup since approval of trastuzumab deruxtecan. However, its relevance as
ongoing Phase (Ph) Ia/Ib dose escalation/expansion trial (NCT04886804) has demon- a distinct subtype remains debated. We aimed to identify HER2-related gene signatures,
strated preliminary clinical activity of zongertinib across a range of HER2-driven solid examine the molecular characteristics of HER2-low mBC, and investigate the potential for
tumors. Here we present comparative preclinical data for zongertinib in breast cancer classifying HER2-low mBC into subgroups with profiles resembling either HER2-positive
(BC) cell lines and cell line-derived xenograft (CDX) models, as well as clinical data from (HER2+) or HER2-negative (HER2-) mBC. Methods: We performed differential gene ex-
patients (pts) with HER2-driven BC who received zongertinib during Phase Ia dose es- pression (DGE) analysis using the TCGA-BRCA dataset, comparing HER2+ and HER2-
calation. Methods: Cell proliferation assays were undertaken in HER2-amplified BC cell samples to identify a 17-gene HER2 expression signature. The findings were validated on
lines (relative copy numbers ranging from 3.2–10.1) exposed to serial dilutions of archival samples from a UAE clinical center, where DGE analysis was repeated to compare
zongertinib and tucatinib. Antitumor activity of zongertinib (5–40 mg/kg QD) in vivo was HER2+ and HER2-zero mBC. HER2-low samples were then classified into two subsets;
assessed in three HER2-amplified BC CDX models. Ph Ia of the trial enrolled pts with positive-like and zero-like based on their HER2 signature gene expression level. Lastly,
confirmed HER2 alterations (mutations, amplification or overexpression) who had Gene Set Enrichment Analysis (GSEA) was conducted to evaluate the molecular char-
exhausted all other treatment (Tx) options. In Ph Ia, zongertinib was administered at acteristics of these subsets. Results: In the TCGA-BRCA dataset (n = 409; median age 59
15–150 mg BID or 60–360 mg QD in 21-day cycles. Primary endpoints were maximum years, range 26–90), 20.5% of samples were HER2+, and 79.5% were HER2-. DGE analysis
tolerated dose (MTD) and dose-limiting toxicities (DLTs). Efficacy (objective response, identified a 17-gene HER2 signature that effectively separated HER2+ and HER2- mBC.
OR) was evaluated as a secondary endpoint using RECIST v1.1. Results: Zongertinib These 17 genes including GSDMB, ERBB2, and MED1, are associated with HER2 ex-
inhibited tumor cell growth in vitro with greater potency (4.5–16.4-fold) than tucatinib pression. In the UAE cohort (n = 69; median age 52 years, range 24–84), comprising 7.25%
(zongertinib IC50: 2.6–40.6 nM; tucatinib IC50: 13.2–664.0 nM). In mice, zongertinib was HER2+, 60.87% HER2-low, and 31.88% HER2-zero mBC, 7 genes (GSDMB, GRB7, ERBB2,
well tolerated and led to a dose-dependent inhibition of BC tumor growth, with tumor STARD3, PGAP3, MIEN1, TCAP) from the 17-gene HER2 signature were similarly upre-
regressions at higher doses ($ 20 mg/kg QD). As of August 29, 2024, 121 pts had been gulated in the HER2+ samples. We observed an increasing trend in HER2 signature ex-
treated in the Ph Ia trial. Two DLTs occurred during the MTD evaluation period; the MTD pression across the groups, from HER2-zero to HER2-low and then HER2+ mBC showing
was not reached. Treatment-related adverse events (TRAEs; all/grade $3) occurred in the highest expression. This trend was supported by separation across HER2 status with
82.6%/12.4% of pts; the confirmed OR rate was 31.4% across all doses and tumor types. In gene-level expression, Gene Set Variation Analysis scores and Principal Component
total, 15 pts with Stage IV BC (HER2 overexpression/amplification: n = 10; HER2 mu- Analysis (PCA). Next, HER2-low mBC were classified into two distinct subgroups based on
tations: n = 4; both: n = 1) received zongertinib (100 mg BID: n = 1; 240–360 mg QD: n = distance-to-centroid approach with PCA. Based on distance to HER2+ and HER2-zero
14). Most were white (60.0%) and had an ECOG PS of 1 (66.7%). Mean (standard deviation, centroids, HER2-low mBC were classified into positive-like and zero-like subgroups re-
SD) age was 58.1 (7.4) years. Mean (SD) Tx duration was 4.0 (3.4) months. TRAEs (any/ spectively. Separation of the positive-like and zero-like samples were observed with DGE
grade $3) occurred in 93.3%/0.0% of pts. In pts with BC, the confirmed OR rate was 26.7% analysis and expression of the HER2 signature genes. GSEA of HER2-low positive-like
(4 partial responses). The confirmed disease control rate was 73.3%. Regardless of samples indicate increased activation of ERBB2 oncogenic pathway and suppression of
confirmation, the OR rate was 46.5%. At the time of data cut-off, 2 pts with responses gene sets involved in immune-mediated pathways compared to HER2-low zero-like mBC.
were still on treatment, and an additional 3 patients had a response lasting longer than 4 Conclusions: This study reveals the potential utility of transcriptomic HER2 signature to
months. Conclusions: Zongertinib potently inhibits HER2-driven BC growth in preclinical characterize HER2-related molecular features in mBC, revealing a gradient of HER2
models in vitro and in vivo. Preliminary Ph Ia data indicate that zongertinib has en- signature expression across HER2+, HER2-low, and HER2-zero. Identification of sup-
couraging clinical activity and manageable safety in pts with advanced, HER2-driven BC. pressed immune-mediated pathways in HER2-low positive-like mBC suggests combination
Clinical trial information: NCT04886804. Research Sponsor: Boehringer Ingelheim. with immune mediators as potential treatment strategy. Research Sponsor: None.

1025 Poster Session 1026 Poster Session

Phase I summary of the C406 (CART) efficacy and safety for an HER- Decoding HER2 dynamics: Exploring HER2 expression across a real-world
2–positive breast cancer population. First Author: Meili Sun, Department of breast cancer cohort. First Author: Michelle Green, Labcorp, Durham, NC
Oncology, Jinan Central Hospital, Shandong University, Jinan, China Background: HER2 expression in breast cancer (BC) has evolved from a binary scoring system to a
Background: C406 is a chimeric antigen receptor (CAR) modified autologous T cell, continuum-based approach, driven by the therapeutic benefits of trastuzumab deruxtecan (T-DXd)
which is a CART targeting HER-2. Here we present the breast cancer results in Phase 1 in patients with HER2-Low tumors. Our understanding of HER2 expression is complicated by
study. Methods: This phase 1 trial to evaluate the safety and efficacy of chimeric several factors, including tumor heterogeneity, altered expression as tumors progress, and var-
iations in assay technique and scoring. Here we describe HER2 expression patterns in a large
antigen receptor (CAR) modified autologous T cells (C406) for pts with Her2-positive cohort of patients with BC, including a subset with longitudinal results. Methods: HER2 expression
recurrent or refractory breast cancer. C406 is administered intravenously at a fixed dose. levels were determined by IHC using the VENTANA 4B5 antibody. HER2 copy number and HER2/
Response was evaluated by RECIST v.1.1 every 4 weeks. Results: As of Jan 18, 2025, 8 CEP17 ratios were determined using a validated dual-probe ISH assay. Tumors were retrospectively
(F) breast cancer pts have received C406 at doses of 3*107/kg (n = 6), 1*108/kg (n = 2). identified as HER2-Zero (IHC 0), HER2-Low (IHC 1+ or 2+ with negative ISH) or HER2-positive
Two patients in the 3*107/kg -dose group received the second transfusion. Overall, the (IHC3+ or IHC2+ with positive ISH) based on the established IHC and ISH values scored at the time
median age was 59 years, the median number of priortreatment lines was 3.125 and the of reporting (Apr 2013 to Nov 2024). Estrogen and progesterone receptor (HR) levels were de-
median number of anti-HER-2 treatment lines was 2. The histological type was invasive termined using validated IHC assays. All testing was performed in a CAP/CLIA accredited referral
breast cainoma, 7 cases of ductal carcinoma and 1 case of mucous carcinoma. These laboratory. Tumor histology, specimen collection site (CS), and patient demographics were ab-
breast cancer hormone receptor types include ER+/PR+, ER-/PR-, and ER+/PR-. Cy- stracted from test requisition forms. Results are descriptive and presented in aggregate. Re-
sults: A total of 30,023 BC specimens from 27,055 patients were included in the analysis. Most
clophosphamide combined with fludarabine was the regimen for lymphocyte clearance
patients were female (99%; 26,687) with a median age at testing of 63.5 years. Among HER2-Low
in all patients. Among them, 1 patient did not receive bridging therapy. Bridging specimens, 87% were HR-positive (HR+) and 13% were HR-negative/HR-Low (ER IHC , 10%).
treatment options for other patients were as follows, attillizumab bridging therapy (n = Longitudinal specimens were available for 1,267 patients with a median of 77 days between
1), albumin-bound paclitaxel combined with carboplatin and attillizumab (n = 1), al- collection dates. 69% (869) of longitudinal specimens demonstrated the same level of HER2
bumin-bound paclitaxel combined with Attillizumab (n = 2), gemcitabine combined with expression, including in 69% (597/868) of cases where both specimens were collected from breast
attillizumab (n = 1), Albumin-bound paclitaxel + epirubicin + cyclophosphamide + and 69% (203/294) of cases where the first specimen was collected from breast and the second
Attilizumab + local radiotherapy (n = 1), docetaxel + Attilizumab + local radiotherapy (n = was collected from another site. Among 398 patients who had HER2 expression levels changes,
1). Among the 8 pts having imaging tumor assessment, the DCR was 75%, which in- 46% went from HER2-Low to HER2-Zero and 33% went from HER2-Zero to HER2-Low. Conclu-
cludes 0 partial responses (PR), 6 stable disease (SD), and 2 progressive disease (PD) sions: HER2-Low status is common in BC, especially among HR-positive cases.HER2 expression
according to RECIST v1.1. Among the 6 SD pts, 1 pt’s progression-free survival (PFS) is 8 status changed between longitudinal specimens in 31% of patients, which may inform therapeutic
eligibility. This reaffirms current guidelines recommending a biopsy at first distant recurrence and
months. All the 8 pts experienced a treatment related adverse events (TRAEs). The most
suggests that serial biopsies might be helpful in detecting HER2 expression in an originally HER2-
common TRAEs included white blood cell count decrease (100%, 8/8), neutrophil count Zero tumor. Research Sponsor: None.
decrease (100%, 8/8), lymphocyte count decrease (100%, 8/8) and cytokine release
Longitudinal HER2 expression levels changes.
syndrome (25%, 2/8). No TRAE leading to discontinuation and death. Conclusions: C406
HER2 Expression Level Change Both CSs Breast Breast to Other CS Other CS combinations Total
therapy showed an acceptable safety profile anti-tumor activity for advanced HER-2
positive breast cancer, but the antitumor activity needs to be further explored. Clinical Positive to Low 27 3 3 33
Positive to Zero 5 2 1 8
trial information: ChiCTR2500096093. Research Sponsor: None. Low to Positive 23 6 3 32
Low to Zero 116 51 17 184
Zero to Positive 6 3 9
Zero to Low 94 26 12 132
No Change (Positive) 71 32 7 110
No Change (Low) 365 104 41 510
No Change (Zero) 161 67 21 249

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 43s

1027 Poster Session 1028 Poster Session

A phase Ib/IIa study of BAT8010+BAT1006, an anti-HER2 monoclonal JSKN003, a biparatopic HER2-targeting ADC, in heavily pretreated HER2-
antibody-exatecan conjugate combined with an ADCC-enhanced HER2 positive breast cancer: A pooled analysis of early-phase studies. First Author:
mAb in patients with advanced solid tumors. First Author: Shusen Wang, Sun Yiqun Du, Department of Medical Oncology, Fudan University Shanghai Cancer Center,
Yat-sen University Cancer Center, Guangzhou, China Shanghai, China
Background: BAT8010 is an ADC argeting HER2, while BAT1006 is a humanized Background: JSKN003 is a biparatopic HER2-targeting antibody-drug conjugate (ADC)
monoclonal antibody targeting another epitope of HER2, with ADCC enhancement activity conjugated to a topoisomerase I inhibitor (TOP1i) via a tetrapeptide linker, designed to
via completely devoid of fucose. This study investigates the combination of BAT8010 and enhance serum stability and anti-tumor activity. The efficacy and safety of JSKN003 in
BAT1006 in patients with advanced solid tumors. Methods: Patients in this open-label, advanced ovarian cancer and other solid tumors have been highlighted in previous
multicenter clinical trial received BAT8010 +BAT1006 on day 1 of a 21-day cycle until reports, and this analysis provides updated insights into its performance in HER2-
intolerable or disease progression occurred. The study objectives included assessing positive breast cancer. Methods: JSKN003-101 is a dose-escalation and -expansion
tolerability, safety, pharmacokinetic characteristics, immunogenicity, and preliminary study in Australia, and JSKN003-102 is a phase I/II study in China, both involving
efficacy. Results: As of January 15, 2025, 20 patients with metastatic breast cancer (mBC, patients with advanced solid tumors. A pooled analysis was performed to assess its
n=14) and gastric cancer (GC, n=6) were enrolled in three cohorts: BAT8010 (2.1mg/kg) + efficacy and safety in HER2-positive advanced breast cancer. Results: As of November
BAT1006, BAT8010 (2.4mg/kg）+ BAT1006 and BAT8010 (2.7mg/kg) + BAT1006, with 29, 2024, the median follow-up duration was 3.52 months (range: 2.99-3.71). A total of
BAT1006 fixed at 15 mg/kg. HER2 positivity on tumor tissue was categorized as IHC2+/ 71 patients with HER2-positive breast cancer were enrolled, with the majority receiving
FISH+ or IHC3+. Two dose-limiting toxicity (grade 4 thrombocytopenia and neutropenia) 6.3 mg/kg or 8.4 mg/kg doses. The median age was 54 years (range: 32-79), with 78.9%
were reported in the BAT8010 (2.7mg/kg) + BAT1006 group. The maximum tolerated dose ECOG 1. All patients had stage IV disease, with 76.1% having visceral metastases. All
was determined to be BAT8010 (2.4mg/kg）+ BAT1006, and expansion studies have been patients had prior anti-HER2 therapy, including 87.3% with prior ADCs or TKIs, and 56.3%
proceeded at this dose. Safety: Among the 20 patients who received at least one dose of having $3 prior lines. Among 62 evaluable patients, 56 were T-DXd naı̈ve. In these 56
BAT8010 + BAT1006, 17/20 (85%) reported at least one treatment-emergent adverse patients, the overall response rate (ORR) was 67.9% (95%CI: 54-79.7), and the disease
events (TEAEs). The most common TEAEs ($5%) included neutropenia, leukopenia, control rate (DCR) was 94.6% (95%CI: 85.1-98.9). In the RP2D subgroup (6.3mg/kg, n =
nausea, thrombocytopenia and anemia. Most TEAEs were Grade 1 or 2; however, 55% of 30), the ORR was 70.0% (95%CI:50.6-85.3). Of 6 patients with prior T-DXd exposure, 1
the patients experienced Grade 3 or greater AEs, including neutropenia 7/20 (35%),
achieved a partial response (PR), 3 had stable disease (SD), and tumor shrinkage was
thrombocytopenia 5/20 (20%), infusion-related reaction 1/20 (5%) and febrile neu-
observed in 3. Both median progression-free survival (PFS) and median overall survival
tropenia1/20 (5%). The infusion-related reaction was mild, and one subject discontinued
(OS) were immature. Treatment-related adverse events (TRAEs) $Grade 3 occurred in
the study treatment due to TEAEs. No cases of interstitial lung disease (ILD)/pneumonitis
11.3% of patients, and serious adverse events (SAEs) in 9.9%, with 2 drug-related. No
were reported. Efficacy: Fourteen mBC patients were recruited across the dose cohorts:
BAT8010 (2.1 mg/kg)+BAT1006 (n=3), BAT8010 (2.4 mg/kg)+BAT1006 (n=7) and BAT8010
TRAEs led to death or treatment discontinuation. The most common TRAEs ($20%)
(2.7 mg/kg)+BAT1006 (n=4). Most had previously undergone 3-7 lines of systemic included nausea, elevated liver enzymes, vomiting, decreased appetite, fatigue, diarrhea,
treatments, including trastuzumab and HER2 ADC regimens. Six GC patients were included and anemia. No$Grade 3 neutropenia was observed. Grade $3 anemia and decreased
in the BAT8010 2.4mg/kg + BAT1006 (n=4) and BAT8010 2.7mg/kg + BAT1006 (n=2) platelet count were each reported in 1 patient (1.4%), both being Grade 3. Interstitial lung
cohort. Sixteen patients had at least one tumor assessment, yielding an ORR of 43.7% (7/ disease (ILD) occurred in three patients (4.2%), all Grade 1-2, with no Grade $3 events.
16) and a DCR of 87.5% (14/16). Among the 12 mBC patients, the ORR is 50% (6/12) with a Conclusions: JSKN003 demonstrated promising efficacy and manageable safety
DCR of 91.66% (11/12), including one CR. In the 4 GC patients, the ORR was 25% (1/4) with in heavily pretreated HER2-positive breast cancer, including T-DXd-experienced pa-
a DCR of 75% (3/4). Conclusions: The combination of BAT8010 and BAT1006 was well- tients. The biparatopic HER2 antibody design likely enhanced its binding efficiency
tolerated, with manageable toxicity, and demonstrated promising preliminary antitumor and contributed to the observed clinical benefit. These findings support the
activity in metastatic breast cancer and gastric cancer. Dose expansion studies are planned Phase 3 trial to further evaluate its therapeutic potential. Clinical trial infor-
ongoing to further detect the safety and efficacy in these population. Clinical trial in- mation: NCT05494918; NCT05744427. Research Sponsor: Jiangsu Alphamab Bio-
formation: CTR20241120. Research Sponsor: None. pharmaceuticals Co., Ltd.

1029 Poster Session 1030 Poster Session

IBI354 (anti-HER2 antibody-drug conjugate [ADC]) in patients (pts) with Efficacy of tucatinib, trastuzumab, and capecitabine (TTC) following tras-
HER2-positive breast cancer (BC) and other solid tumors: Updates from a tuzumab-deruxtecan (T-DXd) in HER2-positive metastatic breast cancer
phase 1 study. First Author: Charlotte Rose Lemech, Scientia Clinical Research, (MBC): Updated results and subgroup analyses from the UNICANCER mul-
Randwick, Australia ticenter retrospective cohort. First Author: Jean-Sebastien Frenel, Institut de
Background: HER2 has been established as an important therapeutic target for BC. Cancerologie de L’Ouest, Saint-Herblain, France
IBI354 consists of trastuzumab (anti-HER2 antibody) conjugated to a camptothecin Background: T-DXd is the standard second-line treatment for HER2-positive MBC with
derivative. In a global, multicenter, phase 1 study, IBI354 was well tolerated and showed TTC being the preferred third-line option. However, the efficacy of TTC following T-DXd
promising efficacy in BC and other solid tumors (2024 ESMO 345MO/720MO/576P). remains unclear. Here, we provide an updated and subgroup analysis of a French cohort
Here, we report updated safety and efficacy of IBI354. Methods: Eligible pts with comprising 101 patients who received TTC after T-DXd. Methods: We conducted a
advanced solid tumors who had failed or were intolerant to standard treatment were retrospective study across 12 French comprehensive cancer centers, including patients
enrolled. Positive HER2 was defined as immunohistochemistry (IHC) 2+/in situ hy- with HER2-positive MBC treated with TTC after T-DXd exposure. The primary endpoint
bridization (ISH)+ or IHC 3+. IBI354 was administered intravenously at 6-15 mg/kg Q3W was progression-free survival (PFS), while secondary endpoints included overall survival
or Q2W. Primary endpoint was safety. Secondary endpoints were objective response rate (OS) and time to next treatment (TTNT). Results: A total of 101 patients who initiated TTC
(ORR), disease control rate (DCR), duration of response (DoR) and progress-free survival between August 2020 and December 2022 were included in the analysis. The median age
(PFS) assessed by investigators per RECIST v1.1 and overall survival (OS). Results: As was 56.4 years (range: 30.8–84.8). Patients had received a median of 4 prior MBC
of Nov 12, 2024, a total of 368 pts with solid tumors were enrolled in China and Australia therapies (range: 2–15), which included pertuzumab (81%) and T-DM1 (93%). 82 patients
(females: 89.4%, median age: 56.0 years [range: 27-82], ECOG PS 1: 75.0%). Median (81%) experienced progression on T-DXd, while 19 discontinued due to toxicity or other
follow-up time was 11.3 months (range: 5-19). Median treatment duration was 25.0 reasons. The data cutoff date was December 1, 2024. For the whole population, with a
weeks (range: 3.1-63.3) and 124 (33.7%) pts remain on treatment. Treatment-related median follow-up of 29.6 months (95% CI [26.0–34.0]), the median PFS was 4.7 months
adverse events (TRAEs) occurred in 331 (89.9%) pts while $grade 3 (G3) TRAEs oc- (95% CI [3.9–5.8]), the median TTNT was 5.2 months (95% CI [4.5–6.6]), and the median
curred in 93 (25.3%) pts. Most common TRAEs included white blood cell count de- OS was 13.9 months (95% CI [12.4–19.0]). For the 86 patients who initiated TTC im-
creased (48.6%, with 7.1% $G3), anemia (46.7%, with 4.9% $G3), nausea (46.2%, with mediately after T-DXd, the median PFS and TTNT were 5.2 months (95% CI [4.4–6.4]) and
0.8% $G3) and neutrophil count decreased (38.3%, with 9.8% $G3). Interstitial lung 5.5 months (95% CI [4.7–7.2]), respectively. HR+ disease was identified in 71.3% (n=72) of
disease occurred in 8 (2.2%) pts (5 treatment-related and 3 treatment-unrelated, all G1- the cohort, with 84.7% receiving TTC immediately post-T-DXd. With a median follow-up of
2). TRAEs led to dose reduction in 5 (1.4%) pts and treatment discontinuation in 4 (1.1%) 29.6 months (95% CI [25.1–NR]), the HR+ population had a median PFS of 4.1 months
pts. No TRAE led to death. Efficacy was evaluable in 88 pts with HER2-positive BC (stage (95% CI [3.5–5.6]) and a median OS of 13.4 months (95% CI [12.3–19.0]). The median
IV: 97.7%; prior systemic therapy regimens$5: 65.9%; IHC 2+/ISH+: 19.3%, IHC 3+: TTNT was 4.7 months (95% CI [4.0–6.3]). Among the 65 RECIST-evaluable HR+ patients,
80.7%). The overall confirmed ORR was 58.0% (95% CI: 47.0-68.4) and DCR was 90.9% best response included progressive disease in 40%, stable disease in 29%, partial re-
(95% CI: 82.9-96.0). Among 51 pts with confirmed responses, median DoR was not sponse in 29%, and complete response in 2%. With a median follow-up of 29.3 months
reached (events rate: 19.6%) and 12-month DoR rate of 71.8% (95% CI: 52.9-84.2). (95% CI [26.0–NR]), the HR- population had a median PFS of 5.8 months (95% CI
Median PFS was not reached with events rate of 37.5%. Median OS was not reached with [4.4–10.5]) and a median OS of 17.5 months (95% CI [10.6–22.9]). The median TTNT was
events rate of 5.7% and 9-month OS rate of 96.2% (95% CI: 88.7-98.8). Conclu- 6.0 months (95% CI [4.9–10.7]). Among the 24 RECIST-evaluable HR- patients, best
sions: IBI354 continues to demonstrate favorable safety profiles with no new safety response included progressive disease in 25%, stable disease in 38%, partial response in
signals. Encouraging efficacy was observed in HER2-positive BC. Clinical trial infor- 33%, and complete response in 4%. Conclusions: This large retrospective cohort with
mation: NCT05636215. Research Sponsor: Innovent Biologics (Suzhou) Co., Ltd. extended follow-up highlights the efficacy of TTC in HER2-positive MBC patients pre-
viously treated with T-DXd. These findings support the role of TTC as a viable treatment
option post-T-DXd and provide insights for optimizing therapeutic strategies in this
setting. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
44s BREAST CANCER—METASTATIC

1031 Poster Session 1032 Poster Session

Eribulin plus pyrotinib In trastuzumab-resistant HER2-positive advanced Quantitative pre-treatment assessment of trastuzumab deruxtecan (T-DXd)
breast cancer: A single-arm, multicenter phase II trial (EPIC trial). First Author: antibody target (HER2) and payload target (topoisomerase 1, Topo1) to
Zhenchuan Song, Fourth Hospital of Hebei Medical University, Shijiazhuang, China, predict outcomes in metastatic breast cancer (MBC). First Author: Paolo
Shijiazhuang, China Tarantino, Dana-Farber Cancer Institute, Boston, MA
Background: This study (ChiCTR2000038832) aimed to evaluate the efficacy and safety Background: The antitumor activity of T-DXd for MBC is sub-optimally predicted by HER2 immunohisto-
chemistry (IHC). We evaluated novel assays to quantify the expression of T-DXd antibody and payload targets
of combining eribulin with pyrotinib in patients with advanced HER2-positive breast
and their association with outcomes. Methods: We retrieved pre-treatment FFPE tumor samples for patients
cancer who had developed resistance to trastuzumab. These patients typically face a (pts) with MBC receiving T-DXd at Dana-Farber Cancer Institute between 2017 and 2023. Patients were
poor clinical prognosis, and evidence-based guidance for treatment decisions remains categorized by HER2 IHC status at start of T-DXd. The RPPA-based protein assessment of HER2 and Topo1
limited. Methods: Eligible patients were those with pathologically confirmed HER2- was performed in a CLIA lab after laser capture microdissection enrichment of tumor epithelium. The HER2DX
positive metastatic breast cancer, an Eastern Cooperative Oncology Group (ECOG) standardized assay was performed after RNA extraction. We evaluated the association of each marker,
continuously and by tertiles/quartiles, with time to next treatment (TTNT) with T-DXd. Cox proportional hazards
performance status of 0 or 1, and prior treatment with trastuzumab and taxanes. models were utilized to estimate hazard ratios and log-rank test p-values were reported. The Kaplan-Meier
Participants received oral pyrotinib (400 mg once daily) and intravenous eribulin (1.4 method was used to calculate median estimates. Results: HER2DX and RPPA testing were conducted for 41
mg/m² on days 1 and 8 of each 21-day cycle) for up to six cycles. Pyrotinib was (25 with HER2+, 16 with HER2- MBC) and 38 pts (24 with HER2+, 14 with HER2- MBC), [Link]
continued until disease progression or intolerable toxicity. The primary endpoint was HER2DX and RPPA HER2 quantitative testing significantly predicted outcomes with T-DXd (Table). The
progression-free survival (PFS). The secondary endpoints include the objective response HER2DX HER2 amplicon mRNA signature was significantly associated with TTNT with T-DXd (p=0.001),
including when divided into tertiles, with a range of 4.7 months in the lowest vs 12.03 months in the highest
rate (ORR), disease control rate (DCR), duration of response (DoR), overall survival (OS), tertile (p=0.02). Similarly, the RPPA-based HER2 protein expression was significantly associated with TTNT
and safety. Results: Between February 2021 and September 2023, 30 patients were when divided into quartiles (p=0.02). Pre-treatment Topo1 protein expression was significantly associated with
enrolled, with a median age of 57 years (range: 30–76). All had received prior tras- outcomes in pts with HER2-negative MBC (n=14), with higher expression of TOPO1 associated with worse TTNT
tuzumab and taxane therapies. As of November 30, 2024, the median follow-up was 20.1 with T-DXd (p=0.04). Conclusions: Higher pre-treatment HER2 mRNA signature (HER2DX) and protein (RPPA)
expression predicted improved outcomes with T-DXd for MBC, whereas higher Topo1 expression was as-
months. Disease progression or death occurred in 15 patients, and the median PFS was
sociated with worse outcomes with T-DXd among pts with HER2- MBC. Research Sponsor: Terri Brodeur
13.73 months (95% confidence interval [CI]: 11.1–14.8). The 12-month PFS rate was Breast Cancer Foundation; Saverin Award; Susan G. Komen Breast Cancer Foundation; Breast Cancer Research
61.7% (95% CI: 44.2%–86.0%). The 12-month OS rate was 75.3% (95% CI: 66.2%–84.4%). Foundation.
The objective response rate (ORR) was 53.3% (16/30), and the disease control rate (DCR) Association of pre-treatment HER2 amplicon mRNA signature, HER2 RPPA and Topo1 RPPA expression
was 80.0% (24/30). Median overall survival was not reached. Common adverse events with outcomes among patients receiving T-DXd.
(AEs) of any grade occurring in . 15% of patients included diarrhea (40%), nausea (20%), Median TTNT
anorexia (16%), and vomiting (16%), with grade 3 diarrhea reported in 3% of patients. No Group (months) 95% CI HR 95% CI p-value
treatment-related deaths were observed. Conclusions: The combination of eribulin and HER2 amplicon 1 unit increase - - 0.70 0.56-0.87 0.001
mRNA expression
pyrotinib shows promise as a therapeutic option for patients with HER2-positive ad- n=41
vanced breast cancer resistant to trastuzumab. While advancements in anti-HER2 HER2 amplicon Low (ref) 4.70 3.27-NA 0.71 0.33-1.55 Log-Rank
therapies continue, further studies are needed to address unresolved challenges in this tertiles Med 5.33 4.7-NA 0.23 0.08-0.69 p=0.019
n=41 High 12.03 7.37-NA
clinical context. Clinical trial information: ChiCTR2000038832. Research Sponsor: None. HER2 RPPA protein 10 unit increase - - 0.95 0.90-1.01 0.083
expression
n=38
HER2 RPPA # 25% (ref) 4.03 2.87-NA 0.64 0.25-1.61 Log-Rank
quartiles .25%-50% 5.83 2.13-NA 0.28 0.10-0.74 p=0.019
.50%-75% 8.00 5.33-NA 0.30 0.12-0.75
n=38 .75% 9.07 5.83-NA
Topo1 RPPA , median (ref) 5.87 4.90-NA 3.49 1.02-11.96 Log-Rank
expression $ median 2.70 2.47-NA p=0.036
(HER2- only)
n=14

1033 Poster Session 1034 Poster Session

Efficacy and safety results of TQB2930, a HER2-targeted bispecific antibody Association of germline homologous recombination deficiency mutations
combined with chemotherapy in patients with HER2-positive breast cancer with HER2 status conversion from negative to positive following neoadju-
(BC) previously treated with ‡2 line treatments: Results from a phase 1b/2 vant chemotherapy in breast cancer. First Author: Colin P. Bergstrom, Stanford
study. First Author: Qingyuan Zhang, Breast Department, Harbin Medical University University School of Medicine, Stanford, CA
Cancer Hospital, Harbin, China Background: Neoadjuvant therapy (NT) is well established in breast cancer management,
Background: TQB2930 is a HER2-targeted bispecific antibody designed to bind two with subsequent adjuvant therapy guided by biomarker status and tumor response. While
distinct HER2 epitopes: the extracellular domain 4 (ECD4), and the extracellular domain 2 rare, biomarker status can change post-NT. The mechanisms driving conversion and
(ECD2). In an ongoing phase 1b/2 clinical trial, TQB2930 has demonstrated favorable optimal treatment strategies remain unclear. This study aimed to investigate clinico-
tolerability alongside durable responses in patients with heavily pretreated metastatic pathological characteristics, including germline mutations, in patients who underwent
HER2-positive BC. In this context, Cohort 4 of the trial was designed to evaluate the safety HER2 status conversion from negative to positive (N-P) after NT. Methods: Patients
and efficacy of TQB2930 in combination with chemotherapy for patients with HER2- treated with NT for breast cancer between 2012 and 2023 were retrospectively identified
positive BC who had received at least two prior lines of treatment. Methods: Cohort 4 from the Stanford STRIDE database. Clinicopathological features, including demo-
enrolled patients aged $18 years with recurrent or metastatic HER2-positive BC who had graphics, genetic data, treatment history, pathological characteristics, and recurrence,
undergone at least two prior systemic therapies. Patients with stable brain metastases were collected through chart review. ER, PR, and HER2 status were assessed using
were permitted. All enrolled patients received TQB2930 intravenously at a dose of 30 mg/ immunohistochemistry (IHC) and fluorescence in-situ hybridization (FISH), with HER2
kg every three weeks (Q3W), administered in combination with one of four investigator- interpretation following ASCO/CAP guidelines relevant at the time of testing. Per in-
selected chemotherapies (capecitabine, eribulin, gemcitabine, or vinorelbine). The pri- stitutional policy, both IHC and FISH were routinely performed for HER2 assessment.
mary endpoint was the overall response rate (ORR), while secondary endpoints Results: A total of 28 patients with HER2 status change were identified; 26 (93%) were
encompassed disease control rate (DCR), progression-free survival (PFS), overall survival female and 2 (7%) male, with a median diagnosis age of 47 years (IQR 39–58). Most
(OS), safety, and immunogenicity. Results: As of December 15, 2024, 55 patients (pts) patients were ER-positive (N=24, 86%) at diagnosis. Among 25 patients who underwent
had been treated with TQB2930 combined with chemotherapy. The median follow-up germline testing, 56% harbored mutations in homologous recombination deficiency
duration was 4.14 months (95% CI: 3.55–4.31). Out of 52 patients evaluable for efficacy, genes, including BRCA2 (5, 20%), BRIP1 (4, 16%), PALB2 (2, 8%), ATM (2, 8%), and BRCA1
the ORR was 48.1% (25/52), with 88.5% (46/52) experiencing a reduction in target lesion (1, 4%). Post-NT, patients were classified into HER2 groups: 36% (10/28) in Group 1, 46%
size. Among patients who had progressed on prior T-DM1 therapy, the ORR was 36.8% (7/ (13/28) in Group 1b (HER2 low amplified with HER2/cell ,6 and ratio .2), 14% (4/28) in
19), whereas patients who had failed other HER2-ADC therapies exhibited an ORR of 50% Group 3, and 4% (1/28) in Group 4. Adjuvant HER2-directed therapy was administered to
(8/16), including RC48, DS-8201, and so on. Notably, the median PFS and OS had not yet 24 patients (89%): trastuzumab in 8 (29%), trastuzumab and pertuzumab in 12 (43%), and
been reached, while the 6-month PFS rate was estimated at 71%. Grade $3 TRAEs were trastuzumab emtansine in 4 (14%); 4 patients (14%) did not receive adjuvant HER2
primarily hematological, encompassing decreased white blood cell count, neutropenia, therapy. Recurrence occurred in 8 patients (29%), including 3 with germline mutations.
thrombocytopenia, and infusion-related reactions. Importantly, there were no grade $3 One non-mutated case involved discordant HER2-positive recurrence with ipsilateral in-
cardiac toxicities, and the incidence of sinus bradycardia or QT interval prolongation was breast recurrence and liver metastases, testing HER2-positive by FISH despite IHC 0.
, 3%. Conclusions: The combination of TQB2930 with chemotherapy demonstrates Conclusions: This study highlights that a majority of patients who underwent HER2
encouraging antitumor efficacy and an acceptable safety profile in patients with HER2- status conversion after NT harbored homologous recombination deficiency mutations,
positive BC who have undergone at least two prior lines of treatment. These results suggesting a potential mechanistic basis. Furthermore, while 85.8% of patients received
support the potential of TQB2930 as a novel therapeutic strategy for HER2-positive BC adjuvant HER2-directed therapy, the rarity of HER2-positive recurrences underscores the
and underscore the need for further clinical exploration. This study was funded by Chia Tai potential for overtreatment. Further studies are needed to elucidate the mechanisms of
Tianqing Pharmaceutical Group Nanjing Shunxin Pharmaceutical Co., Ltd. Clinical trial HER2 status conversion from negative to positive after NAT which will improve the
information: NCT06202261. Research Sponsor: Chia Tai Tianqing Pharmaceutical Group efficacy of treatment strategies and patient outcomes. Research Sponsor: None.
Nanjing Shunxin Pharmaceutical Co., Ltd.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 45s

1035 Poster Session 1036 Rapid Oral Abstract Session

Recurrence risk prediction model in HER2-positive early breast cancer after Racial and ethnic disparities in clinical outcomes of HER2-positive meta-
HER2-targeted therapy. First Author: Qingyao Shang, National Cancer Center/Na- static breast cancer treated with antibody-drug conjugates: A TriNetX real-
tional Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical world evidence study. First Author: Zunairah Shah, Roswell Park Comprehensive
Sciences and Peking Union Medical College, Beijing, China Cancer Center, Buffalo, NY
Background: HER2-positive breast cancer patients treated with adjuvant targeted Background: HER2-positive breast cancer (HER2+ BC) is an aggressive BC subtype driven by
therapy, including trastuzumab or pertuzumab have demonstrated improved outcomes. overexpression of the human epidermal growth factor receptor 2 (HER2). Antibody drug
However, a part of patients still experience recurrence despite targeted therapy. This conjugates (ADCs), such as trastuzumab deruxtecan (T-DXd) and ado-trastuzumab emtansine
study aims to develop a time-dependent model to predict recurrence risk in HER2- (T-DM1), both approved for metastatic HER2+ BC, have significantly improved survival in this
positive early breast cancer patients following targeted therapy, utilizing data from the population. However, racial and ethnic disparities in outcomes with ADCs remain unclear.
Methods: De-identified data from TriNetX, a global federated health research network, were
APHINITY trial. Methods: The APHINITY trial included two arms: trastuzumab-only arm
analyzed for patients with metastatic HER2+ BC receiving HER2-directed ADCs. Kaplan-Meier
(n = 2,400) and trastuzumab + pertuzumab dual-target therapy arm (n = 2,404). Each analysis assessed overall survival. Statistical comparisons of survival rates between groups
group was randomly divided into training (70%) and validation (30%) cohorts, resulting in were made using log-rank tests. Propensity matching was used to adjust for age, comorbidities
3,363 patients in the training set and 1,441 patients in the validation set. The Cox and lines of treatment. Two-sided P #0.05 was used to determine statistical significance.
proportional hazards model and two machine learning models, Random Survival Forest Results: A total 7,462 patients were included in this analysis. The median age was 56.2 years
(RSF) and XGBoost (XGB), were used to predict invasive disease-free survival. Model (range: 43-69), 68% were Non-Hispanic White (NHW), 17% Black, 5% Asian, and 10% Hispanic.
performance was evaluated using Harrell`s C-index and area under the curve (AUC). 4,774 pts received TDM-1, 1,547 received T-DXd, and 1,141 received both. 37.6% patients treated
Results: After selecting clinical variables provided by the APHINITY trial, 12 variables with TDM-1 and 64.3% treated with T-DXd received .2 lines of prior therapy. Overall, the survival
were included in the model training. The predictive performance of Cox model, RSF and rate was 96% at 1 year, 89.3% at 3 years, and 83.7% at 5 years. At 3 years, the survival varied
XGB machine learning models was assessed. Among them, the RSF model demon- significantly by race: 90.3% in NHWs, 87.1% in Black, 92.9% in Asians, and 91.3% in Hispanics (p
strated the best predictive effectiveness. In the training set, the RSF model achieved a C- , 0.001). Similar differences extended to 5 years: NHW 85.0%, Black 80.9%, Asian 89.5%,
index of 0.66, with AUCs of 0.78 for 1-year recurrence risk, 0.70 for 3-year recurrence Hispanic 86.7% (p,0.001) (Table). Among patients receiving T-DM1, the 3-year survival rate
risk, and 0.66 for 5-year recurrence risk. In the validation set, the RSF model achieved a was 71.9%, compared to 45.7% in the T-DXd cohort (p,0.05). After adjusting for age and
comorbidities, there was no difference in survival between patients who received T-DXd and
C-index of 0.68, with AUCs of 0.79 for 1-year recurrence risk, 0.73 for 3-year recurrence
TDM1 (not including who received both) (HR 2.55 (95% CI: 2.20–2.96; p = 0.75). Conclu-
risk, and 0.71 for 5-year recurrence risk. The XGB model performed slightly worse than
sions: Our study shows better outcomes with use of ADCs in heavily pretreated metastatic
RSF, and the machine learning methods significantly outperformed the Cox model. HER2-BC, with 83.7% of patients surviving beyond 5 years. However, significant racial disparities
Conclusions: In this study, a time-dependent recurrence prediction model was were observed, with Asian patients showing the highest survival while Black patients had the
established based on large-sample randomized controlled trial, demonstrating a poorest survival which could be due to multi-level factors. Future studies are needed to un-
favourable short-term recurrence prediction effect, which can serve as a clinical decision derstand the underlying mechanisms behind this racial disparity in outcomes with HER2-di-
assistant for screening patients at high risk of recurrence for intensified adjuvant rected ADCs to inform strategies to improve patient outcomes. Research Sponsor: None.
therapy or follow-up monitoring. Research Sponsor: None.
Survival rates and hazard ratios by race.
3-year survival 5-year survival
Race probability (%) probability (%) Hazard Ratio
NHW 89.60 83.96 1.00
Black 86.90 80.30 1.33 (95% CI: 1.26–1.40)
Asian 92.26 88.92 0.69 (95% CI: 0.6–0.75)
Hispanic 91.20 86.27 0.89 (95% CI: 0.83–0.95)
NHW (Non-Hispanic White).

1037 Poster Session 1039 Poster Session

Characterization of the immune microenvironment and spatial phenotypes Risk of radiation necrosis with concurrent antibody-drug conjugates and
across HER2 subtypes in advanced or metastatic breast cancer. First Author: radiotherapy in HER2-positive breast cancer with brain metastases: A meta-
Ayse A. Koksoy, The University of Texas MD Anderson Cancer Center, Houston, TX analysis. First Author: Zouina Sarfraz, Miami Cancer Institute, Baptist Health South
Background: Breast cancer is defined by HER2 and hormone receptors (HR) status, Florida, Miami, FL
which influence the clinical outcomes. HER2-positive has been traditionally defined as Background: Antibody-drug conjugates (ADCs) have transformed the outcomes of HER2-pos-
HER2 overexpression on immunohistochemistry (IHC score of 3+) or 2+ and ERBB2 itive breast cancer (BC), particularly in patients with brain metastases (BM) due to the use of
amplification on in situ hybridization (ISH). HER2low (IHC 1+ or 2+ and non-amplified ISH) ADCs like trastuzumab emtansine (T-DM1) and trastuzumab deruxtecan (T-DXd), which have
accounts for nearly half of tumors. There is a paucity of data regarding immune sub- demonstrated intracranial efficacy. Radiotherapy (RT), especially stereotactic radiosurgery
populations and spatial phenotypes in HER2 subtypes. We have investigated the (SRS), remains a cornerstone for BM management. However, combining ADCs with RT may
increase the risk of symptomatic radiation necrosis (SRN). This meta-analysis evaluates SRN
characteristics of tumor immune microenvironment contexture across HER2 groups
outcomes in patients receiving concurrent (C-ADC) versus non-concurrent ADCs with RT (NC-
(HER2+ vs HER2low vs HER2- (0 by IHC)) focusing tumor infiltrating lymphocytes (TIL) and ADC). Methods: A systematic search was performed in January 2025 across PubMed, Cochrane,
on immune cell dynamics, including the distribution and spatial proximity to tumor cells and conference proceedings from ASCO, SNO, ESMO, and SABCS. Eligible studies included
to potentially inform treatment selection. Methods: Formalin-fixed paraffin-embedded randomized controlled trials and cohort studies (CS) of C-ADC and NC-ADC in HER2-positive BC
(FFPE) samples of patients with metastatic breast cancer who had HER2 IHC/ISH testing patients with BM. Studies reporting SRN rates or related outcomes were included. A random-
according to ASCO-CAP guidelines were stained and analyzed using an 8-plex immu- effects model was used to calculate pooled proportions and risk ratios (RR) with 95% confidence
nofluorescence (mIF) panel (CD3, CD8, CD69, FOXP3, Ki67, PD-L1, PD1, PanCK). For intervals (CIs). Heterogeneity across studies was assessed using the I² statistic, with values of
neighborhood analysis, samples with an area . 2 mm2 and a phenotypes density with . 0–25% considered low, 26–50% moderate, and greater than 50% high. Results: Out of 884
2 cells/mm² were considered. A novel spatial analysis method was used to quantify the studies screened, 9 CS (N = 421) were included. The median age was 56.3 years (IQR: 49.8–57).
Euclidian distance between tumor cells and surrounding immune cell populations. The Patients receiving prior intracranial RT was 41.28%, 19.28% underwent SRS, and 19.58% received
clustering coefficient was used to determine the connectivity of immune cell node prior whole brain radiotherapy (WBRT). The median time of C-ADC was 8.75 days (IQR: 8.0–18.0)
neighbors. These findings were analyzed in relation to the clinical characteristics. and NC-ADC was 273.5 days (IQR: 225.75–327.75). The pooled proportion of SRN in the C-ADC
Results: Tumor and stromal compartment analysis was done on 44 FFPE samples (10 group was 19.5% (95% CI: 9.2%–29.8%; I² = 39.19%, t² = 0.0061, P = 0.1382) indicating moderate
heterogeneity. NC-ADC group experienced a pooled SRN proportion of 6.9% (95% CI: 2.5%–11.2%;
HER2-, 19 HER2low, and 15 HER2+) with 84% collected from metastatic sites. HER2 status
I² = 51.98%, t² = 0.0014, P = 0.0089), signifying high heterogeneity. The pooled RR showed a
was not significantly associated HR status or overall TIL infiltration into the tumor significantly increased SRN risk in the C-ADC group (RR = 2.726, 95% CI: 1.454–5.109, P = 0.002).
compartment. The dominant TIL subset identified was non-regulatory CD3+ T cells as Heterogeneity for the pooled RR was negligible (I² = 0.0%, Q = 0.20, P = 0.977), indicating
defined as CD3+/FOXP3-/CD8-. HER2- samples were more associated with lack of PD-L1 consistent findings across studies. Conclusions: C-ADC is associated with a significantly higher
expression on intratumoral myeloid cells and PD-L1 low expression on tumor cells as risk of SRN. This risk is concerning but must be balanced against potential improvements in local
compared with HER2low and HER2+ (p = 0.06). For spatial analysis, 33 samples (6 HER2-, control and efficacy outcomes. Prospective studies are needed to optimize treatment schedule
16 HER2low and 11 HER2+) were considered. Macrophages and proliferating tumor cells and sequences to minimize toxicity and optimize survival. Research Sponsor: None.
were more abundant in HER2- samples than HER2low or HER2+ (p = 0.006 and p-0.027, Meta-analytical findings.
respectively). Median distances from tumor cells to macrophages and Tregs were shorter Effect
in HER2- cases compared to HER2low (p , 0.001). Although the clustering coefficient Analysis/Measure Size 95% CI I² P-Value Notes
were similar between HER2 groups, HER2low group clustered mostly around macro- Pooled Proportion (Overall) 0.110 0.050–0.169 76.53% 0.0003 High heterogeneity (t²=0.0047)
phages while HER2+ group preferred cytotoxic T cells (CD8+). Conclusions: The spatial C-ADC Proportion 0.195 0.092–0.298 39.19% 0.0002 Moderate heterogeneity (t²=0.0061)
organization and density of immune cells in the HER2low and HER2+ breast cancer NC-ADC Proportion 0.069 0.025–0.112 51.98% 0.0021 High heterogeneity (t²=0.0014)
microenvironment may provide insight into prognosis and guide therapeutic approaches Risk Ratio (C-ADC vs. 2.726 1.454–5.109 0.0% 0.002 No heterogeneity (t²=0.0000)
NC-ADC)
for combination therapies and HER2-targeted immunotherapies. Research Sponsor:
None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
46s BREAST CANCER—METASTATIC

1040 Poster Session 1041 Poster Session

Molecular and clinical insights of trastuzumab deruxtecan efficacy in ad- Transcriptomic biomarkers of therapeutic response to antibody-drug con-
vanced breast cancer (aBC). First Author: Elias Bou Farhat, Brigham and Women’s jugates in metastatic breast cancer: A comprehensive multi-center study.
Hospital, Boston, MA First Author: Samer Alkassis, UCLA Health Jonsson Comprehensive Cancer Center, Los
Background: In aBC, trastuzumab deruxtecan (T-DXd) has demonstrated efficacy in HER2-positive, Angeles, CA
HER2-low, and HER2-ultralow disease subtypes. However, the impact of molecular markers on Background: The three antibody-drug conjugates (ADCs) — sacituzumab govitecan
treatment outcomes requires further investigation, particularly in the context of HER2 status across (SG), trastuzumab deruxtecan (T-DXd), and trastuzumab emtansine (T-DM1) — that are
biopsies, genetic alterations, and the tumor micro-environment (TME). Methods: We retrospectively
analyzed 477 patients (pts) with aBC (DFCI = 369; Yale = 108) treated with T-DXd. HER2 immu-
FDA-approved for treatment of metastatic breast cancer (MBC) have markedly improved
nohistochemistry (IHC) discordance was defined as a shift in HER2 status across the last two tumor patient outcomes. However, most patients with MBC treated with ADCs ultimately have
samples prior to T-DXd (HER2 0 to 1, 2, or 3+ or vice versa). Concordance was defined as consistent disease progression via either primary or acquired ADC resistance. Here, we charac-
HER2-0, HER2-low (1 or 2+) or HER2 3+ across samples. Outcomes included time-to-next treatment terized the transcriptomic profile of drug efflux genes in MBC prior to ADC treatment (tx)
(TTNT) and overall survival (OS); multivariable Cox proportional hazards models were performed. to elucidate biomarkers of response and resistance to SG, T-Dxd, and T-DM1.
Targeted tumor sequencing was conducted on 163 pts with aBC treated with T-DXd. TME analysis of Methods: We analyzed the transcriptomic tumor profile of six drug efflux pump genes
97 patients used machine learning on H&E slides to classify tumors as inflamed, desert, or altered. (ABCB1, ABCC1-4, ABCG2) generated from pre-tx biopsies collected from patients with
Results: Pts with discordant HER2 (n = 118, 25%) showed similar outcomes to those with concordant
MBC (N = 453; 36% TNBC, 26% HR+/HER2-, 20.5% HER2+, 19% NOS) 1 year prior to or up
HER2-0 (n = 32) expression, both of which had significantly worse OS and TTNT compared to
concordant HER2-low (both 1 or 2+) (n = 202) or HER2-3+ (n = 111) tumors (Table). PTEN mutations to 15 days post-tx with SG (n = 204), T-DXd (n = 178), or T-DM1 (n = 71). RNA-sequencing
(mut; n=13) were associated with significantly shorter TTNT. ERBB2 amplifications or gains (n= 31) data were generated and processed with the Tempus xR assay. The correlation between
correlated with improved outcomes. CDK12 deletions or loss (n = 14) were linked to poorer TTNT duration of treatment (DoT) and gene expression was tested for all genes of interest
(Table). PTEN mut and ERBB2 amplifications were predictive of outcomes with T-DXd, as neither using Pearson’s correlation coefficient. Cox proportional hazards models with risk set
alteration was associated with TTNT in pts receiving non-T-DXd first-line systemic therapy. Tumors adjustment were used to test for associations between pre-tx gene expression and
with inflamed TME had the worst outcomes, followed by deserts and altered TMEs (Table). Con- overall survival (OS), where gene expression was modeled as a continuous linear
clusions: We identify favorable biomarkers of T-DXd efficacy in aBC, including concordant HER2-low predictor. The proportional hazards assumption for OS was tested, and Cox modeling
or HER2-3+ status, absence of PTEN mut, and an altered or desert TME. These findings require
validation to refine treatment strategies across HER2-driven malignancies. Research Sponsor: None.
results were omitted when evidence of non-proportional hazards was detected. Given
the exploratory nature of the analyses, all p-values are uncorrected and nominal sta-
OS T-DXd: p/q- TTNT T-DXd: p/q- TTNT 1st Line: p/q- tistical significance was set at p , 0.05. Results: This diverse cohort had a median age
HR (95% CI) value HR (95% CI) value HR (95% CI) value
of 52 and a range of races (55% White, 14% Black, 7.1% Other, 24% Unknown). Median
Concordant HER2-0 (n = 32) vs 1.07 (0.65- 0.789 1.17 (0.76- 0.474 - - DoT across all patients was 130 days. Higher expression of drug efflux pump genes was
discordant (n = 118) 1.75) 1.79)
Concordant HER2-low (n = 202) vs 0.67 (0.49- 0.012 0.65 (0.50- 0.002 - - associated with shorter DoT for T-DXd (ABCB1: -0.290, p = 0.017; ABCC1: -0.274, p =
discordant 0.92) 0.85) 0.025). Additionally, higher expression of ABCB1 was associated with worse OS for T-
Concordant HER2-3+ (n = 111) vs 0.23 (0.14- , 0.001 0.27 (9.18-0.4) , 0.001 - -
discordant 0.38) DXd (HR: 1.30, 95% CI: 1.10 - 1.53, p = 0.002). In the SG cohort, no significant as-
PTEN mutations (n = 13) vs - - 2.20 (1.20-4.0) 0.068 0.99 (0.76-1.35) 0.93 sociations between efflux pump expression and DoT were found, but higher pre-tx
wild-type (WT) tumors (n =150)
ERBB2 amplifications/gains 0.43 (0.26- 0.045 - - 1.1 ( 0.77 - 1.71) 0.51
ABCC1 and ABCC4 gene expression was associated with worse OS (HR: 1.34, 95% CI:
(n = 31) vs ERBB2 WT ( n = 132) 0.72) 1.02-1.75, p = 0.034; HR:1.19, 95%CI: 1.00-1.41, p = 0.042). In the T-DM1 cohort, no
CDK12 loss/deletions (n = 14) vs - - 2.56 (1.39- 0.014 1.42 (1.04-1.86) 0.016 significant associations were found between efflux pump gene expression and DoT or
CDK12 WT (n = 116) 4.76)
Altered (n = 28) vs Desert (n = 34) 0.58 (0.26- 0.18 0.97 (0.52-1.80 0.91 - - OS. Conclusions: Multi-modal analysis identified drug efflux pump gene expression as a
1.30) potential biomarker of resistance, primarily to T-DXd. These findings should be further
Inflamed (n = 35) vs Desert (n = 34) 2.21 (1.20- 0.011 2.13 (1.21- 0.0084 - -
4.05) 3.74) validated, and combinatorial clinical trial strategies may be explored. Research Sponsor:
Tempus AI, Inc.

1042 Poster Session 1043 Poster Session

Integrating dynamic analysis of serial ctDNA testing to enhance diagnostic TP53 genomic alterations including targetable TP53 Y220C mutation in
and prognostic assessments in patients with metastatic breast cancer. First clinically advanced breast cancer. First Author: Nicole Casasanta, Yale Cancer
Author: Qiang Zhang, Department of Medicine, Division of Hematology and Oncology, Center, Yale School of Medicine, New Haven, CT
CTC Core Facility, Robert H. Lurie Comprehensive Cancer Center, Northwestern Uni- Background: Recent studies demonstrating the ability of drugs such as Rezatapopt to target the TP53
versity, Chicago, IL Y220C mutation motivated us to assess the TP53 mutation landscape in clinically advanced breast
Background: The monitoring of circulating tumor DNA (ctDNA) in patient with metastatic cancer (CABC). Methods: FFPE blocks of 23,760 CABC were analyzed by hybrid capture-based
breast cancer (MBC) plays a critical role in predicting therapy resistance, metastasis, and comprehensive genomic profiling that evaluated broad types of genomic alterations (GA) including
mutations, amplifications, deletions, and fusions. MSI-high (MSI-H) status, tumor mutational burden
prognosis. Our previous studies have highlighted the importance of dynamic ctDNA analysis
(TMB), genomic ancestry, mutational signature, and homologous recombination deficiency signature
correlated with treatment resistance and prognosis in MBC (ASCO 2022 (#1057), AACR 2023 (HRDsig) were determined from sequencing data. PD-L1 expression was determined by IHC (Dako
(#1031), and CCR Zhang Q., 2024). Here, we report that multivariable analysis of ctDNA 22C3, TPS scoring system). GA were compared using Fisher’s exact test with the Benjamini-Hochberg
mutations including P53, Myc, and BRAF, provides a significantly greater prognostic impact multiplicity adjustment. Results: Among analyzed cases of CABC, 12,653 (53%) had TP53 GA and 254
on survival. Methods: This study included 391 MBC patients who received systemic (1.1%) were the Y220C mutation. When compared with TP53 wild type (wt) cases, TP53 GA group were
treatment between 2016 and 2022 (IRB-approved non-interventional trial, NU16B06) at the younger (56 vs 60 years; p , .0001) and had a higher median GA (6 vs 5; p , .0001). Both TP53 Y220C
Robert H. Lurie Cancer Center, Northwestern University. Blood samples (15 ml each) were group (15.7% vs 11.2%; not significant [NS]) and TP53 non-Y220C group (18.3% vs 11.2%; p , .0001)
collected from patients at 3 time points: before treatment, and 3 and 6 months after were more frequently of African genetic ancestry than TP53wt group. The TP53 non-Y220C had
treatment. Plasma ctDNA was analyzed by Guardant 360 using NGS for a 74-gene panel. The significantly less European (TP53 non-Y220C: 65.2% vs 74.7%; p , .0001) genetic ancestries. MSI-H
median follow-up was 26.6 months since enrollment. Causal Inference-Ensemble Learning was rare in all groups, but slightly higher in TP53 Y220C than TP53wt cases (1.7% vs 0.3%; p = .036).
was used for statistical analyses. Results: Among 391 patients (54.4% Luminal-like, 17.7% Median TMB was low for all groups (range 2.41-2.61; NS). An APOBEC genomic signature was more
HER2-positive, 27.9% Triple-negative), the most common ctDNA mutations were TP53Mut common in TP53 non-Y220C mutant than TP53wt (6.2% vs 5.0%; p = .0002) but not in TP53 Y220C
(4.3% vs 5.0%; NS). GA more frequent in TP53 Y220C and non-Y220C groups versus TP53wt included
(160 patients, 40.92%), PIK3CAMut (39 patients, 29.4%), and MycMut (53 patients, 13.55%) at
BRCA1, ERBB2, PTEN, and RB1. GA more frequent in the TP53wt group included BRCA2, CCND1, CDH1,
any time point. Other notable mutations included HER2Mut (49 patients, 12.5%), FGFR1Mut (45 ESR1, and PIK3CA. More TP53 Y220C than TP53 non-Y220C mutant cancers had CDH1 mutations
patients, 11.5%), PTENMut (39 patients, 9.9%), and BRAFMut (35 patients, 8.95%). Less fre- (6.3% vs 2.8%; NS) suggesting the Y220C GA may be more frequent in lobular carcinomas. Con-
quent mutations were FGFR2Mut (13 patients, 3.3%), MAPKMut (8 patients, 2.0%), BRCA1Mut clusions: TP53 Y220C is a relatively rare event in CABC. The TP53 mutant group was associated with
(20 patients, 5.1%), BRCA2Mut (17 patients, 4.3%), and CDH1Mut (21 patients, 5.37%). Patients GA in tumor suppressor genes, including BRCA1, PTEN, and RB1, whereas the TP53wt group was
in mutation groups showed significantly shorter median overall survival (OS) compared to associated with GA in pathways associated with endocrine resistance, including PIK3CA and ESR1.
wild-type groups: TP53Mut vs TP53WT, Hazard Ratio (HR) = 1.91 (P = 0.0002); MycMut vs Research Sponsor: None.
MycWT, HR = 3.24 (P , 0.0001); and BRAFMut vs BRAFWT, HR = 2.45 (P = 0.007). No significant
TP53 non- TP53 TP53 non-
correlations were found between other gene mutations and OS. Analysis of multiple ctDNA TP53wt Y220Cmut P- TP53 wt Y220Cmut P- Y220Cmut Y220Cmut P-
mutations (TP53, Myc, and BRAF) revealed significant prognostic differences. Cohort 1 (no (N=11107) (N=254) value† (N=11107) (N=12399) value† (N=254) (N=12399) value†
mutations, 231 patients) had a significantly longer median OS of 32.2 months compared to
BRCA1 1.4% 7.5% ,.0001 1.4% 6.0% ,.0001 7.5% 6.0% NS
20.5 months in cohort 2 (at least one mutation, 119 patients) and 15.3 months in cohort 3 BRCA2 5.2% 4.7% NS 5.2% 3.6% ,.0001 4.7% 3.6% NS
(two or more mutations, 59 patients). Cohort 3 exhibited the worst prognosis compared to CCND1 24.3% 7.5% ,.0001 24.3% 11.3% ,.0001 7.5% 11.3% NS
both cohort 1 and cohort 2 (Chi-square = 13.3, P = 0.0003). These findings suggest that CDH1 23.8% 2.8% ,.0001 23.8% 6.3% ,.0001 2.8% 6.3% NS
combined analysis of ctDNA mutations enhances the ability to predict prognosis. Conclu- ERBB2 10.1% 16.5% 0.003 10.1% 13.5% ,.0001 16.5% 13.5% NS
ESR1 11.5% 3.1% ,.0001 11.5% 4.2% ,.0001 3.1% 4.2% NS
sions: In this study, we identified multiple ctDNA mutations during long-term follow-up, PIK3CA 44.3% 23.6% ,.0001 44.3% 28.2% ,.0001 23.6% 28.2% NS
which are associated with prognosis. The synergy of multivariable analysis of ctDNA mu- PTEN 9.4% 18.9% ,.0001 9.4% 15.4% ,.0001 18.9% 15.4% NS
tations during treatment enhances the role of single ctDNA alterations in monitoring RB1 2.6% 10.6% ,.0001 2.6% 11.8% ,.0001 10.6% 11.8% NS
metastatic prognosis, thereby supporting clinical decision-making. Research Sponsor: Robert †Benjamini/Hochberg adjustment.
H Lurie Cancer Center, Northwestern University.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 47s

1044 Poster Session 1045 Poster Session

Comprehensive results of ESG401, a TROP2-targeting ADC: Updated phase Effect of ERBB2 activating mutations on enhanced internalization and
1 analysis in advanced solid tumors. First Author: Fei Ma, Cancer Hospital Chinese activity of trastuzumab deruxtecan in HER2-non-amplified metastatic
Academy of Medical Sciences and Peking Union Medical College, Beijing, China breast cancer. First Author: Nicholas Mai, Memorial Sloan Kettering Cancer Center,
Background: ESG401 is a novel ADC comprising a humanized anti-TROP2 IgG1 New York, NY
monoclonal antibody conjugated to the Topoisomerase I inhibitor SN-38 via a stable Background: Trastuzumab Deruxtecan (T-DXd) is a HER2-targeting antibody drug con-
cleavable linker. ESG401-101 is a phase 1, open-label, dose-escalation (1a) and dose- jugate approved for the treatment of HER2 low metastatic breast cancer (MBC). Whether
expansion(1b) study evaluating the safety and antitumor activity of ESG401 in advanced HER2 activating mutations define a distinct clinical subset within HER2 low MBC is
solid tumors. This report summarizes the comprehensive phase 1 results. Methods: unknown. Here we present a single institution retrospective study of patients treated with
Patients (pts) aged 18–75 years with locally advanced/metastatic solid tumors received T-DXd and report real world progression free survival (PFS), in patients with mutant vs. wild
ESG401 until unacceptable toxicity, progressive disease, or consent withdrawal. Phase type (wt) HER2. We further modeled the impact of various HER2 mutations on T-DXd
1a results (n = 40) have been reported previously. Phase Ib comprised three parallel internalization and activity preclinically to characterize mechanistic and mutation-specific
cohorts: late-stage TNBC, late-stage HR+/HER2-, and first-line TNBC. Results: As of Oct differences. Methods: All patients who had received T-DXd for HER2-low and HER2-null
23, 2024, 156 pts were enrolled at 13 sites across China (40 in 1a; 116 in 1b). Most pts MBC at Memorial Sloan Kettering Cancer Center were eligible for inclusion. Clinico-
had metastatic HR+/HER2-BC (n = 65; median prior lines: 3; range: 1–10), followed by pathologic data were abstracted from patient records. PFS was determined clinically and
late-line TNBC (n = 47; median prior lines: 3; range: 1–12), first-line TNBC (n = 40), calculated using the Kaplan-Meier method. Univariable and multivariable associations
HER2+BC (n = 2), and one case each of endometrial cancer (EC) and adenoid cystic between PFS and patient characteristics were assessed using Cox-proportional hazards
carcinoma (ACC). All pts had distant metastases at baseline; 13%, 57%, and 54% had models. ERBB2 mutations were modeled in breast cell lines and examined for kinetics of
brain, liver, and lung metastases, respectively. ESG401 demonstrated efficacy in pts with fluorescence-labeled T-DXd cell internalization and potency of T-DXd antitumor effects.
solid tumor(Table), including those with brain metastases. The safety profile remained Results: We found 278 patients who received T-DXd for HER2 non-amplified MBC. Thirty-
consistent with no new or unexpected signals. The most common any-grade TEAEs were one had triple negative breast cancer and 247 had estrogen receptor positive MBC. Median
leukopenia, neutropenia, anemia, nausea, and vomiting. Grade $3 TRAEs were primarily age was 59 and TDXd was the median 6th line of systemic treatment for MBC. Median PFS
neutropenia and leukopenia, none leading to permanent discontinuation. TRAEs led to for all patients was 6.97 months (95%CI 5.73-8.4). ERBB2 mutations were found in 23
(8.2%) patients on genomic sequencing via MSK-IMPACT. Among mutations, 20 were
delayed dosing, dose reduction, and discontinuation in 38.5%, 7.1%, and 2.6% of pts,
known oncogenic mutations per OncoKB (eg D769Y, L755S, S310F, V777L), while 3 were
respectively. Conclusions: ESG401 demonstrated favorable safety and efficacy benefits
variants of unknown significance (L35R, P378L, R1169K). ERBB2 activating mutations
due to its enhanced linker, showing good safety and promising antitumor activity in
were significantly associated with prolonged T-DXd PFS; median 6.28 months in the wt
advanced solid tumors across settings. These results warrant further clinical investi-
population vs 10.58 months with an ERBB2 mutation (HR 0.55, 95%CI 0.31-0.98, p = 0.04).
gation. Clinical trial information: NCT04892342. Research Sponsor: Shanghai Escugen After adjusting for age, treatment line, and ER status, ERBB2 mutations were independently
Biotechnology Co., Ltd. associated with longer PFS. Among patients with ERBB2 activating mutations, 9 had HER2
Late-line IHC 0 disease, while the remaining 14 were at least HER2 IHC 1+. There was no statistically
First-line
HR+/HER2–BC TNBC HER2+BC EC ACC TNBC
significant difference in PFS between patients with HER2 IHC 0 vs 1+ (HR 1.74, 95%CI 0.53-
5.7, p = 0.35) among those with ERBB2 mutations. Finally, expression of the most common
n 58 37 2 1 1 35 ERBB2 mutants in MCF10A cells lead to more rapid internalization of labeled TDX-d into
ORR% (95% CI) 34.5 (22.5, 48.1) 35.1 (20.2, 52.5) 0 0 0 83.0 (66.4, 93.4)
DCR% (95% CI) 77.6 (64.7, 87.5) 62.2 (44.8, 77.5 100 100 100 100 (90.0, -) cells and lower IC50 for inhibition of proliferation. Conclusions: ERBB2 activating mu-
mPFS Mons (95% CI) 7.4 (4.0, 9.2) 3.7 (2.1, 4.9) 3.8, 21.3a 8.3 b 3.7 b NR tations are associated with longer T-DXd PFS in HER2-non-amplified MBC, even when
mDOR Mons (95% CI) 6.6 (4.6, 14.2) 4.5 (3.1, 13.6) NA NA NA NR HER2 IHC was 0, likely due to enhanced ADC internalization. The data imply that ERBB2
a
The actual value for these two patients is listed. mutant breast cancers may be uniquely sensitive to T-DXd, independent of HER2 ex-
b
The actual value for one patient is listed. pression levels. Research Sponsor: National Cancer Institute; CA009512-34A1; Brian
NA, not applicable. NR, not reached. Piccolo Cancer Research Fund; Sussman Family Fund.

1046 Poster Session 1047 Poster Session

Differences in genomic profiles, targeted treatment use, and overall survival Dissecting primary endocrine resistance through ctDNA profiling of a hybrid
in patients with metastatic breast cancer by Area Deprivation Index. First real-world and clinical trial dataset in hormone receptor-positive, HER2-
Author: Emily L. Podany, Washington University in St. Louis, St. Louis, MO negative (HR+/HER2-) metastatic breast cancer (MBC). First Author: Lorenzo
Background: We previously showed racial differences in circulating tumor DNA (ctDNA) Gerratana, Department of Medical Oncology, Centro di Riferimento Oncologico di Aviano
profiles and PI3K inhibitor (PI3Ki) use in patients (pts) with metastatic breast cancer (CRO), IRCCS, Aviano, Italy
(mBC); however, these findings may be influenced by socioeconomic disadvantage. A Background: New targeted therapies, including novel endocrine agents and antibody drug
validated measure to explore this is the Neighborhood Atlas Area Deprivation Index (ADI, conjugates, are revolutionizing the treatment of HR+/HER2- metastatic breast cancer (MBC).
Kind et al NEJM 2018), which includes 17 measures of neighborhood disadvantage such However, questions surrounding primary and secondary endocrine resistance (R1 and R2, re-
as poverty, employment, and education. We sought to determine differences in genomic spectively) still hinder the development of personalized treatment strategies. This study aimed to
profiles, PI3Ki use, and overall survival (OS) in pts with mBC by ADI. Methods: This investigate R1 by leveraging liquid biopsy in a hybrid real-world and clinical trial dataset.
retrospective cohort study analyzed 1127 pts with mBC and ctDNA testing using the Methods: This study used the nationwide (US-based) deidentified Flatiron Health-Foundation
Medicine MBC clinicogenomic database (FH-FMI CGDB), comprising data originated from ~280
Guardant360 assay who were treated at Washington University in St. Louis (N = 634),
US cancer clinics (~800 sites of care), and analyzed a cohort of 855 patients (pts) profiled through
Massachusetts General Hospital (N = 313), Weill Cornell (N = 109), and Northwestern the FoundationOne Liquid CDx NGS panel, combining it with the first 65 pts enrolled in the GIM-24-
University (N = 71). 9-digit zip codes were converted into national ADI ranks (0-100) PalboBP study (NCT04318223). R1 was defined as 1-line PFS of , 6 months. A 1:3 Propensity
divided into high deprivation (HDep, rank $60) and low deprivation (LDep, rank , 60) Score Matching was applied to balance key factors (i.e. age, ECOG performance status, visceral,
groups based on prior studies. Multivariate models were designed to determine genomic lymph node, multiple metastasis, type of CDK6/4i). Pathogenic alterations with a . 10%
and prognostic differences by ADI. Pts with PIK3CA mutations were evaluated by ADI and prevalence were tested singularly and according to oncogenic pathways based on Sanchez-Vega
use of PI3Ki in the second line or beyond, either through clinical trial enrollment or after F et al, Cell. 2018. Associations between ctDNA alterations, R1, and R2 were assessed using
FDA approval. OS from time of first ctDNA test was stratified by ADI and self-reported logistic regression, while prognosis was evaluated through Cox regression. Results: A set of 528
race. Results: The cohort included 165 Black pts (14.6%) and 335 pts (29.7%) from HDep pts (respectively 132 and 396 for R1 and R2) was selected from the original cohort of 855 pts. Top
zip codes. Black pts were more likely to be from HDep areas (Odds ratio [OR] 3.82, 95% detected alterations were PIK3CA SNV (46%), TP53 SNV (33%) and ESR1 SNV (24%). R1 was
confidence interval [CI] 2.62-5.57, P , 0.001). There were no differences in mBC subtype associated with TP53 SNV (OR = 2.30, P , 0.001), CCND1, FGF19, FGF3 and FGF4 CNVs (re-
between ADI groups. Pts with HR+ HER2- mBC in the HDep group were significantly less spectively OR = 1.75, P = 0.018; OR = 1.62, P = 0.044; OR = 1.80, P = 0.015; OR = 1.73, P = 0.024).
likely to receive PI3Ki vs LDep (8/46, 17.4% vs 33/90, 36.7%, P = 0.02) despite equal On the other hand, ESR1 SNV was associated with R2 (OR = 0.47, P = 0.005). CCND1, FGF19, FGF3
and FGF4 CNVs were significantly co-occurring (P , 0.001) and located in the chromosome (chr)
incidence of PIK3CA mutations. Pts in the HDep group were more likely to have TP53
11q13.3 region. In multivariable analysis, TP53 SNV, ESR1 SNV and chr11q13.3 CNV maintained
single nucleotide variants (snv) (OR 1.58, 95% CI 1.18-2.10, P = 0.002) and less likely to their association with R1 (respectively OR = 2.22, P , 0.001; OR = 0.47, P = 0.006 and OR = 1.94, P
have AKT1 snv (OR 0.29, 95% CI 0.11-0.80, P = 0.017). Among pts in the HDep group, = 0.006). Pathway analysis was consistent, showing a significant association between R1 and
worse prognosis was seen in pts who self-identified as Black (hazard ratio [HR]1.51, 95% SNVs in the P53 and in the ER pathways (respectively OR = 1.92, P = 0.002; OR = 0.58, P = 0.024)
CI 1.02-2.25, P = 0.04), had PIK3CA snv (HR 1.73, 95% CI 1.23-2.44, P = 0.002), or TP53 snv and CNVs in the cell cycle pathway (OR = 2.18, P = 0.001). No differences were observed for 2-line
(HR 1.56, 95% CI 1.12-2.17, P = 0.009). Median OS was significantly shorter in the HDep vs (PFS2) across R1 and R2 (median PFS2 8.89 vs 8.03 months P = 0.589). Chemotherapy was
LDep group (24 months [mos] vs 28 mos, P = 0.04) and significantly lower for Black pts in prevalent in the R1 group (35% vs 15%). Within pts receiving 2-line endocrine therapy, TP53 SNV
the HDep vs Black pts with LDep or White pts with HDep or LDep (15 mos vs 25-28 mos, P was the only prognostic factor in R1 (HR = 2.02, P = 0.008), while SNVs in TP53 and ESR1 had an
= 0.02). Conclusions: In this multi-institutional cohort, we identified significant dis- impact on PFS2 in R2 (respectively HR = 1.50, P = 0.003 and HR = 1.45, P = 0.008). Conclu-
parities in the use of PI3Ki in HDep neighborhoods and higher rates of TP53 snv, which are sions: This study confirms TP53 and ESR1 as key factors for R1 and R2, respectively, with ESR1
associated with aggressive tumor biology. Pts with mBC in HDep areas, especially Black showing a prognostic impact on PFS2 in R2 only. Additionally, chr11q13.3 emerges as a new
pts, had shorter OS. Further research is needed to validate these findings, determine the candidate region for R1. These results provide critical data for both decision making and 2-line
root causes of these disparities, and implement change to achieve equity in precision clinical trial design. Research Sponsor: None.
medicine use. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
48s BREAST CANCER—METASTATIC

1048 Poster Session 1049 Poster Session

Genomic alterations (GAs) associated with durability of benefit from tras- Macroscale genomic alterations in histomolecular invasive lobular carci-
tuzumab deruxtecan (T-DXd), trastuzumab emtansine (T-DM1) and sacitu- noma compared to other breast cancer subtypes. First Author: Jason A.
zumab govitecan (SG) in metastatic breast cancer (MBC). First Author: Tess A. Mouabbi, The University of Texas MD Anderson Cancer Center, Houston, TX
O’Meara, Dana-Farber Cancer Institute, Boston, MA Background: Although invasive lobular carcinoma (ILC) is often classified as a separate
Background: Predictive biomarkers are needed to guide use of T-DXd, T-DM1 and SG in breast cancer (BC) subtype with distinct molecular features, options for diagnosis and
MBC. We used real-world comprehensive genomic profiling (CGP) of tumor tissue and treatment remain similar to other BCs. Using an integrated histomolecular approach to
circulating tumor DNA (ctDNA) to describe pre- and post-treatment somatic GAs in classify 617 BC samples into either histomolecular ILC (hmILC) or histomolecular no
patients receiving these antibody-drug conjugates (ADCs) and evaluated the predictive special type (hmNST) subsets, we compared their macroscale genomic alterations to
value of ERBB2 amplification (ERBB2amp) in MBC treated with T-DXd and T-DM1. describe biological traits of the ILC BC subtype, which may lead to improved approaches
Methods: MBC patients with FoundationOne CDx or FoundationOne Liquid CDx who in BC therapies. Methods: A total of 617 BC FFPE samples were subject to whole-exome
received T-DXd, T-DM1 or SG monotherapy were included. Clinical data originated from 280 and bulk RNA sequencing analysis. hmILC subset was defined based on CDH1 trun-
cancer clinics (~800 sites of care) between 1/2011-4/2024 included in the US-wide de- cation/deletion or low CDH1 expression (z-score , -2.53MAD), while all other samples
identified Flatiron Health-Foundation Medicine MBC clinicogenomic database. For each were classified as hmNST. Copy number variations (CNVs) were assessed using
ADC and MBC subtype, we characterized the GA profile of pre-ADC samples. Pre- and post- Sequenza to detect recurrent amplifications/gains and deletions; homologous re-
treatment GAs were then compared by chi-square, adjusted for multiple comparisons. For combination deficiency (HRD) scores were calculated based on large-scale state
patients who received T-DXd or T-DM1, time to next treatment (TTNT) was compared transitions and loss-of-heterozygosity events; tumor mutational burden (TMB) scores
between patients with or without ERBB2amp by tissue CGP by Cox models, adjusted for were evaluated as percent of mutations per megabase; and mutational signatures were
age, ECOG status, HR status, and line of therapy. Results: We identified 1,177 pre-ADC deconvoluted using maftools R package. Results: Genome-wide CNV analysis revealed
samples (n = 972 tissue, n = 205 liquid biopsy; T-DXd n = 492, T-DM1 n = 167, SG n = 518). distinct patterns in hmILC compared to hmNST. hmILC showed hallmark deletions at
TP53 (59.7%), PIK3CA (34.4%), and ERBB2 (20.6%) were most commonly altered across all regions harboring CDH1 (16q), while gains were observed significantly more frequently
samples. Median TTNT for T-DXd in HER2+ MBC (n = 106) was 16.6 mo; ERBB2 alterations at regions harboring FCGR3A (1q) compared to hmNST. Elevated APOBEC activation
were present in 84.6% of cases above the median vs 47.5% below. Median TTNT for cases signature expression was found in 32% hmILC vs. 19% hmNST samples (p = 0.002, chi-
with somatic BRCA1/2 mutations (n = 6) was 8.48 mo vs 18 mo for BRCA1/2 wildtype.
squared test). HRD-positive cases were less frequent in the hmILC subset (25%)
Formal statistical analyses of baseline GAs and associations with TTNT on each ADC will
compared to hmNST (42%) (p = 0.001). In contrast, hmILC tumors more frequently
be presented at the conference. GAs in ATM (24% vs 3.7%, p , 0.0001), GNAS (6% vs 2%, p
demonstrated high TMB scores compared to hmNST (10% vs. 2%, p = 0.0003). Moreover,
, 0.0001), EGFR (4% vs 0.2%, p , 0.01) and ERCC4 (2% vs 0, p = 0.02) were more prevalent
TMB-positive hmILC samples were mostly [Link] alterations in genes
in samples post-T-DXd (n = 50, 15 tissue/35 liquid) vs pre-T-DXd (n = 492, 378 tissue/114
liquid). GAs in ERCC4 (3.3% vs. 0, p = 0.001) were more prevalent in samples post-SG (n =
like FANCA, FANCD2 and PALB2 were more frequently enriched in hmILC tumors
60, 31 tissue/29 liquid) vs pre-SG (n = 518, 439 tissue/79 liquid). HER2+ MBC with compared to hmNST (p , 0.1). Furthermore, hmILC subset with high HRD scores
ERBB2amp by CGP (n = 54) had more favorable TTNT on T-DXd vs non-amplified HER2+ (n frequently harbored additional DNA repair gene mutations (e.g., BRCA2) compared to the
= 30; 22.5 vs 6.4 mo, HR 0.10, 95% CI 0.04-0.24, p , 0.0001). HER2-low MBC with subset with low HRD scores (p = 0.02). Conclusions: This study revealed macroscale
ERBB2amp by CGP (n = 6) had more favorable TTNT on T-DXd vs non-amplified HER2-low genomic alterations, such as unique CNV patterns, altered distributions of HRD- and
(n = 263; NR vs 7.4 mo, HR 0.22, 95% CI 0.05-0.90, p = 0.035). HER2+ MBC with ERBB2amp TMB-positive cases and increased APOBEC-driven mutational processes, in the hmILC
by CGP (n = 102) had more favorable TTNT on T-DM1 vs non-amplified HER2+ (n = 45; 8.3 BC subset. These distinct genomic architectures highlight the need for innovative trials
vs 2.6 mo, HR 0.50, 95% CI 0.33-0.75, p , 0.001). Conclusions: This real-world analysis using inhibitors of DNA repair and related pathways for hmILC BC patients, particularly in
provides insight into baseline GAs and associations with TTNT on T-DXd, T-DM1 and SG in those with high HRD-high tumors. Research Sponsor: None.
MBC as well as GAs that emerge on treatment. ERBB2amp by CGP carried additional
predictive value to IHC/FISH HER2 status in both HER2+ and HER2-low MBC treated with
T-DXd and HER2+ MBC treated with T-DM1. Research Sponsor: None.

1050 Poster Session 1051 Poster Session

Landscape analysis of proteins in the development of breast cancer brain Evaluation of tumor immune microenvironment in Hispanic and African
metastasis. First Author: Dongyan Xu, The Second Affiliated Hospital, Zhejiang American breast cancer. First Author: Robert Hsu, Norris Comprehensive Cancer
University School of Medicine, Hangzhou, China Center, University of Southern California, Los Angeles, CA
Background: Breast cancer brain metastasis(BCBM) is a major cause of mortality in ad- Background: Hispanics or Latinos (HL) and African Americans or Black (AA) have a higher prevalence
vanced breast cancer patients, and treatment options are limited. In this project, we in- of advanced-stage breast cancer (BC) at diagnosis compared to Non-Hispanic Whites (NHW). To
vestigated the proteomic landscape of primary breast cancer and brain metastasis. understand the role of immune system, we evaluated the tumor immune microenvironment (TIME) by
Methods: We conducted high-throughput proteomic analysis on primary and brain met- race/ethnicity among HL, AA, and NHW BC patients. Methods: 15544BC samples were tested by NGS
(592, NextSeq; WES, NovaSeq) and WTS (NovaSeq; Caris Life Sciences, Phoenix, AZ). Race/ethnicity
astatic breast cancer tissues surgically resected from 21 patients. Following screening and data is self-reported. Immune cell were estimated using WTS deconvolution (Quantiseq). Gene ex-
sample processing using pressure cycle technology (PCT) for peptide extraction, we used pression profiles were analyzed for T-cell inflammation score (TIS) and interferon-gamma (IFN-gamma)
the data-independent acquisition (DIA) mass spectrometry (MS) method to acquire the data. score. Real-world overall survival (OS) was obtained from insurance claims and calculated from date of
Additionally, we used biological function assays, co-culture experiments, and transcriptome tumor biopsy to last contact using Kaplan-Meier estimates. Statistical significance was determined by
sequencing analysis to further investigate the differentially expressed proteins. Re- chi-square and Mann-Whitney U test with p-values adjusted for multiple comparisons (q , .05).
sults: Patients were mostly in clinical stage II, with two excluded from analysis due to no Results: 7170 NHW (35.3%, N = 2528) biopsied (bx) from primary breast cancer (pBC), 64.7% (N = 4642)
surgery. Breast cancer classifications included 47.6% HER2+/HR-, 28.6% HER2+/HR+, 9% metastatic bx (mBC), 1,508 AA (pBC 39.3% N = 592, mBC 60.7% N = 916), and 1,754 HL (pBC 44.1% N =
HER2-/HR+, and 14.3% TNBC. All patients received chemotherapy, with 57% undergoing 774, mBC 55.9% N = 980) cases were included. By subtype, there were 1,956 (60.4% NHW, 20.7% NHB,
neoadjuvant therapy and 12 receiving targeted therapy (all with trastuzumab). A total of 18.9% HL) TNBC, 3425 HR+/HER2- (72.6% NHW, 11.9% NHB, 15.6% HL), and 694 HER2+ (64.6% NHW,
15.7% NHB, 19.7% HL). Across all cases, AA (20.5%) and HL (20.4%) had greater incidence (%) of PD-L1+
9430 protein groups were identified, with 692 showing differential expression in BCBM. cases versus (vs) NHW (17.4%), all q , .05. TMB-High (³10 mut/Mb) was similar in NHW (11.5%), AA
Notably upregulated proteins included CRYAB, GFAP, STXBP1 and significantly down- (10.8%), and HL (10.9%). AA tumors had lower median % cell infiltration of M2-like macrophages (M2
regulated proteins included CACNA1A, AOC3, PMP2, OGN. Most differentially expressed Mw), B cells, and neutrophils vs NHW (Table). HL had a lower fraction of M2 Mw and higher CD8+ T cells
proteins were involved in extracellular matrix (ECM) and cell-cell interactions, with collagen (Table). AA had lower TIS (-8 vs 1, p = .02) while HL had lower IFN-gamma (-0.38 vs. -0.35, q , .05) vs
family proteins (e.g., COL14A1, COL22A1) playing key roles in BCBM. HER2+ BCBM was NHW. By subtype, AA had lower neutrophils (4% vs 4.3%) and increased DC fractions (3.1% vs 2.8%) in
associated with ECM pathways, while TNBC impacted the immune microenvironment. TNBC vs NHW, all q , .05; no significant changes seen in HL vs NHW. AA had worse mOS than NHW
Regardless of the subtype, we identified 11 proteins that collectively contribute to the overall (31.8 vs 36.8 months (mo)), HR 1.1, 95% CI 1 – 1.2, p = , .01), in pBC (40.3 vs 49.9 mo, HR 1.3,
development of BCBM, including CRYAB, ATP6V0A1, HLA-DQB1, TPM2, SERPINB9, NFATC2, 95% CI 1.1 – 1.5, p = , .01), but not mBC (27.4 vs 29.1 mo, HR 1, p = 0.2). HL had similar mOS vs NHW
GRAP, ALDH1L2, DHRS4L2, SEPTIN1 and SAMHD1. We found that high ACOX1 and low overall (37.4 vs 36.8 mo, HR 0.9, p = 0.9) and in mBC (29.1 vs 31 mo, HR 0.96, p = 0.4), but worse mOS in
pBC (44.7 vs 50.0 mo, HR 1.1, 95% CI 1 – 1.3, p = .01). Conclusions: Our study shows worse mOS in AA
KRT9/KRT14 expression were linked to poor prognosis in BCBM. Analyzing whether patients
and HL pBC cases vs NHW, possibly from a less inflamed TIME in AA and HL and lower fraction of
had undergone targeted therapy, we found that resistance might be acquired through the neutrophils and M2 Mw despite higher % of PD-L1+. Targeting Mw and CD8+ T cells and converting cold
oxidative phosphorylation pathway, promoting brain metastasis in HER2+ breast cancer. to hot TIME may lessen race/ethnic disparities, especially in early-stage BC. Research Sponsor: None.
Interestingly, CRYAB expression was prominent across all subtypes of BCBM and targeted
Immune cell fraction of NHW, AA and HL BC.
therapy groups, suggesting it may serve as an essential biomarker for BCBM. We have
NHW AA HL q-value q-value
confirmed that CRYAB can promote the proliferation and migration of HER2+BCBM, and (median %) (median %) (median %) NHW vs AA NHW vs HL
preliminarily verified that CRYAB is closely related to immune infiltration through CXCL5,
B cell 5.2 4.8 5.0 ,.05 0.7
CXCL8 and CCL3 in tumor microenvironment, which may promote the occurrence of DC 2.6 2.7 2.6 0.08 0.4
HER2+BCBM by affecting the NF-kB pathway. Conclusions: Leveraging high-throughput M1 Mw 2.5 2.4 2.4 0.4 0.9
proteomics, we present a detailed analysis that elucidates the biological processes involved M2 Mw 4.6 3.7 4.2 ,.05 ,.05
Neutrophils 3.7 3.5 3.4 ,.05 ,.05
in developing and progressing BCBM from multiple angles. This work offers new directions NK cell 2.9 2.9 2.9 0.7 0.6
for early prediction, treatment, and prognosis of BCBM in clinical practice. Research CD8+ T cell 0.1 0.15 0.26 0.8 ,.05
Treg 1.5 1.5 1.6 0.6 ,.05
Sponsor: National Natural Science Foundation of China; 81602716.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 49s

1052 Poster Session 1053 Poster Session

Circulating tumor DNA (ctDNA)-based minimal residual disease (MRD) Antibody drug conjugates treatment response score (ADC TRS) for sequenc-
measured by Guardant Reveal in patients (pts) with HER2-positive ing trastuzuamb deruxtecan (T-DXd) and sacituzumab govitecan (SG) in
(HER2+) metastatic breast cancer (mBC) with long-term disease control advanced breast cancer (aBC). First Author: Andi Cani, University of Michigan
on first-line trastuzumab-pertuzumab. First Author: Antonio Llombart-Cussac, Medical School, Rogel Cancer Center, Ann Arbor, MI
Hospital Arnau Vilanova - Lliria, FISABIO, Valencia, Spain Background: T-DXd, targeting HER2, is approved in hormone receptor positive or negative (HR+/-)
Background: The CLEOPATRA and PERUSE trials established the combination of a taxane HER2-low (1-2+ by immunohistochemistry [IHC]) and HR+/HER2-ultralow (0+ with membrane
with the antiHER2 monoclonal antibodies trastuzumab and pertuzumab (HP) as the gold staining) aBC. SG, targeting TROP2, is approved in HR+/HER2 negative [ -; , 3+ by IHC] and triple
negative BC (TNBC; HR-/HER2-). Given limitations in determining IHC 0-1+ status and lack of pre-
standard first-line treatment for HER2+ mBC. In both studies, the progression events
dictive biomarkers, biomarkers for sequencing these ADCs in those with HER2- aBC are needed.
reached a plateau after 4 years and up to 30% of pts remained long-term progression-free, Recently, Thomas et al. reported the discovery of ADC TRS—a generalized model combining individual
hypothesizing HP maintenance can be safely discontinued. We therefore evaluated ADC target expression, proliferation, and adhesion to predict multi-ADC target/tumor type clinical
whether epigenomic-based ctDNA MRD analysis can potentially identify pts with a higher benefit—as well as validation of SG and T-DXd TRS models (tuned by pan-tumor response rates) by a
chance of permanent remission. PRE-PHENIX is a multi-center observational study that qRT-PCR based clinical trial assay (CTA; Strata Select ADC [SSA]) for predicting clinical benefit of first
explores the prevalence of MRD measured by Guardant Reveal in HER2+ mBC pts on long- ADC (SG or T-DXd) in those with HER2- aBC (ASCO 2024 #3140; AACR 2025 #1014). Here, we
term first-line HP maintenance. Methods: A total of 40 pts with HER2+ mBC on first-line evaluated ADC target expression and T-DXd/SG TRS status from SSA validation cohort HER2- BC
treatment with HP maintenance for a minimum of 4 years were included. Confirmation of patients stratified by clinical HR/HER2 IHC status. Methods: Adults with aBC from an observational
no progressive disease by CT or PET-scan in the last 3 months previous to study entry was trial (NCT03061305) with valid FFPE tumor tissue results from SSA validation cohort testing were
mandatory. Plasma samples were analysed using Guardant Reveal powered by the included. aBC types were determined by HR/HER2 IHC results (by ASCO/CAP scoring), with those
HER2 3+ or HER2 amplified (by Strata Select testing) considered HER2+. ADC target component
Guardant Infinity platform, a tissue-free epigenomic assay interrogating differentially expression (HER2 or TROP2; pan-tumor scaled absolute expression) and T-DXd and SG TRS statuses (+
methylated regions of DNA optimized to detect breast cancer DNA from normal cell-free associated with more clinical benefit) by SSA were compared by IHC defined aBC types. Results: 230
DNA. Two ctDNA tests were performed on each patient within a 6 – 12-week interval. patients with aBC from SSA validation testing were included (median age 58 yrs, 40% self-reported
Additionally, 11 pts with confirmed disease progression on antiHER2 therapy for mBC were non-European;180 with definitive aBC type [see Table for distribution]). HER2 expression was sig-
included as case controls. The primary objective was to establish the prevalence of positive nificantly increased vs. TROP2 in HER2+ and HR+/HER2 aBC, and did not significantly differ in those
MRD in both populations and the agreement between the two tests for the Long-Term with TNBC (see Table). Across all patients, 42%, 30%, 27% and 0.4% were T-DXd/SG TRS +/+, -/-, +/-
responders. Results: Median age was 63.2 years (range 30.8 – 84.4). The median duration and -/+, respectively. Results were similar in those with HER2 IHC 0+ (n = 37, median HER2 vs. TROP2 =
of first-line HP treatment was 6.9 years (range 4.2 - 11.1). At diagnosis, 26 pts (65%) 2.2 vs. 2.3, p = 0.51; 8% and 3% T-DXd/SG TRS +/- and -/+, respectively. Conclusions: Pan-tumor
presented with “de novo” mBC and 20 (50%) had visceral disease. The last radiological optimized, validated ADC TRS models support sequencing T-DXd before SG in nearly all patients with
advanced HER2- BC, including those with TNBC and HER2 0+ IHC. Prospective evaluation of the CTA is
evaluation categorized 6 pts (15%) as having stable disease (SD), 2 pts (5%) with partial
warranted. Research Sponsor: Strata Oncology.
response (PR), and 32 pts (80%) with complete response (CR). Among the 11 pts with
confirmed progression, 2 presented exclusive Central Nervous System (CNS) disease. ADC target expression (HER2 or TROP2) and T-DXd | SG TRS status by SSA in BC patients.
Guardant Reveal identified MRD in 4 long-term responders (10%), 3 out of 6 pts (50%) with T-DXd/SG TRS Status
SD, 1 of 2 pts (50%) with PR, and no MRD among the 32 pts with CR. A perfect agreement BCa Type (n) HER2^ TROP2^ p-value^ +/+ -/- +/- -/-
was observed between the two tests (Kappa-index of 1). Ten out of the 11 pts (91%) with HER2+ 43 6.4 2.6 ,0.0001 42% 7% 51% 0%
disease progression had MRD, including the two with exclusive CNS involvement. Con- IHC NA* 50 2.6 2.7 0.31 44% 40% 16% 0%
clusions: Our study demonstrates clinically significant performance of a tissue-free MRD HR+/HER2- 76 3.0 2.3 ,0.0001 54% 20% 26% 0%
test, Guardant Reveal, as a potential non-invasive monitoring tool to guide de-escalation TNBC 61 2.3 2.3 0.7 25% 52% 21% 2%
Total 230 2.9 2.4 ,0.0001 42% 30% 27% 0.4%
strategies in pts HER2+ mBC pts with long-term remissions on HP treatment. A prospective
study (PHENIX) to guide HP interruption by ctDNA monitoring is planned. This study was ^Median, pan-tumor scaled HER2 and TROP2 target expression; Wilcoxon test.
funded by a Fundación Contigo full grant (Spain) and Guardant Health. Research Sponsor: *IHC not available; HER2 not amplified.
None.

1054 Poster Session 1055 Poster Session

Estrogen receptor (ER) expression on circulating tumor cells (CTCs) and cell Overall survival in patients with HR+/HER2- advanced or metastatic breast
free DNA (cfDNA) mutational landscape in the PACE randomized phase II cancer treated with a cyclin-dependent kinase 4/6 inhibitor plus an aroma-
study. First Author: Carolina Reduzzi, Weill Cornell Medicine, New York, NY tase inhibitor: A US Food and Drug Administration pooled analysis. First
Background: The PACE (NCT03147287) randomized phase II trial investigates CDK4/6 inhibition Author: Jennifer Gao, Oncology Center of Excellence, U.S. Food and Drug Administration,
beyond progression in combination with endocrine treatment, with or without PD-L1 inhibition, in Silver Spring, MD
hormone receptor-positive (HR+)/HER2- metastatic breast cancer (MBC) (Mayer et al 2024). We Background: Cyclin-dependent kinase 4/6 inhibitors (CDKI) are FDA-approved for use in com-
previously reported that cfDNA alterations and CTC number correlated with survival and treatment bination with aromatase inhibitors (AI) for the treatment of patients with hormone receptor-
response (Jeselsohn et al 2024; Gerratana et al ASCO 2023). Here we investigated ER expression on positive (HR+), human epidermal growth factor receptor 2-negative (HER2-), advanced or met-
CTCs in relation to the cfDNA mutational landscape. Methods: Samples were collected at baseline. astatic breast cancer (MBC) as initial (1L) endocrine-based therapy. We have previously reported
CTC enumeration and ER protein expression on CTCs (by immunofluorescence) was evaluated with the pooled analyses of the benefit in progression-free survival of adding CDKI to AI, and here report
the CellSearch and the ACCEPT software. Samples were classified as CTChigh or CTClow (cutoff $5 the pooled overall survival (OS) results for adults treated with CDKI + AI for 1L HR+/HER2- MBC.
CTC/sample). CTChigh samples were defined ER+ if .15% CTCs/sample expressed ER to ensure good Methods: We pooled individual patient data (N=2252) from 4 randomized trials (MONALEESA-2 &
inter-group stratification. Concurrently, cfDNA was analyzed with the Guardant360 assay. Only 7, MONARCH-3, PALOMA-2) of a CDKI (abemaciclib, palbociclib, ribociclib) or placebo + AI in adults
pathogenic single nucleotide (snv) and copy number variations (cnv) with $3% prevalence were with 1L HR+/HER2- MBC. OS was defined as time from randomization to death from any cause and
included and categorized into oncogenic pathways (Sanchez-Vega et al 2018). Differences in was a key secondary endpoint in all 4 trials. Not all 4 trials reached OS statistical significance, but
distribution across CTC groups were tested through Chi-squared and Fisher’s test. Results: From all OS hazard ratios of the individual trials were ,1. The median OS was estimated using Kaplan-
220 enrolled patients, 167 were evaluable for ER on CTCs. Of these, 91 were CTClow, 30 were CTChigh/ Meier methods, and hazard ratios with 95% confidence intervals (CI) were estimated using Cox
ER-, and 46 CTChigh/ER+. ESR1 mutations were more common in CTChigh /ER+ samples, while regression models. Analyses were prespecified, with patients analyzed collectively and by various
CTChigh/ER- samples had higher incidence of alterations in SMAD4, PIK3CA, BRAF and CDK4 clinicopathological subgroups of interest. Results: Overall results in all patients and various
compared to the other 2 groups (Table 1). CTChigh/ER- had also higher mutant allele frequency clinicopathologic subgroups of interest are shown (Table). Conclusions: In this descriptive ex-
compared to CTChigh /ER+ and CTClow (MAF . 3% in 73% vs 54% and 31%, respectively, p , 0.001).
ploratory pooled analysis, the addition of a CDKI to AI suggested an association with an OS benefit
CTClow samples had overall lower cfDNA alteration incidence. Similarly, alterations in the ER pathway
for this class of drugs used as a component of 1L endocrine-based therapy for adults with HR+/
were more frequent in samples with CTChigh /ER+, whereas alterations in PI3K, cell cycle and P53
HER2- MBC. Additional research is needed to determine which subgroup of patients may benefit
pathways were more common in CTChigh/ER- samples. Alterations in the RTK/RAS/RAF pathway
more or less of the addition of a CDKI to AI. Research Sponsor: None.
were more common in CTChigh samples (23% and 28% for ER- and ER+ vs 9.9% for CTClow, p = 0.017).
Similar results were observed with a 10% threshold. Conclusions: Distinct cfDNA alterations were # Events
identified based on ER expression in CTCs in HR+/HER2- MBC. Integrating CTC enumeration and CDKI/n # Events
cfDNA profiling may help elucidate resistance mechanisms, identify actionable targets, and predict n (%) Placebo/n (%) HR (95% CI)
benefit from continued CDK4/6 inhibition beyond progression. Research Sponsor: Pfizer; Merck All 2252 716/1320 (54) 550/932 (59) 0.81 (0.73, 0.91)
KGaA; CrossRef Funder ID: 10.13039/100009945. PR negative 273 84/155 (54) 89/118 (75) 0.51 (0.38, 0.70)
De Novo 752 233/450 (52) 173/302 (57) 0.82 (0.67, 1.00)
Incidence of cfDNA alterations across the 3 CTC-based groups.
Lobular Histology 144 72/97 (74) 34/47 (72) 0.99 (0.66, 1.50)
cfDNA alterations CTClow1 CTChigh/ER-1 CTChigh/ER+1 P value Bone-Only 493 142/284 (50) 115/209 (55) 0.74 (0.58, 0.95)
ESR12 36 (40) 15 (50) 31 (67) 0.009
Liver/Lung Mets 1111 365/639 (57) 291/472 (62) 0.81 (0.70, 0.95)
SMAD42 0 (0) 3 (10) 2 (4.3) 0.009 Age <40 193 40/106 (38) 44/87 (51) 0.78 (0.51, 1.21)
PIK3CA3 2 (2) 4 (13) 0 (0) 0.012 Age >70 403 159/247 (64) 106/156 (68) 0.86 (0.67, 1.09)
BRAF3 1 (1) 1 (3.3) 5 (11) 0.022 ECOG 1 851 305/499 (61) 239/352 (68) 0.78 (0.66, 0.93)
CDK43 1 (1) 3 (10) 2 (4.3) 0.035 White 1594 529/919 (58) 407/675 (60) 0.88 (0.77, 1.00)
ER pathway2 40 (44) 15 (50) 32 (70) 0.017 Asian 438 113/269 (42) 89/169 (53) 0.59 (0.45, 0.78)
PI3K pathway3 2 (2.2) 4 (13) 0 (0) 0.012 Black or African 43 14/25 (56) 11/18 (61) 0.80 (0.36, 1.76)
Cell cycle pathway3 8 (8.8) 9 (30) 8 (17) 0.019 American
P53 pathway2 27 (30) 16 (53) 13 (28) 0.040
Additional clinicopathologic subgroup analyses conducted with results not shown.
1
n (%); 2snv; 3cnv.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
50s BREAST CANCER—METASTATIC

1056 Poster Session 1057 Poster Session

Prospective cohort study of palbociclib in HR+/HER2- metastatic breast Results of a phase 1 study of vosilasarm (EP0062), a first-in-class oral
cancer in Japan. First Author: Takashi Ishikawa, Department of Breast Surgical selective androgen receptor modulator (SARM) in patients with advanced or
Oncology, Tokyo Medical University, Tokyo, Japan metastatic AR+/ER+/HER-2- breast cancer. First Author: Hyo S. Han, H. Lee
Background: The combination of palbociclib (PAL) with an aromatase inhibitor or ful- Moffitt Cancer Center and Research Institute, Tampa, FL
vestrant has been shown to improve progression-free survival (PFS) in hormone receptor Background: Vosilasarm (EP0062) is an oral, nonsteroidal, Selective Androgen Receptor
(HR)-positive and human epidermal growth factor receptor (HER2)-negative metastatic Modulator (SARM). Initially developed under the code RAD140, EP0062 has been
breast cancer. However, the addition of PAL to endocrine therapy increases toxicity and reformulated with markedly improved bioavailability and pharmacokinetics. Preclinically,
cost compared to endocrine therapy alone. In addition, PAL treatment may affect the vosilasarm has been shown to act as a potent tissue-selective AR agonist, suppressing
efficacy of subsequent treatments, as its benefit in terms of overall survival (OS) has not growth and proliferation of multiple AR+/ER+/HER-2- breast cancer cell lines and patient-
yet been demonstrated. Therefore, it is crucial to prospectively evaluate whether PAL can derived xenograft models, as monotherapy or in combination with standard of care (SoC)
improve clinical outcomes and quality of life (QoL) for patients in a real-world setting. regimens (Clin Can Res 2017 23(24); SABCS 2019 P5-05-01). Here we report results from
Methods: A prospective observational study of PAL is planned in three cohorts (A, B, and the dose finding and optimization cohorts of an ongoing phase 1/2 study (NCT05573126)
C) categorized by line of endocrine treatment (1st, 2nd, or 3rd or later line) for post- in patients (pts) with advanced AR+/ER+/HER-2- breast cancer. Methods: The study
menopausal metastatic or unresectable breast cancer. The primary endpoint is PFS in each recruited post-menopausal women with locally advanced or metastatic AR+/ ER+/HER-2-
line of treatment. For cohort B, PFS2 is defined as time from initiation of first-line therapy breast cancer, $ 18 years of age, with endocrine sensitive disease. AR+ defined as $ 10%
for metastatic disease. Based on the results of the PALOMA-2 and -3 studies, the planned AR nuclei staining by IHC. Primary objectives were to evaluate safety and determine the
sample size was set at 700 cases with confidence intervals: 340 in cohort A, 200 in cohort B optimal dose for evaluation in future combination cohorts. Other endpoints included PK,
and 130 in cohort C. Secondary endpoints include OS, clinical benefit rate, time to ORR, DOR, CBR $6 months and genomic analysis (biopsy- or ctDNA-based NGS). Re-
chemotherapy, adverse events (AEs), patient-reported outcomes and health-related quality sults: A total of 20 pts (Median age 59.5 y, PS 0/1 [70/30%]) were treated across 4 dose
of life, which will also be evaluated during follow-up. This study aims to determine whether cohorts: 20mg QD (n = 2), 10mg BID (n = 10), 10mg QD (n = 5), 15mg BID (n = 3). The 10mg
the efficacy, safety and QoL outcomes of PAL treatment in daily clinical practice are BID cohort was expanded for dose optimization. All pts received prior CDK4/6i and AI and/
comparable to those observed in clinical trials, and whether PAL affects the efficacy and or SERD with a median of 4 prior lines (in any setting). CtDNA analysis showed genomic
safety of subsequent treatments. This report presents PFS results from each cohort. An heterogeneity at baseline, with ESR1 mutations (8/19 pts) and TP53 mutations (9/19 pts)
exploratory analysis of OS rates from the start of 1st-line therapy for metastatic disease is the most frequent. No DLTs were observed. 89% of all TEAEs were G1 or G2 with most
also reported. Results: A total of 700 patients were enrolled from April 2019 to January common being increase in LFTs (55% of pts), nausea (40% of pts) and anemia (25% of
2023. After excluding cases with contraindications, the final cohort distribution was as pts). The LFT increases were transient, asymptomatic and generally occurred in cycle 1,
follows: 246 in cohort A, 282 in cohort B, and 65 in cohort C. The median PFS was 25.8 with 2 pts requiring a dose interruption followed by reduction. Most common $ G3 TEAEs
months (95% CI: 21.4) for cohort A, 18.0 months (95% CI: 14.0-22.7) for cohort B, and 12.0 were ALT increase in 4 pts (20%). No treatment related deaths were observed. 19 pts were
months (95% CI: 7.7-17.4) for cohort C. The median PFS2 for cohort B was 57.9 months evaluable for efficacy. For 11/ 19 (58%) pts the best response was stable disease. 4/19
(95% CI: 45.2-65.1). The 3-year OS rates for cohorts A and B from the start of 1st-line (21%) pts had clinical benefit with CBR $6 mo, corresponding with marked suppression of
metastatic therapy were 76.3% and 93.1%, respectively. Conclusions: The PFS result for CA15-3 in 5/19 (26%) patients. Vosilasarm has a favorable PK profile with good bio-
the 1st-line cohort (Cohort A) was nearly equivalent to the 24.8 months observed in availability and no accumulation. Full data will be reported. 10 mg BID was selected as the
PALOMA-2, while the 2nd-line cohort (Cohort B) showed markedly better results than the optimal dose for Phase 2. Conclusions: Vosilasarm has promising clinical benefit, safety
9.5 months reported in PALOMA-3. Although the background of each cohort needs to be and tolerability in this heterogeneous, heavily pre-treated population. This confirms the
further investigated, the PFS2 result of Cohort B was excellent and the subsequent 3-year potential of vosilasarm, a first in class SARM, as a new treatment strategy for AR+/ ER+/
OS of this cohort was satisfactory. Based on these results, the use of PAL in the 2nd line HER-2- breast cancer. The study is continuing with evaluation of vosilasarm in com-
setting may be clinically acceptable. Clinical trial information: UMIN000035863. Research bination with SoC therapies including oral SERD, mTOR inhibitor and CDK4/6 inhibitors.
Sponsor: Pfizer Inc. Clinical trial information: NCT05573126. Research Sponsor: Ellipses Pharma.

1058 Poster Session 1059 Poster Session

18
Impact of BMI on CDK4/6 inhibitors efficacy and safety in advanced breast Clinical utility of [ F]fluoroestradiol (FES) PET/CT to guide second-line
cancer: Results from a propensity score matched study—CAMELIA. First treatment decision in patients with ER-positive HER2-negative metastatic
Author: Min Tian, Department of Oncology, the First Affiliated Hospital of Nanjing breast cancer progressing on first-line endocrine therapy. First Author:
Medical University, Nanjing, China Hannah M. Linden, University of Washington, Seattle, WA
Background: Body mass index (BMI) is strongly associated with the development and progression of breast cancer. Despite the widespread use of cyclin-dependent kinase
(CDK) 4/6 inhibitors combined with endocrine therapy (ET) in hormone receptor (HR)-positive advanced breast cancer, the effect of BMI on therapeutic outcomes remains Background: Second line treatment options for patients with ER+/HER2- metastatic
poorly understood. Methods: Patients aged $18 years with advanced HR-positive breast cancer who received CDK4/6 inhibitors at six hospitals in China were included. 588
patients admitted between December 2016 and December 2024 were evaluated. Patients were categorized into two groups based on BMI: Group 1 (BMI , 25.0 kg/m²) and
breast cancer (MBC) after progression on 1st line endocrine therapy (ET) continue to
Group 2 (BMI $ 25.0 kg/m²). Propensity score matching with a 3:1 ratio was performed, resulting in 452 patients included in the final analysis. The median follow-up duration expand with novel endocrine agents (oral SERDs) and combination therapies (PIK3CA/
was 21.53 months. Progression-free survival (PFS) and overall survival (OS) across BMI categories were compared using Kaplan-Meier (KM) curves and log-rank test.
Univariate and multivariate cox regression analyses were performed to assess the impact of baseline clinical factors on PFS. Results: Of 452 patients, 339 (66.7%) were in AKTi, CDKi). However, short median PFS reported in clinical trials show that many patients
Group 1, 113 (33.3%) were in Group 2 at baseline. The KM analysis revealed that patients with BMI $ 25 kg/m² had a significantly longer PFS compared to those with BMI ,
25 kg/m². The median PFS was 16.77 months (95% CI: 12.86–20.67) in Group 1, versus 12.93 months (95% CI: 11.38–14.49) in Group 2 (p = 0.036, HR 0.737, 95% CI:
do not benefit from 2nd line ET. Identifying endocrine-resistant MBC at time of progression
0.554–0.981). However, no significant difference in OS was observed between the two groups (Group 1: 55.6 months vs. Group 2: not reached, p = 0.949, HR 1.015, 95% CI:
0.637–1.618). Univariate and multivariate cox regression analyses identified BMI, lymph node, liver, bone, and brain metastasis are independent prognostic factors for the
on 1st line ET can help place patients on potentially more effective non-ET options (e.g.,
entire cohort. Subgroup analyses revealed that BMI $25 kg/m² was associated with improved survival in patients aged , 60 years, Eastern Cooperative Oncology Group chemotherapy / antibody-drug conjugates). FES PET/CT has been approved for clinical use
(ECOG) performance status $ 1, with lung metastases, received $ 1 line of chemotherapy, and received $2 lines of CDK4/6 inhibitors. While Group 1 demonstrated a higher
overall response rate (ORR) (30.4% vs. 26.5%, p = 0.476), Group 2 had a higher disease control rate (DCR) (87.0% vs. 91.2%, p = 0.314), though neither reached statistical in the U.S. and allows whole-body evaluation of ER expression in MBC. Difference in FES
significance. No significant differences were found in the incidence of grade 3/4 hematologic adverse events (AEs) (36.9% vs. 31.5%, p = 0.460) or non-hematologic AEs
(17.1% vs. 14.4%, p = 0.605). Conclusions: In this study, overweight patients (BMI $ 25 kg/m²) with metastatic breast cancer may benefit more from CDK4/6 inhibitors.
uptake across lesions may reflect ER loss/downregulation, a mechanism of endocrine
Moreover, similar adverse events were observed across BMI groups. These findings suggest that BMI could serve as a key predictor of CDK4/6 inhibitors treatment response, resistance. Goal of this clinical trial was to evaluate the impact of FES PET/CT results on
2nd line therapeutic management decisions. Methods: In this multicenter trial in the U.S.
providing valuable insights for personalized therapeutic strategies in metastatic breast cancer. Research Sponsor: None.

[NCT05068726], patients with progression of ER+/HER2- MBC on 1st line ET were pro-
Characteristics of patients before and after matching according to BMI categories.

Unmatched cohort Matched cohort

Characteristics
BMI<25kg/m2
(n=448)
BMI‡25kg/m2
(n=140) P SMD
BMI<25kg/m2
(n=339)
BMI‡25kg/m2
(n=113) P SMD
spectively enrolled to undergo FES PET/CT in addition to standard of care (SOC) imaging
Age grope at study entry, No. (%) 0.134 0.152 0.694 0.036 (CT + bone scan / FDG PET/CT). Treating oncologists completed questionnaires before and
＜60years 286 (63.8) 79 (56.4) 204 (60.1) 70 (61.9)
‡60years
ECOG PS
162 (36.2) 61 (43.6)
0.064 0.189
135 (39.8) 43 (38.1)
0.350 0.078
after FES PET/CT, detailing therapeutic management plans plus their confidence in the
0
‡1
Stage at diagnosis
248 (56.4)
240 (53.6)
52 (37.1)
88 (62.9)
0.217 0.124
149 (43.7)
191 (56.3)
45 (39.8)
68 (60.2)
0.406 0.072
plans. FES PET/CT scans were compared with SOC imaging to assess FES uptake in MBC
I-III
IV
Estrogen receptor status, No. (%)
368 (82.1)
80 (17.9)
108 (77.1)
32 (22.9)
0.131 0.226
279 (82.3)
60 (17.7)
96 (84.7)
17 (15.0)
0.434 0.070
lesions. An FES uptake score (number of FES-positive lesions divided by total number of
1~10%
10~50%
9 (2.0)
37 (8.3)
0 (0.0)
16 (11.4)
0 (0.0)
29 (8.6)
0 (0.0)
12 (10.6)
lesions per patient) was calculated to evaluate ER expression heterogeneity by central
＞50% 402 (89.7) 124 (88.6) 310 (91.4) 101 (89.4)
Progesterone receptor status, No. (%)
Negative 65 (14.5) 21 (15.0)
0.328 0.150
55 (16.2) 19 (16.8)
0.940 0.030 blinded image evaluation. Results: 45 patients underwent FES PET/CT. FES PET/CT
1~20%
＞20%
HER-2 status, No. (%)
121 (27.0)
262 (58.5)
29 (20.7)
90 (64.3)
0.052 0.204
75 (22.1)
209 (61.7)
26 (23.0)
68 (60.2)
0.812 0.053
results led to a change in therapeutic management in 17/45 patients (37.8%; 95% CI 23.8%
Negative
Positive
418 (93.3)
23 (5.1)
124 (88.6)
9 (6.4)
311 (91.7)
21 (6.2)
104 (92.0)
6 (5.3) - 53.5%). Revised management plan was ET in 5/17 and non-ET in 12/17 patients. FES
uptake score was 1 (all lesions FES-positive) in 14/45 patients and , 1 (with FES-negative
Unknown 7 (1.6) 7 (5.0) 7 (2.1) 3 (2.7)
Ki-67 status, No. (%) 0.138 0.184 0.475 0.097
＜20 166 (37.1) 46 (32.9) 120 (35.4) 36 (31.9)
‡20
Unknown
Resistance to previous endocrine treatment, No. (%)
246 (54.9)
36 (8.0)
75 (53.6)
19 (13.5)
0.085 0.234
188 (55.5)
31 (9.1)
64 (56.6)
13 (11.5)
0.924 0.031
lesions, indicating ER expression heterogeneity) in 31/45 patients. Of 14 patients with FES
Primary resistance
Secondary resistance
ET naı̈ve
80 (17.9)
325 (72.5)
42 (9.4)
27 (19.3)
90 (64.3)
23 (16.4)
61 (18.0)
239 (70.5)
39 (11.5)
20 (17.7)
81 (71.7)
12 (10.6)
uptake score of 1, 11 received 2nd line ET. Of 31 patients with FES uptake score , 1, 15
non-sensitive
Bone metastases
1 (0.2) 0 (0.0)
0.281 0.107
0 (0.0) 0 (0.0)
0.816 0.024 were treated with ET and 16 with non-ET, based on guidelines suggesting that an FES-
No 181 (40.4) 64 (45.7) 151 (44.5) 49 (44.1)
Yes
Visceral metastases
267 (59.6) 76 (54.3)
0.432 0.084
188 (55.5) 64 (56.6)
0.310 0.084
negative lesion is predictive of lack of endocrine response. FES PET/CT results led to 25/45
No
Yes
Previous endocrine therapy lines, No. (%)
183 (40.8)
265 (59.2)
63 (45.0)
77 (55.0)
0.136 0.198
146 (43.1)
193 (56.9)
44 (40.5)
69 (61.1)
0.673 0.068
patients avoiding additional tests (20 biopsies, 5 scans) and 7/45 patients receiving further
0
1
206 (46.0)
140 (31.2)
77 (55.0)
40 (28.6)
170 (50.1)
100 (29.5)
59 (52.2)
34 (30.1) testing (4 biopsies, 1 scan, 2 other). Treating oncologist’s confidence in 2nd line treatment
‡2 102 (22.8) 23 (16.4) 69 (20.4) 20 (17.7)
CDK4/6 inhibitors treatment lines, No. (%)
1 155 (34.6) 68 (48.6)
0.014 0.286
140 (41.3) 50 (44.2)
0.696 0.065
decision (measured in n = 45 on 10-point scale with 10 being fully confident) increased on
2 90 (20.1) 22 (15.7) 60 (17.7) 20 (17.7)
‡3
Combination of CDK4/6 inhibitors therapy, No. (%)
203 (45.3) 50 (35.7)
0.644 0.070
139 (41.0) 43 (38.1)
0.943 0.028
average with 2 points from 6.6 (SD = 1.7) pre-FES PET/CT to 8.6 (SD = 1.8) post-FES PET/
Aromatase inhibitors
Fulvestrant
Others
270 (60.3)
172 (38.4)
6 (1.3)
86 (61.4)
51 (36.5)
3 (2.1)
202 (59.6)
132 (38.9)
5 (1.5)
68 (60.2)
43 (38.1)
2 (1.8)
CT. Conclusions: FES PET/CT is a clinically useful tool in the post-first line ER+/HER2-
Disease-free survival, No. (%)
£2years 75 (16.7) 23 (16.4)
0.417 0.126
54 (15.9) 21 (18.6)
0.478 0.092
MBC setting. FES PET/CT results led to a change in management in 37.8% of patients and
increased oncologist’s confidence in 2nd line treatment decision. Clinical trial information:
>2years 293 (65.4) 85 (60.7) 225 (66.4) 75 (66.4)
De novo stage IV 80 (17.9) 32 (22.9) 60 (17.7) 17 (15.0)

BMI: body mass index; ECOG: Eastern Cooperative Oncology Group; HER-2: human epidermal growth factor receptor 2; ET: endocrine treatment; CDK4/6: cyclin-dependent kinase 4/6.
NCT05068726. Research Sponsor: Zionexa SAS, a GE HealthCare Company.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 51s

1060 Poster Session 1061 Poster Session

Imlunestrant with or without abemaciclib in advanced breast cancer (ABC): Giredestrant (G) with atezolizumab (ATEZO), and/or abemaciclib (ABEMA) in
Safety analyses from the phase III EMBER-3 trial. First Author: Joyce O’Sh- patients (pts) with ER+/HER2– locally advanced/metastatic breast cancer
aughnessy, Baylor University Medical Center, Texas Oncology, Dallas, TX and Sarah (LA/mBC): Interim analysis (IA) from the phase I/II MORPHEUS Breast
Cannon Research Institute, Dallas, TX Cancer study. First Author: Mafalda Oliveira, Vall d’Hebron University Hospital,
Background: Imlunestrant is a next-generation, brain-penetrant, oral SERD. The EMBER-3 trial Barcelona, Spain
(NCT04975308) in patients with ER+, HER2- ABC and disease progression on/after aromatase inhibitor Background: Endocrine therapy (ET) + a cyclin-dependent kinase 4/6 inhibitor (CDK4/6i) is a
therapy showed significant progression-free survival improvement with imlunestrant (imlu; 400 mg once therapeutic mainstay for first-line treatment (tx) of ER+ mBC, but selection of effective ET
daily [QD]) over standard therapy (SOC, fulvestrant or exemestane) among patients with ESR1 mutations, combinations after progression remains a challenge. G is a highly potent, non-steroidal, oral (PO),
as well as with imlunestrant+abemaciclib (imlu [400 mg QD] + abema [150 mg twice daily]) over imlu in all
selective ER antagonist and degrader shown to be well tolerated and to achieve robust ER
patients regardless of ESR1 mutation status. Detailed safety analyses are presented. Methods: The
occupancy. Immune checkpoint inhibition has shown a trend towards activity in a number of ER+
safety population included all patients who received at least one dose of study treatment. Analyses
included incidence, severity (CTCAE v 5.0), management, and outcomes of common treatment-emergent BC studies. Additionally, ABEMA (a CDK4/6i) has immunomodulatory activity, making its inclusion
adverse events (TEAEs). Results: Safety analyses included 859 patients: imlu (n=327), SOC (n=324), and a compelling therapeutic approach. Here, we present a 24-week IA of the G + ATEZO 6 ABEMA
imlu+abema (n=208). Incidence of any (imlu: 83%; SOC: 84%; imlu+abema: 98%), $ grade 3 TEAEs (imlu: arms and a 22-week IA of the G + ABEMA arm from MORPHEUS BC (NCT04802759).
17%; SOC: 21%; imlu+abema: 49%), and serious AEs (SAEs; imlu: 10%; SOC 12%; imlu+abema: 17%) were Methods: Eligible pts had ER+, HER2– LA/mBC and had received prior tx with a CDK4/6i and 1–2
similar between imlu and SOC arms and higher in the combination arm. Most common any-grade AEs with lines of ET. Pts were randomized to G (30 mg PO QD) alone (previously reported), G + ATEZO (840
imlu were diarrhea (21%), nausea (17%), and fatigue (23%) and with imlu+abema were diarrhea (86%), mg IV Q2W), G + ATEZO + ABEMA (150 mg PO BID), or G + ABEMA until loss of clinical benefit/
nausea (49%), and neutropenia (48%); majority were grade 1 AEs. Incidence of elevated transaminases unacceptable toxicity. Investigational drug doses were identical across all arms. Primary end-
(any%/$G3%: 16/1 and 20/5), VTE (1/0 and 3/1), ILD (1/0 and 2/0), bradycardia (2/0 and 1/0), and points were safety and objective response rate (ORR). Exploratory analyses included evaluation of
photopsia (0/0 and 0/0) were relatively low or not observed with imlu and imlu+abema, respectively. Dose circulating tumor DNA alterations and tumor gene expression using RNAseq. Results: As of Apr
reduction rates were 2% with imlu and 39% with imlu+abema, and discontinuation rates due to AEs were 24, 2024, 15 pts in the G + ATEZO arm, 30 in the G + ATEZO + ABEMA arm and, as of Jan 9, 2023, 15
low (4% and 6%, respectively). The table characterizes the most commonly observed AEs. Further details in the G + ABEMA arm, were efficacy/safety evaluable. Many pts had prior fulvestrant (60%, 43%,
will be presented. Conclusions: Imlunestrant had a favorable safety profile, similar to SOC, with mostly and 27%, respectively) and prior CDK4/6i duration $12 mo (73%, 77%, and 53%). Safety data are
grade 1 AEs. Safety of imlunestrant + abemaciclib was consistent with the known abemaciclib profile, shown in the Table. In the triplet arm, the most common grade $3 adverse event (AE) was
without additive toxicity. AEs were manageable with supportive medications and/or dose adjustments,
neutropenia/neutrophil count decreased (20%). No grade 5 AEs were reported. Confirmed ORR (all
resulting in few discontinuations in all arms. Imlunestrant, as monotherapy or in combination with
partial responses) were 20%, 33%, and 7% in the G + ATEZO, G + ATEZO + ABEMA, and the G +
abemaciclib, provides a safe, tolerable, all-oral targeted therapy option for patients with ER+, HER2- ABC.
Clinical trial information: NCT04975308. Research Sponsor: Eli Lilly and Company. ABEMA arms, respectively. 7/9 confirmed responses (in ESR1-evaluable pts) in the G + ATEZO +
ABEMA arm were in pts with ESR1-mutated disease. Data with longer follow-up, including
Characterization of commonly observed AEs. progression-free survival, detailed safety, and exploratory biomarker analyses, will be presented.
Diarrhea Nausea Conclusions: The combinations of G + ATEZO, G + ATEZO + ABEMA, and G + ABEMA were
Imlu SOC Imlu+abema Imlu SOC Imlu+abema tolerable, with no unexpected safety signals including no high-grade interstitial lung disease/
N=327 N=324 N=208 N=327 N=324 N=208 pneumonitis and low rates of high-grade liver toxicity. Clinical activity was observed, with a trend
Grade 1 AE, % 18 9 50 14 8 31 towards improved ORR with G + ATEZO + ABEMA, particularly in tumors with ESR1 mutations.
Grade 2 AE, % 3 3 28 3 5 15 Clinical trial information: NCT04802759. Research Sponsor: F. Hoffmann-La Roche Ltd.
Grade ‡3 AE, % 0.3 0 8 0.3 0 2
Median time to onset 30 52 5 20 57 15 G + ATEZO G + ATEZO + G + ABEMA
(Q1–Q3), days (15–129) (17–132) (2–17) (4–56) (10–147) (3–48) n (%) (n = 15) ABEMA (n = 30) (n = 15)
Median duration of Grade 2 3 5 13 16 10 19
AE (range), days (1–28) (1–55) (1–87) (4–89) (1–90) (2–266) Any AE 14 (93) 30 (100) 15 (100)
Median duration of Grade ‡3 8 0 9 24 0 7 AE: highest grade 3 8 (53) 15 (50) 8 (53)
AE (range), days (8–8) (1–47) (24–24) (6–13) AE: highest grade 4 0 (0) 0 (0) 1 (7)
Dose reduction/discontinuation, % 0/0 0/0 18/1 0.3/0 0/0 5/0 Any-grade tx-related AE (TRAE) 12 (80) 30 (100) 13 (87)
Antidiarrheal medication/ 10 7 68 10 10 21 TRAE leading to discontinuation of any tx 3 (20) 5 (17) 0 (0)
Antiemetic, %

1062 Poster Session 1063 Poster Session

Phase Ib study of inavolisib (INAVO) + weekly paclitaxel (wP) in patients PARPi effectiveness after CDK4/6i in BRCA1- and BRCA2-associated HR+/
(pts) with locally advanced/metastatic (LA/m) incurable solid tumors: HER2- advanced breast cancer: Results from the multicenter real-world
Safety, pharmacokinetics (PK), and preliminary antitumor activity. First PAMBRACA study. First Author: Emma Zattarin, University of Modena, Modena, MO,
Author: Seock-Ah Im, Seoul National University College of Medicine, Cancer Research Italy
Institute, Seoul National University, Seoul, South Korea Background: Poly(adenosine diphosphate–ribose) polymerase inhibitors (PARPi) are the
Background: wP is commonly used for treating solid tumors as a single agent or in paramount of personalized therapy for BRCA1 and BRCA2 pathogenic/likely pathogenic
combination with targeted agents. However, it has an unfavorable benefit–risk profile when variant (P/LPV) carriers with hormone receptor-positive (HR+)/HER2-negative (HER2-)
given with pan-PI3K inhibitors or alpelisib. INAVO, a potent and selective PI3Ka inhibitor that advanced breast cancer (aBC). Nevertheless, data on the efficacy of PARPi following
also promotes mutated p110a degradation, was FDA approved in combination with pal- cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) combined with endocrine therapy (ET) are
bociclib + fulvestrant for hormone receptor-positive, HER2-negative (HR+, HER2–), endo- limited. Methods: The PAMBRACA study is a multicenter, hospital-based, retrospective-
crine-resistant advanced breast cancer (BC) following recurrence on/after completing prospective cohort study enrolling BRCA1 and BRCA2-P/LPV carriers with HR+/HER2- aBC
adjuvant endocrine therapy. We report data from INAVO + wP in pts with LA/m solid tumors treated with ET+CDK4/6i and/or PARPi. In this analysis, the real-world Progression-Free
from a Phase Ib study (CO42800; ISRCTN45319897). Methods: Eligible pts had progressed Survivals (rwPFS) of ET+CDK4/6i and subsequent lines were evaluated through Kaplan-
after standard systemic therapy. In part 1 (dose-escalation phase; 3+3 design), pts with LA/m Meier method and compared with the log-rank test. Median follow-up was calculated using
incurable solid tumors received INAVO 6 mg/9 mg orally daily (PO QD) + wP (80 mg/m2). In the reverse Kaplan-Meier method. Multivariate Cox regression model was used to adjust
part 2 (dose-expansion phase), pts with LA/m incurable PIK3CA-mutated solid tumors (triple- the association between treatment regimens and rwPFS for clinically relevant variables.
negative BC [TNBC]; HR+, HER2– BC; others) received INAVO 9 mg PO QD (recommended Results: We included12 BRCA1 and 57 BRCA2-P/LPV carriers who were diagnosed with
dose from part 1) + wP. Primary endpoint: Safety/tolerability in parts 1 and 2. Secondary HR+/HER2- aBC between January 1998 and December 2023 in six Italian Institutions. All
endpoints: Preliminary antitumor activity in part 2 (only TNBC and HR+, HER2– BC data are the patients (pt) received CDK4/6i+ET for aBC (85.5% as first line, 7.2% as second line, 7.3%
available); PK in parts 1 and 2. Results: Of 66 pts enrolled (parts 1 and 2), four received no as third or subsequent line). At CDK4/6i starting, median age was 45 years (range 28-80);
treatment and eight were still on treatment at clinical cutoff (Oct 11, 2024). Reasons for study 52.2% of pts had visceral metastases and 17.4% had de novo aBC. Median follow-up was
discontinuation were per protocol study completion (56.1%), death (16.7%), pt withdrawal 39.5 months (mo). Among pts treated with CDK4/6i as first or second line, median rwPFS
(9.1%), loss to follow-up (1.5%), and other (4.5%). There were no dose-limiting toxicities. In was 15.1 mo (95%CI 11.8-18.5) and 3.1 mo (95%CI 2.1-NA), respectively. Among the 49
safety-evaluable pts (n = 62), grade 3, 4, and 5 adverse events (AEs) occurred in 59.7%, 3.2%,
patients who progressed to first or second-line CDK4/6i, 17 (34.7%) received a PARPi as
and 0% of pts, respectively. One pt discontinued INAVO due to AEs; INAVO dose modifications
first line post-CDK4/6i, 12 (24.5%) a monochemotherapy (monoCT), 8 (16.3%) an ET (+/-
(reduction/interruption) due to AEs occurred in 61.3% of pts. The most common AEs ( . 10%
everolimus), 8 (16.3%) a polychemotherapy (polyCT) and 4 (8.2%) died without receiving a
of pts) were diarrhea (61.3%), hyperglycemia (51.6%), and anemia (45.2%). Neutropenia
subsequent line. No significant differences in clinicopathological characteristics were
(24.2%) and diarrhea (8.1%) were the most common grade 3–4 AEs. Serious AEs occurred in
observed among the treatment groups, except for the number of metastatic sites (,3 vs .
30.6% of pts (mostly single AEs in individual pts, and unrelated to study treatment).In part 2,
confirmed overall response rate in pts with TNBC (n = 20) was 50.0% and in pts with HR+, 3), which was higher for pts receiving mono/polyCT (p = 0.053). PARPi treatment was
HER2– BC (n = 19) it was 36.8%. Median duration of confirmed response was 7.4 mo (95% associated with significantly higher median rwPFS (13 mo vs 4.5 mo for monoCT vs 3 mo
confidence interval 5.2, 11.5) and 12.8 mo (3.7, not evaluable), respectively; median pro- for ET vs 6 mo for polyCT, p , 0.001), also after adjusting for the number of metastatic
gression-free survival, 7.0 mo (3.5, 9.3) and 7.4 mo (6.2, 14.7). PK of INAVO and wP at Cycle 1, sites [for PARPi vs other lines, adjusted hazard ratio (aHR) 0.20, 95%CI 0.09-0.49, p ,
Day 15 were comparable to historic data. Conclusions: In CO42800, INAVO + wP was well 0.001]. 17 pts received PARPi as later treatment lines, which were independently asso-
tolerated in pts with LA/m solid tumors, including those with a PIK3CA mutation, with no new ciated with lower median rwPFS vs PARPi as first post-CDK4/6i line (6 vs 13 mo, aHR 2.81,
safety signals or drug–drug interactions observed. Encouraging preliminary antitumor ac- 95%CI 1.15-6.90, p = 0.024). Conclusions: AfterCDK4/6i+ET, PARPi were independently
tivity shown in pts with HR+, HER2– BC or TNBC supports further investigation. Clinical trial associated with longer rwPFS compared to other systemic therapies in BRCA1 and BRCA2-
information: ISRCTN45319897. Research Sponsor: Genentech, Inc. P/LPV carriers with HR+/HER2- aBC. Earlier PARPi use after CDK4/6i was associated with
greater clinical benefit. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
52s BREAST CANCER—METASTATIC

1064 Poster Session 1065 Poster Session

A phase I/IIa study to evaluate the tolerability, safety, pharmacokinetics and Quantifying the clinical impact of tissue reflex testing for liquid biopsy ESR1
efficacy of eciruciclib (BPI-1178) alone in advanced solid tumors and in mutation–negative cases with low ctDNA tumor fraction (TF) in HR(+)
combination with endocrine therapy for advanced or recurrent HR+/HER2- HER2(-) breast cancer. First Author: Jing Du, Yale School of Medicine, New Ha-
breast cancer. First Author: Yiqun Du, Phase I Cinical Trial Center, Fudan University ven, CT
Shanghai Cancer Center, Shanghai, China Background: ESR1 mutations (ESR1mut) commonly drive acquired resistance to estrogen
Background: Eciruciclib (BPI-1178) is a new cyclin-dependent kinases (CDKs) 2/4/6 inhibitor, deprivation by aromatase inhibitors, a first-line standard of care for HR(+)HER2(-) meta-
which has shown strong inhibition on the expression of CDK2/4/6 in pre-clinical studies. This static breast cancer (MBC). We previously published that approximately 63% of patients with
first-in-human phase I/IIa study aimed to assess the preliminary efficacy, safety and toler- HR(+)HER2(-) MBC at progression have a liquid biopsy (LBx) negative for ESR1mut. The
ability of eciruciclib monotherapy for advanced solid tumors or in combination with endocrine absence of an ESR1mut may either accurately reflect the tumor genotype (true negative) or
therapy (ET) for HR+/HER2- advanced breast cancer (ABC). Methods: Patients with advanced represent a false negative due to insufficient ctDNA shedding, with the risk of missing an
solid tumors included in phase I received eciruciclib alone at doses of 25~500 mg in a 3+3 actionable mutation. Among these patients, 40% exhibit a high ctDNA TF $1% (informative
dose-escalation or expansion manner. Phase IIa consisted of two cohorts, A and B. Cohort A negative), while 60% have a low ctDNA TF , 1% (indeterminate negative). This suggests
included patients with HR+/HER2- ABC who had progressed after ET receiving eciruciclib in that up to 38% of all patients with HR(+)HER2(-) MBC in this context could potentially benefit
combination with fulvestrant, and treatment-naive patients with HR+/HER2- ABC in cohort B from reflex tissue biopsy (TBx) for ESR1mut in cases deemed indeterminate negative by LBx
were treated with eciruciclib in combination with letrozole. All patients administered eciruciclib due to low ctDNA shedding. The goal of this study is to determine the rate of ESR1mut
with either intermittent (21 days on, 7 days off) or continuous (28 days on) dosing schedule in a detection in a new TBx after an indeterminate negative result from FoundationOne Liquid
28-day cycle until disease progression, unacceptable toxicity, etc. Safety was assessed as per CDx (F1LCDx). Methods: This study included a cohort of patients with BC who underwent
CTCAE 5.0. Efficacy endpoints included confirmed objective response rate (ORR), disease tissue and liquid Foundation Medicine comprehensive genomic profiling (CGP) within an
control rate (DCR), progression-free survival (PFS), etc. assessed by investigators per RECIST interval of up to 90 days during routine clinical care. Clinical data of a subset of patients with
1.1. Results: As of August 9, 2024, a total of 129 patients have been enrolled. And 33 patients confirmed HR(+)HER2(-) MBC was obtained from the US-wide deidentified Flatiron Health-
were enrolled in Phase l study. No DLT was observed. In Phase IIa Cohort A, 70 patients were Foundation Medicine MBC clinicogenomic database (CGDB). The data originated from ~280
enrolled and in which 64 patients were evaluable for efficacy. 26 patients were enrolled into cancer clinics (~800 sites of care) between 01/2014-09/2024. False negative rate (FNR) and
cohort B with 25 efficacy-evaluable patients. In Cohort A, the top three treatment related positive percent agreement (PPA) for ESR1mut detection were calculated with tissue CGP as
adverse events (TRAEs) of grade $ 3 were neutrophil count decreased (35.7%), white blood reference. Results: A total of 522 BC patients underwent TBx and LBx. Among these, 229
cell count decreased (14.3%), and hypertriglyceridemia (14.3%) while in Cohort B those were
(43.9%) had ctDNA TF , 1%. Without accounting for TF, the overall FNR for ESR1mut was
neutrophil count decreased (23.3%), hypertriglyceridemia (16.7%), alanine aminotransferase
6.3% and the PPA was 67.1%. In LBx with TF $1%, the FNR for ESR1mut was 0.9% and PPA
increased (10.0%), and white blood cell count decreased (10.0%). No TRAE leading to per-
was 96.0%. In contrast, for TF , 1% samples, the FNR was 12.0% and the PPA 25.7%. 101
manent discontinuation or death occurred in this trial. Conclusions: Eciruciclib in combination
patients were included in the CGDB and had a confirmed HR(+)HER2(-) MBC, in which 56
with ET demonstrated promising efficacy and manageable safety profile in patients with HR+/
HER2- ABC. Clinical trial information: NCT04282031. Research Sponsor: Beta Pharma (55.4%) had LBx with ctDNA TF , 1%. The overall FNR for ESR1mut in this subset of patients
(Suzhou) Co., Ltd. was 9.5% and the PPA was 61.9%. In LBx with TF $1%, the FNR was 0% and PPA was 100%.
And for TF , 1%, the FNR was 15.1% and PPA 20.0%. Conclusions: BC patients with
Cohort A (n = 64) Cohort B (n = 25) informative negative ESR1mut (defined as LBx ESR1mut negative with TF $1%) are unlikely
400 mg a 300 mg a 300 mg b 200 mg b 400 mg a 300 mg a to have ESR1mut detected on tissue CGP testing. However, patients with indeterminate
(n = 20) (n = 16) (n = 20) (n = 8) (n = 19) (n = 6) negative ESR1mut (defined as LBx ESR1mut negative with TF , 1%), 12-15% were found to
ORR, n (%) 9 (45.0) 7 (43.8) 11 (55.0) 0 15 (78.9) 5 (83.3) be false negatives. This suggests that approximately 5% of all HR(+)HER2(-) MBC patients
DCR, n (%) 17 (85.0) 13 (81.3) 20 (100.0) 8 (100.0) 18 (94.7) 6 (100.0) with ESR1mut could be missed without reflex testing with a TBx. ctDNA TF levels offer
Median PFS, 18.3 7.3 NR NR NR NR
months (95% CI) (7.2, NR) (2.4, NR) (12.8, NR) (16.7, NR) critical guidance in deciding when a reflex to tissue is warranted, ensuring accurate
a
treatment. Research Sponsor: None.
intermittent dosing schedule;
b
continuous dosing schedule; NR, not reached.

1066 Poster Session 1067 Poster Session

Thymidine kinase activity (TKa) as independent predictor of outcome in Prognostic role of estrogen receptor (ER) expression in breast cancer (BC)
metastatic breast cancer (MBC) patients in the GEICAM/2013-02 PEARL metastases and its dynamics from primary to metastatic disease: Results
trial. First Author: Angel Guerrero, Instituto Valenciano de Oncologı́a (IVO). GEICAM from a large multicentric cohort of patients with phenotypically stable
Spanish Breast Cancer Group, Valencia, Spain ER+(‡10%)/HER2- BC. First Author: Federica Miglietta, University of Pado-
Background: TKa is a proliferation biomarker measurable in blood via the DiviTum™ TKa vaOncology 2 Unit, Istituto Oncologico Veneto IRCCS, Padova, Italy, Italy
assay. Levels of TKa before and during treatment can provide prognostic, predictive and Background: ER expression is one of the main determinants of prognosis in patients (pts) with BC.
monitoring information in MBC. The PEARL trial (NCT02028507) was a phase III, multicenter, Phenotypic conversion from ER+ (ER.=10%)/HER2- primary BC towards ER,10%/HER2- advanced
open-label, randomized study that compared endocrine therapy (ET) + CDK4/6 inhibitor BC has a well-known negative impact on outcome. However, in the specific context of phenotypically
Palbociclib (Palbo) vs. Capecitabine (Cape) in aromatase inhibitor-resistant HR+/HER2- MBC stable ER+/HER2- BC, the prognostic impact of ER expression in metastases or ER dynamics during
patients (pts). ET + Palbo did not improve median progression-free survival (mPFS 17.8 vs. disease evolution, remains largely understudied. Methods: We enrolled pts with advanced BC un-
dergoing biopsy of a metastatic site. ER+ was defined as ER.=10%. ER expression was evaluated both
17.3 months (m.), p = 0.9) or overall survival (mOS 31.1 vs. 32.8 m., p = 0.5) over Cape. We
as continuous and categorical variable (categories: 10-30%, 30-50%, 50-100%). Overall survival (OS)
explored whether TKa levels could predict better response to ET + Palbo vs Cape. was the primary study endpoint. Cox multivariable models included covariates associated with OS in
Methods: Plasma from 555 pts (92%) was collected at baseline (BL) and on treatment univariate analysis. Results: Among 1114 pts, 410 had ER+/HER2- phenotype in both primary and
(C1D15, C2D15). 1129 samples were analyzed using the DiviTum TKa assay (FDA approved/ metastatic tumor specimens. In this subgroup, ER expression (both continuous and categoric) in
CE labelled, Biovica, Sweden). Cutoffs: 250/400 DiviTum units of Activity (DuA) for BL, 50 metastases had significant prognostic value: for each 10% lower ER expression, the risk of death
DuA or fold change (C1,C2/BL) . 2 for on-treatment. The Kaplan-Meier method estimated increased by 6.7% (p=0.011). Pts with ER 10-30% had significantly worse OS than those with ER 50-
median PFS and OS. Adjusted hazard ratio (HR) with 95% confidence interval (CI) were 100% (HR 0.62, p=0.023), and numerically shorter than ER 30-50% (HR 0.51, p=0.063). The table shows
calculated using Cox proportional hazards regression model, considering relevant prognostic the evolution of ER categories from primary BC to metastases. ER expression dynamics also had
clinical variables. Results: BL TKa # 250 DuA predicted better mPFS (11.4 vs. 4.04 m., aHR prognostic impact. Regarding continuous ER expression changes in paired primary vs. metastatic BC,
2.1; 95% CI 1.7-2.6, p , 0.0001) and mOS (38.47 vs. 17.31 m., aHR 3.2; 95% CI 2.45-4.19, p , each 10% ER increase corresponded to 5.9% decrease in the risk of death (p=0.008). Pts whose tumors
0.0001) regardless of therapy. After starting therapy, Cape and ET + Palbo elicited distinct showed an increased ER expression from 10-30% to 50-100% had the most favorable outcome overall,
with better OS compared to pts with persistently low ER levels (HR 6.73, p=0.002), or to pts with
TKa responses due to their different mechanisms. At C1, C2, pts on Cape had higher mTKa vs
decreased ER expression in metastases - particularly pts whose ER levels dropped from 50-100% to
ET + Palbo (448 vs. 28 DuA, p , 0.0001). In the CT arm, an increase of TKa at C1 or C2 greater 10-30% (HR 2.89, p=014). These pts also had superior OS when compared to pts with persistently high
than 2-fold from BL predicted for better mPFS (13.04. vs. 6.34 m., aHR 0.59; 95% CI 0.43- ER levels (HR 2.08, p=0.043). The prognostic impact was preserved at the multivariate analysis
0.81, p = 0.0013) and mOS (39.26 vs. 23.23 m., aHR 0.31; 95% CI 0.2-0.5, p , 0.0001). In the (including age, grade, visceral/non-visceral disease, biopsy site). Conclusions: Intratumor ER ex-
ET + Palbo arm, a TKa at C1 or C2 . 50 DuA predicted a shorter mPFS (3.68 v 11.27 m, aHR pression and dynamics may in part explain the prognostic heterogeneity of pts with ER+/HER2- stable
2.81; 95% CI 2.08-3.8, p , 0.0001) and mOS (18.73 vs. 45.11 m., aHR 3.44; 95%CI 2.34-5.06, phenotype from primary to advanced BC. Pts with lower ER levels in metastases are prognostically
p , 0.0001). Similar results are observed regardless of BL TKa value. Exploring a BL TKa . disadvantaged. Dynamic changes in ER expression provide additional insights beyond those captured
400 DuA demonstrated a better response to Cape compared to ET+ Palbo, despite overall by single-point assessment. Interestingly, pts whose tumors shifted from ER 10–30% to ER 50–100%
very poor outcomes: mPFS 4.04 m on Cape vs 2.01 on ET + Palbo, (aHR 1.72; 95% CI 1.14- showed the most favorable prognosis, even outperforming those with consistently high ER levels.
2.59, p , 0.0096), and showed a similar trend in mOS, 15.4 m on Cape vs 14.6m on ET+Palbo, Research Sponsor: None.
(aHR 1.29 95% CI 0.84-1.99, p = 0.24). Conclusions: These data demonstrate that CT vs a Metastases, ER
CDK4/6 inhibitor influence TKa response differently, and the direction and magnitude of the
10-30% 30-50% 50-100% Total
TKa response can predict for benefit to a specific therapy. The original PEARL study analysis
showed no outcome differences between Cape vs ET + Palbo in HR+/HER2- MBC pts, Primary BC, ER n % n % n % n %
10-30% 5 1.2 0 0 18 4.4 23 5.6
however assessment of TKa before and during therapy identified which patients had the 30-50% 7 1.7 4 1.0 11 2.7 22 5.4
highest probability of responding. Utilization of TKa as a predictive biomarker may allow for 50-100% 24 5.9 21 5.1 320 78.0 364 89
better personalized treatment selection. Clinical trial information: NCT02028507. Research Total 36 8.8 25 6.1 349 85.1 410 100
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 53s

1068 Poster Session 1069 Poster Session

SIM0270 in combination with palbociclib in patients with ER+/ HER2- First-line (1L) ribociclib (RIB) + endocrine therapy (ET) vs combination
advanced breast cancer: The phase Ib study. First Author: Jiong Wu, Fudan chemotherapy (combo CT) in clinically aggressive hormone receptor
University Shanghai Cancer Center, Shanghai, China (HR)+/HER22 advanced breast cancer (ABC): A subgroup analysis of pa-
Background: SIM0270 is a highly potent oral selective estrogen receptor degrader tients (pts) with or without liver metastases (mets) from RIGHT Choice. First
(SERD) which has shown ER degradation and robust antitumor activity across variety of Author: Nagi S. El Saghir, American University of Beirut Medical Center, Beirut, Lebanon
preclinical models. Here, we present results of SIM0270 combined with palbociclib Background: The phase II RIGHT Choice trial reported a statistically significant progression-free
cohort(dose escalation and dose expansion) from Phase I study in patients with ER+/ survival (PFS) benefit at the primary prespecified analysis and a 9-month (mo) benefit at the final
HER2- advanced breast cancer (NCT05293964). Methods: Patients with ER+/HER2- analysis with 1L RIB + ET over combo CT in pts with clinically aggressive HR+/HER2– ABC. As liver
mets in ABC indicate a worse prognosis, an analysis by liver mets status was performed (data cutoff:
advanced breast cancer were enrolled. The key inclusion criteria for dose escalation and
May 10th, 2023). Methods: Pre- and perimenopausal women (N = 222) with no prior systemic
dose expansion were the same as follows: $ 1 prior endocrine therapy (ET) with disease therapy for clinically aggressive HR+/HER22 ABC were randomized 1:1 to receive RIB + letrozole or
recurrence/ progression while being treated with adjuvant ET for $ 24 months and/or anastrozole + goserelin or physician’s choice of combo CT. Enrolled pts had ABC for which combo CT
first line ET for $ 6 months in advanced setting; #2 prior chemotherapies in advanced was clinically indicated by physician’s judgment. Results: In total, 107 (RIB + ET, n = 54; combo CT, n
setting; and prior fulvestrant was allowed. A Bayesian Optimal Interval design (BOIN) = 53) and 115 (RIB + ET, n = 58; combo CT, n = 57) pts presented with or without liver mets,
was adopted for dose escalation. The key endpoint of dose escalation was dose limiting respectively. Pts with liver mets had PFS of 18.3 vs 12.7 mo and median time to treatment failure
toxicities (DLT), and the key endpoints of dose expansion included safety and toler- (mTTF) of 13.2 vs 8.3 mo with RIB + ET vs combo CT, respectively (Table). A clinical benefit rate
ability, pharmacokinetics (PK) and efficacy. Results: As of December 26, 2024, 44 (CBR) of 77.8% vs 67.9%, overall response rate (ORR) of 64.8% vs 60.4%, and median time to
patients were enrolled including 12 from dose escalation and 32 from dose expansion, response (mTTR) of 6.4 vs 3.0 mo were seen in the RIB vs CT arm, respectively. Pts without liver mets
had PFS of 25.2 vs 15.4 mo and mTTF of 24.0 vs 10.1 mo in the RIB vs CT arm, respectively, and
with a median follow up of 11.8 months. No DLT was reported in dose escalation. In total, similar CBR, ORR, and TTR regardless of treatment (tx). No new safety signals were observed in pts
38 patients (86.4%) had visceral disease, and 8 patients (18.2%) had ESR1 mutation at with liver mets. A numerically longer median time to deterioration (mTTD) in FACT-B total score was
baseline. 22 patients (50%) received prior endocrine therapy in the advanced setting, of seen with RIB + ET vs combo CT in pts with and without liver mets. Conclusions: This analysis from
which, 15 patients (34.1%) had aromatase inhibitor (AI), 12 patients (27.3%) had ful- RIGHT Choice showed similar clinically meaningful efficacy and quality-of-life benefits and no new
vestrant. 13 patients (29.5%) received prior chemotherapy in the advanced [Link] safety signals for RIB + ET vs combo CT between pts with and without liver mets. These results
most common treatment emerged adverse events (TEAEs) were white blood cell count support the 1L use of RIB + ET in pts with clinically aggressive HR+/HER2– ABC even in the presence
decreased (95.5% ) and neutropenia (95.5%). Sinus bradycardia was reported in 77.3% of liver mets. Clinical trial information: NCT03839823. Research Sponsor: Novartis Pharmaceuticals
(34/44) of the patients, 85.3% (29/34) were grade 1 (asymptomatic) requiring no dose Corporation.
modification. Grade 3/4 treatment-related AEs (TRAEs) occurred in 77.3% of the patients mPFS, mTTF, CBRa, ORRa, mTTRa, mTTD (FACT-B
Liver mo HR mo HR % % mo HR total score)b, HR
with most commonly reported events including neutropenia (70.5%) and white blood cell mets Tx arm n (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) mo (95% CI)
count decreased (40.9%). No fatal AEs were reported. TRAEs led to dose reduction were Yes RIB + ET 54 18.3 0.68 13.2 0.60 77.8 64.8 6.4 0.68 37.7 0.68
reported in 38.6% for palbociclib and 9.1% for SIM0270. No TRAEs led to treatment (10.3-
24.0)
(0.42-
1.11)
(10.2-
21.2)
(0.39-
0.92)
(64.4-
88.0)
(50.6-
77.3)
(4.6-
23.9)
(0.42-
1.10)
(0.34-
1.34)
discontinuation. And 24 patients remain on study treatment. In the response evaluable Combo CTc 53 12.7 8.3 67.9 60.4 3.0 18.4
(7.5-21.0) (5.3-12.8) (53.7- (46.0- (2.6-
patients, confirmed overall response rate (ORR) was 41.5% (17/41) and clinical benefit 80.1) 73.5) 6.7)
rate (CBR, defined as complete response, partial response or stable disease $ 24 weeks) No RIB + ET 58 25.2
(18.6-NE)
0.57
(0.34-
24.0
(16.4-
0.44
(0.29-
84.5
(72.6-
67.2
(53.7-
4.6
(2.8-
0.81
(0.52-
NE 0.59
(0.29-
was 82.5% (33/40). Median progression free survival (PFS) was not reached (NR). In c
0.93) 32.2) 0.69) 92.7) 79.0) 10.2) 1.28) 1.21)
Combo CT 57 15.4 10.1 80.7 63.2 4.5 37.1
patients with ESR1 mutation at baseline, ORR and CBR were 87.5% (7/8) and 100% (8/8), (8.8-20.0) (7.8-13.6) (68.1- (49.3- (1.4-
90.0) 75.6) 8.2)
respectively. Conclusions: SIM0270 in combination with palbociclib showed acceptable
safety and tolerability, promising clinical activity in patients with ER+/HER2- advanced NE, not evaluable.
a
Without confirmation;
breast cancer. Clinical trial information: NCT05293964. Research Sponsor: Simcere b
$7 point decrease;
c
Zaiming Pharmaceutical Co., Ltd. docetaxel + capecitabine, paclitaxel + gemcitabine, or capecitabine + vinorelbine.

1070 Poster Session 1071 Poster Session

Elacestrant (Ela) combinations with ribociclib (Ribo) and everolimus (Eve) in Differential genomic landscape of estrogen receptor (ER)-low versus ER-
patients (pts) with ER+/HER2- locally advanced or metastatic breast cancer positive (ER+) and ER-negative (ER-) metastatic breast cancer (MBC). First
(mBC): Update from ELEVATE, a phase (Ph) 1b/2, open-label, umbrella Author: Chiara Corti, Dana-Farber Cancer Institute, Boston, MA
study. First Author: Hope S. Rugo, University of California, San Francisco Helen Dil- Background: Guidelines define ER+ breast cancer (BC) as $1% tumor nuclei staining positive by IHC.
ler Family Comprehensive Cancer Center, San Francisco, CA Data on managing ER-low tumors (1–10% ER staining) is limited, with mixed evidence suggesting
Background: Progression of ER+/HER2- mBC on 1L endocrine therapy (ET) + CDK4/6i is outcomes similar to ER- but a higher risk of death with adjuvant endocrine therapy (ET) omission. We
associated with several mechanisms of resistance that impact efficacy and subsequent aimed to examine the genomic landscape of ER-low MBC compared to ER+ and ER-. Methods: This
retrospective study included consecutive patients (pts) with MBC who consented to clinicopathologic
therapy. Treatment options include endocrine monotherapy, continuing ET+CDK4/6i, or PI3K/
data collection and genomic profiling (OncoPanel) on tumor samples with matched ER IHC through the
AKT/mTOR pathway–ET combination regimens. Acquired ESR1 mutations emerge in up to 50% EMBRACE (Ending Metastatic Breast Cancer for Everyone) program. For multiple sequencing
of patients and continuing SOC ET is limited by resistance to ET due to these mutations. timepoints, the first was analyzed. SNVs, CNVs and TMB were compared among ER groups. Genes
Several trials have shown improved mPFS with the addition of Eve: 3.6-6.8 mo (Cook 2021, altered in . 3% of pts were analyzed for ER status association, with Benjamini-Hochberg adjusted p ,
Vasseur 2024) or switch in CDK4/6i: 5.3 mo (Kalinsky 2023). In the Ph 3 EMERALD trial, single- 0.2 subjected to Holm-corrected pairwise testing. Results: Between 10/2000-12/2020, 1199 pts were
agent Ela significantly improved PFS vs SOC ET (ESR1-mut tumors HR 0.55; 95% CI 0.39-0.77; identified:48 ER-low, 797 ER+, and 354 ER-. Median age at diagnosis was 63.8 (34.8-86.8), 64.4 (30.8-
P=0.0005; all pts HR 0.70; 95% CI 0.55-0.88; P=0.0018) with manageable safety in pts with 96.3), and 62.6 (30.3-92.7) years for ER-low, ER+, and ER- groups, respectively. De novo stage IV
ER+/HER2- mBC who had prior ET+CDK4/6i (Bidard 2022). This analysis reports updated safety disease was observed in 8.3% (4/48) of ER-low, 27.0% (215/797) of ER+, and 17.2% (61/354) of ER-
and preliminary efficacy for Ela in combination with Ribo or Eve. Methods: ELEVATE evaluates cases. Overall, 801/1199 (66.8%) had metastatic and 398/1199 (33.2%) had primary samples se-
Ela in combination with everolimus (Eve), alpelisib (Alp), capivasertib (Capi), ribociclib (Ribo), quenced. 27/48 ER-low (56.3%), 451/797 ER+ (56.6%), and 73/354 ER- (20.6%) pts received ET before
palbociclib (Palbo), or abemaciclib (Abema) to address different resistance mechanisms. Pts sequencing. CDK4/6i were administered in 8/48 ER-low (16.7%), 107/797 ER+ (13.4%), and 5/354 ER-
with ER+/HER2- mBC and 1-2L of prior ET are eligible regardless of ESR1-mut status. Ob- (1.4%) pts prior to sequencing. The most clinically relevant genomic alterations are shown in the Table.
jectives are to identify the RP2D (Ph 1b) and evaluate PFS (Ph 2) with each combination. TP53 mutations (mts) were more frequent in ER-low vs ER+ BC, and not significantly different between
Results: Elacestrant combinations with Ribo or Eve showed safety consistent with the known ER-low and ER- tumors. PIK3CA and CDH1 alterations were more frequent in ER-low than ER- BC, with
profiles of each drug + SOC ET. The most common AEs ($30%) with Ela + Ribo (n=32) from Ph no significant difference compared to ER+. AKT1 and RB1 alterations were significantly higher in ER-
low vs ER+ BC. ESR1 mts were not significantly different between ER+ and ER-low BC. Median TMB
1b were neutropenia (38%; 25% $Gr3) and nausea (31%; 0 $Gr3). The most common AEs for
was higher in ER-low vs ER+ cases, without significant differences between ER-low and ER- cases.
Ela + Eve (n=72) from Ph 1b + Ph 2 were nausea (54%; 6% $Gr3), diarrhea (43%; 7% $Gr3),
Conclusions: ER-low BC has a distinct genomic profile, with high TP53 mts (similar to ER-) and
stomatitis (38%; 3% $Gr3), and fatigue (32%; 6% $Gr3). Median PFS for Ela + Ribo was 7.2 frequent PI3K pathway alterations (typical of ER+). Ongoing analyses of clinicopathologic features and
months, while for Ela + Eve was 8.5 months. Table 1 summarizes mPFS from Ph 1b in efficacy- survival across ER-low, ER+, and ER- cohorts will be presented. Research Sponsor: Terri Brodeur
evaluable pts who received prior ET+CDK4/6i, as of Dec 2024. Updated data will be presented. Breast Cancer Foundation; The Benderson Family Fund; NIH/NCI grant; 1P50CA168504.
Conclusions: Elacestrant plus Ribo or Eve demonstrates promising Ph 1b efficacy in pts with
ER+/HER2- mBC with progressive disease after ET+CDK4/6i in all patients. Ela 345 mg + Ribo ER+ ER-low ER- ER-low vs ER+ ER-low vs ER- ER+ vs ER-
400 mg QD was determined as the RP2D. Previously, Ela 345 mg + Eve 7.5 mg was identified as Characteristic (N = 797) (N = 48) (N = 354) (p value) (p value) (p value)
the RP2D. Elacestrant has the potential to become an ET backbone for various targeted agents, TP53 24% 79% 83% 1.45 x 10^-14 0.689 1.85 x 10^-79
offering an all-oral treatment regimen in pts with ER+/HER2- mBC, delaying chemo or ADC- PIK3CA 39% 31% 13% 0.291 0.00292 1.96 x 10^-21
CDH1 19% 17% 3% 0.85 0.00193 6.66x10^-14
based regimens. Clinical trial information: NCT05563220. Research Sponsor: None. AKT1 2% 15% 4% 0.0155 0.0155 0.635
Ph 1b mPFS in prior ET+CDK4/6i, efficacy-evaluable population. RB1 1% 12% 9% 0.00354 0.35 0.000116
PTEN 9% 10% 14% 0.828 0.828 0.828
Ela (86-345 mg) + Ela (258-345 mg) + ESR1 11% 6% 0% 0.612 0.00322 3.06x10^12
N Ribo (400-600 mg) N Eve (5-10 mg) CCND1 (CNV) 17% 12% 3% 0.551 0.0258 2.83 x 10^-11
mPFS, mo (95% CI) 32 7.2 (3.52 - 12.78) 22 8.5 (7.23 - 16.07) TMB, median (IQR) 6.844 8.365 7.604 0.018 0.212 0.001
(4.562) (6.917) (4.562)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
54s BREAST CANCER—METASTATIC

1072 Poster Session 1073 Poster Session

Molecular and prognostic convergence of HR+/HER2– metastatic breast Ultrasensitive ctDNA monitoring during CDK4/6 inhibitor therapy for met-
cancer (MBC) to a TNBC-like profile: Insights from circulating tumor DNA astatic breast cancer. First Author: Julia Ah-Reum An, Memorial Sloan Kettering
(ctDNA)-based genomic analysis across treatment lines. First Author: Lorenzo Cancer Center, New York, NY
Foffano, Universita degli Studi di Udine, Udine, Italy Background: The combination of CDK4/6 inhibitor (CDK4/6i) and endocrine therapy
Background: While the transition to a triple negative (TNBC)-like profile represents a (ET) is the standard first-line treatment for patients with hormone receptor-positive/
recognized mechanism of treatment resistance for hormone receptor-positive, HER2- HER2-negative (HR+/HER2-) metastatic breast cancer (MBC). However, it exhibits highly
negative (HR+/HER2-) MBC, the molecular mechanisms of this phenomenon remain largely variable efficacy, with some cancers progressing within 3–6 months while many others
unknown. This analysis investigated the genomic and prognostic differences between achieve durable and potentially indefinite complete responses (CRs). While pharma-
HR+/HER2– and TNBC across treatment lines through ctDNA profiling analysis cologic strategies to escalate or deescalate this therapy exist, diagnostic tools to
Methods: This retrospective study analyzed a multi-institutional cohort of 1071 patients identify the patients who would benefit from each approach are needed. Ultrasenstive
(pts) with HER2 negative MBC and ctDNA testing with the Guardant360 NGS panel within a ctDNA offers the potential to assess disease burden dynamically and with more pre-
large academic consortium (PMAC). HR and HER2 status were defined based on the most cision. In this study, we evaluate the validity of an ultrasensitive assay capable of
recent biopsy, pts with ER-low profile (ER , 10% regardless of PR status) were excluded. detecting ctDNA levels in the parts per million range for monitoring patients with HR+/
Associations across single nucleotide and copy number variations (SNVs and CNVs), HR+/ HER2- MBC. Methods: Patients from the MSK-LINC prospective ctDNA monitoring
HER2– and TNBC subtypes across treatment lines were tested by multinomial logistic study, who received CDK4/6i+ET for HR+/HER2- MBC were included in the study. MRD
regression (MLR) in terms of Relative Risk Ratio (RRR). The impact of prognosis was monitoring was performed using personalized tumor-informed panels designed from
evaluated through Cox regression for overall survival (OS), defined from time of baseline whole genome sequencing (WGS) of matched tumor and normal specimens to identify
ctDNA collection. Results: There were 827 pts with HR+/HER2- MBC (77.2%) and 244 pts up to 2,000 somatic alterations for each patient using the Precise MRD assay (Myriad
with TNBC (22.8%). Multivariable MLR, designed with first line HR+/HER2- as the reference, Genetics). Results were reported as an overall ctDNA detection status and a quantitative
investigated genomic alterations across treatment lines. In second line, ESR1 SNVs (RRR tumor fraction. Results: 29 patients with HR+/HER2- MBC (8 de novo, 21 recurrent)
7.34, p , 0.001) and EGFR CNVs (RRR 0.15, p = 0.01) were significantly associated with were included in this ongoing study. The median progression-free survival (PFS) was
HR+/HER2-, while TP53 SNVs had a higher prevalence in TNBC (RRR 2.71, p = 0.009). In 48.8 months (range 2.6 – 78.5) with 17/29 of patients experiencing disease progression.
third line, ESR1 SNVs were significantly enriched in HR+/HER2- (RRR 5.44, p , 0.001),
ctDNA panels were successfully designed for all cases, and 140/146 (95.9%) plasma
while TP53 SNVs emerged for TNBC (RRR 5.26, p , 0.001). From fourth line onward ($
samples passed QC. All pre-treatment samples had detectable ctDNA with a median
4L), ESR1 SNVs (RRR 8.09, p , 0.001), TP53 SNVs (RRR 1.81, p = 0.022) and PIK3CA CNVs
tumor fraction of 1.4% (range 0.00093%, 14.0%). An early decrease in ctDNA levels, .
(RRR 5.93, p = 0.003) showed higher prevalence in HR+/HER2- relative to first line HR+/
50% reduction from baseline or levels , 0.01% in the second sample collected within 3
HER2-, while TP53 SNVs were also associated with TNBC (RRR 10.43, p , 0.001).
Compared to TNBC, HR+/HER2– had a favorable prognostic impact in terms of OS in first
months, was significantly associated with longer PFS (p , 0.001). We focused on 7
(HR 0.32, p , 0.001), second (HR 0.35, p , 0.001) and third line (HR 0.37, p , 0.001). patients who achieved radiographic CR all with PFS . 3y. Notably, 3 patients had
However, in $ 4L, no significant differences emerged (HR 0.79, p = 0.282), with similar continued to have ultra low levels of ctDNA (median: 0.0086%, range 0.00032%, 0.11%),
results observed with respect to TNBC across all lines (HR 1.01, p = 0.929). MYC CNVs had indicating stable viable micrometastatic disease below the threshold of imaging, ef-
an unfavorable prognostic role for both HR+/HER22 $4L (HR 2.41, p = 0.004) and TNBC in fectively controlled by treatment. In contrast, 4 patients also achieved molecular CR
all lines (HR 2.14, p = 0.014). Conclusions: Our study suggests a dynamic molecular (mCR) defined as sustained undetectable ctDNA suggesting that metastatic disease was
evolution of HR+/HER2– MBC, with a progressive acquisition of molecular and prognostic either eradicated or rendered dormant without significant cell turnover. Conclu-
features compatible with a TNBC-like profile and loss of endocrine sensitivity. These sions: Ultrasensitive ctDNA monitoring is a promising tool for monitoring disease
findings highlight the need for comprehensive biological characterization of this subtype burden and treatment response. Our results highlight the ability of ctDNA to distinguish
across treatment lines to better understand its evolution under therapeutic pressure and between stable molecular disease vs. mCR, highlighting the potential of ctDNA as a
consequently adapt treatments. Research Sponsor: None. biomarker for tailoring treatment strategies in patients who achieve outstanding clinical
responses. Research Sponsor: Myriad Genetics, Inc.; Susan G. Komen.

1074 Poster Session 1075 Poster Session

Effectiveness comparison of palbociclib, ribociclib and abemaciclib in pa- Use of baseline plasma circulating tumor DNA (ctDNA) to predict duration of
tients with HR+/HER2- aBC: Updated results from the real-world, Italian endocrine therapy (ET) and CDK4/6 inhibitor (CDK4/6i) therapy (tx) and to
study PALMARES-2. First Author: Claudio Vernieri, Fondazione IRCCS Istituto analyze intrinsic vs acquired endocrine resistance. First Author: Pietro
Nazionale dei Tumori, Milan, Italy De Placido, Dana-Farber Cancer Institute, Boston, MA
Background: The Cyclin Dependent Kinase 4/6 inhibitors (CDK4/6i) Palbociclib (P), Ribociclib Background: ET + CDK4/6i is standard-of-care for patients (pts) with hormone receptor-
(R) and Abemaciclib (A) combined with Endocrine Therapy (ET) are the standard 1st line therapy positive/HER2-negative (HR+/HER2-) metastatic breast cancer (MBC). We aimed to identify
for patients (pts) with Hormone Receptor positive, Human Epidermal growth factor Receptor 2- predictors ET + CDK4/6i tx duration and to compare genomic profiles in pts with intrinsic vs
negative, advanced Breast Cancer (HR+/HER2- aBC). However, based on conflicting results of acquired resistance. Methods: Plasma samples were collected from pts with HR+/HER2-
large real-world (RW) studies (Vernieri C et al. Abstr 1014, ASCO 2024; Rugo H et al. PS2-03, MBC enrolled in the EMBRACE cohort study who had plasma collection within 3 months (mo)
SABCS 2024), it remains unclear whether P, R and A are similarly effective. Methods: prior to CDK4/6i initiation to 14 days after initiation. The primary outcome was duration of
PALMARES-2 is a multicenter, observational Italian RW study comparing the effectiveness of 1st ET+CDK4/6i tx, defined as time from tx initiation to end of tx. Intrinsic resistance was defined
line P, R or A in female pts with HR+/HER2- aBC. The primary endpoint is overall survival (OS); RW as pts with tx duration , 180 days. Plasma samples were analyzed using the Guardant360
progression-free survival (rwPFS) and time to chemotherapy (TTC) are secondary endpoints. assay, which includes genotyping of . 700 genes and tumor fraction (TF) score. TF was
rwPFS, TTC and OS were defined as the time between 1st line ET+CDK4/6i initiation and disease estimated by normalizing cancer-specific differentially methylated regions with matched
progression/death, initiation of 1st chemotherapy line/death, or patient death, respectively. We control regions in each sample. The predictive value of baseline TF (0 vs . 0) was tested
used Inverse Probability of Treatment Weighting (IPTW) to balance 14 prognostic covariates
using a Cox regression model including age, line of tx, and liver metastases. For comparison
related to patients (age, ECOG PS, menopausal status), tumor biology (ER, PgR, HER2, Ki67,
of pts with intrinsic vs acquired resistance, analysis was limited to samples with TF .1% to
grading, histology, endocrine sensitivity/resistance/de novo metastatic) and metastatic sites
minimize the impact of variation in tumor shed. Gene frequency between intrinsic and
(liver, bone, lung, serosal) in P, R and A cohorts. Effectiveness comparisons were reported as
adjusted Hazard Ratio (aHR) and 95% confidence interval (CI). Results: With a cutoff date of Jan acquired resistance samples were compared using q-tests(q , 0.25). Results: A total of 188
10th, 2025, we enrolled 3598 pts, of whom 1392 (38.7%), 1408 (39.1%) or 798 (22.2%) received P, pts were included. Median age at MBC diagnosis was 57.5 yrs. ET+CDK4/6i was given in the
R or A, respectively. Pts receiving A were more likely to have endocrine-resistant disease, liver first-line (1L) in 115 pts, second-line (2L) in 37 pts, and . 2L in 36 pts. Of 167 pts, TF was
metastases and lower PgR expression, and less likely to have de novo metastatic disease (p , undetectable (TF = 0) in 19 (11%) and detectable (TF . 0) in 148 (89%). In Cox regression,
0.001). Median follow-up was shorter in R/A cohorts (31.8/29.6 months) than in the P cohort baseline TF (p = 0.001), line of tx (n = 0.002), and presence of liver metastasis (p = 0.014), but
(52.4 months). Median rwPFS, TTC and OS in the whole population were 26.1, 39.4 and 67.1 not age, were predictors of duration of tx. Median duration of tx was 44.6 mo in pts with
months, respectively. After IPTW adjustment, R and A were associated with better rwPFS and baseline TF = 0 vs. 5.8 mo in pts with baseline TF . 0 (HR 0.28, 95% CI 0.14-0.56). Similar
TTC when compared to P, while only R was associated with better OS (Table). R and A did not results were found when restricting the analysis to those receiving tx in the 1L or 2L. There
show significant rwPTS, TTC or OS differences (Table). Conclusions: The three CDK4/6i have were notable differences in the frequency of ESR1 (63% vs 48%), CDH1 (38% vs 18%), PTEN
different effectiveness in HR+/HER2- aBC pts. Longer follow-up of PALMARES-2 study and more (21% vs 9%), RB1 (32% vs 20%), and CDKN2A (20% vs 5%) alterations in pts with intrinsic vs
pts/events in the A cohort are needed to perform definitive OS comparisons between P, R and A. acquired resistance, though these did not reach statistical significance in the setting of small
Research Sponsor: None. sample size. ESR1 fusions were seen in 14% (8/56) pts with intrinsic resistance vs 7% (3/44)
pts with acquired resistance. Among pts with intrinsic resistance, ERBB2 copy number loss
P (N = 1392) R (N = 1408) A (N = 798)
N° events (%) N° events (%) N° events (%) was present in 7(13%) (6 het loss, 1 homozygous deletion), RB1 copy number loss in 10 (18%)
(all het loss), and CDKN2A copy number loss in 9(16%) (7 het loss, 2 homozygous deletions).
rwPFS 992 (71.2%) 711 (50.5%) 418 (52.4%) Conclusions: Baseline TF in ctDNA is highly predictive of time on ET+CDK4/6i tx in pts with
TTC 845 (60.7%) 516 (36.6%) 332 (41.6%)
OS 586 (42.1%) 270 (19.2%) 189 (23.7%) HR+/HER2- MBC. Baseline genomic profiles differ qualitatively between pts with intrinsic vs
R vs P: aHR (95% CI; P) A vs P: aHR (95% CI; P) A vs R: aHR (95% CI; P) acquired resistance. If validated, baseline plasma may provide a valuable tool in tx selection.
rwPFS 0.88 (0.80-0.97; 0.02) 0.88 (0.77-0.99; 0.04) 0.99 (0.87-1.13; 0.91) Research Sponsor: Guardant Health; Breast Cancer Research Foundation; Saverin Breast
TTC 0.83 (0.74-0.94; ,0.01) 0.86 (0.75-0.99; 0.04) 1.04 (0.89-1.20; 0.65)
OS 0.75 (0.64-0.87; ,0.01) 0.91 (0.76-1.09; 0.3) 1.21 (0.99-1.49; 0.06)
Cancer Research Fund; Pan Mass Challenge; NCCN-Pfizer Collaborative Grant.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 55s

1076 Poster Session 1077 Poster Session

Comparing clinical benefit of trastuzumab deruxtecan (T-DXd) and sacitu- Steroid receptor expression and overall survival in breast cancer patients
zumab govitecan (SG) in a large cohort of HER2-negative metastatic breast with ER+ bone metastasis: A retrospective review. First Author: Anthony Michael
cancer (MBC). First Author: George W. Sledge Jr., Caris Life Sciences, Phoenix, AZ Rossi, University of Arizona College of Medicine, Tucson, AZ
Background: T-DXd and SG are antibody-drug conjugates (ADCs) increasingly used in HER2- Background: Endocrine therapy (ET) resistance is common in estrogen receptor-positive
negative BC, however, there are insufficient data to guide ADC sequencing and use in tumors of (ER+) metastatic breast cancer (BC), where bone metastases (BMET) are usually the first
various HER2 expression levels. Methods: A total of 4033 HER2 negative MBC treated with SG sign of spread. ER signaling and ET effects can depend on other steroid hormones re-
or T-DXd that underwent tumor profiling at Caris Life Sciences (Phoenix, AZ) were studied. HER2 ceptors (SHRs), such as progesterone receptors (PR) and androgen receptors (AR).
low (Her2-L), ultra low (Her2-UL) and null (Her2-N) were tested by IHC and CISH. Hormone However, the roles of these receptors in ER+ BC BMET are underexplored. To address this
receptor status (HR+/-) was tested by ER and PR IHC. Real-world clinical data were obtained gap, PR and AR protein expression in HER2-/ER+ BMET and associations with overall
from insurance claims. Time on treatment (TOT) was determined as the interval from start to end survival (OS) were examined. Methods: In a retrospective analysis on BC BMET samples
of the ADCs. Cox proportional hazards model was used for hazard ratio (HR) and log-rank for p analyzed at Caris Life Sciences, n = 2038 HER2- BMETS were identified by immunohis-
values. Results: Overall, 1444 (36%) were treated with T-DXd but not SG (T-only) while 1808
tochemistry (IHC) (#1+intensity or #10% staining, or 2+ & . 10% with CISH-null reflex
(45%) with SG but not T-DXd (S-only). HR+ cases comprise 64% of T-only and 24% of S-only
test). HER2- ER+ (by IHC, $1+ & $1%) BMETs (“ER+ BMET” n = 1700; 84.5% of total) were
cohorts; 75% and 68% of T-only and S-only tumors were taken from metastatic sites. As
expected, HER2-L, HER2-UL and HER2-N cohorts treated with T-DXd had decreased TOT (4.8
then examined for prevalence of IHC+ SHR expression (PR: $1+ & $1%; AR: $1+ &
months (m), 4.1m and 3.5m, p, .001) while HER2 status had no impact on SG TOT (3.0m, 2.8m $10%) and associated pathogenic/likely pathogenic ESR1 or PIK3CAgene mutations
and 3.4m). Interestingly, even in HER2-N group, T-only showed a borderline better TOT than S- (mut). Associations of SHR status with clinical outcomes were tested by inferring OS from
only (Table, p=.053); this effect was significant in HR+ HER2-N subset but not significant in HR- biopsy collection or start of therapy to last contact. Results: Most ER+ BMET expressed
HER2-N. Similarly, in HER2-UL and L, T-DXd TOT was significantly longer than SG TOT; when PR (59.3%) or AR (87.1%). Only 9.4% of PR+/ER+ BMET were AR-null, while 38.2% of AR+/
further stratified by HR status, the effect was highly significant in HR+ and not seen in HR- ER+ BMET were PR-null. Overall, “triple positive” (AR+/PR+/ER+) BMET comprised the
cohorts. In cohorts crossed over from one ADC to another, patients treated with T-DXd first largest group (53.7%), followed by PR-null AR+/ER+ BMET (33.2%). AR-null ER+ BMET with
(N=420) or SG first (N=361) showed no TOT difference (10.4m vs. 10.8m, p= .4); although the (5.6%) or without PR (7.3%) were less common. AR+ status was associated with better
HER2-N subset had moderate preference of SG first (11.5m vs. 8.5m, HR=0.66 [0.52-0.84], p, outcomes for ER+ BMET patients (pts) with longest OS for “triple positive”, while “loss” of
.001, while TOT were similar in HER2-UL (HR=0.93, p=.7) and HER2-L (HR=1.04, p=.7) groups. PR was associated with shorter OS (Cox proportional hazards ratio (HR) = 1.37, p ,
Conclusions: We report outcome from a large real world dataset and demonstrate that T-DXd 0.0001). AR-null status was associated with worse OS compared to “triple positive” re-
shows statically significant improved outcome in HR+ tumors across HER2 subgroups while in gardless of PR status (AR-/PR- HR = 1.66, p , 0.001; AR-/PR+ HR = 1.85, p , 0.0001). In
TNBC, both agents exhibit comparable benefit. In patients treated with both ADC’s, SG first ER+ BMET with ESR1mut (16.3%), OS for triple positive tumors, which were more prevalent
showed preferred outcome in HER2-N group but not in HER2-UL or L. We provide important (74.0%) due to increased PR expression, was reduced (HR = 1.92, p , 0.0001 vs ESR1wt).
insight on clinical benefit of the two widely used ADCs in breast cancer and warrants further For PIK3CA, “loss” of PR abrogates benefits associated with AR+ status only in PIK3CAmut
validation in independent cohorts. Research Sponsor: None. pts (48.7%) with patterns between SHR groups otherwise maintained. Among ER+ BMET
T-DXd (T-only) TOT and SG TOT (S-only) in cohorts stratified by HER2 and HR status. pts who received aromatase inhibitors or fulvestrant, AR+ status was associated with the
All HR+ HR- best OS, regardless of PR status, where treatment was associated with significantly longer
TOT (T;S, HR [95% OS (vs no treatment) across SHR groups, except in PR+/AR-null. For ER+ BMET pts treated
months) N (T;S) CI] p TOT N HR p TOT N HR p with CDK4/6 inhibitors, OS was highest in the “triple positive” cohort, with onlyAR+ SHR
HER2- 4.7; 3.4 262; 1.1 [1- 0.053 4.8; 209; 1.5 [1.2- <0.001* 4.6; 48; 1.1 [0.8- 0.6 groups demonstrating improved OS with treatment. Conclusions: Based on this analysis,
N 1116 1.3] 3.0 277 1.8] 3.5 822 1.4] AR expression is more prognostic of OS than PR, regardless of treatment, in pts with ER+
HER2- 4.8; 3.0 295; 1.4 [1.2- <0.001* 5.1; 245; 1.8 [1.4- <0.001* 3.2; 46; 1.1 [0.8- 0.4
UL 289 1.7] 2.5 99 2.2] 3.0 188 1.6] BC BMETs, with “triple positive” BMETs generally associated with the best OS. Research on
HER2- 4.9; 3.5 707; 1.2 [1- 0.011* 5.1; 595; 1.5 [1.2- <0.001* 4.2; 100, 1.1 [0.8- 0.7 SHRs as mediators vs biomarkers of risk in ER+ BC BMETs is needed to provide direction
L 244 1.4] 3.1 77 1.9] 3.9 164 1.3] for possible therapeutic targeting. Research Sponsor: None.
*: Significant

1078 Poster Session 1079 Poster Session

Evaluating accuracy and concordance of pathologists and the utility of AI Elacestrant combinations in patients (pts) with ER+/HER2- locally advanced
assistance software for digital HER2 IHC assessment in breast cancer or metastatic breast cancer (mBC): Safety update from ELEVATE, a phase
including HER2-ultralow scoring: An international multicenter observational (Ph) 1b/2, open-label, umbrella study. First Author: Nancy Chan, NYU Langone
study. First Author: Gabriela Acosta Haab, Pathology Department, Maria Curie Hospital, Health, New York, NY
Buenos Aires, Argentina Background: Tumors develop resistance following 1L endocrine therapy (ET) + CDK4/6i in ER+/
Background: The emergence of novel therapeutic agents demonstrating improved HER2- mBC. Elacestrant (Ela) significantly improved PFS vs standard-of-care (SOC) ET (ESR1-mut
progression-free survival (PFS) and overall survival (OS) in breast cancer patients with tumors HR 0.55; 95% CI 0.39-0.77; P=0.0005; all pts HR 0.70; 95% CI 0.55-0.88; P=0.0018) with a
low HER2 expression underscores the need for accurate and reproducible HER2 status manageable safety profile for pts with ER+/HER2- mBC and prior ET+CDK4/6i (Bidard 2022).
ELEVATE (NCT05563220) evaluates Ela in combination with everolimus (Eve), alpelisib (Alp),
assessment. However, challenges such as subjective interpretation of immunohisto- capivasertib (Capi), ribociclib (Ribo), palbociclib (Palbo), or abemaciclib (Abema) to address
chemistry (IHC) staining and variability in assay quality hinder diagnostic consistency. different resistance mechanisms. Prior analyses have demonstrated safety consistent with the
AI-based decision support software could enhance diagnostic accuracy and repro- known profiles of each agent in combination with SOC ET. Ph 1b safety and efficacy evaluations
ducibility. To date, systematic evaluation of pathologist performance in scoring low reported the RP2D and antitumor activity with the following combinations: Ela + Eve (RP2D: Ela 345
HER2 expression, as well as the role of AI assistance, remains limited in real-world, mg QD + Eve 7.5 mg QD) and Ela + Abema (RP2D: Ela 345 mg QD + Abema 150 mg BID)(Ciruelos
multicenter settings. Methods: Six academic centers from different countries provided ESMO 2024, Rugo ESMO 2024). Ela 345 mg + Palbo 125 mg was determined as the RP2D (Rugo
digital HER2 IHC-stained breast cancer images (n = 728) generated with five whole-slide SABCS 2024). Herein, we report updated safety that includes additional pts/dose levels, and longer
scanner models and one microscope camera. In a two-arm observational study, con- observation time for Ela in different combinations. Methods: Eligible pts have ER+/HER2- mBC and
sensus ground truth (GT) scores were established by two expert pathologists per center 1-2L of prior ET regardless of ESR1-mut status. Objectives are to determine the RP2D (Ph 1b) and
evaluate PFS (Ph 2) with each combination. Results: Table 1 reports the most common all-grade
without AI assistance. Subsequently, two additional pathologists (scorers) evaluated
AEs from Ph 1b (Ribo, Alp, Capi combinations) and Ph 1b + Ph 2 (Eve combination) as of Dec 2024.
each case both without and with AI support. Scoring followed ASCO/CAP 2023 HER2 Ela 345 mg + Ribo 400 mg QD was identified as the RP2D. Updated data will be presented.
interpretation guidelines, with an additional subclassification of IHC 0 cases into "null" Conclusions: Elacestrant combinations continue to demonstrate safety consistent with the known
(IHC 0 with no staining) and "ultralow" (IHC 0 with membrane staining). Results: For the profiles of each drug + SOC ET without increased risk of associated AEs. Elacestrant has the
HER2-low decision range, AI software alone achieved 91.0% accuracy in distinguishing potential to become an ET backbone for multiple targeted agents, providing an all-oral treatment
HER2 0 from 1+/2+/3+ scores against GT. Across the four categories, AI achieved 80.3% option in pts with ER+/HER2- mBC, delaying chemo or ADC-based regimens. Clinical trial in-
accuracy compared to 77.6% for scorers alone and 81.4% with AI assistance. AI support formation: NCT05563220. Research Sponsor: None.
improved inter-reader agreement from 73.5% to 86.4%. When the HER2 ultralow cat- Treatment-emergent AEs (‡30%).
egory was included, AI assistance increased scorers’ average accuracy across all classes Ph 1B Ph 1B + Ph 2 Ph 1B Ph 1B
from 70.4% to 74.7% and boosted inter-reader agreement from 65.6% to 80.6%. For Ela (86-345 mg) + Ela (258-345 mg) + Ela (258 mg) + Ela (258-345 mg) +
differentiating HER2 null from HER2 ultralow, AI improved scorers’ accuracy from 68.6% Ribo (400-600 mg) Eve (5-10 mg) Alp (150-250 mg) Capi (320 mg)
(n=32) (n=72) (n=11) (n=9)
to 77.9%, resulting in 40% more cases being classified as HER2 ultralow and 65%
All grades, n (%) Neutropenia* 12 (38) Nausea 39 (54) Nausea 8 (73) Nausea 6 (67)
reduction in the number of incorrectly scored HER2 null cases. Conclusions: This first Nausea 10 (31) Diarrhea 31 (43) Vomiting 6 (55) Fatigue 5 (56)
international multicenter study on HER2 IHC diagnosis, including HER2 ultralow scoring Stomatitis 27 (38) Rash† 4 (36) Diarrhea 5 (56)
highlights the challenges faced by pathologists and the significant benefits of AI de- Fatigue 23 (32) Vomiting 3 (33)
Grade ‡3, n (%) Neutropenia* 8 (25) Diarrhea 5 (7) Rash† 2 (18) 0
cision-support systems in real-world settings. AI assistance improved pathologist Nausea 4 (6) Nausea 1 (9)
concordance and accuracy, particularly at the HER2 null vs. ultralow boundary, reducing Fatigue 4 (6)
diagnostic errors. Incorporating AI into routine clinical diagnostics has the potential to Stomatitis 2 (3)
optimize treatment selection for breast cancer patients. Research Sponsor: *Combined terms;
†
AstraZeneca. Maculopapular rash.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
56s BREAST CANCER—METASTATIC

1080 Poster Session 1081 Poster Session

Enhancing precision oncology for Haitian breast cancer patients through Cell-free circulating chromatin profiling for epigenomic characterization of
deep learning-enabled computational pathology tools. First Author: Rebecca mechanisms of response and resistance to sacituzumab govitecan in breast
Henderson, University of Alabama at Birmingham, Birmingham, FL cancer. First Author: Ana Christina Garrido-Castro, Dana-Farber Cancer Institute,
Background: While early-stage breast cancer is often curable in high-resource settings, Boston, MA
mortality-to-incidence ratios remain unacceptably high for women in lower- and middle- Background: The efficacy of sacituzumab govitecan (SG), a TROP2-directed antibody-
income countries (LMIC). This disparity is due to an inability within LMICs for patients to drug conjugate (ADC), in hormone receptor-positive/HER2-negative (HR+/HER2-) meta-
access basic cancer diagnostics (e.g., IHC). Consequently, there is a major need to static breast cancer (mBC) has been demonstrated, yet biomarkers predicting response
develop innovative approaches for the cancer diagnostic-therapeutic pipeline to deliver and resistance remain an unmet clinical need. We applied a novel multimodal epigenomic
high-quality care for LMIC patients. To this end, deep learning (DL) has shown con- liquid biopsy assay to characterize tumor-specific transcriptional activation of relevant
siderable promise in identifying clinically relevant biology within histopathology (H&E). genes of interest and resistance mechanisms in the phase 2 SACI-IO HR+ trial
Therefore, we have curated an unprecedented dataset of H&E whole slide images (WSIs) (NCT04448886). Methods: Baseline plasma samples were collected from patients (pts)
and tissue-matched estrogen receptor (ER) status for 5500 breast cancer slides from with HR+/HER2- mBC enrolled in SACI-IO HR+, which compared SG alone to SG combined
Zanmi Lasante (Haiti). Methods: Using The Cancer Genome Atlas (TCGA) breast cancer with pembrolizumab (SG-pembro). Genome-wide signals from promoters, enhancers and
and Haitian datasets, we trained a DL-enabled tool, using H&E WSIs, to predict ER status DNA methylation were profiled from 1 mL of plasma from 95 pts, of which 80 met the assay
for each patient. As the TCGA dataset predominantly comprises patients of European quality control thresholds and ctDNA metrics required for downstream analysis (ctDNA $
ancestry, we assessed whether a TCGA-trained model would generalize to Haitian 0.5%, NSG= 42, NSG-pembro= 38). We used epigenomic and RNA-seq datasets from 23 breast
patients. After WSI processing and feature extraction, attention-based weakly super- cancer cell lines to train a model to predict TROP2 expression (r = 0.66, P , 0.01) and
vised multiple instance learning was used to train a classification model. To assess tested for association with progression free survival (PFS). We used Gene Set Variation
performance, both the TCGA and Zanmi Lasante datasets were split into training (70%), Analysis to score samples for HALLMARK gene set activities using gene-proximal epi-
validation (15%), and testing (15%) sets, and the results were compared across both genomic signals and tested for independent association of those activities with PFS via
patient populations. Results: Using the TCGA dataset (2100 H&E WSIs), we trained an CoxPH models, with baseline ctDNA fraction included as a known prognostic covariate.
ER classification model. This model demonstrated a performance of an area under Statistical significance was determined based on improved model fit compared to ctDNA
alone. Results: Compared to healthy donors, plasma from trial pts was enriched for breast
receiver operating characteristic (AUROC) of 0.92 on the “held-out” TCGA test set, but
cancer specific signatures such as estrogen response and hedgehog signaling (FDR ,
only an AUROC of 0.71 on the Haitian “held-out” test set for ER status prediction. This
0.05), highlighting the ability of the liquid biopsy platform to extract tumor specific signal.
drop in model performance, or domain shift, is consistent with known biological dif-
Baseline ctDNA fraction was prognostic in both treatment arms (Hazard Ratio [HR]SG=
ferences between breast cancers enriched in Black women compared to those in
0.38, P , 0.01; HRSG-pembro= 0.28, P , 0.01). Conversely, predicted TROP2 expression was
Caucasian women. Notably, pre-training our model on the TCGA dataset and then fine- not associated with PFS in either treatment arm. In the SG arm, higher activity of pathways
tuning on a portion of the Haitian training set (2800 WSIs) substantially improved such as epithelial to mesenchymal transition and Wnt signaling were associated with
predictive performance to an AUROC of 0.85 on the Haiti test set. Conclusions: This shorter PFS (FDR , 0.1), highlighting potential mechanisms of resistance. Gene sig-
study illustrates the potential of DL to advance precision oncology in low-resource natures related to cell cycle such as mitotic spindle and E2F targets were associated with
settings and highlights the need for adequate training data from LMIC patients. We longer PFS (FDR , 0.1). The above pathway associations with PFS were not statistically
anticipate tools from this work will be deployed for use in Haitian breast cancer patients significant in the SG-pembro arm. Conclusions: This study demonstrates the feasibility of
to inform precision-based use of endocrine therapies. Research Sponsor: None. a multimodal epigenomic liquid biopsy platform for non-invasive characterization of
therapeutic response and resistance to SG with or without pembrolizumab in HR+/HER2-
mBC. By providing real-time insight into transcriptional regulation, this approach may
improve patient stratification and guide ADC treatment strategies. Clinical trial infor-
mation: NCT04448886. Research Sponsor: Gilead Sciences; Merck; Komen; METAvivor;
Gateway for Cancer Research; Mehlman Family Fund.

1082 Poster Session 1083 Poster Session

Treatment patterns in HR+/HER2- metastatic breast cancer (MBC) with co- Targeting PARP-1 in ER-positive endocrine-resistant breast cancer. First
occurring PIK3CA and ESR1 mutations. First Author: Jimmitti Teysir, Memorial Author: Azzurra Zicarelli, University of Chicago, Chicago, IL
Sloan Kettering Cancer Center, New York, NY Background: Resistance to endocrine therapy (ET) in breast cancer (BC) patients is
Background: PIK3CA and ESR1 mutations co-occur in 12-15% of patients with HR+/ frequently associated with acquired ESR1 gene mutations like the Y537S, which triggers
HER2- MBC. While FDA-approved PI3K inhibitors (PI3Ki) and selective estrogen receptor a constitutive estrogen receptor (ERa) activation. Therefore, the identification of novel
degraders (SERDs) have advanced care for patients with biomarkers, treatment outcomes therapeutic strategies is crucial for the management of ER-positive, ET-resistant BC. In
and optimal sequencing in those with co-occurring mutations remain unclear. this context, PARP1 poly(ADP-ribose) polymerase 1 (PARP-1) has emerged as a
Methods: We conducted a retrospective analysis of patients with HR+/HER2- MBC and promising therapeutic target, based on its involvement in the regulation of oxidative DNA
co-occurring PIK3CA and ESR1 mutations treated at Memorial Sloan Kettering (MSK) damage in BC cells. Methods: Data from the METABRIC dataset were used to assess the
between 2010 and 2024. Mutations were identified via MSK-IMPACT (tumor tissue) and clinical relevance of PARP-1 in ER-positive BC patients. As experimental models, MCF7
MSK-ACCESS (ctDNA), with additional genomic data integrated from Guardant and and T47D BC cell lines expressing ERa wild type (wt) or Y537S mutation were used.
Foundation One. Clinical and treatment data, including PI3Ki (alpelisib, inavolisib or PARP-1 regulation was investigated by western blotting, immunofluorescence, and
investigational agents) and/or SERDs (fulvestrant or oral agents) were abstracted. Median chromatin immunoprecipitation (ChIP) assays. Gene expression, promoter assays, and
progression-free survival (mPFS) was assessed in patients with mutations identified prior chromatin immunoprecipitation sequencing (ChIP-seq) studies allowed us to analyze the
to therapy initiation. Results: 3,166 patients with HR+/HER2- MBC were identified, in- transcriptional activity mediated by ERa. Cell cycle, proliferation and colony formation
cluding 1,444 (46%) with PIK3CA mutations, 664 (21%) with ESR1 mutations and 243 (8%) experiments as well as in vivo studies were performed to evaluate the biological effects
with co-occurring mutations. After excluding 26 patients with incomplete records or of the PARP-1 inhibitor niraparib. Results: We observed that the up-regulation of PARP-
concurrent malignancies, the final cohort included 217 patients (7%), with 206 having 1 upon exposure to 17b-estradiol (E2) occurs through ERa in BC cells expressing either
MSK-IMPACT, 58 MSK-ACCESS, 49 Guardant, and 3 FoundationOne data. Of 217 patients, ERa wt or Y537S mutation. Moreover, we assessed that the transcriptional activity of
77 (36%) received a PI3Ki after a median of 4 prior lines (range: 1–16), with 68 (88%) ERa relies on PARP-1, as demonstrated by the ability of nirabarib to prevent the
having prior CDK 4/6i. Single-agent SERD was administered to 46 patients (21%) after a transactivation of ERa and the regulation of ERa target genes. In addition, niraparib
median of 3 prior lines (range: 2–18), including 34 (74%) with prior CDK 4/6i. 8 patients halted the proliferation and cycle progression of BC cells expressing either ERa wt or
received PI3Ki followed by SERD, and 7 received SERD followed by PI3Ki, either con- Y537S mutation. Of note, niraparib suppressed primary tumor growth in xenograft
secutively or with intervening treatments. The mPFS was 7.1 m with PI3Ki (95% CI: tumors derived from ERa Y537S mutated MCF7 cells. Conclusions: Our data suggest
4.6–9.3) and 4.0 m with single-agent SERD (95% CI: 3.4–9.8). In the SERD cohort, earlier that crosstalk between PARP-1 and ERa is involved in the proliferative responses of ERa
line of treatment (1–2 vs. 3+; HR 0.29, 95% CI 0.09–0.90) and liver metastasis (HR 3.42, wt or Y537S mutated BC cells. Therefore, targeting PARP-1 could provide a promising
95% CI 1.10–10.6) were independently associated with PFS. CDK 4/6i duration $12 m in strategy to overcome the ET resistance of BC cells. Research Sponsor: None.
the SERD cohort was associated with improved mPFS in stratified analysis (9.8 vs. 2.7 m;
log-rank p = 0.002). For patients treated with PI3Ki followed by SERD (n = 8), mPFS1 was
8.9 m (95% CI: 5.6, NE) and mPFS2 was 6.0 m (95% CI: 2.3, NE). SERD to PI3Ki (n = 7)
yielded a mPFS1 of 3.4 m (95% CI: 1.8, NE) and mPFS2 of 10.0 m (95% CI: 1.6, NE).
Conclusions: Prior CDK4/6i $ 12 m in patients with co-occurring mutations treated with
a SERD was associated with significantly improved PFS, potentially reflecting a con-
ditioning effect of prior therapy. The total PFS (PFS1 + PFS2) in patients treated with
sequential targeted therapies was similar; however, small sample sizes and potential
confounders in this retrospective cohort limit definitive interpretations. Larger pro-
spective studies are needed to determine optimal sequencing strategies. Research
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 57s

1084 Poster Session 1085 Poster Session

Racial and ethnic differences in biomarker testing for targetable alterations Therapeutic impact of novel agents in patients with stage IV de novo HR+ve/
among patients with HR+ HER2- metastatic breast cancer (mBC). First Author: Her2-ve breast cancer: Results from a real world dataset. First Author:
Catherine Keane, Flatiron Health, New York, NY Shaheenah S. Dawood, Mediclinic City Hospital, Dubai, United Arab Emirates
Background: In recent years novel therapies have been approved for patients (pts) with Background: The objective of this retrospective analysis was to look at the therapeutic
mBC and ESR1, AKT1, PTEN, PIK3CA, and gBRCA alterations. Given limited evidence on impact of CDK4/6i and novel agents among pts with stage IV Denovo HR+ve/HER2-ve
which patients are receiving standard of care, this study assessed racial and ethnic breast cancer (BC). Methods: We utilized a federated network of de-identified health
inequities in biomarker testing and the role of social determinants of health (SDOH) in data representing approximately 165 million pt lives available through the TriNetX
explaining potential inequities. Methods: This study leveraged the US nationwide Research Network. We identified 41,843 pts with HR+ve/HER2-ve stage IV Denovo BC
Flatiron Health electronic health record (EHR)-derived, deidentified database of . 750 treated diagnosed between Jan 2005 - Jan 2025. Propensity score matching analysis by
000 pts with BC. Adult female pts diagnosed (dx) with HR+ HER2- mBC between 1/1/ age and site of metastases was carried out. OS was computed using the Kaplan Meier
2011, and 4/30/2024, with a geocodeable address were included. Testing rates for product limit method. The index event is the date of diagnosis. Results: 8,541(20.4%)
alterations in ESR1, PIK3CA, AKT1, PTEN, and gBRCA were measured over time from received a CDK4/6i. Among pts treated with CDK4/6i 1,396(16.3%), 6,169 (72.2%) and
mBC dx using variables extracted from unstructured clinician documentation in the EHR 2,157 (25.2%) pts received Ribo, Palbo and Abema respectively. Over time there has been
using machine learning. Fine and Grey models accounting for competing risks were used a significant decrease in use of palbocilcib with significant increase in use of abema and
to estimate subdistribution hazard ratios (HR) and 95% confidence intervals (CI) for ribo. Median OS was similar between Ribo and Abema(HR 0.88; 95%CI (0.75,1.04)
biomarker access. Models were adjusted for covariates including age, stage, ECOG p=0.14). Compared to pts receiving palbo median OS was signficantly better among pts
status, and dx year, followed by practice setting and area-level SDOH factors (ie, English receiving abema (HR 0.77; 95%CI (0.69,0.85) p,0.0001) or ribociclib (HR 0.69; 95%CI
language proficiency, residential segregation, vehicle ownership, urbanicity, and resi- (0.60,0.81) p,0.0001).628 pts received more than one CDK4/6i. (152 palbo + ribo, 118
dence in medically underserved areas). Results: The cohort included 36 316 pts (61.5% ribo + abema, 368 abema+ palbo). 5-year OS was 63.9% and 70.4%(HR 1.37; 95%CI
non-Latinx [NL]-White, 6.1% Latinx, 9.7% NL-Black, 1.9% NL-Asian, and 20.8% NL-Other/ 1.06,1.77) respectively among pts who received 1 vs .1 CDK4/6i. Among patients
Unknown). Overall, Asian, Black, and Latinx pts were less likely than White pts to undergo treated with CDK4/6i, OS was significantly longer among pts who received elacestrant vs
biomarker testing (adjusted HR [95% CI]: Latinx, 0.88 [0.82-0.95]; NL-Black, 0.87 [0.82- those who did not (HR 0.401; 95%CI (0.215,0.745). 228 pts treated with a CDK4/6i
0.93]; NL-Asian, 0.87 [0.76-0.98]). Racial/ethnic inequities in overall biomarker testing received an antibody drug conjugate (ADC). Median time to use of an ADC was 45m. 5yr
were partially explained by SDOH factors. Specifically, the White-Latinx inequity in OS was 75.8% vs 58.7% among those who did and did not receive an ADC respectively
testing was mediated by residential segregation ie, association attenuated towards the (HR 0.45, 95%0.30,0.67). 5yr OS was 76.2% vs 52.7% among those who did and did not
null (mediated HR [95% CI], 0.94 [0.87-1.02]), limited English proficiency (0.92 [0.85- receive Trastuzumab deruxtecan respectively (HR 0.37, 95%0.19,0.70). 5yr OS was
1.00]), and lack of vehicle ownership (0.91 [0.84-0.98]). Compared with White pts, NL- 76.2% vs 60.4% among those who did and did not receive Sacituzumab govetican
Black pts were less likely to be tested for ESR1 (HR [95% CI], 0.86 [0.77-0.95]) and respectively (HR 0.40, 95%0.25,0.65). Conclusions: Among pts with stage IV Denovo
PIK3CA (0.86 [0.80-0.92]). Latinx pts were less likely to be tested for PIK3CA (0.87 [0.80- HR+ve/HER2-ve BC treated with a CDK4/6i using a CDK4/6i beyond progression is an
0.95]) and this inequity was mediated by residential segregation (0.96 [0.87-1.05]) and option. Novel agents such as oral SERDS and ADCs are also associated with improved
limited English proficiency (0.92 [ 0.84-1.00]). Conclusions: Asian, Black, and Latinx pts prognostic outcome in the real world setting. Research Sponsor: None.
were generally less likely than their White counterparts to receive biomarker testing after
a mBC dx, especially for PIK3CA and ESR1. SDOH factors explained some of these
biomarker testing inequities. Equitable access to biomarker testing should be prioritized
to ensure patients have access to the most effective therapies. Future research should
examine whether racial/ethnic inequities in biomarker testing are associated with in-
equities in treatment and outcomes. Research Sponsor: Flatiron Health.

1086 Poster Session 1087 Poster Session

Updated efficacy of mutant-selective PI3Ka inhibitor RLY-2608 in combi- Impact of body weight and body composition on survival and toxicities in
nation with fulvestrant in patients with PIK3CA-mutant HR+HER2- advanced patients receiving CDK4/6 inhibitors for ER+/HER2- metastatic breast
breast cancer: ReDiscover trial. First Author: Sarah L. Sammons, Dana-Farber cancer. First Author: Jorge Avila, Montefiore Medical Center, Bronx, NY
Cancer Institute, Boston, MA Background: Body composition influences treatment outcomes and adverse events in
Background: Oncogenic PIK3CA mutations constitutively activate PI3Ka and drive many oncologic conditions including metastatic breast cancer (mBC). The objective of
approximately 40% of HR+HER2- breast cancer (BC); however, the toxicity (hypergly- our study was to evaluate the impact of body composition measures on progression-free
cemia, rash, diarrhea, stomatitis) of non-selective inhibitors (i) limits their tolerability survival (PFS) and adverse events in patients treated with cyclin-dependent kinase 4/6
and efficacy. RLY-2608 is the first oral, pan-mutant-selective, allosteric PI3Kai designed (CDK4/6) inhibitors for estrogen receptor-positive (ER+)/HER2- mBC. Methods: A single
to overcome these limitations. We report efficacy and safety of RLY-2608 + standard- institution retrospective analysis of 207 patients treated with CDK4/6 inhibitors was
dose fulvestrant (F) in pts with PIK3CA-mutant, HR+HER2- BC treated in the FIH study, conducted. Baseline body weight and body composition measures (total fat, visceral fat,
ReDiscover (NCT05216432). Methods: Previously treated adult pts with advanced subcutaneous fat, skeletal muscle area, skeletal muscle density and muscular adiposity)
HR+HER2- BC and PIK3CA mutation per local assessment were eligible. Pts were eligible were analyzed. PFS was evaluated using Cox proportional hazard models, and logistic
to enroll with measurable or non-measurable disease. Key objectives were investigator- regression was used to evaluate the relationship between these variables and adverse
assessed efficacy per RECIST 1.1 and adverse events (AEs) per CTCAE v5.0. Results: As events. Early changes in body composition were defined as variations in muscle or fat
of 4NOV24, safety was assessed in 118 pts treated across RLY-2608 doses 100-1000 mg compartments within 3 months. Low muscle quality was defined as muscle power index
BID, and efficacy in the 52 pts without detectable PTEN/AKT co-alterations treated at the value of 1 to 2 Standard Deviations below the normal range. Results: The median age of
RP2D (600 mg BID). All pts received prior endocrine therapy and CDK4/6i with 48% our cohort was 61 years, with 77% of patients being postmenopausal and 46% identifying
having $ 2 prior systemic therapies for advanced disease including 56% with prior F/ themselves as Black. Most patients received palbociclib (76%), followed by abemaciclib
SERD and 25% with prior chemotherapy or antibody-drug conjugate. Median follow-up (14%) and ribociclib (10%) as part of their treatment. Median BMI was 27.97 kg/m2, with
was approximately 9.5 months. The RP2D provided exposure in the target therapeutic 36% being classified as obese. Sarcopenia was present in 18% of our patients, and 71%
range and rapid clearance of mutant PIK3CA ctDNA. 31/52 pts had measurable disease had low muscle quality. Higher BMI (HR, 0.96; 95% CI, 0.93-0.99; p 0.01), obesity (HR,
with 26/31 (83.9%) achieving disease control, 23/31 (74.2%) experiencing radiographic 0.60; 95% CI, 0.42-0.86; p = 0.01) and weight (HR, 0.99; 95% CI, 0.98-0.99; p = 0.01) were
tumor reduction and 12/31 achieving an objective response (38.7%, 95% CI 21.8-57.8) significantly associated with improved PFS. Modest but statistically significant asso-
with median time-to-response 8 weeks. mPFS was 9.2 months (95% CI 5.8,18.4) across ciations with PFS were observed for total fat (HR 0.99; 95% CI 0.98-0.99; p = 0.01) and
all 52 RP2D pts, and 11.4 months (95% CI 7.2-NR) in 32 pts receiving RLY-2608 at the subcutaneous fat (HR, 0.99; 95% CI, 0.98-0.99; p = 0.01); however, this was not the case
RP2D as 2L treatment. Treatment-related AEs (TRAEs) were generally low-grade, for visceral fat and muscle adiposity. There was no significant association between body
manageable and reversible, most commonly hyperglycemia (42.4% any grade; 2.5% Gr muscle compartment distribution (including sarcopenia) and PFS. Early changes in
3), nausea (41.5%; 0.8% Gr 3), fatigue (40.7%; 8.5% Gr 3), creatinine increased (34.7%; skeletal muscle density were associated with improved PFS (HR, 0.95; 95% CI, 0.91-0.99;
0.8% Gr 3), and diarrhea (30.5%; 1.7% Gr 3). There were no grade 4/5 TRAE; and severe, p = 0.01), while sarcopenia and low muscle quality were not significant predictors. Grade
off-target stomatitis and rash were absent or rare. Conclusions: RLY-2608 demon- 3/4 hematologic toxicities were associated with lower muscle area (p = 0.02), but no
strates favorable safety/tolerability along with highly encouraging PFS observed across significant association between fat compartments and adverse events was found.
PIK3CA genotypes in pts with advanced PIK3CA-mutant HR+HER2- BC previously Conclusions: Obesity is associated with improved survival outcomes in patients re-
exposed to CDK4/6i. These data validate RLY-2608 as the first allosteric pan-mutant ceiving CKD4/6 inhibitors for ER+/HER2- mBC; furthermore, this effect is driven by
selective PI3Kai and support advancing RLY-2608 + F to pivotal testing, which is subcutaneous fat. Early changes in skeletal density during CDK4/6 inhibitors treatment is
planned for later this year. Clinical trial information: NCT05216432. Research Sponsor: a potential predictor of improved outcomes. Muscle area is a potential predictor of
Relay Therapeutics. treatment toxicities. Our findings suggest that body composition plays an important role
in outcomes and adverse events in this group of patients. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
58s BREAST CANCER—METASTATIC

1088 Poster Session 1089 Poster Session

Phase 1 trial of exercise as first-line therapy for hormone receptor (HR)– Pulmonary toxicities in patients (pts) with metastatic breast cancer (mBC)
positive advanced breast cancer (TBCRC 054). First Author: Neil M. Iyengar, treated with trastuzumab deruxtecan (T-DXd): The Mayo Clinic Enterprise
Memorial Sloan Kettering Cancer Center, New York, NY Experience, updated. First Author: Jenna Elizabeth Hoppenworth, Mayo Clinic,
Background: Observational studies show post-diagnosis exercise is associated with Rochester, MN
reduced risk of cancer death in patients with HR-positive breast cancer. We conducted a Background: T-DXd has become an important treatment option in mBC and other
phase 1a dose-finding trial of exercise therapy as first-line treatment for HR-positive malignancies. Interstitial lung disease/pneumonitis (ILD) occurred in 10-14% of pts in
advanced breast cancer (TBCRC 054). Methods: This multicenter trial was conducted the DESTINY-Breast trials (0-2% G5 ILD), and with potential lower incidence/severity in
using a patient-centric, decentralized platform (NCT03988595). Non-exercising patients earlier line settings (Krop et al, ASCO 2023). We previously reported the Mayo Clinic
receiving first line endocrine plus CDK4/6 inhibitor therapy were allocated using an Rochester, MN experience with T-DXd related ILD in mBC (Hoppenworth et al, ASCO
adaptive continual reassessment design to one of four escalated exercise therapy dose 2024). Here, we expand the data to include patients treated at all locations of the Mayo
levels (range: 90 to 300 min/week) of individualized, moderate-intensity treadmill Clinic Enterprise. Methods: We retrospectively identified pts with mBC who received $1
walking for 6 consecutive months. The trial was later amended to add a fifth dose level of dose of T-DXd across the Mayo Clinic enterprise (Rochester, MN; Mayo Clinic Health
375 min/week and to allow dose cohort backfilling. Exercise therapy sessions were System locations in MN/WI; Scottsdale, AZ; and Jacksonville, FL) between July 2022-
conducted remotely in patient’s homes with real-time monitoring. The primary objective December 2023. Demographic, mBC characteristics, and pulmonary clinical variables
was to identify the recommended phase 2 dose (RP2D) as determined by feasibility and were abstracted from the clinical records. Data were summarized using descriptive
preliminary clinical efficacy. Feasibility was evaluated by relative exercise dose intensity statistics. Diagnosis of ILD was determined by treating clinicians, and severity was
(REDI). A dose level was considered feasible if $70% of patients achieved a REDI $75%. approximated to CTCAE V5 based on clinical documentation. Results: 252 pts with mBC
One-year progression free survival (PFS) rates were assessed by the Kaplan-Meier received T-DXd during the study period. The majority were Caucasian (86%) and female
method. Results: Fifty-four women (median age 53 [46 to 63] years) were enrolled (99%) with a median age of 63. 91 pts (36%) were current/formers smokers and 97 (39%)
between August 2019 and April 2024; 23 (43%) had visceral metastases, 43 (80%) had prior pulmonary comorbidities. The majority had HER2 low (155, 62%) and HR
received an aromatase inhibitor, 11 (20%) received fulvestrant, and 53 (98%) received a positive (164, 65%) mBC. 35 (14%) developed any grade ILD [G1: 6 (17%), G2: 13 (37%),
CDK4/6 inhibitor. The proportion of patients with REDI $75% in each dose level was: 90 G3: 3 (8.6%), G4: 3 (8.6%), G5: 10 (29%)], with 15 (43%) presenting with $G3. 29 (83%)
min/week (n = 10): 80%, 150 min/week (n = 10): 60%, 225 min/week (n = 11): 82%, 300 presented with at least 1 symptom (cough or SOB). Among those with previous pul-
min/week (n = 13): 62%, and 375 min/week (n = 10): 40%. Among the two feasible dose monary toxicity, 4 (33%) had previous pneumonitis in the ILD cohort. The median prior
levels (90, 225 min/week), 1-year PFS rate was 70% (95% CI, 47% to 100%) in the 90 min/ lines of all therapies (including endocrine therapy) were 5, with a median of 3 prior lines
week dose level and 91% (95% CI 75% to 100%) in the 225 min/week dose level. No of chemotherapy in both the ILD (range 1-13) and non-ILD cohorts (range 1-13). Median
serious adverse events were observed. Overall, 225 min/week (~ 45 minutes per onset to any grade ILD was 7 cycles. 20 pts (57%) received steroids for ILD. 14 (40%) had
treatment at 5 times weekly) was selected as the RP2D. Conclusions: This multicenter a bronchoscopy, all had a CT chest, 16 (46%) were hospitalized, 7 (20%) were intubated
phase 1 trial showed that exercise therapy doses of 90 and 225 min/week are feasible and 1 (3%) had a lung biopsy. Pts with G5 ILD had a median of 3 lines of chemo and a
and safe in patients receiving first line endocrine-based therapy. These data also support median of 8 cycles prior to onset of ILD, compared to 3 lines of chemo and a median of 7
the rationale for a phase 2 trial testing the preliminary clinical efficacy of exercise at the cycles for those with G1-4 ILD. G5 pts were all Caucasian females, 5 were former
RP2D of 225 min/week in HR-positive advanced breast cancer. Clinical trial information: smokers, 5 were non-smokers, 1 had a history of pneumonitis. T-DXd was rechallenged
NCT03988595. Research Sponsor: National Cancer Institute; R01CA235711. in 4 pts (G1-2 ILD) without ILD recurrence. Conclusions: In this retrospective case
series,14% of pts treated with T-DXd experienced any grade ILD (4% G5) which is in
alignment with the rate observed in the pivotal DESTINY-Breast trials, though with higher
rates of G5 toxicity. Among those with ILD, increased lines of therapy were not as-
sociated with increased risk. Further research is needed to correlate risk. Research
Sponsor: None.

1090 Poster Session 1091 Poster Session

Preliminary efficacy and safety of TQB2102 in patients with HER2 low- Standard-dose vs fixed-dose capecitabine in patients with advanced gas-
expressing recurrent/metastatic breast cancer: Results from a phase 1b trointestinal and metastatic breast cancer. First Author: Nanuli Gvazava, The
study. First Author: Shusen Wang, Department of Medical Oncology, Sun Yat-Sen University of Kansas Cancer Center, Westwood, KS
University Cancer Center; State Key Laboratory of Oncology in South China; Collabo- Background: Capecitabine at the FDA-approved standard dose (SD) of 1250 mg/m² twice daily for 14
rative Innovation Center for Cancer Medicine, Guangzhou, China days with a 7-day break, has significant toxicities. We conducted a randomized trial comparing SD and
Background: TQB2102 is a novel antibody-drug conjugate (ADC) comprised of a recombinant fixed dose (FD) Capecitabine 1500 mg twice daily, 7 days on, 7 days off in patients with metastatic
humanized anti-HER2 bispecific antibody that simultaneously binds to two distinct HER2 breast cancer (MBC) and advanced gastrointestinal (GI) cancers. We previously reported that in MBC
cohort progression-free survival and overall survival were similar, and FD had significantly lower
epitopes (ECD4 and ECD2), an enzyme-cleavable linker, and a DNA topoisomerase I inhibitor
toxicities. We now present time to treatment failure (TTF) and toxicity in MBC and GI cohorts.
payload. This study aims to evaluate the efficacy and safety of TQB2102 for patients (pts) Methods: Patients with MBC or advanced GI cancers (colorectal, small bowel, gastroesophageal,
with HER2-expressing relapsed/metastatic breast cancer. Methods: This 1b phase, open- pancreatic and bile duct) with any prior lines of therapy were randomized 1:1 to either FD-7/7 or SD-14/
label, multicenter, randomized trial was divided into two cohorts: Pts in cohort 1 were HER2 7. Post hoc analysis was performed to determine TTF, and landmark analysis was performed for
low-expressing breast cancer and in cohort 2 were HER2 positive BC, all pts were refractory or Freedom from Treatment Failure (FFTF). Capecitabine-related toxicities [diarrhea, hand foot syndrome
intolerant to standard therapy. In cohort 1, HER2 low-expressing pts were randomly assigned (HFS) and stomatitis] were solicited and graded at each visit. Results: 182 patients were enrolled
to receive TQB2102 monotherapy at a dose of 6.0 mg/kg (Q3W, IV) or 7.5 mg/kg (Q3W, IV). (N=93 FD, N=89 SD) of which 153 had MBC and 29 had an advanced GI cancer. Median TTF was 4.92
The primary endpoint was ORR per RECIST v1.1, and the secondary endpoints were PFS, DCR months (3.02, 5.93) in FD arm, and 3.11 months (2.49, 3.90) in SD arm (log rank p=0.0111). Landmark
and safety etc. Results: 73 HER2 low-expressing female pts were randomized to receive at analysis of FFTF is shown in Table 1. At 24 months, the FFTF in the FD arm was 15.6%, while in the SD
least one dose of TQB2102 6mg/kg (n = 37) or 7.5mg/kg (n = 36), the median age was 53. All arm it was 2.5% (p=0.0054). Grade 2 and higher toxicities were more common in SD compared to FD,
pts had received chemotherapy in the metastatic setting, and hormone receptor positive pts including HFS, diarrhea, and stomatitis (Table 1) Conclusions: Fixed-dose capecitabine at 1500 mg
(n = 50) also had received prior CDK4/6 inhibitors. In cohort 1, pts had undergone a median of twice daily for 7 days on and 7 days off demonstrates a longer time to treatment failure compared to
the standard FDA-approved dosing in patients with MBC and advanced GI cancers and is associated
4 prior treatment lines (range: 1-10) in the metastatic setting, the median prior lines of
with significantly lower toxicities. Clinical trial information: NCT02595320. Research Sponsor: None.
chemotherapy therapies were 2 (range: 1-5), and 12.3% (n = 9) pts had received prior ADCs. As
of data cutoff on Nov 1, 2024, median follow-up time was 7.16 months. ORR was 53.4% (39/ Landmark freedom from treatment failure at 12, 24 and 36 months; solicited adverse events.
73) in cohort 1, and the ORR of 7.5mg/kg (58.3%) was better relative to 6.0mg/kg (48.7%). FD-7/7, N=93 SD-14/7, N=89
Objective responses were observed in subgroups with HR positive pts (27/50, ORR 54.0%), HR Time Survival Probability Estimate P-value
negative pts (12/23, ORR 52.2%); of which the ORR of 7.5mg/kg with HR+ and HR- was 66.7% 3-month 61.3% 51.4% 0.1917
(14/21) and 46.7% (7/15), respectively. For HER2 low-expressing pts who received prior ADC 6-month 39.6% 30.6% 0.2224
therapies, the ORR was 44.4% (4/9). In cohort 1, DCR was 86.3% (63/73, among 6 pts was not 12-month 23.8% 14.1% 0.1188
24-month 15.6% 2.5% 0.0054
available), and median PFS was not yet mature. TRAEs were reported in 71 (97.3%) HER2 low- Adverse Event Number (proportion)
expressing pts. Grade$3 TRAEs and serious TRAEs were reported in 30 (41.1%), 13 (17.8%) Diarrhea
pts, respectively. The main TRAEs were neutropenia, leukopenia, anemia, nausea, vomiting. Any Grade 51 (54.8%) 59 (66.3%) 0.1142
Grade 2-4 8 (8.6%) 35 (39.3%) ,0.0001
The common TRAEs and grade $3 TRAEs occurred similarly at both doses, with hematologic Grade ‡ 3 3 (3.2%) 20 (22.5%) ,0.0001
toxicities such as anemia were slightly higher at 7.5mg/kg than at 6.0mgkg, but all were HFS
tolerable. No Interstitial lung disease was reported in cohort 1 pts. Conclusions: TQB2102 Any Grade 46 (49.5%) 61 (68.5%) 0.0090
was well-tolerated and showed promising antitumor activity in heavily pretreated HER2 low- Grade 2-4 13 (14.0%) 40 (44.9%) ,0.0001
Grade ‡ 3 1 (1.08%) 15 (16.9%) 0.0002
expressing recurrent/metastatic breast cancer pts. The recommended phase 3 dose of Stomatitis
TQB2102 in HER2 low-expressing r/m BC was 7.5 mg/kg Q3W. A phase III trial to evaluate the Any Grade 24 (25.8%) 48 (53.4%) 0.0001
efficacy and safety of TQB2102 versus investigator-selected chemotherapy in HER2 low- Grade 2-4 2 (2.2%) 13 (14.6%) 0.0023
Grade ‡ 3 0 (0.0%) 6 (6.7%) 0.0125*
expressing r/m BC is currently ongoing (NCT06561607). Clinical trial information:
NCT06115902. Research Sponsor: None. *Fisher’s exact test.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 59s

1092 Poster Session 1093 Poster Session

Limited changes in the CNS immune microenvironment in patients with A phase II study to evaluate the safety and efficacy of BB-1701 in subjects
breast cancer metastasis and capturing these changes using machine with HER2 expression locally advanced/metastatic breast cancer previously
learning. First Author: Andrew Ip, John Theurer Cancer Center, Hackensack Univer- treated with HER2-ADC containing TOP-I inhibitor. First Author: Xiaoxiang Guan,
sity Medical Center, Hackensack, NJ The First Affiliated Hospital of Nanjing Medical University, Nanjing, China
Background: Metastasis of breast cancer to the central nervous system (CNS) is Background: BB-1701 is an HER2-targeting antibody-drug conjugate (ADC) containing
common, especially in triple negative and HER2-positive tumors. The CNS is considered eribulin. In phase I study, BB-1701 had shown promising antitumor activity in breast
immune specialized and likely the brain and brain-border immune microenvironment cancer (BC) patients with HER2 high/low expression and manageable safety profile.
creates a sanctuary site for breast cancer CNS metastasis. A better understanding of the Currently, there are no approved treatment options for metastatic BC patients with high/
immune microenvironment may allow for better utilization of immunotherapy to treat low expression who have received HER2-ADC containing TOP-I inhibitor, especially for
CNS metastasis. Toward this goal, we evaluated the cellular transcriptomic profile of trastuzumab deruxtecan. We report the preliminary efficacy and safety results from the
cerebrospinal fluid (CSF) cells and compared between patients with documented ongoing phase 2 study of BB-1701 in advanced or metastatic breast cancer patients with
metastatic tumor by cell-free DNA (cfDNA) testing of the CSF fluid (cfCSF-Pos) and HER2 expression previously treated with HER2-ADC containing TOP-I inhibitor.
patients without evidence of cfDNA metastasis (cfCSF-Neg) (Charifa et al., https:// Methods: Patients enrolled were $18 years of age; had confirmed locally advanced/
[Link]/10.1016/[Link].2024.100281). Methods: RNA was extracted from the CSF cells of metastatic HER2 expressing breast cancer; disease progression after previous HER2-
63 cfCSF-Pos patients and 93 cfCSF-Neg patients. The RNA was sequenced and targeting ADC (containing TOP-I inhibitor) therapy; an ECOG PS , 2; and measurable
quantified using a targeted RNA panel of 1600 genes by next generation sequencing lesion(s) (per RECIST v1.1). HER2 expression was confirmed by IHC before patient
(NGS). We used two thirds of the samples for training a machine learning (ML) system enrollment. BB-1701 is administered at 1.6 mg/kg Q3W. Results: As of 28 January
and one third for testing. The ML system uses Bayesian statistics with k-fold cross- 2025, 23 patients with HER2 high/low-expressing breast cancer have been enrolled and
validation (with k = 12) to first rank the top biomarkers distinguishing CSF-Pos from CSF- treated. Median age is 51 years, all patients are female, and 26.1%/73.9% patients have
Neg samples. Then Random Forest was used to distinguish between the two classes ECOG PS 0/1. The median number of prior systemic therapy lines was 4.0, 21.7%/78.3%
using the top ranked biomarkers. Results: cfCSF-Neg contains mainly T-cells with HER2 status are high expression/low expression. All patients experienced at least one
median CD2:CD22 RNA ratio of 70.07 (range 0.01-14820). This was not significantly treatment-emergent adverse events (TEAEs). The most common ($10%) all grade
different (p = 0.19) from cfCSF-Pos cases (median: 41.31, range: 0.4-10172). The ratio of TEAES are neutrophil count decreased, platelet count decreased, Aspartate amino-
CD4:CD8A in cfCSF-Neg (median: 4.89, range: 0.47-2522) was also not significantly transferase increased, and white blood cell count decreased. One grade 3 TEAEs is
different (p = 0.31) from that in cfCSF-Pos cases (median: 5.0, range: 0.29-48). While peripheral neuropathy, and another grade 3 TEAE is neutrophil count decreased. There
significant variation in the levels of T- and B-cells is noted within each group, there was has been no grade 4 or grade 5 events as of data cut-off date. One treatment emergent
no significant difference in the individual cell population (T-cells, B-cells, plasma cells, serious adverse event is peripheral neuropathy. Of the 23 patients, 14 were evaluable for
natural killer cells, neutrophils, monocytes, or dendritic cells) overall after adjusting for efficacy. Among 14 evaluated patients, 4 patients achieved partial response (PR) and 9
multiple testing. Despite this lack of difference in cell populations, the testing set patients had stable disease (SD), with disease control rate (DCR) of 92.8%. Among 3
showed that cfCSF-Pos patients can be readily distinguished from cfCSF-Neg patients HER2 high-expressing patients who were previously treated with trastuzumab der-
with AUC of 0.886 (CI: 0.797-0.976) using 30 top genes selected by ML. Except for KRT8 uxtecan, 1 patient achieved PR and 2 patients had SD with DCR of 100.0%. Among 8
and KRT19, the majority of the top genes selected by the ML algorithm suggests HER2 low-expressioning (IHC 1+) patients, 3 patients achieved PR (2 patients received
modulation of T-cell activation including but not limited to TBX21, CD3D, CD5, IKZF3, prior trastuzumab deruxtecan and 1 patient received prior SHR-A1811) and 4 patients
NFATC2, and INPP5D. Conclusions: The data suggest the CNS remains immunologi- had SD with DCR of 87.5%. More data will be presented at the ASCO meeting. Con-
cally specialized with modulating the adaptive immune response in the setting of breast clusions: BB-1701 shows promising antitumor activity and a manageable safety profile
cancer metastasis. Significant modulation in the CSF T-cells suggests selective tar- in HER2 expressing breast cancer patients who had previously been treated with HER2-
geting with immunotherapy may prove beneficial with the potential for active monitoring ADC (containing TOP-I inhibitor). Clinical trial information: CTR20241422. Clinical trial
using this combination CSF cellular and cfDNA approach. Research Sponsor: None. information: CTR20241422. Research Sponsor: None.

1094 Poster Session 1095 Poster Session

Results of a phase I study of alpelisib and sacituzumab govitecan (SG) in Assessing the impact of scalp cooling in patients receiving trastuzumab
patients with HER2-negative metastatic breast cancer (MBC). First Author: deruxtecan for metastatic breast cancer. First Author: Elahe Salehi, Dana-Farber
Priyanka Sharma, University of Kansas Medical Center, Westwood, KS Cancer Institute, Boston, MA
Background: Sacituzumab govitecan (TROP2-directed antibody drug conjugate, ADC) is Background: Outcomes for patients (pts) with metastatic breast cancer (MBC) have improved with
effective in treatment of pretreated HER2-negative MBC. PI3K is the most frequently novel antibody drug conjugates like trastuzumab deruxtecan (T-DXd). While T-DXd has been
altered pathway in breast cancer. This phase I trial investigated the combination of SG associated with increased risk of alopecia, there are limited data describing the efficacy of scalp
plus alpelisib (oral a-specific PI3K inhibitor) in HER2-negative MBC. Methods: Eligible cooling in preventing alopecia and improving quality of life among pts receiving T-DXd.
Methods: This prospective, phase II study enrolled pts with MBC without alopecia at baseline who
patients had HER2-negative MBC and had received $1 prior line of chemotherapy in the were initiating treatment with T-DXd; pts elected to participate in a scalp cooling (SC) arm with the
advanced or neo/adjuvant setting and had not received prior PI3K/AKT inhibitor. The Paxman scalp cooling system or a non-SC arm. The primary endpoint was hair loss, defined as
study was 3+3 dose escalation design: dose level 1: alpelisib 250 mg+SG 8 mg/kg; dose locally assessed CTCAE v5.0 grade $1 alopecia occurring at C3D1, C5D1, or end of treatment
level 2: alpelisib 250 mg+SG 10 mg/kg; dose level 3: alpelisib 300 mg+SG 10 mg/kg. (EOT), whichever occurred first. The impact of SC on quality of life (QOL) was assessed using the
Alpelisib was dosed PO daily and SG IV, D1 and 8 every 21 days. Antidiarrheal prophylaxis Chemotherapy-Induced Alopecia Distress Scale (CADS) and body image scale (BIS) at baseline,
was utilized for the first two cycles. Primary endpoint was recommended phase 2 dose C3D1, C5D1, and EOT. The study aimed to enroll 20 pts per arm to provide at least 80% power to
(RP2D). Additional endpoints included adverse events (AEs), objective response rate detect a 28% decrease in hair loss rate between the SC vs non-SC arms (8% vs 36%) using a
(ORR), progression-free survival (PFS), and pharmacokinetics. Results: 12 patients difference in proportions test (one-sided type I error 10%). Results: A total of 40 evaluable pts were
were enrolled between 2022-2024 (dose level 1: N = 3, dose level 2: N = 6, dose level 3: N enrolled: 20 in SC arm and 20 in non-SC arm. Median age was 61 (33-77); 2 (5%) were Black, 2 (5%)
were Hispanic. Twenty-eight (70%) pts had hormone receptor positive disease, 11 (27.5%) HER2+, 2
= 3). 7/12 (58%) had triple-negative breast cancer (TNBC), and 5/12 (42%) had hormone
(5.0%) triple-negative. Thirty-five (87.5%) had prior chemotherapy for MBC; median prior lines was
receptor (HR)-positive disease. All patients had visceral disease. 8/12 (67%) had $1 1 [range: 0-5]. 27 (67.5%) had prior endocrine therapy, 28 (70.0%) had prior CDK4/6 inhibition; 7
prior metastatic chemotherapy; 4/12 (33%) had prior immunotherapy; 3/12 (25%) had (17.5%) had prior use of SC. Thirty-three (82.5%) pts (18 [90%] in SC arm, 15 [75%] in non-SC arm)
prior ADC. One dose-limiting toxicity (hyperbilirubinemia) was observed at dose level 2. experienced .grade 1 alopecia, with similar rates observed in both arms (p = 0.41). Grade 2
RP2D was alpelisib 300 mg + SG 10 mg/kg. The most frequent grade $3 treatment- alopecia was the main reason (45%) for SC discontinuation. Median time to G2 alopecia was 2.76
related AEs were electrolyte imbalance (G3 17%, G4 17%), neutropenia (G3 33%, G4 0%), months (95% CI: 1.64-NA) in SC arm and 4.60 months (95% CI: 2.53, NA) in non-SC arm (p = 0.8).
diarrhea (G3 17%, G4 0%), and hyperglycemia (G3 8%, G4 8%). There were no G5 AEs. Median CADS scores trended upward from baseline to EOT (Baseline: 3.50; C3: 7.22; C5: 9.00; EOT:
Pharmacokinetics of SG and its metabolites were consistent with previous reports. 11.5) in the SC arm and were more variable in the non-SC arm (baseline: 3.00; C3: 5.00; C5: 2.56;
Among 11 patients evaluable for response, ORR was 36% (4/11) (complete response [CR] EOT: 6.50); median BIS scores trended upward in both the SC (baseline: 3.00; C3: 8.00; C5: 9.00;
= 1, partial response [PR] = 3) and clinical benefit rate (CR + PR + stable disease . 24 EOT: 11.5) and non-SC arms (baseline: 3.00; C3: 5.00; C5: 5.00; EOT: 9.00), with no statistically
significant difference. Conclusions: In this prospective phase II trial, the use of SC with T-DXd did
weeks) was 64% (7/11). ORR in TNBC was 50% and in HR-positive disease was 20%.
not show a benefit in hair preservation vs no SC. QOL analysis was not significantly different for
Among patients with prior ADC, ORR was 33% and clinical benefit rate was 67%. Three of those receiving SC vs no SC. Small sample size and lack of randomization may limit interpretation
four patients with ORR had response lasting . 6 months, with median duration of of results. Further work is planned to investigate strategies to improve efficacy of SC with ADCs.
response of 14.9 (range 3.2 to 28.1) months. Median PFS was 5.7 months. Tumor and Clinical trial information: NCT04986579. Research Sponsor: AstraZeneca and Daiichi Sankyo;
serial ctDNA analysis are ongoing. Conclusions: Combination of SG + alpelisib was Paxman Scalp Cooling; Friends of Dana-Farber.
feasible, with manageable side effects. The toxicity profile of the combination was CTCAE v5 alopecia.
consistent with the known safety profiles of the two agents. This combination dem- All SC Arm Non-SC Arm
onstrated encouraging efficacy in HER2-negative MBC, with prolonged duration of CTCAE v5 Alopecia N=40 N=20 N=20 P-value
responses, and warrants further evaluation in larger studies. Clinical trial information: No alopecia 7(17.5%) 2(10.0%) 5(25.0%) 0.41
NCT05143229. Research Sponsor: Novartis; Gilead Sciences. Grade 1 11(27.5%) 7(35.0%) 4(20.0%)
Grade 2 22(55.0%) 11(55.0%) 11(55.0%)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
60s BREAST CANCER—METASTATIC

1096 Poster Session 1097 Poster Session

Bria-IMT + checkpoint inhibitor: Phase I/II survival results compared to First-line durvalumab plus chemotherapy with or without oleclumab for
benchmark trials in metastatic breast cancer. First Author: Saranya Chumsri, locally advanced or metastatic triple-negative breast cancer: SYNERGY
Mayo Clinic Florida, Jacksonville, FL overall survival and circulating tumor DNA analysis. First Author: Elisa Agos-
Background: Bria-IMT is a combination immunotherapy consisting of allogeneic whole cell cancer tinetto, Université Libre de Bruxelles (ULB), Hôpital Universitaire de Bruxelles (HUB),
vaccine (SV-BR-1-GM) administered w/ immune checkpoint inhibitor (CPI). SV-BR-1-GM breast Institut Jules Bordet,, Brussels, Belgium
cancer cells are engineered to directly stimulate anti tumor immunity via expression of tumor Background: SYNERGY (NCT03616886) is a randomized, investigator-initiated, phase I/II trial
associated antigens and secretion of GM-CSF to enhance dendritic cell activation. Addition of CPI testing if targeting the immunosuppressive adenosine pathway with the anti-CD73 antibody
potentiates SV-BR-1-GM to overcome the immune suppressive tumor microenvironment. oleclumab, plus the anti-PD-L1 durvalumab and chemotherapy, enhances antitumor activity in
Methods: This Ph I/randomized Ph II study evaluated the Bria-IMT regimen in pts w/ metastatic untreated locally advanced or metastatic triple-negative breast cancer (TNBC). Here, we report the
breast cancer; CTX (300 mg/m²) on day -2/-3, SV-BR-1-GM and CPI on Day 0, w/ low dose peg overall survival (OS) and circulating tumor DNA (ctDNA) analysis. Methods: 133 patients received
interferon a at inoculation sites on day 2 (61) . Phase II pts were randomized 1:1 to receive CPI at weekly carboplatin and paclitaxel x12 plus durvalumab, with (arm A) or without (arm B) oleclumab
cycle 1 or cycle 2. Two SV-BR-1-GM formulations (w/ vs w/o IFNg incubation) were evaluated. (6 in phase I, 63 in arm A, 64 in arm B). Maintenance with durvalumab +/- oleclumab was continued
Biomarkers included cancer-associated macrophage-like cells, circulating tumor cells, PD-L1 until disease progression or unacceptable toxicity. The primary endpoint was clinical benefit rate at
scores, and delayed-type hypersensitivity skin tests. Results: 54 pts (22 Ph I, 32 Ph II) enrolled; 11 week 24 (previously reported, Nat Commun. 2023;14(1):7018). Secondary endpoints included OS
received pembrolizumab, 44 retifanlimab (1 crossover). 33 (61%) pts were ER+/PR+/HER2-, 18 and progression-free survival (PFS). Exploratory endpoint included ctDNA analysis. Circulating cell
(33%) TNBC, 3 (6%) HER2+. Median OS, PFS, ORR, and CBR were evaluated against two pivotal Ph 3 free DNA (cfDNA) was extracted from 343 plasma samples collected at baseline (n = 128), week 3
trials, ASCENT1 (SG in TNBC) and TROPiCS-022 (SG in HR+/HER2- MBC) (see Table 1). In ran- (n = 122), and week 13 (n = 93). For ctDNA detection, we performed low-coverage genome-wide
domized pts, C1 vs C2 CPI had PFS (3.7 vs 3.2 mos, P=0.09) and OS (11.4 vs 7.4 mos, P=0.19). Pts sequencing of cfDNA from the above-mentioned samples and from 55 plasma samples from
receiving Ph 3 formulation (w/o IFNg; N=37) had greater PFS (3.6 vs 2.6 mos, P=0.01) and OS (13.4 healthy donors to correct a possible batch effect. Results: Data cut-off was September 13, 2024,
vs 6.9 mos, P=0.01). Bria-IMT was well tolerated w/ no Tx related D/Cs. Conclusions: The Bria-IMT with a median follow-up of 21.7 months. Median OS was not significantly different between arms:
Ph 3 formulation cohort OS was comparable to ASCENT and TROPiCS-02 (13.43 vs 11.8, 14.4 mos), 25.1 vs. 20.9 months in arm A vs. B, respectively; HR 0.97 (95%CI 0.63-1.50, p = 0.90). The updated
exceeding TPC arms (6.9, 11.2 mos). CBR (61%) compared favorably to ASCENT (40%) and median PFS showed similar PFS: 4.8 vs. 5.4 months in arm A vs. B, respectively; HR 1.22, (95%CI
TROPiCS-02 (34%); ORR (14%) matched or exceeded TPC arms (4%, 14%). These outcomes were 0.84-1.78, p = 0.29). Among the 343 plasma samples analyzed, ctDNA detection declined from 77%
observed in a more heavily pretreated population, demonstrating Bria-IMT’s clinical activity. at baseline to 46% at week 3, to 18% at week 13. At any timepoint, CtDNA detection was sig-
Randomized Ph 2 results suggest efficacy and safety in heavily pretreated MBC, w/ no significant nificantly associated with worse PFS and OS (Table). Twelve patients across both arms exhibited
OS difference between C1 and C2 CPI initiation and 22% of pts still in active survival follow up. exceptional long responses, without progressive disease and still receiving the study treatment at
Superior outcomes w/ the Ph 3 formulation support its continued evaluation. A randomized Ph 3 data cutoff; all exceptional responders evaluable for ctDNA analysis had ctDNA clearance.
trial is ongoing, comparing Bria-IMT vs treatment of physician’s choice (NCT06072612). Clinical Conclusions: Theaddition of oleclumab to chemo-immunotherapy did not improve PFS or OS in
trial information: NCT03328026. Research Sponsor: BriaCell Therapeutics Corp. advanced TNBC. However, a subgroup of patients experienced exceptional long-lasting response,
Age (Median, Prior Therapies OS (Median, PFS (Median, ORR CBR
indicating potential benefit in selected cases. CtDNA detection was strongly associated with poorer
Trial (Cohort) Range) (Median) mos) mos) (%) (%) outcomes at all timepoints, underscoring its potential as a biomarker in this disease/setting.
Clinical trial information: NCT03616886. Research Sponsor: AstraZeneca.
Bria-IMT (Overall Cohort) 61 (38-81) 6 (2-13) 9.9 (1.8-30.3) 3.6 10% 55%
Bria-IMT 62 (44-80) 6 (2-13) 13.43 (1.8-30.3) 3.6 (1.8-16.5) 14% 61%
(Ph 3 Formulation) Timepoint ctDNA detection mPFS (95% CI)* P value mOS (95% CI)* P value
ASCENT (SG) 54 (27-82) 4 (2-17) 11.8 4.8 31% 40% Baseline Yes 4.6 (4.2-5.4) 0.002 19.3 (16.6-25.9) 0.007
ASCENT (TPC) 53 (27-81) 4 (2-14) 6.9 1.7 4% 8% No 9.2 (5.7-16.4) 37.5 (22.5-NE)
TROPiCS-02 (SG) 57 (49-65) 3 14.4 5.5 21% 34% Week 3 Yes 3.9 (2.7-4.7) ,.001 18.1 (12.8-26.5) 0.003
TROPiCS-02 (TPC) 55 (48-63) 3 11.2 4 14% 22%
No 5.7 (4.6-9.3) 25.7 (19.0-NE)
References: Week 13 Yes 0.7 (0.4-2.4) ,.001 8.4 (7.3-24.2) ,.001
Bardia A et al. J Clin Oncol. 2024 May 20;42(15):1738-1744Rugo, Hope S et al. The Lancet, Volume 402, Issue 10411, No 3.7 (3.4-6.2) 25.6 (22.1-34.5)
1423 – 1433.
*PFS and OS for Week 3 and 13 estimated from this time point.

1098 Poster Session 1099 Poster Session

Multiomic profiling of LRRC15 in triple negative breast cancer (TNBC). First Acute circulating tumor DNA dynamics during and after infusional therapy
Author: Dan Morgenstern Kaplan, University of Miami/Jackson Memorial Hospital, initiation. First Author: Briana To, The Ohio State University Comprehensive Cancer
Miami, FL Center, Columbus, OH
Background: Leucine-rich repeat-containing protein 15 (LRRC15) has emerged as a po- Background: For patients with advanced breast cancer, the proportion of tumor-derived
tential biomarker and therapeutic target for various cancers due to its high expression in DNA in circulation (‘tumor fraction’/TF) has been shown to be prognostic. There is
cancer-associated fibroblasts (CAFs) and role in tumor progression. High LRRC15 ex- evidence that early change in TF may correlate with response to therapy, potentially
pression is associated with poor prognosis in TNBC. This study aims to define the multiomic providing a rapid, minimally-invasive predictive biomarker. However, there is little known
profile of LRRC15 in TNBC. Methods: 3,038 TNBC samples were analyzed via Next-Gen- regarding TF dynamics during and in the hours immediately after infusion of targeted or
eration Sequencing (592, NextSeq; Whole Exome Sequencing, NovaSeq) and Whole cytotoxic therapies, including whether there is a ‘surge’ in ctDNA corresponding to acute
Transcriptome Sequencing (NovaSeq; Caris Life Sciences, AZ). Immune cell fractions were cell death. We hypothesized that tracking TF change peri-infusion would provide insight
estimated using WTS deconvolution (Quantiseq). Stromal cell abundance in the tumor regarding acute ctDNA dynamics. Methods: Banked plasma samples were derived from
microenvironment (TME) was estimated from RNA expression profiles using MCP Counter. a phase 1b trial of HSP90 inhibitor onalespib with paclitaxel in patients with advanced
LRRC15-high (H) and -low (L) tumors were classified by RNA expression above or below the triple negative breast cancer. Plasma was collected during the first cycle of therapy on
25th percentile. Real-world overall survival (OS) and treatment-related survival were derived study pre-infusion, end-of-infusion (EOI), then 0.5/1/2/4/6/8/24 hours post-infusion for
from insurance claims and calculated from tissue collection or treatment initiation to last 1) onalespib alone (day -7); 2) paclitaxel alone 7 days later (day 1); 3) onales-
contact using Kaplan-Meier. Statistical significance was assessed using chi-square and pib+paclitaxel 7 days later (day 8) for a total of maximum 26 time points per patient. 317
Mann-Whitney U tests with multiple comparison adjustments (q , .05). Results: LRRC15-H samples from 14 patients underwent shallow whole genome sequencing (sWGS) and TF
TNBC tumors had higher frequency of PIK3CA (25.9% vs 16.8), PIK3R1 (6.2% vs 1.6%), PTEN determination. The objective was to evaluate change in TF from pre-infusion to 6-hours
(11.3% vs 5.8%), but lower frequency of RB1 (7.8% vs 12.1%) and KMT2D (2% vs 4.4%)
and 24-hours post-infusion. Exploratory objectives included association of TF dynamics
compared to LRRC15-L, all q , 0.05. LRRC15-H had higher PD-L1 positivity (32.3% vs 24.5%,
with progression-free survival (PFS) and overall survival (OS). Results: 313/317 (98.7%)
q , 0.05). Analysis of immune cells showed LRRC15-H TNBC had higher infiltration of B
of available plasma samples completed sWGS. Of these, 104/313 (33.2%) were collected
cells (4.2% vs 3.5%), M1 macrophages (4.3% vs 2%), M2 macrophages (4% vs 2.5%), Tregs
(1.9% vs 1.1%), neutrophils (4.9% vs 3.9%), CD8+ T cells (0.4% vs 0.1%), but lower dendritic
on onalespib alone, 114/313 on paclitaxel alone (36.4%), and 95/313 (30.4%) on ona-
cells (2.5% vs 3%), all q , 0.05. LRRC15-H tumors had greater abundance of CAFs (575.6 vs lespib+paclitaxel. For the co-primary objectives, there was a significant decline in TF
93.78, 6.14 fold change (FC)) and endothelial cells (7.3 vs 3.7, 1.97 FC), all q , 0.05. LRRC15- from pre-infusion to 6 hours for paclitaxel alone (Wilcoxon signed rank p = 0.03) but no
H had higher T-cell inflamed score (71.5 vs -77) and IFNg score (-0.14 vs -1.72), all q , 0.05. significant change for onalespib alone/onalespib+paclitaxel or from pre-infusion to 24
LRRC15-H tumors had higher expression of immune checkpoint genes (CD274, PDCD1, hours for any treatment group (all Wilcoxon signed rank p . 0.05). There was a sig-
PDCD1LG2, CTLA4, LAG3, HAVCR2, FOXP3, IDO1, TNFSF14, TIGIT, BTLA, CEACAM1, CD47, nificant decline in TF from pre-infusion day -7 (median TF 16%) to 24 hours after C1D8
CD80, CD86, CD160, CD274; FC 1.2-2.5, q , 0.05). LRRC15-H was associated with better OS (median TF 6.5%, Wilcoxon signed rank p = 0.004). Baseline TF$20% was associated
(mOS: 24.7 vs 13.6 months; HR 0.61, 95% CI 0.53-0.7, p , 0.001). Post-pembrolizumab with significantly worse PFS (log-rank p = 0.002) with a trend toward worse OS (log-rank
survival was longer for LRRC15-H patients (mOS: 27.2 vs 19.4 months; HR 0.61, 95% CI 0.42- p = 0.067) but categorization of TF change using ctDNA-RECIST was not associated with
0.89, p = 0.01). Conclusions: LRRC15-H TNBC exhibited better outcomes with pem- significant differences in PFS or OS. Conclusions: In this study of . 300 plasma
brolizumab, likely due to higher immune cell fractions and increased CAFs. These findings timepoints during the first cycle of treatment on a phase Ib clinical trial, there was no
highlight TNBC heterogeneity and position LRRC15 as a potential biomarker for tumor significant ‘surge’ in ctDNA TF within minutes to 24 hours of infusion of onalespib,
stratification, a possible adverse prognostic biomarker and a positive predictive biomarker. paclitaxel or both in combination. However, there was a significant decline in TF over the
Ongoing phase I trials targeting LRRC15 show promise. Combining LRRC15-targeted first full cycle of therapy. This suggests that despite ctDNA half-life of minutes-to-hours,
therapies with immunotherapy may improve TNBC outcomes, warranting further validation consistent change in TF may not be detectable for days or weeks, providing important
in breast cancer models. Research Sponsor: None. insight in the design of studies evaluating ctDNA change as a minimally-invasive
biomarker. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 61s

1100 Poster Session 1101 Poster Session

A phase 1b study of Plk1 inhibitor onvansertib in combination with paclitaxel Camrelizumab plus nab-paclitaxel and cisplatin as first-line treatment for
in metastatic triple-negative breast cancer (mTNBC) patients. First Author: metastatic triple-negative breast cancer: A prospective, single-arm, open-
Antonio Giordano, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA label phase II trial. First Author: Biyun Wang, Department of Medical Oncology, Fudan
Background: TNBC represents 15-20% of all breast cancer and is characterized by a more ag- University Shanghai Cancer Center, Shanghai, China
gressive clinical course compared to other subtypes with response rate , 10% after 2-3 lines of Background: Platinum-based chemotherapy plays an important role in the treatment of
chemotherapy. Onvansertib is an oral polo-like kinase 1 (PLK1) ATP-competitive inhibitor with TNBC. Our previous research has demonstrated the superiority of nab-paclitaxel/cisplatin
preclinical data showing synergy when combined with paclitaxel (P) in TNBC models. Here, we report (AP) regimen as the initial treatment for metastatic TNBC(mTNBC; Xichun Hu,
safety and outcome data for subjects enrolled in a phase 1b clinical trial of onvansertib and P for
patients (pts) with mTNBC. Methods: Eligible pts received escalating doses of onvansertib, studied
2020ESMO). Camrelizumab is a humanized monoclonal antibody against PD-1. Herein, we
using a Bayesian Optimal Interval (BOIN) design, with a fixed dose of P to determine the maximum conducted this prospective, single arm, open-label phase II study to evaluate the efficacy
tolerated dose and recommended phase 2 dose (RP2D) of onvansertib. The primary objective was and safety of camrelizumab in combination with AP regimen as the first-line treatment of
the characterization of dose-limiting toxicity (DLT). Onvansertib was tested at 9, 12, and 18 mg/m² mTNBC (NCT04537286). Methods: Patients with untreated mTNBC received camreli-
dose levels (DL). Onvansertib was administered orally, once daily for 21 consecutive days, followed zumab (200 mg D1), nab-paclitaxel (125 mg/m2 D1,D8) and cisplatin (75 mg/m2 D1)
by 7 days off; P was administered intravenously at 80 mg/m2 once on days 1, 8, and 15 of every 28- intravenously every 3 weeks until disease progression or intolerable toxicity. The primary
day cycle. Exploratory objectives included pharmacokinetic (PK) and circulating tumor DNA ana- endpoint was progression-free survival (PFS). Secondary endpoints included objective
lyses. Results: 17 pts enrolled from September 2022 to August 2024. Median line of chemotherapy response rate (ORR), disease control rate (DCR),overall survival (OS) and safety. Ex-
for mTNBC was 3 (range 1-11), 14/17 pts received prior taxane, 7/17 immunotherapy (IO), and 13/17 ploratory analyses included immunohistochemistry and RNA sequencing of archival
a prior antibody drug conjugate (ADC). There were 3 pts enrolled at DL0 (9 mg/m²), 4 at DL1 (12 mg/
tumour samples. Results: A total of 90 patients were enrolled. Overall, median age was
m²), and 10 at DL2 (18 mg/m²). One pt in DL2 remains on treatment, and 16 are off study (11 pts
discontinued due to disease progression (PD) per RECIST 1.1; 3 due to clinical PD; 1 due to un- 51 years; 46.7% of patients had three or more metastatic sites; 78.9% of patients had
acceptable toxicity; 1 death unrelated to the study drug). DLTs were observed in 0/3 pts at DL0, 1/4 visceral involvement; 82.2% of patients had taxanes exposure. As of data cutoff (July 10th
(25%) at DL1, and 3/10 (30%) at DL2. Common adverse events were anemia (47% $ Grade 2, 12% 2024), median duration of follow-up was 18.1 months. Median PFS was 11.8 (95%CI 10.1-
Grade 3), decreased neutrophil count (47% $ Grade 2, 24% Grade 3-4), and fatigue (24% $ Grade 2, 13.6) months and median OS was 27.1 (95%CI 22.1-33.6) months. ORR was 71.1% and
6% Grade 3). Best responses included 24% partial response (PR, 2/4 confirmed) and 24% stable DCR was 86.7%. Median time to response was 1.5 months. TRAEs were reported in all
disease (2/4 SD $ 12 weeks) (Table 1). All 4 responders were treated in DL2 (18mg/m2), 3/4 pts patients while grade 3-4 TRAE occurred in 55.6% patients, including neutropenia (34.4%),
received prior P (2/4 in mTNBC setting) and IO (all in mTNBC), 2/4 received an ADC. The RP2D of leukemia (24.4%), and anemia (10.0%). irAEs were reported in 57.8% patients, including
onvansertib in combination with P is 18 mg/m². PKs and other biomarkers will be presented. RCCEP (45.5%), rash (11.1%), pneumonitis (10.0%), while grade 3-4 irAEs only occurred in
Conclusions: The combination of onvansertib and P demonstrated a safe toxicity profile and 4.4% patients. Three-months landmark analyses showed that patients with irAE have
promising clinical activity in pretreated mTNBC pts and warrant further exploration of the com-
bination at the RP2D. Clinical trial information: NCT05383196. Research Sponsor: METAVIVOR;
significantly longer OS than those without (29.3 vs. 22.1 months, P = 0.018). Exploratory
Cardiff Oncology. analyses demonstrated that patients with PDL1 CPS $10 had significantly longer PFS
(13.7 vs 11.4 months, P = 0.039). Patients with high TILs had significantly longer OS (23.1
Best response per RECIST 1.1 among different DL.
vs.10.3 months, P = 0.003). The proportion of PDL1 positive (CPS $1) patients was 81.8%
Response All Pts (N=17) DL0 (N=3) DL1 (N=4) DL2 (N=10) in basal compared to 0% in non-basal subtype (P = 0.023). Hallmark pathway analysis
PR 4 (23.5%) 0 (0.0%) 0 (0.0%) 4 (40.0%) showed that the activation of DNA repair pathway (HR,11.6, 95%CI 2.4-55.7, P = 0.002)
Confirmed PR 2 (11.8%) 0 (0.0%) 0 (0.0%) 2 (20.0%) and MYC target pathway (HR,7.4,95%CI 1.9-28.2, P = 0.004) was significantly associated
Unconfirmed PR 2 (11.8%) 0 (0.0%) 0 (0.0%) 2 (20.0%)
SD 4 (23.5%) 2 (66.7%) 2 (50.0%) 0 (0.0%) with shorter PFS, while the activation of KRAS signaling (HR,3.2, 95%CI 1.1-9.7, P = 0.035)
SD > 12 wks 2 (11.8%) 2 (66.7%) 0 (0.0%) 0 (0.0%) was significantly associated with worse OS. Conclusions: Camrelizumab plus AP as first-
SD < 12 wks 2 (11.8%) 0 (0.0%) 2 (50.0%) 0 (0.0%) line treatment in patients with mTNBC demonstrated satisfying efficacy with manageable
PD 9 (52.9%) 1 (33.3%) 2 (50.0%) 6 (60.0%)
By RECIST 1.1 8 (47.1%) 1 (33.3%) 1 (25.0%) 6 (60.0%)
toxicity. Randomized controlled trial is warranted in the future. Clinical trial information:
Clinical PD 1 (5.9%) 0 (0.0%) 1 (25.0%) 0 (0.0%) NCT04537286. Research Sponsor: None.

1102 Poster Session 1103 Poster Session

Safety and efficacy of the anti-TROP2 antibody-drug conjugate (ADC) SHR-A1811 plus adebrelimab in unresectable or metastatic triple-negative
IBI130 in patients (pts) with advanced triple-negative breast cancer (TNBC) breast cancer: Results from a phase 1b/2 expansion cohort. First Author: Yan
and other solid tumors: Results from the phase 1 study. First Author: Fan Wu, Liang, The First Affiliated Hospital of Nanjing Medical University, Nanjing, China
Fujian Cancer Hospital, Fuzhou, China Background: SHR-A1811 is a novel HER2-directed antibody-drug conjugate with promising
Background: TROP2 is a promising therapeutic target in various solid tumors. IBI130 is antitumor activity in breast cancer (BC), yielding a confirmed objective response rate (ORR) of
composed of an anti-TROP2 antibody conjugated to the camptothecin derivative NT1. 79.4% in HER2 positive BC, 60.9% in HER2 low-expressing BC, and 52.0% in triple-negative BC
Herein, we report the multi-regional, first-in-human, phase 1 study of IBI130. (TNBC) as monotherapy. We evaluated SHR-A1811 in combination with adebrelimab (anti-PD-L1
Methods: Eligible pts with unresectable locally advanced or metastatic solid tumors who antibody), pyrotinib (irreversible, pan-HER receptor tyrosine kinase inhibitor), pertuzumab, or
failed or intolerant to standard treatment were enrolled. The study included dose esca- albumin-bound paclitaxel in unresectable or metastatic breast cancer in an open-label, dose-
finding and efficacy expansion phase 1b/2 study. Here, we report the safety and efficacy results of
lation and dose expansion. IBI130 was intravenously administered at 1/2/4/6/8/10/12 mg/
the SHR-A1811 plus adebrelimab cohort. Methods: Patients (pts) with unresectable or metastatic
kg Q3W during dose escalation, which was guided by modified continuous reassessment TNBC and $1 line of prior treatment received intravenous SHR-A1811 at an escalating dose of 4.8
method (mCRM) according to Bayesian logistic regression model (BLRM) and escalation mg/kg and 5.6 mg/kg Q3W, in combination with adebrelimab (1200 mg Q3W) in phase 1b part. In
with overdose control (EWOC) principle. Primary endpoint was safety. Secondary endpoint phase 2 part, TNBC pts with no systematic antitumor therapy in the recurrent or metastatic setting
was efficacy assessed by investigator per RECIST v1.1 including objective response rate were treated with SHR-A1811 at 4.8 mg/kg Q3W plus adebrelimab. Primary endpoints were safety
(ORR), disease control rate (DCR), duration of response (DoR) and progression-free survival and ORR. The data cutoff date was Nov 30, 2024. Results: Fifty TNBC pts were enrolled in total. In
(PFS). Results: As of Dec 15, 2024, 71 pts were enrolled from China and Australia (median phase 1b, 8 pts were enrolled and treated. No DLT was observed. The confirmed ORR was 66.7%
age: 60 years [range: 30-81], female: 85.9%, Caucasian: 16.9%, ECOG PS 1: 48.6%; prior (2/3) and 60.0% (3/5) in the 4.8 mg/kg and 5.6 mg/kg dose group, respectively. In phase 2, 42
lines of anticancer treatment$2: 63.8%). Median follow-up of the study was 4.6 months treatment naive TNBC pts were enrolled. 13 (31.0%) pts had $3 metastases sites, 22 (52.4%) pts
(range: 0.8-9.7). No dose-limiting toxicity (DLT) was observed across all dose levels during were HER2-low (IHC 2+/ISH- or IHC 1+)/ultra-low (IHC 0-1), 20 (47.6%) pts were HER2-nul (IHC 0),
dose escalation (n = 18). Median treatment duration was 18 weeks (range: 3-45) with 40 and 30 (71.4%) pts were PD-L1-positive (CPS $1). At the time of data cutoff, the median follow-up
(56.3%) pts still on treatment. Treatment-emergent adverse events (TEAEs) occurred in 68 time was 4.6 mo (range, 0.2-10.4). Among efficacy evaluable TNBC pts, the overall ORR was 66.7%
(95.8%, with 90.1% treatment-related adverse events [TRAEs]) pts including grade 3 (G3) (26/39) (Table). ORR was 77.8% (21/27) in the PD-L1-positive subgroup. The 6-month PFS rate
events in 17 (23.9%, with 15.5% TRAEs) pts. No grade 4-5 events occurred. Common TEAEs was 86.2%. SHR-A1811 plus adebrelimab was well tolerated with no new safety concerns
identified. Treatment-emergent adverse events of grade $3 occurred in 26 (61.9%) out of 42 pts in
($30%) were stomatitis (52.1%, with 9.9% G3), nausea (31.0%, with 2.8% G3) and rash
phase 2, with decreased neutrophil count (45.2%), decreased white blood cell count (33.3%), and
(31.0%, with 1.4% G3). Interstitial lung disease occurred in 1 pt (1.4%, G1). Only 1 pt (1.4%) decreased lymphocyte count (9.5%) being the most common. Conclusions: SHR-A1811 plus
had G3 lymphocyte count decreased. Other $G3 hematological toxicities were not ob- adebrelimab had a good safety and tolerability profile. The combination showed encouraging
served. TRAEs led to dose reduction in 5 (7.0%) pts and treatment discontinuation in 1 antitumor activity in unresectable or metastatic TNBC, irrespective of HER2 or PD-L1 expression
(1.4%) pts. Efficacy of IBI130 was evaluable in 30 pts with TNBC treated at 4/6/8/10 mg/kg status. Clinical trial information: NCT05353361. Research Sponsor: Jiangsu Hengrui
(all stage IV, and 96.7% had failed or were intolerant to taxanes). The overall ORR was Pharmaceuticals.
50.0% (95% CI: 31.3-68.7) and DCR was 83.3% (95% CI: 65.3-94.4). As for different dose
Phase 2 preliminary efficacy summary1.
levels, ORR and DCR were 40.0% (95% CI: 5.3-85.3) and 60.0% (95% CI: 14.7-94.7) for 4 mg/
SHR-A1811 4.8 mg/kg + adebrelimab
kg (n = 5), 40.0% (95% CI: 5.3-85.3) and 80.0% (95% CI: 28.4-99.5) for 6 mg/kg (n = 5),
50.0% (95% CI: 18.7-81.3) and 100% (95% CI: 69.2-100.0) for 8 mg/kg (n = 10), 60.0% (95% CPS ‡1 (N = 30) CPS <1 (N = 12) Total (N = 42)
CI: 26.2-87.8) and 80.0% (95% CI: 44.4-97.5) for 10 mg/kg (n = 10). DoR and PFS data were ORR, Overall2 21/27 (77.8) 5/12 (41.7) 26/39 (66.7)
not mature as of the cutoff date. Conclusions: IBI130 was well tolerated featured by HER2-low/-ultralow 11/13 (84.6) 4/9 (44.4) 15/22 (68.2)
HER2-nul 10/14 (71.4) 1/3 (33.3) 11/17 (64.7)
superiority of hematological safety, and encouraging efficacy of IBI130 was observed in 6-mo PFS rate, % (95% CI) 88.9 (43.3, 98.4) 78.8 (38.1, 94.3) 86.2 (60.7, 95.7)
advanced TNBC, supporting its potential as a best-in-class TROP2 ADC. Clinical trial 1
HER2 and PD-L1 results were based on central lab assessment.
information: NCT05923008. Research Sponsor: Innovent Biologics (Suzhou) Co., Ltd. 2
Data are n/N1(%) with N1 = the number of efficacy evaluable patients.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
62s BREAST CANCER—METASTATIC

1104 Poster Session 1105 Poster Session

ETER901: A randomized, open-label, phase III trial of anlotinib in combi- Clinical, sociodemographic, and facility-related determinants of immuno-
nation with anti-PD-L1 antibody benmelstobart (TQB2450) versus nab- therapy use in metastatic triple-negative breast cancer. First Author: Ismail
paclitaxel in first-line treatment of recurrent or metastatic triple-negative Ajjawi, Yale School of Medicine, New Haven, CT
breast cancer. First Author: Jiayu Wang, National Cancer Center/National Clinical Background: Immunotherapy has emerged as a promising treatment option for met-
Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and astatic triple-negative breast cancer (mTNBC), yet the factors influencing its adoption
Peking Union Medical College, Beijing, China remain poorly understood. This study investigates the clinical, sociodemographic, and
Background: Recurrent or metastatic triple-negative breast cancer (TNBC) represents an facility-related determinants of immunotherapy use in patients with mTNBC from 2015
aggressive malignancy with unfavorable prognoses. Benmelstobart (TQB2450) is a hu- to 2020, utilizing data from the National Cancer Database (NCDB). Methods: We
manized monoclonal antibody targeting PD-L1, and anlotinib (ALTN) is an anti-angiogenic conducted a retrospective cohort study of mTNBC patients from the NCDB between
oral multi-target tyrosine kinase inhibitor. Herein, we present the findings of a randomized, 2015 and 2020, categorizing them into two groups: those who received immunotherapy
open-label, phase 3 study comparing the combination of benmelstobart plus ALTN with and those who did not. Patients were excluded if they had missing data on key variables
nab-paclitaxel as first-line treatments for patients (pts) with recurrent or metastatic TNBC. such as immunotherapy receipt and clinical characteristics (e.g., tumor stage, subtype).
Methods: In this phase 3 trial, patients with previously untreated stage IV or recurrent/ Univariable and multivariable logistic regression analyses were performed to identify
metastatic TNBC were randomly allocated in a 1:1 ratio. One group received 1200 mg of factors influencing immunotherapy adoption. The impact of immunotherapy on overall
intravenous benmelstobart on day 1, along with 12 mg of oral ALTN from days 1 to 14, survival was assessed using Cox proportional hazards regression analysis. Overall
following a 3-week cycle. The other group was administered 100 mg/m² of intravenous survival between the two groups was compared using the log-rank test. Results: A total
nab-paclitaxel on days 1, 8, and 15 within a 4-week cycle. Randomization was stratified of 1,887 mTNBC patients were included in the study: 1,656 (87.8%) did not receive
based on whether patients had received neoadjuvant or adjuvant taxane therapy and the immunotherapy, and 232 (12.2%) received immunotherapy. Multivariable logistic re-
presence or absence of liver or brain metastases at baseline. The primary endpoint was gression identified several factors associated with immunotherapy use. Later year of
progression-free survival (PFS), evaluated by the blinded independent central review by diagnosis (2018-2020: OR 5.35, p , 0.001) and academic facilities (OR 1.43, p = 0.044)
RECIST version 1.1. Results: The initial plan was to enroll 332 pts in this trial. However, were positively associated with immunotherapy use. In contrast, older age (71+: OR 0.49,
due to the COVID-19 pandemic, the enrollment process was delayed, and recruitment was p = 0.019), facilities in rural areas (OR 0.43, p = 0.042), Black race (OR 0.73, p = 0.039),
terminated in January 2023. Eventually, 147 pts were randomized (with a median follow-up Hispanic ethnicity (OR 0.53, p = 0.026), and higher Charlson comorbidity scores (OR 0.31,
of 14 months), among whom 75 were assigned to the benmelstobart plus ALTN group and p = 0.035 for scores $2) were associated with a lower likelihood of receiving im-
72 to the nab-paclitaxel group. In the intention-to-treat analysis, as assessed by the
munotherapy. Insurance status did not significantly influence immunotherapy use. Log-
investigators, the median PFS was 7.85 months for the benmelstobart plus ALTN
rank test showed that patients receiving immunotherapy had significantly improved
combination, in contrast to 5.55 months for nab-paclitaxel (hazard ratio, 0.70; 95%
survival compared to those who did not (Figure 1). The median survival for patients
confidence interval, 0.46 to 1.06; P = 0.1687). The median overall survival was 35.81
receiving immunotherapy was 2.21 years (95% CI: 1.80–2.96), compared to 1.01 years
months for study group and 21.03 months for control group (hazard ratio, 0.78; 95%
(95% CI: 0.93–1.11) for those not receiving immunotherapy (log-rank p , 0.001). Cox
confidence interval, 0.49 to 1.24; P = 0.2625). Grade $3 drug-related adverse events
occurred in 56.5% of the patients in the study group and 36.6% in the control group. The regression analysis showed that immunotherapy use was associated with a significantly
most prevalent grade $3 adverse events in the study group were hypertension (28.0%) and reduced risk of death (HR 0.59, 95% CI: 0.46–0.77, p , 0.001). Conclusions: Immu-
hypertriglyceridemia (13.3%). Conclusions: The combination of benmelstobart plus ALNT notherapy use in mTNBC has increased in recent years, with clinical, sociodemographic,
might extend both progression-free survival and overall survival in the first-line treatment and facility-related factors influencing its adoption. Patients receiving immunotherapy
of patients with recurrent or metastatic TNBC. The adverse events were in line with the had significantly better survival outcomes. Our findings highlight the importance of
previously established safety profiles of each individual agent. (Funded by Chia Tai addressing disparities in access to immunotherapy, particularly related to race, age,
Tianqing Pharmaceutical Group Co., Ltd. [Link] number, NCT04405505). ethnicity, and comorbidity burden, to ensure equitable treatment and outcomes for all
Clinical trial information: NCT04405505. Research Sponsor: Chia Tai Tianqing Pharma- mTNBC patients. Research Sponsor: None.
ceutical Group Co., Ltd.

1106 Poster Session 1107 Poster Session

Chemokines as predictive biomarkers for immune checkpoint inhibitor (ICI) Comprehensive molecular and immune characterization of adrenergic
efficacy in triple negative breast cancer (TNBC). First Author: Shipra Gandhi, stress-signaling receptor ADRB2 in triple negative breast cancer (TNBC).
Roswell Park Comprehensive Cancer Center, Buffalo, NY First Author: Sachin Kumar Deshmukh, Caris Life Sciences, Phoenix, AZ
Background: TNBC, although an aggressive breast cancer (BC) subtype, is highly immunogenic and Background: Chronic stress-mediated b2-adrenergic receptor (b2-AR) signaling promotes
the only BC subtype where the ICI pembrolizumab is approved. However, predictive biomarkers for tumor growth via immunosuppression in the tumor microenvironment (TME) in preclinical
pembrolizumab benefit are limited. The chemokines CXCL9 and CXCL10 attract CD8+ T cells into the models. Blockade of b2-AR has shown higher survival benefit in patients with TNBC in
tumor microenvironment (TME) and are associated with chemotherapy benefit, but little is known observational studies compared to other breast cancer (BC) subtypes. However, the mo-
about their role in prediciting pembrolizumab benefit in TNBC. We investigated the association of
lecular and immunological features associated with ADRB2 (gene for b2-AR) gene ex-
CXCL9, CXCL10 and their cognate receptor CXCR3 with TME and ICI efficacy. Methods: 3,038 TNBC
samples were analyzed via NGS (592-gene panel, NextSeq; WES/WTS, NovaSeq; Caris Life Sciences,
pression in TNBC are unknown, prompting this investigation. Methods: 3,038 TNBC
Phoenix, AZ). Tumor mutational burden (TMB) totaled somatic mutations per tumor (high . 10 mt/ samples were analyzed via NGS (592-gene panel, NextSeq; WES/WTS, NovaSeq; Caris Life
MB). Immune cell fractions were estimated using WTS deconvolution (Quantiseq). CXCL9/CXCL10/ Sciences, Phoenix, AZ). Immune cell fractions were calculated by deconvolution of WTS:
CXCR3-high (H) and -low (L) tumors were classified by RNA expression above or below the 50th Quantiseq. TNBC ADRB2-high(H) and ADRB2-low(L) RNA expression were classified as
percentile. Real-world overall survival (OS) was derived from insurance claims and calculated from above or below the 50th percentile, respectively. Real-world overall survival (OS) was
start of pembrolizumab to last contact using Kaplan-Meier. Statistical significance was assessed obtained from insurance claims and calculated from tissue collection to last contact using
using chi-square and Mann-Whitney U with multiple comparison adjustments (q , .05). Results: Kaplan-Meier estimates. Statistical significance was assessed using chi-square and Mann-
TNBC expressed higher levels of CXCL9 and CXCL10 (median (TPM): 5.3 and 14.7) compared to N = Whitney U tests with multiple comparison adjustments (q , 0.05). Results: ADRB2 gene
1,082 HER2+ (4.8 and 10.3, p , 0.05) and N = 4,918 HR+HER2- (2.7 and 7, q , .05) BC. CXCR3 expression was lowest in TNBC (median (TPM: 1.6) compared to N = 453 HR-HER2+ (1.9), N
expression was higher in TNBC compared to HR+HER2- (1.9 vs 1.7, q , .05), but no difference when = 629 HR+HER2+ (2.0) and N = 4,918 HR+HER2- (2.2) BC (all q , 0.05). African American or
compared to HER2+ (1.9 vs 2, q = 0.97) BC. CXCL9/CXCL10/CXCR3-H TNBC had higher median OS post
Black patients (N = 670) had lower expression of ADRB2 compared to European American or
pembrolizumab [CXCL9-H vs -L: 26.5 vs 15.7 months (mo), HR: 0.65 (95% CI 0.5-0.84); CXCL10-H vs -L:
26.0 vs 20.6 mo, HR 0.74 (0.57-0.95); CXCR3-H vs -L: 32.6 vs 18.3 mo, HR 0.68 (0.52-0.88), all p , White (N = 1,412) TNBC patients (1.3 vs 1.7, q , 0.05). ADRB2-H TNBC had higher mutation
0.05]. CXCL9-H, CXCL10-H and CXCR3-H had higher PD-L1 positivity (22C3), TMB high, higher T cell frequency of PIK3CA (21% vs 15.4%), CDH1 (7% vs 3.5%), NF1 (8% vs 4%), AKT1 (3.5% vs
inflamed score, TP53 mutations, elevated B and CD8+T cells infiltration, but not neutrophils, and higher 2.1%), but lower frequency of TP53 (81.6% vs 87.5%), NOTCH1 (2.5% vs 4.5%) and NOTCH3
expression of immune checkpoint genes (Table). Conclusions: High CXCL9/CXCL10/CXCR3 ex- (4.4% vs 11.7%) compared to ADRB2-L, all q , 0.05. ADRB2-H had greater PD-L1 (22C3)
pression is associated with longer survival in patients with TNBC post pembrolizumab, and char- positivity (39.1% vs 30.2%, q , 0.05), higher % of B cells (4.5 vs 3.4), M1 Mw (3.4 vs 2.8), M2
acterized by an immune-enriched TME. Further investigation is needed to evaluate this chemokine axis Mw (3.9 vs 2.2), Tregs (2.2 vs 1.3), NK cells (3.1 vs 2.6), DC (3.1 vs 2.9), CD8+ T cells (0.9 vs
in TNBC and its potential as a therapeutic target to enhance ICI efficacy. Research Sponsor: NIH 0.2), all q , 0.05. ADRB2-H TNBC had higher T-cell inflamed score (95 vs -80), IFNg score
(NCATS, NCI); K08CA279766-01A1. (-0.23 vs -0.37), MAPK activation score (-0.46 vs -1.7), all q , 0.05; and higher expression of
TME characteristics by CXCL9, CXCL10 and CXCR3 expression. immune checkpoint genes (CD274, PDCD1, PDCD1LG2, CTLA4, LAG3, HAVCR2, FOXP3,
CXCL9 CXCL10 CXCR3 IDO1, TNFSF14, TIGIT, BTLA, CEACAM1, CD47, CD274; fold change: 1.6-3.7, all q , 0.05).
ADRB2-H tumors had higher expression of genes related to inflammatory response, IFNg
High Low q-value High Low q-value High Low q-value
response, IL6-JAK-STAT3 signaling (normalized enrichment score (NES): 1.9 – 2.1), while
PD-L1 % 54 14 ,.05 54 14.7 ,.05 49 19 ,.05 ADRB2-L had enrichment of MYC targets V1, MYC targets V2, E2F targets and G2M
TMB high % 14.3 8 ,.05 12.6 9.7 ,.05 12.4 9.8 ,.05
B cell (median %) 4.4 3.5 ,.05 4.3 3.6 ,.05 5 3.4 ,.05 checkpoint (NES: 2.5 – 4.2), all FDR , 0.01. ADRB2-H TNBC had better OS (mOS: 23.6 vs
CD8+T cell (median %) 1.2 0 ,.05 1 0 ,.05 1.3 0 ,.05 18.6 months; HR 0.81, 95% CI 0.73-0.89, p , 0.0001) compared to ADRB2-L. Conclu-
Neutrophil (median %) 4.2 4.5 ,.05 4.2 4.4 ,.05 4.2 4.3 0.2 sions: High ADRB2 expression in TNBC is associated with better survival and an immune
T cell inflamed score 100 -84 ,.05 98 -84 ,.05 108 -100 ,.05 enriched TME, elevated immune checkpoints and other targetable vulnerabilities. Future
CTLA4 (median TPM) 3.2 0.7 ,.05 3 0.8 ,.05 3.4 0.7 ,.05
LAG3 (median TPM) 6.6 2.1 ,.05 6.9 2 ,.05 6.8 2 ,.05 studies are needed to investigate ADRB2 as a potential stress biomarker and therapeutic
TP53 mutation % 88 81 ,.05 90.5 78.6 ,.05 86 83 0.08 target. Research Sponsor: NIH (NCATS and NCI); K08CA279766-01A1.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 63s

1108 Poster Session 1109 Poster Session

Enhanced efficacy of inavolisib combined with anti-PD-1 or anti-HER2 Efficacy and safety of RC48-ADC in triple-negative breast cancer subtypes:
antibody in treating brain metastases from breast cancer. First Author: FUSCC-TNBC-umbrella trial results. First Author: Yin Liu, Fudan University
Jian-Li Zhao, Sun Yat-sen Memorial Hospital, Sun Yat-Sen University, Guangzhou, China Shanghai Cancer Center, Shanghai, China
Background: The PI3K/AKT/mTOR signaling pathway is a crucial regulatory pathway Background: RC48-ADC is a novel HER2-targeting antibody-drug conjugate. This study evaluated
involved in cell proliferation, survival, migration, and metabolism. This dysregulation can RC48-ADC in pretreated triple-negative breast cancer (TNBC) patients with low HER2 expression,
occur through various mechanisms, such as PIK3CA gene mutations and PTEN gene stratified by AR status. Methods: In this phase Ib/II trial, pretreated metastatic TNBC patients with low
HER2 expression were enrolled: RL group (LAR subtype, n=20) and RO group (non-LAR subtype, n=20).
loss. The research of PI3K inhibitors has made significant progress in the treatment of All received RC48-ADC 2.0 mg/kg intravenously every 2 weeks. Primary endpoint: objective response
ER-positive and HER2-negative breast cancer. Alpelisib is the only approved PI3K in- rate (ORR). Secondary endpoints: progression-free survival (PFS), overall survival (OS), disease control
hibitor for treating PIK3CA mutation-positive breast cancer. The SOLAR-1 trial dem- rate (DCR), and safety. Results: 40 heavily pretreated patients were enrolled (median 2 previous lines,
onstrated that Alpelisib combined with endocrine therapy significantly prolongs range 1-7). In the overall population, best ORR was 35.0% (confirmed ORR: 32.5%), DCR 47.5%, median
progression-free survival in these patients. However, despite improving PFS, the side PFS 4.0 months. RL group showed better outcomes: best ORR 45.0% vs 25.0%, confirmed ORR 40.0%
effects of PI3K inhibitors pose limitations on their widespread application. Conse- vs 25.0%, DCR 50.0% vs 45.0%, median PFS 4.9 vs 3.1 months. Median OS was not reached in RL group
quently, researchers are exploring next-generation PI3K inhibitors with improved safety vs 16.6 months in RO group. Most common treatment-related adverse events (TRAEs) were AST
increased (70% vs 40%) and ALT increased (65% vs 20%), mostly grade 1-2. Peripheral neuropathy
and efficacy. Inavolisib is a novel, highly selective PI3Ka inhibitor that shows better
occurred in 15% (RL) and 10% (RO) patients. Hematologic toxicities were mild. No treatment-related
tolerability and safety compared to existing PI3K inhibitors and has demonstrated deaths occurred. Conclusions: RC48-ADC showed promising antitumor activity with manageable
promising antitumor effects in clinical trials. Building on this, our study aims to identify safety in pretreated TNBC patients with low HER2 expression, particularly in LAR subtype. The 35.0%
the optimal treatment regimen combining Inavolisib with various breast cancer ther- ORR in heavily pretreated TNBC warrants further investigation in biomarker-selected populations.
apies to effectively target brain metastases. Methods: We established a brain me- Clinical trial information: NCT03805399. Research Sponsor: None.
tastasis model in C57BL/6 mice by intracardiac injection of control (triple-negative) and Efficacy and key safety outcomes of RC48-ADC in overall population and by subgroups.
hHER2+ Py8119 breast cancer cells. In addition to the Inavolisib monotherapy and Outcomes Overall (n=40) RL Group (n=20) RO Group (n=20)
vehicle control groups, Inavolisib was combined with a PD-1 antibody or albumin-bound
Efficacy
paclitaxel in the triple-negative model. In the HER2+ model, Inavolisib was combined Confirmed ORR, n (%) 13 (32.5) 8 (40.0) 5 (25.0)
with Tucatinib, trastuzumab, or SHR-A1811 (an ADC drug targeting HER2). We moni- Best ORR, n (%) 14 (35.0) 9 (45.0) 5 (25.0)
tored changes in body weight and survival rates in each group and assessed brain DCR, n (%) 19 (47.5) 10 (50.0) 9 (45.0)
Median PFS, months 4 4.9 3.1
metastasis using IVIS small animal in vivo imaging. Results: In the triple-negative Median OS, months NR NR 16.6
model, Inavolisib monotherapy or its combination with albumin-bound paclitaxel re- Best Response, n (%)
duced intracranial tumor size but did not significantly extend mouse survival. Con- CR 1 (2.5) 1 (5.0) 0 (0.0)
versely, the combination of Inavolisib and PD-1 antibody significantly prolonged overall PR 13 (32.5) 8 (40.0) 5 (25.0)
SD 5 (12.5) 1 (5.0) 4 (20.0)
survival in triple-negative breast cancer mice. In the HER2+ breast cancer model, all PD 21 (52.5) 10 (50.0) 11 (55.0)
three combination therapies reduced tumor burden and extended survival compared to Selected TRAEs, n (%)
monotherapy. However, overall, the combination with trastuzumab achieved unex- AST increased
- Grade 1-2 22 (55.0) 14 (70.0) 8 (40.0)
pectedly good results, which were comparable to SHR-A1811 and superior to Tucatinib. - Grade ‡3 0 (0.0) 0 (0.0) 0 (0.0)
Conclusions: Our findings suggest that the combination of the PI3K inhibitor Inavolisib ALT increased
with anti-PD-1 or anti-HER2 antibody therapy may offer an effective strategy for treating - Grade 1-2 17 (42.5) 13 (65.0) 4 (20.0)
- Grade ‡3 0 (0.0) 0 (0.0) 0 (0.0)
brain metastases in breast cancer. This discovery provides new insights and possibilities Peripheral neuropathy
for improving treatment options in breast cancer brain metastasis. Further research is - Grade 1-2 3 (7.5) 2 (10.0) 1 (5.0)
needed to validate the efficacy and safety of this combination therapy. Research - Grade ‡3 2 (5.0) 1 (5.0) 1 (5.0)
Sponsor: None.

1110 Poster Session 1111 Poster Session

Artificial Intelligence-based tumor microenvironment and PD-L1 analysis Geographic access to triple negative breast cancer (TNBC) clinical trials: Are
using digital pathology to predict pembrolizumab response in metastatic trials located near Black women? First Author: Laura Burns Amin, The University of
triple-negative breast cancer. First Author: Jee Hung Kim, Division of Medical Texas MD Anderson Cancer Center, Houston, TX
Oncology, Department of Internal Medicine, Gangnam Severance Hospital, Yonsei Background: Despite recent treatment advances, TNBC has a poor prognosis relative to other
University College of Medicine, Seoul, NA, South Korea breast cancer subtypes. Black women in the U.S. are more likely to be diagnosed with TNBC, are
Background: The combination of pembrolizumab and chemotherapy improves survival in diagnosed at more advanced stages, and have higher mortality even after controlling for
programmed death ligand 1 (PD-L1) positive metastatic triple-negative breast cancer socioeconomic variables than women of other races. Although clinical trials are essential to
(mTNBC). However, responses vary even among PD-L1 positive, and predictive biomarkers improving TNBC treatment, Black women are underrepresented. We investigated the geographic
remain undefined. This study investigates the predictive biomarkers to pembrolizumab availability of TNBC clinical trials for Black women in the U.S. to elucidate potential trial access
through digital pathology and artificial intelligence (AI)-based tumor microenvironment limitations. Methods: All trials registered on [Link] as of 9/30/2024 were queried (N=
(TME) and PD-L1 analysis. Methods: We retrospectively analyzed 53 PD-L1 positive, mTNBC 510,397). Phase II and III interventional treatment trials in active status in the U.S. including
patients treated with pembrolizumab at Gangnam Severance Hospital (2017–2024). PD-L1 “breast” in the title or disease variable (n =449) were considered. We narrowed results to TNBC
positivity was defined as a combined positive score (CPS) $ 10. Immune phenotypes and trials through keyword searches of title and disease variables (e.g., TNBC, HER2 negative and
immune cell density in both tumor and stroma were analyzed in 67 H&E images using Lunit hormone receptor negative). We tabulated the number of trials per county and supplemented
with 5-year population estimates (2018-2022) and county adjacency data from the U.S. Census
SCOPE IO, an AI-powered whole slide image analyzer. PD-L1 CPS was analyzed in paired PD-
Bureau to evaluate for geographic and demographic differences in TNBC trial availability.
L1 staining images by both Lunit uIHCv2 analyzer and pathologists. Samples were cate-
Results: We identified 108 active TNBC trials (58 metastatic [54%], 50 nonmetastatic [46%]),
gorized as pre-(pre) or post-treatment (post). Pre-samples were collected before any therapy
including 87 Phase II (81%) and 21 Phase III (19%). There were 1,230 U.S. study sites, of which
exposure, while post-samples were obtained after recurrence following neo/adjuvant 217 had one active TNBC trial (18%), 529 had 2-4 trials (43%), and 484 had $5 trials (39%). Most
therapy. These features were analyzed for their association with pembrolizumab response sites had metastatic and nonmetastatic offerings (700, 57%) while 450 sites had only non-
and clinical outcome. Results: With a median follow-up of 13.2 months, the median age was metastatic trials (37%) and 80 had only metastatic trials (7%). State-level differences in trial
53 years, and 16 patients (22.5%) were de novo stage IV TNBC. AI-assessed PD-L1 positivity availability were observed (see Table). For example, 37% of Black and 34% of non-Black women
was seen in 52.2% (35/67) of cases, compared to 74.6% (50/67) by pathologist. Overall, AI- 18+ in Alabama had no TNBC trials in their or neighboring counties while all women 18+ in nine
based PD-L1 positive cases had a median progression-free survival (mPFS) of 8.8 months states had a trial available in at least a neighboring county. Conclusions: A geographical
(mo) vs 6.7 mo in PD-L1 negative (p = 0.028), while pathologist-reported cases showed 7.9 analysis of active Phase II and III therapeutic TNBC clinical trials found uneven trial availability
mo vs 6.3 mo, respectively (p = 0.17). AI-based PD-L1 positivity in pre-samples was as- across the country. Most study sites had , 5 TNBC trials available; a third of sites had no
sociated with better PFS with pembrolizumab (mPFS 7.7 mo vs 4.4mo, HR 0.32, p = 0.014), metastatic trials, suggesting that many women may have difficulty finding an applicable trial
while post-samples showed no significant association (mPFS 7.3 mo vs 6.4mo, HR 0.69, p = even when near to a site. On a national scale, distance does not appear to be a primary reason for
0.4). Notably, post-samples (50.0%) had a higher proportion of cases with AI-based CPS $ 10 disparities in TNBC trial participation for Black and non-Black women. Nevertheless, millions of
compared to pre-samples (36.4%), primarily driven by increased PD-L1-expressing mac- women live in areas without any trials, therefore expanding geographic reach is a necessary but
rophages, as revealed by AI-based cell composition analysis (22.7% vs 7.7%, p = 0.0007). insufficient approach to improve access. Research Sponsor: None.
When categorized by AI-based immune phenotype, notable differences were seen between
Black women Non-Black women Total in
pre/post-samples despite PD-L1 positivity. Post-samples showed a higher prevalence of the Category (Number of Counties) 18+ in millions (%) 18+ in millions (%) millions (%)
immune-desert phenotype, reflecting significant changes in the TME following prior therapy
exposure (40.0% vs 20.0%, p = 0.06). Conclusions: This study highlights the role of the TME Counties with trial (748) 14 (82%) 88 (77%) 102 (78%)
Trials only in adjacent 2 (12%) 21 (18%) 22 (17%)
and PD-L1 assessed by AI in predicting pembrolizumab response in mTNBC. While PD-L1 county (1452)
positivity in pre-samples was associated with outcome, PD-L1 positivity in post-samples No trials in county nor 1 (6%) 6 (5%) 7 (5%)
showed limited association with PFS, potentially influenced by immune desert phenotype adjacently (944)
and increased PD-L1-expressing macrophages. Research Sponsor: None. Total (3144) 17 (100%) 115 (100%) 131 (100%)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
64s BREAST CANCER—METASTATIC

1112 Poster Session 1113 Poster Session

68
Immunotherapy vs. chemotherapy run-in followed by pembrolizumab plus Trop2-targeted PET/CT with Ga-MY6349 for the diagnosis of primary and
nab-paclitaxel in metastatic triple negative breast cancer (mTNBC): Results metastatic breast cancer and evaluation towards patient stratification in
from a phase II study. First Author: Alessandro Leal, Perlmutter Cancer Center, NYU Trop2-targeted ADCs. First Author: Liang Zhao, The First Affiliated Hospital of
Langone Health, New York, NY Xiamen University, Xiamen, China
Background: Pembrolizumab (pembro) with chemotherapy has shown survival benefit in PD- Background: Trop2-targeted ADCs have demonstrated promising efficacy and have been
L1+ (CPS10+) mTNBC, but many responses are not durable, and patients with PD-L1 negative/ approved in patients with HR+HER2- and triple-negative breast cancer (TNBC). However,
low tumors do not benefit from the combination. Data from GeparNuevo and TONIC suggest that not all patients within these subtypes benefit equally from such treatment, highlighting the
induction therapy can remodel the tumor immune environment and improve responses. We have urgent need for developing tools for patient selection and stratification. We have pre-
conducted a trial with two run-in cohorts and mandatory serial tissue and blood collections in 50 viously developed a novel PET/CT imaging agent (68Ga-MY6349) that specifically targets
mTNBC patients, comparing pembro vs. nab-paclitaxel (nab-P). Methods: Single-arm, single Trop2, which has shown high specificity for Trop2 in preclinical and clinical studies (DOI:
institution phase II study (NCT02752685) to evaluate safety and clinical activity of nab- 10.1172/JCI185408). Methods: This study enrolled patients with newly diagnosed or
P+pembro in PD-L1 unselected mTNBC, 0-2 prior lines of chemotherapy allowed. Patients (n = previously treated breast cancer at the First Affiliated Hospital of Xiamen University
50) were enrolled sequentially into two cohorts: chemotherapy run-in (cTNBC, nab-P before nab-
between January 2024 and December 2024. All patients underwent paired 18F-FDG PET/CT
P+pembro) and immunotherapy run-in (iTNBC, pembro before nab-P+pembro). Serial tumor
and 68Ga-MY6349 PET/CT imaging. SUVmax derived from the two PET/CT modalities and
biopsies assessed by IHC (Dako 22C3), quantitative multiplex immunofluorescence (qMIF), and
gene expression (NanoString). Overall response rates assessed using irRECIST. Tumoral T- and
pathological results were recorded to evaluate the tumor uptake pattern and lesion de-
myeloid-cell phenotypes, peripheral lymphocyte-to-neutrophil ratio (LNR), and monocyte-to- tectability of the two imaging modalities. Results: A total of 61 patients were pro-
lymphocyte ratio (MLR) were correlated with overall response rate (ORR) and survival outcomes. spectively enrolled, including 7 true-negative and 54 true-positive cases. Among the 562
Results: 50 patients enrolled and completed treatment, for 80% of patients: treatment was 1L true-positive lesions, 68Ga-MY6349 uptake (SUVmax) was significantly associated with
for metastatic disease. Median follow-up is 19.9 months, clinical results for cTNBC and iTNBC breast cancer subtypes (P,0.001, Kruskal-Wallis H=34.9). SUVmax values were highest in
cohorts shown in table. Across both cohorts higher LNR was associated with improved OS (R= HR+/HER2- [7.2 (4.4–9.4)], followed by TNBC [5.2 (3.8–6.4)], HER2+ [4.8 (1.7–7.4)], and
0.37, p= 0.0075), conversely, higher MLR was associated with poorer OS (R= -0.46, p= 0.00087). HR+/HER2+ [3.3 (2.1–8.1)]. In HR+/HER2- subtypes, 68Ga-MY6349 demonstrated sig-
Tumor immune cell subpopulations showed no significant differences between iTNBC and nificantly higher uptake compared to 18F-FDG [7.2 (4.4–9.4) vs. 3.6 (2.3–5.5), P,0.001].
cTNBC at baseline. Analyses of on-treatment samples will be presented at the meeting. PD-L1 However, no significant difference regarding tumor uptake was observed in other sub-
expression, while not different at baseline, remained unchanged in cTNBC but increased types. In 27 patients with HR+/HER2- subtypes, 18F-FDG PET/CT detected 139/208 lesions
significantly in iTNBC (p, 0.02), possibly reflecting pembrolizumab-driven immune modulation. (missing 2 primary, 40 visceral and bone metastases, and 27 lymph node metastases),
With the caveat of comparing sequential cohorts, the iTNBC cohort showed a trend for higher while 68Ga-MY6349 PET/CT detected 202/208 lesions (missing 1 visceral and bone
ORR (47% vs. 23%, p= 0.08) and longer median PFS (8.4 vs. 5.5 months, HR = 0.68, 95%CI: 0.37- metastasis and 5 lymph node metastases). Interestingly, among 215 lesions in 16 TNBC
1.24), with significantly longer OS (25.8 vs. 18 months, HR = 0.50, 95%CI: 0.26-0.98, p = 0.043) patients, 18F-FDG PET/CT detected 206 metastatic lesions (missing 9 lymph node me-
compared to cTNBC. Conclusions: Timing of pembro administration may influence PD-L1 tastases), whereas 68Ga-MY6349 PET/CT detected all lesions. For HER2+ (7 patients with
expression and clinical outcomes in mTNBC. We show that the immunotherapy run-in strategy 74 lesions) and HR+/HER2+ (4 patients with 65 lesions) subtypes, the two tracers exhibited
converts more PD-L1-negative/low into PD-L1-positive tumors, possibly rendering more patients comparable lesion detectability. Conclusions: 68Ga-MY6349 PET/CT demonstrated su-
eligible for chemoimmunotherapy and improving outcomes. Clinical trial information: perior uptake and greater lesion detectability compared to 18F-FDG PET/CT in patients with
NCT02752685. Research Sponsor: Merck & Co., Inc.
HR+/HER2- breast cancer. The high uptake of 68Ga-MY6349 in HR+/HER2- and TNBC
cTNBC (n=30) iTNBC (n=20) lesions may partially explain the favorable clinical outcomes observed with Trop2-targeted
PD-L1 CPS >/=10 5/23 (22%) 2/16 (12%)
ADCs in these subtypes, suggesting its potential role for patient selection and stratification
PD-L1 CPS conversion (CPS<10 to CPS>/=10) 2/14 (14%) 4/13 (31%) for Trop2-targeted therapies. However, the observed heterogeneity in uptake warrants
Confirmed ORR (CR+PR) 7/30 (23%) 9/19 (47%) further investigation to clarify its applications across different patient populations. Clinical
mPFS (months) 5.5 8.4 trial information: NCT06188468. Research Sponsor: None.
mOS (months) 18.0 25.8

1114 Poster Session 1115 Poster Session

Comprehensive characterization of interleukin-enhanced factor 2 (ILF2) in Concurrent GLP1R-agonist use with chemoimmunotherapy for early-stage
triple-negative breast cancer (TNBC). First Author: Matias Alberto Bustos, Saint triple-negative breast cancer. First Author: Bethania Santos, UT Southwestern
John’s Cancer Institute at Providence Saint John’s Health Center, Santa Monica, CA Medical Center, Dallas, TX
Background: While treatment and management of TNBC has improved, there is a need for Background: Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have emerged as a
novel prognostic biomarkers to better inform outcomes and guide therapeutic options. ILF2 key class of drugs for treating type 2 diabetes mellitus (DM2) and obesity. GLP-1 is rapidly
is a poorly characterized protein with pleiotropic functions that is highly expressed in TNBC. degraded by DPP4, which led to the development of DPP4 inhibitors (DPP4i). Prior work
Here we evaluated the associations of ILF2 with 1) genomic and transcriptomic data, 2) has shown GLP-1R in tumor cells activates key growth signaling and GLP-1RA likely
tumor microenvironment (TME), and 3) clinical outcomes in TNBC. Methods: 15,544 breast dampen inflammation. This suggests that GLP-1R activation may influence response
cancer (BC) samples, including 3,038 TNBC, were tested by NGS (592, NextSeq; WES, rates to chemoimmunotherapy. This study aims to investigate the impact of GLP-1RAs
NovaSeq) and WTS (NovaSeq; Caris Life Sciences, Phoenix, AZ). ILF2-high (H) and ILF2- and DPP4i (GLP1 drugs) exposure on pathological complete response (pCR) rates for
low(L) TNBC were defined by respective quartiles. Immune cell fractions were estimated by patients with early-stage triple negative breast cancer (TNBC) receiving neoadjuvant
WTS deconvolution (Quantiseq). Real world overall survival (OS) was obtained from in- chemoimmunotherapy. Methods: Patients with early-stage TNBC diagnosed between
surance claims and calculated from tissue collection to last contact using Kaplan-Meier July 1, 2021, and December 31, 2023, who received the KEYNOTE-522 regimen were
estimates. Statistical significance was determined by chi-square, Fisher’s exact, and Mann- identified at three institutions. Patients using GLP-1RAs and DPP4i at breast cancer
Whitney U test with p-values adjustments (q , .05). Results: ILF2 expression (median diagnosis and throughout the neoadjuvant period, alone or with other diabetes medi-
Log2(TPM+1) was higher (all q , .05) in key subgroups: ductal compared to lobular cations, were included. Those who started or discontinued GLP-1 drugs during
carcinoma (6.4 vs 6.0); primary compared to metastatic BC (6.4 vs 6.3); African American chemoimmunotherapy were excluded. Group comparisons were made using Chi-square
compared to White (6.4 vs 6.3); basal compared to luminal A, luminal B, HER2 PAM50 and two-sample t-tests. Human TNBCs were analyzed by IHC and CosMx 6000-plex spatial
subtypes (6.9 vs 5.8, 6.3, 6.3); and TNBC compared to HR+HER2+, HR-HER2+, HR+HER2- transcriptomics. Results: Among 343 patients, 7.5% were using GLP-1 drugs. The pCR
subtypes (6.7 vs 6.3, 6.4, 6.2). Biopsied tissues from primary TNBC (pTNBC) and metastatic rate among patients exposed to GLP-1 drugs was 30.8% compared to 64.4% in those not
TNBC (mTNBC) patients were stratified into ILF2-H and ILF2-L groups. In both mTNBC and
exposed (p = 0.001). For patients using other classes of DM2 medications (n = 46), the
pTNBC, ILF2-H groups had 1) higher percentage of young patients (age , 50) (pTNBC: 35.5%
pCR rate was 65.2%, while for those not taking any DM2 medications (n = 271), the pCR
vs 19.8%; mTNBC: 28.1% vs 17.3%; all q , .05); 2) higher mutation frequency of TP53
rate was 64.2%. In univariate analysis, patients exposed to GLP-1 drugs were significantly
(pTNBC: 94.5% vs 79.6%; mTNBC: 92.4% vs 74.4%), but lower frequencies for PIK3CA
older than non-exposed (median age: 60 vs. 51 years; p = 0.009), had a higher BMI (35.0
(pTNBC: 5.1% vs 23.4%, mTNBC: 8.8% vs 27.4%), CDH1 (pTNBC: 0.8% vs 6.1%; mTNBC: 2.8%
vs. 28.9 kg/m²; p = 0.002), and had higher rates of DM2, hypertension, and hyperlipidemia.
vs 12.2%; all q , .05); 3) higher infiltration of NK cells (pTNBC: 3% vs 2.6%; mTNBC: 2.8% vs
2.6%), but lower infiltration of M2 Mw (pTNBC: 2.5% vs 3.3%; mTNBC: 2.6% vs 3.2%) and In multivariate analysis, only age was associated with pCR (OR: 0.97, 95% CI: 0.96-0.99, p
Tregs (pTNBC: 1.5% vs 1.9%; mTNBC: 1.4% vs 1.7%; all q , .05); 4) higher expression levels = 0.007). When comparing patients taking GLP-1 drugs with those using other DM2
of immune checkpoint (CD274, PDCD1LG2, CTLA4, LAG3, HAVCR2, FOXP3, IDO1, CD276, FC: medications, no significant differences were observed regarding age, BMI, or clinical T or
1.2-3.1; all q , .05), stem cell genes (CD44, NANOG, POU5F1, KLF4, ALDH1A1, FC: 1.4-2.4; N stage. To evaluate tumor-intrinsic factors that may influence treatment response, we
all q , .05), and drug efflux genes (ABCC3, ABCC11, ABCC2, ABCB1, ABCG2, ABCC1, FC: 1.1- examined TNBC specimens (n = 84) and identified GLP-1 receptor expression in tumor
4.5; all q , .05) compared to ILF2-L group. In pTNBC, ILF2-H had significantly shorter OS vs cells in 35.7% of cases and in the tumor microenvironment in 60.7% of cases. A spatial
ILF2-L group (22.3 vs 28.9 months, HR 1.2 [95% CI 1-1.5], p = .03), but no significant transcriptomics atlas of GLP-1 drug-exposed tumors (469,029 cells) provides evidence of
differences were observed between mTNBC ILF2 groups (HR 1.1 [95% CI 0.93-1.3], p = .2). GLP-1 pathway activity in both malignant and non-malignant cells of the tumor mi-
Conclusions: ILF2-H TNBC patients showed differential genomic and transcriptomic al- croenvironment. Conclusions: We observed significantly lower pCR rate among patients
terations that relate to therapy resistance, immune suppressive TME, and shorter OS. taking GLP1 drugs during neoadjuvant chemotherapy for TNBC. These effects were not
Further studies are warranted to validate the effects of ILF2 upregulation on therapeutic observed with other diabetic medications. Detection of GLP1R expression in TNBC
efficacy. Research Sponsor: The Fashion Footwear Association of New York (FFANY) specimens indicates there may be direct and indirect effects of agonists to the GLP1
Foundation; None. pathway on chemoimmunotherapy response rates. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 65s

1116 Poster Session 1117 Poster Session

HAI-score, an objective HER2 artificial intelligence method for accurate H- Impact of HER2-ultralow heterogeneity and optimal threshold on trastuzu-
score estimation from IHC-stained breast cancer samples. First Author: Sahar mab deruxtecan (T-DXd) efficacy in metastatic breast cancer: A national
Almahfouz Nasser, Emory University, Atlanta, GA multicenter cohort study (HEROIC). First Author: Yutian Zou, Sun Yat-sen Uni-
Background: Accurate HER2 assessment is essential for breast cancer (BC) treatment, as it directs targeted versity Cancer Center, Guangzhou, China
therapy decisions and predicts patient prognosis. While immunohistochemistry (IHC) is widely used, its manual Background: Trastuzumab deruxtecan (T-DXd) has been approved for patients (pts) with
scoring is susceptible to inter-observer variability. RNAscope, an RNA in situ hybridization (ISH)-based
technique, has shown to have a strong correlation with HER2 protein levels and has outperformed AQUA, a HER2-ultralow metastatic breast cancer (MBC). HER2 discordance commonly occurs
high-throughput quantitative immunofluorescence imaging system, in detecting HER2-low cases. However, between primary and metastatic lesions within the same patient; however, its incidence
RNAscope is constrained by its higher cost and technical complexity compared to IHC staining assays. To remains unknown in the HER2-ultralow era. Additionally, there is still controversy about
address this, we propose the HAI-Score, an objective, robust, accessible, non-tissue disruptive, and fast which specimen to use to determine HER2-ultralow status and optimal threshold to guide
Artificial Intelligence method for evaluating the H-score from IHC images, validated using RNAscope values. T-DXd therapy. Methods: This national, multicenter cohort study included MBC pts treated
Methods: The dataset comprises 526 tissue microarray (TMA) cores for RNAscope and IHC evaluations. The
dataset includes 100 commercially available BC cores (from [Link]) and 426 cores from 243
with T-DXd (5.4 mg/kg) with HER2 status available for both primary tumors and matched
patients at MD Anderson Cancer Center. We digitized TMA cores stained with HercepTest (Dako) (S1 dataset, metastases between January 2020 and October 2024 (NCT06551220). HER2 status was
n=566) and Ventana Pathway 4B5 (Roche) (S2 dataset, n=580) assays. Half of the images were randomly determined according to the DB-06 protocol. Pts were divided into three cohorts based on
allocated for training and the remaining half were used for validation. A computer vision algorithm detects cell HER2 discordance patterns: cohort 1 (HER2-positive/low/ultralow in both primary and
membranes using a specially designed image filter based on domain knowledge. Different visual features were metastases), cohort 2 (HER2-positive/low/ultralow in primary and HER2-null in metas-
then extracted from these detected cell membranes, including perimeter, normalized area, Feret diameter,
tases), and cohort 3 (HER2-null in primary and HER2-positive/low/ultralow in metastases).
fractal dimension, porosity, and staining intensities. These features were used to train a neural network to
predict the HAI-Score. The ground truth was defined as HER2 RNA levels measured by RNAscope. We Endpoints included progression-free survival (PFS), overall survival (OS), objective re-
evaluated the HAI-Score accuracy by correlating it (Pearson, R²) with RNA values and compared it to cor- sponse rate (ORR), disease control rate, and clinical benefit rate. Results: From 24 centers
relations from AHSQ (a state-of-the-art deep learning model), an expert pathologist, and FDA-approved HER2 nationwide, 3546 pts met the criteria and were included. The incidence of HER2 dis-
IHC assays (HercepTest, Ventana PATHWAY). Results: HAI-Score yielded a correlation of 0.85 and an R2 value cordance between primary and matched metastases has changed across eras of HER2-
of 0.711 on the testing dataset, which includes images from both the S1 and S2. This performance surpassed positivity definitions: HER2-positive era (9.8%, K = 0.78), HER2-low era (25.0%, K = 0.39),
AHSQ, the H-score by a breast pathologist, and the scores of two FDA assays with RNA values (Table 1).
Conclusions: HAI-Score provides an objective alternative to evaluate HER2 expression. It is strongly cor-
and HER2-ultralow era (20.2%, K = 0.16). Among T-DXd-treated pts (n = 1052), a higher
related with HER2 RNA levels and was superior to evaluations by an experienced breast pathologist. Following response rate was observed in cohort 1 (ORR = 55.7%) and cohort 3 (ORR = 53.1%)
additional independent multi-site validation, HAI-Score could enable treatment personalization, optimize better compared to cohort 2 (ORR = 13.0%). ORR is positively correlated with HER2 expression if
surgical planning, and reduce overtreatment. Research Sponsor: National Cancer Institute; the National Heart, metastatic lesions are used as the examined tissue (positive 62.9%, low 49.8%, ultralow
Lung and Blood Institute; the National Institute of Allergy and Infectious Diseases; the National Institute of 47.0%, null 13.0%). However, the correlation between ORR and HER2 expression is not
Dental and Craniofacial Research; the National Library of Medicine; the National Instutute on Aging; the VA
Research and Development Office through the Lung Precicision Oncology Program; the Office of the Assistant
significant when primary lesions were examined (positive 57.8%, low 41.5%, ultralow
Secretary of Defense for Health Affairs through the Prostate Cancer Research Program; the Kidney Mapping 54.4%, null 53.1%). Additionally, cohort 1 (mPFS = 11.6 mo, mOS = 30.7 mo) and cohort 3
and Atlas Project (KMAP); sponsored research agreements from Astrazeneca, Bristol Myers Squibb, the (mPFS = 10.9 mo, mOS = 18.4 mo) exhibited significantly superior PFS and OS compared to
Prevent Cancer Foundation, Innovation in Cancer Informatics, and the Scott Mackenzie Foundatio; the National cohort 2 (mPFS = 6.1 mo, mOS = 12.3 mo). Faint incomplete membrane staining per-
Institute of Diabetes and Digestive and Kidney Diseases; the Kidney Mapping and Atlas Project (KMAP); the VA centage $5% in metastatic lesion was the best threshold to distinguish PFS (HR = 0.54, P=
Biomedical Laboratory Research and Development Service. 0.02; mPFS, 11.4 vs 8.6 mo) and ORR (OR = 4.00, P= 0.01; 60% vs 27%) among HER2-
Comparison of methods with RNAscope using Pearson correlation and R². ultralow MBC treated with T-DXd. Conclusions: A high HER2-ultralow discordance rate
Method Pearson Correlation R2 was observed between primary tumors and matched metastases. HER2 status in met-
Roche 0.58 0.33 astatic specimens more accurately predicts T-DXd efficacy compared to primary spec-
Dako 0.76 0.57 imens. A staining threshold of $5% tumor cells in metastatic lesions may optimize T-DXd
Pathologist 0.76 0.58
AHSQ 0.83 0.69 treatment in HER2-ultralow MBC. Therefore, re-evaluating HER2 status in metastatic
HAI-Score 0.85 0.71 lesions is recommended for T-DXd treatment decision. Research Sponsor: None.

1118 Poster Session 1119 Poster Session

Opioid use disorder among females with breast cancer: A comprehensive Treatment patterns, genomic characteristics, and outcomes among patients
analysis of prevalence in the United States and associated factors. First with metastatic lobular breast cancer. First Author: Sherry Shen, Memorial Sloan
Author: Aneri Sanepara, Pandit Deendayal Upadhyay Medical College, Rajkot, Gujarat, Kettering Cancer Center, New York, NY
India Background: Invasive lobular carcinoma (ILC) is characterized by loss of E-cadherin
Background: Patients with breast cancer (BC) are frequently prescribed opioids for pain expression and accounts for 10-15% of breast cancer diagnoses. ILC differs from the
management, placing them at risk of opioid use disorder (OUD). This study analyzes the more common invasive carcinoma of no special type in the pattern of metastatic spread
prevalence of OUD and identifies factors contributing to its risk among BC patients in the and genomic characteristics; however, clinicopathologic characteristics among patients
United States. Methods: We conducted a retrospective analysis using the National In- with metastatic ILC are not well described. Here, we present comprehensive treatment,
patient Sample (NIS) database, a Healthcare Cost and Utilization Project (HCUP) com- genomic, and outcome data in a large single-center cohort of patients with metastatic ILC.
ponent. Females with BC were identified through ICD-10 codes. The Cochran-Armitage Methods: Patients were identified for inclusion in this retrospective study if they had ILC
trend test assessed OUD prevalence trends from 2016 to 2022. Multivariable regression histology on early-stage breast biopsy/surgical pathology and/or CDH1 mutation on
models estimated the impact of multiple patient demographics and comorbidities on the metastatic site biopsy; all patients were required to have MSK-IMPACT somatic next
presence of OUD. Results: Among 1,189,884 females aged$18 with BC, 2.3% (27,500) had generation sequencing (NGS) data available. Clinicopathologic characteristics were
OUD. The mean age of OUD patients was 58.38 years, compared to 64.46 years in the non- abstracted from the EMR. The Kaplan-Meier method was used to estimate overall survival
OUD cohort. OUD prevalence was highest in those aged 18–50 years (3.8%), followed by (OS). The log-rank test was used to compare OS by ILC subtype and by genetic mutations.
51–60 years (3.0%), and lowest in those . 60 years (1.7%). Between 2016 and 2020, OUD Wilcoxon rank sum test and Kruskal-Wallis test were used to compare number of
prevalence increased from 1.9% to 2.8%, followed by a decline to 2.4% in 2022 (p-trend , treatment lines by receptor status. Results: 654 patients were included, of whom 99.8%
0.01). Factors that were linked with higher OUD involved patients with neoplasm-related were female, 89% were white, and 96% non-Hispanic. 438 (67%) had recurrent disease
pain(NRP)(aOR 5.718, 95% CI 5.549-5.893, p , 0.01), on palliative care (aOR 1.397, 95% CI whereas 212 (33%) had de novo metastatic disease. Among 307 with ILC histologic
1.353-1.443, p , 0.001), with metastasis (aOR 1.573, 95% CI 1.526-1.621, p , 0.01), subtype data available, 139 (45%) had classic type, 65 (21%) had pleomorphic, 45 (15%)
depression (aOR 1.447, 95% CI 1.400-1.495, p , 0.01), bipolar disorder (aOR 2.173, 95% CI had mixed, and 58 (19%) had other ILC subtypes. 454 (87%) had hormone receptor-
2.038-2.317, p , 0.01), suicidality (aOR 3.228, 95% CI 2.938-3.546, p , 0.01), and anxiety positive (HR+) disease, 45 (9.1%) had HER2+ disease, and 50 (9.5%) had triple negative
(aOR 1.617, 95% CI 1.572-1.664, p , 0.01). Moreover, substance use such as cocaine (aOR disease at metastatic diagnosis. In the total cohort, median number of treatment lines for
5.252, 95% CI 4.708-5.859, p , 0.01) and amphetamine (aOR 3.948, 95% CI 3.443-4.527, p metastatic disease was 4 (IQR 2-7) and median number of chemotherapies was 2 (IQR 1-
, 0.01) was also associated with higher odds, while cannabis users (aOR 0.876, 95% CI 3). In the HR+ cohort, median number of endocrine therapies was 2 (IQR 1-3). Among
0.793-0.968, p , 0.01) had lower odds of OUD. Our study further found racial disparities, patients with genomic data from a biopsy obtained within 2 months of metastatic di-
with reduced odds among Blacks ( vs Whites, aOR 0.933, 95% CI 0.901-0.967, p , 0.01) and agnosis, 79% had a CDH1 mutation, 48% had a PIK3CA mutation, 5.6% had an AKT1
Hispanics ([Link], aOR 0.866, 95% CI 0.827-0.906, p , 0.01). Socio-economic dif- mutation, 11% had a PTEN mutation, 9.3% had an ESR1 mutation, 17% had a HER2
ferences were also noted, with lower odds among those of the 26th-50th (vs. 0-25th, aOR mutation, and median tumor mutation burden (TMB) was 4 (IQR 3-7); 17% had TMB ³10.
0.932, 95% CI 0.9-0.966, p , 0.01), 51st-75th (vs. 0-25th, aOR 0.951, 95% CI 0.918-0.986, p Median OS in the total cohort was 4.4 years (95%CI 4.1-4.8). OS did not differ significantly
, 0.01), and 76th-100th (vs. 0-25th, aOR 0.916, 95% CI 0.882-0.951, p , 0.01) household by ILC subtype (p= 0.8). OS differed significantly by CDH1 mutation status (wt 5.3 years,
income quartiles. Conclusions: This study showcases the significant prevalence and 95%CI 4.1-6.6; mut 3.7 years, 95%CI 3.5-4.2, p= 0.01), PIK3CA status (wt 4.6 years, 95%CI
impact of OUD among BC patients, identifying socioeconomic and racial disparities, and 4.0-5.8; mut 3.4 years, 95%CI 3.1-3.9, p, 0.001), and PTEN status (wt 4.2 years, 95%CI
key risk factors such as NRP, psychiatric comorbidities, and concurrent substance use, like 3.7-4.5; mut 3.4 years, 95%CI 2.2-4.3, p= 0.008). OS did not differ significantly by HER2,
cocaine and amphetamines. Interestingly, cannabis use was associated with a lower risk of AKT1, or ESR1 mutation status. Conclusions: In a large single-center cohort of patients
OUD, which may reflect its role as an alternative pain management strategy. Overall, this with metastatic ILC, OS did not vary by ILC subtype, but did differ significantly by CDH1,
study suggests the need to adopt crucial preventative measures against OUD in patients PIK3CA, and PTEN mutation status. This underscores the prognostic importance of NGS
exhibiting these characteristics. Research Sponsor: None. in metastatic ILC. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
66s BREAST CANCER—METASTATIC

TPS1120 Poster Session TPS1121 Poster Session

A phase III randomized trial of radiotherapy optimization for low-risk HER2- IND.241: A Canadian Cancer Trials Group liquid-biopsy informed platform
positive breast cancer (HERO): NRG-BR008. First Author: Lior Zvi Braunstein, trial to evaluate treatment in CDK4/6-inhibitor resistant ER+/HER2- meta-
Memorial Sloan Kettering Cancer Center, New York, NY static breast. First Author: David W. Cescon, Princess Margaret Cancer Centre/UHN,
Background: Breast radiotherapy (RT) is the standard of care for patients with early-stage Toronto, ON, Canada
breast cancer (BC) who undergo breast-conserving surgery (BCS). However, the magnitude Background: The combination of a CDK4/6 inhibitor + endocrine therapy (CDK4/6i+ET)
of benefit of RT is less clear in BCS patients with low-risk disease who receive effective is standard first-line systemic treatment for patients with ER+/HER2-negative meta-
systemic therapy. Among patients with early-stage HER2-positive (HER2+) BC, 10-year static breast cancer (MBC). Beyond this initial therapy, there are numerous therapeutic
locoregional recurrence has been reported as low as 1.5% following BCS, adjuvant agents available/ in development for subsequent lines of treatment. Circulating tumor
chemotherapy and HER2-targeted therapy, and RT. Given these exceedingly favorable DNA (ctDNA) via liquid biopsy is a promising, non-invasive approach for blood-based
outcomes, with the addition of HER2-directed therapy, we seek to evaluate the feasibility of tumor genotyping, patient stratification and response assessment with the potential to
omitting RT among patients with early-stage HER2+ BC following BCS and appropriate enhance biomarker-driven strategies and aid in development of new therapeutics.
systemic therapy. Methods: This is a phase III randomized trial for patients $18 years Methods: IND.241 is a master protocol platform design consisting of independent
with early-stage, node-negative, HER2+ (IHC/FISH) BC treated with BCS with negative substudies monitoring patients with ER+/HER2- MBC prior to progression (PD) on CDK4/
margins and sentinel lymph node biopsy or axillary dissection. Patients undergoing 6i+ET and investigating novel agents or drug combinations in 2nd/3rd lines after pro-
primary surgery must have pathologic T1-2 (#3 cm) N0 disease, whereas patients re- gression on CDK4/6i+ET. The primary objective of the novel drug/combination sub-
ceiving neoadjuvant therapy must have clinical T1-2 (with radiographically T#5 cm) N0 studies is to centrally interrogate ctDNA (Tempus xF+, a 523-gene liquid biopsy panel)
disease and exhibit a pathologic complete response (ypT0N0) at surgery (residual DCIS and evaluate whether biomarker selection improves ORR or CBR as assessed by RECIST
[ypTis] spanning #1 cm is permitted, and surgical margins are negative for DCIS). All 1.1. Secondary objectives include safety and toxicity profile for each drug/combination,
patients must receive cytotoxic chemotherapy and HER2-targeted therapy, either in the PFS, and OS. The monitoring substudy (Substudy A) enrolls patients currently on CDK4/
adjuvant or neoadjuvant setting. Stratification is by age (,60; $60), tumor size (#1 cm; 6i+ET treatment and aims to characterize the molecular profile, clinical features, and
.1 cm), estrogen-receptor status (positive; negative), and systemic therapy sequencing ctDNA dynamics of acquired resistance. This platform trial enables creation and
(adjuvant v neoadjuvant). Patients will be randomized to standard breast RT in addition to
maintenance of a tissue and data bank including clinical data, genomics, and radiomics
continuation of trastuzumab to complete one year of treatment (Arm 1), or trastuzumab
from all substudies to evaluate surrogates of treatment outcomes and potential bio-
alone (Arm 2). Endocrine therapy will be recommended for patients with hormone-re-
markers of response, resistance, and disease progression. Patients with specific bio-
ceptor-positive tumors. The primary endpoint is the recurrence-free interval (RFI). Sec-
markers detected in ctDNA will be enrolled into corresponding biomarker positive
ondary endpoints include time to ipsilateral breast recurrence, locoregional recurrence,
disease-free survival, and overall survival, in addition to the 7-year ipsilateral breast re-
cohorts of substudies. Patients with no substudy-specific biomarkers are randomized to
currence rate among those not receiving RT. A health-related quality of life sub-study will biomarker negative cohorts of available substudies. Treatment substudies follow a 2-
assess differences in patient-reported breast pain and worry. We estimate a 7-year RFI of stage design. Currently, the monitoring substudy A is actively accruing. Substudy B is
97.5% with RT and allow for a clinically acceptable decrement of 3.63% without RT (7-year evaluating lunresertib (PKMYT1 inhibitor) + gemcitabine in patients +/-CCNE1 over-
RFI of 93.87%; HR 2.5) to establish omission of RT as non-inferior. NRG-BR008 aims to expression / amplification. Substudy C is evaluating niraparib (PARP inhibitor) + ful-
enroll 1,300 patients over 7.25 years, yielding 80% power to detect the non-inferiority of RT vestrant (ET) in patients +/- alterations in BRCA1/2 (germline/somatic) or PALB2
omission with a one-sided a=0.05. We expect to observe the required 38 RFI events within (germline). These latter two substudies have now closed to accrual, with efficacy and
4.5 years of additional follow-up. The NRG-BR008/HERO trial opened to accrual in March safety evaluation ongoing. Substudy D, which has recently been added, is evaluating
2023. Accrual is 64/1,300 as of 1/23/24. NCT #: NCT05705401. Support: U10 CA180868, lunresertib + camonsertib (ATR inhibitor) in patients +/- CCNE1 overexpression/am-
-180822, UG1 CA189867, U24 CA196067. Clinical trial information: NCT05705401. plification, FBXW7 or PPP2R1A alterations. Additional substudies are in development for
Research Sponsor: National Cancer Institute; U10CA180868; National Cancer Institute; inclusion in this platform trial. Clinical trial information: NCT05601440. Research
U10CA180822; National Cancer Institute; UG1CA189867; National Cancer Institute; Sponsor: Repare Therapeutics; GSK.
U24CA196067.

TPS1122 Poster Session TPS1123 Poster Session

Phase II study evaluating 68Ga-FAPI PET uptake heterogeneity as a pre- SOLTI-2201 ACROSS-TROP2 trial: A phase II study to identify predictive
dictor of T-DXd treatment response in HER2-positive breast cancer brain biomarkers of sacituzumab govitecan benefit and to understand resistance
metastases. First Author: Biyun Wang, Department of Medical Oncology, Fudan mechanisms in HR+/HER2- advanced or metastatic breast cancer. First
University Shanghai Cancer Center, Shanghai, China Author: Eva Maria Ciruelos, University Hospital 12 de Octubre/ SOLTI Cancer Research
Background: Breast cancer is a leading cause of metastasis to the central nervous Group, Madrid, Spain
system (CNS). Approximately 30% of patients with HER2-positive breast cancer develop Background: Sacituzumab Govitecan (SG) is a TROP2-directed antibody-drug conjugate
brain metastases, which are associated with a poor prognosis and limited treatment (ADC) linked to a topoisomerase I inhibitor via a hydrolysable CL2A linker. It is approved
options. T-DXd has shown promise in treating brain metastases from HER2-positive for the treatment of metastatic triple-negative breast cancer (mTNBC) patients who have
breast cancer. However, up to 10% patients showed brain metastasis progression at the undergone at least two prior systemic therapies, including one for advanced disease, and
initial evaluation of treatment and may require radiotherapy or neurosurgery immedi- of hormone receptor-positive (HR+)/HER2-negative metastatic breast cancer (mBC)
ately. Identifying predictors of treatment response and determining the timing for local patients after endocrine therapy (ET) and two systemic treatments. Currently, no bio-
therapy intervention is crucial for personalized medicine. Heterogeneity in tumor markers, including TROP2 protein expression, have been identified to predict SG re-
metabolism, as assessed by (68Ga)–labeled fibroblast-activation protein inhibitor sponse, highlighting the need to explore biomarkers of efficacy and to identify key
(68Ga-FAPI) PET-CT which displays the activity of cancer-associated fibroblasts (CAFs) resistance mechanisms to the drug. The ACROSS-TROP2 study aims to address this
in the tumor microenvironment, with good sensitivity and specificity in brain metastasis unmet medical need. Methods: ACROSS-TROP2 (NCT06236269) is a phase II, open-
imaging, may serve as a biomarker for treatment response. This study aims to in- label, single-arm trial investigating SG in HR+/HER2-negative mBC patients. The study
vestigate the predictive value of 68Ga-FAPI PET uptake heterogeneity for T-DXd initially planned to enroll 50 pre- or post-menopausal female or male participants who
treatment response in HER2-positive breast cancer brain metastases. Methods: This progressed during or after treatment with CDK4/6 inhibitors and received up to one prior
open-label, single-center, phase II clinical trial will investigate the heterogeneity of brain chemotherapy or ADC regimen for metastatic disease. Due to high recruitment rates and
metastasis and analyze the difference between stable and active brain metastasis promising findings demonstrating ADC benefits in earlier treatment lines (Bardia et al.,
evaluated by 68Ga-FAPI uptake in HER2-positive MBC. Patients with HER2-positive NEJM 2024), a protocol amendment was introduced to expand the sample size to 100
metastatic breast cancer and confirmed brain metastases by MRI were enrolled; at least patients. Participants will receive SG at 10 mg/kg via IV infusion on Days 1 and 8 of each
one measurable intracranial lesion ($ 1.0 cm) that has not previously been treated with 21-day cycle until disease progression (PD). Fresh tumor biopsies will be obtained at
radiation. Radiotherapy or neurosurgery is allowed with an interval $ 4 weeks. Patients baseline, after 2–3 weeks of treatment (C2D1), and at PD. The primary endpoint is to
will receive T-DXd treatment and undergo 68Ga-FAPI PET-CT scans before and after two measure changes in the CelTIL score—a composite of tumor cellularity and tumor-in-
cycles of treatment. The primary endpoint is the difference in baseline heterogeneity filtrating lymphocytes—between baseline and C2D1 biopsies, as CelTIL is associated with
index by 68Ga-FAPI PET-CT between cerebral lesions achieving ORR and those that long-term efficacy. Secondary endpoints include overall response rate, progression-free
do not. Secondary endpoints include 68Ga-FAPI PET-CT value changes (SUVmax, survival, duration of response, time to response, safety, and tolerability. Correlative
SUVmean) at baseline and after treatment; difference in baseline heterogeneity index analyses of molecular markers in tissue and blood will be conducted to correlate bi-
for PFS, CBR and OS; difference of baseline heterogeneity index, SUVmax and ological findings (e.g., CelTIL, Ki67, TROP2, PD-1/PD-L1, PAM50) with clinicopathological
SUVmean between active or stable brain metastasis; 68Ga-FAPI PET-CT value changes data, evaluate the predictive value of early dynamic changes in ctDNA, identify genomic
(heterogeneity index, SUVmax, SUVmean ) at baseline and 2 cycles after T-DXd alterations linked to treatment response and resistance, and explore changes from
treatment of whole body metastasis lesions. The study plans to enroll 50 patients and is baseline to PD to identify mechanisms of resistance. A paired t-test will assess whether
actively enrolling. Clinical trial information: NCT06797622. Research Sponsor: CSCO- the mean change in CelTIL score is statistically different from zero. The study has been
LingHang Oncology Research Foundation (Y-2022HER2AZQN-0378). approved in Spain and is actively enrolling participants at 10 sites within the SOLTI
network. Previously presented at ESMO Breast 2024, FPN: 265TiP, Eva Ciruelos et al. -
Reused with permission. Clinical trial information: NCT06236269. Research Sponsor:
None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 67s

TPS1124 Poster Session TPS1125 Poster Session

LITESPARK-029: A phase 2, randomized, open-label study of belzutifan plus ALISertib in combination with endocrine therapy in patients with hormone
fulvestrant in participants with estrogen receptor–positive, HER2-negative receptor-positive (HR+), HER2-negative (HER2–) recurrent or metastatic
unresectable locally advanced or metastatic breast cancer after progression breast cancer: The phase 2 ALISCA-Breast1 study. First Author: Pooja Prem
on previous endocrine therapy. First Author: Bora Lim, Department of Breast Advani, Mayo Clinic, Jacksonville, FL
Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX Background: Despite the many available treatments for patients (pts) with HR+, HER2–
Background: Endocrine-based therapy (ET), with or without cyclin-dependent kinase 4/6 recurrent/metastatic breast cancer (MBC), optimal treatment after progression on
inhibitors (CDK4/6i), prolongs PFS and OS in participants (pts) with metastatic hormone CDK4/6 inhibitors (CDK4/6i) is unclear. One possible CDK4/6i resistance mechanism is
receptor–positive (HR+) and human epidermal growth factor receptor 2–negative (HER2–) increased expression of Aurora kinase A (AURKA), a key mitosis regulator associated
breast cancer. After PD on first-line therapy, next-line therapy options provide limited PFS with poor prognosis. Further implicated in CDK 4/6i resistance, high c-Myc or RB1 loss of
gains, in part due to resistance mechanisms (eg, hyperactive FOXA1). The transcription function (LOF) are associated with transcriptional co-regulation or synthetic lethality,
factor hypoxia-inducible factor 2a (HIF-2a), a major target of FOXA1, regulates key respectively, with AURKA. Alisertib is a highly selective, reversible, ATP-competitive,
components of angiogenesis and subsequent development of metastasis. Preclinical orally administered, small-molecule AURKA inhibitor with antiproliferative activity in
studies show suppression of tumor growth with an HIF-2a antagonist, particularly when HR+ BC-derived cell lines and BC xenograft models. Models with elevated AURKA or c-
combined with fulvestrant. Belzutifan, an HIF-2a inhibitor, is approved for the treatment of Myc expression, or RB1 LOF show greater alisertib sensitivity. Alisertib had activity in
pts with advanced renal cell carcinoma following a PD-(L)1 inhibitor and vascular endothelial phase 1 and 2 trials, including objective response rates (ORRs) of 19.6–20% and median
growth factor tyrosine kinase inhibitor. LITESPARK-029 (NCT06428396) evaluates belzu- progression-free survival (PFS) of 5.4–5.6 months alone or with fulvestrant in pts with
tifan + fulvestrant vs everolimus + fulvestrant or exemestane in pts with estrogen HR+/HER2–, endocrine-resistant MBC. The most common treatment-related grade $3
receptor–positive (ER+)/HER2– unresectable locally advanced or metastatic breast cancer.
adverse events (AEs) were neutropenia, anemia, and leukopenia. Methods: ALISCA-
Methods: This phase 2, randomized, active-controlled, open-label, multicenter study is
Breast1 (NCT06369285) is a randomized phase 2 study. Primary objective: to determine
enrolling pts ($18 y) with locally confirmed ER+/HER2– unresectable, locally advanced or
metastatic disease who have had radiographic PD on $12 mo of ET + CDK4/6i therapy in the
the optimal alisertib dose administered with endocrine therapy (ET) based on AEs and
noncurative setting or received $2 lines of ET in the noncurative setting including CDK4/6i serious AEs per CTCAE v5.0 and efficacy (ORR, duration of response, disease-control
where the CDK4/6i was discontinued due to intolerance (not due to progression). Pts must rate, PFS, overall survival). Secondary objectives: to identify biomarkers of efficacy and
also be eligible for additional ET with everolimus plus either fulvestrant or exemestane per alisertib pharmacokinetics (PK). Key inclusion criteria: $18 years; ECOG performance
local investigator assessment, have an ECOG PS of 0 or 1, and provide a new or recent core status 0 or 1; confirmed HR+, HER2–, recurrent/metastatic breast adenocarcinoma not
biopsy for central determination of ER and HER2 status. Prior treatment with chemotherapy, amenable to curative therapy; available tumor tissue for biomarker analyses; pro-
antibody-drug conjugates, or PARP inhibitors in the noncurative setting is prohibited. Pts are gression on or after $2 prior ET lines in recurrent/metastatic setting; prior CDK4/6i with
randomized 1:1 to receive oral belzutifan 120 mg once daily + fulvestrant 500 mg on days 1 ET in recurrent/metastatic setting. Key exclusion criteria: prior chemotherapy in re-
and 15 of cycle 1 and on day 1 of all subsequent 28-day cycles or oral everolimus 10 mg once current/metastatic setting; prior AURKA-specific or pan-Aurora-targeted agents; un-
daily + fulvestrant (as above) or oral exemestane 25 mg once daily until PD or unacceptable stable brain metastases. Eligible pts will be randomized [Link] to alisertib 30 mg, 40 mg,
toxicity. Randomization is stratified by treatment with prior ET + CDK4/6i therapy ( , 18 mo or 50 mg orally twice daily on days 123, 8–10, and 15–17 every 28 days, plus physician’s
duration before PD vs $18 mo duration before PD or no PD). Tumor imaging is performed at choice of anastrozole, letrozole, exemestane, fulvestrant, or tamoxifen not previously
screening, Q8W from randomization through week 56, and Q12W thereafter. The primary used in recurrent/metastatic setting or progressed upon in adjuvant setting; #50 pts will
endpoint is PFS per RECIST v1.1 by blinded independent central review (BICR). Secondary be enrolled per arm in the USA and Europe. All pts will undergo sparse PK sampling.
endpoints include PFS rate per RECIST v1.1 by BICR at 6 and 12 mo, OS, ORR per RECIST Tumor tissue will be centrally assessed for biomarkers, including RB1, MYC, TP53, ESR1,
v1.1 by BICR, clinical benefit (CR, PR, or stable disease for $24 weeks), and safety. The PI3K/AKT pathway, HER2 and AURKA genomic alterations/expression levels. The study
study start date was July 2024. Clinical trial information: NCT06428396. Research Sponsor: will determine the optimal alisertib dose to combine with ET and may identify bio-
Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. marker(s) defining pts with the greatest benefit from alisertib-based therapy. Clinical
trial information: NCT06369285. Research Sponsor: Puma Biotechnology Inc.

TPS1126 Poster Session TPS1127 Poster Session

Integrating gene signatures to guide HR+/HER2- MBC therapy in a diverse ELCIN: Elacestrant in women and men with CDK4/6 inhibitor (CDK4/6i)-
cohort (INSIGHT). First Author: Sonya A. Reid, Vanderbilt-Ingram Cancer Center, naı̈ve estrogen receptor-positive (ER+), HER2-negative (HER2-) metastatic
Vanderbilt University Medical Center, Nashville, TN breast cancer (mBC)—An open-label multicenter phase 2 study. First Author:
Background: Black women with breast cancer (BC) have a 40% higher mortality rate Virginia G. Kaklamani, University of Texas Health Science Center at San Antonio, San
compared to Non-Hispanic White (NHW) women. Worse outcomes have been observed Antonio, TX
among Black women with hormone receptor positive (HR+), human epidermal growth Background: Endocrine therapy (ET) plus a CDK4/6i is the mainstay treatment in first-
factor receptor 2 negative (HER2-) BC despite comparable systemic therapies. Gene line ER+/HER2- mBC; however, a subset of patients are unable to tolerate CDK4/6i, and
expression profiling has been used in early-stage BC to provide prognostic and predictive resistance to ET emerges. Intrinsic resistance mechanisms include alterations in the
information beyond standard immunohistochemical classifications. BluePrint, an 80- PI3K/AKT/mTOR or cell cycle pathways; acquired resistance mechanisms include
gene molecular subtype signature, and MammaPrint (Agendia), a 70-gene risk of distant estrogen receptor gene 1 mutations (ESR1-mut), which emerge in up to 50% of patients
recurrence signature, further classify HR+/HER2- BC into luminal A, luminal B, HER2- during prolonged aromatase inhibitor therapy in mBC. In the phase 3 EMERALD trial,
enriched, and basal-type tumors. Non-Luminal A (Luminal B, HER2-enriched, and Basal- elacestrant significantly prolonged PFS vs standard-of-care (SOC) ET and was asso-
type) tumors are more aggressive and are associated with worse survival compared to ciated with a manageable safety profile in patients with ER+/HER2- mBC previously
Luminal A tumors. Our preliminary data demonstrate that non-Luminal A tumors are treated with ET+CDK4/6i, leading to its approval as the first clinically available oral
overrepresented in Black women (11% Black vs. 5% White). The role of molecular SERD. Elacestrant significantly reduced the risk of progression or death vs SOC ET by
subtyping in guiding therapy for patients with HR+/HER- MBC is not defined. Retro- 30% in the overall population (HR 0.70; 95% CI 0.55-0.88; P=0.002) and by 45% in
spective studies have shown that non-Luminal A HR+/HER2- tumors derive less benefit patients with ESR1-mut tumors (HR 0.55; 95% CI 0.39-0.77; P=0.0005) [Bidard, 2022].
from endocrine therapy (ET). We hypothesize that patients with non-Luminal A, HR+/ Preclinical studies demonstrated that elacestrant is equally active in both in vitro and in
HER2- MBC progressing on ET +/- CDK4/6 inhibition derive more benefit from che- vivo models of ER+/HER2- breast cancer, regardless of prior exposure to CDK4/6i. Based
motherapy than ET in the second line. Furthermore, the impact of the intervention will be on preclinical models and clinical efficacy data, elacestrant may improve clinical
more pronounced in Black women compared to NHW women. INSIGHT is a randomized outcomes in CDK4/6i-naı̈ve patients and provide a convenient all-oral treatment option if
phase II study evaluating the anti-tumor effect of capecitabine versus physician’s choice combined with CDK4/6i. The ELCIN trial will evaluate efficacy and safety of elacestrant
ET as second line for patients with non-Luminal A HR+/HER2- MBC (NCT05693766). in patients with ER+/HER2- mBC who received prior ET and no prior CDK4/6i in the
Methods: In this study, patients progressing on 1st line ET +/- a CDK4/6i are enrolled. metastatic setting. Methods: ELCIN (NCT05596409) is an open-label, multicenter,
Archival primary or metastatic tumor samples are analyzed using MammaPrint and single-arm phase 2 trial. Eligible patients are women or men with ER+/HER22 mBC who
BluePrint. Patients with non-Luminal A tumors are randomized (1:1) to receive phy- received 1-2 lines of prior ET and no prior CDK4/6i or chemo in the metastatic setting.
sician’s choice ET versus capecitabine, stratified by molecular subtype and race. Disease Patients must have measurable disease per RECIST v1.1 or a mainly lytic bone lesion (for
assessments are performed every three months. The primary endpoint is progression bone disease only), ECOG PS #1, adequate bone marrow and organ function, and no
free survival (PFS). Secondary endpoints include overall response rate, clinical benefit active or newly diagnosed CNS metastases or visceral crisis. Patients will receive
rate, overall survival, and patient reported outcomes. The study has 80% power to detect elacestrant 345 mg once daily. The primary objective is investigator-assessed PFS.
a minimal hazard ratio of 0.5 in 5-year PFS with one-sided a = 0.05. Exploratory Secondary objectives are ORR, DoR, CBR, OS, PROs-QoL, and safety. Exploratory ob-
correlative studies are planned. This trial enriches for racial/ethnic minority patients jectives include elacestrant efficacy according to ESR1-mut status, changes in bio-
through collaborations with the University of Texas Southwestern and the University of markers, including allele mutation frequencies (cfNAs), and relationship between
Alabama at Birmingham, health systems that serve large minority populations. Seven of efficacy endpoints. Status: ELCIN has a planned sample size of 60 patients; recruitment
the 64 planned patients have been enrolled. Clinical trial information: NCT05693766. is ongoing worldwide. Clinical trial information: NCT05596409. Research Sponsor:
Research Sponsor: Susan G. Komen. Menarini Group.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
68s BREAST CANCER—METASTATIC

TPS1128 Poster Session TPS1129 Poster Session

SIMRISE: A randomized phase III trial evaluating SIM0270 in combination ADELA: A double-blind, placebo-controlled, randomized phase 3 trial of
with everolimus versus treatment of physician’s choice in patients with ER+/ elacestrant (ELA) + everolimus (EVE) versus ELA + placebo (PBO) in ER+/
HER2- advanced breast cancer, previously treated with CDK4/6 inhibitors. HER2- advanced breast cancer (aBC) patients with ESR1-mutated tumors
First Author: Jiong Wu, Fudan University Shanghai Cancer Center, Shanghai, China progressing on endocrine therapy (ET) + CDK4/6i. First Author: Antonio
Background: CDK4/6 inhibitors (CDK4/6i) in combination with endocrine therapy (ET) Llombart-Cussac, Medica Scientia Innovation Research (MEDSIR), Hospital Arnau de
have demonstrated sustained benefits in the first-line treatment of HR+/HER2- ad- Vilanova, FISABIO, Translational Oncology Group, Universidad Cardenal Herrera-CEU,
vanced breast cancer(aBC). However, effective ET options are limited for patients who Alfara Del Patriarca, Spain
progressed after the treatment of ET in combination with CDK4/6i. SIM0270 in com- Background: ET+CDK4/6i is standard-of-care (SOC) in 1L ER+/HER2- aBC; however, tumors
bination with everolimus exhibited promising anti-tumor activities in a Phase I study. eventually develop resistance. Constitutive activation in the PI3K/AKT/mTOR pathway can
Methods: SIMRISE is an ongoing randomized, open label, Phase III trial designed to contribute to endocrine resistance in breast cancer. ESR1 mutations are a common type of
evaluate SIM0270 in combination with everolimus versus the treatment of physician’s acquired resistance that emerges in 40-50% of patients in the metastatic setting after
choice (TPC) for patients with ER+/HER2- aBC progressed on previous ET and CDK4/6i. A prolonged aromatase inhibitor exposure. There is an unmet need for novel therapeutic ap-
total of 460 patients will be enrolled across approximately 50 sites in China. Patients are proaches to overcome resistance mechanisms and improve outcomes in patients with ER+/
HER2- aBC with ESR1-mutated tumors progressing after ET+CDK4/6i. ELA is a next-generation
randomized in a 1:1 ratio to receive either SIM0270 + everolimus or TPC (exemestane +
oral SERD that binds to ER-alpha, inducing its degradation. In EMERALD, ELA improved PFS vs
everolimus or fulvestrant). Stratification factors include: visceral metastasis (yes or no);
SOC ET in patients with ESR1-mutated tumors (HR 0.55; 95% CI 0.39-0.77; P=0.0005) [Bidard
prior fulvestrant (yes or no); baseline ESR1 status (mutation detected or not detected). 2022]. Differences were notable among patients who received prior ET+CDK4/6i $12 mo;
Key eligibility criteria include: ER+/HER2- aBC patients having measurable disease per median PFS with ELA was 8.6 mo vs 1.9 mo with SOC ET (HR 0.41; 95% CI 0.26-0.63) [Bardia
RECIST 1.1 or bone only disease with at least one predominant lytic bone lesion or mixed 2024]. Crosstalk between ER and PI3K/AKT/mTOR pathways provides a rationale for eval-
lytic-blastic lesion; postmenopausal women and pre-/perimenopausal women or men uating ELA+EVE (a mTORC1 inhibitor). In ELEVATE phase 1b (NCT05563220), ELA+EVE
receiving luteotropic hormone releasing hormone agonist(LHRHa) therapy per local demonstrated ORR 22% and CBR at 24 weeks 72% in patients with ER+/HER2- aBC progressing
prescribing information; patients must have received at least one line and no more than after ET+CDK4/6i; ELA 345 mg + EVE 7.5 mg was identified as the RP2D [Rugo ESMO 2024].
two lines of ET; recurrence while on or within 12 months of completion of adjuvant ET for Safety was consistent with the known profile of EVE+SOC ET. ADELA compares ELA+EVE vs
$24 months is considered as first-line ET, or first line ET in advanced setting for $6 ELA+PBO in ER+/HER2- aBC patients with ESR1-mutated tumors progressing on ET+CDK4/6i.
months. Patients must have previously received CDK4/6i combined with ET for $6 Methods: ADELA (NCT06382948) is an international, multicenter, double-blind, placebo-
months; one line chemotherapy for aBC is allowed. The Primary endpoint is progression- controlled phase 3 trial. Eligible patients are adults ($18 yrs) with ER+/HER2- aBC and ESR1-
free survival (PFS) as assessed by blinded independent review committee (BIRC). The mutated tumors, previously treated with 1-2 lines of ET for aBC, and evidence of disease
secondary endpoints include PFS (assessed by investigators), overall survival (OS), progression on prior ET+CDK4/6i for aBC after $6 mo. Patients receiving CDK4/6i-based
objective response rate (ORR), duration of response (DoR), clinical benefit rate (CBR), adjuvant therapy are eligible (disease progression must be confirmed after $12 mo of
time to progression (TTP), safety, pharmacokinetics (PK) and patients-reported out- treatment but ,12 mo following CDK4/6i completion). Other criteria include adequate organ
function and ECOG PS 0-1. Exclusion criteria include prior chemotherapy for aBC and active
comes (PRO). The analysis of primary endpoint will use a stratified log-rank test at an
uncontrolled/symptomatic brain metastasis. Patients will be randomized 1:1 to 28-d cycles of
overall of 0.05 significance level (two-sided). Futility analyses are planned, and an
ELA 345 mg + EVE 7.5 mg QD or ELA 345 mg + PBO QD until disease progression or un-
independent data monitoring committee will be in place. Clinical trial information: acceptable toxicity. Patients will receive dexamethasone mouthwash during the first 8 wks.
NCT06680921. Research Sponsor: Simcere Zaiming Pharmaceutical Co., Ltd. Stratification factors are presence of visceral metastases (yes vs no) and duration of prior
CDK4/6i ($12 mo vs ,12 mo). The primary objective will be to evaluate PFS based on blinded
independent review committee. Secondary endpoints include investigator-assessed PFS, OS,
ORR, CBR, DoR, TTR, best percentage change in tumor burden, safety, and HRQoL. Status:
Planned enrollment is 240 patients; recruitment is ongoing. Clinical trial information:
NCT06382948. Research Sponsor: Menarini Group.

TPS1130 Poster Session TPS1131 Poster Session

Dauntless-1, a phase 2 clinical trial to evaluate PMD-026, a first-in-class OPERA-01: A randomized, open-label, phase 3 study of palazestrant (OP-
pan-RSK inhibitor, combined with fulvestrant to overcome resistance to 1250) monotherapy vs standard-of-care for ER+, HER2- advanced or met-
CDK4/6 inhibitors in advanced or metastatic HR+/HER2- breast cancer. First astatic breast cancer patients after endocrine therapy and CDK4/6
Author: Sandra Elaine Dunn, Phoenix Molecular Designs, Vancouver, BC, Canada inhibitors. First Author: Barbara Pistilli, Breast Cancer Unit, Gustave Roussy, Ville-
Background: Resistance to CDK4/6 inhibitors (CDK4/6i) is common for many patients juif, France
with HR+/HER2- advanced or metastatic breast cancer, therefore new strategies are Background: Endocrine therapy(ET) resistance is a major challenge in treating estrogen
urgently needed to overcome this challenge. Ribosomal S6 kinases (RSK1-4) are im- receptor-positive (ER+), human epidermal growth factor receptor 2-negative (HER2–)
plicated in breast cancer growth and resistance, and they are activated by the PI3K and metastatic breast cancer (MBC); estrogen receptor (ESR1) mutations are an important
MAPK pathways, which are linked to CDK4/6i resistance. As a convergence point of these mechanism of resistance. The standard of care (SOC) first-line treatment for ER+, HER2–
pathways, RSK drives resistance by promoting the G2/M phase of the cell cycle and MBC is ET plus a cyclin-dependent kinase 4/6 inhibitor (CDK4/6i). Despite the benefit of
bypassing G1/S control. Inhibiting RSK with PMD-026, a first-in-class oral small molecule ET and CDK4/6i, disease progression and acquired resistance to the combination remain a
inhibitor, halts G2/M progression and blocks growth in CDK4/6i-resistant models, in- challenge. Novel, more effective ETs that can overcome resistance are needed to improve
cluding those cross-resistant to abemaciclib and palbociclib. RSK also complexes with outcomes and delay time to chemotherapy. Palazestrant (OP-1250) is a novel oral,
estrogen receptor alpha (ERa), enhancing transcription and tumor growth. PMD-026 complete estrogen receptor antagonist (CERAN) and selective ER degrader (SERD) that
disrupts this interaction, showing activity in both ESR1 wild-type and mutant HR+/HER2- acts by blocking both transcriptional activation function domains, AF1 and AF2, re-
models, making it a promising partner for endocrine therapies. It synergizes with ful- gardless of ESR1 mutation status. As monotherapy, palazestrant showed a tolerable
vestrant and oral SERDs, achieving significant growth inhibition in preclinical models, safety profile, favorable pharmacokinetics and encouraging antitumor efficacy in heavily-
including a 7000-fold improvement with fulvestrant in soft agar assays. Nuclear pretreated patients during phase 1/2 studies, regardless of ESR1 mutation status
translocation of RSK is a key driver of breast cancer in mice and serves as a biomarker for (NCT04505826; Lin et al. ESMO 2023 MO382). Methods: OPERA-01 (NCT06016738) is a
RSK signalling activity. In the Phase 1/1b monotherapy study, PMD-026 was generally multicenter, randomized, open-label, phase 3 clinical trial comparing the efficacy and
well-tolerated, and it reduced the risk of progression or death in patients by 93% in a safety of palazestrant as a single agent to SOC ET (fulvestrant, anastrozole, letrozole, or
subset of RSK2 high metastatic breast cancer patients. Methods: Dauntless-1 is a Phase exemestane) in patients with ER+, HER2– MBC that relapsed or progressed on 1-2 prior
2a study for locally advanced or metastatic HR+/HER2- breast cancer patients previously lines of ET, including a CDK4/6i. Adult patients are eligible with a diagnosis of evaluable
treated with a CDK4/6i in combination with endocrine therapy (NCT04115306). It is ER+, HER2– inoperable locally advanced or MBC and an Eastern Cooperative Oncology
designed to prospectively enroll RSK2+ ($50% nuclear staining with $2+ staining in- Group performance status of 0 or 1. Prior treatments must include 1 or 2 prior lines of ET
tensity) patients to evaluate PMD-026 in combination fulvestrant. Fulvestrant will be with the last ET duration of $6 months; must have received and have disease progression
dosed per the package insert (500 mg IM, Day 1 and 15 of the first 28-day cycle, then Day 1 on CDK4/6i with ET for MBC. Prior chemotherapy for MBC is not allowed. The study
of every cycle thereafter) in combination with PMD-026 at the RP2D (200 mg, PO, Q12h), included a dose selection phase, where participants were randomized to 90 mg qd or 120
determined in the dose-finding portion of the study. The combination regimen will have a mg qd palazestrant or SOC; enrollment in this phase is complete. After the dose selection
safety lead-in cohort of 6 patients. The SRC will review the safety data after the sixth of palazestrant, the study will continue with the selected dose compared to SOC ET at a 1:
patient has been treated for at least 28 days. If determined to be safe, up to 14 additional 1 randomization. Overall, 510 patients will be randomized to palazestrant or SOC ET
patients will receive the combination for a total of 20 patients. A Bayesian safety during the study. The primary endpoint of progression-free survival will be assessed by
monitoring rule will be used to evaluate the rate of DLTs during expansion. Primary blinded independent central review in patients with and without ESR1 mutations (dual
objectives will be safety, pharmacokinetics, and progression free survival. Secondary primary endpoint). Secondary endpoints include overall survival, antitumor activity
objectives include duration of response, overall response and overall survival. Exploratory (objective response rate, clinical benefit rate, and duration of response), safety, exposure
objectives will evaluate PMD-026 in the context of mutations (ESR1, PIK3CA, AKT1, p53, and patient-reported outcomes in patients with and without ESR1 mutations. Study
KRAS) at baseline using ctDNA. Clinical trial information: NCT04115306. Research recruitment began in November 2023. Clinical trial information: NCT06016738. Research
Sponsor: None. Sponsor: Olema Oncology.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 69s

TPS1132 Poster Session TPS1133 Poster Session

Immunologic targeting of native and mutated ESR1 receptor for treatment of Adaptive designed eniluracil + capecitabine phase 2 trial in advanced or
hormone receptor expressing metastatic breast cancer. First Author: Aixa Elena metastatic breast cancer patients. First Author: David Young, Processa Phar-
Soyano Muller, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL maceuticals, Inc., Hanover, MD
Background: Hormone receptor positive (HR+) HER2 negative metastatic breast cancer Background: Ethynyl-uracil (eniluracil or 6422), an irreversible inhibitor of the dihy-
(MBC) remains a difficult clinical problem. Endocrine therapies have remained the dropyrimidine dehydrogenase enzyme that metabolizes 5-FU to catabolites, eliminates the
mainstay of therapy for decades and despite combination with targeted agents and formation of 5-FU catabolites and catabolite side effects while exposing cancer cells to
development of novel targeted therapies, the 5-year survival rate of MBC remains low. more 5-FU and more cancer killing 5-FU anabolites. The combination of a single 40 mg day
Resistance to ET can occur due to the development of point mutations in the estrogen 1 dose of 6422 followed by Day 2 Capecitabine (6422+Cap) on a 7 day on + 7 day off
receptor alpha type I (ESR1), which constitutively activates the receptor, making it schedule (7+7) of Capecitabine (Cap) is being evaluated. The dose of Cap used in 6422+Cap
resistant to anti-estrogen. We produced a peptide library of the entire wild type (WT) will be approximately 15% of the therapeutic dose of Cap used in clinical practice for breast
ESR1 and identified four promiscuous peptide epitopes that routinely drive a CD4 Th1 cancer. A Phase 1B study in patients with refractory gastrointestinal (GI) cancer has been
response in healthy donors and breast cancer patients. Additionally, we created completed. The Maximum Tolerated Dose of Cap in 6422+Cap was determined to be 225
overlapping peptides around each known ESR1 mutation site and pulsed them on type I mg BID. The Recommended Phase 2 Dose Range of Cap in 6422+Cap was determined to be
polarized dendritic cells (DC1) also resulting in an increased CD4 Th1 response. We from 75 mg BID to 225 mg BID. based on the FDA’s Optimal Design Guidance and Project
hypothesize that ER alpha receptor can serve as a target for the immune response and Optimus initiative to define the dose-response relationship for both safety and efficacy.
ESR1 mutations that develop in HR+ breast cancer patients can be utilized as a Since FDA believed that determining the optimal dosage regimen following the Principles
neoantigen that will drive CD4 Th1 responses and antibodies that can be developed as an of Project Optimus would be extremely difficult for 6422+Cap in GI cancer given the
immune based therapy for patients with HR+ MBC. Combining DC1 vaccination with standard combination chemotherapeutic treatments, the target population was changed to
novel endocrine therapies such as Elacestrant, we expect an increase in ESR1 deg- breast cancer patients. FDA also determined that a Phase 1B study in breast cancer would
radation and enhanced antigen presentation leading to an expanded immune and clinical not be required given the GI cancer Phase 1B data. Methods: Several Project Optimus
response. Methods: In this open pilot study, up to 18 patients with HR+ HER2 negative, focused Phase 2 designs were evaluated. Based on guidance from the FDA, an adaptive, 3
arm, 30 patients/arm, phase 2, open-labelled, randomized trial was selected which would
ESR1 mutated MBC with measurable or evaluable disease will be enrolled to determine
compare 2 regimens of 6422+Cap vs. standard dose of Cap alone. The study would initially
the feasibility and safety of the combination of DC1 vaccines and Elacestrant. Prior use
enroll patients into 2 treatment arms. The 2 arms are: a 1000 mg/m2 BID monotherapy Cap
of elacestrant is exclusionary. Eligible patients will undergo apheresis of peripheral
control group and a 6422+Cap regimen of 40 mg on day 1 followed by day 2 Cap dose of
blood to collect and create DC1 vaccines. DC1 will be pulsed with ESR1 WT and mutated
150 mg BID on 7+7 schedule. Upon completing the enrollment and evaluation of 9-10
peptides. Patients will be injected in their groin nodes (or accessible tumor if available) patients in each of the first 2 arms, an interim analysis will be conducted to determine the
weekly with these pulsed DC1 (20-50 million) for eight consecutive weeks. They will Cap dose to be used in the 6422+Cap 3rd arm. Depending on the interim results, Cap dose in
alternate between WT ESR1 DC1s and mutated ESR1 DC1s. Patients will receive the 3rd arm will either be increased to 225 mg BID or decreased to 75 mg BID. Patients with
combination of DC1 vaccinations and Elacestrant at 345 mg orally daily concurrently triple-negative or HR positive/HER2 negative, advanced or metastatic breast cancer are
during vaccination and continued after. After the initial vaccination series, patients will eligible for the study. Patients should have measurable disease in accordance with RECIST
undergo radiological assessment of their disease, and if no evidence of progression they 1.1. Patients with stable brain metastases are eligible. The primary endpoint of the study is
will receive booster DC1s every four weeks x 3 doses. The primary objective of this pilot Objective Response Rate. This will be assessed based on the null hypothesis that the
study is safety and feasibility. Secondary objectives include preliminary efficacy, endpoint in each 6422+Cap arm is less than or equal to the monotherapy arm. Additionally,
biomarkers assessment, safety and patient reported outcomes. Tumor tissue and blood safety will be assessed by the incidence and severity of adverse events across treatment
samples will be collected for correlative analyses including ctDNA and changes in groups. The pharmacokinetics of Cap, 5-FU, and the FBAL catabolite will be evaluated
variant allele frequency of ESR1 during treatment. The study is open at H. Lee Moffitt using population PK analysis. Currently 3 patients have been enrolled in the study. Clinical
Cancer Center. Clinical trial information: NCT06691035. Research Sponsor: V trial information: NCT06568692. Research Sponsor: Processa Pharmaceuticals Inc.
Foundation.

TPS1134 Poster Session TPS1135 Poster Session

DATO-Base: A phase II study of DATOpotamab deruxtecan for patients with The efficacy and safety of eutideron, etoposide, and bevacizumab in pa-
breast cancer brain metastases or leptomeningeal disease. First Author: Paolo tients with brain metastases from breast cancer. First Author: Bishal Tiwari,
Tarantino, Dana-Farber Cancer Institute, Boston, MA Nassau University Medical Center, East Meadow, NY
Background: Approximately half of patients with metastatic TNBC and one fifth of Background: Brain metastases (BM) from breast cancer (BC) are a significant thera-
those with estrogen receptor (ER)+/HER2-negative metastatic breast cancer (MBC) peutic challenge, with limited systemic treatment options capable of crossing the blood-
eventually develop breast cancer brain metastases (BCBM), with an adverse prognostic brain barrier (BBB). Median overall survival (OS) ranges from 4 to 16 months, influenced
effect. Intracranially penetrant systemic therapies in HER2-negative MBC are very by molecular subtype and treatment modality. Triple-negative and HER2-positive
limited. In this setting, antibody-drug conjugates (ADCs) have shown promise, with subtypes are associated with higher BM incidence. There is a crucial need to explore
impressive intracranial activity observed with trastuzumab deruxtecan. Datopotamab systemic therapies that address both intracranial and extracranial disease. Eutideron, a
deruxtecan (Dato-DXd) is a novel anti-Trop2 ADC with robust antitumor activity in HER2- novel small-molecule inhibitor with robust CNS penetration, has demonstrated activity
negative MBC. In the TROPION-Breast01 phase 3 trial, Dato-DXd outperformed che- against metastatic BC models involving the brain. Early clinical studies suggest its
motherapy for ER+/HER2-negative MBC, and promising early-phase data was also efficacy in advanced BC, with intracranial activity and manageable toxicity. Combining
reported in triple-negative MBC. Preclinical data in tumor models found favorable in- eutideron with etoposide, a cytotoxic agent, and bevacizumab, an anti-VEGF monoclonal
tracranial penetration for Dato-DXd, and encouraging clinical data were reported in antibody, may enhance therapeutic outcomes. Bevacizumab, known for reducing BM-
patients with lung cancer brain metastases. Based on the relevant unmet need and the associated edema and improving quality of life, has shown promise in combination
promising preclinical and clinical data seen with Dato-DXd, there is a strong rationale in regimens but has not been evaluated alongside eutideron and etoposide. This Phase II
testing Dato-DXd for patients with HER2-negative MBC and BCBM or leptomeningeal trial investigates this novel three-drug regimen in patients with recurrent BC and
disease (LMD). Methods: DATO-Base is an ongoing, open label, multicenter, investi- measurable BM. Methods: This open-label, single-arm Phase II trial evaluates the
gator-initiated phase II trial for patients with HER2-negative MBC with active BCBM and/ efficacy and safety of eutideron, etoposide, and bevacizumab in female patients aged
or LMD. Eligible participants are women and men with HER2-negative active (newly $18 years with recurrent metastatic BC and measurable BM. Eligible patients had an
diagnosed/untreated or treated/progressive) brain metastases or LMD. Patients are ECOG performance status of 0–2, life expectancy $12 weeks, and progressed untreated
enrolled in one of three cohorts: Cohort A (n = 24) for HR+/HER2-negative BCBM; Cohort or previously treated BM not requiring immediate local treatment. Baseline brain MRIs
B (n = 24) for triple-negative BCBM; Cohort C (n = 10) for HER2-negative LMD (any ER confirmed at least one measurable CNS lesion per RANO-BM criteria. The treatment
status). Patients in Cohort A require prior treatment with at least one line of endocrine regimen includes eutideron (30 mg/m²/day, IV, Days 1–5 of a 21-day cycle), etoposide
treatment in the metastatic setting; no prior treatment is required for Cohorts B and C. (30 mg/m²/day, IV, Days 1–3 of a 21-day cycle), and bevacizumab (10 mg/kg, IV, Days 1
Prior treatment with approved or investigational ADCs is allowed. Participants receive and 21 of each cycle). After 4–6 cycles, responders continued bevacizumab mainte-
Dato-DXd 6 mg/kg IV on day 1 of each 21-day cycle until progression, unacceptable nance until progression or intolerable [Link] endpoint: CNS Objective Re-
toxicity, withdrawn consent, noncompliance, or death. The primary endpoint for Cohorts sponse Rate (CNS-ORR) per RANO-BM. Secondary endpoints: CNS Clinical Benefit Rate,
A and B is intracranial objective response rate per RANO-BM criteria. Patients in each CNS Progression-Free Survival, Overall Survival, and systemic ORR by RECIST 1.1.
cohort will be enrolled based upon Simon two-stage designs: if $1/9 patients respond, a Safety was monitored using CTCAE v5.0. The trial, targeting 43 patients across Chinese
total of 24 patients will be enrolled. Cohort C is exploratory, with description of overall centers, aims to inform future strategies for BC patients with BM. Clinical trial infor-
survival and exploratory endpoints. Blood and cerebrospinal fluid is being collected at mation: NCT05781633. Research Sponsor: None.
baseline, C2D2, and at progression for translational studies. The trial was activated in
December 2023, with enrollment ongoing. Clinical trial information: NCT06176261.
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
70s BREAST CANCER—METASTATIC

TPS1136 Poster Session TPS1137 Poster Session

Trial in progress: A study of Bria-OTS cellular immunotherapy in metastatic Trial in progress: ENCORE—Multicenter prospective registry of sequential
recurrent breast cancer. First Author: Neal Shiv Chawla, Sarcoma Oncology Center, antibody drug conjugates (ADCs) in HER2 negative metastatic breast cancer
Santa Monica, CA (MBC) (TBCRC-067). First Author: Laura Ann Huppert, University of California San
Background: Metastatic breast cancer is almost always fatal. Objectives: Primary: To Francisco, San Francisco, CA
evaluate the safety of BC1 cell line immunotherapy in patients with advanced late-stage Background: Antibody-drug conjugates (ADCs) have demonstrated substantial im-
metastatic breast cancer; Secondary: To evaluate the tumor response to BC1 cellular provement in progression free survival (PFS) and overall survival (OS) in phase III clinical
immunotherapy; Exploratory: To evaluate progression-free (PFS) and overall survival trials in patients with metastatic triple negative breast cancer (mTNBC) and hormone
(OS); To evaluate the immune responses elicited by BC1 cellular immunotherapy; To receptor positive/HER2-negative (HR+/HER2-) metastatic breast cancer (MBC), offering
evaluate patient and tumor characteristics that may be predictive of responses to HLA- an effective new treatment strategy. Several outstanding questions impact the use of
matched cellular immunotherapy; To evaluate time to subsequent therapy; and To these drugs clinically, and prospective real-world data is needed. First, it is important to
evaluate PFS 2 on subsequent therapy. Methods: Study Population: Patients with understand the safety and efficacy of these agents in a real-world population with
metastatic recurrent breast cancer after progression on prior therapies. Key Inclusion diverse patient characteristics. Second, it is critical to understand the safety and ef-
Criteria: Histologically-confirmed metastatic breast cancer after failure of standard ficacy of these ADCs in sequence. Third, it is essential to identify biomarkers that can
therapies; $18 years old; Expected survival of .4 months; Adequate performance help clarify mechanisms of response and resistance to ADCs, which may inform future
status (ECOG #2); Adequate hematologic and organ function; Clinically stable with sequencing and treatment strategies. Methods: This is a multicenter prospective
resolution of toxicities from previous treatment to baseline with the exception of al- registry study of patients with HER2-negative MBC who are treated with sequential ADCs
opecia. Key Exclusion Criteria: Concurrent anti-cancer treatment or concurrent cancer; per standard of care (SOC) with the goal to understand the safety and efficacy of
Anti-cancer treatment within 3 weeks of first treatment; History of hypersensitivity to sequential ADCs in a real-world setting (NCT06774027). A total of 100 participants with
study therapies; New York Heart Association stage 3-4 cardiac disease; Moderate- HER2-negative MBC will be enrolled in this study, either prior to starting their first ADC
severe pleural or pericardial effusion; Pregnant or nursing; HIV+; Known immunode- per SOC (cohort 1 = HR+/HER2-; cohort 2 = mTNBC) or prior to starting their second ADC
ficiency or ongoing treatment with immunosuppressive therapy .10 mg/day prednisone per SOC (cohort 3 = HR+/HER2-; cohort 4 = mTNBC). The dual primary endpoints are real-
equivalent; Severe psychiatric or other clinically progressive major medical problems. world progression free survival (rwPFS) of ADC1 and rwPFS of ADC2. Secondary
Study Design: This is an open-label study. Phase 1: BC1 cell line alone; Phase 2, Bria-OTS endpoints include overall response rate (ORR), duration of response (DOR), best overall
regimen with check point inhibitor (CPI). Phase 1: Patient 1: 20 million cells BC1 in- response (BOR), disease control rate (DCR), and real-world overall survival (rwOS), and
tradermally q2 wks x 8 wks (4 doses); Patient 2: 40 million cells of BC1; Patient 3: 60 safety for each ADC. Exploratory endpoints include translational correlates of response/
million cells BC1. If no DLT with BC1 monotherapy, the combinational phase of the study resistance to ADCs (e.g., circulating tumor DNA, circulating tumor cells, and tissue
will begin with BC1 and the Bria-OTS regimen q3 wks + CPI. During the Phase 1 spatial correlates) and patient-reported outcomes (PROs). rwPFS and rwOS will be
combination and Phase 2 expansion phases, all patients will be treated with BC1 cells as estimated by the Kaplan-Meier method. Statistics will be descriptive. Enrollment to start
part of the Bria-OTS regimen, which includes cyclophosphamide 300 mg/m2 2-3 days in the first quarter of 2025. Clinical trial information: NCT06774027. Research Sponsor:
prior to BC1 cell inoculation, and concurrent peg-interferon 0.6 mcg s.c. on the day of Gilead.
BC1 cell inoculation. Imaging studies: At screening, after monotherapy phase, before
combination phase, and q9 weeks thereafter for 6 months, then q12 weeks. Patients who
had PD but with clinical benefit may continue treatment. Subjects will continue to be
followed for time on subsequent therapy (PFS2) and survival q3 mos. for 2 years. The
phase 1 monotherapy part of the study has enrolled and treated 3 patients. Clinical trial
information: NCT06471673. Research Sponsor: BriaCell Therapeutics Corp.

TPS1138 Poster Session TPS1139 Poster Session

Update on phase III pivotal trial of Bria-IMT + CPI vs physician’s choice in Efficacy and safety of disitamab vedotin in combination with RC148 versus
advanced metastatic breast cancer (BRIA-ABC). First Author: Saranya Chumsri, albumin-bound paclitaxel 6 toripalimab for patients with HR-negative
Mayo Clinic Florida, Jacksonville, FL HER2-low-expressing unresectable locally advanced or metastatic breast
Background: The SV-BR-1-GM breast cancer cell line activates anti-tumor immunity by cancer: An open-label, randomized, controlled phase II study. First Author:
expressing tumor associated antigens and secreting GM-CSF which enhances dendritic Jiayu Wang, Cancer Hospital, Chinese Academy of Medical Sciences, Beijing, China
cell activation and promotes adaptive (T-cell mediated) and innate (dendritic and NK Background: Patients (pts) with hormone receptor (HR)-negative and HER2-low-
cell) immune responses. The cells are also engineered to optimize immune recognition expressing (defined as IHC 1+, or IHC 2+/ISH-) advanced breast cancer have poor
through pt specific HLA antigen matching. SV-BR-1-GM acts through direct antigen prognosis and more effective treatment options are needed. Disitamab vedotin (DV) is a
presentation and CD4+ T-cell activation and, when combined w/ checkpoint inhibitors novel humanized anti-HER2 antibody conjugated with monomethyl auristatin E (MMAE)
(CPIs), has demonstrated clinical benefit in 54 heavily pretreated metastatic breast via a cleavable linker. DV alone or in combination with a PD-1 inhibitor showed en-
cancer (MBC) pts. In pts w/ disease progression following CPI therapy, similar or couraging antitumor activities with manageable safety in pts with HER2-low-expressing
improved progression free survival (PFS) compared to their prior treatment regimen. (IHC 1+, or IHC 2+/ISH-) advanced or metastatic breast cancer, gastric cancer and other
Disease control following antibody drug conjugates was observed in 40%of pts. Clinical solid tumors (Wang J., et al., Cancer Commun, 2024; Wang Y., et al., eClinicalMedicine,
benefit was seen in 5 out of 8 pts w/ untreated intracranial metastases. CD8+ Immuno- 2024). RC148 is a bispecific monoclonal antibody directed against programmed death
PET imaging suggests systemic activation, w/ increased CD8+ tumor infiltrating receptor-1 and vascular endothelial growth factor receptor. DV+RC148 combination is
lymphocytes at both primary and metastatic tumor sites, as well as lymphoid organs. expected to exert a synergistic antitumor effect by improving the tumor immune mi-
Optimized sequencing of CPI w/ SV-BR-1-GM and its latest phase 3 formulation have croenvironment. We aim to evaluate the efficacy and safety of DV plus RC148 versus
shown enhanced clinical outcomes, including improved overall survival (OS) (median albumin-bound paclitaxel 6 toripalimab in pts with HR-negative HER2-low-expressing
13.4 mos), PFS (3.6 mos), and clinical benefit rate (CBR; 61%).These findings have advanced breast cancer in this randomized phase II trial (NCT06642545). Methods: The
informed refinements to the ongoing pivotal, registration enabling Phase 3 trial, key eligibility criteria are pts aged 18 years or older with unresectable stage III or stage IV
designed to optimize pt selection and treatment sequencing strategies. Methods: This breast cancer, negative HR status, low HER2 expression (defined as IHC1+, or IHC2+/
ongoing multicenter, randomized, open label Phase 3 trial evaluates Bria-IMT + CPI vs. ISH-), no previous chemotherapy for locally recurrent or metastatic disease, and no
Treatment of Physician’s Choice (TPC) in MBC pts lacking approved curative therapies. disease recurrence within 6 months after treatment completion (within 12 months if
Pts are randomized [Link] to Bria-IMT + CPI, TPC, or Bria-IMT monotherapy (discontinued using taxanes) if with radical treatment. Pts who previously received anti-HER2 therapy
after 150 enrollments to prioritize combination arms). The Bria-IMT regimen consists of: or immunotherapy are excluded (except pts receiving neoadjuvant/adjuvant PD-[L]1
Day -2: Cyclophosphamide 300 mg/m², Day 0: 20M irradiated SV-BR-1-GM cells, Day 2/3: inhibitors 12 months prior to recurrence or progression). Pts will be randomized
0.1 mcg pegylated a interferon at each inoculation site. CPI infusion is administered Day (stratified by PD-L1 expression status: positive or negative) in a ratio of 1:1 to receive DV
-3 to 3. Cycles q3w. TPC regimens follow site specific SOC. Imaging q6w (first 2 cycles), (2.0 mg/kg) plus RC148 (20 mg/kg) intravenously once every two weeks or to receive
then q8w. Eligibility includes all MBC subtypes, including CNS mets, and permits prior albumin-bound paclitaxel (125 mg/m2 day 1 and day 8) 6 toripalimab (240 mg day 1)
CPI therapy ( .21 days pre-treatment). There will be 100 sites across the U.S., Canada, intravenously every three weeks until occurrence of disease progression or intolerable
and ex-North America w/ an enrollment target of 404. The trial is currently active at 59 toxicity. The primary endpoint is objective response rate (ORR) in all pts per investi-
sites with 217 sub investigators. To date, 67 pts screened: 46 randomized (median age gator’s assessment according to RECIST v1.1. The secondary endpoints are ORR in the
56 yrs [34–83], median 6 [2–13] prior lines of therapy. The primary endpoint is OS, with PD-L1-positive pts; investigator-assessed progression-free survival, disease control
an interim analysis at 144 events targeting a hazard ratio of 0.6. Secondary endpoints: rate, duration of response, and overall survival in all pts and the PD-L1-positive pts. This
PFS, overall response rate, CBR, CNS event free survival, and TWiST. Safety analyses study was initiated in August 2024. Clinical trial information: NCT06642545. Research
ongoing; pt reported outcomes assess subjective treatment impact. Clinical trial in- Sponsor: RemeGen Co., Ltd.
formation: NCT03328026. Research Sponsor: BriaCell Therapeutics Corp.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 BREAST CANCER—METASTATIC 71s

TPS1140 Poster Session TPS1141 Poster Session

A phase Il trial to assess the impact of b2 adrenergic receptor (b2-AR) Emiltatug ledadotin (Emi-Le): A B7-H4-directed dolasynthen antibody-drug
blockade in metastatic triple negative breast cancer (mTNBC). First Author: conjugate (ADC) being investigated in phase 1 dose expansion in patients
Shipra Gandhi, Roswell Park Comprehensive Cancer Center, Buffalo, NY with triple negative breast cancer who received at least one prior top-
Background: In PD-L1+ mTNBC patients (pts), the standard of care treatment is oisomarase-1 inhibitor ADC. First Author: Hyo S. Han, H. Lee Moffitt Cancer
chemotherapy and pembrolizumab (P) in the first-line setting. Our group and others have Center and Research Institute, Tampa, FL
demonstrated that chronic b2-AR signaling suppresses CD8+ cytotoxic T lymphocytes Background: Breast cancer is the leading cause of cancer death for women worldwide,
(CTL) function, drives their exhaustion, and increases the number of immunosuppressive with triple-negative breast cancer (TNBC) considered one of the more aggressive breast
myeloid-derived suppressor cells (MDSCs) and regulatory T cells (Tregs) in the tumor cancers, accounting for ~15% of all cases. Unfortunately, there remains an unmet
microenvironment (TME), thus supporting tumor proliferation. Consequently, abrogation medical need for effective and well-tolerated treatments for advanced/metastatic TNBC;
of b-AR signaling using the pan b-blocker propranolol or b-AR-/- knockout mice in- in heavily pretreated patients, standard-of-care single-agent chemotherapy has limited
creased the intratumoral frequency of CTLs and elevated the CTL:Treg ratio (Bucsek et efficacy, with response rates of ~5%, PFS ~7 weeks. Emiltatug ledadotin (Emi-Le; XMT-
al. Cancer Res. 2017; PMID: 28819022). Similarly, mouse tumor models also demon- 1660) is a B7-H4-directed Dolasynthen ADC designed with a precise, target-optimized
strated decreased exhaustion markers (PD1, TIM3, LAG3) on CTLs when b-AR was drug-to-antibody ratio (DAR 6) and a proprietary auristatin F-HPA microtubule inhibitor
blocked, via propranolol (Qiao G et al. Cancer Immunol Res. 2021, PMID: 33762351). payload with controlled bystander effect. The FDA has granted Emi-Le two Fast Track
Confirming this phenomenon, we have shown, in a prospective clinical trial in metastatic designations for the treatment of adult patients with breast cancer, including patients
melanoma, that b-AR blockade with propranolol significantly increased response to P with TNBC who have previously been treated with topoisomerase-1 inhibitor (topo-1)
with an objective response rate (ORR) of 78%, as opposed to 30-40% with P alone ADCs. Initial dose escalation clinical data from the ongoing Phase 1 trial at doses
(Gandhi et al. Clin Cancer Res 2021, PMID: 33127652). Moreover, clinical b-AR blockade ranging from 38.1-67.4 mg/m2 per cycle demonstrated a 23% confirmed response rate
was associated with higher immune infiltration in the TME (Hiller JG, Clin Cancer Res in patients with B7-H4 high TNBC who were heavily pretreated all of whom received at
2020, PMID: 31754048). Therefore, we hypothesize that using propranolol with che- least one prior topo-1 ADC. Methods: Based on encouraging clinical activity and tol-
motherapy and P should improve response for pts with newly metastatic PD-L1+ TNBC. erability data in the initial dose escalation data, the expansion portion (EXP) of the Phase
Methods: This is a phase II single-arm, non-randomized multi-center study. Pts are 1 trial has been initiated and is actively enrolling patients. EXP has a Simon 2-stage
women $18 yrs with PD-L1+ mTNBC, who will receive propranolol, chemotherapy design and will evaluate two doses in patients with advanced/metastatic TNBC who have
(paclitaxel, nab-paclitaxel, gemcitabine-carboplatin) and P in the upfront setting: received 1-4 prior lines of systemic therapy, including at least one topo-1 ADC. Patients
chemotherapy on days 1, 8 and P on day 1 every 3 weeks in addition to propranolol 30 mg will be evaluated for B7-H4 expression prospectively by IHC and will be stratified into B7-
BID, with intra-pt propranolol dose-escalation by 10 mg BID weekly to a total of 80 mg H4 TPS “high” and B7-H4 TPS “low” cohorts. The first EXP dose is 67.4 mg/m2 Q4W.
BID as tolerated by blood pressure and heart rate as natural biomarkers for dose. Dose exploration is ongoing to identify a potential second higher EXP dose. The protocol
Treatment will continue until disease progression per RECIST. The primary endpoint is includes the option for multiple additional indications, including HR+/HER2- breast
ORR, defined as complete or partial response. The secondary endpoint is safety, 6-month cancer, endometrial cancer, ovarian cancer, and ACC-1. Clinical trial information:
progression-free and overall survival. As an exploratory endpoint, changes in TME and NCT05377996. Research Sponsor: Mersana Therapeutics.
blood immune markers will be assessed. In stage 1, n1=23 evaluable pts will be enrolled.
If $ 13/23 responses are observed, then the study will continue to enroll another n2=14
pts for a total of n=37, otherwise will be suspended for futility. If $ 24/37 responses are
observed, then the proposed therapy will be considered promising. Pre- and 6-week post-
treatment tumor biopsies and blood samples will be analyzed for changes in stress-
induced biomarkers (epinephrine, norepinephrine, and frequency of CTL, MDSC, Treg)
and exhaustion markers (PD1, TIM3, LAG3). The study is currently open and has accrued
one patient. Clinical trial information: NCT05741164. Research Sponsor: NIH (NCI).

TPS1142 Poster Session TPS1143 Poster Session

TBCRC 058: A randomized phase II study of enzalutamide, enzalutamide AXALAP: Phase Ib study of axatilimab in combination with olaparib in
with mifepristone, and treatment of physician’s choice in patients with BRCA1/2 and PALB2-associated metastatic HER2-negative breast cancer
androgen receptor-positive metastatic triple-negative or estrogen recep- (BC). First Author: Filipa Lynce, Dana-Farber Cancer Institute, Boston, MA
tor-low breast cancer (NCT06099769). First Author: Tiffany A. Traina, Memorial Background: Poly(ADP-ribose) polymerase (PARP) inhibitors (PARPi) have revolutionized
Sloan Kettering Cancer Center, New York, NY the treatment of patients (pts) with germline BRCA1/2 (gBRCA)-associated HER2-negative
Background: Triple-negative breast cancer (TNBC) refers to a heterogenous group of BC. However, resistance eventually occurs in almost all pts. Tumor-associated macrophages
breast cancers (BC) that lack expression of ER, PR, and HER2. Despite recent advances (TAMs), a key component of the BC tumor microenvironment, are highly immunosuppressive
with immunotherapy (IO) and antibody-drug conjugates (ADCs), TNBC remains the most and associated with poor clinical outcomes. In preclinical immunocompetent models of
aggressive subtype, characterized by a high risk of recurrence and a short overall survival BRCA-associated BC, PARP inhibition induces suppressive CSF-1R+ TAMs, contributing to
in the metastatic setting. BCs with low levels of ER and PR expression (1-10%) clinically resistance, so that combining anti-CSF-1R therapy with PARPi significantly enhances
behave like TNBC, and clinical management follows the TNBC treatment (tx) paradigm. progression free survival (PFS) compared to PARPi monotherapy (190d vs 92d). Axatilimab
We and others have identified a subset of ER/PR/HER2 negative breast cancers (BC) that (SNDX-6352; Ab969.g2), a humanized IgG4 monoclonal antibody targeting CSF-1R, reduces
express the androgen receptor (AR). Enzalutamide (enza), an AR-antagonist, has dem- TAMs, potentially slowing tumor growth and enhancing anti-tumor immunity. In the Phase I
onstrated activity in AR+ metastatic TNBC (Traina et al, JCO 2018). Activation of the SNDX-6352-0502 study for advanced solid tumors, axatilimab showed tolerability at the
glucocorticoid receptor (GR) has been implicated as a mechanism of resistance to AR highest dose (6 mg/kg), with biomarker modulation observed at doses as low as 1 mg/kg.
inhibition in prostate and BC (Kach et al, Sci Transl Med 2015). Advanced TNBC remains Axatilimab is FDA-approved for chronic graft-versus-host disease (cGVHD). Methods:
an area of unmet need, particularly in patients who are ineligible for or progress following AXALAP (NCT06488378) is a non-randomized open-label, proof-of-concept phase 1 study
a checkpoint inhibitor. This randomized study will evaluate the efficacy of enza or enza designed to evaluate axatilimab 1mg/kg or 3mg/kg every 2 weeks in combination with
plus the GR antagonist mifepristone (mif) as compared to physician’s choice chemo- olaparib 300 mg twice daily in pts with somatic or germline BRCA1/2- and PALB2-associated
therapy (TPC). Methods: This is a randomized phase II trial; 201 patients (pts) will be HER2-negative metastatic BC. Patients must be PARPi naı̈ve, or have not progressed on
prior PARPi, and have received up to 2 prior lines of chemotherapy for metastatic disease.
randomized [Link] to enza, enza with mif, or TPC (carboplatin, paclitaxel, eribulin, or
Pts will receive a two-week lead-in of olaparib monotherapy, followed by combined olaparib
capecitabine). The primary endpoint (endpt) is progression free survival (PFS), and the
and axatilimab. Pts will undergo mandatory tumor biopsies pre-treatment, after the 2-week
trial is designed to test the hypothesis that PFS in the pooled enzalutamide arms is
olaparib lead-in, and after 2 cycles of olaparib/axatilimab, with an optional biopsy at time-of-
superior to TPC; there is 80% power to detect a hazard ratio (HR) of 0.70, corresponding to
progression. Primary objectives are to establish the maximum tolerated dose (MTD) and
increase in PFS from 3.5 months (mos) with TPC to 5.0 mos with enza-based tx. recommended phase 2 dose and to assess the safety and tolerability of axatilimab and
Secondary endpts include pairwise comparisons of PFS among the 3 arms and evaluation olaparib. Secondary objectives include assessment of changes in CSF-1R+ CD163+
of response rate, clinical benefit rate, duration of response, overall survival, safety/ macrophage levels after olaparib monotherapy and after 2 cycles of combination treatment
toxicity, and patient-reported outcomes by arm. Exploratory endpts include correlation of at the MTD; to determine the objective response rate and the median PFS of the combination
tumor and circulating markers (constitutively active AR variants in circulating tumor cells per RECIST 1.1 criteria. The MTD will be determined by Bayesian Optimal Interval (BOIN)
and circulating tumor cell DNA) with tx response. Eligible pts must have: ECOG 0-2, design. Pts will be treated in cohorts of 3 with a maximum of 10 at each dose. If the MTD is
metastatic ER/PR low or negative, HER2 0-2+ (FISH not amplified) (BC), measurable or identified as 3mg/kg, we will complete enrollment of 10 pts at 1 mg/kg to assess biological
evaluable disease (dz) per RECIST v1.1, , 3 prior lines of chemotx, any # prior endocrine effectiveness and clinical efficacy of the lower dose. We expect to treat up to 20 pts, who will
txs, no prior anti-AR tx or CYP17 inhibition, no prior mif. Pts with PD-L1+ BC must have receive study treatment until development of unacceptable toxicity or disease progression.
received prior IO if not contraindicated. Tumors must have AR .10%, normal organ Enrollment began on 5/2024 at Dana-Farber Cancer Institute and the trial will also open at
function, no history of brain mets. As of 1/23/25, 11 of 201 pts have begun protocol- Beth Israel Deaconess Medical Center and Mayo Clinic-Rochester. Clinical trial information:
specified tx. Clinical trial information: NCT06099769. Research Sponsor: Breast Cancer NCT03604692. Research Sponsor: Incyte; U.S. National Institutes of Health; U.S. National
Research Foundation; TBCRC; Astellas; Corcept; The TaTa Sisterhood Foundation; Pfizer. Institutes of Health.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
72s CARE DELIVERY/MODELS OF CARE

1500 Oral Abstract Session 1501 Oral Abstract Session

Pharmacist-led medication reconciliation televisit (MRT) for phase 1 on- Remote clinical pharmacist impact on reducing total cost of care in En-
cology clinical trials: A telemedicine model. First Author: Sarah O’Neill, hancing Oncology Model–enrolled oncology practices. First Author: Daniel
Massachusetts General Hospital Cancer Center, Boston, MA Kendzierski, McKesson Specialty Health, The US Oncology Network, The Woodlands, TX
Background: Accurate medication reconciliation is critical in oncology trials to mitigate drug- Background: The national cost of cancer care is estimated to exceed $245 billion by 2030,
drug interactions (DDIs), address eligibility and enhance patient safety. Cancer patients often primarily due to the high cost of cancer drugs. The Enhancing Oncology Model (EOM) is a
experience polypharmacy ($5 medications), increasing the risk of additive toxicities when voluntary 6-month, 2-sided, risk-based payment model implemented by the Centers for Medicare
investigational agents are introduced. Medication lists in Electronic Medical Records (EMRs) are and Medicaid Services (CMS) to improve cancer care while simultaneously reducing the total
often incomplete or outdated, complicating eligibility evaluations. To streamline screening for cost of care (TCOC). The US Oncology Network (The Network) comprises approximately 50% of
Phase 1 trials, a pharmacist-led medication reconciliation televisit (MRT) model was imple- all providers participating in EOM nationwide across twelve practice sites. In The Network, drug
mented to improve baseline medication list accuracy and reduce in-clinic time. Methods: From costs represented an average of 63% of a patient’s TCOC. This study aims to demonstrate the
December 2022 to January 2025, one MRT was scheduled for each patient screening for Phase 1 impact of a pharmacist in reducing TCOC in the EOM model. Methods: Medication initiatives
trials ($18 years, English-speaking or with interpreter support), after trial consent but before were clinically evaluated and adopted at an individual practice level and included: monoclonal
registration. Pharmacists reviewed EMRs and dispense histories, then conducted structured antibody (moAB) dose rounding, pembrolizumab dose banding, biosimilar therapeutic inter-
phone interviews with patients to review prescription (RX), over-the-counter (OTC), herbal and change (TIC) to preferred products, use of a preferred PD-1 agent in metastatic NSCLC, de-
cannabis product usage, including name, strength, dose, frequency, start dates, indications and creased upfront usage of long-acting growth factor (GF) in metastatic cancer, and preferred use
ingredients. Inactive medications were discontinued, allergies were updated, patient concerns of zoledronic acid over alternatives. ClinReview pharmacists (CRP) remotely reviewed oncology
and medication details were documented in the EMR. Results: A total of 525 MRTs across 82 treatment orders for cost-savings opportunities. CRPs updated eligible treatments per practice
trials had a median turnaround time of 2 days. Patients (median age 61 years) reported a median protocols or reviewed with the treating oncologist. Interventions were submitted by the CRP
of 12 medications (range 2–41) and a median total time spent of 45 minutes (range 5–330) into a tracking system and marked as an EOM-related intervention. TCOC reduction was
including documentation. 4.8% of patients required interpreter support. Primary cancer types calculated using the difference between the CMS allowable for the original treatment ordered
(12 total) included gastrointestinal (32%), breast (20%) and head and neck (11%). Table 1 and the new order. Results: From July 1, 2023, to December 31, 2024, seven CRPs within five of
summarizes MRTs April 2024 and later which captured additional data: 64% required an EMR- The Network’s EOM participating practices evaluated over 5,600 patients for medication ini-
prompted outside source reconciliation, 32% modified allergies, a median of 3 medications were tiatives. A total of 1,271 interventions were identified, with 1,180 accepted. The sum of TCOC
added (range 0–37), 2 were changed (range 0-13), 3 were discontinued (range 0-18) and the reduction amounted to 8,982,235. Further breakdown of each initiative and average TCOC
median call time was 18 minutes. Conclusions: Pharmacist-led MRTs provided substantial reduction per intervention are shown in Table 1. In addition to the six initiatives, the CRP
value for investigators and research nurses, enabling them to focus on patient care. Flexible contributed an additional 1,201,326 in medication savings associated with drug selection.
scheduling of remote MRTs supported presence of caregivers and medications, reducing list Conclusions: CRP’s medication initiatives within The Network’s EOM participation reduced
omissions and hospital chair time for patients. Sponsors and study teams gained more accurate TCOC by nearly $9 million across five practices. Key initiatives such as pembrolizumab dose
baseline records, lowering the risk of unknown prohibited medications. This scalable tele- banding and preferred use of zoledronic acid were the largest contributors. These findings
medicine model offers potential for broader use in oncology trials, improving efficiency and demonstrate the potential for pharmacist-driven interventions to lower costs and drive the
patient safety. Research Sponsor: None. success of value-based care models in oncology practices. Research Sponsor: None.
MRT identification of previously undocumented medications with DDI potential (n=235). TCOC Average TCOC Reduction
Before MRT After MRT EOM Initiative n (%) Reduction, $ per intervention, $
> 1 medication on EMR categorized as: n (%) n (%) Absolute Change % Increase
moAB dose rounding 443 (35) 1,537,273 3,470
RX 208 (98%) 232 (99%) +24 +12% Pembrolizumab dose banding 106 (8) 1,962,105 18,510
OTC Non-herbal 179 (84%) 224 (95%) +45 +25% TIC 356 (28) 1,510,945 4,244
OTC Herbal 26 (12%) 46 (20%) +20 +77% Preferred PD-1 agent for NSCLC 26 (2) 153,117 5,889
Antacid/H2 Blocker/PPI 81 (38%) 107 (46%) +26 +32% Decrease GF use 37 (3) 109,822 2,968
Cannabis 11 (5%) 66 (28%) +55 +500% Zoledronic acid use 181 (14) 2,157,895 11,992

1502 Oral Abstract Session 1503 Oral Abstract Session

Video-based genetic counseling to reduce physician workload and enhance Electronic patient-reported outcome-based weight management versus
consulter understanding: A prospective randomized clinical trial. First Author: usual care during induction chemotherapy followed by concurrent chemo-
Georg Pfeiler, Department of Obstetrics and Gynecology and Center for Breast Health, radiotherapy in nasopharyngeal carcinoma: A phase II randomized con-
Comprehensive Cancer Center, Medical University of Vienna and Austrian Breast and trolled trial. First Author: Qiu-Yan Chen, Department of Nasopharyngeal Carcinoma,
Colorectal Cancer Study Group, Vienna, Austria Sun Yat-Sen University Cancer Centre, Guangzhou, China
Background: Genetic counseling is an essential part of germline genetic testing, which is Background: Nearly almost patients with nasopharyngeal carcinoma (NPC) experience
reccomended for use of PARP inhibitors in breast cancer treatment. Despite its importance, weight loss (WL) and malnutrition during treatment. However, effective patients’ weight
only about 40-60% of breast cancer patients receive genetic counseling. A video tool has management remains a challenge, as no standardized approach currently exists. This study
been developed to provide genetic counseling, aiming to reduce the physician’s workload aimed to evaluate the feasibility and efficacy of an electronic patient-reported outcome
and improve the patient’s understanding. Methods: Advice seekers at increased risk for (ePRO) based weight management in mitigating WL and malnutrition in patients with locally
hereditary breast and ovarian cancer as well as breast and ovarian cancer patients were advanced NPC (LA-NPC) receiving induction chemotherapy (IC) followed by concurrent
included in the trial. They were randomly assigned 1:1 to either standard of care (physician chemoradiotherapy (CCRT). Methods: In this phase II randomized controlled trial, eligible
only, PO) or video based followed by physician genetic counseling (VPO). A 15-minute video patients aged 18 to 70 years with stage III-IVa (AJCC 8th) NPC were randomly assigned in a 1:
tool was created for VPO participants, who watched it on an iPad and answered 6 com- 1 ratio to either ePRO-based weight management (ePRO group) or usual care (UC group).
prehension questions online. The physician then clarified any misunderstood topics. In both Body weight, NRS2002, PG-SGA, EORTC QLQ-C30 and hematological indicators were col-
groups, counseling time was measured from conversation start to blood donation. Af- lected at six timepoints during treatment. Patients in ePRO group used the ePRO weight
terwards, participants completed a questionnaire with 9 comprehension questions (16 management system (ePRO-WMS). When predefined e-alert were triggered (e.g., WL . 5%),
points total). Data analysis included Bernard’s and Pearson’s tests for categorical data, the MDT promptly communicated and responded with patients. Patients in UC group did not
Kendall’s test for correlations, and ordinal logistic regression for multivariable analysis. use the ePRO-WMS and had no e-alerts. The primary outcome was the proportion of patients
Results: A total of 110 participants with a median age of 47 years were randomized into two experiencing WL . 10% at one month after completing CCRT (Post-1m). The secondary
groups: PO counseling (55 participants, 50%) and VPO counseling (55 participants, 50%). outcomes included WL . 10% and WL . 20% at the end of CCRT (W7-CCRT), as well as
Among them, 29% (32 participants) received therapeutic counseling for breast/ovarian comparisons of NRS2002, PG-SGA, EORTC QLQ-C30, and hematological indicators. WL was
cancer, while 71% (78 participants) received predictive counseling with no cancer diagnosis. compared using the x² test, while other secondary outcomes were analyzed with linear mixed-
Participant characteristics were well balanced between groups, with no significant dif- effects models ([Link], NCT05834712). Results: From May 2022 to August 2023,
ferences in age, indication for counseling, sex (90% female), level of education, or German- 112 patients were enrolled in each group (76.8% male, 86.6% with $ junior high school
speaking proficiency. Participants reported their sources of genetics knowledge as previous education). 110 patients (UC group) and 109 patients (ePRO group) completed IC+CCRT.
knowledge (30%) and physician counseling (70%) in the PO group, and as previous Among ePRO group, 107 patients successfully completed the ePRO-WMS, generating 561 e-
knowledge (20%), physician counseling (30%), and the video tool (45%) in the VPO group. alerts. Compared to UC group, ePRO group had significantly fewer patients with WL . 10% at
The video significantly improved comprehension scores from 62.5% (10 points, PO group) to Post-1m (37.5% vs. 57.1%, P = 0.003) and W7-CCRT (22.3% vs. 38.4%, P , 0.001). Addi-
81.3% (13 points,VPO group) (p , 0.0001). Additionally, the video tool significantly reduced tionally, ePRO group showed improved nutritional status (Retinol Binding Protein, 38.3 vs.
the time physicians spent on counseling from 6.6 minutes to 2.4 minutes (p , 0.0001). 34.8, P = 0.011) and quality of life. Notably, reductions in inflammation (C-reactive protein,
According to the logistic regression model, both the level of education (estimate -1.34, p = 10.4 vs. 15.9, P = 0.027) and immune suppression (CD4+CD25+ regulatory T cells, 20.1 vs.
0.002) and German language comprehension level (estimate 1.62, p = 0.001) significantly 22.2, P = 0.034) were also observed. No significant differences were found between groups for
influenced the genetic counseling comprehension score. Conclusions: In this prospective acute adverse and progression-free survival (median follow-up time: 16 months). However, 2
randomized trial, video genetic counseling improved comprehension and reduced physician cases of nasopharyngeal necrosis were observed in the UC group during follow-up.
counseling time. The genetic video tool, which can be translated into various languages, Conclusions: Compared to usual care, ePRO-based weight management in LA-NPC pa-
facilitates genetic counseling by decreasing the workload for physicians, which may in- tients mitigated weight loss, alleviated inflammation, and significantly improved nutritional
crease the genetic counseling rate in the clinic. Research Sponsor: None. status and quality of life. Clinical trial information: NCT05834712. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 73s

1504 Oral Abstract Session 1505 Oral Abstract Session

A novel virtual reality supportive care intervention (BMT-VR) for patients Geriatric assessment and management with a question prompt list using a
undergoing hematopoietic stem cell transplantation (HSCT): A pilot ran- web-based application to reduce treatment toxicity in older patients with
domized clinical trial. First Author: Hermioni L. Amonoo, Department of Supportive cancer: A randomized controlled trial (J-SUPPORT 2101 study). First Author:
Oncology, Dana-Farber Cancer Institute; Department of Psychiatry, Mass General Ayumu Matsuoka, Division of Survivorship Research, National Cancer Center Institute for
Brigham; Harvard Medical School, Boston, MA Cancer Control, National Cancer Center, Tokyo, Japan
Background: Patients with hematologic malignancies undergoing HSCT experience im- Background: Older adults with cancer experience aging-related physical, psychosocial and
mense physical and psychological symptom burden during their extended transplant hos- cognitive challenges that require comprehensive communication with their oncologists.
pitalization. Interventions that help manage patients’ psychological distress and improve their Geriatric assessment (GA) can assess these aging-related problems and guide manage-
quality of life (QOL) during this inpatient stay are limited. Virtual reality (VR), with its three- ment. Communication support may further facilitate the implementation of GA-guided
dimension capabilities for user engagement, offers a novel delivery modality for scalable, management (GAM). We report secondary outcomes of a single-blind, parallel-group,
targeted, and patient-centered supportive care interventions aiming to address the persistent multicenter, randomized controlled trial evaluating the efficacy of a program combining
unmet psychosocial needs of these patients. Methods: We conducted a pilot randomized GAM recommendations and communication support to facilitate aging-related communi-
clinical trial (RCT) of a VR supportive care intervention (BMT-VR). Patients undergoing HSCT cation between older Japanese patients with cancer and their oncologists. Methods: This
were randomly assigned to BMT-VR or usual care during their 3-4-week hospitalization. BMT- study included patients aged $70 years with advanced or recurrent gastrointestinal cancers
VR consisted of five self-directed modules addressing 1) supportive psychoeducation and who were scheduled to receive first- or second-line systemic therapy and had impairment in
managing expectations during HSCT; 2) effective coping; and 3) acceptance and gratitude at least one GA domain as assessed with a web-based application at baseline. In the
while dealing with uncertainty. The primary endpoint was feasibility ($60% of eligible pa- intervention group, GAM recommendations and a question prompt list were given to patients
tients enrolling, and $60% of BMT-VR participants completing $3/5 modules). To assess by trained intervention providers to be shared with their oncologists at the first outpatient
BMT-VR’s acceptability, we used the System Usability Scale ( . 80 = excellent acceptability). visit after randomization. During 5 months after the initial intervention, the implementation
We assessed psychological distress (Hospital Anxiety and Depression Scale), QOL (Functional of the GAM recommendations was reviewed monthly by the intervention providers with the
Assessment of Cancer Therapy-BMT), post-traumatic stress symptoms (PTSD-Checklist), patients and their oncologists. The control group received usual care. Secondary outcomes
coping (Measure of Current Status-A), and self-efficacy (Cancer Self-Efficacy Scale) at included the incidence of grade 3-5 adverse events (National Cancer Institute Common
baseline (i.e., 3 days post-HSCT) and 4-, 12-, and 24-weeks post-HSCT. We used analysis of Terminology Criteria for Adverse Events, ver. 5.0), dose modifications, early treatment
covariance (ANCOVA) to explore the preliminary effects of BMT-VR on outcomes. discontinuation, unplanned hospital utilization during 3 months, overall survival rate at
Results: We enrolled 58.3% (81/139) of eligible patients (BMT-VR (n = 40); usual care (n = 6 months, and health-related quality of life and patient satisfaction at 3 and 6 months.
41)) with a mean age of 57.9 (SD = 14.7) and 51.9% women. 74.4% of BMT-VR participants Results: Between September 2021 and September 2023, 215 patients (99 women, 116 men;
completed $3/5 modules and 65.1% completed 5/5 modules, with median acceptability median age 75 [range 70-88] years) were randomized (n=108/107 in the intervention/control
score = 81.2. At 4-weeks, BMT-VR vs. usual care participants reported improved anxiety (5.3 group). No differences were found between the groups in patient background character-
vs. 3.6, P = 0.016), QOL (108.2 vs. 96.8, P = 0.014), coping (36.6 vs. 32.4, P = 0.023), and self- istics. The incidence of any grade 3-5 adverse event was significantly lower in the inter-
efficacy (144.6 vs. 131.9, P = 0.019). Although BMT-VR vs. usual care participants reported vention group than in the control group (50.9% vs. 66.4%, P,0.05); the incidence of
sustained improvements in QOL (B = 3.8, P = 0.002), coping (B = 1.8, P = 0.011), and self- hematologic toxicities was significantly reduced (37.0% vs 55.1%, P,0.01), while that of
efficacy (B = 4.5, P = 0.017), BMT-VR effects became more pronounced for depression (B = non-hematologic toxicities remained similar (29.6% vs 36.4%, P=0.31). Early treatment
-0.5, P , 0.001), and PTSD (B = -1.7, P , 0.001) symptoms longitudinally across all time discontinuation was also significantly lower in the intervention group than in the control
points. Conclusions: A novel VR-delivered supportive care intervention tailored to the group (26.9% vs. 43.0%, P,0.05). No significances were found in other secondary outcomes
psychosocial needs of HSCT recipients is feasible and acceptable and demonstrated pre- including overall survival rate at 6 months (77.7% vs 78.7%, P=0.88) Conclusions: Our
liminary efficacy for improving psychological distress and QOL. A subsequent multi-site RCT program combining GAM recommendations with communication support significantly
will evaluate BMT-VR’s efficacy for improving outcomes in diverse HSCT settings. Clinical trial reduced severe treatment toxicity without compromising survival in older patients with
information: NCT05629676. Research Sponsor: Doris Duke Charitable Foundation; Clinician cancer. Clinical trial information: UMIN000045428. Research Sponsor: Japan Agency for
Scientist Development Award; National Cancer Institute; K08CA251654. Medical Research and Development.

1506 Oral Abstract Session

A technology-enabled clinical trial program’s impact on patient screening
and trial enrollment in 2024. First Author: Samantha Mallahan, Tempus AI, Inc.,
Chicago, IL
Background: Typical workflows for clinical trial start-up and screening are time- and resource-
intensive. The Tempus AI TIME program offers a novel clinical trial solution, collaborating with
clinical sites to increase trial access and alleviate site burden by streamlining study activation and
screening methods. Methods: The TIME program consists of an algorithmic trial screening platform
(TApp), team of oncology nurses, diverse trial portfolio, and rapid study activation processes.
Patient-level clinical information is centralized within the TIME database and includes structured and
unstructured data generated from Electronic Medical Record integration, next generation se-
quencing results, and natural language processing models. The TApp used this data combined with
trial eligibility criteria to algorithmically match patients to TIME trials. TApp searches were triggered
by changes to study criteria and/or updates to clinical data. Algorithmic matches were filtered based
on site capabilities, site interest in the trial, and trial lookback criteria, which defined the required
recency of a patient’s latest clinical document or encounter. Qualifying matches were then reviewed
by a Tempus nurse and sent to sites if confirmed eligible. Trial activations followed TIME’s
streamlined operational methods using a pre-negotiated rate card for site reimbursement of all
clinical trial activities, standardized clinical trial agreement, and central IRB. Trials could be activated
prospectively before the first eligible patient was identified, or in a “just-in-time” (JIT) manner if a
patient was ready to consent. Data collected included TIME network information, TApp and nurse
screening results, activation timelines, and enrollments across all active TIME sites and trials from
01/01/2024 - 12/31/2024. Results: During 2024, the TIME network consisted of 87 sites (79
Community, 8 Academic) and 98 trials. The TApp completed 1,323,259,353 searches across
1,281,676 patients resulting in 2,251,505 potential TApp trial matches. After applying site capability
and trial lookback filters, TIME nurses screened 35,912 of these matches with 5,034 confirmed.
ABSTRACT
These matches led to 186 activations (82 JIT, 104 prospective) and 573 consents. Conclusions: The
Tempus AI TIME program facilitated the screening of 1.28M+ patients for over 95 clinical trials,
averaging 1.57 consents per day over 1 year. Future trial matching strategies should utilize al-
WITHDRAWN
gorithmic screening and rapid activation processes to improve patient access and trial success.
Research Sponsor: Tempus AI, Inc.
2024 patient screening and consents.
Patient Population 1,281,676
TIME Trials 98
TApp Searches 1,323,259,353
Algorithmic Matches 2,251,505
Matches Screened 35,912
Matches Confirmed 5,034
Interventional Consents 225
Observational Consents 348
Total Activations JIT: 82, Prospective: 104
Avg Activation Time (business days) JIT: 16.1, Prospective: 39.6

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
74s CARE DELIVERY/MODELS OF CARE

1508 Oral Abstract Session 1509 Clinical Science Symposium

Effect of human-AI teams on oncology prescreening: Final analysis of a Disparities in pediatric oncology outcomes in the occupied Palestinian
randomized trial. First Author: Ravi Bharat Parikh, Emory University, Atlanta, GA territories: A retrospective study from Augusta Victoria Hospital. First Au-
Background: Eligibility assessment for oncology clinical trials – “prescreening” – relies on manual review of thor: Ru’a Rimawi, Dana-Farber Cancer Institute, Boston, MA
unstructured clinical notes, which is error-prone and time-consuming. Artificial intelligence (AI) language Background: Childhood cancer survival rates exceed 80% in high-income countries, but
models that merge deep learning with oncologist-derived rules (neurosymbolic AI) can enhance prescreening
by automating eligibility extraction from longitudinal electronic health records (EHRs), yet real-world over 80% of the global burden occurs in low- and middle-income countries, where
evaluations are limited. We compared the accuracy and efficiency of traditional vs. AI-augmented survival rates are significantly lower. The occupied Palestinian territories (OPT)—
(Human+AI) prescreening. Methods: In this randomized non-inferiority trial, two research coordinators comprising the West Bank, Gaza, and East Jerusalem—face additional and location-
(RCs) abstracted 12 common trial eligibility criteria from complete EHRs from patients with advanced non- specific challenges, including political instability, movement restrictions, and frag-
small cell lung cancer (NSCLC) or colorectal cancer (CrC) treated in a community oncology practice. Before mented healthcare, that would be expected to further negatively impact care. The aim of
the trial, gold-standard abstraction was performed by 3 independent oncologist reviewers. Charts were
randomized in blocks of 20 to be viewed alone (Human-alone) or augmented by a pretrained neurosymbolic our study is to report on outcomes of children treated at the only specialized pediatric
model (Human+AI) in a paired design, such that each RC reviewed each patient chart. The primary aim was to oncology cancer center, Augusta Victoria Hospital (AVH), located in East Jerusalem.
evaluate noninferiority (margin 65%) and subsequent superiority of chart-level accuracy (proportion of Methods: This was a retrospective IRB approved study conducted by Dana-Farber
correctly abstracted elements per chart relative to gold standard) between Human+AI vs. Human-alone. Cancer Institute and AVH. Chart review was performed to obtain diagnoses, treatments
Secondary outcomes were criterion-level accuracy (proportion of correctly abstracted elements across and outcomes of all pediatric oncology patients with histologically confirmed cancer
charts for each eligibility criterion), and efficiency (median abstraction time per chart). Paired t-tests and
Wilcoxon rank-sum tests assessed differences between Human+AI vs. Human-alone. We descriptively
admitted at AVH from January 2018 to June 2024. Results: A total of 424 patients were
compared accuracy of both arms vs. the AI algorithm (AI-alone). Results: Among 356 charts (196 NSCLC, included, with a median age at diagnosis of 6.95 years (IQR, 3.35–11.2). Of these, 51.2%
160 CrC), Human+AI had noninferior and superior accuracy than Human-alone (76.1% vs. 71.5%, p , 0.001); were male. Patients resided in the West Bank (51.2%), Gaza (43.4%), and East Jerusalem
both Human arms were superior to AI-alone (59.9%). Human+AI had greatest criterion-level accuracy for 7 of (3.8%). The median diagnostic delay, defined as date of symptom onset to date of
12 criteria. Efficiency was similar between Human arms (32.1 vs. 31.8 min, p = 0.51). Conclusions: AI- diagnosis, was 24.5 days (IQR, 10-45), with significant variation by gender (males:
augmented prescreening was more accurate than RC or AI prescreening alone. Human+AI teaming most
improved accuracy for biomarker, staging, and response criteria. While Human+AI did not save time, ef-
30 days; females: 20 days, p = 0.018), age group (,5 years: 14 days; $5 years: 30 days, p
ficiency gains may be realized as RCs become more familiar with AI eligibility models. AI language models = 0.003), and oncology diagnosis (leukemia: 14 days, lymphoma and solid tumors:
can enhance CRC prescreening and identification of trial-eligible patients. Clinical trial information: 30 days, p = 0.014). Treatment delays, defined as the time form diagnosis to treatment
NCT06561217. Research Sponsor: Mendel AI. initiation, was 14 days (IQR, 4-30), with the shortest duration for leukemia (2 days) and
Accuracy across arms. the longest for solid tumors (30 days, p = 0.008). There were no significant differences in
Accuracy (%) time to treatment by region. The majority of patients lost to follow-up were living in Gaza
Criteria Human- Alone Human+AI AI-Alone p-value (15/21, 71%). Among the 13% deaths, 15.4% were treatment-related, primarily due to
Overall 71.5 76.1 59.9 ,0.001 infection, and most of these treatment-related deaths (65%) occurred in patients living in
Neoplasm Cancer Type 86.9 86.4 73.3 0.80 Gaza. The 3-year overall survival (OS) rate was 76.31%, and event-free survival (EFS) rate
Stage Group 71.7 73.4 57.0 0.57
M Stage 43.9 57.0* 60.2 ,0.001 was 62.36%. Gaza patients had the lowest 3-year EFS rate (37.12%) compared to the
N Stage 50.5 66.3* 52.6 ,0.001 West Bank (70.37%) and East Jerusalem (77.78%; p , 0.0001). Conclusions: This study
T Stage 56.3 71.6* 54.3 ,0.001
Biomarker Biomarker Tested? 84.6 93.2** 88.1 ,0.001
is the first to report on outcomes of pediatric oncology patients treated at the only
Biomarker Result 67.9 79.0* 32.5 ,0.001 specialized center in the OPT. Males and patients older than 5 years experienced longer
Biomarker Result Interpretation 80.8 91.3* 35.7 ,0.001
Other Outcome 23.7 35.9* 55.2 0.004
diagnostic delays, while patients with solid tumors faced treatment delays 15 times
Response 47.1 51.7 60.4 0.20 longer than those with leukemia. Gaza patients had higher lost to follow-up rates and
ECOG 84.7** 78.1 34.4 0.10
Medications 89.0 89.1 59.4 0.92
treatment-related deaths with significantly inferior EFS. Research Sponsor: None.
Bold indicates arm with greatest accuracy for a given criterion. *>10% accuracy difference between Human1AI vs Human-
alone. **5-10% accuracy difference between Human1AI vs Human-alone.

1510 Clinical Science Symposium 1511 Clinical Science Symposium

Analysis of evidence in NCCN harmonized guidelines for sub-Saharan Integrated health system for resolving breast cancer screening actions:
Africa. First Author: Scott Swartz, Department of Medicine, University of California, Multicenter randomized clinical study—Itaberaı́ randomized trial, ReBEC,
San Francisco, San Francisco, CA RBR-39vm2nd. First Author: Ruffo Freitas-Junior, CORA – Advanced Center for
Background: The National Comprehensive Cancer Network (NCCN) Harmonized Guidelines for Sub- Diagnosis of Breast Diseases Federal University of Goias, Goiania, Goias, Brazil
Saharan Africa (SSA) have emerged as leading cancer treatment guidelines in SSA. The NCCN-SSA Background: The ITABERAÍ Project involves an intervention through training of Community
guidelines, derived by adapting NCCN guidelines for the SSA context, offer standardized recommendations
Health Workers (CHW), based on evidence from clinical breast examinations (CBE)
to guide cancer care and shape policy in SSA. This study examines the evidence cited in support of the
NCCN-SSA guidelines, with a focus on population characteristics and generalizability to SSA. screening. It is a randomized, prospective, phase III, multicenter clinical study. The target
Methods: Two reviewers independently examined the NCCN-SSA guidelines for the eight most common population is divided into a Control Group (CG) and an Intervention Group (IG), where the CG
cancers in SSA to identify all studies cited in support of treatment/management recommendations. Study receives the Brazilian Ministry of Health’s (MS) recommendations for breast cancer
selection discrepancies were resolved by discussion. Full-text articles were reviewed and data on age, sex, screening, The IG, in addition to the MS recommendations, receives the CBE. Among the
race, and recruitment geography were abstracted. Descriptive analyses were performed using R statistical stages of the project are the training of CHW and the development of tools for data
software. Results: Overall, 4,589 citations were reviewed, and 2,938 (64.0%) studies with individual-level collection. Objective: This study aims to evaluate the functioning of the integrated system
data were included, representing .10.4 million study participants. Of the 2,061 studies reporting geo-
as a tool for the resolution of breast cancer screening actions, according to the ITABERAÍ
graphic information (70%), 50 (2.4%) recruited in SSA; of these, 39 (83.0%) recruited from South Africa.
Three studies (0.2%) recruited exclusively in SSA. Most studies (95.3%) recruited exclusively from high or Project. Methods: Information stored in the integrated system database developed for the
upper-middle income countries. Only 29.2% of studies with race data included .10% Black study par- project, from 2022 to 2024, was analyzed. The system comprises a set of services and
ticipants. Conclusions: Most studies cited in the NCCN-SSA guidelines were conducted in high-income applications that integrates actions from the registration of participants to the diagnosis
countries outside SSA. A small minority of all study participants were Black. Our findings underscore and treatment of altered cases. It involves the integration of the App Rosa with the Web
potential limitations in the generalizability of the NCCN-SSA guidelines to SSA and highlight a pressing need System (RosaWatch). The App Rosa was created for exclusive use by the CHW to collect
to generate and incorporate context-specific data to guide care and inform policy. Research Sponsor: UCSF data from participants who agree to participate in the study, while RosaWatch is for use by
Helen Diller Family Comprehensive Cancer Center. the Family Health Team (FHT), specialist doctors, nurse navigators, and researchers, and
WHO region of recruitment, race, and country-level income of studies included in NCCN-SSA guidelines. was created to collect information on the follow-up of altered cases. To evaluate the
# (%) studies functioning of the system, data on the completeness of the information stored according to
representing
# (%) studies # (%) studies by # (%) studies by # (%) studies by re- % partici- % partici- % partici- only HMIC the Study Group and the follow-up of cancer cases were checked. Results: At the end of 3
reporting recruitment re- recruitment re- cruitment region*: pants, pants, pants, (per World
Cancer type geography gion*: Americas gion*: Europe Sub-Sah. Africa White Black Other race Bank) years of project implementation, data from 3,670 randomized women were reported, of
Overall 2061 1177 (57%) 921 (45%) 50 (2%) 79% 9% 11% 1964 which 92% (3,359) were active in the project, out of this 1,780 (53%) in the CG and 1,579
(70%) (95%) (47%) in the IG (p , 0.05). Regarding the completeness of the information, by the end of
B-cell 325 173 (53%) 178 (55%) 5 82% 7% 10% 307
(57%) (2%) (94%) 2024, it was found that 2,984 (88.8%) had consistent data. Stratifying by year of data
Breast 206 123 (60%) 124 (60%) 12 (6%) 80% 9% 11% 179
(44%) (87%)
collection, a significant increase in completeness was observed: in 2022, of the 425 records,
Cervical 163 100 (61 %) 44 3 64% 14% 22% 158 278 (65%) were compliant; in 2023, of the 2,255 records, 2,117 (94%) were compliant; and in
(90%) (27%) (2%) (97%)
Colorectal 584 278 (48%) 292 (50%) 11 (2%) 80% 10% 10% 571 2024, of the 679 records, 654 (96%) were compliant. During the study period, 3,143 breast
(82%) (98%) exams were performed by the CHW of the IG, with 594 (19%) altered cases. Of these, after
Hepato- 183 80 68 0 69% 9% 22% 169
cellular (74%) (44%) (37%) (0%) (92 %) screening at the Family Health Units (ESF), 90 (15%) participants received care from a
Kaposi 51 25 (49%) 19 5 38% 51% 10% 44
sarcoma (80%) (37%) (10%) (86%) specialist doctor (mastologist), and 10 had a confirmed breast cancer. In the CG, ESF
Ovarian 115
(56%)
76 (66%) 35 (30%) 0
(0%)
83% 4% 13% 115
(100%)
doctors referred 33 women for specialist consultations, and six diagnoses were confirmed.
Prostate 434 322 (74 %) 161 (37%) 14 (3%) 79% 9% 12% 421 Conclusion: The integrated system proved to be effective in monitoring the actions of the
(88 %) (97%)
ITABERAÍ Project. Additionally, it contributed to the adherence of the CHW and the quality
HMIC = high and upper-middle income countries. standard of the information, facilitating better data monitoring. Clinical trial information:
*Percent of studies includes as the denominator only the number of studies with reported geographies.
RBR-39vm2nd. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 75s

LBA1512 Rapid Oral Abstract Session 1513 Rapid Oral Abstract Session
Anesthesia type during surgery for treatment of biologically aggressive General practitioner-led vs surgeon-led colon cancer survivorship care: A
cancers: Results of the GA-CARES randomized, multicenter trial. First Au- randomized clinical trial. First Author: Julien Vos, Amsterdam UMC, Amsterdam,
thor: Elliott Bennett Guerrero, Stony Brook University School of Medicine, Stony Brook, Netherlands
NY Background: The role of general practitioners (GPs) in providing survivorship care for
cancer patients remains debated. In 2015, the randomized I CARE trial was initiated to
evaluate the impact of GP-led vs. surgeon-led survivorship care on quality of life (QoL) and
assess the effect of the eHealth application Oncokompas. An interim analysis after
12 months revealed no clinically relevant differences in QoL changes. However, patients
continued to have follow-up consultations for up to 60 months after treatment. This study
addressed the long-term QoL outcomes of the trial. Methods: The I CARE trial was a
pragmatic, 2x2 factorial, open-label, randomized controlled trial. The trial was conducted in
8 hospitals and 225 general practices across the Netherlands. The trial included patients
who underwent primary surgical treatment for stage I-III colon cancer or rectosigmoid
carcinoma, and who were eligible for routine follow-up according to national guidelines.
Inclusion lasted from March 26, 2015, to Nov 21, 2018. Patients were randomized using
variable block randomization, stratified by age and tumor stage, into four groups ([Link]):
usual surgeon-led care, surgeon-led care with Oncokompas, GP-led care, and GP-led care
The full, final text of this abstract will be available at with Oncokompas. The primary outcome was QoL at 5 years, as measured by the change
[Link] on the day of presentation and in the from baseline in the EORTC QLQ-C30 summary score (range 0-100). Generic and disease-
online supplement to the June 10, 2025, issue of the Journal specific QoL were measured at baseline, 3, 6, and 12 months, and annually up to 60 months
post-treatment. Differences in QoL changes were analyzed using piecewise linear mixed-
of Clinical Oncology. effects models with a knot at 24 months to capture potential deviations in QoL recovery. A
10-point difference was considered clinically relevant (superiority design with a = 0.05,
power of 80%, and 15% dropout). The trial is registered with the Netherlands Trial Register
(NTR4860). Results: In total 303 patients were enrolled; 79 were randomized to surgeon-
led care, 83 to surgeon-led care with Oncokompas, 73 to GP-led care, and 68 to GP-led care
with Oncokompas. Patients were male (67%) with a mean age of 68.0 years (SD 8.4). Of the
151 patients assigned to Oncokompas, 51 (36%) reported using the app at least once in the
first year. Baseline QoL was high in all groups. No clinically meaningful differences in QoL
were observed between GP-led and surgeon-led groups at 24 months (difference of -0.5
[95% CI -1.6 to 0.5]) and 60 months (-0.01 [-0.8 to 0.8]). Oncokompas also had no
meaningful effect (difference of 0.8 [0.0 to 1.6] at 60 months). Conclusions: In this
pragmatic, randomized controlled trial conducted in the Netherlands, GP-led survivorship
care did not improve long-term QoL compared to traditional surgeon-led care among non-
metastatic colorectal cancer survivors. Due to low usage rates, the impact of Oncokompas
is inconclusive. Survivorship care models can be tailored to fit individual preferences.
Clinical trial information: NTR4860. Research Sponsor: Dutch Cancer Society; grant BMA
5954.

1514 Rapid Oral Abstract Session 1515 Rapid Oral Abstract Session
Effect of post-discharge symptom monitoring on hospital readmissions: A SNF-CLIMEDIN: A HECOG prospective randomized trial of digital support
randomized trial. First Author: Robert Michael Daly, Memorial Sloan Kettering Cancer and intervention in patients with advanced non-small cell lung cancer
Center, New York, NY (NSCLC)—Final results. First Author: Paris A. Kosmidis, Department of Medical
Background: There is growing interest to improve patient care transitions from hospital Oncology, Hygeia Hospital, Athens, Greece
to home and to prevent readmissions but effective interventions are lacking. Background: This trial aims to investigate the feasibility and effectiveness of online digital
Methods: We conducted a randomized clinical trial among patients with cancer dis- intervention to NSCLC patients in terms of adverse events (AEs), quality of life (QoL), cost,
charged after an unplanned hospital admission at a specialty cancer center. Hospitalized and the interrelation with clinical and molecular characteristics. Methods: This prospective
patients on medical oncology services were randomized at discharge to receive either a randomized trial recruited 200 advanced NSCLC patients (3/22-10/23). Final analysis was
digital symptom monitoring and management intervention or to usual care. Patients undertaken in 12/24. All had NGS tissue analysis for 161 genes, and received standard
randomized to the intervention received a daily electronic symptom assessment for treatment (predominantly immuno-chemotherapy). Through the CareAcross online plat-
10 days post-discharge, consisting of 9 common symptoms from the National Cancer form, they received information about their disease and treatment, and periodically reported
Institute Patient Reported Outcomes version of the Common Terminology Criteria for any of the 22 preplanned AEs. Patients were randomized 1:1 in the Intervention (A) and
Adverse Events and an open-ended question to allow for patients to provide further Control (B) arm; patients in arm A received digitally, additionally, evidence-based guidance
symptom context and for two-way engagement with their clinical team. Patients also for the reported AEs. The study was designed to assess AE improvement (measured per
received customized self-management education delivered through the patient portal patient as reduction of AEs reported at last contact, compared to those previously reported)
based on their reported symptoms. Primary oncologists received alerts through the and QoL. EQ5D-5L scores were collected. Patient-case level hospitalization data were
collected and costs were estimated based on reimbursed cost as defined by the Ministry of
portal for moderate and severe symptoms. Usual care consisted of symptom monitoring
Health. Results were correlated with patients’ clinical and molecular characteristics.
at the discretion of the primary clinical team, generally comprising an oncologist and an
Results: Clinical and molecular characteristics will be presented during ASCO Congress.
office practice nurse. The primary outcome was the 30-day readmission rate, analyzed Comparing arms A vs B: ORR: 42.1% vs 41.7%; Median PFS: 11m (8.0-15) vs 10m (7.0-13), 1-
using cumulative incidence functions and Gray’s test with death as a competing risk. year PFS: 43% (31%-54%) vs 42% (31%-53%) (p = 0.4). Median OS: 15m (12-20) vs 16m (12-
Secondary endpoints included 90-day readmission rate and 30-day emergency room 21), 1-year OS: 59% (48%-68%) for both arms (p = 0.9). PFS and OS were improved for those
visit without admission rate. Results: Between 04/19 and 09/19/2024, 1,713 patients with best responses (p , 0.001). Patients with EGFR mutations had better OS (p = 0.05). The
were randomized with median age 66 years (range: 19 – 99), 66% white, 12% African most common AEs reported in both arms were fatigue, cough, anorexia, nausea. More AEs
American, 11% Asian, and 11% Hispanic with 53% female. The most common cancer were reported online vs to clinicians (89% vs 68% of patients; p , 0.01). Baseline EQ5D-5L
diagnoses were gastrointestinal (26%), thoracic (11%), genitourinary (10%), gynecologic was similar for both arms; when compared with data at best response, Anxiety/Depression
(10%), leukemia (9%), and lymphoma (9%). In the intervention group, the most frequently showed the biggest difference in improvement for arm A vs B. Among the 22 AEs, 17
reported moderate and severe symptoms were fatigue and pain. The two arms had improved more in arm A, 1 improved equally, and 4 improved more in Arm B. The com-
roughly similar proportions of patients who died before a hospital readmission. The 30- parative improvements of rash and stomatitis in arm A vs B were statistically significant (p =
day readmission rate was 30% in the intervention group compared to 37% in the usual 0.0073 & p = 0.0447). The mean hospitalization cost (arm A vs B, in Euros) was 455.4 (95%CI:
care group (p = 0.001). The decrease in readmission rate was maintained at 90 days (45% 91.9-941.5) vs 779.5 (346.6-1328.5) (p , 0.001); the mean diagnostics cost was 20.3 (0.5-
vs. 52%, p = 0.002). Emergency room visits without admission at 30 days were also lower 50.8) vs 73.3 (1.3-186.1) (p , 0.001). Conclusions: Digital oncology is feasible, cost-
in the intervention group (12% vs. 17%, p = 0.007). Conclusions: Digital post-discharge effective by reducing hospitalizations and tends to improve QoL (especially anxiety and
symptom monitoring and customized patient self-management education for 10 days depression) and most AEs of NSCLC patients regardless of clinical and molecular status.
post discharge reduced hospital readmissions in patients with cancer. Further research Patients report, digitally, more informative AEs for clinical and research analysis. Through
is necessary to identify the precise mechanisms that contribute to the success of this the digital transformation of healthcare, digital oncology can be a complementary tool to the
intervention. Research Sponsor: National Cancer Institute; Emerson Collective Digital Oncology team and warrants further exploration. Clinical trial information: NCT05372081.
Oncology Care. Research Sponsor: Stavros Niarchos Foundation.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
76s CARE DELIVERY/MODELS OF CARE

1516 Rapid Oral Abstract Session 1517 Rapid Oral Abstract Session
Breast cancer diagnosis, management, and outcomes in transgender, non- Inclusion of people living with HIV in Food and Drug Administration (FDA)
binary, and gender-diverse individuals: A multicenter cohort. First Author: oncology pivotal registration trials from 2020 to 2024. First Author:
Chandler Scott Cortina, Division of Surgical Oncology, Department of Surgery, Medical Alberto Giovanni Leone, Department of Oncology and National Centre for HIV Malig-
College of Wisconsin & MCW Cancer Center, Milwaukee, WI nancy, Chelsea and Westminster Hospital, London, United Kingdom
Background: Paucity of data on breast cancer (BC) in transgender, nonbinary, and gender- Background: People living with HIV (PLWH) have an increased risk of developing cancers compared
diverse (TGD) individuals leads to suboptimal screening and treatment algorithms. We to the population without HIV, with cancer being the leading cause of death for this population in high-
developed a national multicenter retrospective cohort to describe demographic, clinico- income countries. Previous research by the ASCO-led HIV working group found only 11% of clinical
pathologic, and treatment characteristics of TGD individuals with BC and report outcomes. trials supporting FDA cancer therapy approvals from 2010-2014 allowed for inclusion of PLWH, leading
Methods: The cohort included TGD persons age $18yrs with stage 0–IV BC treated at 22 US to clinical uncertainty in the efficacy-safety profile of new cancer treatments in this group. To address
this gap, ASCO (2017) and FDA (2020) issued guidelines encouraging inclusion of PLWH in cancer
centers from 1990–2023. Demographic and clinicopathologic characteristics were evaluated
clinical trials. To explore their impact, we examined inclusion of PLWH in pivotal cancer trials post-
and compared to BC patients in the SEER 2016–21 dataset. Wilcoxon rank sum tests, x2 tests, guidelines release. Methods: We reviewed all new FDA-approved indications of the past five years
and KM analysis were used to compare variables and estimate 5-yr BC-specific survival (Jan/2020-Nov/2024). For each new approval, two authors independently assessed the inclusion/
(BCSS). Results: 112 TGD persons with 113 BCs were included. Median age at diagnosis was exclusion criteria outlined in the primary protocol of each pivotal trial. We analyzed data on the FDA
42.5yrs (IQR 36.5–51), 92.9% were female sex at birth (FSAB), 73.2% were NH-White, and label, cancer type, therapeutic modality, inclusion/exclusion criteria, sponsor, and the protocol
38.4% used gender-affirming hormones pre-BC. Of those FSAB (n=104), 61.5% were pre- (version 1) publication date. Results: We identified 244 new therapy indications, based on supporting
menopausal and 11.5% had undergone gender-affirming top surgery (GATS) pre-BC. Most BCs data from 259 pivotal clinical trials. 27% of trials permitted inclusion of PLWH. Pivotal trials for
(51.8%) were self-detected, 27.7% were screen-detected (48.2% underwent screening pre- hematological cancers, compared to solid cancers, were significantly more likely to exclude PLWH
BC), 13.4% were incidentally found on GATS pathology, 3.6% were provider-detected, and [unadjusted Odds Ratio (OR) 3.15, 95% confidence interval (CI): 1.51-6.56; p=0.0012]. Pivotal trials of
2.7% were incidentally found on other imaging. Of 84 (75%) tested patients, 16/84 (19%) had a immunomodulatory agents were significantly more likely (OR 3.87, 95% CI: 1.91–7.83; p,0.0001) to
pathogenic germline variant, with BRCA2 (25%) and BRCA1 (18.8%) being most common. exclude PLWH compared to other cancer therapies. The inclusion rate was 10.3% for AIDS-defining
Most tumors were HR+ (85.7%) and early stage (25.7% DCIS and 45.1% stage I). Regarding cancers and 29.8% for non-AIDS-defining cancers. Trials funded only by industry were significantly
more likely (OR 2.80, 95% CI: 1.36-5.77; p=0.0078) to exclude PLWH, compared to non-industry funded
local treatment, most (61.6%) underwent mastectomy, with 63.8% omitting reconstruction;
trials. Inclusion rate of PLWH was higher in protocols published after 2020 (39.1%) compared to those
after lumpectomy, 29% omitted radiation (RT). 41.1% received systemic chemotherapy and
before (26.3%). Conclusions: Our analysis indicates an improvement in the inclusion of PLWH in
while endocrine therapy (ET) was recommended for 79, only 81% (64/79) received ET. There oncology pivotal trials following ASCO and FDA guidance. However, nearly three out of four pivotal
was no difference in surgery type (p = 0.22) or ET receipt (p = 0.32) by SAB. Compared to cancer trials continue to exclude PLWH. This highlights an unmet need, resulting in uncertainty for
patients in SEER (N = 401,311), the TGD cohort was younger (median age 42.5 vs 62yrs), more healthcare professionals regarding the safety and clinical utility of new cancer treatments in PLWH.
frequently NH-White (75.2% vs 65.1%), more often had PR+ disease (79.5% vs 70.7%), and Additional strategies must be considered to address this disparity. Research Sponsor: None.
had a higher proportion of males (MSAB) (7.1% vs 0.8%) (p ,0.01 for all); but there was no
Rates of PLWH inclusion in oncology pivotal trials.
difference in disease stage (p = 0.39). At 38 months median follow-up: 12 (10.7%) had a
Include PLWH (%)
locoregional recurrence (LRR), and 2 died of metastatic BC. 5-yr BCSS probability was 96.2%
(95% CI 85.3–99.0%). Conclusions: The first multicenter cohort study of TGD individuals with Total (n=259) 27.4
BC identified they were younger and had a higher proportion MSAB compared to BC patients in Solid Cancers (185) 33
Haematological malignancies (74) 13.5
SEER. Most tumors were self-detected, and the pathogenic germline variant rate was high – Immunomodulatory agents (89) 12.4
suggesting a possible role for earlier screening in high-risk TGD persons regardless of SAB. Other agents*(170) 35.3
The elevated LRR rate along with low ET and RT uptake indicates opportunities to improve Industry-funded (223) 24.2
adherence to guideline concordant care. Findings underscore the necessity for prospective Non-industry funded (36) 47.2
AIDS-Defining Cancers (30) 10
research to inform gender-inclusive evidence-based BC screening and treatment guidelines. Non-AIDS Defining Cancers (45) 29.8
Research Sponsor: National Cancer Institute.
*Chemotherapy, targeted therapy, drug-antibody conjugates, hormone therapy, radionuclide therapy.

1518 Rapid Oral Abstract Session 1519 Rapid Oral Abstract Session
Economic modelling to inform pricing for LMICs of immune checkpoint Association of court-documented major adverse financial events before
inhibitors in advanced PD-L1-high non-small cell lung cancer. First Author: cancer diagnosis and mortality risk in the US. First Author: Robin Yabroff,
Giulia Segafredo, Medicines Patent Pool, Genève, Switzerland Surveillance and Health Equity Science, American Cancer Society, Atlanta, GA
Background: Lung cancer is the most common cancer and cause of cancer death, encompassing Background: Cancer diagnosis is associated with increased risk of financial hardship in
for 16.8% of all cancer-related deaths worldwide. Anti-PD(L)1 Immune Checkpoint Inhibitors (ICIs) the US. This study examined the associations of court-documented major adverse
as monotherapy currently represent the standard of care (SoC) for advanced Non-Small Cell Lung financial events (AFEs) of bankruptcies, liens, and evictions prior to cancer diagnosis
Cancer with high ($50%) PD-L1 expression in high-income countries. Despite their efficacy, ICIs and risks of all-cause and cancer-specific mortality. Methods: Individuals aged 21 to 69
remain largely inaccessible in low- and middle-income countries (LMICs), with affordability
years diagnosed with common cancer types, including bladder, female breast, colorectal,
being a significant barrier. Providing evidence on cost-effective (CE) price ranges for ICIs in LMICs
is critical for global health policymakers to devise strategies to enhance access. Methods: A
kidney, lung and bronchus, oral cavity/pharynx, or prostate cancers or melanoma during
partitioned-survival model was used to estimate CE price targets for three ICIs (atezolizumab, 2014-2015 were identified from the SEER population-based registries for Seattle,
cemiplimab, and pembrolizumab) as single agents compared to platinum-based combination Louisiana, and Georgia. Registry data were linked with LexisNexis consumer data to
chemotherapy (current SoC in several LMICs). Treatment duration was assumed of up to 35 cycles identify history of court-documented AFEs of bankruptcies, liens, and evictions. Vital
or until disease progression. Cost-effectiveness thresholds were set at 1, 2, and 3 times the gross status and cause of death were examined through December 31, 2021. The association
domestic product (GDP) per capita per quality-adjusted life year (QALY) gained. Case studies were of pre-diagnosis AFEs and risk all-cause and cancer-specific mortality was assessed
modelled in two LMICs (India and South Africa – not all ICIs were registered in both countries) to with separate multivariable Cox proportional hazards models for each survival outcome,
determine the maximum price at which ICIs would be CE from the perspective of publicly-funded stratified by cancer site. Models were adjusted for stage, age, race and ethnicity, marital
health systems. Primary efficacy data were sourced from phase III clinical trials (KEYNOTE-024, status, registry, registry-specific income categories, and the interaction between income
IMpower110, and EMPOWER-Lung 1), and country-specific data were collected through interviews and registry. Results: Of 58,796 individuals diagnosed with one of the 8 selected
with key technical stakeholders. Values were reported in USD for 2023. Results: The analysis cancers, 21,694 (36.9%) had a pre-diagnosis AFE and there were 16,714 deaths (28.4%)
determined that the maximum acquisition costs for ICIs to be cost-effective at 1-, 2-, and 3-times
GDP per capita in India and South Africa, range from $14.20 to $648.00 per cycle per patient.
during the study period between 2014 and 2021. Pre-diagnosis AFEs were associated
Current reference prices would require discounts of up to 93.3% to meet the 3 GDP threshold. with higher risk of all-cause mortality for individuals diagnosed with female breast
Dose-optimization strategies such as low-dose and vial sharing were identified as feasible and (hazard ratio (HR): 1.18; 95% confidence interval (CI): 1.09-1.28), colorectal (HR: 1.14;
evidence-based approaches to achieve partial price reduction (sensitivity analysis will be pro- 95% CI: 1.06-1.23), oral cavity/pharynx (HR: 1.14; 95% CI: 1.06-1.23 ) and prostate (HR:
vided). Conclusions: To make ICIs cost-effective in LMICs, significant discounts from current 1.33; 95%CI: 1.20-1.47) cancer and early- and late-stage melanoma (HR: 2.23; 95% CI:
reference prices are needed. Similar price reductions (up 93%) have been achieved for other 1.89-2.99 and HR:1.34; 95% CI:1.01-1.80, respectively), in adjusted models. Pre-
monoclonal antibodies, such as trastuzumab, in India and South Africa, also driven by the diagnosis AFEs were also associated with significantly higher risk of cancer-specific
availability and uptake of quality-assured biosimilars. A comprehensive approach, combining mortality for these five cancers. Conclusions: Court-documented AFEs of pre-diagnosis
accelerated biosimilar availability, also leveraging voluntary licensing and technology transfer, bankruptcy, lien, or eviction was associated with increased risk of all-cause and cancer-
with dose and treatment-duration optimization strategies could help achieve target price levels specific mortality for multiple cancer types in this study using a novel SEER cancer
and improve accessibility. Research Sponsor: Medicines Patent Pool. registry-LexisNexis consumer data linkage. The association of pre-diagnosis AFEs and
Country WTP threshold (xGDP) Pembrolizumab Atezolizumab Cemiplimab mortality risk underscores lasting adverse consequences of patient financial vulnera-
India 1 $72.6 $74.2 $66.5 bility prior to incurring high out-of-pocket costs of cancer treatment. Our findings are
2 $199.7 $166.9 $181.5 especially timely, with growing efforts by health care providers to screen and address
3 $308.6 $259.6 $300.0 patient health-related social needs as part of comprehensive oncology care. Research
South Africa 1 $51.5 $53.1 $14.2
2 $349.9 $289.0 $314.7
Sponsor: None.
3 $648.2 $525.0 $615.2

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 77s

1520 Rapid Oral Abstract Session 1521 Poster Session

Effect of broad-based genomic sequencing on survival outcomes in ad- Financial toxicity and drivers of delayed care among cancer survivors across
vanced non-small cell lung cancer: A national cohort study, 2011-2023. First the lifespan. First Author: Justine Po, Keck School of Medicine of University of
Author: Patricia Mae Garcia Santos, Division of Health Services, Outcomes, and Policy, Southern California, Los Angeles, CA
Department of Radiation Oncology, Winship Cancer Institute, Emory University, Atlanta, Background: Financial toxicity has been increasingly recognized as a cause of poor
GA outcomes among cancer survivors. Cancer survivors have been shown to face greater
Background: Use of broad-based genomic sequencing (BGS) for advanced non-small cell lung healthcare access and affordability issues than the general population, and available
cancer (aNSCLC) is rising. While older studies have found no survival benefit with BGS, its impact evidence suggests that adolescent/young adult (AYA) cancer survivors may face espe-
on survival outcomes in the era of modern targeted therapy is unknown. Methods: In this cially high financial toxicity. However, existing research is limited by small sample sizes
retrospective cohort study, the 2011-2023 Flatiron Health Database—a nationally representative and methodological constraints. Understanding age-specific healthcare barriers could
database of electronic health records from . 280 US cancer clinics—was queried for patients with inform targeted interventions to improve outcomes for cancer survivors across the
Stage IIIB-IV NSCLC who received at least one line of systemic therapy with $12 months follow- lifespan. Methods: Participants aged 18 and older were enrolled with informed consent in
up. Primary exposure was receipt of BGS vs. “Focused” biomarker testing (i.e., ALK FISH, EGFR the All of Us Research Program, an NIH database integrating multiple health information
PCR) within 90 days of first- and second-line therapy start. To address baseline confounding, we
sources. Included participants were cancer survivors who had available data for age, sex
used 1:1 nearest-neighbor propensity score matching based on age at initial diagnosis, sex, self-
and other demographics. Cancer diagnoses were identified using ICD and SNOMED codes.
reported race/ethnicity, histology (squamous vs. non-squamous), insurance status, smoking
status, ECOG performance status, practice type (academic vs. community), stage at diagnosis, Survey data were used to assess outcomes of financial toxicity and drivers of delayed care,
advanced diagnosis year, and practice rate of BGS. Adjusted Cox proportional hazards models as well as covariates of race, ethnicity, income, education, marital and insurance status. As
compared median progression-free survival (mPFS) and median overall survival (mOS) between all participants were insured, insurance status was dropped from the final models. Age
groups. Sensitivity analyses adjusted for biomarker status and used an instrumental variable groups were coded as 18-39, 40-49, 50-64, 65-74 and 75+. Statistical analysis was
approach. Results: Our initial unmatched cohort consisted of 35,060 patients (BGS, n = 14,192; conducted using logistic regression by age group, adjusted for covariates. Results: 15,637
Focused, n = 20,868; 52% female, 3.5% Asian, 9.3% Black, 3.8% Hispanic, 79% community cancer survivors were included for analysis, including 1,090 participants aged 18-39 and
practice). In the propensity-matched first-line therapy cohort (BGS vs. Focused, n = 10,008 in each 2,104 participants over age 75. Compared to 18-39 year olds, odds of being unable to afford
group; all standardized mean differences , 0.1), BGS was associated with greater mPFS (6.4 vs. specialist care decreased with age: [40-49] OR = 0.57 (p = 0.01), [50-64] OR = 0.18 (p ,
6.0 months; adjusted HR [95%CI], 0.96 [0.92-1.0], p = 0.046) and mOS (16 vs. 14 months; 0.91 0.0001), [65-74] OR = 0.07 (p , 0.0001), [75+] OR = 0.006 (p , 0.0001). Results were
[0.86-0.95], p , 0.001). Sensitivity analyses were consistent with primary results. Patients similar for primary care, with more extreme effect sizes. Causes of delayed care differed by
receiving BGS had higher rates of ALK/EGFR positivity (18% vs. 13%) and receipt of targeted age group. 18-39 year olds were more likely to report elderly caregiving responsibilities and
therapy (20% vs. 17%). Upon adjustment for biomarker status, however, BGS remained associated inability to get time off work as drivers of delayed care, while 40-49 year olds were more
with improved OS (1.01 [0.88-0.98], p = 0.004) but not PFS (0.98 [0.94-1.03], p = 0.4). In the likely to report difficulties accessing child care as a driver (OR = 5.25, p , 0.0001). Older
propensity-matched second-line therapy cohort, no associations between BGS and survival cancer patients were also more likely than AYAs to report lack of transportation access
outcomes were observed. Conclusions: This is the first national analysis of survival outcomes in
as a cause of delayed care ([65-74] OR = 2.33, p , 0.0001). Conclusions: To date, this
aNSCLC to demonstrate a survival benefit with BGS. These findings support guideline en-
dorsement and payer coverage of BGS prior to 1st line therapy. Research Sponsor: None.
study represents the largest and most comprehensive analysis of healthcare barriers
among cancer survivors across the lifespan. Targeted interventions based on the most
Median (95%CI) Univariate HR (95%CI) Adjusted HR (95%CI) significant barriers by age group may more effectively improve healthcare access and
First Line PFS, months Focused 6.0 (5.8-6.1) — — outcomes for all ages. Our results support that AYA cancer survivors may benefit most
BGS 6.4 (6.2-6.5) 0.95 (0.91-0.99) 0.96 (0.92-1.00) from financial support, elderly care resources and medical notes to facilitate time off work.
OS Focused 14 (14-15) — — By comparison, expanding childcare support may be most important for increasing access
BGS 16 (16-17) 0.90 (0.86-0.95) 0.91 (0.86-0.95)
Second Line PFS Focused 3.8 (3.5-4.1) — — among those aged 40-49. Among older cancer survivors, ensuring reliable transportation
BGS 4.2 (4.0-4.5) 0.91 (0.82-1.01) 0.92 (0.83-1.02) presents the greatest opportunity for improving healthcare access. Research Sponsor:
OS Focused 13 (12-14) — — National Institutes of Health, Office of the Director.
BGS 12 (12-14) 1.02 (0.91-1.15) 1.01 (0.89-1.14)

1522 Poster Session 1523 Poster Session

Spillover effects of Medicaid expansion on insurance coverage, diagnosis, Bolstering access to clinical trials: Sociodemographic characteristics of
and survival among low-income elderly patients with cancer. First Author: patients enrolled in interventional clinical trials within a large, NCORP-
Kewei Sylvia Shi, American Cancer Society, Atlanta, GA designated community oncology practice setting. First Author: Meera Vimala
Background: Medicaid expansion under the Affordable Care Act is associated with in- Ragavan, Kaiser Permanente San Francisco Medical Center, San Francisco, CA
creased health insurance coverage and improved outcomes among patients with Background: Access to clinical trials is an important aspect of cancer care given rapidly
cancer , 65 years. Although not the target population, individuals $65 years may also changing treatment [Link] patients with cancer including those belonging to
benefit from Medicaid expansion through “welcome mat” effects, referring to the indirect racial/ethnic minoritized groups are treated in community oncology settings, but only a
increase in Medicaid enrollment from increased public awareness and streamlined en- minority ( , 5%) of these patients are enrolled on clinical trials. Barriers to trial en-
rollment procedures. This study examines the associations of Medicaid expansion with rollment in community oncology settings have been well described at the patient,
Medicaid coverage, stage at diagnosis, and survival among cancer patients $65 years. provider, and system levels. The National Cancer Institute Community Oncology Pro-
Methods: Using the National Cancer Database, we identified patients $65 years newly gram (NCORP) aims to address these barriers- and improve equity in trial enrollment- by
diagnosed with cancer between 2010-2022 residing in areas with median household providing support and infrastructure for community oncology practices to conduct
income below 200% of the federal poverty level. We applied a quasi-experimental clinical trials. Kaiser Permanente Northern California (KPNC) is a large integrated health
difference-in-differences design, with multivariable linear probability models to com- system comprising thirty cancer trial sites and is part of the Kaiser Permanente NCORP,
pare the changes in the percentage of dual-eligible or Medicaid-only coverage, stage at one of the largest NCORP-designated sites. In this study, we compared sociodemo-
diagnosis, and two-year survival post (vs. pre) Medicaid expansion in expansion states graphic characteristics of patients enrolled in interventional trials with the overall cancer
compared with non-expansion states. Results: A total of 1,468,116 patients with cancer population within KPNC. Methods: We evaluated all patients enrolled on interventional
were identified, with 885,671 patients from expansion states and 582,445 patients from cancer clinical trials within KPNC between 1/1/2015-12/31/2022. We abstracted de-
non-expansion states. After adjusting for sociodemographic characteristics, the per- mographic and clinical characteristics from the KPNC cancer registry. We compared
centage of patients with dual or Medicaid-only coverage increased from 10.27% to 11.33% characteristics to the incident cancer population diagnosed over the study period across
in expansion states and decreased from 9.4% to 8.11% in non-expansion states, resulting KPNC. We used Pearson chi squared tests (categorical), binomial tests (binary) and one-
in a net increase of 1.34 percentage points (ppt, 95% confidence interval [CI]: 1.12, 1.56) sample t tests (continuous) to compare sociodemographic characteristics. Results: We
associated with Medicaid expansion. Differences were more pronounced among patients identified 1,341 patients who were enrolled onto interventional clinical trials and
with stage III-IV cancers, females, non-Hispanic Black, metropolitan residents, and those compared them to 97,764 patients diagnosed with cancer over the study period. Patients
with $2 comorbidities. The percentage of early-stage (I/II) cancer diagnoses decreased enrolled on interventional trials were younger (mean age 60.2 vs 62.9 years, p , 0.001),
more in non-expansion states (49.82% to 47.19%) than in expansion states (47.63% to more likely to have Stage IV disease (22.7% vs 11.2%, p , 0.001), more likely to reside in
46.06%), resulting in a net increase of 0.96 ppt (95% CI: 0.58, 1.34). The protective effects high-socioeconomic status neighborhoods (27.7 vs 23.8%, p , 0.001), and less likely to
of Medicaid expansion were stronger for late-stage (III/IV) non-small cell lung and uterine speak a language other than English (4.7 vs 7.1%, p , 0.001). There were no differences
cancers, as well as early-stage thyroid and bladder cancers. Two-year overall survival in race, ethnicity or sex distributions between the trial population and overall population.
rates increased from 58.86% to 62.39% in expansion states and from 59.18% to 62.55% in Conclusions: Across a large NCORP-designated community oncology trials program,
non-expansion states, leading to a net increase of 0.95 ppt (95% CI: 0.61, 1.29). Im- the racial and ethnic makeup of patients enrolled on trials was similar to the broader
provements were most notable for prostate, lung, kidney, and bladder cancers. cancer population within KPNC. These findings suggest that the trial infrastructure
Conclusions: Medicaid expansion was associated with an increase in Medicaid coverage, provided by NCORP may surmount the structural barriers that drive low access to trials
early-stage cancer diagnoses, and improved two-year survival among patients diagnosed among racial/ethnic minorities. Small differences in enrollment based on age, language
with cancer $65 years. Findings underscore the spillover benefits of Medicaid expansion and socioeconomic factors persisted. Efforts to bolster clinical trial portfolios in
in supporting low-income elderly populations and the importance of indirect benefits community oncology settings may address existing barriers to enrollment. Research
when evaluating Medicaid expansion’s broader impact. Research Sponsor: None. Sponsor: KPNC Community Benefit Program- Cancer Section Pilot Funding.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
78s CARE DELIVERY/MODELS OF CARE

1524 Poster Session 1525 Poster Session

Integrating collaborative care in oncology: Improving quality of life and From food deserts to clinical trial deserts: Challenges in access to breast
mental health for patients with cancer. First Author: Kyle N. Lavin, University of cancer trials. First Author: Rachel Ann Sachs, Robert Wood Johnson Medical School,
North Carolina at Chapel Hill, Chapel Hill, NC New Brunswick, NJ
Background: In the U.S., 22 million patients with cancer face unmet behavioral health Background: Patients who live in food deserts have high mortality rates from breast
needs, contributing to up to $245B in preventable healthcare costs. These unique challenges cancer and stand to benefit from participating in research studies, yet they may face
stem from physical, emotional, practical, and relational stressors associated with their complex barriers to doing so. This study explores the relationship between living in a
cancer journey. Collaborative care, an evidenced based model that integrates behavioral food desert and breast cancer clinical trial enrollment and assesses the contribution of
health into medical settings, has been shown to improve access, outcomes, and reduce transportation barriers and distance. Methods: The national Vizient Clinical Database
overall healthcare costs in primary care. Cerula Care offers a virtual collaborative care model (which includes 98% of academic and 1,000 community hospitals) was queried for
that seamlessly integrates with oncology practices to improve quality of life and behavioral women treated for breast cancer between January 2022 and June 2024. Patients who
health outcomes. The team includes a Consulting Psychiatrist, Behavioral Health Care traveled . 4 hours to get to treatment were excluded (“destination care”). The outcome
Manager, and Behavioral Health Coach working in coordination with oncology teams. of interest was participation in a clinical trial. A “clinical trial desert” was defined as . 2
Methods: This study analyzed data from 127 patients with cancer enrolled in Cerula Care’s hour driving time to the nearest hospital enrolling patients in clinical trials. Multivariable
12-week virtual care program between 1/31/24 and 8/31/24. Behavioral health outcomes, analysis evaluated the association between living in a level 1 (most severe) USDA food
including anxiety (GAD-7), depressive symptoms (PHQ-9), and quality of life (FACT), were desert census tract, a clinical trial desert, a census tract with high transportation
assessed monthly. Analyses focused on baseline to two-month outcomes due to limited
vulnerability (households with limited access to a car or public transit), and breast
sample sizes beyond two months (many patients were still in earlier stages of the program).
cancer clinical trial participation. Interaction analysis was performed between food
Correlations and paired t-tests were conducted to evaluate changes in outcomes.
Results: Of the 127 patients enrolled, mean age was 57.1 years (range: 28–78), 83.5% were
desert and clinical trial desert status. Results: Of 1,317,269 patients, 103,790 (7.9%)
female, 52.8% identified as White, 33.9% Black or African American, 0.8% Asian, and 3.1% lived in a food desert, and 22,779 (1.7%) participated in a clinical trial. Patients living in a
other races. Breast cancer was the most common diagnosis (52.8%), followed by lung, food desert comprised 6.2% of patients enrolled in a trial vs 7.9% of patients not enrolled,
colorectal, pancreatic, and ovarian cancers. Behavioral health outcomes from baseline to p , .0001. 41.0% of patients living in a food desert also resided in a clinical trial desert.
month 2 are shown in the table. Conclusions: Within just two months of care, Cerula Care On multivariable analysis, living in a food desert was associated with decreased odds of
significantly improved depressive symptoms and anxiety, as well as quality of life for cancer trial participation (aOR 0.87, 95% CI 0.82-0.92, p , .0001), as was living in a clinical trial
patients. Notably, improvements in quality of life were correlated with minority status, desert (aOR 0.89, 95% CI 0.84-0.94, p , .0001), and living in the most vulnerable quartile
highlighting the program’s potential to reduce disparities in cancer care. By integrating for neighborhood transportation (aOR 0.89, 95% CI 0.85-0.93, p , .0001; ref: least
behavioral health into oncology treatment, this model demonstrates a scalable, impactful vulnerable). Medicaid insurance also decreased the odds of enrollment (aOR 0.84, 95%
solution to enhance patient outcomes. Future research should focus on long-term sus- CI 0.80-0.89, p , .0001; ref: private). Conversely, receiving care at an academic hospital
tainability, cost-effectiveness, and broader implementation across diverse oncology set- increased the odds of enrollment (aOR 2.98, 95% CI 2.62-3.15, p , .0001; ref community
tings. Research Sponsor: None. hospital). Living in both a food desert and a clinical trial desert decreased the odds of
Behavioral health outcomes from baseline to month 2.
clinical trial participation by 27% (aOR 0.73, 95% CI 0.70-0.76, p , .0001) compared to
19% if living in a food desert alone (aOR 0.81, 95% CI 0.78-0.84, p , .0001); p , .0001 for
Sample Statistical
Outcome Measure Size (n) Change Significance (p-value) interaction. Conclusions: Neighborhood transportation barriers, clinical trial deserts,
and food deserts all independently confer a similar lower likelihood of participation in a
Depressive Symptoms PHQ-9 Score 50 ↓ 4.47 points p < .001 clinical trial. Living in a clinical trial desert compounds the negative impact of living in a
Anxiety Levels GAD-7 Score 52 ↓ 2.06 points p = .007
Quality of Life FACT Score N/A Improved, especially in r = .439, p = .017 food desert alone, further taxing already disadvantaged populations. Interventions such
racial minority patients as patient navigation, food banks, and opening clinical trials near communities ex-
periencing food insecurity may mitigate these challenges. Research Sponsor: None.

1526 Poster Session 1527 Poster Session

Low-dose anti-PD-(L)1 inhibitor strategies: A systematic review. First Author: Characterizing transportation need and missed visits among patients re-
Pablo Jiménez Labaig, Head and Neck Unit, The Royal Marsden NHS Foundation Trust, ceiving radiotherapy. First Author: Kathleen Cui, University of California, San
London, United Kingdom Francisco, San Francisco, CA
Background: Immune checkpoint inhibitors (ICIs) targeting the PD-(L)1 pathway have revolutionized cancer Background: Patients with cancer receiving radiotherapy (RT) are vulnerable to treat-
therapy, but their high costs significantly limit accessibility, particularly in low- and middle-income countries ment interruptions, which affect oncologic outcomes. To prioritize program development
(LMICs). Low-dose regimens may offer a viable solution to this challenge. This systematic review analysed
study designs, dosing strategies, clinical outcomes, and potential cost savings of low-dose ICIs. Methods: A for improved RT access, we aimed to characterize transportation need via standardized
PRISMA/EQUATOR compliant systematic search of WebOfScience, Cochrane Central Register of Clinical Trials, department screening and its association with missed RT visits. Methods: We pro-
ASCO and ESMO conference databases was conducted until October 10th, 2024. Studies investigating reduced- spectively identified a cohort of 552 consecutive patients with cancer who received RT
dose anti-PD(L)1 from those FDA/EMA-approved regimens were included. Data were categorized by dose and at a single academic institution between September 2023 and March 2024, during which a
country income level. Results: From 1,751 records, 25 studies (4 clinical trials, 21 observational studies) quality improvement program for standardized transportation needs screening was
involving 1,793 participants met inclusion criteria, with 1,202 receiving low-dose ICIs. Two studies used non-
inferiority designs, and 21 evaluated participants across multiple treatment lines. The population had a median implemented. Missed RT visits were determined computationally and counted if a patient
age of 53.1 years (range 19–84), 26% female, 88% with advanced disease, and 9% ECOG $2. Most studies were did not arrive for a scheduled RT visit; these missed visits may have been due to pre-
conducted in Asia (81.4%, n = 1,459), with head and neck (22.8%, n = 409) and non-small cell lung cancers planned reasons, unplanned hospital admissions, or other logistical causes. Clinical data
(17.3%, n = 311) being most studied. 48% studies were from LMICs, 44% from high-income countries (HIC), and were extracted from the electronic medical record system. Univariable and multivariable
8% from upper middle-income countries (UMIC). India contributed the most (studies, k= 12, 839 participants). logistic regression analyses were used to determine associations between demographic
Nivo (k= 21), pembro (k= 6) and atezo (k= 1) were assessed. The most common regimens were Nivo40mg Q2W
(k= 7), Nivo20mg Q3W (k= 6) and Nivo20mg Q2W (k= 5) [Table 1]. A radiological response rate between 5-75% variables and transportation need. Fisher’s exact test was used to compare missed RT
was noted when low-dose ICI was used as monotherapy (k= 10). High variability in participant selection and visits between patients with varying social needs. Results: Median age was 66 years
interventions restricts further conclusions about efficacy and safety. The median projected savings were 83.3% (IQR: 53.75-74). Most patients were English-speaking (85.1%), male (54%), and white
(25–99.40%), with $70% savings in half of the studies. Conclusions: This review described the use of low- (51.8%). Common planned transportation modes were driving (76.2%) and public transit
dose anti-PD(L)1 drugs, especially in healthcare settings with limited resources, highlighting radiological (8%), as well as Veterans Affairs transportation, rideshare, and taxi. Of all patients, 26.4%
responses observed with monotherapy. Non-inferiority or near-equivalence randomized clinical trials will be
helpful in establishing their clinical validity. Research Sponsor: None. missed $1 RT visit and 19.9% reported transportation need. Overall, 39.1% versus 23.3%
of patients with versus without transportation need missed $1 RT visit (p = 0.001). Other
Tumor type
(origin from participants HNSCC NSCLC HCC RCC HL Multiple Melanoma Gastric/GEJ Gyne CCR Cervical Thymic social needs were found among 17.6% of patients, the most common being housing
assessed) n=409 n=311 n=93 n=73 n=70 n=57 n=56 n=42 n=35 n=30 n=20 n=6
(80.3%). Of those with transportation need, 45.5% had additional social needs. Frequency
LMIC 409 57 57 42 30 20
UMIC 23 6 of missed RT visits were similar between patients with sole transportation need (41.7%)
HIC
Number of studies (k) k: 6
311
k: 4
93
k: 2
16
k: 2
47
k: 3 k: 2
56
k: 1 k: 1
35
n: 1 n: 1 n: 1 n: 1
and those with additional social needs (39.1%) (p = 0.563). On univariable analysis, there
assessing each dose
per tumor
was increased transportation need among patients identifying as Asian (OR = 1.87, 95% CI
Nivo 0.3 mg/Kg Q2W
Nivo 10 mg Q2W 2 1
1 1.05-3.27, p = 0.030), Latinx (OR = 3.14, 95% CI 1.74-5.63, p , 0.001), unknown/declined
Nivo 10 mg Q8W 1 race/ethnicity (OR = 2.84, 95% CI 1.29-5.99, p = 0.007), non-English speaking (OR = 3.01,
Nivo 20 mg Q2W 1 2 1 1
Nivo 20 mg Q3W 3 1 1 95% CI 1.80-4.98, p , 0.001), with Medicaid insurance (OR = 3.61, 95% CI 1.94-6.71, p ,
Nivo 40 mg Q2W 1 3 2 1
Nivo 40 mg Q3W 3 1 1 1 0.001), and with Medicare insurance (OR = 1.81, 95% CI 1.11-2.97, p = 0.019). On
Nivo 40 mg Q4W 1
Nivo 80 mg Q4W 1 multivariable analysis, Latinx (OR = 2.18, 95% CI 1.02-4.52, p = 0.040) and Medicaid versus
Nivo 100 mg Q2W
Nivo 100 mg Q3W 1
1 1 1
private insurance (OR = 2.28, 95% CI 1.12-4.61, p = 0.022) were independent predictors of
Nivo 100 mg Q4W
Nivo 140 mg Q2W
1
1 1
transportation need. Conclusions: Our findings support the utility of transportation
Pembro 1 mg/Kg Q3W
Pembro 1 mg/Kg Q6W
1
1
screening as a tool for anticipating and providing resources to minimize missed RT
Pembro 50 mg Q3W 1 treatments. Future initiatives toward improving RT access may benefit from proactive
Pembro 100 mg Q3W 3 1 1
Pembro 300 mg Q6W 1 assessment and support of social needs, including but not limited to transportation.
Atezo 1 mg/Kg Q3W 1
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 79s

1528 Poster Session 1529 Poster Session

Disparities in breast cancer screening for the Brazilian Unified Health Sys- Differences in insurance status among Asian Americans with cancer: A
tem (SUS): A warning of the need to change public policies. First Author: Ruffo disaggregated analysis by ethnic subgroup. First Author: Lilac Nguyen, Memorial
Freitas-Junior, CORA – Advanced Center for Diagnosis of Breast Diseases Federal Sloan Kettering Cancer Center, New York, NY
University of Goias, Goiania, Goias, Brazil Background: Asian Americans, the fastest-growing racial/ethnic group in the U.S.,
Background: Breast cancer is the most prevalent form of cancer in Brazilian women, experience significant variation in barriers to cancer care access, yet most research
contributing significantly to cancer-related mortality, particularly when diagnosed at treats them as a monolithic group. Insurance is crucial for accessing cancer care, but
advanced stages. Public policies of the Ministry of Health have been not changed for the limited data exist on non-insurance rates among Asian Americans. This study inves-
last 2 decades. Methods: This ecological, temporal series study evaluated breast cancer tigates the heterogeneity in insurance status across Asian American subgroups in the
screening coverage, clinical staging, and the time from diagnosis to treatment initiation context of common cancers. Methods: We analyzed data from the National Cancer
in women of 40-49, 50-69 and 70 years of age in Brazil as a whole, its geographical Database (NCDB), and focused on breast, prostate, and non-small-cell lung cancers
regions, and states between 2013 and 2022. The data were extracted from databases of among patients under 65. Insurance status was dichotomized as non-insured/Medicaid
the Unified Health System (DATASUS). Results: There was a decreasing trend in versus private/government/Medicare/other, and binary logistic regression was used to
screening coverage for the 40-49-year age group between 2013 and 2020 (APC = -10.79; calculate odds ratios (ORs) for insurance status by Asian American subgroups.
p , 0.001), followed by stability in 2020-2022. Rates for the 50-69-year group remained Results: There were 2,161,947 patients with breast cancer, 603,172 with lung cancer,
stable, while coverage fell for women 70 years of age (APC = -6.27; p , 0.001) between and 964,423 with prostate cancer. We found significant heterogeneity in insurance
2013 and 2022. Cases of advanced stages at diagnosis tended to increase in all age status among Asian American subgroups; such heterogeneity was mirrored across the
groups: 40-49 (APC = 1.71; p , 0.001), 50-69 (APC = 1.43; p , 0.001) and 70 years (APC three most common cancer types. Consistently, Japanese Americans were less likely to
= 1.82; p = 0.001). Breast cancer screening coverage was low for all the age groups and be uninsured or have Medicaid than White Americans (4.5%, 2.5%, and 12.2%, for
all geographical regions, with lower rates found for the 40-49 and 70-year age groups. patients with breast, prostate, and lung cancer, respectively, among Japanese-
The poorest coverage was in the north, northeast and Midwestof the country, revealing Americans vs. 10.2%, 5.5%, and 20.7% among White Americans, (odds ratio
regional disparities. The proportion of cases diagnosed at advanced stages (III/IV) 0.54(95% CI: 0.47-0.61) for breast, 0.55(0.37-0.83) for prostate, 0.72(0.56-0.91) for
increased,particularly in younger women (40-49 years) and the elderly (70 years). Time lung), while every other Asian subgroup was significantly more likely to be uninsured or
from diagnosis to treatment initiationexceeded 60 days in . 50% of cases in all age have Medicaid than White Americans (OR range 1.24 to 7.23 for breast, 1.25 to 10.42 for
groups, with an increasing trend in women of 50-69 (APC = 1.27; p , 0.001) and 70 years prostate, 1.12 to 6.39 for lung, all p-values ,0.05 except for Laotians, Hmongs, and Thai
of age (APC = 1.83; p , 0.001). Conclusions: This study highlightsthe urgent need for with prostate cancer). Pakistani Americans were the group most likely to be uninsured or
public policies to increase breast cancer screening coverage beyond the 50-69-year age on Medicaid among patients with breast cancer (OR 7.22 (6.17-8.46) and prostate
group, and to guarantee equitable access to early diagnosis and timely treatment, cancer (OR 10.41 (7.08-15.33) and were the second-most likely subgroup among pa-
particularly in less affluent areas. Dealing with these disparities is crucial to improving tients with lung cancer (OR 4.82 (3.21-7.23). Hmong and Kampuchean were among the
breast cancer outcomes in Brazil. Research Sponsor: None. top three groups with breast or lung cancer who were uninsured or on Medicaid.
Conclusions: Significant heterogeneity exists in insurance coverage among Asian
American subgroups, which highlights the diversity of disparities Asian American pa-
tients face; these findings call into question the model minority myth, as every other
Asian American subgroup besides Japanese-American was significantly less likely to be
insured than White Americans. Asian American patients have broad differences in
histories, social determinants of health, and barriers to accessing care, which may merit
differentially targeted health interventions among certain subgroups. Research Sponsor:
None.

1530 Poster Session 1531 Poster Session

Breast cancer screening mammography among transgender and gender Implementation of an academic precision oncology service in a community
diverse (TGD) individuals: A nationwide study of >10,000 TGD individuals. setting. First Author: Junlone Moy, Eshelman School of Pharmacy, The University of
First Author: Elizabeth Jane Cathcart-Rake, Mayo Clinic, Rochester, MN North Carolina at Chapel Hill, Chapel Hill, NC
Background: Regular breast cancer screening reduces mortality. The American College of Background: Implementation of precision oncology (PO) in community practice remains
Radiology (ACR) recommends annual screening mammography for asymptomatic, average challenging due to insufficient resources to order and interpret genomic test results. In 2023,
risk 1) transgender and gender diverse (TGD) men (individuals assigned female sex at birth our PO program (located at an academic hub) partnered with a community site to optimize
but identify as men) age 40+ and with residual breast tissue and 2) TGD women (individuals clinical workflows around molecular testing and to provide documented expert review of
assigned male sex at birth but identify as women) age 40+ and on gender-affirming hormone testing results. To date, the service has reviewed over 150 cases. We undertook a retrospective
therapy for 5+ years. This nationwide study investigated screening mammography in TGD review to determine the impact of our intervention on the rates of next generation sequencing
individuals. Methods: Four cohorts were identified with complex algorithms of medical and (NGS) ordered at this site. Methods: Patients with visits to the community cancer center
pharmacy data in the OptumLabs Data Warehouse, a longitudinal, administrative insurance between 1/2022 and 7/2024 were screened for inclusion (n = 900). Eligible patients had an
claims database that includes patients with commercial insurance and Medicare Advantage: advanced/metastatic solid tumor and were managed at the community site. Patients were
1) TGD men, 2) TGD women, 3) individuals with gender dysphoria not meeting other TGD placed in the historical or interventional cohort based on the date that their cancer became
criteria (NMOT), and 4) cisgender women – all of whom met ACR screening criteria. The advanced/metastatic relative to the service implementation (1/1/2023). Due to a historical
primary endpoint was the percentage of individuals with high adherence, defined as lack of genomic data integration within the EHR, evaluation of NGS results and subsequent
completing .75% of recommended screenings; comparisons were made across the cohorts treatment required manual chart review. The study was approved by the IRB of the University of
(Chi-square test). Multivariable logistic regression was used to compare adherence between North Carolina. Results: We identified 109 historical patients and 76 interventional patients
TGD men and cisgender women with adjustment for demographics and with 1:4 matching who met eligibility. NSCLC patients comprised the majority of the population (27%), followed
(on race/ethnicity, age group, year of healthcare plan enrollment, and duration of follow-up). by prostate (12%), breast (10%), and colon (7%) cancers. We observed a significantly increased
Results: 10,478 TGD individuals (3,778 TGD men; 1,294 TGD women; and 5,406 with gender rate of NGS testing in solid tumors after implementation of the service (40.4% versus 57.9%
before and after, p = 0.0192). The rate of NGS testing within 30 days of diagnosis of advanced/
dysphoria NMOT) and 6,218,369 cisgender women were identified. For TGD men, TGD
metastatic disease also improved in the interventional cohort (21.1% versus 36.8%, p =
women, individuals with gender dysphoria NMOT, and cisgender women, high adherence
0.0187). The median time from first clinic visit to results was 20.5 days in those who were
was observed in 41.1% (95% confidence interval (CI): 39.5, 42.6%); 7.4% (95% CI: 6.0, 8.8%);
testing in the historical cohort versus 10.5 days in the interventional cohort. RNA tran-
11.9% (95% CI: 11.0, 12.7%); and 38.3% (95% CI: 38.3, 38.4%), respectively, p , 0.0001. No
scriptome sequencing was used more frequently in the interventional cohort (40.9% versus
screening ever in these same populations was observed in 23.6% (95% CI: 22.2, 24.9%); 6.8%, p = 0.0003). Conclusions: Utilization of a PO program at an academic hub to support
81.3% (95% CI: 79.2, 83.4%), 72.8% (95% CI: 71.6, 74%), and 35.9% (95% CI: 35.8, 35.9%), NGS testing at a community site resulted in increased testing rates and more timely access to
respectively, p , 0.0001. When TGD men were matched and compared to 15,112 cisgender results in advanced cancer patients. Although limited by sample size, the data emphasizes a
women, a differential association was observed between gender and high adherence by age continued need for infrastructure to support the application of PO in community settings.
group (p , 0.0001); compared to cisgender women, the likelihood of high adherence in TGD Research Sponsor: None.
men ages 40-49 (n = 2,472), 50-59 (n = 852), 60-69 (n = 320), and 70+ (n = 134) were as
follows: odds ratios (OR‘s): 1.38 (95% CI: 1.28, 1.50), p , 0.0001; 1.57 (95% CI: 1.35, 1.81), Historical Interventional
p , 0.0001; 0.812 (95% CI: 0.619, 1.065), p = 0.1316; and 1.277 (95% CI 0.835, 1.951), p = (Prior to 1/1/2023) (After 1/1/2023) P-value
0.2595, respectively. Conclusions: This study found that TGD men manifest relatively high Rate of NGS testing in all patients 44/109 (40.4%) 44/76 (57.9%) 0.0192
adherence to screening mammography, whereas TGD women and individuals with gender Rate of NGS testing within 30 days of diagnosis 23/109 (21.1%) 28/76 (36.8%) 0.0187
Rate of NGS testing in NSCLC 15/24 (62.5%) 22/26 (84.6%) 0.109
dysphoria NMOT manifest low adherence. Future research should focus on improving breast Rate of blood-only NGS testing in those with results 27/44 (61.4%) 19/44 (43.2%) 0.0896
cancer screening in these underserved populations. Research Sponsor: National Institutes Rate of genome-informed therapy (all patients) 5/109 (4.6%) 8/76 (10.5%) 0.121
of Health, National Institute of Aging; K23 MD019644; National Institutes of Health, NCI; K07 Rate of genome-informed therapy (NSCLC) 1/24 (4.2%) 4/26 (15.4%) 0.3508
AG076401. Median time from clinic visit to NGS testing results 20.5 days 10.5 days ———

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
80s CARE DELIVERY/MODELS OF CARE

1532 Poster Session 1533 Poster Session

Development and fairness assessment of machine learning models for Fall risk assessment by QTUG device in geriatric cancer patients on che-
predicting 30-day readmission after lung cancer surgery. First Author: motherapy at a tertiary care hospital. First Author: Tilak Tvsvgk, Command
Atulya Aman Khosla, Department of Internal Medicine, William Beaumont University Hospital (AF), Bengaluru, India
Hospital, Royal Oak, MI Background: Age-related changes, such as declines in muscle strength, balance, and coordination,
Background: Predicting post-surgical readmissions is essential for improving patient are exacerbated by cancer and its treatments elevating the risk of falls. Cancer chemotherapy often
outcomes and reducing healthcare costs. While machine learning (ML) models offer high lead to adverse effects which can increase fall risk. The consequences of falls in older cancer patients
can be severe impacting the recovery process. The aim of the study was to estimate the baseline fall
predictive accuracy, they may perpetuate healthcare disparities if not rigorously risk and frailty in geriatric patients and comparing the same after 03 cycles of chemotherapy and
evaluated for algorithmic bias. In this study, we examine the limitations of ML-based correlate the factors contributing to the increased risk. Methods: A prospective observational study
readmission prediction models, highlighting how bias can persist despite strong per- enrolling elderly cancer patients, who are not bed-bound or using a limb prosthesis were enrolled.
formance metrics. We also explore the impact of integrating fairness constraints to Patients with brain/spinal cord tumours were excluded. Fall risk assessment was done using the
mitigate these disparities, ensuring equitable clinical decision-making across racial and Kinesis Q-TUG device. Based on the scores the patients were classified into low-risk, moderate-risk
ethnic groups. Methods: We analyzed National Surgical Quality Improvement Program and high-risk of falls. The fall risk assessment was repeated after three cycles of chemotherapy.
(NSQIP) data (2016–2020) for 23,843 lung cancer surgery patients. Multiple ML models Results: A total of 94 males and 122 females (n = 216) patients, 25% of whom were more than 70 yrs
of age were enrolled. Overall the females were associated with higher fall risk and frailty compared to
were developed using demographic, clinical, and laboratory variables. Model perfor-
males. The combined fall risk estimates at baseline and post 3 cycles did not reveal a significant fall
mance was assessed using standard accuracy metrics alongside fairness evaluations, risk increase, however the timed-up-and-go (TUG) times were significantly lower post 3 cycles [Table-
including Demographic Parity and Equalized Odds, to measure disparities across racial 1]. The average stride length, stride velocity did not show any significant difference but the number of
groups. Results: The cohort had 56.5% females; 66.4% of cases belonged to the White steps to turn increased post 3 cycles. Though the frailty did not show a significant difference, the
race, 6.3% were Black, and 2.9% belonged to the Hispanic ethnicity. The median [Q1, Q3] correlation between fall risk and frailty showed a significant positive correlation (r = -0.91).
was 69.0 [62.0, 74.0] years, and the overall readmission rate was 7.5%. The median Conclusions: While there was a slight increase in fall risk and frailty among geriatric cancer patients
operation time was higher among readmitted cases (171 minutes vs. 157.0; p , 0.001). undergoing chemotherapy, these changes were not statistically significant. However, the decline in
However, there was no clinically significant difference between median [Q1, Q3] LOS mobility underscores the adverse effects of chemotherapy on functional performance. The strong
correlation between fall risk and frailty emphasizes the need for assessments and targeted inter-
between the two groups (4.0 [2.0, 6.0] vs. 4.0 [3.0, 7.0]; p , 0.001). The best-performing ventions to mitigate fall risks. With increasing number of older adults with cancer, the findings
model (CatBoost) achieved high accuracy but showed disparities in prediction rates advocate for enhanced monitoring and early management strategies to improve safety and outcomes
across racial groups (Demographic Parity Difference: 0.030, Equalized Odds Difference: for elderly cancer patients. Research Sponsor: None.
0.333) since the model disproportionately flagged Hispanic patients for readmission risk Characteristics of QTUG device in elderly cancer patients.
while potentially under-identifying risk in other groups. Significant predictors included
Parameter Category Median IQR Min Max P value
operative time, preoperative sodium (139 vs. 140 mmol/L, p , 0.001), and COPD status
Combined Fall Risk (%) Pre 49.18 73.44 - 31.53 11.78 79.46 0.098
(33.8% vs. 25.3%, p , 0.001). After implementing fairness constraints, the model Post 47.75 60.28 – 36.15 13.96 84.64
maintained strong predictive performance while reducing demographic disparities, with Combined Frailty Risk (%) Pre 63.48 74.98 – 47.27 12.33 93.15 0.743
selection rates balancing across racial groups (range: 0.51%-3.50%). Post 60.25 75.29 – 39.98 11.31 94.41
TUG (s) Pre 12 17 – 9.69 6.5 28.29 0.002
Conclusions: Despite their high accuracy, ML models for predicting post-surgical Post 13.44 15.8 – 11.19 7.19 274.29
readmissions can reinforce existing healthcare disparities. Our findings underscore Average Stride Velocity (cm/s) Pre 93.73 105.67 – 73.37 48.31 135.67 0.062
the importance of fairness-aware modeling to mitigate bias, ensuring equitable clinical Post 95.73 115.23 – 77.29 54.22 132.66
Time taken to turn (s) Pre 2.78 3.4-2.24 1.10 36.32 <0.001
decision support. While fairness constraints improved demographic balance, residual Post 6.18 7.2-5.3 3.66 56.48
disparities persisted, highlighting the need for ongoing scrutiny when deploying AI in Number of steps to turn Pre 2.5 3-2 1 9 <0.001
clinical settings. This study emphasizes the critical need for continuous fairness Post 3.5 4-3 2 10
evaluation in medical AI applications to prevent unintentional harm to vulnerable patient
populations. Research Sponsor: None.

1534 Poster Session 1535 Poster Session

APP-first as a strategy to increase new patient access and treatment of Gender, place of death, and racial disparities in the reporting odds ratio of
patients with gastrointestinal malignancies within a cancer center. First cardiovascular disease burden in leukemia across age groups (15–85) in the
Author: Amalia Stefanou, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL U.S.: A CDC WONDER disproportionality analysis. First Author: Tehmasp Mirza,
Background: Quality cancer care depends on timely and efficient evaluation of patients to determine Shalamar Medical and Dental College, Lahore, Pakistan
next steps in treatment. To support this goal, we implemented a care delivery model where advanced Background: Leukemia, a hematologic malignancy, often coexists with cardiovascular
practice providers (APP) were tasked with seeing new patients as an entry into the system, with a goal disease (CVD), worsening outcomes due to shared risk factors like diabetes, hyper-
of completing testing and referrals prior to a subsequent visit with a surgical oncologist. Methods: A
retrospective review was performed focusing on four hepatobiliary and pancreatic surgeons and their
tension, and smoking. Despite CVD’s known impact on leukemia mortality, research on
corresponding APP teams. The APP-First program was deployed in 2021. The purpose was to increase demographic and geographic disparities remains limited. This study examines dis-
capacity and best prepare the patient’s work-up prior to an appointment with the surgeon to facilitate parities in the Reporting Odds Ratio (ROR) of CVD burden among leukemia deaths across
treatment initiation. We compared NPs during two distinct time periods, before (7/2018-6/2020) and age, gender, place of death, and race/ethnicity using CDC WONDER data (1999–2020).
after (7/2022-6/2024) implementation of APP First. The primary outcome, impact on access, was Methods: A disproportionality analysis of CDC WONDER death certificate data for U.S.
determined by change in the number of NP seen by the group, and the secondary outcomes were adults (15–85) was conducted. Records were grouped into four variables: leukemia
number and proportion of patients receiving treatment at our institution. Patients were excluded if deaths with CVD (A), leukemia deaths (B), CVD deaths (C), and all deaths (D). RORs were
time to care was . 180 days or they received non-GI related care. Changes in outcomes of interest
calculated as (A/B) / (C/D) and stratified by gender, race/ethnicity, urbanization, and
before and after implementation of the program were compared by Chi-square with significance set at
p = 0.05. Results: A total of 2585 NPs were seen during the study period with 1797 beginning place of death. Age groups were categorized into 15–24, 25–64, and 65+ years.
treatment at our facility (69.5%). During the pre-intervention period 1091 NPs were seen by the group, Joinpoint regression was used to compute annual percentage change (APC) and average
including 277 (25.5%) initially evaluated by an APP. Following the model implementation, 1494 NPs annual percentage change (AAPC) to identify trends. Results: Analysis of 22 years of
were seen, 915 (61.2%) by an APP and 579 (38.8%) by an MD (p , 0.001). There was no change in data revealed disparities in RORs across demographics. The 15–24 age group had the
percent of NPs choosing to pursue care at our institution (68.7% vs 68.7%, p = 0.970), however after highest ROR for males (2.51) and females (1.59), indicating a greater CVD burden in
implementation, patients were more likely to be scheduled for operations after their initial visit (11.4% leukemia deaths than all-cause mortality. Middle-aged (25–64) and older adults (65+)
vs 14.3%, p = 0.031). Conclusions: Implementation APP-first led to increased NPs capacity translating had lower RORs (,1), suggesting a reduced CVD-leukemia burden in these groups.
into a 36.9% higher volume of NPs seen, without change in patient satisfaction as demonstrated by an
unchanged percentage of patients choosing treatment at our institution. The established workflow
Trends: The young female cohort showed a sharp ROR decline (2018–2020, APC:
within this model facilitated expedited care, resulting in higher proportion of patients treated and an -22.02%; 95% CI -36.00 to -1.23; p=0.039), while the elderly male cohort (85+) had a
increase in the number and proportion of patients receiving surgery. These data support the steady rise (APC: 6.93%; 95% CI 1.74–9.24; p,0.0001). Place of Death: Medical facilities
implementation of delivery of care models leveraging the role of APPs in a well-integrated system, with had the highest CVD burden in leukemia deaths, especially in younger cohorts (15–24;
overall improved capacity, access, and treatment for patients with cancer. Research Sponsor: None. ROR: 1.47). Hospice facilities had the lowest RORs across age groups. The "Other/
Outcomes of APP-First program. Unknown" category had an outlier ROR of 4.81 in the youngest cohort, suggesting data
FY 19-20 (n,%) FY 23-24 (n,%) p value limitations. Race/Ethnicity: Hispanics had the highest ROR (2.64) in the 15–24 age
Total # NP 1,091 1,494 –
group, followed by Asians (2.21). In middle and older age groups, RORs declined for all
NP Distribution races. Hispanics (65–74) showed an increasing ROR trend (AAPC: 0.92%; 95% CI
APP 277 (25.4) 915 (61.2) <0.001 0.45–1.36; p=0.002). Conclusions: Significant disparities exist in CVD burden among
MD 814 (74.6) 579 (38.8)
Treated leukemia deaths, with younger cohorts, males, Hispanics, and patients in medical fa-
Yes 750 (68.7) 1026 (68.7) 0.970 cilities showing the highest RORs. These findings highlight the need for targeted cardio-
No 341 (31.3) 468 (31.3)
Surgery 124 (11.4) 213 (14.3) 0.031 oncology strategies to address CVD risk in leukemia patients. Further research on
Chemotherapy 143 (13.1) 194 (13.0) 0.928 chemotherapy-related cardiotoxicity and healthcare disparities is crucial to reducing
Endoscopy 376 (34.5) 466 (31.2) 0.080 inequities and improving outcomes. Research Sponsor: None.
Radiation 112 (10.3) 154 (10.3) 0.972

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 81s

1536 Poster Session 1537 Poster Session

Oncofertility practice patterns in NCCN cancer centers after the overturn of Trade-off preferences in older adults with newly diagnosed acute myeloid
Roe v. Wade. First Author: Nikita V. Baclig, University of California Los Angeles, Los leukemia. First Author: Kah Poh Loh, University of Rochester Medical Center,
Angeles, CA Rochester, NY
Background: The Supreme Court decision of Dobbs v. Jackson to overturn Roe v. Wade Background: Treatment decisions for older adults with acute myeloid leukemia (AML) are highly
gave states authority to regulate reproductive health. This has led to concerns about preference-sensitive, requiring a balance between survival and other important outcomes such as
access to assisted reproductive technologies used in oncofertility. The impact of this toxicities and quality of life (QoL). Understanding patients’ trade-off preferences is critical for guiding
personalized treatment planning. We examined the trade-off preferences of older adults newly di-
legal climate on oncofertility practices remains unknown. This study aims to understand agnosed with AML and factors influencing these preferences. Methods: We collected data from two
how academic cancer centers across the United States have experienced changes in clinical trials evaluating an AML communication tool. Older adults completed questionnaires at
oncofertility care in a post-Roe world. Methods: In collaboration with the National diagnosis assessing trade-offs between survival and two key outcomes: a) maintaining QoL and b)
Comprehensive Cancer Network (NCCN) Best Practices Committee, a survey was de- treatment-related toxicities (nausea/vomiting, bedbound status, assistance with daily activities,
veloped to evaluate changes in oncofertility access and utilization in the 2 years since worsening memory, and confusion). The survival-QoL trade-off was categorized as agree vs. disagree.
the Dobbs decision. In July 2024, the survey was sent to NCCN Member Institutions, Trade-offs for treatment-related toxicities were scored from 0 to 5, with higher scores indicating a
which represent 23 states with varied post-Roe protections for reproductive care. The greater willingness to endure toxicities for survival. We used binary logistic regression to identify
factors associated with survival-QoL trade-off, while ordinal logistic regression was used for survival-
survey responses were de-identified for analysis. Questions were both multiple choice
toxicity trade-off. Results: We included 95 older patients with newly diagnosed AML; mean age was
and free response. Results: The survey was sent to 33 NCCN Member Institutions and 73.7 (SD 7.6), 38% female, and 94% White. Approximately 37% received intensive and 54% received
yielded 24 responses (72.7%). A majority (83.3%) indicated that reproductive care was a lower-intensity treatment. Only 15% prioritized survival over maintaining QoL, 39% neutral, and 46%
moderate-high priority for their cancer center. Most (62.5%) reported an increase in the prioritize maintaining QoL over survival (Table). Over 60% would decline treatments leading to
number of cancer patients receiving fertility preservation. According to the Center for confusion, and 45% would avoid treatments causing bedbound status. On multivariable analyses,
Reproductive Rights, 13 institutions are in states that have restricted reproductive rights patients enrolled with a caregiver had a significantly higher odds of prioritizing survival over QoL [Odds
since the Dobbs decision. However, only 4 (16.7%) survey respondents reported that Ratio (OR): 5.92, p=0.04]. Employed patients were more likely to endure treatment-related toxicities for
reproductive health laws had become more restrictive. The remaining reported that laws survival compared to those who were unemployed, retired or homemaker (OR: 7.76, p,0.01).
Conclusions: Trade-off preferences among older adults with AML vary widely and are influenced by
had not changed, were less restrictive, or abstained (83.3%). All 4 respondents who caregiving support and employment status. Actively eliciting these preferences is essential to align
indicated more restrictive laws reported moderate-high priority placed on reproductive treatment decisions with individual patient values. Research Sponsor: Conquer Cancer Foundation
care and half reported an increase in cancer patients receiving fertility care. In the states Walther Cancer Foundation; American Cancer Society.
that reported no or neutral change (n=14, 58.3%) or less restrictive laws (n=5, 20.8%),
Preferences of older adults with AML.
most (68.4%) reported an increase in patients receiving fertility care. One respondent
I would like to try treatments for my cancer Agree/ Disagree/
who indicated more restrictive laws reported a decrease in resources for fertility care. if they could help me live longer, Strongly Strongly
Many centers have prioritized oncofertility by developing oncofertility programs, even if it is very likely they would agree Neutral disagree
assigning fertility navigators, and creating electronic health record-assisted referral Have high level of side effects (e.g., nausea/vomiting) 60% 24% 16%
alerts and clinical pathways. Conclusions: Large academic NCCN Member Institutions, Make me require more assistance from family and friends with 53% 22% 25%
most in states with no change or less restrictive reproductive laws since the Dobbs completing daily activities (e.g., shopping, managing money)
Make me bedbound and unable to use the bathroom without 35% 20% 45%
decision, reported an increase in number of patients who accessed fertility care. Ad- assistance
ditional studies will clarify whether this reflects underlying trends or increased fertility Make my memory worse 29% 29% 40%
care due to a fear of limited future access. Only a minority of the institutions in restrictive Cause me to become confused often so that I am not aware 15% 23% 61%
of my surroundings
states responded to the survey and most who did reported similar or improved access to Living longer is more important to me than maintaining my 15% 39% 46%
oncofertility care. The lack of response from restrictive states needs to be examined quality of life
further as it may reflect concerns about oncofertility care in the new political landscape.
Research Sponsor: None.

1538 Poster Session 1539 Poster Session

OP-35: Does a tool designed to measure potentially preventable chemo- Qualitative findings from providers and patients for planning implementa-
therapy toxicities do so effectively? First Author: Ryan W. Huey, The University of tion of screening clinical breast examination in Soweto, South Africa. First
Texas MD Anderson Cancer Center, Houston, TX Author: Daniel O’Neil, Yale Cancer Center, Yale University, New Haven, CT
Background: OP-35 is a measure developed by the National Quality Foundation that the Center for Background: Over 50% of South African women with breast cancer (BC) are diagnosed at
Medicare and Medicaid Services (CMS) uses to evaluate the quality of care for patients undergoing stages III & IV. To inform an Implementation Mapping process to design strategies for
outpatient chemotherapy treatment. Launched in 2021, it was intended to measure rates of potentially implementing screening clinical breast exam (CBE) in primary care facilities in Soweto,
preventable complications of chemotherapy treatment. The tool assesses the rate of emergency
department visits and admissions (EDV/A) visits for patients receiving outpatient intravenous (IV)
South Africa, we gathered qualitative data from local primary care providers and patients
systemic anti-cancer therapy (SACT) and defines potentially preventable by the presence of $1 of 10 on barriers to CBE and possible implementation approaches. Methods: We conducted
diagnoses: anemia, dehydration, diarrhea, emesis, fever, nausea, neutropenia, pain, pneumonia, or semi-structured interviews with administrators, nurses, doctors, and community health
sepsis. However, it is unknown if these diagnoses track truly preventable visits and therefore if it is a workers (CHWs) and focus groups with women potentially eligible for screening CBE at
valid measure of quality. Methods: We conducted a retrospective review of patients who received four Soweto primary care facilities that do not offer BC screening. Our discussion guide
outpatient IV SACT (the denominator for OP-35) at the University of Texas MD Anderson Cancer Center explored BC screening perceptions among both groups. We also asked providers for
between January 2023 and December 2023. All patients who had an EDV/A were assessed to un- recommendations about how to best implement a future CBE screening program, and we
derstand the primary and secondary diagnoses associated with their encounters. Results: The total
asked patients about factors that would motivate them to participate in such a program.
number of patients included in the population who received outpatient IV SACT was 10,353. Of these,
2,401 (23.2%) had an EDV/A within 30 days of receiving outpatient IV SACT. Of patients with an EDV/A, We analyzed transcripts deductively in parallel with data collection. We organized themes
67% of patients had one EDV/A, 21% had 2, and 12% had $3. The most common diagnosis groups were using the Consolidated Framework for Implementation Research (CFIR). To support
pain (83%), anemia (69%), nausea (44%), and fever (35%). 82% of patients had more than one qualifying Implementation Mapping’s emphasis on addressing the needs of individual stakeholders,
diagnosis. For 68%, the qualifying diagnosis was a secondary diagnosis. Of patients with a qualifying our analysis focused on the CFIR Individuals domain and Characteristics subdomain taken
EDV/A, only 15% did not have a qualifying OP-35 diagnosis. Conclusions: While OP-35 was designed to from the COM-B system. Results: We analyzed 27 interviews with 8 administrators, 1
measure potentially preventable chemotherapy-related complications, the exclusion of relatively few medical officer, 14 nurses, 3 CHWs, and 1 clerk, and 4 focus groups with 23 total women.
patients among those with an EDV/A suggests that a significant proportion of qualifying events may not Providers (i.e., deliverers) and patients (i.e., recipients) alike expressed enthusiasm for
truly be preventable. The qualifying diagnosis list may need to be tailored to exclude non-preventable
admissions. This would improve the metric’s specificity and validity as measure of care-quality. Future CBE’s potential to decrease BC mortality and morbidity. Both groups also cited CBE’s
work should clarify which diagnoses lead to misclassification of non-preventable EDV/As. Research potential to overcome and counteract patients’ limited knowledge of breast health and BC
Sponsor: None. symptoms. The primary barrier to CBE, according to both groups, is the high patient
volume at public facilities. Providers described staff shortages limiting opportunity to
Patients with qualifying EDV/A N = 2,401
perform CBE and patients cited long wait times as a barrier to pursuing “extra” services.
Type of Cancer Solid Malignancy 2,020 (84%)* Providers often recommended hiring new personnel designated for CBE screening.
Heme Malignancy 383 (16%)*
Qualifying OP-35 Diagnosis Pain 1,998 (83%) Patients suggested various approaches to expanding access, such as screening in both
Anemia 1,652 (69%) the clinic and community, opportunistic screening while patients wait for other clinic
Nausea 1,065(44%)
Fever 833 (35%)
services, and walk-in access for “screening only” visits. Patients also emphasized the
Dehydration 679 (28%) need to improve trust in the clinics and their staff. Regarding educational outreach,
Pneumonia 552 (23%) providers focused on expert-delivered teaching in both the clinic and community. Patients
Neutropenia 509 (21%)
Diarrhea 534 (22%) valued experts but also recommended engaging BC survivors and other community
Sepsis 364 (15%) members to promote screening through word-of-mouth. Conclusions: Soweto’s primary-
Emesis 148 (6%)
Number of EDV/A Per Patient 1 1,614 (67%)
care clinicians and patients expressed enthusiasm for the health benefits of BC screening,
2 502 (21%) but successful implementation must address barriers faced by both groups, including
$3 285 (12%) long clinic wait times and personnel shortages. Research Sponsor: National Cancer
*Two patients categorized with both solid and heme malignancies. Institute.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
82s CARE DELIVERY/MODELS OF CARE

1540 Poster Session 1541 Poster Session

Development of an early sepsis treatment-decision algorithm in children and Impact of the eSyM symptom monitoring program on nurse telephone
adolescents with cancer in a middle-income country: Results from a mul- encounters across six cancer centers. First Author: Michael J. Hassett, Dana-
tinational modified Delphi consensus. First Author: Paula Aristizabal, Division of Farber Cancer Institute, Boston, MA
Hematology/Oncology, Department of Pediatrics, University of California San Diego and Background: After developing an ePRO-based, EHR-integrated symptom monitoring program (eSyM)
Rady Children’s Hospital-San Diego and University of California San Diego Moores and implementing it across 6 health systems, we found lower odds of acute care utilization among
Cancer Center, San Diego, CA those who used eSyM to report symptoms. Facilitating communication between patients and cli-
nicians is a potentially important mechanism by which symptom monitoring programs may improve
Background: Despite global efforts, striking childhood cancer survival gaps between low-
outcomes. We measured the association between eSyM deployment, symptom reporting and severe
and middle-income countries (LMIC) and high-income countries persist. Based on data symptom reporting on the frequency and number of nurse telephone encounters (TELs).
from the Colombian childhood cancer clinical outcomes surveillance system, VIGICANCER, Methods: eSyM was deployed in a stepped wedge RCT from 2018-2023 for adults who started
sepsis accounts for approximately 90% of all preventable deaths. In response, we chemotherapy (CHEM) or were discharged following surgery (SURG) for a suspected or confirmed GI,
developed a consensus-based Treatment-Decision Algorithm (TDA) for early sepsis de- GYN or thoracic cancer. We analyzed three cohorts: 1) for all patients, we compared those treated
tection and treatment, adapted to the local context and balancing optimal clinical before vs. after eSyM deployment; 2) for post-deployment patients (eSyM eligible), we compared those
management and resource utilization. Methods: We used RAND/UCLA Delphi method who did vs. did not report symptoms within 30 days of first eSyM prompt; and 3) for symptom reporters
(preparatory phase, literature review, rating) to consult experts on the appropriateness of (eSyM users), we compared those who did vs. did not report severe symptoms. Outcomes of interest
the proposed risk/alert definitions, critical steps, evidence-based interventions, and were the proportion of patients with at least one TEL and total number of TELs within 30 days of first
eSyM prompt. Poisson regression was used to estimate the number of TELs within 30 days accounting
decision-making trees in the adapted TDA. Consensus involved: a) Pilot anonymous voting
for cancer and treatment type, as well as age, gender, and other factors. Results: In total, 18,830
on a 22-statement online survey (5-point Likert scale, open-ended questions); b) Reading/ patients were and 21,112 were not exposed to eSyM (median age 64, 66% female). Among eligible
rating in a hybrid meeting (in-person/virtual); and c) Voting on statements where con- patients, 8,298 (44%) reported symptoms within 30 days. Among eSyM users, 3,666 (44%) reported
sensus was not reached. Consensus was defined as: $70%, strong; 51%-69%, moderate; one or more severe symptoms within 30 days. The proportion of patients with TELs and the number of
#50%, no consensus. Results: Preparatory phase and literature review: A multinational TELs per patient are below (Table). In regression analyses, there were more TELs within 30 days after
(Colombia, Mexico, US), interprofessional panel of 19 members, including from pediatric eSyM deployment (SURG 0.09 [95%CI 0.07-0.11; P , .0001], CHEM 0.26 [95% CI 0.23-0.28; P , .0001])
cancer centers and academic societies (oncology, emergency medicine, hospital medicine, and among severe symptom reporters (SURG 0.42 [95%CI 0.38-0.45; P , .0001], CHEM 0.43 [95% CI
critical care, infectious diseases, nursing) ensured geographic and resource represen- 0.38-0.48; P , .0001]). Conclusions: eSyM exposure and reporting severe symptoms were associated
tation. A Colombian internal expert taskforce (n = 6) completed a comprehensive literature with a greater likelihood and larger number of TELs, supporting nurse intervention as a mediator of the
review, met biweekly, and developed the Colombian Protocol for Early Sepsis Detection in association between ePRO monitoring and clinical outcomes. Future studies should explore the
impacts of ePRO systems on nursing workload and health system costs. Clinical trial information:
Children with Cancer and accompanying TDA with 3 domains: sepsis screening, sepsis NCT03850912. Research Sponsor: National Cancer Instiute; 1UM1CA233080-01.
huddle, and early treatment. Consensus results: a)Pilot(n = 19 members), strong
agreement was obtained for 81% of statements in one round; b) Reading/rating (n = 8 % patients reporting Mean number of TELs among those
members), strong agreement was reached in 90% of statements after two rounds; and c) Cohort Treatment within 30 days, P value with at least one TEL (SD), P value
Voting on statements without consensus (n = 11 members, online), moderate agreement All patients Control Intervention Control Intervention
was reached in 100% of statements (2) in one round. Conclusions: Through collaborative Surg 63% 69% ,.0001 2.6 (2.5) 2.9 (2.5) ,.0001
Chemo 60% 79% ,.0001 3.9 (3.6) 4.0 (3.5) 0.08
consensus, we successfully developed an evidence-based, user-friendly TDA for early eSyM Non-reporter Reporter Non-Reporter Reporter
sepsis detection and treatment, tailored to resource-constrained settings. The diverse, eligible
interprofessional panel facilitated contextual adaptations of the TDA. The proposed TDA Surg 66% 71% ,.0001 3.0 (2.6) 2.9 (2.4) 0.23
Chemo 77% 81% ,.0001 4.1 (3.6) 3.9 (3.4) 0.04
provides clinicians serving children with cancer in Colombia with an easy-to-follow TDA eSyM No severe Severe No severe Severe
that is clear, exhaustive, and suitable for adaptation to individual local settings. Next steps users symptoms symptoms symptoms symptoms
involve applying improvement science methodology to implement the TDA in Colombia and Surg 67% 77% ,.0001 2.6 (2.1) 3.3 (2.7) ,.0001
Mexico and evaluating its predictive value for prompt sepsis detection in children with Chemo 76% 89% ,.0001 3.4 (3.0) 4.5 (3.5) ,.0001
cancer, contributing to reduction of survival gaps in LMIC. Research Sponsor: None.

1542 Poster Session 1543 Poster Session

Cost and resource utilisation for liquid biopsy vs tissue biopsy genotyping in Impact of a collaborative care-based symptom intervention model on che-
advanced NSCLC: A micro-costing model. First Author: David O’Reilly, Beaumont motherapy adherence in patients with breast cancer. First Author: Michael H.
RCSI Cancer Centre, Dublin, Ireland Storandt, Mayo Clinic, Rochester, MN
Background: For patients with advanced non-small cell lung cancer, tumour genotyping Background: Treatment toxicity may limit the ability of cancer patients to receive all
identifies actionable variants that inform targeted therapeutic choices, that improve recommended cycles of therapy. Four to six cycles of docetaxel plus cyclophosphamide
outcomes. Liquid biopsy genotyping (LBG) is a non-invasive approach to tissue biopsy (TC) is a common adjuvant chemotherapy regimen for early-stage breast cancer. We
genotyping (TBG) that reduces turnaround, avoids repeat tissue biopsy, and can identify assessed the impact of routine collection of patient-reported outcomes (PROs), coupled
additional actionable variants. However, despite these benefits, patient access to LBG is with a collaborative care model-based symptom management intervention, on the number
not universal in a range of healthcare systems. While others have developed models of cycles of TC received by patients with breast cancer. Methods: The Enhanced, EHR-
evaluating the cost-effectiveness of LBG, these have are limited by assumptions re- facilitated Cancer Symptom Control (E2C2) trial was a cluster-randomized, pragmatic
garding frequency of oncogenic variants and treatment utilisation. We utilised a micro- clinical trial, conducted between March 2019 and January 2023 at Mayo Clinic Rochester
costing model (MCM) to quantify the cost/resources of LBG and TBG in a prospective trial and within the Mayo Clinic Health System in Minnesota and Wisconsin. Patients regularly
(PLAN; [Link] Identifier: NCT05542485) aimed at investigating the feasibility reported the severity of 6 SPPADE symptoms (Sleep deficit, Pain, Physical function
of LBG in a tertiary cancer centre. Methods: A deterministic MCM was developed to impairment, Anxiety, Depression, and Energy deficit/fatigue) on 11-point numerical rating
enumerate the cost to generate a genomics report for both LBG and TBG in NSCLC. Capital scales. Each symptom score was interpreted as none to mild (0-3), moderate (4-6), or
costs were calculated based on up-front investment and annual depreciation/ severe (7-10). The E2C2 intervention included symptom management education modules,
maintenance. Costs of consumables and staff time associated with each procedure clinician decision support aids, and the option to discuss severe symptoms with a nurse,
was sourced from relevant hospital departments (e.g. Medical Physics) and evaluated for physical therapist, or social worker. Patients with breast cancer who received at least one
accuracy by a health economist and medical oncologist. We calculated the cost of sample cycle of TC were included in this analysis. We compared the number of cycles of docetaxel
acquisition (endobronchial ultrasound-guided biopsy or phlebotomy), processing, and and cyclophosphamide completed by patients in the control condition versus those in the
genotyping for both LBG and TBG, from patients enrolled on the PLAN study (n = 100) intervention condition. Results: We identified 198 patients with breast cancer who
between 08/2023-07/2024. Finally, we performed an exploratory analysis investigating received TC during the control condition and 128 who received TC during the intervention
potential reduction in staff time associated with automated library preparation, using condition. Median age was 61.3 years in the control group and 60.3 years in the in-
currently available technology. Results: We identified that TBG requires more staff time tervention group, and 94% and 95% were white, respectively. Those receiving treatment
(V534 vs V330), capital investment (V326 vs V16), and consumables (V1544 vs V788), during the control condition, on average, completed 3.53 cycles of docetaxel, while those
resulting in an overall increased cost, compared with LBG (V2404 vs V1135). Automation receiving TC during the intervention condition completed 3.77 cycles (p = 0.014). Seventy-
of library preparation would reduce staff time required for LBG (Reduced to V191; 33% eight percent in the control group completed at least 4 cycles of docetaxel, compared to
reduction) with less of an impact on TBG (Reduced to V485; 10% reduction). This 84% in the intervention group. Those receiving TC during the control condition completed
difference was due to the increased wet-lab time with LBG and greater staff time for an average of 3.71 cycles of cyclophosphamide, compared to 3.78 cycles in the inter-
sample acquisition in TBG vs LBG (V298 vs V8). Finally, in the PLAN study, LBG resulted vention condition (p = 0.231). Eighty-four percent of patients in the control group and 86%
in cancellation of 12 repeat tissue biopsies, resulting in further savings. in the intervention group completed at least 4 cycles of cyclophosphamide.
Conclusions: LBG is a cheaper alternative to TBG. Our data indicates LBG saves cost in Conclusions: Routine PRO surveillance, coupled with guideline-based collaborative care
the areas of healthcare staffing and capital infrastructure with further savings made interventions, was associated with completion of a greater number of cycles of docetaxel.
through avoidance of repeat tissue biopsies. Thus, the resources required for LBG and These findings suggest that routine symptom surveillance and management may enhance
TBG are different and should be considered in service planning for tumour types such as the docetaxel tolerance profile and improve treatment adherence in patients with early-
NSCLC in which genotyping is standard-of-care. Clinical trial information: NCT05542485. stage breast cancer, allowing for delivery of an optimized treatment course. Research
Research Sponsor: AstraZeneca; Amgen; Novartis; Irish Cancer Society; Charitable In- Sponsor: NCI of the National Institutes of Health; UM1CA233033 (PI Cheville, Mayo Clinic,
firmary Charitable Trust. Rochester, MN).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 83s

1544 Poster Session 1545 Poster Session

Successful accrual of a cluster randomized controlled trial (RCT) comparing Electronic patient-reported outcomes with vital sign monitoring versus
an educationally enhanced genomic tumor board (EGTB) intervention to usual care during trastuzumab deruxtecan treatment for metastatic breast
usual practice (S2108CD, NCT# 05455606). First Author: Meghna S. Trivedi, cancer: Updated results from the PRO-DUCE study. First Author: Yuichiro
Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New Kikawa, Department of Breast Surgery, Kansai Medical University Hospital, Hirakata,
York, NY Japan
Background: Observational studies show genomic tumor boards (GTBs) can enhance clinician Background: The PRO-DUCE study (jRCTs031200387), a multicenter, randomized controlled
knowledge and application of genomic tumor test (GTT) results. An RCT is needed to evaluate the study, evaluated the impact of ePRO plus body temperature (BT)/SpO2 monitoring (ePROm)
impact of GTBs on treatment and outcomes. Conducting such a trial requires physician engagement and [Link] care (UC) on the quality of life (QoL) for patients (pts) with HER2-positive metastatic
recruitment of an unbiased study population over a short period of time due to rapid changes in cancer breast cancer (MBC) treated with T-DXd. The primary results presented at ASCO 2024 showed
genomics. S2108CD prospectively evaluates the impact of an educationally enhanced GTB (EGTB) that at week 24, ePROm demonstrated better results for global health status (GHS) vs. UC.
intervention through a cluster RCT design across rural and community oncology practices in the United
Based on recent studies showing that symptom monitoring and alert notifications via ePROs
States. Methods: Recruitment Centers (RCs) were selected from the NCI Community Oncology
Research Program (NCORP) with a focus on rural and minority/underserved (RMU) sites and cluster could improve pt QoL and overall survival (OS), we consider the effects of long-term QoL scores
randomized 1:1 to EGTB intervention (virtual GTB and physician educational materials) versus usual and OS. Methods: The pts population was observed from March 1, 2021, to February 29, 2024
practice (UP). Oncologists and their patients with advanced solid tumors who had GTT ordered were (data cutoff date: May 20, 2024). We randomized pts with HER2-positive MBC eligible for T-DXd
enrolled. The primary aim is to compare the proportion of patients receiving evidence-based genome to the ePROm or UC group. ePROm involved weekly reporting of symptoms based on PRO-
informed therapy by arm. Here, we describe accrual and baseline characteristics of the RCs and the CTCAE and daily monitoring of BT and SpO2 reports via a smartphone at home. If any reported
enrolled physicians and patients. Results: Between 8/2022 and 11/2024, 18 RCs were randomized to symptoms exceeded the predetermined thresholds, an email alert was sent to the medical
UP versus EGTB intervention. A median of 5 clinics (range, 1-13) comprised each RC. Six RCs were staff; the ePRO data were then reviewed, and, if necessary, a phone consultation was provided.
classified as RMU in each arm (n = 12). Overall, 121 physicians registered to the study and they Endpoints in this updated analysis included QoL scores and cancer-related fatigue beyond
registered 1284 patients (median 47.5 patients/month), of which 983 (77%) were registered at RMU week 24 using the EORTC QLQ-C30 (C30) and EORTC QLQ-FA12 (FA12), OS, and safety.
RCs. In the UP arm, 614 patients and 61 physicians registered, and in the EGTB arm, 670 patients and 60 Results: Between March 2021 and January 2023, 111 pts were enrolled (ePROm: 56; UC: 55).
physicians registered (demographics of registered patients in table; unknown race/ethnicity [3%] not Throughout the observation period, 1045 alert notifications were generated, of which 279 were
shown), meeting accrual goal within target timeframe. The 3 most common patient neoplasms were
considered necessary to contact pts and 231 telephone counseling was conducted by
lung, mediastinal and pleural (n = 352, 27.4%), gastrointestinal (n = 344, 26.8%), and breast (n = 152,
11.8%). Conclusions: Timely accrual to S2108CD demonstrates the feasibility of conducting a cluster
healthcare providers (response rate: 82.8%). Throughout 48 weeks, the PRO-CTCAE response
RCT to evaluate an EGTB intervention. There was high participation of RMU sites with a study population rate was . 89% in the ePROm group. The number of evaluable pts at 48 weeks was 37 in the
reflecting cancer types relevant for GTT; however, there were fewer than anticipated non-White and ePROm and 41 in the UC group. At week 48, changes in GHS from baseline in the ePROm and
Hispanic patients registered. Study endpoint data are currently maturing. The final analysis of S2108CD UC groups were -3.7618.6 and -12.7623.8, and those in FA12 total score were -4.9614.3 and
will be conducted and reported in 2026. Clinical trial information: NCT05455606. Research Sponsor: 8.1616.4, respectively. The median OS was 24.5 months (95% CI: 21.8, not reached) in the
NIH/NCI/NCORP; UG1CA189974; The Hope Foundation for Cancer Research. ePROm group with a median follow-up duration of 18.3 months, whereas in the UC group, the
median OS was not reached (95% CI: 20.7, not reached) with a median follow-up duration of
Overall UP EGTB
Patient characteristic (N=1284, %) (N=614, %) (N=670, %) 18.1 months [HR = 1.39; 95% CI: 0.75, 2.59]. The majority of any-grade treatment-emergent
adverse event using CTCAE were reported in a higher proportion of pts in the ePROm group
Age, median (range) 67.7 (20.3, 96.7) 67.6 (20.3, 96.7) 67.7 (22.4, 94.2)
Sex Female 646 (50) 325 (53) 321 (48) than in the UC group during the observation period. Interstitial lung disease incidence was
Race American Indian or 10 (1) 4 (1) 6 (1) similar in both groups (7.4% in ePROm group vs. 9.3% in UC group), with all cases being grade
Alaska Native 1. Conclusions: Long-term findings from the PRO-DUCE study demonstrated that QoL ob-
Asian 56 (4) 51 (8) 5 (1)
Black or African American 88 (7) 26 (4) 62 (9) served at week 24 in the ePROm group was maintained over time. While there was no dif-
Native Hawaiian or other 23 (2) 23 (4) 0(0) ference in OS, these findings might support the integration of ePROm in clinical practice to
Pacific Islander optimize QoL for pts receiving T-DXd. Clinical trial information: jRCTs031200387. Research
White 1047 (82) 490 (80) 557 (83)
Multiracial 25 (2) 8 (1) 17 (3) Sponsor: Daiichi Sankyo.
Ethnicity Hispanic 106 (8) 25 (4) 81 (12)

1546 Poster Session 1547 Poster Session

Advancing equity and collaboration in a dedicated young onset cancer clinic: Automated conversational artificial intelligence (AI) for outpatient malig-
A prospective study. First Author: Ilit Turgeman, Lin Medical Center, Haifa, Israel nant bowel obstruction (MBO) symptom monitoring. First Author: Ainhoa
Background: Coordinated young onset cancer (YOC) clinics improve patient experience Madariaga, Hospital Universitario 12 de Octubre, Madrid, Spain
and access to supportive services. However, geographic and ethnic disparities hinder Background: MBO is a severe complication of advanced cancer. A Canadian ambulatory
equitable representation and outcomes. This study describes the design and impact of a MBO program with nurse-led proactive call management demonstrated reduced hospi-
YOC clinic serving a socioeconomically diverse population. Methods: A YOC clinic was talization rates and improved survival. To overcome resource limitations, a smartphone
established at a peripheral cancer center for patients aged 18-49, staffed with an app was developed, achieving 65% adherence. Building on this foundation, automated
oncologist, nurse, and social worker. Tailored referrals were facilitated to designated phone calls offer a promising approach to enhance adherence and improve symptom
psychosocial and integrative services. Monthly team meetings and biweekly patient monitoring. Methods: We conducted a prospective pilot study at a tertiary Spanish
support activities were implemented. Data on demographics, cancer characteristics, hospital to remotely monitor MBO signs and symptoms using a conversational AI-based
coping styles (BASIC-Ph model), anxiety and depression (HADS-A/D), referral patterns platform (Lola-Tucuvi). Patients (pts) with cancer with an active MBO or at risk of de-
and satisfaction were prospectively collected and analyzed across ethnic and clinical veloping it (per PMMBO criteria) were enrolled. Automated, interactive phone calls were
subgroups. Results: From 6/2022 - 4/2023, 104 patients enrolled (mean age 38; 76% performed by the platform (Lola) weekly or biweekly. Lola performed structured MBO
female; 62.4% Jewish, 34.7% Arab). Most were married (63.5%), had children (81.4%), symptom assessments utilizing advanced natural language processing and AI algorithms,
unemployed (72.3%), with up to high school education (50.7%). Most (76%) were on to analyze responses in real time. Alerts were generated for moderate or severe symptoms,
active treatment and 38% had metastatic disease. Breast cancer was most common which were flagged on a dashboard. Nurses contacted pts based on alerts. The primary
(46.2%), followed by gastrointestinal and thoracic malignancies. Family cancer history objective was feasibility measured by adherence (% of answered calls), with a hypoth-
(61.5%) actionable somatic (25%) and germline (6.7%) alterations were noted. Referrals esized adherence of $65% considered optimal. Results: From January 2024 to January
were most frequent for integrative medicine (50%), genetics (41.3%), psychology, oc- 2025, 54 pts were enrolled, with 25 still active at the time of analysis. Median age was 60
cupational therapy, with adherence rates of 88%. Ethnic and cancer-type variations in years (range 29-86), and 96% of pts are female. Type of tumors included gynecologic (87%)
referral patterns were observed, but high uptake was consistent. HADS-A correlated with and gastrointestinal (13%). All pts were on systemic therapy: chemotherapy (50%), im-
munotherapy (24%), ADC (15%), targeted (11%). Median prior lines of therapy were 2 (1-6),
lack of exercise and HADS-D with unemployment and non-metastatic disease (p ,
and 41% (22/54) of pts had an active MBO prior to enrollment. Lola performed 716 phone
0.05). Females tended more to belief-based coping, males favored imagination, aca-
calls and 645 were answered, with an adherence of 90%. This resulted in an estimated
demics employed cognitive strategies, alcohol use to affect and cannabis to imagination
183.2 hours of nursing call time saved. Median time on the program was 117 days (7-356),
(p , 0.05), while those with advanced cancer preferred physical coping (p = 0.82).
and pts received a median of 14 calls. Of answered calls, the 36% (234/645) generated
HADS-A correlated with psychology referrals (M = 9.78 vs. 6.86, p = 0.013), belief-based alerts, with 44% classified as severe. Most frequent severe and moderate alerts were
coping to spirituality, social coping to physiotherapy and nutrition (p , 0.05). Social constipation and abdominal pain, respectively. Nurses acted on 73% (171/234) of the
coping and sexuality services were less utilized in ethnic minorities and those on active alerts, providing interventions such as dietary modifications, medication adjustments,
treatment. Nearly all (94.2%) had no prior support group yet satisfaction was high (mean clinical or emergency assessments. During follow-up in the program 31.5% (17/54) of pts
5/5) with patients citing community, supportive services and side effect management as had $1 active MBO and 18.5% (10/54) required admissions for MBO. Feedback was
key benefits. Conclusions: A structured, multidisciplinary YOC clinic leverages existing received from 26 pts, indicating a high satisfaction (4.6/5), and 96% would recommend the
services to enhance collaboration, equity, and access to tailored resources, meeting the use of Lola. Conclusions: This conversational AI platform demonstrated excellent fea-
unique needs of diverse young cancer patients. Beyond addressing disparities, this sibility with 90% adherence, higher than prior app-based solutions. It effectively monitored
model fosters a sense of community among patients, promoting engagement with MBO symptoms, enabling timely clinical interventions and enhancing patient engagement.
supportive services that may enhance treatment outcomes. Sharing this approach These results highlight the potential of AI-driven remote monitoring system to improve
globally may inspire broader adoption of YOC clinics to benefit underserved populations outcomes in cancer care. Further validation through randomized studies is warranted.
and advance cancer care equity. Research Sponsor: None. Research Sponsor: Spanish Society of Medical Oncology (SEOM).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
84s CARE DELIVERY/MODELS OF CARE

1548 Poster Session 1549 Poster Session

Oncologists’ perspectives on challenges using chemotherapy, immune Improving access to cancer screening through national telehealth-based
checkpoint inhibitors, and targeted kinase inhibitors for metastatic lung and colorectal cancer screening programs. First Author: Deanna Brockman,
cancer. First Author: Christine M. Veenstra, University of Michigan, Ann Arbor, MI Color Health, Burlingame, CA
Background: Compared to cytotoxic chemotherapies, the standard of care for metastatic Background: Access to routine cancer screening remains a significant barrier to early
cancers for decades, use of immune checkpoint inhibitors (ICI) and targeted kinase inhibitors detection, especially among historically underserved populations. However, little is
(TKI) has rapidly expanded with evolving indications for patients with advanced disease. We known about how virtual care can improve access to cancer services. Color and the
evaluated oncologists’ report of challenges to using each type of therapy and associations American Cancer Society (ACS) collaborated to develop two community-based, tele-
between challenges and oncology practice resources. Methods: From 2023-24 we surveyed 824
medicine programs for national colorectal and lung cancer screening. Here we evaluate
medical oncologists, identified using SEER registry data, in Georgia and Los Angeles. We asked
oncologists about challenges to using chemotherapy, ICIs and TKIs for the treatment of met-
the impact of these programs in expanding access. Methods: The Colorectal Cancer
astatic cancer. We also asked about resources available at their practice, including administrative (CRC) Screening Program (launched in June 2025) provides at-home fecal immuno-
and clinical support staff, financial counselors, dedicated pharmacists, social workers, genetic chemical tests (FIT) to eligible individuals aged 45-75 years. Kits are distributed through
counselors, and interpreters. We generated descriptive statistics of challenges and assessed federally qualified health centers and other community locations. The Lung Cancer
bivariate associations between challenges and practice resources. Results: We present results Screening Program (launched in November 2023) offers eligible individuals access to
for a preliminary sample (N = 370). The Table shows the proportion of oncologists who endorsed scheduling support for low-dose screening CTs based on ACS guidelines. Both programs
each challenge. Compared to cytotoxic chemotherapy, oncologists were more likely to endorse leverage a virtual-first approach: patients provide health history information for a cancer
moderate to big challenges using ICIs and TKIs related to insurance approval/prior authorization, risk assessment through an online platform where they can also access educational
co-pay assistance and out-of-pocket costs, as well as keeping up with clinical guidelines and resources and schedule appointments with physicians to discuss cancer risk (note:
familiarity with dosing and side effects; symptom management was more likely to be a moderate screening eligibility is determined based on self-report). Care advocates provide per-
to big problem for chemotherapy and TKIs than for ICIs (all p , 0.01). Challenges with insurance
sonalized support, including step-by-step guidance on completing tests, navigating
approval/prior authorization, co-pay assistance, and out-of-pocket costs were associated with the
availability of administrative and clinical support staff, dedicated pharmacists, and social workers
screening guidelines, and coordinating follow-up care. Results: Across both cancer
(all p , 0.05). Challenges with symptom management were associated with the availability of screening programs (n = 548), participants were predominantly female (CRC: 62.0%;
clinical support staff and dedicated pharmacists (all p , 0.05). Conclusions: Oncologists en- lung: 62.9%) and had similar average ages (CRC: 50.3 years, range 18-90; lung: 53.9
dorsed more challenges using ICIs and TKIs compared to cytotoxic chemotherapy in the years, range 18-81). In the CRC screening program, 397 participants picked up or
treatment of metastatic cancer. TKIs were associated with the most challenges, including requested kits, with the highest demand in California (10.1%), Texas (9.1%), and Florida
problems keeping up with clinical guidelines and familiarity with dosing and side effects for nearly (7.1%). In total, 126 participants were eligible and activated a kit; participant ineligibility
20% of respondents. It is reassuring that many practice resources exist to help address these was largely due to already being up-to-date with screening (29.1%) or age (28.5%). Of the
challenges. Oncologists may benefit from more educational resources—particularly related to 94 completed tests, 4 (4.3%) were abnormal results. In the lung cancer screening
TKIs—in their practices. Research Sponsor: National Cancer Institute; CA251464. program, 71 participants (47%) across 28 states met ACS eligibility criteria; the
Oncologist-reported challenges. remaining participants were ineligible due to age (21.1%), less than 20-pack year
% reporting moderate to big smoking history (25.4%), no smoking history (25.4%), or a combination of age and
challenge smoking history (28.2%). A total of 17 participants (23.9%) subsequently completed a
Cytotoxic lung CT. Average time to appointment was 29 days (range 9-93), and all appointments
chemotherapy ICI TKI P were located within 11 miles of the participant’s preferred location. Clinically significant
Obtaining insurance approval/prior authorization 13 26 35 ,0.01 findings included Lung-RADS 3 or 4 nodules (29.4%, n = 5) and other incidental findings
Obtaining co-pay assistance for patients 19 27 43 ,0.01 requiring follow-up care. Conclusions: These results illustrate how targeted,
Out-of-pocket costs to patients 25 42 66 ,0.01
Symptom management 50 17 43 ,0.01
community-based approaches can bridge critical gaps in cancer screening by simpli-
Keeping up with clinical guidelines 7 14 19 ,0.01 fying logistics, reducing costs, and providing tailored support through virtual and
Familiarity with dosing and side effects 6 7 17 ,0.01 community-based solutions. Research Sponsor: None.

1550 Poster Session LBA1551 Poster Session

Current treatment patterns for early breast cancer among healthcare pro- Cancer Care Beyond Walls (CCBW): A randomized pragmatic trial of home-
fessionals and concordance with expert recommendations: Analysis of an based versus in-clinic cancer therapy administration. First Author:
online interactive decision support tool. First Author: Timothy Quill, Clinical Care Roxana Stefania Dronca, Mayo Clinic, Rochester, MN
Options, Reston, VA
Background: The treatment paradigm for HER2-negative early breast cancer (EBC) now
includes pembrolizumab and targeted therapies such as olaparib, abemaciclib, and most
recently, ribociclib in the adjuvant setting for eligible patients. Here, we assess current
intended treatment patterns among healthcare professionals (HCPs) for EBC and compare
them with those of experts using an online Interactive Decision Support Tool (IDST).
Methods: We developed an online IDST in July 2024 with input from 5 breast cancer
experts providing therapy recommendations for 12 unique patient case scenarios based on
presentation characteristics including disease subtype, disease burden, treatment history,
BRCA mutation status, and risk of recurrence. HCPs entered specific patient charac-
teristics to define a case along with their intended management for that case. The IDST
then showed each expert’s recommendation for that case scenario and asked the HCPs if
the recommendations affected their intended approach. Here, we report a comparison of
the expert recommendations and HCP-selected therapy for different EBC case scenarios. The full, final text of this abstract will be available at
Results: Between August 2024 and January 2025, 140 HCPs entered 182 cases. Among [Link] on the day of presentation and in the
the 138 HCPs who indicated their treatment plan, plans were concordant with experts for
59% of the cases. Of note, the 5 experts showed complete concordance in their treatment online supplement to the June 10, 2025, issue of the Journal
recommendations for all 12 unique case scenarios. Concordance with the experts was of Clinical Oncology.
higher among HCP treatment plans for cases of hormone receptor–positive (HR+)/HER2-
negative (HER2-) cases compared with triple-negative cases (65% vs 49%). High con-
cordance was seen for HR+/HER2- cases receiving adjuvant AI with BRCA WT and high risk
of recurrence per the monarchE trial criteria (86%; n = 37) with lower concordance for this
setting without a high risk of recurrence (65%, n = 26) with HCPs often choosing endocrine
therapy plus a CDK4/6 inhibitor. In the setting of TNBC, HCPs entered cases predominantly
related to adjuvant therapy after neoadjuvant chemotherapy plus pembrolizumab (n = 44).
Among cases without a pathologic CR in this setting (n = 29), concordance was 50% with or
54% without a pathologic germline BRCA variant, respectively. Among TNBC cases with a
pathologic CR in this setting (n = 15), concordance was 40% with overtreatment by HCPs
evident in 33% of cases. HCPs indicated that expert recommendations changed their
intended treatment in 28 of 84 (33%) cases and confirmed their choice in 45 of 84 cases
(53%). Conclusions: These data suggest ongoing challenges with incorporating pem-
brolizumab and the newest targeted therapies into adjuvant treatment plans for high-risk
EBC, particularly TNBC. Continued education and development of resources for HCPs,
including online IDSTs, may be increasingly important as the treatment of high-risk EBC
continues to evolve. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 85s

1552 Poster Session 1553 Poster Session

Evaluation of longitudinal image-derived AI prognostication as a predictor of Real-world side effects of targeted therapies: High-throughput association
overall survival (OS) in a phase 3 advanced non-small cell lung cancer studies leveraging the CancerLinq Discovery lung cancer database. First
(aNSCLC) trial. First Author: Javier Montalt-Tordera, Bayer, Sant Joan Despi, Spain Author: Joseph Vento, Department of Internal Medicine, Division of Hematology and
Background: Confidently anticipating an overall survival (OS) benefit in cancer care and Oncology, UT Southwestern, Dallas, TX
therapeutic research is a defining challenge. AI tools may offer longitudinal mea- Background: Targeted therapies have unique side effect profiles distinct from other cancer drugs. Chal-
surements that predict OS differences objectively from existing data. Longitudinal lenges of collecting generalizable toxicity data on these medications include the self-reporting infrastructure
of existing post-market surveillance databases, as well as the infrequent use of many of these drugs by single
imaging-based prognostication (IPRO-D), a fully automated deep learning system, was cancer centers. The CancerLinQ Discovery (CLQD) database synthesizes deidentified electronic health record
independently trained on real-world imaging data to predict survival from pairs of (EHR) data from millions of U.S. cancer patients. Diagnosis codes in this database can be used to study
longitudinal computed tomography (CT) scans. Methods: We retrospectively assessed treatment side effects. Methods: We developed high-precision phenotyping algorithms to identify non-small
and compared how IPRO-D and percent change in RECIST sum of longest diameters cell lung cancer (NSCLC) patients receiving targeted therapies in the CLQD database. We then performed
(DSLD) predicted OS from 165 pairs of baseline and week 13 CT scans acquired in NExUS phenome-wide association studies (PheWAS) comparing new diagnosis codes in patients receiving each
targeted therapy to new codes in patients receiving chemotherapy or immunotherapy. Codes with significant
(NCT00449033), a phase 3 randomized controlled trial evaluating chemotherapy in associations were compared to toxicity data reported in clinical trials and the FDA Adverse Event Reporting
combination with either sorafenib or placebo for first-line treatment of subjects with System (FAERS) database. Results: We identified 5,278 NSCLC patients who received targeted therapies
aNSCLC. The two study arms did not show a difference in OS and were combined for this with the latest CLQD data pull in 2022. For each of the 18 targeted therapies with five or more patients in the
analysis. To examine the association of IPRO-D and DSLD with OS, we measured the database, descriptive statistics and two PheWAS analyses are reported: one for diagnosis codes relative to
concordance index (c-index) and the standardized hazard ratios (HRs, change in risk chemotherapy, and one relative to immunotherapy. These analyses identified significant associations
corresponding to known toxicity profiles as well as potentially underreported side effects. Conclusions: This
for a one-standard-deviation increase in the marker). We also report median OS (mOS) high-throughput framework augments the characterization of side effect profiles for existing targeted
for patients with partial response (PR, n = 60), stable disease (SD, n = 83) and progressive therapies and can proactively monitor for toxicity signals as novel therapies and treatment indications
disease (PD, n = 22) at week 13, as defined by RECIST 1.1 guidelines (zero patients had a emerge. The importance of collecting real-world data across institutions is highlighted in the ability to find
complete response). To explore the stratification potential of IPRO-D, we also define clinically relevant associations even in targeted therapies directed against rare mutations. Research
equivalent strata by ordering patients by their IPRO-D score and maintaining the same Sponsor: National Library of Medicine.
proportions (e.g., the top 60 patients by IPRO-D would be IPRO-PR, while the bottom 22 PheWAS analysis sample for side effects of osimertinib and capmatinib.
patients would be IPRO-PD), and report the mOS for these strata. Results: For the Drug Control Side Effect Code p-value OR
combined trial arms, median OS was 10.9 months (95% CI: 8.6 – 13.6), 108 (65.4%) were Osimertinib IO Abnormal EKG 3.68E-14* 2.91
Dermatitis 1.91E-09* 1.98
male, and 145 (87.9%) were diagnosed as stage IV. Table 1 reports the c-index, HR and Disorders of muscle 8.18E-06* 2.13
stratified mOS values for both survival markers. Conclusions: At week 13 in NExUS, Other CNS disorders 2.41E-05* 2.17
Thrombosis 2.90E-05* 1.49
IPRO-D predicted OS differences significantly better than DSLD. Future work will explore Chemo Dermatitis 1.42E-62* 7.90
how IPRO-D could serve as the basis for a surrogate endpoint in aNSCLC trials. Skin symptoms 5.00E-45* 3.70
Abnormal EKG 3.95E-33* 4.93
Research Sponsor: None. Other musculoskeletal symptoms 5.61E-27* 6.14
Joint pains 1.09E-22* 2.43
Summary of association of IPRO-D and DSLD with OS. Capmatinib IO Edema 2.34E-08* 4.19
Other soft tissue disorders 9.93E-05* 3.81
IPRO-D @ DSLD @ Hypocalcemia 1.06E-02 3.82
Week 13 (95% CI) Week 13 (95% CI) p-value Pleural effusion 1.96E-02 2.22
Abnormal LFT results 4.97E-02 3.24
C-Index 0.654 (0.604 – 0.713) 0.543 (0.495 – 0.599) ,0.01 Chemo Edema 2.05E-14* 7.03
Other soft tissue disorders 1.75E-11* 10.10
HR (1SD) 1.72 (1.38 – 2.15) 1.14 (0.94 – 1.38) ,0.01 Respiratory failure 4.20E-08* 7.12
OS (months, PR / SD / 16.5 / 10.9 / 5.7 12.5 / 12.3 / 4.6 - Abnormal LFT results 3.12E-04* 8.70
PD or IPRO equivalent) Stroke 1.48E-03 6.68

IO-immunotherapy, OR- odds ratio.

*Statistically significant after Bonferroni correction.

1554 Poster Session 1555 Poster Session

Transforming oncology clinical trial matching through multi-agent AI and an Breath-based VOC analysis leveraging canine olfaction for multi-cancer
oncology-specific knowledge graph: A prospective evaluation in 3,800 detection: Insights from a 1000-sample study. First Author: Akash Kulgod,
patients. First Author: Arturo Loaiza-Bonilla, St. Luke’s University Health Network, Dognosis, Inc., Bangalore, India
Easton, PA Background: Volatile organic compound (VOC) analysis is a validated approach for
Background: Clinical trial enrollment in oncology is often hampered by the manual, time- identifying disease-specific metabolic alterations through exhaled breath. The non-
intensive process of matching patients to trials with highly specific eligibility criteria. invasive and low-cost nature of breath sample collection makes it particularly suitable
Advances in artificial intelligence (AI)—particularly multi-agent large language models for large-scale cancer screening in resource-limited settings, such as those commonly
(LLMs) and oncology-specific knowledge networks—hold promise for streamlining this found in the Global South. Canine olfaction has been demonstrated in prior controlled
workflow and minimizing human labor. This abstract presents a prospective evaluation of studies to detect VOCs with high accuracy across a range of pathologies, including
an AI platform that automates medical data extraction, leverages an oncology-specific malignancies. This study evaluates the performance of trained biomedical detection
knowledge graph, and provides real-time trial recommendations, demonstrating a sig- dogs in identifying multiple cancer types using VOC analysis and examines the inte-
nificant reduction in staff effort while maintaining high clinical accuracy. Methods: Multi- gration of neurobehavioral data to support real-world diagnostic applications.
Agent AI & Oncology Knowledge Graph 1. OncoAgents: Specialized LLMs (data extraction, Methods: A retrospective case-control study was conducted involving 1000 partici-
eligibility, trial matching), collaborating to outperform generic or zero-shot AI. 2. pants across three clinical sites in Hubli, India. Exhaled breath samples (n = 105 cancer-
OncoGraph: Domain-specific knowledge graph uniting patient data, molecular profiles, positive, n = 895 healthy controls) were collected using standardized protocols designed
and clinical guidelines for context-aware matching. 3. OncoRecommend: Real-time to maintain VOC integrity. Trained biomedical detection dogs analyzed these samples,
engine processing new data, trials, and guidelines, delivering rapid, relevant sugges- with their behavioral responses recorded via motion sensors, video data, and elec-
tions. 4. OncoSet: Expert-curated dataset (.2,000 patient records, 14,000+ trials, 50+ troencephalography (EEG) systems. A consensus-based decision framework was
tumor subtypes) ensuring robust AI performance. Prospective Analysis (Jan–Dec 2024): implemented to account for variability among individual dogs. Preliminary machine
Cohort: 3,804 patients (ECOG 0–2) with metastatic/progressing malignancies seeking learning models were trained using the recorded neurobehavioral data to evaluate their
trial options. Data Extraction: 157,367 pages (~86.5M tokens) processed for tumor type, potential for augmenting detection accuracy; however, these models remain in the
stage, treatment lines, and biomarkers. Trial Matching: Automated application of validation phase. Results: The detection system demonstrated a sensitivity of 96%
inclusion/exclusion criteria; oncologists validated AI-generated matches. Efficiency: and a specificity of 100% across multiple cancer types in the test set, including oral,
Manual matching for large cohorts can require thousands of hours; this AI approach breast, esophageal, and cervical cancers. Sensitivity for early-stage cancers was 85%.
condensed it to ~1 hour of expert review. Results: 1. Screening & Identification: 3,804 The consensus-based approach among dogs enhanced reliability and minimized indi-
patients screened; 23,912 trials identified; 17,912 confirmed after expert review. 2. Time- vidual variability. Preliminary analysis of neurobehavioral data indicates potential for
to-Recommendation: Under one week from screening to final recommendations via real- machine learning applications to refine diagnostic interpretation. Conclusions: Breath-
time AI prioritization. 3. Performance Metrics: Sensitivity (Recall): 0.8375, Specificity: based VOC analysis combined with canine olfaction demonstrates high accuracy in
0.8359, Precision: 0.8121, F1 Score: 0.8246. Demonstrates advantages over zero-shot or multi-cancer detection, including early-stage cancers. Its suitability for non-invasive and
frontier GPT-based models. 4. GPT Comparison: Extraction Accuracy: 80.29% vs up low-cost implementation, particularly in resource-constrained settings like the Global
to 63.15% (GPT-4o); Trial Matching Accuracy: 82.06% vs 47.00% (GPT-4o). South, highlights its potential for addressing disparities in cancer screening access.
Conclusions: This multi-agent AI platform, underpinned by an oncology-specific Future research will focus on validating machine learning models and comparing the
knowledge graph, significantly boosts efficiency and accuracy in oncological trial system’s performance with existing diagnostic standards to further support global
matching. By cutting manual workloads from thousands of hours to near-automated scalability and clinical adoption. Research Sponsor: None.
speeds, recommendations allow for just-in-time, decentralized and patient-centric trial
activation. Ongoing enhancements—such as deeper biomarker integration, expanded
knowledge graph coverage, and seamless EHR interoperability—promise further gains in
personalized oncology care. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
86s CARE DELIVERY/MODELS OF CARE

1556 Poster Session 1557 Poster Session

PRESCIENTai, an AI-based digital histopathological image signature for risk Computational pathology to predict docetaxel benefit for high-risk localized
of late distant recurrence and extended endocrine therapy (EET) benefit in prostate cancer in NRG/RTOG 0521 (NCT00288080). First Author: Sebastian R.
hormone receptor–positive breast cancer. First Author: Eleftherios P. Mamounas, Medina, Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and
NSABP and AdventHealth Cancer Institute, Orlando, FL Emory University, Atlanta, GA
Background: A subset of patients (pts) with hormone receptor-positive (HR+) breast Background: The benefit of adding docetaxel (DTX) to standard of care (SOC) for high-risk localized
cancer (BC) experiences late distant recurrence (DR) and is more likely to benefit from EET. prostate cancer remains debated. The NRG/RTOG 0521 randomized phase III trial demonstrated that
Clinical practice guidelines recommend use of genomic assays such as Breast Cancer docetaxel, when added to SOC—comprising radiotherapy (RT) and long-term androgen deprivation
therapy (ADT)—improved overall survival (OS). However, while RTOG 0521 demonstrated improved OS
Index (BCI) to identify these pts. We developed an updated AI-based digital histopath- with DTX, the observed improvement did not meet the predetermined threshold for clinical significance,
ological risk score model to predict risk of late DR and extended letrozole therapy (ELT) leaving the role of DTX intensification uncertain. Enhanced stratification methods are needed to identify
benefit in this population. Methods: The AI model, PRESCIENTai, was trained on eligible aggressive disease phenotypes and guide patient selection for adjuvant chemotherapy. This study aims
samples (N = 2,271) from the National Surgical Adjuvant Breast and Bowel Project to develop and validate a computational AI derived pathology image classifier (APIC) to quantify the
(NSABP) B-42 cohort, which randomized postmenopausal women with HR+ BC who were tumor-immune microenvironment from diagnostic biopsy specimens and predict DTX benefit in patients
disease-free after 5 yrs of endocrine therapy (aromatase inhibitor (AI) or tamoxifen from the NRG/RTOG 0521 trial. Methods: The study included patients with available high-quality biopsy
followed by AI) to either 5 yrs of letrozole or placebo. A transformer-based end-to-end deep images from the NRG/RTOG 0521 trial. Primary outcome was OS, median follow-up was 5.7 years. After
segmenting nuclei and identifying lymphocytes, we derived features that captured immune-tumor spatial
learning model predicted risk score from H&E whole-slide images (WSI) in conjunction with patterns and nuclear diversity in the tumor microenvironment to construct APIC. DTX benefit was
clinical information (age at randomization, surgery type, node status, prior use of ta- evaluated using Cox proportional hazards models with interaction terms, log-rank tests and Kaplan-Meier
moxifen, race, lowest bone mineral density T-score, HER2 status). CTransPath was used for analyses by comparing OS between treatment arms within APIC-stratified groups. Results: Among NRG/
feature extraction from WSI tiles. 5-fold cross validation was performed with data split into RTOG 0521 trial participants, 350 patients had evaluable quality biopsy slide images. Half of the SOC
training, validation, and test sets ([Link]). The risk score threshold was defined by the (RT+ADT) arm was used for training (84 patients), and 266 patients were used for validation (SOC: 85
50% quantile of the training set for each fold. Cox regression and Kaplan-Meier analysis patients, and SOC+DTX arm: 181 patients). DTX significantly improved OS in APIC-positive (n = 119, 45%)
evaluated late DR and ELT benefit in high- and low-risk pts. Results: Hazard ratio (HR) was patients (HR = 0.49, 95% CI: 0.26-0.92, p = 0.023) but not in APIC-negative (n = 147, 55%) patients (HR =
1.17, 95% CI: 0.59-2.3, p = 0.66). APIC-positive patients derived 22% 10-year OS benefit (95% CI: 1.7%-
computed for DR in low- vs. high-risk pts [HR = 0.198 (95% CI: 0.124, 0.317); p , 0.001],
41.6%) from DTX. The 10-year OS was 74% in the DTX arm compared to 52% with RT and ADT alone in the
with absolute difference of 7.61% in 10-yr DR (1.84% vs. 9.46%). High-risk pts experienced APIC-positive group. A significant interaction (p = 0.024) was observed between APIC status and
greater ELT benefit over placebo (HR = 0.622; 95% CI: 0.416–0.929; p = 0.02) than low-risk treatment. Conclusions: We validated APIC as a predictive biomarker for DTX benefit in high-risk lo-
pts (HR = 0.727; 95% CI: 0.305–1.733; p = 0.471), with 10-yr absolute benefit of 3.74% vs. calized prostate cancer patients from NRG/RTOG 0521, identifying a subset who achieved significant
0.66%. Even among node(+) pts, PRESCIENTai identified greater ELT benefit for high-risk survival improvement from treatment intensification – a benefit not reached in the unselected trial
pts (HR = 0.521; 95% CI: 0.329–0.827; p = 0.006) than low-risk pts (HR = 0.53; 95% CI: population. Further investigation is warranted to evaluate APIC’s predictive potential of DTX intensi-
0.048-5.905, p = 0.606), with 10-yr absolute benefit of 6.66% vs. 1.89%. ELT benefit was fication in metastatic disease settings. Research Sponsor: NCORP; UG1CA189867; NCI/NIH;
also observed for high-risk pts in other clinical subgroups such as age #60 years and prior R01CA202752-01A1; NCI/NIH; R01CA208236-01A1; NCI/NIH; R01CA216579-01A1; NCI/NIH;
R01CA220581-01A1; NCI/NIH; R01CA257612-01A1; NCI/NIH; 1U01CA239055-01; NCI/NIH;
tamoxifen. However, p-interaction for ELT benefit in high- vs. low-risk groups was not 1U01CA248226-01; NCI/NIH; 1U54CA254566-01; National Heart, Lung and Blood Institute;
significant for all pts (p = 0.791) or node(+) pts (p = 0.889). Conclusions: This novel digital 1R01HL15127701A1; National Heart, Lung and Blood Institute; R01HL15807101A1; NRG Oncology
signature predicts risk of late DR in pts with HR+ BC. Although absolute ELT benefit was Operations; U10CA180868; National Institute of Biomedical Imaging and Bioengineering; 1R43EB028736-
greater in high- vs. low-risk pts, the treatment by risk score interaction was not statistically 01; VA Merit Review Award; IBX004121A; Breast Cancer Research Program; W81XWH-19-1-0668;
significant. This is, to our knowledge, the first AI model to predict long-term outcomes in Prostate Cancer Research Program; W81XWH-20-1-0851; Lung Cancer Research Program; W81XWH-18-
pts with HR+, early BC using a single slide image and clinical information. Successful 1-0440; Peer Reviewed Cancer Research Program; W81XWH-20-1-0595; Peer Reviewed Cancer Research
validation in additional pt cohorts will confirm the clinical utility of PRESCIENTai for Program; W81XWH-18-1-0404; Peer Reviewed Cancer Research Program; W81XWH-21-1-0345; Peer
Reviewed Cancer Research Program; W81XWH-21-1-0160; NRG Oncology SDMC; U10CA180822; NRG
prediction of late DR risk and EET benefit. Research Sponsor: U.S. National Institutes of Specimen Bank; U24CA196067; National Cancer Institute (NCI) and Sanofi; NCI/NIH; R01CA268287A1;
Health; U10CA180868; U.S. National Institutes of Health; UG1CA189867; U.S. National NCI/NIH; U01CA269181; NCI/NIH; R01CA26820701A1; NCI/NIH; R01CA249992-01A1.
Institutes of Health; U24CA196067; U.S. National Institutes of Health; U10CA180822.

1558 Poster Session 1559 Poster Session

Use of a large language model (LLM) for pan-cancer automated detection of Association of deep learning CT response assessment and interpretable
anti-cancer therapy toxicities and translational toxicity research. First Author: components with overall survival in advanced NSCLC: Validation in a trial of
Ziad Bakouny, Memorial Sloan Kettering Cancer Center, New York, NY sasanlimab and a real-world dataset. First Author: Chiharu Sako, [Link], San
Background: Understanding why patients develop adverse events to anti-cancer therapies and Carlos, CA
predicting the occurrence of these toxicities has lagged behind tumor response biomarker de- Background: Identifying advanced non–small cell lung cancer (aNSCLC) patients who
velopment. This critical gap is primarily due to limited availability of large-scale curated toxicity derive long-term benefit from immune checkpoint inhibitors (ICIs) remains a significant
data. Here, we leverage advances in natural language processing (Jee J et al., Nature, 2024), pooled challenge. Radiomic analyses, particularly leveraging deep learning, hold promise for
clinical trial data, and associated germline sequencing to detect adverse event data and determine
improving prognostic accuracy beyond tumor size metrics. We developed serialCTRS, a
clinical and genomic correlates. Methods: We utilized the Llama 3.1 LLM to automatically an-
notate patient adverse event data for 5 of the most common anti-cancer therapy related adverse
novel biomarker using deep learning to quantify thoracic CT changes from baseline to
events (adrenal insufficiency, hyperthyroidism, hypothyroidism, colitis, and pneumonitis). To 3 months post-treatment, predicting overall survival (OS) in patients receiving PD-(L)1
validate LLM predictions at the patient-level, we used a pooled institutional dataset with gold inhibitors. Methods: SerialCTRS was previously trained and validated on a multi-
standard prospectively collected adverse event data from 1,754 patients with solid tumors across institutional Real-World Dataset (RWD) (training: 1,171 aNSCLC patients, 14,424 CT
675 individual clinical trials. We further validated the LLM predictions at the clinical note-level scans; validation: 612 patients; Sako et al. SITC, 2024). For this study, we retrospectively
using a subset of 100 manually curated notes. We evaluated note-level and patient-level pre- validated serialCTRS in two distinct cohorts of aNSCLC patients: (1) a clinical trial (N = 52)
dictions using sensitivity and specificity. Patient-level time-to-adverse event development pre- treated with the PD-1 inhibitor sasanlimab in the second or later line and (2) a fully blinded
dictions were evaluated using Pearson R2 coefficients. Common Terminology Criteria for Adverse RWD from Baylor Scott & White Health system (N = 147), an institution not used for
Events v 5.0 was used for toxicity definitions. Results: The patients’ average age (standard training. The pipeline—spanning image quality control, preprocessing, feature extraction,
deviation) was 61.6 (14.5) years and 836 (47.7%) were female. The most common cancers were and survival modeling—operated without manual annotations. To enhance interpret-
non-small cell lung cancer (N= 194, 11.1%), soft tissue sarcoma (N=171, 9.7%), breast cancer (N=
ability, we developed 3D submodels for prognostic signals related to (i) tumor burden, (ii)
155, 8.8%), and melanoma (N=129, 7.4%). 44 (2.5%) patients had adrenal insufficiency, 88 colitis
(5.0%), 253 hypothyroidism (14.4%), 66 hyperthyroidism (4.4%), and 146 pneumonitis (8.3%).
body composition, and (iii) lung vasculature. Predictive performance was compared to
Among 1258 patients with complete systemic therapy information available, 422 (33.5%) were RECIST 1.1 using concordance index (c-index) and ROC-AUC for 24-month OS (OS24
treated with immunotherapy and 563 (44.8%) with chemotherapy. The performance metrics for AUC). Results: SerialCTRS outperformed RECIST in OS prediction and remained a sig-
LLM predictions at the note and patient levels are summarized in the table. Conclusions: We nificant predictor after multivariate adjustments with other known predictors including
demonstrate the ability of an LLM to accurately annotate anti-cancer therapy toxicity data across a age, sex, PD-L1 TPS, and NLR across both validation cohorts. In the sasanlimab cohort,
large number of patients. This approach is scalable to other toxicities and promises to spur adverse serialCTRS achieved a c-index of 0.77, surpassing RECIST (0.72), with an OS24 AUC of
event research. Clinical and genomic correlates of anti-cancer therapy adverse events, using data 0.86 (95% CI: 0.74–0.98). In the Baylor cohort, serialCTRS demonstrated a c-index of 0.68
from all patients with solid tumors with MSK-IMPACT data, will also be presented at the meeting. vs. RECIST (0.62) and an OS24 AUC of 0.76 (0.67–0.86). Submodels targeting individual
Research Sponsor: National Cancer Institute; T32CA009512-35; National Cancer Institute; P30- components achieved c-indices of 0.65 (tumor burden), 0.61 (body composition), and 0.61
CA008748. (vasculature) in the sasanlimab cohort, and 0.63, 0.61, and 0.59, respectively, in the Baylor
Performance metrics for LLM model. cohort. Combining the submodels improved c-indices to 0.69 (sasanlimab) and 0.66
Toxicity
Note-level (N= 100 notes) Patient-level (N= 1,754 patients) (Baylor), demonstrating complementary signal among radiographic features.
Sensitivity Specificity Sensitivity Specificity R2 Conclusions: SerialCTRS outperformed RECIST 1.1 in predicting OS in independent
Adrenal insufficiency 100.0% 97.8% 97.7% 94.7% 98.2%
clinical trial and RWD datasets. Interpretable submodels highlighted the prognostic value
Colitis 66.7% 99.0% 94.3% 80.4% 89.2% of tumor burden, body composition, and vasculature changes. SerialCTRS offers a
Hyperthyroidism 57.1% 100.0% 74.0% 91.4% 98.7% promising tool for personalizing therapy and accelerating drug development in aNSCLC,
Hypothyroidism 100.0% 88.9% 88.1% 74.0% 96.1%
Pneumonitis 76.9% 97.7% 98.6% 70.1% 83.9%
with a fully automated pipeline for robust and scalable clinical use. Future work will focus
on larger, more diverse cohorts to validate utility in guiding precision oncology. Research
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 87s

1560 Poster Session 1561 Poster Session

Computational pathology to predict docetaxel benefit in patients with met- Frailty in motion: How Fitbit data reflects patients with cancers’ functional
astatic hormone-sensitive prostate cancer from the CHAARTED trial (ECOG- status in the All-of-Us database. First Author: Kenan Najjar, Carle Illinois College of
ACRIN E3805). First Author: Sebastian R. Medina, Wallace H. Coulter Department of Medicine, Champaign, IL
Biomedical Engineering at Georgia Tech and Emory University, Atlanta, GA Background: Frailty is a vital determinant of outcomes in adults with cancer, leading to
Background: The CHAARTED trial demonstrated the efficacy of early docetaxel (DTX) in combination with heightened toxicity, morbidity, and reduced survival. Despite its significance, compre-
androgen deprivation therapy (ADT) for metastatic hormone-sensitive prostate cancer (mHSPC), partic- hensive frailty assessment remains constrained by time, resources, and knowledge. Deficit-
ularly in patients (pts) with high volume (HV) disease. Volume of metastases has been shown to assist pts based frailty indices (FIs) require extensive surveys, limiting real-world use. Wearable
selection, however more precise biomarkers are needed to select pts that might benefit from early DTX,
especially in low volume (LV) metastasis. We aim to validate a computational AI pathology image classifier
devices (e.g., Fitbit) that track daily steps offer a straightforward measure of physical
(APIC), that quantifies the tumor-immune microenvironment from biopsy images to predict DTX benefit in function, yet their associations with FIs are uncertain in oncology. Leveraging the All of Us
pts from the CHAARTED trial. Methods: The study included a subset of pts from CHAARTED for whom high- Research Program (Controlled Tier Dataset Version 7), one of the most diverse NIH re-
quality biopsy images were available. Outcomes were defined as overall survival (OS) and time to castration sources, we examined the association between steps and the validated All of Us Frailty Index
resistance (CRPC). After segmenting nuclei and identifying lymphocytes, we derived features that captured (AoU-FI) [Wong et al. J Gerontol A Biol Sci Med Sci. 2023]. We hypothesized that higher step
immune-tumor spatial patterns and nuclear diversity to construct APIC and stratify patients into positive or counts would correspond to lower (fitter) AoU-FI scores. Methods: We identified
negative groups. DTX benefit was evaluated using Cox proportional hazards with interaction terms, log-rank adults $50 years old with cancer and 3–7 days of Fitbit data between 2017–2025. The AoU-
tests and Kaplan-Meier analyses by comparing endpoint estimates between treatment arms within APIC-
FI comprises 33 deficits encompassing lifestyle, comorbidities, overall health, and
stratified groups. Results: Among CHAARTED trial pts, we analyzed H&E images that met quality control
and excluded prostatectomies (N = 286/790, 36.2%). Half of the ADT arm was used for training (78 pts), and healthcare access. Frailty was categorized as Fit ( , 0.15), Pre-Frail (0.15–0.25), or Frail ( .
208 pts were used for validation (ADT = 77, DTX = 131). Among these, 118 pts (56%) were classified as 0.25). Scores were matched with wearable data 6 30 days; outliers ( , 300 or . 20,000
APIC-positive and 90 pts (44%) as APIC-negative. In APIC-positive, DTX significantly improved OS (HR = steps/day) and records missing $20% of FI items were excluded. Demographics (age, sex,
0.52 [95% CI: 0.31–0.85], p = 0.0075, interaction p , 0.05) and delayed time to CRPC (HR = 0.48 [95% CI: race/ethnicity) and cancer diagnoses were documented. Mean step counts were compared
0.33–0.71], p = 0.00019, interaction p , 0.05). APIC-positive pts who received DTX derived a 24.3% higher via t-tests. A linear regression assessed the step–frailty link, and a multivariable logistic
5-year OS and remained castration-sensitive 21.8% longer than those who received ADT alone. In APIC- model for discerning Fit vs Pre-Frail or Frail (age$65, male, . one cancer) employed
negative pts, no prolongation effects of OS or time to CRPC were observed from the addition of DTX. APIC
normalized step counts (per 1,000). Results: Among 361 participants (mean age 65.767.8;
was able to identify pts benefiting from DTX in the HV group for OS (HR = 0.43 [95%CI: 0.24-0.77], p =
0.0035), and in both HV (HR = 0.50 [95%CI: 0.31-0.79], p = 0.0027) and LV (HR = 0.42 [95%CI: 0.20-0.91], p = 66.8% female; 92.5% White; 97.5% non-Hispanic/Latino), 86.2% had one cancer (skin 55.7%,
0.023) groups for time to CRPC. While CHAARTED showed modest CRPC delay in LV pts and no clear OS breast 24.8%, prostate 9.9%). Frailty categories were 44.0% Fit, 40.4% Pre-Frail, and 15.5%
benefit, APIC identified a subset of LV pts who derived substantial CRPC delay from DTX. Conclusions: We Frail. Mean steps declined with increasing frailty (Fit vs. Pre-Frail 8,13663,251 vs.
validated APIC as a predictive biomarker for DTX benefit in mHSPC pts from CHAARTED. Notably, in 7,06763,598; p = 0.007; vs. Frail 5,48663,623; p , 0.0001). A linear regression confirmed
unselected pts with LV disease where the benefit of DTX is less, APIC identified an LV subset who derived an inverse link between frailty and steps. (slope = –7.38310⁻⁶; p , 0.0001). Adjusted
significant delayed progression to CRPC from adding DTX to ADT. This work, validated in the context of ADT analysis showed steps remained highly significant (OR = 1.14, 95% CI: 1.07–1.21, p ,
alone, warrants investigation alongside androgen receptor axis-targeted agents. Research Sponsor: NCI/ 0.0001), whereas gender, . one cancer, and age$65 were not. Conclusions: All-of-Us is
NIH; R01CA268287A1, U01CA269181, R01CA26820701A1, R01CA249992-01A1,R01CA202752-
01A1,R01CA208236-01A1,R01CA216579-01A1,R01CA220581-01A1,R01CA257612-01A1, 1U01CA239055- designed to represent many populations, and our sample showed older adults adopting
01, 1U01CA248226-01, 1U54CA254566-01; National Heart, Lung and Blood Institute; 1R01HL15127701A1, wearable technologies, highlighting their potential for clinical use. However, the sample was
R01HL15807101A1; National Institute of Biomedical Imaging and Bioengineering; 1R43EB028736-01; primarily White individuals with skin cancer, reflecting coverage gaps in wearable data
United States Department of Veterans Affairs; IBX004121A; Biomedical Laboratory Research and De- across the cancer population. Step counts strongly correlated with fitness, supporting
velopment Service the Office of the Assistant Secretary of Defense for Health Affairs, through the Breast wearable-based functional assessment in oncology. Chronological age did not correlate with
Cancer Research Program; W81XWH-19-1-0668; Biomedical Laboratory Research and Development frailty, aligning with guidelines prioritizing holistic assessments. Future efforts should
Service the Office of the Assistant Secretary of Defense for Health Affairs, through the Prostate Cancer integrate broader populations, minute-level metrics, and cancer and treatment information
Research Program; W81XWH-20-1-0851; Lung Cancer Research Program; W81XWH-18-1-0440, W81XWH-
20-1-0595; Peer Reviewed Cancer Research Program; W81XWH-18-1-0404,W81XWH-21-1-0345, W81XWH-
to refine frailty classification as recruitment grows. Research Sponsor: “The All of Us
21-1-0160; Kidney Precision Medicine Project; Sanofi. Research Program is supported by the National Institutes of Health, Office of the Director.
The program would not be possible without the partnership of its participants.”

1562 Poster Session 1563 Poster Session

DeepSeal: Empowering clinical researchers to analyze clinicogenomic data Virtual oncology collaborative tumor board using multiple artificial intelli-
with an intuitive chat-based interface. First Author: Gaurav Sharma, Ocean Ge- gence agents. First Author: Jiasheng Wang, Comprehensive Cancer Center & James
nomics, Pittsburgh, PA Solove Research Inst., The Ohio State University Medical Center, Columbus, OH
Background: Traditional clinicogenomic analysis workflows in oncology require sub- Background: Clinical guidelines are complex documents with tables and figures. Finding
stantial bioinformatics expertise and custom coding, creating a bottleneck where clinical specific answers to questions within these guidelines can be challenging and time-
researchers must rely on analysts for data interpretation. This dependency hinders consuming. Current AI tools often struggle to extract information from these PDF
hypothesis development, exploration and discovery, as researchers cannot directly guidelines. To address this, we developed a novel system that uses a group of artificial
interact with their data in real-time. DeepSeal addresses these challenges by providing a intelligence (AI) agents, where each agent is designed to perform a distinct job, like
user-friendly, chat-based interface that enables immediate analysis of clinicogenomic finding information, reading documents, or summarizing findings. These agents com-
data and seamless generation of results, empowering clinical researchers to inde- municate with each other to analyze complex clinical guidelines, working as a team to
pendently explore and validate hypotheses. Methods: DeepSeal, an integration of a answer physician questions like trained oncologists discussing cases in a tumor board
large language model with clinical and molecular databases and bioinformatics tools, environment. Methods: Publicly available PDF guidelines published by the ASCO from
was evaluated by replicating the findings of Riaz et al. (Cell, 2017), in advanced Jan 2021 to Dec 2024 were acquired. A three-agent framework was constructed using the
melanoma patients treated with nivolumab. Through natural language prompts, AutoGen platform, comprising a Coordinator Agent, a PDF Viewer Agent, and a Reviewer
DeepSeal performed multiple analyses including differential gene expression analysis Agent. The Coordinator Agent selects the appropriate guideline based on a user’s
and Gene Set Enrichment Analysis (GSEA) to generate comprehensive molecular profiles question; the PDF Viewer Agent extracts information from the selected guideline file, and
of responders versus non-responders and to identify molecular signatures associated the Reviewer Agent generates a summary of the findings answering the original question.
with treatment response. Results: DeepSeal successfully replicated the key findings of The agents were powered by Anthropic’s Claude 3.5 Sonnet. The primary objective of the
Riaz et al., identifying significant differential expression of immune-related genes in study was to evaluate the platform’s accuracy in selecting the relevant guideline based on
treatment responders. GSEA executed through DeepSeal’s chat interface further user questions and in subsequently answering those questions accurately. Results: A
revealed enrichment of immune-related pathways critical for response, including B cell total of 34 ASCO guidelines were obtained, covering a range of cancer types: breast (15),
activation (GO:0042113), T cell activation (GO:0042110), and regulation of adaptive GI (4), head and neck (4), thoracic (4), neuro-oncology (3), GU (2), melanoma (1), and
immune response (GO:0002819). The chat interface enabled rapid hypothesis testing gynecologic (1). One hundred question-answer pairs were created by board-certified
and visualization generation, with analyses completed in minutes without the need for oncologists based on these guidelines to evaluate the system’s performance. It’s im-
programming expertise. Conclusions: DeepSeal demonstrates the feasibility of en- portant to note that these answers were based directly on the information in the
abling clinical researchers to independently analyze complex clinicogenomic data guidelines and may not always reflect the most current clinical knowledge, thus serving
through natural language interaction. By successfully replicating and validating findings as a rigorous test of the framework’s ability to adhere to the provided documents. Our
from Riaz et al., it generates reliable insights without programming expertise, offering a multi-agent framework achieved a 93% accuracy rate in matching user questions with the
transformative approach to accelerate translational research. This removal of technical correct guideline and answered 88% of the questions accurately. Comparatively, when the
barriers between researchers and their data has the potential to substantially speed same questions were evaluated using OpenAI’s GPT-4o (ChatGPT) and Claude 3.5 Sonnet
hypothesis testing and discovery in oncology, ultimately enhancing the pathway from without the multi-agent framework, the accuracy was significantly lower at 48% and 49%,
molecular insights to improved patient care. Research Sponsor: None. respectively. The total computational cost of processing all questions using the multi-
agent framework was 13.44 USD. The complete code, dataset, and detailed results are
publicly accessible at [Link]
Conclusions: This study demonstrates that a collaborative AI agent system can ac-
curately provide answers from clinical guidelines that is more accurate than ChatGPT and
similar software. Our results suggest a promising way to develop more effective AI tools
for clinicians to use in their practice. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
88s CARE DELIVERY/MODELS OF CARE

1564 Poster Session 1565 Poster Session

Comparison of artificial intelligence to expert physician assessments of Evaluating fairness and mitigating bias in models predicting financial tox-
real-world oncology cases. First Author: Olivia Main, NYU Perlmutter Cancer Center, icity among patients with genitourinary cancers. First Author: Atulya Aman
NYU Grossman Long Island School of Medicine, Mineola, NY Khosla, Department of Internal Medicine, William Beaumont University Hospital, Royal
Background: Artificial intelligence (AI) is increasingly being incorporated into the Oak, MI
oncology field as a tool to support clinical decisions. AI tools such as ChatgGPT or Background: Financial toxicity, the economic burden patients face from healthcare
OpenEvidence provide responses to user-generated queries, whereas some institutions expenses, is a growing concern in cancer care. Recognizing the high costs of diagnosis
or companies such as Primum, Inc, offer consultations with actual experts who provide and treatment of genitourinary (GU) cancers, this study aims to (1) comprehensively
personalized responses to clinician-submitted real-world cases. However, the value of AI characterize the socioeconomic, demographic, and care-related factors associated with
tools to augment expert consultations continues to evolve. We report results of a study financial toxicity in patients with GU cancers, and (2) evaluate bias in the predictive model
comparing AI versus expert oncologists’ responses to 107 real-world hematology/ developed using these patient factors. Methods: The 2019–2022 Medical Expenditure
oncology cases. Methods: Among 107 cases, inquiries included lymphomas (30), Panel Survey (MEPS) data was used to identify patients with GU cancers. MEPS captures
myeloma (24), leukemias (11), myeloid disorders (10), as well as classical hematology utilization, frequency, cost, and payment sources of U.S. health services alongside health
(32), assessed among 20 experts. Responses to de-identified cases submitted by insurance coverage characteristics and accessibility in the workforce. Financial toxicity
practicing clinicians to Primum ([Link]) between June 2022-July 2023 were was defined as patient-reported difficulties paying medical bills, high out-of-pocket
compared to GPT-4 responses ([Link]/chatgpt). The instructional prompt to GPT-4 expenses ( . 10% of total income), and high self-pay ratios ( . 20% of total healthcare
was, "You are an expert oncologist conversing with another oncologist as a peer. You expenditure). Predictive modeling was performed using logistic regression using age, sex,
prefer to rely on guidelines and data published in reputable medical journals when race/ethnicity, income, insurance status, and expenditure-related predictors. To address
responding.” Five expert faculty at our institution adjudicated the blinded comparative potential algorithmic bias, Fairlearn’s ThresholdOptimizer, a postprocessing algorithm,
responses, including their preference, quality and practical value scores, and prediction was applied to this predictive model, adjusting predictions to ensure equalized odds
of which response was AI generated. Comparison of scores was by t-test to generate P- across racial groups. Performance metrics, including accuracy, precision, and recall, were
values between expert and AI groups, and Pearson correlation was used for comparisons evaluated overall and by racial group. Results: Overall, we identified 1131 patients with
between adjudication scores. Results: Expert responses were preferred by . 50% of GU cancers (weighted n = 11,723,024) in the MEPS data; median age 72 yrs; sex 93.4%
adjudicators in 75% of cases (deviation 625%). Randomized AI responses were cor- male; 71.6% White, 18.3% Black, 6.5% Hispanic, and 3.5% Other. 22.2% of patients re-
rectly identified 90% of the time. Mean expert vs AI scores (Likert scale 0-4) for quality ported financial toxicity with a median [Q1, Q3] total healthcare expenditure of $2,645.0
(2.0 vs 2.1, P = 0.9) and practical value (2.1 vs 2.1, P = 0.9) were equivalent. Interestingly, [$898.5, $5328.0] vs. $503.5 [$171.0, $1286.8]. Logistic regression achieved an overall
AI responses were preferred in 46% (n = 15) of classical hematology and 31% (n = 9) of accuracy of 95%, with a precision of 97% and recall of 77% for financial toxicity cases.
lymphoma cases, largely due to being more concise. However there was no concordance Fairness metrics of the unadjusted predictions revealed bias to specific communities with
between high practical value scores and disease subtype for either group. lower recall for Black (46.2%) and Other Races (33.3%) compared to Hispanic (75.0%) and
Conclusions: : Expert physician responses were preferred over AI responses for most of White (90.4%) patients. After threshold optimization, recall improved to 61.5% for Black
the cases based on the level of detail presented, suggesting an implicit value of and 50% for Other Races, while Hispanic (84.6%) and White (100%) patients maintained
personalized responses compared to AI. Results showed no significant differences in high performance. However, disparities persisted, as evidenced by an equalized odds
quality or practical utility between AI generated responses and those from experts, difference of 0.21. Conclusions: This study underscores the critical need for responsible
reflecting a similarity in the information extracted from standardized guidelines, and development of predictive models impacting cancer care. Our findings show that a bias-
potentially adding value of AI in supporting clinical decision making. Our findings are correcting postprocessing algorithm can be an essential tool since it can be applied to
limited by the broad coverage of hematologic conditions for which experts and existing models without requiring retraining; however, these algorithms do not represent a
guidelines vary. Overall, these data suggest that while AI can supplement knowledge of definitive solution since this model’s underlying bias persists, highlighting the need to
management paradigms by providing basic management strategies, at present it cannot ensure models learn from fair data sets that are representative of the US population.
replace personalized expert consultation in clinical practice. Research Sponsor: None. Research Sponsor: None.

1566 Poster Session 1567 Poster Session

Machine learning model to forecast patient availability for oncology clinical Magnetic resonance imaging (MRI) radiomics as predictor of clinical out-
trials. First Author: Rajeev Kulkarni, ConcertAI, Cambridge, MA comes to neoadjuvant immunotherapy in patients with muscle invasive
Background: Automating eligibility criteria assessment for oncology clinical trials is an bladder cancer undergoing radical cystectomy. First Author: Andrea Necchi,
emerging application of machine learning (ML). However, machine learning applications IRCCS San Raffaele Hospital, Vita-Salute San Raffaele University, Milan, Italy
to predict patient availability – the likelihood of a patient beginning a new treatment (time Background: Muscle-invasive bladder carcinoma (MIBC) is a deadly disease, for which
to next treatment) in a prespecified time window – are not well described. We used a large we pioneered the use of neoadjuvant immune-checkpoint inhibitors (ICI) in a clinical trial
clinicogenomic database of patients diagnosed with solid cancer indications to train ML (PURE-01, NCT02736266) testing 3 cycles of neoadjuvant pembrolizumab before radical
model to predict patient availability for clinical trials. Methods: This was a retrospective cystectomy (RC). The objective of this study is to assess the ability of radiomic features
study based on data drawn from the ConcertAI Oncology Research database, enriched by extracted from a robust MRI processing pipeline to predict the pathological response to
key variables derived from unstructured data. Line of therapy was derived from expert neoadjuvant pembrolizumab. Methods: A total of 120 patients (pts) with MIBC (102M/
rules applied to structured medications data. Our cohort consisted of patients with 18 F), with median age of 68 years, a clinical stage T2N0 (n = 53; 44%) or T3-4N0 (n = 67;
confirmed diagnosis of solid cancers without a second malignancy. Patient follow-up 56%), who were enrolled in PURE-01 study were analyzed. Patients had matched pre-
period started on the date of diagnosis of metastasis and ended on the earlier of last date and post-ICI MRIs, and tumors were segmented on both T2w images by GU radiologists.
of activity / date of death. Random observation date was set between start and end dates The MRI signal intensities were standardized by N4-bias field correction and robust z-
to label patients. Patients administered a new treatment after the random observation scores. IBSI-compatible pyCERR software was used to extract radiomics features. A
date were labelled evet, else censored (no new treatment began). Label date is start of total of 289 radiomic features, including shape, first-order statistics, and higher-order
new treatment and end dates for event & censored cases respectively. The time to event textures, were analyzed for associations with pathological complete response (pCR at
(TTE) was defined as the duration between the random observation and the label dates. In RC). An additional association was also investigated for major response groups, i.e., CR
the event cases, this duration is the time to next treatment (TTNT). Over 2000 features and partial response (PR, i.e. downstaging to ypT#1N0) versus no response (NR). We
based on variables broadly grouped as tumor-specific biomarkers (PTEN, KRAS, etc.), employed Elastic Net, a machine learning technique that blends the strengths of Lasso
ECOG, staging, disease status, medications, and imaging (evidence of image, not report) and Ridge regression and is particularly effective for datasets with many correlated
were employed to build multiple ML models. Temporal validation of the models was features such as in our study. The endpoint was modeled by training Elastic Net logistic
performed by setting up a simulated index date and predicting the probability of patient regression models separately for pre- and post-ICI MRI features, as well as clinical T-
beginning a new treatment within 60 days of the simulated index date. Patients receiving stage. Models were evaluated on a 30% held-out test set using ROC curves (AUC).
new treatment within the 60 days were true positives. Results: TTE models were trained Results: For pCR, the best-performing model included four post-ICI MRI features: shape
on a cohort comprised of 90K patients across 12 cancer indications with 54% patients (flatness) and texture features from Gray Level Co-occurrence Matrix (GLCM: homo-
starting a new treatment. Median age and overall survival (OS) of the cohort was 73 years geneity, sum average, and sum entropy), and had a test AUC of 0.83 (95%CI: 0.66 - 0.99).
and 703 days respectively. Temporal validation was performed on 25K patients with Separate models fit on pre-ICI MRI features selected two important pre-ICI MRI features:
similar demographics/OS and 58% patients starting new treatment. Multiple ML methods shape (surface-to-volume ratio) and first order (robust mean absolute deviation), but the
were used to train models, with boosted gradient model demonstrating highest c-index of overall performance was lower than post-ICI models (test AUC 0.66; 95%CI: 0.42 - 0.89).
0.73 based on 87 features. Temporal validation demonstrated AUC and weighted F1 of For major response assessment, the best-performing model included two post-ICI MRI
87% and 67% respectively. True positive cases were assigned high predicted probability in features: shape (flatness) and texture (GLCM sum average) and had a test AUC of 0.92
75% of the cases. Conclusions: AI models supporting 12 solid cancer indications ac- (95%CI: 0.8-1.0). Conclusions: This is one of the first machine learning models using
curately predicted patient availability. These models can be integrated into real-time MRI radiomics to predict neoadjuvant immunotherapy response in pts with MIBC. These
clinical workflows alongside patient eligibility models to provide clinicians and patients results could be instrumental for improving the way we can predict the pathological
visibility in ascertaining a patient’s likelihood of being eligible for a clinical trial. Research response in these pts. Clinical trial information: NCT02736266. Research Sponsor:
Sponsor: None. Associazione Italiana per la Ricerca sul Cancro (AIRC); IG 27746.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 89s

1568 Poster Session 1569 Poster Session

Acoustic biomarkers and AI: Transforming NSCLC detection and personal- Real-time AI-based computer-aided detection/diagnosis (AI-CAD) for breast
ized care. First Author: Chiara Giangregorio, Department of Electronics, Information ultrasound: A prospective, multicenter, multinational study. First Author:
and Bioengineering, Politecnico di Milano, Milan, Italy Jeeyeon Elizabeth Lee, Department of Surgery, Kyungpook National University Chilgok
Background: Implementing a mass screening program for lung cancer using low-dose Hospital, School of Medicine, Kyungpook National University, Daegu, South Korea
chest CT presents significant challenges, including financial constraints and concerns Background: To evaluate the effectiveness of a real-time AI-based computer-aided
about radiation exposure. Nonetheless, recent evidence reveals that lung cancer is not detection/diagnosis (AI-CAD) system as a diagnostic decision support tool for breast
limited to smokers, as it also affects non-smoker populations who are currently excluded ultrasound in a real-world clinical setting, conducted as a prospective, multicenter, and
from existing screening programs. As part of the I3LUNG study (NCT05537922), we multinational study. Methods: From May to December 2024, a total of 75 patients
investigated the use of AI-based forced cough analysis as a non-invasive approach to undergoing breast ultrasound were enrolled in a prospective study conducted in Korea (n
distinguish NSCLC patients undergoing immunotherapy (IO) from healthy individuals. = 38) and Hong Kong (n = 37). In this study, six experts operated a real-time AI-CAD
Additionally, we examined whether cough features could differentiate patients based on system (CadAI-B, BeamWorks Inc., Korea) on a tablet PC connected to a handheld
their baseline clinical features. Methods: Machine Learning-based preprocessing ultrasound device during breast ultrasound examinations. Image and clinical data were
isolated meaningful cough events and extracted 39 acoustic features from the time and collected from patients with established ground truth through follow-up, biopsy, or
frequency domains. To reduce redundancy and improve model performance, highly surgery. The AI-CAD system highlights suspicious areas during scanning to assist
correlated features ( . 85%) were eliminated. Support Vector Machines (SVM) and Deep physicians in detecting breast cancer and supports big data-driven differential diagnosis
Learning (DL) models were then employed to distinguish NSCLC patients from healthy by providing BI-RADS categories and malignancy scores (0–100%) when the user
controls. Additional statistical analyses of acoustic features were conducted on cough freezes the image. The diagnostic performance of experts and the real-time AI-CAD
recordings from patients to evaluate differences based on smoking status (current, system was evaluated using the area under the receiver operating characteristic curve
former, or never smokers) using the Kruskal-Wallis test with Benjamini-Hochberg post- (AUC), along with sensitivity and specificity. Results: The analysis included 75 patients
hoc correction. Similarly, differences based on the presence or absence of lung me- (mean age 55 years, IQR 46–66) with 24 malignancies (32.0%), 45 benign lesions
tastases were assessed using the Mann-Whitney test. Results: A total of 200 individuals (60.0%), and 6 normal cases (8.0%). The mean breast mass size was 1.2 cm (61.0 cm):
were enrolled in the study, including 91 stage IIIB-IV NSCLC patients undergoing IO and benign 0.8 cm (60.7 cm), malignant 1.8 cm (61.3 cm). The BI-RADS category dis-
109 healthy controls. Cough recordings were analyzed, with the SVM model achieving an tribution was as follows: for experts—category 1 (4.0%), 2 (21.3%), 3 (24.0%), 4a (16.0%),
accuracy of 82% and a specificity of 92% on the test set. The DL model demonstrated 4b (18.7%), 4c (4.0%), 5 (12.0%); and for AI-CAD—category 1 (32.0%), 2 (5.3%), 3 (9.3%),
superior performance, with an accuracy of 95% and a specificity of 100%. Significant 4a (17.3%), 4b (21.3%), 4c (13.3%), 5 (1.3%). The overall diagnostic performance of
differences were observed in the peak-to-root-mean-square value ratio and cough experts and AI-CAD, as AUCs calculated by BI-RADS, were 0.801 and 0.751, respectively
duration among smokers (current, former, or never), with P-values of 0.026 and 0.042, (P = .679). The sensitivity and specificity were 91.7% (22/24) and 68.6% (35/51) in
respectively. Furthermore, spectral features - including centroid, rolloff, spread, kurtosis, experts and 87.5% (21/24) and 57.8% (32/51) in AI-CAD, respectively (P = .481).
bandwidth, and flatness - differed significantly between patients with and without lung Conclusions: In this real-world clinical study conducted across multiple centers and
metastases (P , 0.01). Conclusions: These findings highlight the potential of cough countries, CadAI-B demonstrated performance comparable to that of experts and
as a valuable digital biomarker for NSCLC diagnosis. The tool’s high sensitivity facilitates showed its potential as a valuable diagnostic tool. Clinical trial information:
the effective identification of individuals at risk for lung cancer, while its exceptional NCT06622967. Research Sponsor: None.
specificity makes it a promising initial screening method, efficiently triaging positive
cases for follow-up chest CT scans. Future studies should validate these results on
larger cohorts. Moreover, the correlation of specific cough features with smoking status
and the presence of lung metastases suggests that this tool could extend beyond
screening to monitoring disease progression over time. Research Sponsor: Hori-
zonEurope; Grant agreement ID: 101057695.

1570 Poster Session 1571 Poster Session

Cancer patients’ messages about radiology/pathology reports: Insights for Analysis of a large language model-based system versus manual review in
AI. First Author: Susan Chimonas, Memorial Sloan Kettering Cancer Center, New York, clinical data abstraction and deduction from real-world medical records of
NY patients with melanoma for clinical trial eligibility assessment. First Author:
Background: Cancer patients often use portals to view results prior to discussing with Christine Vecchio, Cleveland Clinic Taussig Cancer Institute, Cleveland, OH
physicians, leading to messages with questions or concerns.1 These messages vary Background: Manual chart review (MCR) is the gold standard for assessment of in-
widely in content and urgency, creating challenges for healthcare providers to respond formation from electronic medical records (EMRs) for clinical trial eligibility. However, this
effectively.2 Categorizing and triaging these messages through AI-enhanced tools could method is labor-intensive, prone to error, and limited in scalability with high volumes of
streamline communication and improve patient care and satisfaction. Methods: This unstructured EMR data. Large language models (LLMs), have shown promise in natural
study assessed common themes in 1 week (April 1-8, 2023) of patients’ portal messages language understanding, and automating chart review and abstraction would significantly
about “rapidly read” pathology and radiology reports (viewed by patients within 6 hours of improve efficiency and accuracy in data review for clinical research. In this evaluation
posting to the portal, as a proxy for viewing before discussing with physicians) at project, we compared the performance of Synapsis LLM, a medically-specialized LLM,
Memorial Sloan Kettering Cancer Center in New York City. Results: Five notable themes with medical professionals at answering questions tied to eligibility criteria of relevant
emerged across a total of 48 messages about rapidly read radiology and pathology clinical trials, by reading clinical notes of patients with melanoma. Methods: We
results: Interpretation (24/48, 50%): Half of the messages contained questions like, “What conducted a comparative analysis using records of randomly selected patients with
does this mean?” Patients sought explanations of pathology and radiology findings, melanoma from the Cleveland Clinic. Two parallel processes were assessed: (1) MCR
reflecting a need for clear, accessible interpretations. Implications (14/48, 29%): With conducted by a melanoma and an oncology specialized research nurse (2) Automated
questions like, "What are the next steps?” patients often asked how findings might alter chart review using the Synapsis LLM. Both processes ran on two cohorts: Cohort (A)
treatment plans, highlighting a need for guidance on the care implications of their reports. consisting of 25 EMRs that were posed 23 eligibility questions each, and Cohort (B)
Concern (5/48, 10%): Some patients expressed worry or pessimism about pathology and consisting of 25 different EMRs, posed 22 eligibility questions each. In total, there were
radiology reports: "I am very worried" and “Maybe it’s time to give up.” Such statements 1,125 questions answered by each of the research nurses as well as Synapsis AI. The
indicated a need for supportive communication. Relief (3/48, 6%): In other messages, questions addressed focused on melanoma-specific clinical characteristics, including but
patients shared positive emotions regarding favorable results – “It is a huge weight off my not limited to treatment approaches, related surgical procedures, imaging findings, and
mind.” These responses offer clinicians opportunities to reinforce patient satisfaction. genetic testing. Performance metrics included accuracy of answers to the questions, and
Errors/Omissions (3/48, 6%): Occasionally, patients perceived errors or omissions in their time required to complete the abstraction process. Discrepancies between the responses
reports – "The radiologist totally misread the size of the lesion" – which impacted their of the two research nurses and the LLM were analyzed in comparison to the established
trust in the information. Addressing these concerns promptly can help strengthen the ground truth, which was determined through a consensus review by physicians to ensure
patient-provider relationship. Conclusions: This novel study highlights opportunities for the validity and reliability of the results. Results: Synapsis LLM performed the task with
AI-enhanced tools to triage messages and facilitate timely, effective responses. This 95.73% accuracy in 2.5 minutes while the melanoma specialized nurse responded with
study found common themes in patients’ diverse questions about rapidly read pathology 95.11% accuracy in 427 minutes. The oncology specialized research nurse’s accuracy was
and radiology reports. By implementing AI to categorize and triage message patterns, 88.09%, and the tasks was completed in 540 min. The comparison demonstrated sig-
providers could support patients more efficiently. Methods in development could be used nificant time savings and medical-grade accuracy for the application of this LLM-based
to classify the message content.3 For instance, AI-driven natural language processing technology compared to manual methods. Conclusions: This is the first project that
tools could recognize queries related to "What does this mean?" and offer clear, ac- compares an LLM-based system vs research nurses in deducing clinical characteristics
cessible explanations of medical terms. Similarly, AI could be trained to flag high-priority from patients’ EMRs for clinical trial eligibility. Synapsis LLM accurately completed the
messages based on distress signals, ensuring that these messages are addressed swiftly. abstraction process, outperforming in accuracy and time the clinical personnel. This
Implementing such AI-based solutions could help meet patients’ immediate needs while study highlights the potential of LLMs like Synapsis AI in scalable clinical research
they await conversations with their providers. Research Sponsor: None. applications that currently rely solely on MCR. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
90s CARE DELIVERY/MODELS OF CARE

1572 Poster Session 1574 Poster Session

Leveraging AI to enhance symptom capture and reduced hospitalizations. The role of healthcare system distrust in shaping patients’ attitudes and
First Author: Arman Koul, Cancer Center, Stanford Healthcare, Stanford, CA beliefs of artificial intelligence (AI) use in oncology. First Author: Marco
Background: Unplanned hospitalizations and ED visits for cancer patients impose Santos Teles, Memorial Sloan Kettering Cancer Center, New York, NY
significant morbidity, financial costs, and reduced quality-of-life. Efficient resource Background: AI offers significant potential to improve cancer care, yet little is known
allocation in oncology care requires proactive strategies to identify patients at high risk about patients’ attitudes and beliefs around its use and which factors influence ac-
for preventable admissions, optimizing bed availability and outpatient management. ceptance of this new technology. Distrust in healthcare is increasingly prevalent and
CMS classifies acute care utilization (ACU), hospital stays or ED visits, within 30 days of may hinder patient perceptions of innovations such as AI. This study aimed to evaluate
chemotherapy for certain conditions, as “preventable” under OP-35, emphasizing the the relationship between healthcare system distrust and acceptance of AI in oncology.
need for better risk stratification. By leveraging artificial intelligence (AI) including Methods: We conducted a cross-sectional survey study with patients at an urban
machine learning and large language models (LLM), predictive models can analyze the academic cancer center. We developed an 8-item AI Patient Acceptance scale, where
electronic health record (EHR) to identify patients who may benefit from early outpatient patients rated their comfort with AI in different aspects of oncologic care (e.g. diagnosis,
interventions. This study leverages AI to assess preventable admissions and quantify the treatment planning) on a 5-point Likert scale (range 8-40, higher scores indicate greater
benefits of proactive management, enhancing patient outcomes and care efficiency in comfort; Cronbach’s a = 0.94). The survey also included a 10-item Health Care System
oncology. Methods: This study analyzed data from 18,187 patients across a multisite Distrust (HCSD) scale (range of 10-50, higher scores indicate greater distrust). Multiple
cancer center to develop predictive models for ACU within 30 days of any systemic linear regression was performed to evaluate the association between HCSD and AI
therapy following OP-35 criteria. Models incorporated structured data and clinical notes Patient Acceptance scores, adjusted for demographic and clinical factors. Results: Of
using LLMs. Using 2010–2019 data for training and 2020–2024 for validation, XGBoost 383 patients approached, 330 (86%) participated. Among these, 49.4% were age 65 or
and Random Forest models were developed to maximize sensitivity while maintaining older, 55.9% male, 68.1% non-Hispanic white, 77.4% had a college degree or more. The
acceptable specificity. Estimates of preventable hospital bed days were modeled to most common tumor types reported were prostate (34.5%) and breast (26.4%) cancer,
evaluate the impact of AI-driven risk stratification and targeted interventions on hospital with 70.6% currently receiving treatment. Patients were most comfortable with AI use in
resource utilization. Results: The 30-day hospital visit prediction model demonstrated cancer screening (80.2% somewhat or very comfortable), and supportive care appli-
strong performance, with the XGBoost algorithm achieving an AUROC of 0.84 (95% CI: cations, such as exercise (78.2%) and diet (74.8%). They were least comfortable with AI
0.83–0.86). Incorporating later therapy lines improved accuracy by accounting for the use to assist with diagnosis (70.4%) and other clinical decision-making applications,
complexity of advanced disease. A threshold was set to prioritize sensitivity for including treatment planning (64.8%) and prognosis (61.5%). Higher levels of distrust
identifying high-risk patients while maintaining specificity to minimize unnecessary measured by the HCSD scale were negatively associated with the AI Patient Acceptance
interventions. Model implementation was estimated to prevent 22% of hospital visits scale scores after adjusting for co-variates (B = -0.263, p = 0.002). Younger patients
when paired with timely intervention, saving 1,160 of the 5,370 bed days observed in the (age , 65) were more likely to report lower scores on the AI acceptance scale (B =
2021–2024 cohort. Conclusions: This study underscores the potential of AI-driven -1.996, p = 0.021), while sex, race/ethnicity, and education level were not associated with
prediction models to enhance precision oncology by identifying patients at risk for AI acceptance. Conclusions: Higher distrust in the healthcare system is associated with
unplanned hospital visits following systemic cancer therapy—a critical quality indicator lower acceptance of AI in cancer care. As we integrate new technologies like AI into
impacting patient outcomes, healthcare costs, and operational efficiency. By incor- oncology, mitigating distrust in the medical community will be essential to ensure
porating multi-line therapy data and leveraging advanced modeling techniques, the patient-centered implementation. Research Sponsor: None.
approach effectively captures disease progression and personalized treatment histories.
Utilizing LLMs to structure fragmented data across care systems addresses a prevalent
challenge in oncology. Accurate risk prediction of hospitalization facilitates proactive
interventions, improves care coordination, reduces bed occupancy, and supports in-
formed decision-making, ensuring timely and targeted support for high-risk patients.
Research Sponsor: None.

1575 Poster Session 1576 Poster Session

Cross-sectional study on the impact of receiving potentially sensitive test Utilization and impact of a digital care platform on cancer patients in India.
results online on the emotional health of patients with breast cancer. First First Author: Vamshi Krishna Muddu, AIG Hospitals, Hyderabad, India
Author: Anezka Carvalho Rubin de Celis Ferrari, Hospital Sirio Libanês, S~
ao Paulo, Brazil Background: Cancer patients experience complex symptoms and treatment-related side
Background: Internet has changed the way people communicate. In healthcare, pa- effects that impair quality of life (QoL) and increase healthcare utilization. Digital health
tients have now easy access to their test results online; however, this has been par- tools have the potential to improve symptom management and patient outcomes by
ticularly challenging in oncology because of potentially sensitive results present in facilitating real-time communication and support, especially for patients with limited
imaging and pathology reports and tumor markers. In breast cancer, survival rates have access to immediate care. This study aimed to assess the usability and impact of a digital
improved significantly and more people are getting screened regularly. Providing online cancer care platform (Alivius) among Indian cancer patients, their primary caregivers, and
access to test results to patients may have ambiguous outcomes, such as increasing care teams. Methods: We conducted a prospective, interventional, real-world study at a
their engagement, but also their anxiety levels. This study aims to investigate asso- single tertiary cancer center in India from Jan 30, 2024, to Jun 11, 2024. We enrolled 100
ciations between online access to potentially sensitive test results and effects on the cancer patients aged 18–80 years who had access to a smart phone, can read English,
emotional health of patients with breast cancer, particularly for symptoms of anxiety and had a confirmed cancer diagnosis and had undergone at least one cycle of chemotherapy
depression. Methods: We conducted a cross-sectional study with 385 patients who had or radiotherapy. Patients were asked to communicate their symptoms and concerns to
been diagnosed with breast cancer in the past 5 years. Participants completed a printed the care team and got health education materials for their disease. Primary endpoints
questionnaire on their approach to receiving test results (whether or not they accessed included the Monthly Active Users (MAU) and Net Promoter Score (NPS) a metric used to
the results online), their feelings of anxiety about the results, and validated ques- measure customer loyalty and satisfaction by asking customers how likely they are to
tionnaires assessing symptoms of anxiety (GAD-7) and depression (PHQ-9). Descriptive recommend a company or product. Secondary endpoints included the number of chat
data analysis was performed using simple and cross tabulations for qualitative variables interactions, patient-logged symptoms, and use of educational content. Data were
and measures such as mean, median, mode and standard deviation for quantitative collected through in-app user activity and user feedback surveys. Descriptive statistics
variables. The normality of the data was checked using the Shapiro-Wilk test, while summarized the data with chi-square test for categorical variables and t-tests or non-
associations between qualitative variables were assessed using the Pearson chi-square parametric tests for continuous variables. Results: Out of 100 registered patients (mean
test or Fisher’s exact test. The significance level was set at 5%, using SPSS software age 58.7 6 12.5 years; 61% male), the majority had gastrointestinal (55%) and geni-
version 22.0. Results: Partial results presented here include 329 patients, representing tourinary (18%) cancers. 66 caregivers participated in survey, with 84.8% being immediate
85.45% of the total sample. A statistically significant association was found between the family members. Average MAU during the study period was 58. The NPS for recom-
habit of accessing test results online and higher levels of anxiety (p , 0.05). Patients mending the hospital and the digital application was 30.9 and 32.4 respectively, indicating
with positive screening for anxiety on GAD-7 reported greater anxiety while waiting for positive user perception. User engagement was high during the study period: 85%
results (p , 0.05), while no such association was found for depression. In addition, accessed educational content, 82% updated health status, and 78% tracked mood. A total
online access to results showed a significant association with education level (p , 0.05) of 2,825 app activities were recorded, including 888 health status updates and 436 chat
and patient age (p , 0.05), with younger patients and those with a higher level of interactions. 507 side effects alerts generated, 29.5% were high severity, primarily related
education showing a greater propensity for this practice. However, no significant as- to tiredness and fatigue (29.2%), gastrointestinal issues (27.2%), and pain (17.5%).
sociations were found between tumor stage (early vs. metastatic) or time from diagnosis Customer Satisfaction Score (CSAT) was 50 and 48.5 among care teams (Doctors &
and the perception of anxiety (p . 0.05). Conclusions: There is an apparent association Nurses) and patients, respectively. The app facilitated real-time communication between
between the habit of receiving test results online and anxiety levels in patients with patients, caregivers, and care teams. Conclusions: Digital platform demonstrated high
breast cancer. The recruitment is now complete, and results of the entire cohort will be user engagement and positive perceptions among both the patients and their caregivers.
available for presentation. Research Sponsor: None. The app effectively facilitated communication, symptom tracking, and management of
treatment side effects. Patient centered digital platforms hold promise for improving
cancer care support. Research Sponsor: Dr reddys Laboratories.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 91s

1577 Poster Session 1578 Poster Session

Biologically interpretable pathomics-driven transformer model with self- Artificial intelligence based music therapy intervention in cancer patients
supervised training for outcome prediction of immunotherapy in non-small undergoing chemotherapy in oncology day care (MUSICC). First Author:
cell lung cancer. First Author: Butuo Li, Shandong Cancer Hospital and Institute, Sujith Kumar Mullapally, Apollo Proton Cancer Centre, Chennai, India
Shandong First Medical University and Shandong Academy of Medical Science, China Background: Studies on the role of music therapy for improving the quality of life of
Institution or Organization, Jinan, China cancer patients are scarce. Artificial intelligence (AI) is being increasingly utilised in
Background: Only a subset of non-small cell lung cancer (NSCLC) patients experiences healthcare universally. Echo Care ©DigiNxtHlt Solutions leverages AI technology multi-
durable benefit from immune checkpoint inhibitors (ICIs), and precise biomarkers remain class neural networks to make personalised recommendations by learning continuously
scarce. Meanwhile, computational pathology, based on digital pathology, has led to from patients’ usage and interactions. We aim to study impact of this AI based music
significant advancements in NSCLC prognosis prediction. Yet limited generalization and therapy in health-related quality of life (HR-QOL) in cancer patients undergoing chemo-
interpretation remain critical challenges in current clinical practice. Self-supervised therapy. Methods: A pilot interventional study for 50 consecutive cancer patients being
learning, pretrained on unlabeled data to comprehensively capture underlying biological administered chemotherapy at daycare facility was conducted. After informed consent,
information in pathological images, may enable expandable, interpretable pathomics each patient underwent assessments during chemotherapy on week 1, week 3 and week 6.
models for outcome prediction of ICIs. Methods: H&E-stained slides from pan-cancer Vitals assessment, FACT G7 and HADS assessment was done at baseline, 3rd week and 6th
patients were digitized as whole slide images (WSIs), then segmented for preprocessing. week.30 minutes experience of sounds rendered by “echo Care” AI based application
A self-supervised foundation model (patho-GPT) were performed to extract WSI fea- ©DigiNxtHlt Solutions was provided at each session of chemotherapy at 1st week, 3rd week
tures, and further fine-tuned for outcome prediction of immunotherapy in NSCLC with and 6th week. Vitals were checked on admission to daycare unit followed by HR-QOL
progression-free survival (PFS) labels. Its performance was evaluated by accuracy, assessments which were done on digital platform by patients using unique ID and
sensitivity, specificity, and AUC in internal and external validation. Patients were password. After completing the FACT G7 and HADS assessment, option was given to select
classified as immunotherapy-resistant (R) or immunotherapy-sensitive (S), and ROC/ from multiple music modules generated by AI based on data entered by patients. Once they
survival curves were generated. The model was also tested in an operable cohort on completed the 30min session, post session vitals were measured. Qualitative feedback
neoadjuvant immunotherapy. Finally, single-cell RNA (scRNA) sequencing analyses also was collected. The primary objective was quality of life as defined by FACT G7/ HADS
scores and secondary endpoint was change in physiological vital parameters at baseline,
were performed to provide biological insights. Results: A total of 13770 whole slide
3 weeks, 6 weeks. Statistical analysis done by SPSS version 22.0. Results: A total of 50
images (WSIs) from 6589 patients were included to construct the self-supervised patho-
patients were enrolled for the study during study period from March 2023 to Dec 2024.
GPT model, which utilizes a context encoder, target encoder, and predictor based on the
Mean HADS score for depression was 10.2 +/-4 and for anxiety was 7.5 +/- 3. 40% patients
Vision Transformer architecture. There were 771 WSIs from 511 NSCLC patients re-
had high HADS score for anxiety. 18% had high HADS score for depression.50% of patients
ceiving immunotherapy, labeled using 5.23 months as the cut-off value. All patients were had improvement in HADS score for anxiety and depression whereas 50% did not have
divided into training and validation sets at a 7:3 ratio. For downstream outcome much change in their scores. Mean FACT-G7 score is 13.5+/-2 with majority having value
prediction of immunotherapy, the weight of patho-GPT was fine-tuned in the training set, less than 16 (70%). From qualitative feedback, 25% reported “very satisfied”, whereas 75%
and performance was evaluated in internal and independent external validation sets. The reported to be “satisfied” with music intervention. 54% patients felt less anxious, 50% had
accuracy was 0.828 (AUC 0.774) in the internal set and 0.758 (AUC 0.752) externally. calming effect and masking of pain, 37.5% felt mood elevation, positivity and closer to their
Survival analyses showed the model’s risk group was significantly associated with family. 30% patients felt muscle relaxing effect. 37.5% felt sleepy during music session
survival after immunotherapy. In contrast, the ViT-ViT model using initial weights with 25% not feeling any change in them during the sessions. Conclusions: Artificial
achieved 0.677 accuracy (AUC 0.547) in the external set, which was significantly inferior. intelligence-based music therapy intervention in cancer patient undergoing chemotherapy
ScRNA sequencing and differential analysis between R and S groups were performed, shows promising results in terms of health-related quality of life, satisfaction and ex-
and a high level of H1.2hi Teffs was found in R, linked to dysfunction of cytotoxicity- perience and can be offered to cancer patients as a non-pharmacological intervention to
related genes and immune pathways. Conclusions: The self-supervised patho-GPT can improve their quality of life during their treatment. Clinical trial information: CTRI/2023/03/
be used for the accurate prediction of immunotherapy outcome, with well generalization 050509. Research Sponsor: None.
and biological interpretation. Research Sponsor: None.

1579 Poster Session 1580 Poster Session

A randomized, controlled pilot trial of a neuromodulatory digital therapeutic Exploring social determinants of health and immunotherapy utilization in
for individuals with breast cancer. First Author: Samantha Adler, Click Thera- patients with stage III non-small cell lung cancer following definitive
peutics, Inc, New York City, NY chemoradiation. First Author: Chaewon Hwang, Beth Israel Deaconess Medical
Background: Fatigue, pain, and mood symptoms are common side effects of cancer Center, Boston, MA
therapy and cancer patients. Frontoparietal circuitry has been implicated in Background: Since the approval of immunotherapy (IO) for maintenance therapy in stage
chemotherapy-induced cognitive impairment as well as post-treatment fatigue. We III non-small cell lung cancer (NSCLC) in 2018, its use has expanded rapidly. This study
hypothesized that a smartphone-based multimodal multistable bias modification aimed to evaluate patterns of IO use among diverse patient populations who received
(MMBM) intervention could improve fatigue, mood, and pain-related attentional biases in definitive chemoradiation (CRT) followed by adjuvant IO for stage III NSCLC. Methods: A
breast cancer patients by correcting the neurocircuitry alterations caused by uncertainty retrospective analysis was performed using the National Cancer Database for patients $
stress induced by cancer diagnosis, as well as alterations resulting from chemotherapy, 18 years old with stage III NSCLC diagnosed between 2018 to 2021. Patients receiving $
correcting frontoparietal circuitry changes resulting from cancer or its treatment. 60 Gy of radiation in $ 30 fractions and IO delivered 50-150 days from starting CRT were
Methods: A randomized, single-blinded exploratory study in patients with breast cancer included. Patients who underwent surgery were excluded. Social determinants of health
(n = 81) was conducted. Participants were randomized 1:1 to an active MMBM app or a included race, ethnicity, insurance, treatment site, Charlson-Deyo comorbidity index (CCI),
control app and instructed to use it for 7 minutes daily over 4 weeks. Clinical endpoints high school graduation rate (HSGR; lowest, low, mid, high per US Census data), income
included PROMIS-29+2, Brief Pain Inventory (BPI), and Pain Catastrophizing Scale (categorized similarly to HSGR), sex, and age. Odds ratios (OR) were calculated for IO use.
(PCS). Feasibility and acceptability of app usage were also evaluated. Results: The Logistic regression analyses were performed for each variable followed by multivariable
MMBM app group showed significant improvements over control in PROMIS-29 fatigue analysis with statistically significant factors (P , 0.05). IRB exemption was granted.
(-3.4, p , 0.05), depressive symptoms (-2.8, p , 0.05), and symptoms of anxiety (-3.0, Results: 25,746 patients were included: 21,848 White, 3,236 Black, 558 Asian, and 104
p , 0.05) domains. There were also several pain measures that showed significant Native American. 620 were Hispanic. On univariate analysis, the following were predictive
improvement over time in the MMBM group, but not the control group, such as the of IO receipt: Black vs White race (OR 0.86, P , 0.001), Hispanic vs non-Hispanic ethnicity
PROMIS-29 pain intensity measure (-0.9, p , 0.001), the BPI average pain intensity (-0.7, (0.72, P , 0.001), Medicare or other governmental insurance vs private (OR 0.94, P =
p , 0.001) and current pain intensity (-0.7, p , 0.01), and the PCS total score (-3.3, p , 0.037 and 0.85, P = 0.034), CCI 1 or 2 vs 0 (1.11, P , 0.001 and 1.13, P = 0.005),
0.01). Conclusions: This study provides preliminary evidence that the MMBM inter- community site vs academic/research center (0.88, P = 0.007), lowest and low vs high
vention may alleviate fatigue and mood-related symptoms in breast cancer patients, HSGR (0.78, P , 0.001 and 0.87, P , 0.001), lowest vs high income (0.90, P = 0.005),
with potential for improving pain-related symptoms as well. These findings underscore female vs male sex (1.06, P = 0.028), and age (-0.88%/year, P , 0.001). On multivariate
the potential of digital neuromodulation as an innovative approach to enhance the analyses, Black patients were 3.3% less likely to receive IO than White patients (0.88, P ,
quality of life of patients with complex conditions, specifically breast cancer. Clinical 0.001). Hispanic patients were 8.2% less likely to receive IO than non-Hispanic patients
trial information: NCT06136923. Research Sponsor: None. (0.72, P , 0.001). Medicare insurance (0.93, P = 0.012) and treatment at community sites
(0.86, P = 0.002) were associated with reduced IO use. Patients from areas with lower
HSGR were less likely to receive IO than those with the highest HSGR (lowest: 0.75, P ,
0.001; low: 0.83, P , 0.001; mid 0.91, P = 0.039). Higher CCI (CCI 1: 1.12, P , 0.001; CCI 2:
1.13, P = 0.004; CCI 3: 1.10, P = 0.046), lower income (lowest: 1.11, P = 0.048; low: 1.10,
P = 0.041; mid: 1.10, P = 0.025), and female sex (1.05, P = 0.048) were associated with
increased IO use. Conclusions: This study highlights disparities in IO use following CRT
for stage III NSCLC patients. Black race, Hispanic ethnicity, Medicare or governmental
insurance, treatment at community sites, and lower education were associated with
decreased IO use. These findings emphasize the need for strategies to ensure equitable
access to advanced cancer therapies. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
92s CARE DELIVERY/MODELS OF CARE

1581 Poster Session 1582 Poster Session

Food insecurity and disparities in mental health symptoms and severity Real world treatment patterns and outcomes in metastatic EGFR mutation–
among breast cancer survivors. First Author: Kent Schechter, Ben May Department positive NSCLC patients: A retrospective study from a tertiary care cancer
for Cancer Research, The University of Chicago, Chicago, IL center in India. First Author: Jyothis P. Jose, MVR Cancer Centre & Research Institute,
Background: Breast cancer survivors (BCS) often experience mental health challenges Calicut, India
during survivorship. Food insecurity is a growing public health concern in the US, which may Background: Lung cancer is the leading cause of cancer related mortality worldwide
exacerbate the challenges faced by BCS. However, little is known about the associations with Non small lung cancer (NSCLC) constituting the majority of cases. EGFR mutations,
between food insecurity and mental health symptoms and severity among BCS. predominantly exon 19 and Exon 21 L858R are actionable targets in approximately 25-
Methods: We conducted a secondary analysis of the 2022 National Health Interview Survey 30% Indian patients. Third generation EGFR tyrosine kinase inhibitors have set new
that used stratified clustering sampling to interview US adults aged $18. This study was benchmarks in the treatment of metastatic EGFR mutation positive NSCLC. However
limited to women with a breast cancer history. Food insecurity (secure/insecure) was real world adoption in India faces significant barriers including financial constraints and
measured using a 10-item questionnaire assessing past 30-day household food situations. healthcare disparities. Methods: Our objective was to evaluate whether patients with
Ever having anxiety was self-reported (yes/no), and the severity level in the past 2 weeks was EGFR mutation positive metastatic NSCLC receiving the standard of care treatment and
examined using the 7-item Generalized Anxiety Disorder scale. Women reported ever analyze demographic, clinical and molecular characteristics of EGFR mutation positive
having a depressive symptom (yes/no), and the 8-item Patient Health Questionnaire de- lung cancer patients treated at our center and compare treatment patterns and survival
pression scale was used to evaluate the severity level in the past 2 weeks. We compared outcomes with national and global data. This retrospective study analyzed 894 stage 4
weighted percentages using Rao-Scott Chi-squared tests and computed P-trends using the
NSCLC patients treated between 2018-2023 of these 252 (28.1%) were EGFR mutation
Cochran-Armitage test. We performed weighted logistic regression to estimate adjusted
positive. Data on EGFR types, treatment modalities and survival outcomes were col-
odds ratios (AOR). All analyses accounted for survey weights. Results: We obtained an
unweighted sample of 644 BCS (weighted sample 4,234,520). Overall, 7.6% experienced food
lected and analyzed. Kaplan Meir survival analysis was performed to estimate pro-
insecurity. BCS who were food insecure were younger than those secure (mean age 61 vs 69 gression free survival (PFS) and overall survival (OS). Results: First line Osimertinib was
years; P,.002). Black (20.4%) or Hispanic (24.3%) BCS were more likely than White BCS median OS was 30 months and PFS was 20 months. The standard of care at that time
(4.7%) to experience food insecurity (P,.001). BCS who were food insecure reported a received by 13.5% only. Other TKI alone (Gefitinib, Erlotinib and Afatinib) median OS was
higher percentage of depression than those secure (52.9% [95% CI: 35.5-70.3%] vs 24.4% 21 months while PFS was 12 months. Other TKI with chemotherapy median OS 27
[95% CI: 20.2-28.6%]; P,.001). The proportion of anxiety was higher among BCS who were months and PFS was 18 months. Mutation specific outcomes showed exon 19 deletions
food insecure than those secure (45.4% [95% CI: 27.7-63.1%] vs 19.8% [95% CI: 16.0-23.6%]; had better overall survival 27 months PFS 16 months compared to Exon 21 L858R (OS
P,.001). After controlling for demographic and socioeconomic factors, BCS who were food 16; PFS 11 months). T790M were identified in 71% patients progressing after first line
insecure had greater odds of depression (AOR 3.19, 95% CI: 1.42-7.17) or anxiety (AOR 3.14, treatment. Second line osimertinib achieved a median PFS of 9.5 months. Conclu-
95% CI: 1.34-7.34) than those secure. Compared with BCS who were food secure, those sions: Despite the efficacy of osimertinib demonstrated in global trials, its adoption was
insecure were more likely to experience moderate (26.0% vs 6.0%) or severe (7.5% vs 0.6%) limited to 13.5% of eligible patients in this cohort due to economic barriers. These
depression (P-trend,.001), as well as moderate (6.4% vs 4.4%) or severe (14.5% vs 1.3%) findings emphasize the urgent need for systemic interventions to improve access to
anxiety (P-trend,.001). BCS who were food insecure also reported a higher proportion of advanced therapies in India. Policymakers and healthcare systems must address these
forgoing mental health counseling due to cost than those secure (6.8% vs 1.5%; P=.02). gaps through measures like government subsidies, expanded insurance coverage and
Conclusions: Our findings highlight food insecurity prevalence and associated disparities in cost effective diagnostic platforms to ensure equitable access to precision oncology in
mental health symptoms and severity among BCS. Interventions and policies, e.g., food resource limited settings. Research Sponsor: None.
pantries/meals programs and nutrition assistance, are needed to address food insecurity.
Cancer centers should also consider routine mental health screening and offer proper
services to reduce racial disparities among BCS. Research Sponsor: National Institute on
Aging; T32AG000243; Susan G. Komen Breast Cancer Foundation; TREND21675016.

1583 Poster Session 1584 Poster Session

Racial disparities in oncology clinical trials by absolute neutrophil count Factors associated with receipt of surveillance breast MRI among racially/
eligibility criteria: A single center retrospective analysis. First Author: Arvind ethnically diverse women with a personal history of breast cancer. First
Suresh, Department of Medicine, University of California San Francisco, San Francisco, Author: Preeti Kakani, Columbia University Irving Medical Center, New York, NY
CA Background: Surveillance breast imaging among breast cancer (BC) survivors is rec-
Background: Black individuals have been historically underrepresented in clinical trials. The Duffy-null ommended for early detection of local recurrence or contralateral BC. Breast MRI is
phenotype leads to clinically insignificant lower absolute neutrophil counts (ANC) and is found in 66% of non- increasingly used as a screening modality given its heightened sensitivity compared to
Hispanic Blacks (NHB) and , 1% non-Hispanic Whites (NHW) in the United States. Although the ANC lower
limit of normal for Duffy-null individuals is estimated to be 1,210/mL, little is known about the impact of ANC
mammography, and in 2018, the American College of Radiology released guidelines
criteria on racial underrepresentation in clinical trial eligibility and enrollment. Methods: We conducted a recommending supplemental breast MRI among women diagnosed with BC under age 50
single-center retrospective study of patients newly diagnosed with five cancers (lung, breast, prostate, or with dense breasts. We investigated demographic and clinical factors associated with
colorectal, non-Hodgkin lymphoma) between 1988-2024. Baseline organ function data (Hgb, ANC, platelets, receipt of surveillance breast MRI among women with a personal history of BC.
AST, ALT, creatinine, and bilirubin) were compared between NHB and NHW. We also identified ANC Methods: We conducted a retrospective cohort study of women with stage 0-III BC
thresholds above which eligibility proportions differ between NHB and NHW. Results: We identified 23,854 between January 2018-June 2023 at Columbia University Irving Medical Center in New
patients. 14,585 had available baseline lab data within one year prior to starting anti-cancer therapy. The
table lists the percentage of NHW and NHB who meet ANC eligibility criteria at thresholds of 1,000, 1,500, and York, NY. We excluded women with bilateral mastectomies. The primary outcome was
2,000/mL. Among patients with breast cancer, the median ANC was 3,500 for NHB and 3,890 for NHW (p = receipt of at least one breast MRI . 1 year after initial diagnosis. Data from the electronic
0.63) and fewer NHB were eligible at all ANC thresholds of 1,500/mL or greater (94.5% NHB vs 97.2% NHW, p health record included age at diagnosis, race/ethnicity, primary language, health in-
= 0.03). For those with prostate cancer, the median ANC was 3,780 for NHB and 4,450 for NHW (p = 0.03) and surance, first degree family history of BC, germline genetic testing results, stage at
significantly fewer NHB were eligible at all ANC thresholds of 1,400/mL or greater (97.6% NHB vs 99.6% NHW, diagnosis, mammographic density (MD), and BC treatments (surgery, radiation, che-
p = 0.04). For those with non-Hodgkin lymphoma (NHL), the median ANC was 4,260 for NHB and 4,020 for
NHW (p = 0.34) and significantly fewer NHB were eligible at all ANC thresholds of 1,600/mL or greater (90.4%
motherapy, and hormonal therapy). We performed multivariable logistic regression to
NHB vs 97.1% NHW, p = 0.008). Patients with lung cancer (ANC NHB 4,820 vs NHW 4,920; p = 0.85) and assess factors associated with receipt of surveillance breast MRI. Results: Among 1,990
colorectal cancer (ANC NHB 3,930 vs NHW 4,150; p = 0.27) did not have significant ANC eligibility differences evaluable patients, mean age was 59.7 years (SD 12.2), and the cohort included 14% non-
between NHB and NHW. No significant differences were found for other baseline labs. Conclusions: Our Hispanic Black, 33% Hispanic, and 6% Asian women. About 22% of women were diag-
findings provide further evidence that ANC criteria may contribute to differences in clinical trial eligibility nosed before age 50, 53% had dense breasts (BIRADS C-D), and 18% received at least one
between NHB and NHW for breast cancer, prostate cancer, and NHL. Further work is needed to identify surveillance breast MRI . 1 year after diagnosis. On adjusted analysis, younger age at
optimal cutoffs for each disease group. Limitations include lack of Duffy status for most patients and race is
an imperfect approximation. Future clinical trials should proactively address these differences by using ANC diagnosis, higher MD, first-degree family history of BC, receipt of germline genetic testing,
eligibility criteria based on Duffy phenotyping. Research Sponsor: UCSF Department of Epidemiology and and having a germline pathogenic variant were associated with receipt of breast MRI.
Biostatistics. Compared to non-Hispanic White women, Hispanic and non-Hispanic Black women had
Cancer NHB NHW p NHB NHW p NHB NHW p
lower odds of receiving breast MRI (odds ratio [OR] = 0.44, 95% confidence interval [CI] =
(n=NHB; NHW) ANC >2000 ANC >2000 value* ANC >1500 ANC >1500 value* ANC >1000 ANC >1000 value* 0.31-0.63 and OR = 0.57, 95% CI = 0.38-0.86, respectively). However, compared to patients
Lung 90.4% 94.4% 0.27 93.2% 97.4% 0.10 95.9% 99% 0.08 with commercial insurance, those with Medicaid were more likely to undergo breast MRI
(n=73; 726) (OR = 1.57, 95% CI = 1.10-2.25). Results were similar when restricting the analysis to
Breast 86.1% 92.7% <0.001 94.5% 97.2% 0.03 98.7% 99% 0.99
(n=237; 2,905) those diagnosed before age 50 or with dense breasts. Conclusions: Hispanic and non-
Prostate 88.6% 97.8% <0.001 96.8% 99.4% 0.02 100% 99.9% - Hispanic Black women with BC were less likely to receive surveillance breast MRI than
(n=124; 961)
Colorectal 87.5% 93.5% 0.08 100% 98.6% - 100% 99.9% - their non-Hispanic White counterparts. Also, as patients with Medicaid were more likely to
(n=80; 800) undergo breast MRI than those with commercial insurance, there may be varying health
NHL 76.7% 85.6% 0.02 87.4% 90.2% 0.39 94.2% 93.6% 1.0
(n=103; 1,414) insurance coverage for breast MRI. These results highlight the need for more stan-
*Calculated by Chi-squared tests.
dardized guidelines surrounding surveillance breast MRI among BC survivors, which may
inform public health initiatives aimed at promoting equitable breast imaging practices in
this population. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 93s

1585 Poster Session 1586 Poster Session

Meeting enrollment targets in IMbrave152/SKYSCRAPER-14, a global Intervention adherence, engagement and tool utilization in the breast cancer
phase 3 study in patients with unresectable hepatocellular carcinoma weight loss (BWEL) trial by race and ethnicity (Alliance A011401). First
(HCC). First Author: Christopher Cotter, Genentech, Inc., South San Francisco, CA Author: Ashley Odai-Afotey, Dana-Farber Cancer Institute, Boston, MA
Background: The FDA has released guidance that the patient population in clinical studies should be Background: Black and Hispanic breast cancer (BC) survivors have a higher prevalence of obesity and
representative of the intended-use population and the epidemiology of the disease, particularly in terms of experience less success with weight loss interventions (WLI) than White BC survivors. The BWEL trial
race and ethnicity. IMbrave152/SKYSCRAPER-14 (NCT05904886) is a global phase 3 trial in patients with (Alliance A011401; NCT02750826) is a phase III randomized trial evaluating the impact of a 2-year
HCC, a disease with a globally high Asian and African prevalence, and which significantly impacts un- telephone-based WLI on invasive disease-free survival in participants (pts) with stage II-III HER2-
derrepresented (including Black and Hispanic) patients in the USA. Enrollment goals (by race/ethnicity), as negative BC and a BMI $ 27 kg/m2. At 12-months, the WLI induced significant weight loss across
well as operational and protocol-driven tactics to meet those goals, were implemented for IMbrave152. As demographic factors, including race and ethnicity. However, Black and Hispanic pts lost less weight
of January 2, 2025, the targets for Black and Hispanic patients were met. Methods: Global recruitment and completed fewer calls than White pts. Here, we evaluate intervention adherence, engagement and
began on September 14, 2023. Operational tactics included: feasibility questions about the ability to recruit tool utilization in BWEL pts by race and ethnicity. Methods: BWEL randomized pts to a WLI plus health
underrepresented populations; utilization of internal and external databases to identify sites that could education (HE) or HE alone. WLI pts received semi-structured telephone-based health coaching,
enroll underrepresented patients; incorporation of patient input; and enhanced patient-support services to delivered in English or Spanish, and received an activity monitor and wireless scale. Pts self-reported
facilitate recruitment and retention. Protocol-driven tactics included: modifying inclusion and exclusion
race and ethnicity. Mean values for call duration, call density (time to complete the initial 12-week
criteria (considering race/ethnicity); streamlining study assessments; flexibility to use decentralized
intensive intervention phase), intervention attrition, and frequency of Fitbit use and weight tracking
processes; and inclusion of an Africa Extension Cohort, which allowed for the recruitment of patients in
over 12-months were compared by race and ethnicity, comparing least squares means with Tukey-
Africa beyond the intent-to-treat population. Results: As of January 2, 2025, the randomized, global
population included 8% Black and 14% Hispanic patients (US recruitment included 16% Black and 19% Kramer adjustment for multiple comparisons with adjusted p-values. Results: Of 3181 pts ran-
Hispanic patients), exceeding other global HCC studies, which recruited 2% or fewer Black patients and 11% domized to the study between 08/2016 and 02/2021, 1591 pts were allocated to the WLI arm. 80.5% of
or fewer Hispanic patients. Only two out of seven phase 3 studies in recent years have reported on Hispanic pts were White, 12.8% Black, and 7.1% Hispanic. Average BMI was 34.5 (65.7) kg/m2. Compared to
patient recruitment. Conclusions: IMbrave152 recruited the highest percentage of Black and Hispanic White pts, Black pts had fewer days of Fitbit usage (113.6 vs. 159.8, p,0.0001) and weight tracking
patients to date in a global phase 3 HCC study. These results demonstrate that recruitment of an un- (77.9 vs. 135.6 days, p, 0.0001). Hispanic pts had fewer days of Fitbit usage (108.8 vs. 154.9, p=
derrepresented population is feasible if operational and protocol-driven tactics are utilized. Clinical trial 0.001) and weight tracking (87 vs. 129.5 days, p=0.0002) compared to non-Hispanic pts. There were no
information: NCT05904886. Research Sponsor: F. Hoffmann-La Roche Ltd. differences in attrition rate, average call duration, or call density by race or ethnicity. Conclusions: In a
phase III WLI trial, engagement with tools designed to support weight loss was significantly lower in
Global enrollment from Black and Hispanic pts. Future work is needed to explore ways to enhance engagement and improve
Study White, % Asian, % Black, % Hispanic, % Asia (excl. Japan), %
weight loss outcomes for racial and ethnic minority pts. Support: U10CA180821, U10CA180882,
IMbrave152 32.7 54.7 8.0 13.6 46.7 UG1CA189823; [Link] Clinical trial information: NCT02750826.
IMbrave1501 34.9 56.7 2.0 NR 40.1 Research Sponsor: National Cancer Institute/U.S. National Institutes of Health.
REFLECT2 28.9 69.2 NR NR 67.1*
HIMALAYA3 44.5 50.9 1.6 4.7 40.9 Race Ethnicity
LEAP-0024 43.5 43.5 1.6† 11.2 30.7
CheckMate-4595 53.2 44.7 0.7 NR 25.3 Non-
RATIONALE-3016 NR NR NR NR 63.1 White Black Hispanic Hispanic
COSMIC-3127 50.9 31.4 1.7 NR 28.8 Measure of engagement N=1281 N=204 p-value N=1459 N=113 p-value
1 Withdrew from 58 (4.5%) 13 (6.4%) 0.56 70 (4.8%) 6 (5.3%) 0.51
Finn et al. NEJM 2020;
2
Kudo et al. Lancet 2018; intervention
3
Abou-Alfa et al. NEJM Evid 2022; n (%)
4
Llovet et al. Lancet 2023; Call Duration (min) 34.5 (7.6) 34.6 (9.1) 0.99 34.6 (7.8) 33.3 (7.9) 0.25
5 Mean (SD)
Yau et al. Lancet Oncol 2022;
6 Call density 14.2 (7.2) 14.2 (8.8) 0.99 14.2 (7.3) 14.2 (9.7) 0.99
Qin et al. JAMA Oncol 2023; (weeks)
7
Kelley et al. Lancet Oncol 2022. Mean (SD)
NR, not reported. Days of Fitbit Use 159.8 (132.2) 113.6 (124.7) , 0.0001 154.9 (131.8) 108.8 (125.7) 0.001
*Asia-Pacific region. Mean (SD)
†
Multiple races, including 9 additional patients who were Black plus either Asian or White, were also reported. Days of weight tracking 135.6 (109.5) 77.9 (87.5) , 0.0001 129.5 (108.8) 87 (96.5) 0.0002
Mean (SD)

1587 Poster Session 1588 Poster Session

Breast cancer optimal care timeframes for culturally and linguistically di- Continuous financial toxicity screening in community oncology. First Author:
verse populations and First Nations People: A regional centre experience in Thomas Gregory Knight, Atrium Health Levine Cancer Institute, Wake Forest University
Australia. First Author: Matthew Hon, Townsville University Hospital, Douglas, QLD, School of Medicine, Charlotte, NC
Australia Background: Financial Toxicity (FT) has been repeatedly linked with adverse cancer
Background: Culturally and linguistically diverse (CALD) populations and First Nations clinical outcomes. However, screening practices vary widely, especially in community
People are at-risk communities who face unique challenges in cancer diagnosis and settings where , 50% routinely proactively engage patients to discuss care costs. This
management resulting in inequities. Optimal Care Pathways (OCP) established by Cancer quality improvement pilot examined the feasibility and impact of continuous FT screening
Council Australia aim to address these disparities. The Breast cancer OCP outlines an in a community-based clinical practice. Methods: Using PDSA methodology, an electronic
integrated model of care with optimal timeframes such as time from general practitioner distress screening (EDS) tool was implemented at each visit at two rural oncology
(GP) referral to specialist surgical review, time from decision to treat to surgery or practices. Evidence of FT was defined as answering “yes” to the question “Do you have
neoadjuvant chemotherapy (NAC), and time from completion of NAC to surgery. insurance/financial problems or concerns?” The EDS tool would automatically email the
Methods: Retrospective data was collected for all CALD (migrant from non-English financial navigation (FN) team on “yes” response and patients were contacted by FN within
speaking country and/or primary language identified as not English) and First Nations 48 hours. Contact was attempted at least 3 additional times if unable to be reached. Four
patients diagnosed with breast cancer treated at a regional centre in Australia (Townsville successive monthly PDSA cycles ran from April to July 2024. In addition to demographic
University Hospital) from 2018 – 2022. A comparison cohort (control) of consecutive non- trends, success metrics were: % of screened patients with FT; % of FT patients new to the
CALD, non-First Nations patients was included. Data collected included patient demo- FN team; number and types of resolutions of FT concerns; and satisfaction and feasibility
graphics, tumour characteristics, stage, and identified timeframes which were compared survey of clinical teams and patients. Results: In the 4-month study,1071 patients were
with OCP. Results: 133 patients were included with 43 CALD (32%), 41 First Nations (31%) screened using the EDS tool: 169 (16%) affirmed FT. Of those with FT concerns, 140 (83%)
and 50 control (37%). CALD and First Nations cohorts had higher rates of stage IV disease were new to FN. The FN team provided a primary resolution to 85 patients of 169 (50%)
at diagnosis (12 v 15%) compared to control cohort (0%). They were also more likely to be who alerted. Of the remainder, 45 (27%) could not be contacted after multiple attempts and
diagnosed via emergency department admission (CALD 16 v First Nations 7%) compared to 39 (23%) reported clicking in error. Primary resolutions included: Charity Care Program
control cohort (0%) suggesting later presentation. Of those referred through OCP defined Referral (36%), Financial Teaching (29%), Billing Changes (11%), Social Work Referral (9%),
GP pathway, a similar percentage were reviewed by specialist surgeon within optimal 2- Medication Assistance (6%), and Marketplace Insurance Obtained (5%). The patients
week timeframe in all groups (CALD 47%; First Nations 39%; control 44%). Median time receiving FN services were majority female (75%) and between 35 and 64 yo (57%). The
from decision to treat to surgery were longer in CALD versus control groups (19 v 13 days; p most prevalent cancer types were Blood/Marrow (35%) and Breast (31%). 66% were white,
= 0.03), and in First Nations versus control groups (22 v 13 days; p = 0.02). Less CALD (89%, 24% African American, and 14% Hispanic. Payors included 40% commercial insurance, 31%
n = 24) and First Nations (82%, n = 18) patients underwent surgery within optimal 5-week Medicare, 19% Medicaid, and 9% other. Geographically, 62% of patients resided in rural
timeframe compared to control (98%, n = 40). Similarly, median time from decision to treat areas, 24% suburban, and 4% urban. Patient satisfaction with FN was high across all
to NAC were longer in CALD versus control groups (19 v 14 days; p = 0.05), and First categories; 55% agreed or strongly agreed that FN services helped lower stress about bills.
Nations versus control groups (20 v 14 days; p = 0.03). Most patients (91%, n = 29) Scores were highest for “FN cared about my concerns and needs” (69%); “would rec-
commenced NAC within optimal 4-week timeframe; 2 CALD and 1 First Nations patients did ommend it to others in need” (63%); and “information from FN was clear and easy to
not. Median time from completion of NAC to surgery was longer in CALD versus control understand” (61%). The clinic teams survey in participating locations felt the EDS
groups (29 v 24 days; p = 0.15), and in First Nations versus control groups (35 v 24 days; p = screening tool was feasible in their practice environment (67%) and reported they felt
0.04). Of those who recieved NAC, 100% CALD (n = 9), 69% First Nations (n = 9), and 89% routine FT screening was useful for patients (67%) (n=9). Conclusions: Structured
control (n = 8) patients underwent surgery within optimal 4-week timeframe. implementation of routine FT screening with an EDS tool in a rural, oncology community
Conclusions: Achievement of key OCP timeframes was lower in both CALD populations practice is feasible with high patient and clinical team satisfaction and may allow for earlier
and First Nations People. Strategies need to be further developed to address the delays identification of at-risk patients. Future directions include screening questionnaire re-
and health outcome disparities in these vulnerable cohorts. Research Sponsor: None. finements and expansion to other clinical sites. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
94s CARE DELIVERY/MODELS OF CARE

1589 Poster Session 1590 Poster Session

Socioeconomic- and insurance-based inequities in Oncotype DX testing and Empowering minority patients: A tailored education initiative for clinical trial
score-guided treatment. First Author: Courtney Williams, University of Alabama at awareness in thoracic oncology. First Author: Tadana Angelica Vazquez Rothschuh,
Birmingham, Birmingham, AL Ponce Health Sciences University, Ponce, PR, Puerto Rico
Background: Personalized approaches to breast cancer treatment are increasingly guided Background: Minority populations are underrepresented in clinical trials. Underrepresentation stems
by expensive, lab-based genomic testing like Oncotype DX (ODX) Breast Recurrence Score from multiple barriers that include restrictive eligibility criteria, systemic inequities, and patient-related
[Link] is known about how socioeconomic and insurance status may affect utilization factors. Effective educational interventions are needed to address barriers toward increasing minority
patient understanding and participation in clinical trials. Despite their potential, video-based, culturally
of ODX testing and subsequent ODX score-guided treatment. Methods: This retrospective tailored interventions for minority patient education on clinical trials remain understudied and
cohort study included women diagnosed with an early-stage, HR+/HER2- breast cancer underutilized. Methods: We developed 3 brief (,5 min) culturally tailored educational videos about
from 2011-2023 within the nationwide Flatiron Health electronic health record-derived clinical trials in thoracic oncology for Black (English) and Hispanic/Latino patients (English and Spanish).
deidentified database. Socioeconomic status was measured by the Yost index, a census Patients completed pre- and immediate post-video surveys to assess knowledge about, attitudes toward,
block-level measure of neighborhood deprivation. Insurance status was captured at time of and willingness to participate in clinical trials. Inclusion criteria included age $18 years, diagnosis of
diagnosis. Utilization of ODX testing was compared descriptively. Likelihood of receiving stage I-IV lung cancer, ECOG performance status 0-2, self-identification as Black or Hispanic/Latino,
adjuvant chemotherapy by neighborhood deprivation or insurance status was estimated proficiency in English or Spanish, and access to an electronic device. Feasibility was measured by
participant recruitment rates, completion of pre-and post-video surveys, and video acceptability via self-
using relative risk, predicted probabilities, and 95% confidence intervals from adjusted reported satisfaction. Results: Between April and September 2024, 54 patients were approached, and 30
Poisson models with robust variance estimates. Analyses were stratified by age due to (56%) were successfully enrolled (Table). Among those approached, 4 were ineligible, and 20 declined
differing recurrence risk score categorizations. ODX scores indicating low or low/medium participation. All enrolled patients completed the pre- and post-video surveys. Over 90% reported high
recurrence risk suggests chemotherapeutic benefit will likely not outweigh risk of side satisfaction with the videos. The total knowledge score increased significantly following the intervention
effects, while scores indicating medium or high recurrence risk suggests chemothera- (p , .001). For specific items, significant improvements were observed in understanding pre-clinical
peutic benefit will likely outweigh risk of side effects. Results: Of 4,367 patients eligible studies (p = .012), placebo use (p = .001), and clinical trial registration (p = .012). After the videos, over
for ODX testing, mean diagnosis age was 62 years (SD 12), 77% were white, 69% had stage I 90% of patients believed that clinical trials are useful and play an important role in developing new drugs
and improving lung cancer outcomes. Potential barriers were observed in ~50% of patients, including
cancer, 8% had $1 comorbidity, 48% were commercially insured, and 30% lived in a highly
concerns with informed-consent language, randomization, costs, and healthcare team communication
deprived neighborhood. Compared to those without, patients with an ODX test (46%, n = style. There was a significant increase in willingness to participate in a clinical trial after watching the
2,026) were more often white (81% vs. 74%), commercially insured (51% vs. 45%), or lived educational videos (p=.04). Conclusions: Our culturally tailored, patient-centered educational videos on
in a neighborhood of low deprivation (73% vs. 67%). For patients aged #50 with ODX clinical trials in thoracic oncology were both feasible and well-accepted. Future studies should explore
testing (n = 370), 51%, 25%, 12%, and 12% had low, low/medium, medium, and high whether such interventions can increase minority patient enrollment in therapeutic clinical trials.
recurrence risk. Of those with low recurrence risk, patients who resided in neighborhoods Research Sponsor: Moffitt Cancer Center Foundation; Bristol Myers Squibb Foundation; Fundacion
of low vs. high deprivation had 9% higher probability of receiving potentially inappropriate Intellectus; National Cancer Institute/U.S. National Institutes of Health; U54 CA163068.
overtreatment with adjuvant chemotherapy (15%, 95% CI 10-25% vs. 6%, 95% CI 2-24%). Of Patient demographics.
those with low/medium recurrence risk, publicly vs. commercially insured patients were Variable Level N = 30 %
2.7x more likely to receive adjuvant chemotherapy (RR 2.71, 95% CI 1.00-7.39). For Preferred Language English 23 76.7
patients aged . 50 with ODX testing (n = 1,656), 83% and 17% had low and high recurrence Spanish 7 23.3
Age Group 30-49 5 16.7
risk. Of those with high recurrence risk, patients who resided in neighborhoods of high vs. 50-69 12 40.0
low deprivation were 18% less likely to receive recommended adjuvant chemotherapy, 70-89 13 43.3
suggesting undertreatment (RR 0.82, 95% CI 0.67-1.00). Conclusions: Socioeconomic- Sex at Birth Female
Male
19
11
63.3
36.7
and insurance-based inequities, including both overtreatment and undertreatment, were Race White 10 33.3
observed in this national cohort of women with early stage breast cancer eligible for ODX Black/AA 16 53.3
Other 4 13.3
testing, indicating opportunities to increase care quality. Research Sponsor: Flatiron Ethnicity Not Hispanic 15 50.0
Health. Hispanic 15 50.0

1591 Poster Session 1592 Poster Session

Development of a culturally tailored educational tool designed to increase Comparative analysis of demographics and outcomes in young versus
access to somatic testing among Black men with metastatic prostate cancer average onset hospitalized gastrointestinal cancer patients in New York
(mPCa). First Author: Christopher Johns, University of California, San Francisco, San State. First Author: Mrinalini Ramesh, University at Buffalo, Buffalo, NY
Francisco, CA Background: Recent studies have shown an increasing incidence of gastrointestinal (GI)
Background: Somatic testing (ST), also known as molecular profiling, has become in- cancers among young patients. This study investigates clinical outcomes and healthcare
creasingly important for therapy selection in men with mPCa. Interventions targeting utilization among young ( , 50) versus average onset ($50) GI cancer patients admitted
known racial inequities in the use of ST are lacking. Our objective was to design a culturally to hospitals in New York State (NYS). Methods: We performed a retrospective analysis
tailored educational tool to augment patient education with the ultimate goal of increasing using the Statewide Planning and Research Cooperative System (SPARCS) database from
equitable access to ST for Black patients with mPCa. Methods: We used principles of 2009 to 2022. Patients were divided into two groups: young-onset GI cancer patients
human-centered design to develop a ST educational tool. We first designed a prototype (YOGIC, , 50 years) and average-onset GI cancer patients (AOGIC, $50 years). GI cancers
with key stakeholders, then conducted a qualitative study of Black patients with mPCa at included anal, biliary tract, colorectal, esophageal, gallbladder, liver, pancreatic, peritoneal,
three sites – a tertiary care academic center, a VA medical center, and a safety net small intestine, and stomach cancers. The study population was further stratified by
oncology clinic. Trained interviewers conducted semi-structured interviews to explore demographic and clinical characteristics. All variables were compared using the Kruskal-
patients’ perceptions of ST and elicit feedback about the educational tool until data Wallis test or Fisher’s exact test, along with multivariate linear and logistic regression in
saturation was met. Based on this feedback, we iteratively revised the tool. Interviews RStudio version 4.4.2, with a significance level of p #0.05. Clinical characteristics, in-
were transcribed, and two coders qualitatively analyzed transcripts using the COM-B cluding severity of illness and risk of mortality, were defined using the All Patient Refined
framework to identify barriers/facilitators of tool use. Results: For the initial prototype, (APR) grading system. Results: A total of 256,924 patients were identified (26,071 YOGIC
four physicians, one genetic counselor, and two patients contributed to the design of a 7- and 230,853 AOGIC) with a primary admitting diagnosis of a GI cancer from 2009 to 2022.
min video of a Black oncologist with informative animations followed by a text-based The AOGIC group had a higher proportion of white and female patients, whereas the YOGIC
decision aid. We approached 18 patients, of whom 11 (61%) consented to review the tool group included a higher proportion of black and male patients (p , 0.001). Anal, peritoneal,
then complete the interview. All participants expressed a positive perception of the tool and stomach cancers were more prevalent in YOGIC, while pancreatic and esophageal
and comprehension of the information. Tool facilitators included 1) trust in the tool, 2) cancers were more common among AOGIC (p , 0.001). AOGIC patients were more likely to
actionability of the tool’s content, and 3) appreciation for a Black physician featured in the have extreme (13% vs. 9.7%) or major (39% vs. 32%) severity of illness compared to YOGIC
video. Barriers included 1) difficulty navigating the electronic interface, 2) negative (p , 0.001). However, the median total cost of stay for YOGIC was significantly higher than
emotions from reflection on their cancer diagnosis or racial inequities, 3) too much that for AOGIC ($21,421 vs. $19,658 respectively, p , 0.001). YOGIC patients were more
information in the decision aid, particularly about biopsy risks and testing costs, and 4) likely to undergo procedures during hospitalization (95%) compared to AOGIC (93%) (p ,
content concerns on risks associated with ST/biopsy. For participants’ perceptions of ST, 0.001). YOGIC patients were more likely to be discharged home (66%) vs. AOGIC (48%) (p ,
facilitators included 1) desire to learn more about mPCa and 2) motivation to improve their 0.001). Longer hospital stays were associated with patients diagnosed with esophageal,
health, their family’s health, or the health of the Black community. Barriers included 1) peritoneal, and stomach cancers (p , 0.001) versus the reference group (anal cancer), and
misinformation about ST and mPCa, 2) difficulty accessing affordable healthcare, 3) in AOGIC [log fold-change 0.029 (95% CI: 0.021-0.038), p , 0.001]. Despite a higher
mistrust of the healthcare system due to prior negative experiences, and 4) belief that mortality risk in AOGIC, mortality rates decreased faster over the study period [OR 0.97
mPCa and its treatments are emasculating. Overall, 10/11 patients reported planning to (95% CI: 0.96-0.98), p , 0.001]. Conclusions: AOGIC patients experience a higher risk for
either discuss ST with their oncologist (7/11) or obtain ST (5/11). Conclusions: We mortality and longer hospital stays. However, YOGIC patients undergo further procedural
successfully designed an educational tool for pre-test education about ST for Black men interventions and have been found to have higher inpatient admission costs. Policies
with mPCa. This educational tool was well received and may have contributed to PCa focused on earlier outpatient interventions may alleviate the burden on inpatient care.
patients further pursuing somatic testing. Further evaluation of the feasibility, accept- Research Sponsor: This work was supported by funding from the National Cancer Institute.
ability, and efficacy of this educational tool is warranted. Research Sponsor: Prostate The study’s design and decision to publish were independent of any involvement from the
Cancer Foundation. funding sources.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 95s

1593 Poster Session 1594 Poster Session

Assessing the effects of financial toxicity on quality of life among hema- Early post-operative opioid fills after cancer-directed surgery among Medi-
topoietic stem cell transplantation recipients. First Author: Grace Ann Hanvey, care beneficiaries by race and ethnicity. First Author: Ashley Odai-Afotey, Dana-
Mayo Clinic, Rochester, MN Farber Cancer Institute, Boston, MA
Background: “Financial toxicity” refers to the financial burden imposed by treatment Background: Post-operative pain is often underestimated and undertreated. Researchers
costs on individuals with cancer, constituting a major barrier to achieving equitable have shown that Black patients receive fewer opioids than White patients across multiple
cancer outcomes. Recent literature increasingly demonstrates the detrimental impacts of settings and conditions; however, little is known about whether post-operative opioid pre-
financial toxicity on quality of life (QOL) among individuals with cancer, including in- scribing for cancer-directed surgery differs by race and ethnicity. We characterized racial and
dividuals who have undergone hematopoietic stem cell transplantation (HSCT). This ethnic differences in opioid fills among Medicare beneficiaries undergoing cancer-directed
study evaluates associations among treatment cost burden and various aspects of QOL surgery. Methods: Using 100% Medicare data for fee-for-service beneficiaries enrolled in
following HSCT. Methods: Seven hundred one HSCT recipients completed a survey parts A, B, and D, we identified episodes of cancer-directed surgeries from 2012-2021 among
examining their biopsychosocial health one year following transplant. The survey included adults who survived . 30d after surgery and were discharged home. We used Part D claims to
the Functional Assessment of Cancer Therapy – Bone Marrow Transplantation (FACT- identify opioid prescriptions filled in the 30d after outpatient surgeries and the 30d after
BMT), a multifactorial measure of QOL specific to this population. Treatment cost burden hospital discharge for inpatient surgeries, overall and among non-Hispanic White (NHW),
endorsement was measured on a 5-item Likert scale. Hierarchical regression models were Black (NHB), Hispanic, and Asian patients. Results: Among 981,702 surgical episodes (mean
age 73 [SD 8] years, 36% male), 83% were NHW, 8% NHB, 4% Hispanic and 2% Asian patients.
developed to assess the incremental effects of demographic characteristics (i.e., Block 1),
Most surgeries were for breast (38%), colorectal (15%), prostate (13%), or lung (10%) cancers.
clinical predictors (Block 2), and cost burden (Block 3) on physical, emotional, social,
Most (67%) patients with surgical episodes had an opioid fill within 30d, with a mean dose of
functional, BMT-specific, general, and composite QOL outcomes. Results: Significant
246 morphine milligram equivalents (MMEs) in the first prescription, and a mean total dose of
model improvement was observed with the addition of clinical factors (DF(2,650) = 20.28, 360 MMEs filled within 30d of surgery or discharge. On average, patients filled 1.4 opioid
p , .001), and subsequently, treatment cost burden (DF(1,649) = 110.29, p , .001). In the prescriptions, with a mean of 9-days’ supply in the 30d after surgery or discharge. NHB
final model, higher cost burden was associated with poorer physical (b = -0.323, p , patients had the highest rate of opioid prescription, doses, and days’ supply; Asian patients
.001), emotional (b = -0.301, p , .001), social (b = -0.250, p , .001), functional had the lowest rate of opioid prescriptions, doses, and days’ supply (Table). Findings among
(b = -0.317, p , .001), BMT-specific (b = -0.341, p , .001), general (b = -0.377, p , .001), Hispanic patients mirrored NHW patients. Conclusions: Among Medicare beneficiaries
and composite QOL (b = -0.381, p , .001). Poorer performance score was associated with undergoing cancer-directed surgeries in 2012-2021, two-thirds of surgical episodes were
each QOL indicator (p , .001), with allogeneic transplant type associated with poorer associated with an opioid fill. In contrast to prior studies of opioid fills among cancer patients,
functional (b = -0.001, p = .002), but higher emotional (b = 0.118, p = .002), wellbeing. we observed the highest opioid fills among Black patients. Future work is needed to un-
Older age (b = 0.113, p = .003) and female sex predicted higher (b = 0.183, p , .001), while derstand the association of time and patient, physician, and healthcare factors on post-
Hispanic ethnicity predicted poorer (b = -0.095, p = .010), social wellbeing. Female sex operative opioid prescribing and to understand the association of opioid prescribing with pain
was associated with poorer QOL specific to BMT concerns (b = -0.118, p = .001). control. Research Sponsor: National Cancer Institute/U.S. National Institutes of Health.
Conclusions: Higher treatment cost burden is associated with poorer overall QOL and its
Outcome in 30d Overall NHW NHB Hispanic Asian
physical, emotional, social, functional, and BMT-specific components one year following Mean (SD) (n=981702) (n=815966) (n=77932) (n=42663) (n=21087)
HSCT, after controlling for demographic and clinical characteristics. This reflects a critical Filling an opioid 653883 (67%) 538183 (66%) 58323 (75%) 28866 (68%) 12937 (61%)
barrier to equitable cancer care, suggesting that financial toxicity may perpetuate n (%)
preexisting inequities in QOL, treatment, disease, and survival outcomes that dispro- MME, first prescription 246 (320) 245 (318) 271 (350) 243 (318) 211 (210)
Total MME in 30d 360 (566) 356 (563) 420 (620) 358.9 (565) 276 (374)
portionately impact the underserved. Future research should prioritize 1) better un- # of fills 1.4 (0.8) 1.4 (0.8) 1.5 (0.8) 1.4 (0.8) 1.3 (0.6)
derstanding relationships among complex indicators of financial toxicity, QOL, and their Days prescribed 9 (9) 9 (9) 10 (10) 9 (9) 8 (7)
underpinning mechanisms and 2) developing solutions to mitigate financial toxicity of MME=morphine milligram equivalents.
HSCT and overall cancer care. Research Sponsor: U.S. National Institutes of Health.

1595 Poster Session 1596 Poster Session

Disaggregating Asian American and Pacific Islander subgroups to evaluate The development of the cost of cancer in 31 European countries. First Author:
disparities in breast cancer characteristics and outcomes. First Author: Andrea Manzano, IHE, Stockholm, Sweden
Shawn Michael Doss, Medical College of Georgia, Augusta, GA Background: The estimated number of new cancer cases in Europe has risen from 2.6
Background: When Asian American and Pacific Islander (AAPI) subgroups are aggregated million in 1995 to 4.1 million in 2022. Around every fourth death is due to cancer, yet
under the labels “AAPI” or “Asian,” subgroup-specific differences in breast cancer (BC) survival rates have been improving due to advances in early detection, diagnosis, and
presentation and outcomes may be masked. Potential overlooked disparities among these treatment. The implications of these epidemiological changes and medical advances for
diverse groups remain understudied. We analyzed BC characteristics and outcomes among the overall cost of cancer are not well documented. Methods: This cost-of-illness study
regional AAPI subgroups. Methods: From the National Cancer Database (2009–2020), we estimated the direct and indirect costs of cancer across 31 European countries (the EU-27
identified patients diagnosed with stage I–IV BC, excluding patients missing race or stage. countries, plus Iceland, Norway, Switzerland, and the UK) from 1995 to 2023. For direct
Multivariate logistic regression examined odds of advanced stage (III–IV), high-grade his- costs, information on cancer-specific health expenditure was searched for all countries
tology, and triple-negative BC (TNBC) at diagnosis. Multivariate Cox regression assessed three- and combined with data from Eurostat and the OECD. Extrapolations were made for
year overall survival (OS). Age, comorbidity index, diagnosis year, and zip code income quartile countries with missing information. For indirect costs, productivity losses due to pre-
were analyzed to account for confounding. Results: Of 1,956,145 total patients, there were mature mortality were calculated using mortality data from the World Health Organization,
61,731 pooled AAPI patients, comprised of East Asian (n=23,643), South Asian (n=13,642), Eurostat, and the Office for National Statistics, combined with labor market data from
Southeast Asian (n=19,000), and Pacific Islander (PI) (n=5,446) subgroups. Non-Hispanic Eurostat. Productivity loss due to morbidity were estimated using data from prior studies
White (NHW) (n=1,639,814) served as the reference group. Compared to NHW, pooled AAPI had
and changes in population structure. The human-capital approach was employed to
higher odds of advanced stage (adjusted odds ratio [aOR] 1.09; 95% CI 1.07–1.12; p,0.001)
calculate indirect costs. Total costs, costs per capita, and costs per new case were
and high-grade histology (aOR 1.21; 95% CI 1.17–1.25; p,0.001). There was no difference in
estimated for Europe as a whole and for each individual country. Results: Between 1995
odds of TNBC (0.98; 0.92–1.03; p=0.41). Despite this, AAPI had better three-year OS (hazard
ratio [HR] 0.79 (0.76–0.82; p,0.001). Compared to NHW, each AAPI subgroup had higher odds and 2023, the combined direct and indirect costs of cancer across all countries increased
of high-grade histology and better OS except for PIs, who had similar odds of both. South by 43% from EUR 159 billion to EUR 228 billion (all figures adjusted to 2023 prices). Direct
Asians and Southeast Asians had higher odds of advanced stage (aOR 1.27; 1.21–1.34; costs grew by 135%, from EUR 62 to EUR 146 billion, while indirect costs, which fell below
p,0.001 and 1.22; 1.17–1.27; p,0.001, respectively), while East Asians had lower odds (0.86; direct costs after 2005, decreased by 16% from EUR 97 to EUR 82 billion. The decline in
0.83–0.90; p,0.001). South Asians showed increased odds of TNBC (aOR 1.11; 1.00–1.23; indirect costs reflects a reduction in potential years of working life lost due to premature
p=0.04), whereas Southeast Asians (0.82; 0.73–0.91; p,0.001) and PIs (0.81; 0.66–0.98; mortality of working-age patients. Direct costs of cancer accounted for 4-8% of total health
p=0.036) showed lower odds. Conclusions: Disaggregating AAPI subgroup data is necessary expenditures in all countries, with modest increases over time in some but not all countries.
to understand disparities in BC among these heterogenous populations. Further studies are Per capita costs for all countries combined rose by 35% from EUR 313 to 423. However,
warranted to evaluate disparities in healthcare delivery and its efficacy in these subgroups. there was a seven-fold difference between countries in 2023 ranging from around EUR 150
Research Sponsor: None. in Bulgaria and Romania to EUR 1,011 in Switzerland. This disparity represents a reduction
Breast cancer characteristics by AAPI and subgroup.
from the twelve-fold difference observed in 1995. The cost per new cancer case remained
relatively stable in Europe as whole at around EUR 78,000 between 1995 and 2010, before
Advanced Stage High-Grade TNBC Three-Year
(aOR, 95% CI) (aOR, 95% CI) (aOR, 95% CI) OS (HR, 95% CI) declining slightly to EUR 72,000 in 2015 and stabilizing thereafter. Conclusions: The
societal cost of cancer in Europe has been steadily increasing. The growing number of
NHW (Reference) 1.00 1.00 1.00 1.00
AAPI 1.09 (1.07–1.12)** 1.21 (1.17–1.25)** 0.98 (0.92–1.03) 0.79 (0.76–0.82)** cancer patients is spurring this development, rather than changes in cost per patient, which
East Asian 0.86 (0.83–0.90)** 1.12 (1.07–1.18)** 1.06 (0.97–1.15) 0.71 (0.67–0.75)** has remained mostly stable for nearly 30 years. Although the direct costs of cancer have
Southeast Asian 1.22 (1.17–1.27)** 1.30 (1.24–1.38)** 0.82 (0.73–0.91)** 0.84 (0.78–0.89)** risen the most, this has been partially offset by reductions in indirect costs. Changes in
South Asian 1.27 (1.21–1.34)** 1.28 (1.21–1.36)** 1.11 (1.00–1.23)* 0.75 (0.69–0.82)**
PI 1.31 (1.21–1.42)** 1.10 (0.99–1.21) 0.81 (0.66–0.98)* 1.06 (0.96–1.18) indirect costs and epidemiological trends should be considered in debates about the rising
costs of cancer. Research Sponsor: European Federation of Pharmaceutical Industries and
*p-value ,0.05. Associations (EFPIA), Brussels, Belgium (unrestricted grant).
**p-value ,0.001.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
96s CARE DELIVERY/MODELS OF CARE

1597 Poster Session 1598 Poster Session

Trends in female breast cancer among adolescent and young adults in The impact of race on the association between structural racism and the
Southeast Asia. First Author: Jenny Chen, Memorial Sloan Kettering Cancer Center, quality of non-small cell lung cancer (NSCLC). First Author: Jacquelyne Janean
New York, NY Gaddy, Yale School of Medicine, New Haven, CT
Background: Breast cancer is the leading cancer among women globally and poses a Background: Structural racism encompasses multiple intricate systems that generate
growing public health challenge, particularly in adolescents and young adults (AYAs), and reinforce inequities amongst minoritized communities. Given the complexities
defined as individuals aged 15–39 years. In Southeast Asia (SEA), rising breast cancer associated with measuring structural racism, we sought to evaluate the relationship
rates among AYAs are compounded by unique biological, socioeconomic, and healthcare between structural racism and racial inequities amongst Black and White patients with
barriers, including late-stage diagnosis and limited access to screening and treatment. NSCLC using an established structural racism index. Methods: We conducted a ret-
However, regional data on incidence and mortality trends remain scarce. This study aims rospective analysis using Surveillance, Epidemiology, and End Results -Medicare data.
to analyze temporal trends in AYA breast cancer incidence and mortality across 11 SEA Outcomes were: localized stage at diagnosis, stage appropriate evaluation and treat-
countries from 1990 to 2021 using data from the Global Burden of Disease (GBD) ment, and 2-year survival. We used the County Structural Racism (CSR) index, which
database. Methods: We extracted breast cancer incidence and mortality data for AYAs assesses racial inequity within counties across various domains including criminal
in SEA from the GBD database (1990–2021) for Brunei, Cambodia, Indonesia, Lao PDR, justice, education, employment, housing, and health care. We categorized counties into
Malaysia, Myanmar, the Philippines, Singapore, Thailand, Timor-Leste, and Vietnam. quintiles of the CSR index and estimated multivariable mixed effects logistic regression
Age-standardized rates (ASRs) were calculated, and temporal trends were evaluated models to determine the adjusted association between structural racism and each
using Estimated Annual Percent Change (EAPC) based on log-linear regression. outcome. We included interaction terms between patient race (Black versus White) and
Results: Breast cancer incidence among AYAs increased significantly across SEA from CSR to determine whether structural racism moderates the association between patient
1990 to 2021. Thailand reported the highest ASR in 2021 (11.78 per 100,000) and the race and outcomes. We used the results of the regression models to calculate the
most pronounced rise in incidence (EAPC 4.06). Significant increases were also ob- adjusted predicted probabilities of each outcome across strata of patient race and CSR
served in Vietnam (EAPC 2.92), Cambodia (2.63), and Laos (2.57). Mortality trends were quintile. Results: The cohort included 54,344 individuals (10.3% Black, 89.7% White)
heterogeneous: Singapore achieved a significant decline (EAPC -2.00), attributed to diagnosed with NSCLC from 2013-2019. When compared to White patients, Black
advancements in early detection and treatment, while Thailand, Indonesia, Vietnam, and patients were less likely to be diagnosed at a localized stage (30.9% vs 38.4%), undergo
Cambodia experienced rising mortality rates. In 2021, the highest mortality rates were stage appropriate evaluation and treatment (20.3% vs 28.0%), and survive two years
recorded in Myanmar (2.54 per 100,000), Thailand (2.36 per 100,000), and the Philippines after diagnosis (29.2% vs 37.3%) (all p , 0.001). Black patients were more likely to live in
(2.17 per 100,000). Conclusions: The growing burden of AYA breast cancer in SEA counties with higher structural racism (8.2% of the population in lowest quintile vs 19.2%
reflects a combination of epidemiologic transitions, socio-economic shifts, and regional in highest quintile). We did not find a clear association between structural racism and our
healthcare disparities. Rising incidence is linked to changes in reproductive behavior, outcomes. However, we did find that patient race moderated the association between
lifestyle factors, and urbanization, while increased mortality highlights gaps in structural racism and two-year survival. Specifically, Black patients in areas in the lowest
healthcare access and screening infrastructure. Urgent public health interventions quintile of structural racism had a predicted probability of two-year survival of 28.3%
tailored to AYA populations are needed to enhance early detection, improve treatment (95% CI, 25.2-31.4) compared to 31.1% (95% CI, 29.8-32.4) amongst White patients, a
accessibility, and address disparities across SEA. Regional collaboration and invest- difference of 2.8% (p = 0.08). In areas with the highest structural racism, Black patients
ments in healthcare systems are critical to mitigating the growing burden of breast had an even more pronounced reduction in the probability of two-year survival (27.4%,
cancer among AYAs in this dynamic region. Research Sponsor: None. 95% CI, 24.6-30.2 vs. 37.5, 95% CI, 34.9-40.1 for White patients), resulting in a disparity
of 10.1% (p , 0.001). Conclusions: Increased structural racism exacerbates the racial
disparity in two-year survival experienced by Black patients with NSCLC. Research
Sponsor: R01MD017569.

1599 Poster Session 1600 Poster Session

30 year trends in racial disparities for early stage lung cancer treatment. First At-risk cancer genetic syndrome identification (ARCAGEN-ID): Novel EHR
Author: Olivia Frances Lynch, Yale School of Medicine, New Haven, CT integrated system to overcome disparities in identification and testing for
Background: Racial disparities in lung cancer treatment have been recognized for over 30 cancer genetic syndromes. First Author: Vinit Singh, Roswell Park Comprehensive
years. Our prior work showed that among Medicare beneficiaries diagnosed during 1992-2002, Cancer Center, Buffalo, NY
Black patients were less likely to receive curative therapy than White patients. As treatment Background: Identifying individuals at-risk for a hereditary cancer syndrome (HCS) is
approaches have evolved and increased attention has been paid to healthcare disparities, it is crucial to prevent cancer deaths. While there are established guidelines for genetic
unclear whether this pattern has changed over time. We assessed temporal trends in racial testing, less than 30% eligible individuals are tested, with consistently worse rates among
disparities in receipt of curative therapy from 2005 to 2019, and compared findings to estimates underserved. The complexity of guidelines and providers’ unconscious bias contribute to
from over 25 years earlier. Methods: Using the SEER-Medicare data linkage, we conducted a
these disparities. This project aimed to enhance the identification and testing of at-risk
retrospective cohort study of Medicare fee for service beneficiaries diagnosed with stage I-II
NSCLC during 3 time intervals: 2005-07, 2011-13, and 2017-19. Consistent with the prior study,
individuals, focusing on underserved populations. Methods: NCCN/ACMG criteria for
we restricted the sample to Non-Latinx Black and Non-Latinx White patients. Curative therapy genetic testing were translated into three distinct rule-based conditional logic statements
was defined as either surgery and/or radiation within 6 months of diagnosis. In our prior study, in the EHR. A total of 218 rules that serially evaluate each aspect of an individual criteria,
curative treatment was limited to surgery, as radiation was rarely used as curative therapy at and together roll up into a logic statement of “at-risk for HCS. The rules evaluate personal
that time. We performed multivariable logistic regression with receipt of any curative therapy as and/or family history, determine age at onset, and categorize family relationships. A
our outcome, controlling for sociodemographic and clinical covariates. We included a time*race proof-of-concept automated outreach initiative was developed that allowed patients to
interaction to evaluate whether receipt of treatment differed by race over time, and calculated opt into genetic testing after an informational video was watched was developed.
the predicted probability of treatment across time and race group. Results: We identified Relevant data were extracted and compared using chi-square test. Results: Out of 1.3
28,287 patients (7.5% Black) for study inclusion. Black and White patients were similar across million individuals, ARCAGEN-ID identified 59,377 (4.8%) at-risk of an HCS. Of those,
most demographic variables; however, Black patients were more likely to have $3 comorbidities 47,000 (79.2%) had not been previously evaluated: 43,051 (79.3%) at-risk for Breast,
and to have been hospitalized in the year prior to diagnosis. Overall, receipt of curative therapy Ovarian, Pancreas, Prostate related mutation; 3,308 (70.2%) at-risk for Lynch syndrome,
was lower among Black patients (69.9%) compared to White patients (83.3%) throughout the and 1,144 (80.5%) at-risk for other HCSs. Among previously identified individuals, 2,340
study period and across all time intervals (Table). In adjusted analyses, Black patients were less (18.9%) had a pathogenic variant (PV). Compared to overall population in health system,
likely to receive curative treatment in all 3 time intervals, with a Black-White difference of -17.6% ARCAGEN group had a higher proportion of female (82% vs 55%, p , 0.01), White (78% vs
in 2005-07, -14.2% in 2011-13, and -9.7% in 2017-19 (p-value for time*race interaction = , 65%, p , 0.01) and non-Hispanic (89% vs 84%, p , 0.01) individuals, and had less often
0.001). Compared to the 1992-94 and 2000-02 time intervals from the prior study where Black-
Medicaid (16.7% vs 28%, p , 0.01). Within ARCAGEN, comparing previously identified
White difference in receipt of treatment was -11.8% and -14.4% respectively, disparities have
persisted. Conclusions: Racial disparities in receipt of curative treatment for early stage lung
individuals with newly identified ones, the latter were significantly more often male (19.9
cancer in Medicare beneficiaries have persisted 30 years, with minimal improvement. Research vs 11.13%, p , 0.01), younger (#45y) (33.6% vs 27.2%, p , 0.01), Non-White (22.9& vs
Sponsor: None. 20.5%, p , 0.01), and more often on Medicaid (31.5% vs 13%, p , 0.01). For the pilot, 126/
504 outreached individuals (25%) viewed the video and completed a questionnaire. 43/
Receipt of curative treatment by race and time period, adjusted.
504 (8.5%) pursued testing, and 7 (16%) had a PV. A total of 7% had prior testing not
1992-1994a 2000–2002a 2005-2007b 2011-2013b 2017-2019b recorded in discrete fields; 2% declined testing; and 6% sought genetic counseling prior to
Black 73.1 64.9 66.2 (59.6, 71.6 (65.8, 77.3) 77.7 (72.3, 83.1) testing. A higher proportion of African American (AA) individuals opted for testing through
72.7) this strategy (11%) compared to the overall percentage of this population that was
White 84.9 79.3 83.8 (82.4, 85.8 (84.4, 87.1) 87.4 (86.2, 88.6)
85.1) outreached (6%, p = 0.05). Conclusions: Through this automated system, we were able to
Black:White -11.8 -14.4 -17.6 (-24.3, -14.2 (-20.1, -8.3) -9.7 (-15.2, -4.2) identify more non-White individuals and add more Medicaid-insured individuals for
Difference -10.9) testing. Uptake after outreach was higher among AA. Thus, a system like ARCAGEN can
a
Data from prior 2008 study. help overcome disparities in HCS identification without a relevant increase in resources.
b
Current Study. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 97s

1601 Poster Session 1602 Poster Session

Clinical implications and prevalence of benign ethnic neutropenia (BEN) in Community-based patient navigation and preventative care among women
breast cancer patients of Middle Eastern ethnicity. First Author: Shruti surviving breast cancer. First Author: Anthony Zisa, Lombardi Comprehensive
Prem Sudha, Bahrain Oncology Center, Muharraq, Bahrain Cancer Center, Georgetown University Medical Center, Washington, DC
Background: Benign ethnic neutropenia (BEN) commonly affects patients of African and Background: Routine physical exams, mammograms, and Pap smears are essential to
Middle-Eastern descent and is not a true neutropenic state. Patients with BEN have the long-term follow-up for breast cancer survivors, enabling detection of recurrence and
Duffy-null phenotype on red cells and Duffy phenotyping has been used as a surrogate guiding overall health maintenance. Barriers to preventive care adherence over time can
marker for diagnosis. There is evidence that cancer patients with BEN are not at increased be addressed by community-based organizations (CBOs) by providing information and
risk of infection with chemotherapy. The primary aim of this study was to assess the supportive services, including navigation to screening. We studied adherence to pre-
prevalence of BEN among breast cancer patients in Bahrain using Duffy antigen phe- ventive care and screening among breast cancer survivors who engaged a national
notyping on red cells. The secondary aims were to study treatment delays and infectious cancer control CBO–examining how sociodemographics, cancer care factors, and
complications in BEN patients. Methods: We conducted this retrospective study after quality of life (QoL) were associated with adherence. Methods: A secondary data
obtaining IRB approval. We reviewed records of 493 consecutive breast cancer patients analysis was conducted among N = 777 breast cancer survivors who contacted a CBO for
treated in our setting from January 2018 to January 2024. We included patients with resources, including no-cost patient navigation. Patient-reported outcomes were
neutropenia at presentation (defined as having an absolute neutrophil count [ANC] of , assessed after 30 days, along with survivorship care planning (SCP) and QoL. An index
1.5 3103/mL). Patients with Duffy-null phenotype and no identifiable secondary causes of score was created based upon women’s self-reported adherence to receiving routine
neutropenia were presumed to have BEN. Clinical details studied included drug, and family physical exams, mammograms, and Pap smears at recommended intervals.
history, treatment interruptions for neutropenia, filgrastim responsiveness, and episodes Results: Among survivors, 37% were age , = 46, 19% were non-white, 63% were in a
of febrile neutropenia. Overall survival (OS) and progression-free survival (PFS) estimation partnered relationship, 23% rated their QoL (general health) as fair/poor, and 47%
using the Kaplan-Meier method and Cox regression analysis of prognostic factors were carried a pathogenic variant in BRCA. Medical providers caring for these survivors
performed using R software version 4.2.0. Results: Of 493 patients, 72 (14.6%) had a included primary care physicians (53.6%) and oncology specialists (46.4%). For index
presumed diagnosis of BEN. The median age at presentation was 45 yrs, and the median scores, 66% were adherent to all 3 recommendations for follow-up, 29% to 2 recom-
follow-up duration was 3.6 yrs. 13% patients had metastatic disease at presentation and
mendations, and 6% to , = 1 recommendation. The most adhered to recommendation
11% had triple-negative breast cancer (TNBC). The median ANC at diagnosis was 1.2 3
was a physical exam (97%), and the least was a Pap smear (73%): 88% of survivors
103/mL (range 0.4─2.1 3 103/mL). The 4-yr OS was 95% (95% CI, 89–100%) and the 4-yr
reported mammograms at recommended intervals. At the bivariate level, breast cancer
PFS was 75% (95% CI, 63–89%). Treatment was interrupted due to low ANC in 65% of
survivors who were younger (t, df = 4.59, 711, p , .001), non-white, (t, df = -3.27, 267,
patients, and the median ANC at which treatment was delayed was 0.83103/mL. 89%
patients had received filgrastim and all were filgrastim responsive. Only one patient had
p , .001), in a partnered relationship (t, df = 1.76, 54, p , .05), and with better QoL (r =
uncomplicated neutropenic fever. On multivariable analysis, inferior PFS was seen in -.09 p , .01) were more adherent to guideline-based care. A trend was observed for SCP:
patients with metastatic disease (HR, 6.2; 95% CI, 2.17–17.9; p , 0.001), and TNBC (HR, survivors who received care summaries (56%), including follow-up instructions (64%),
7.73; 95% CI 1.79–33.3; p = 0.006). We did not find any effect of treatment delay on the and in written form (45%), were more likely to adhere (r = 0.05, p , .10). In a multi-
PFS. Conclusions: Ethnic neutropenia is prevalent among breast cancer patients in variable regression model adjusting for partnership status and SCP, younger survivors (B
Bahrain. Duffy phenotyping can be used in place of more invasive tests to identify these = 1.13, p , .001), who were non-white (B = 1.0, p , .01), and with better QoL (B = .09, p ,
patients. Treatment delays due to apparent neutropenia are common, however, response .05) were more adherent. Conclusions: Survivors can benefit from guideline-based
to filgrastim is universal, and febrile neutropenia episodes are rarely seen. Since these cancer prevention and screening with CBO-led support. Tailored SCP is essential to
patients are not at increased risk of infection, larger studies to identify unique neutrophil reinforce life-saving health behaviors and enhance follow-up adherence. Research
thresholds for holding chemotherapy in BEN can help avoid compromising therapy. This Sponsor: Centers for Disease Control and Prevention #U58DP005408; #P30CA051008.
can have far-reaching implications in populations with a high prevalence of BEN. Research
Sponsor: None.

1603 Poster Session 1604 Poster Session

Mutation rate differences across populations and association with perfor- Gentian violet compared with methylene blue for sentinel lymph node biopsy
mance disparities in pathology AI diagnostic models. First Author: Po-Jen Lin, in breast cancer: A retrospective analysis from a resource-limited setting.
Department of Biomedical Informatics, Harvard Medical School, Boston, MA First Author: Mehwish Mooghal, Cancer Foundation Hospital, Karachi, Pakistan
Background: Previous studies have established artificial intelligence (AI) algorithms to classify cancer Background: Sentinel lymph node biopsy (SLNB) minimizes morbidity in breast cancer
types, providing real-time diagnostic support. In addition, AI models have identified previously unknown surgeries compared to axillary lymph node dissection. Standard tracers like vital blue
pathology patterns associated with cancer genomic profiles. However, these models exhibit variable (VB), methylene blue (MB) and radioisotopes (RI) are effective but often costly and
performance in different demographic groups, and the causes remain largely unknown. To address this
challenge, we investigated the relationships between biases in AI diagnostic models and mutation rate
logistically challenging in low- and middle-income countries (LMICs). Gentian violet
disparities across populations and evaluated the efficacy of a fairness-aware contrastive learning (GV), a low-cost alternative, offers potential for resource-constrained settings.
(FACL) framework in reducing performance disparities. Methods: We obtained whole-slide pathology Methods: A retrospective cohort study was conducted at Cancer Foundation Hospital
images, mutation rates of the 5 most frequently mutated genes in each cancer type, age, sex, and race Karachi, Pakistan between January 2023 - December 2024. The study included 28 pa-
from 9,217 patients in The Cancer Genome Atlas across 10 cancer types. We identified tasks with tients with breast cancer who underwent SLNB using GV and RI. Sentinel lymph node
performance disparities across demographic groups, and employed generalized linear models to (SLN) detection rates, concordance between GV and RI, and safety profiles were
quantify the relationship between mutation rates and model bias in each cancer type. We further assessed. Detection was analyzed across tumor grades, histology, receptor statuses, and
developed an FACL framework, and evaluated its effectiveness in mitigating these disparities using
neoadjuvant chemotherapy (NACT) status. Results: The majority of patients were aged
metrics including differences in accuracy (DIA) and equal opportunity. Results: Six genomic profile
prediction tasks showed significant performance disparities across population groups (Table). Vari- 41–50 years (n = 8) and . 70 years (n = 8) and had a BMI in the range of 21–30. Tumor
ations in TP53 mutation are associated with differential error rates in serous UCEC v. nonserous UCEC, size analysis revealed that T2 tumors were most common (57.1%, n = 16), followed by T3
mixed IDC v. ILC, LUAD v. LUSC, and GBM v. LGG classification tasks. Differences in CDH1 mutation tumors (21.4%, n = 6). Neoadjuvant chemotherapy (NACT) was administered to 35.7% (n =
rates were linked to racial disparity in mixed IDC v. ILC and age discrepancy in IDC v. ILC classification 10) of patients and all patients were clinically node-negative (cN0) at diagnosis with no
tasks. Our FACL framework mitigated performance disparities across demographic groups in 5 out of 6 distant metastases. Stage II disease predominated (78.6%, n = 22), and invasive ductal
tasks where standard AI model exhibited significant bias (p , 0.05). Conclusions: Biases in AI-driven carcinoma (IDC) was the most common type (71.4%, n = 20), and 53.6% (n = 15) of tumors
cancer pathology diagnosis stem from disparities in somatic mutation prevalence across demographic
were poorly differentiated (Grade 3). Receptor status analysis showed ER/PR positivity in
groups. Addressing these biases is critical to ensuring fairness and the global applicability of AI tools.
Our findings demonstrate that the FACL-based framework effectively reduces performance disparities, 64.3% (n = 18) of cases, triple positivity in 21.4% (n = 6), while HER2/neu-positive and
making AI-powered cancer diagnostics more reliable. Research Sponsor: None. triple-negative subtypes each accounted for 7.1% (n = 2). GV successfully identified SLNs
in 96.4% of cases, with moderate concordance with RI (Kappa = 0.512). GV identified more
Sensitive Groups and Standard (S) v. nodes in 32.1% of patients, while RI identified more in 14.3%; both identified the same
Tasks Mutation Attribute Mutation rates FACL (F) models DIA
number in 50%. The false-negative rate for GV was low (4.2%). Detection rates were
sUCEC v. nsUCEC TP53 Race W 0.34 p,0.001 S: 0.1360.10, p,0.001
B 0.45 F: 0.1060.05, p=0.088 consistent across histological types (e.g., invasive ductal carcinoma: 1.94 nodes by GV vs.
Mixed IDC v. ILC CDH1 Race W 0.29 p,0.001 S: 0.0760.02, p,0.001 RI), tumor grades (Grade 3: 1.78 by both), and receptor statuses (triple-positive cases:
B 0.46 F: 0.1160.04, p=0.233 ~2.4 nodes by GV vs. 2 by RI), with no significant differences. NACT did not impact SLN
TP53 Race W 0.15 p=0.047 S: 0.1260.02, p=0.023
A 0.11 F: 0.1060.05, p=0.196 detection (p = 0.844). No complications or adverse events related to GV dye were ob-
IDC v. ILC CDH1 Age $59 yrs 0.11 p=0.038 S: 0.0560.02, p=0.001 served intraoperatively or during the postoperative follow-up at days 0, 3–7, and 30. The
,59 yrs 0.17 F: 0.0560.05, p=0.370
LUAD v. LUSC TP53 Sex F 0.75 p=0.002 S: 0.1260.02, p,0.001 safety profile of GV dye demonstrated no staining-related complications, dermatitis,
M 0.58 F: 0.0160.01, p=0.154 tattooing, or skin necrosis. Conclusions: Gentian violet is a safe, effective, and affordable
GBM v. LGG TP53 Race W 0.50 p=0.021 S: 0.2060.01, p,0.001 alternative to MB/VB for SLNB in breast cancer. It demonstrates high detection rates
B 0.33 F: 0.2960.02, p=0.005
and a favorable safety profile, making it particularly suitable for LMICs. Broader studies
W: White; are encouraged to validate these findings and further its clinical adoption. Research
B: Black;
A: Asian. Sponsor: No funding received.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
98s CARE DELIVERY/MODELS OF CARE

1605 Poster Session 1606 Poster Session

Comparison of mobile mammography versus urban hospital-based breast Radiotherapy utilization at the fourth most populous province in Indonesia:
cancer screening. First Author: Carla Zeballos Torrez, Perelman School of Medicine, A single centre study. First Author: Vito Filbert Jayalie, Murni Teguh Memorial
University of Pennsylvania, Philadelphia, PA Hospital, Medan, North Sumatra, Indonesia
Background: Breast cancer screening via mobile mammography units (MMU) is used to improve Background: Cancer continues to grow as a health burden in Indonesia, with the current
access in medically underserved communities. This study aims to evaluate factors associated with cumulative cancer risk at 14.0%. Along with the rising rate of cancer discovery,
site of screening, recall rates and time to diagnostic resolution for MMU vs hospital-based sites. treatment effectiveness should be constantly improved. This study was conducted to
Methods: This retrospective study analyzed screening mammography examinations performed in a
explore the actual radiotherapy utilization rate (aRUR) of the 10 most common cancers
MMU and at our large, urban hospital sites during overlapping 2-week periods in 2022 and 2023. BI-
RADS, recall and cancer detection rates were assessed. For BI-RADS 0 patients, time intervals
at Murni Teguh Memorial Hospital (MTMH), a cancer referral hospital in North Sumatra.
between screening and diagnostic imaging and, when indicated, between diagnostic imaging and Methods: This was a retrospective study utilizing MTMH medical records in 2019. Data
biopsy, were collected. Area of Deprivation Index (ADI), an index of socioeconomic status for on the 10 most common cancers in Indonesia (based on GLOBOCAN 2020) were
communities, was calculated for each patient. Diagnostic resolution was defined as time from collected. Completed data underwent double filtering and cleaning before analysis with
screening to completion of diagnostic work-up. Statistical analyses were performed with chi- the Statistical Package for the Social Sciences (SPSS) and Microsoft Excel to calculate
square, analysis of variance, and Kruskal-Wallis tests. Cox regression analysis was used to assess aRUR. Further elaboration was made to compare with the Collaboration for Cancer
factors associated with diagnostic resolution. Results: In the MMU cohort (n=516) vs the hospital- Outcomes Research and Evaluation (CCORE) RUR data. Results: A total of 3,928
based cohort (n=2401), more patients identified as Non-Hispanic Black (68% vs 40%, p , 0.001), samples were collected with 74% of patients being female; the mean age is 50 (0,6 to 96)
reported no insurance (71% vs 2.1% p , 0.001), had no PCP (35% vs 9.8%, p , 0.001), and were in
years old; 38% of cancer staging reported (6% stage I; 32% stage II; 34% stage III; and
the highest ADI percentile (70% vs 27%, p , 0.001). Regardless of screening site, most patients with
longer time to diagnostic resolution had a higher ADI percentile; 58% of patients with . 80 ADI
28% stage IV). The most to least common cancers are breast, colon, rectal, nasopharynx,
percentile (p , 0.001) had diagnostic resolution in . 60 days. The MMU cohort had a higher recall cervix, ovary, leukemia, prostate, lung, and lymphoma. Out of 457 irradiated cases,
rate (18.8% vs 9.9%; p , 0.001) and trend towards a higher cancer detection rate (13.6 vs 8.7 per radiotherapy was most used in breast (36%), nasopharyngeal (23%), and cervical cancer
1000 examinations, p = 0.32) than the hospital-based cohort. Among BI-RADS 0 patients (n=333), (23%). There was gap between these three cancers’ aRUR to their optimal RUR (oRUR)
there were longer delays to diagnostic resolution in the MMU vs the hospital-based cohort (Table 1). calculated by CCORE (54.4% for cervix; 52% for nasopharynx; 41.27% for rectum).
Patients with no insurance were less likely to have diagnostic resolution compared to insured Optimal radiotherapy utilization was reached only for colon and ovarian cancers.
patients (HR: 0.43, 95%CI [0.26,0.71], p = 0.001). Conclusions: Compared to hospital-based Conclusions: Our study shows that a gap between aRUR and oRUR was observed for
screening, MMU-screened patients experienced longer times to diagnostic resolution and had most cancers treated at MTMH. Several factors may contributed to this result, including
higher recall rates. Although MMU offers an effective strategy to improve screening access, our
patient factors, clinical factors, and administrative/bureaucracy factors. Further study is
study highlights opportunities for improved patient navigation, social work support, and financial
assistance to promote more equitable follow-up of abnormal screening mammograms. Research
needed to address the cause and to plan any measures to shorten the gap and optimize
Sponsor: None. radiotherapy. Research Sponsor: None.
BI-RADS 0 outcomes.
Facility Mobile Overall
N = 236 N = 97 N = 333 p-value
Median Days from Screening to Diagnostic (IQR) 11 (7, 20) 28 (13, 43) 13 (7, 28) ,0.001
Median Days from Diagnostic to Biopsy (IQR) 12 (7, 18) 11 (6, 24) 12 (7, 19) 0.5
Median Days to Diagnostic Resolution (IQR) 14 (7, 29) 29 (16, 52) 17 (8, 34) ,0.001
Days to Diagnostic Resolution ,0.001
<=30 178 (75%) 42 (43%) 220 (66%)
30-60 33 (14%) 20 (21%) 53 (16%)
60+ 16 (6.8%) 17 (18%) 33 (9.9%)
No Follow up 9 (3.8%) 18 (19%) 27 (8.1%)

1607 Poster Session 1608 Poster Session

Rates and predictors of cancer screening in California (CA) prisons. First Implementation of a clinical trial navigation program for cancer patients:
Author: Christopher Manz, Dana-Farber Cancer Institute, Boston, MA Barriers and facilitators identified through stakeholder perspectives. First
Background: Cancer is the leading cause of death in state prisons. Patients diagnosed with Author: Milica Paunic, Temerty Faculty of Medicine, University of Toronto, Toronto, ON,
cancer in prison are more likely to have Stage IV diagnoses and have worse survival. Rates and Canada
predictors of cancer screening in prison and the relationship to stage at diagnosis are un- Background: Patient navigation has been highlighted as a solution to improve clinical
known. Methods: This retrospective study evaluated patients incarcerated in CA prisons in trial access. The Clinical Trial Navigator (CTN) Program is a Canadian cancer clinical trial
2014-2023 who met screening criteria for breast, cervical, colon, liver and lung cancers during navigation program that can be accessed online by patients or healthcare professionals
periods tracked by the prison system (Table). Correctional data were used to identify (HCP). Trained individuals search and provide patients and/or oncologists a report of
screening-eligible patients incarcerated during the study period, screening eligibility dates, potentially eligible trials for free. Over 550 patients have used the Program since its
receipt of screening and periods of incarceration. These data were matched to CA cancer launch in 2019, but systemic implementation within cancer centers has yet to occur. We
registry data from 2014-2021 using name and date of birth; cancers were identified between
aimed to identify facilitators and barriers to implementing the CTN Program in Canadian
the start of the tracking period for each cancer and 2021. For each cancer, we calculated the
cancer centers by gathering insights from key stakeholders. Methods: Thirty-three 45-
proportion of: patients who ever received cancer screening, time covered by a screening test
(sum of non-overlapping time of screening intervals [e.g., 10 years for colonoscopy] divided by minute, virtual, semi-structured interviews were conducted with healthcare/clinical
time eligible for screening), and patients diagnosed at Stage IV (stratified by ever-receipt of research professionals (CRP; n = 9) and patient-focused stakeholders (n = 24). Inter-
screening prior to diagnosis). Generalized estimating equations models with logit link adjusted views were guided by the Consolidated Framework for Implementation Research (CFIR)
for demographic and incarceration characteristics (e.g. incarcerated in the past year) clustered and analyzed by two independent researchers using thematic analyses with deductive and
at the prison+yard level were used to determine predictors of receipt of screening for each inductive coding. Results: Participants highlighted the importance of patient navigation
cancer and the association of ever-receipt of screening with Stage IV diagnosis. Results: The to address barriers related to the limited availability of clinical trials and difficulty in
study included 83,174 individuals who were 79% male, had a median age of 51 when first identifying them, noting that navigation can significantly reduce this workload. CRP: “We
eligible for any screening, and were 33% Non-Hispanic White. Rates of ever receiving cancer need a program dedicated to look at trials across the board. [The clinical trial unit team]
screening ranged from 43-87%, and mean proportion of time covered ranged from 30-75% has no time or tools to be able to do this for patients.” Key barriers to implementing
(Table). In adjusted models, receipt of outpatient mental health services, higher security level navigation were the financial and logistical stressors for patients who may want to enroll
and recent change in prison or primary care clinician were associated with higher screening onto trials that the navigator finds, particularly when only available in another institution.
rates for most cancers. 597 screenable cancers were diagnosed from the start of each HCP: “[Our province] covers only travel for the consultation, so [financing] is a big barrier
screening tracking period through 2021. 17% of cancers were diagnosed as Stage IV. In the and needs to be thought through.” Another commonly cited barrier was obtaining the
adjusted model including all cancers, patients who ever received screening prior to diagnosis required medical information for the CTN Program to perform high quality clinical trial
were 60% less likely to be diagnosed with Stage IV disease (OR 0.40, 95% CI 0.24-0.69). searches. Cancer advocacy group leader: “It’s got to be very physician structured because
Conclusions: Cancer screening rates in CA prisons are high and may explain why rates of [the CTN Program intake form] needs patient records. I’ll ask patients what stage they are
Stage IV diagnoses in CA prisons are comparable to the general population and lower than in at and they don’t know, so asking them for their medical information [to perform a clinical
other state prisons. Research Sponsor: None. trial search], they just don’t know that.” When the clinical trial search is initiated by
Breast Cervical Colon Liver Lung patients and the report of potential eligible trials returned to them, patients felt they
Start of tracking period (all end 6/2023) 1/2014 1/2016 1/2014 10/2015 7/2022 needed extra support in discussing the report with their oncologist. Patient: “Every
N 3,218 16,238 68,086 5,644 3,406 oncologist is different. Some are very easy to talk to...one was extremely difficult...so to
Ever screened, n (%) 2,408 (75) 11,326 (70) 49,023 (72) 4,935 (87) 1,479 (43) have a discussion is very difficult.” Conclusions: Our findings provide critical consid-
Mean proportion of time covered 72% 74% 62% 75% 30%
Cancer diagnosed, n 58 12 280 247 - erations for the successful implementation of the CTN Program in cancer centers across
Stage IV at diagnosis, screened, % 5% 0% 21% 12% - Canada. We have planned program adaptations to address these results and will evaluate
Stage IV at diagnosis, not screened, % 7% 0% 35% 21% - changes in uptake and effectiveness of the CTN Program. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 99s

1609 Poster Session 1610 Poster Session

Association of West African ancestry, reproductive factors, and deprivation Effects of socioeconomic status on access to next generation sequencing in
with incidence of triple-negative breast cancer among Black women in the patients with metastatic breast cancer. First Author: Conchita
U.S. First Author: Neha Hippalgaonkar, University of Illinois Chicago, Chicago, IL Martin de Bustamante, UT Southwestern Medical Center, Dallas, TX
Background: Black women have the highest incidence of triple negative breast cancer Background: Metastatic breast cancer is difficult to treat and a major cause of mortality
(TNBC) of any racial or ethnic group in the U.S. The TNBC incidence rate among Black related to breast cancer. Standard treatment includes therapeutic options that target
women varies substantially by state of residence (Sung H, et al, JAMA Oncol 2023), and specific molecular signals and pathways responsible for cancer growth. For metastatic
the underlying factors driving state level variation are unknown. Genetic ancestry of West breast cancer, focused next-generation sequencing (NGS) on DNA isolated from the
Africans and U.S.-born Black Americans (whose ancestry is primarily admixed West tumor tissue or circulating tumor DNA in the blood has quickly become standard of care
African) is significantly different from East Africans, and approximately 10% of the U.S. to create actionable and personalized treatment plans. In ER+ disease, NGS helps to
Black population identify as African immigrants. We used state level data on the number determine potential second-line therapies. However, these tests are often expensive,
of Black residents who identify as East African immigrants to estimate the proportion of limiting their clinical implementation. We hypothesized that limited access to these
the Black population in that state with West African ancestry (defined as not East African therapies increases health disparities in clinical oncology. Methods: Data from 187
immigrant), and we conducted a mediation analysis of state TNBC incidence data to patients with recurrent MBC were obtained from the Dallas Metastatic Breast Cancer
investigate whether West African ancestry, reproductive patterns, and socioeconomic Study, a clinical database that was established in 2021 at a single academic medical
deprivation influence the relationship between race and TNBC incidence. Methods: We system to track patient demographics, area deprivation index (ADI), treatments, and
obtained state level TNBC incidence rates for Black women (2011-2021) from the U.S. other variables. Commercial NGS testing was performed on patient tumor tissue or
Cancer Statistics Public Use database. State level data on country of birth from the 2020 tumor DNA from blood samples by Tempus and FoundationOne between the years 2014
U.S. census, rates of breastfeeding and fertility for Black residents, and socioeconomic through 2022. Results: Overall, 39% of patients in our dataset received NGS testing.
indicators (the 2015 official poverty measure (OPM) and Multidimensional Deprivation Patients who are not Hispanic/Latino (n=140, OR: 3.99, 95% CI: 1.66-9.61) are 4 times
Index (MDI)) were obtained from the U.S. Census Bureau. Correlations between TNBC more likely to receive NGS compared to those who are Hispanic/Latino. We then ex-
incidence rates and variables of interest were tested with Spearman’s correlation test. plored whether ADI correlated with access to NGS testing. ADI measures education level,
Causal mediation analysis of TNBC incidence rate differences was performed by esti- employment, housing quality, and income to rank neighborhoods by SES disadvantage; a
mating coefficients for direct and indirect effects with linear regression models. We higher quartile ADI equates to a greater disadvantage. Our data showed that patients in
calculated estimates of the proportion mediated with 95% confidence intervals (CI) the lowest quartile ADI are 2.5 times more likely to have NGS testing compared to those
accounting for a priori confounders and potential effect modification at the state level. in the highest quartile (OR: 2.54, 95% CI 1.07-6.20). To account for different clinical
Results: State TNBC incidence rates among Black residents were inversely correlated indications for receiving NGS testing, we then looked at NGS trends between tumor
with the proportion of residents identifying as East African immigrants (r = -0.42, p = .006) molecular subtypes. We observed that ER+ patients were 3 times more likely to have
and the rate of breastfeeding (r = -0.35, p = .03). There was no correlation with fertility NGS testing compared to ER- patients (n=118, OR: 2.77, 95% CI: 1.45-5.29) and that
rates, OPM or MDI. In unadjusted analyses, East African immigrant proportion at the state TNBC patients were less likely to receive NGS testing compared to ER+ patients,
level mediated 15.3% (p = 0.01) of the differences in TNBC incidence rates. After ad- however this difference was not significant (n=113, OR: 0.36, 95% CI: 0.13-1.01). In ER+
justment for rates of breastfeeding, fertility, and socioeconomic indicators, East African patient population, we found that Hispanics/Latinos were 79% less likely to undergo NGS
immigrant proportion was associated with 30.1% (p = 0.02) of the difference in TNBC testing compared to their non-Hispanic counterparts (n=76, OR: 0.21, 95% CI: 0.06-0.73).
incidence rates for Black women at the state level. Conclusions: Proportion of East Conclusions: These results suggest that even in clinically indicated ER+ disease, NGS
African immigrants and rate of breastfeeding are inversely correlated with TNBC inci- testing is disproportionately offered to patients with a higher SES, particularly those who
dence rates in Black women. These factors transmit a portion of state level differences in are not Hispanic/Latino. Whether these discrepancies stem from the recent adoption of
TNBC incidence, suggesting that West African ancestry partially mediates the higher NGS as standard of care or from actual barriers to accessing care should be defined in
incidence of TNBC in Black women. Research Sponsor: None. future studies. However, identifying that these disparities exist promotes awareness for
clinicians to offer NGS more broadly. Research Sponsor: None.

1611 Poster Session 1612 Poster Session

Trust and communication among sexual and gender minority (SGM) cancer Quantifying financial toxicity in oncology: A comprehensive analysis of
survivors. First Author: Brandon M. Godinich, Texas Tech Health Science Center El prescription cost disparities using public data from 2022. First Author:
Paso, El Paso, TX Charishma Bhimineni, Jefferson Einstein Montgomery, East Norriton, PA
Background: Trust and effective communication in healthcare are essential for delivering Background: Financial toxicity in cancer care poses a significant burden for patients
high-quality cancer care, especially for marginalized groups like SGM patients. This study and healthcare systems. This study analyzed public data to evaluate the financial impact
examines the differences in trust and communication among LGBTQIA cancer survivors of oncology treatments, focusing on factors such as demographics, insurance type,
and those without (w/o) a cancer history to inform strategies for improving cancer care gender disparities, or geographics. Methods: The 2022 Medical Expenditure Panel
equity. Methods: Data from the nationally representative Health Information National Survey (MEPS) data was analyzed, focusing on antineoplastic and immunologic agents.
Trends Survey (HINTS) from 2018-2022 was used to evaluate questions on communi- Cost distributions, payment sources, and financial burden—defined as prescription costs
cation, quality of care, and trust in those who self-identified as homosexual, gay, or exceeding 20% of annual household income—were assessed across various insurances
lesbian, bisexual, or “something else” for sexual orientation. Two cohorts compared SGM and demographic groups (age, gender, race, income, and education). Geographic an-
participants with prior cancer diagnosis (survivors) and those w/o cancer. Demographic alyses utilized Federal Information Processing Standard (FIPS) state codes.
data included: age, gender, race/ethnicity, education, employment, and household in- Results: The study included 1,379 oncology prescriptions. Prescription costs averaged
come. Analysis was done in STATA with Chi-squared and T-tests testing between SGM 1,569 per prescription (median 388), ranging from 19 to 7,208. Annual costs spanned
survivors and those w/o cancer and multivariate analysis (MVA) focused on SGM sur- 2,507 (25th percentile) to 26,857 (75th percentile), with a median of 3,561 and a mean of
vivors. Results: In total, 1,258 SGM participants were included, of which 144 (11.4%) 14,138. Non-prescription medical costs, such as procedures and hospitalizations, had a
were SGM cancer survivors. SGM survivors were older than those w/o cancer history median of 45,475 and a mean of 47,509, with some exceeding 132,396 annually. Un-
(median 64 vs. 46, p , 0.001) but had no significant differences in employment (25.0% vs. insured patients faced the highest average costs (78,439 annually), followed by
47.2%) or race (White: 79.9% vs 69.5%; Black 21.9% vs 19.6%) (p = NS). Less than half of Medicare patients (67,979). Medicaid patients had the lowest total costs (53,469). VA/
SGM survivors (44.9%) reported they always or usually had the chance to ask all of their TRICARE patients showed moderate costs (56,619) but higher prescription expenses
health-related questions during provider visits, this was better than SGM w/o cancer (16,758). Low-income patients faced the greatest financial burden, spending 11.71% of
(36.8%, p = 0.03). About a third reported that providers always/usually gave adequate their income on prescriptions, compared to 5.89% for middle-income and 2.66% for high-
attention to their emotions and feelings (37.2% survivors vs 37.5% w/o cancer, p = NS). income patients. Private insurance beneficiaries faced the highest costs (5,500-6,000),
Most felt they were always/usually adequately involved in decisions about their health particularly among Black and White patients, followed by Medicare (4,500-5,000).
care (82.5% survivors vs 82.0% w/o cancer, p = NS). Only 33.1% SGM survivors rated their Medicaid beneficiaries incurred lower costs (3,500-4,000), while uninsured patients had
overall quality of care as excellent/very good within the past year; this was slightly better the lowest mean costs (2,500-3,000), reflecting limited access to comprehensive
than surveyed SGM w/o cancer (24.6%, p = NS). Half of SGM survivors (50.9%) trusted treatments. Female patients had higher costs for breast and lung cancer treatments.
information about cancer from doctors, slightly more than for those w/o cancer (46.2%, p Black and Hispanic patients relied more on Medicaid. Patients with graduate degrees
= NS). In a MVA limited to SGM survivors, only education was associated with decreased had higher average costs (4,226) than those with a high school education or less (3,707).
trust of cancer information from a doctor; those with at least some college (OR = 0.51 95% Geographically, financial burden was highest in the Midwest, moderate in the Northeast,
CI 0.26-0.99, p = 0.048) or postgraduate education (OR = 0.36 95%CI 0.14-0.92, p = 0.034) and mixed in coastal and Southern states. Conclusions: Significant disparities in fi-
had less trust compared to those with a high school degree or less. Conclusions: This nancial toxicity exist across demographics, insurance types, and regions. The gap
national study shows that patient-reported overall healthcare to SGM survivors is poor. between mean and median costs underscores the disproportionate financial strain faced
Less than a third of SGM survivors reported good quality of care and less than half felt by some patients. Policy interventions, including expanded insurance coverage, capped
providers answered all their questions; only half trusted cancer information from a doctor. out-of-pocket costs, and targeted subsidies, are needed to improve equity and af-
Concerningly, those with higher education levels were less likely to trust doctors. Future fordability in cancer care. Research Sponsor: None.
efforts should focus on ensuring that all patients benefit from high quality cancer care and
communication. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
100s CARE DELIVERY/MODELS OF CARE

1613 Poster Session 1614 Poster Session

Racial differences in cardiovascular outcomes among cancer patients re- Impact of sociodemographic factors and Medicaid expansion on postop-
ceiving immune checkpoint inhibitors. First Author: Cho Han Chiang, Department erative outcomes for glioblastoma, 2004-2021. First Author: Bhav Jain, Stanford
of Medicine, Mount Auburn Hospital, Harvard Medical School, Cambridge, MA University School of Medicine, Stanford, CA
Background: Immune checkpoint inhibitors (ICIs) increase the risk of major adverse cardio- Background: Glioblastoma (GBM), the most aggressive primary brain tumor in adults,
vascular events (MACE). We aimed to evaluate disparities in MACE across racial groups. has a median survival of ~15 months despite treatment and exhibits significant dis-
Methods: We conducted a propensity score-matched study using the TriNetX Analytics Network parities in care access. Sociodemographic factors and policy interventions, such as
database, which includes de-identified data from over 140 healthcare institutions. Adult cancer Medicaid expansion under the ACA, show potential to mitigate inequities in other
patients treated with ICIs were included and those with prior MACE were excluded. Patients were
cancers. However, their impact on GBM outcomes remains underexplored.
grouped into White, Black, Asian, and Hispanic cohorts. The primary outcome was incident MACE,
defined as the composite of myocarditis, pericarditis, myocardial infarction, ischemic stroke, heart Methods: Using the National Cancer Database, we conducted a retrospective study of
failure, atrial fibrillation, and venous thromboembolism (VTE) within 12 months of ICI. Matching 85,631 GBM patients treated with surgery between 2004 and 2021. Multivariate re-
was performed using variables: age, sex, cancer type, metastatic disease, comorbidities, and gression models and Kaplan-Meier survival analyses evaluated associations between
cardiovascular medication. Results: A total of 58,217 eligible patients were identified, including sociodemographic factors (e.g., race, income, education, rurality, insurance status) and
44,151 White, 5,876 Black, 5,347 Asian, and 2,843 Hispanic individuals. After matching, cohorts outcomes, including postoperative hospital stay, 30-day readmission, 90-day mortality,
were adequately balanced. Black patients had the highest risk for MACE, with an 18% increased and overall survival. All models adjusted for key clinical (e.g., tumor size, comorbidities,
risk compared to White (HR 1.18 [95% CI: 1.06-1.30]) and Hispanic patients (HR 1.18 [95% CI: 1.03- receipt of chemotherapy/radiation therapy) and patient (e.g., age, sex) covariates. A
1.35]) and an 80% increased risk compared to Asian patients (HR 1.80 [95% CI: 1.57-1.35]). This difference-in-differences analysis assessed the effects of Medicaid expansion on these
increased risk among Black patients appeared to be driven by higher rates of heart failure and VTE. outcomes. Results: Regarding postoperative length of hospital stay, disparities were
The risks of MACE were similar between White and Hispanic individuals while Asian patients had
observed by race (Black vs. White b = 1.45 days [1.22–1.68]; Asian American and Pacific
the lowest risk. Conclusions: There were racial differences in immune-related MACE, with Black
patients experiencing the highest risk of cardiotoxicity following ICI treatment. Research Sponsor:
Islander [AAPI] vs. White b = 0.86 days [0.50–1.22]), rurality (urban vs. metro b =
None. -0.31 days [-0.47 to -0.15]), insurance status (private vs. uninsured b = -1.10 days [-1.41
to -0.80]), and education (highest vs. lowest quartile b = -0.28 days [-0.48 to -0.09]).
Hazard ratio for effects of race on cardiovascular outcomes.
Unplanned 30-day hospital readmission rates demonstrated disparities by race (Black
Black vs. Black vs. Black vs. White vs. White vs. Hispanic vs.
White Hispanic Asian Asian Hispanic Asian
vs. White OR = 1.19 [1.04–1.35]), income (highest vs. lowest quartile OR = 0.84
[0.75–0.96]), and education (highest vs. lowest quartile OR = 1.19 [1.05–1.34]).
Outcomes n=5,109 n=2,636 n=3,095 n=2,842 n=4,876 n=2,157
each each each each each each Moreover, 90-day mortality indicated disparities by race (Black vs. White OR = 0.85
MACE 1.18 1.18 1.80 1.44 0.98 1.60 [0.77–0.95]; AAPI vs. White OR = 0.64 [0.53–0.77]), income (highest vs. lowest quartile
(1.06-1.30) (1.03-1.35) (1.57-2.07) (1.28-1.62) (0.86-1.12) (1.35-1.91) OR = 0.81 [0.74–0.89]), education (highest vs. lowest quartile OR = 1.13 [1.03–1.23]),
Myocarditis 0.50 0.16 0.61 1.85 1.08 1.66
(0.20-1.23) (0.04-0.69) (0.15-2.57) (0.74-4.63) (0.49-2.37) (0.61-4.58) and insurance status (private vs. uninsured OR = 0.71 [0.62–0.82]). Finally, overall
Pericarditis 1.10 1.36 2.31 2.99 1.20 0.99 survival demonstrated disparities by race (Black vs. White HR = 0.88 [0.85–0.91]; AAPI
(0.47-2.56) (0.31-6.09) (0.71-7.51) (1.09-8.21) (0.37-3.92) (0.20-4.91)
Myocardial infarction 1.09 1.18 1.80 1.42 1.18 1.47
vs. White HR = 0.77 [0.73–0.82]), income (highest vs. lowest quartile HR = 0.83
(0.86-1.39) (0.83-1.67) (1.28-2.53) (1.06-1.91) (0.84-1.65) (0.96-2.24) [0.81–0.86]), education (highest vs. lowest quartile HR = 1.12 [1.09–1.15]), rurality (rural
Ischemic stroke 1.23 1.13 1.34 1.17 0.81 1.45 vs. metro HR = 1.06 [1.00–1.12]), and insurance status (Medicaid vs. no insurance HR =
(0.99-1.54) (0.83-1.54) (1.00-1.80) (0.91-1.50) (0.59-1.10) (1.01-2.08)
Heart failure 1.33 1.32 1.78 1.38 1.05 1.33 1.09 [1.04–1.15]). Medicaid expansion did not significantly impact any outcomes, in-
(1.11-1.61) (1.01-1.73) (1.37-2.31) (1.10-1.73) (0.80-1.39) (0.96-1.84) cluding overall survival (DID HR = 0.95 [0.84–1.07]). Conclusions: Significant socio-
Atrial fibrillation 0.96 1.19 1.29 1.36 1.26 0.95 demographic disparities persist in GBM postoperative outcomes, with no improvement
(0.79-1.16) (0.86-1.64) (0.98-1.70) (1.08-1.70) (0.93-1.70) (0.66-1.36)
Venous thromboembolism 1.28 1.19 2.34 1.62 0.97 2.09 from Medicaid expansion. Targeted socioeconomic interventions are needed to address
(1.12-1.47) (1.00-1.41) (1.92-2.84) (1.37-1.91) (0.81-1.15) (1.63-2.69) inequities in access to specialized neuro-oncological care and improve outcomes for
underserved populations. Research Sponsor: None.

1615 Poster Session 1616 Poster Session

Symptom burden, quality of life (QoL), social and behavioral characteristics Prevalence of and factors associated with financial toxicity among gastro-
in young patients (<40 years old) with cancer: A prospective cohort of 7323 intestinal cancer patients in Pakistan. First Author: Sehar Salim Virani, Aga Khan
patients across 110 sites in France and Belgium. First Author: Kaı̈ssa Ouali, University Hospital, Karachi, Pakistan
Gustave Roussy, Drug Development Department (DITEP), Villejuif, France Background: Financial toxicity (FT) impacts cancer care in low- and middle-income countries
Background: Cancer in individuals under 40 years old is increasingly recognized as a (LMICs), affecting treatment adherence and quality of life. This study assesses FT prevalence and
public health concern, characterized by unique etiologies, biology, and clinical behaviors associated factors among gastrointestinal (GI) cancer patients across distinct healthcare systems in
compared to older populations. Young patients may also face specific physical, psy- Pakistan. Methods: A cross-sectional study was conducted across three tertiary care centers in
Karachi: Aga Khan University Hospital (AKUH, private, fee-for-service), Jinnah Postgraduate Medical
chosocial, and socioeconomic challenges that can influence outcomes which are often Center (JPMC, public, free), and Cancer Foundation Hospital (CFH, private-philanthropy, subsidized).
suboptimally addressed in routine care. Digital health and specifically remote patient FT was assessed using the Urdu version of the Comprehensive Score for Financial Toxicity–
monitoring (RPM) offer a way to track and manage these challenges effectively, in- Functional Assessment of Chronic Illness Therapy (COST-FACIT). Multivariable negative binomial
creasing access to supportive care. Methods: Prospective, observational cohort of 7323 regression identified factors linked to high FT. Results: Of 375 patients, 44.5% were from AKUH,
adult patients with cancer participating in an RPM pathway across 110 hospitals in France 33.6% from JPMC, and 21.9% from CFH. Mean age was 50.8 6 14.4 years, with 62.4% males. Only
and Belgium between Jun-2022 and Dec-2024. Patients were grouped by age ( , 40 8.3% had health insurance, and the median International Wealth Index (IWI) was 79.9 (IQR:
vs. $40). At baseline, demographic, clinical, and social and behavioral data were col- 57.1–95.1). Catastrophic healthcare expenditure affected 41.7%. The mean COST-FACIT score was
lected. Longitudinal symptom burden (e.g., anxiety, pain, nausea) was assessed using 16.0 6 7.4, with 46.1% experiencing mild FT (score: 14–26) and 41.9% moderate FT (score: #14).
validated patient-reported outcomes (PRO-CTCAE). High symptom burden was defined as Patients delaying or forgoing care had higher FT (p , 0.001). Borrowing money, selling assets, or
cutting essential expenses were strongly associated with increased FT (p , 0.001). Patients at
PRO-CTCAE grade $3. QoL was measured by EORCT QLQ-C30 summary score. Linear
AKUH reported higher FT than JPMC (IRR = 0.84, 95% CI: 0.74–0.97). Younger patients (21–50
mixed models, adjusted for relevant covariates, were used to compare changes in years) (IRR = 0.66, 95% CI: 0.46–0.95) and those receiving chemotherapy (IRR = 0.89, 95% CI:
symptom burden over 12 weeks between age groups. Results: Younger patients (n = 350 0.81–0.98) experienced higher FT. Females (IRR = 1.36, 95% CI: 1.17–1.58) and higher socio-
pts , 40 years) were more often female (77 vs 62%, p , 0.001) and with localized disease economic status (IRR = 1.39, 95% CI: 1.06–1.83) were associated with lower FT.
(55 vs 45%, p , 0.001). The rate of breast, lung, colorectal and pancreatic cancers Conclusions: Nearly 85% of GI cancer patients faced FT. Younger age, male gender, lower so-
represented respectively 157%, 27%, 57% and 17% of the proportion in older adults (p , cioeconomic status, and systemic therapy were associated with higher FT. Subsidized care, financial
0.001). Younger patients also presented with more unfavorable social and behavioral support, and institution-specific strategies are critical to mitigating FT in LMIC healthcare systems.
characteristics including higher alcohol consumption (26.3 vs. 15.3%, p= 0.005), tobacco Research Sponsor: None.
consumption (23.2 vs. 13.8%, p , 0.001), unemployment (17.7 vs 7.2%, p , 0.001) and Patient characteristics and treatment costs across hospitals.
financial insecurity (23.2 vs 10.9%, p , 0.001). Similar QoL was found at baseline (mean AKUH JPMC CFH
[SD] 76.9 [16.8] vs 76.1 [17.0], p = 0.84). Adherence to RPM surveys was lower in the Gender, n (%)
younger group (74 vs 84%, p = 0.001), who took also on average longer to answer (22 vs Male 109 (65.3) 73 (57.9) 52 (63.4)
10h, p,0.001). Younger patients had higher early (week 1 to 3) symptom burden (anxiety Female 58 (34.7) 53 (42.1) 30 (36.6)
Age (yrs), mean (SD) 54.3 (1.1) 44.2 (1.3) 53.6 (1.4)
16.7 vs 10.6%, p = 0.001], fatigue [36.1 vs 30.6, p = 0.05], nausea [30.4 vs 18.6%, p , IWI score, mean (SD) 84.0 (1.5) 52.8 (2.6) 80.4 (2.0)
0.001], anorexia [16.7 vs 10.6, p = 0.004] and performance status decline [26.8 vs 18.7%, p COST-FACIT score, mean (SD) 16.7 (7.9) 15.3 (7.6) 15.7 (5.8)
EORTC QLQ summary score 78.1 (15.5) 79.3 (17.9) 75.1 (16.0)
= 0.012). At 12 weeks, symptom burden improved in both groups, and the between-group (Quality of Life score), mean (SD)
difference was no longer significant. Conclusions: In this large, multi-institutional cohort, Monthly household income 358.4 (179.2-716.8) 107.5 (43.0-179.2) 233.0 (71.7-358.4)
younger patients faced unique physical, psychological and behavioral challenges and (USD), median (IQR)
Out of pocket costs (USD), median (IQR)
experienced higher early symptom burden. Interestingly, by 12 weeks, both groups Surgery and associated inpatient 2509.0 (1003.6-4569.89) 0 (0-0) 1433.7 (1075.3-1792.1)
demonstrated symptomatic improvement, with no remaining differential across age Chemotherapy 1433.7 (573.5-3225.8) 0 (0-304.7) 1469.5 (358.4-1881.7)
Radiotherapy 896.1 (255.4-2150.5) 0 (0-0) 1075.3 (716.8-1792.1)
groups. These findings suggest that RPM and supportive interventions may help mitigate
disparities in symptom burden over time. Research Sponsor: Resilience Care. *Costs converted using 1 USD = 279 Pakistani Rupee.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 101s

1617 Poster Session 1618 Poster Session

The effect of a vertically integrated health system on the disparity of Dihydropyrimidine dehydrogenase (DPD) deficiency–related variants
socioeconomic status seen in cancer stage at presentation. First Author: among Mexican patients with gastrointestinal (GI) malignancies. First Au-
Robert Michael Cooper, Kaiser Permanente Los Angeles Medical Center, Los Angeles, CA thor: Enrique Soto Pérez de Celis, University of Colorado Anschutz Medical Campus,
Background: Earlier stage of diagnosis may lead to more curable disease and less Aurora, CO
intensive treatment. Vertically integrated health care systems through screening and Background: DPD deficiency is the most important risk factor for developing
integrated care delivery model may provide benefit in identifying cancer patients at fluoropyrimidine-related adverse events. Genetic variants causing DPD deficiency are
earlier stage of disease. Methods: We examined an insured Southern California cohort found in 6-8% of Caucasian patients. However, there is limited information on their
of 503,279 patients diagnosed with invasive cancer between Jan 1 2015 and Dec 31 prevalence in underrepresented ethnic groups, such as Hispanics and Latinos, and testing
2020 provided by the state SEER Cancer Registry. Stage at diagnosis provided was SEER for these variants is not routinely recommended in Latin America. Our goal was to assess
summary stage in which patients were defined as having local disease or advanced the allele frequency of clinically actionable dihydropyrimidine dehydrogenase (DPYD) risk
(regional or metastatic) disease. We used geocoded socioeconomic status, race/ variants defined by the Clinical Pharmacogenetics Implementation Consortium (CPIC) and
ethnicity, and hospital of diagnosis as independent variables. For hospital of diagno- the European Medicines Agency (EMA) among admixed Mexican patients with GI ma-
sis, patients were divided into those diagnosed in Kaiser Foundation Hospitals or not. lignancies. Methods: Patients with recently diagnosed GI cancer candidates for flur-
The first cohort was for cancers with robust screening programs and included breast, opyrimidine therapy were recruited from a single institution in Mexico City. After providing
cervical and colon (CBC group). The second cohort was all other cancers (non CBC informed consent, a blood sample and clinical characteristics were collected. We utilized
group). We evaluated each of these groups independently. The prevalence of local the Illumina Infinium Global Screening Array (GSA)-to genotype 34 DPYD variants, six of
disease was determined and multilinear regression was used to determine the adjusted which are known to lead to an increased risk of fluoropyrimidine toxicity and are con-
odd ratios of being diagnosed with local disease. Results: Compared to patients not sidered clinically actionable. Results: Two hundred and eight patients with a mean age of
diagnosed in Kaiser Foundation hospitals (non KFH), patients diagnosed in Kaiser 62 years (SD 13.2) were included. 47% were female. The most common type of cancer was
Foundation hospitals (KFH) were more likely to be diagnosed with local disease. 1.14 colorectal (38%) followed by pancreas (22%) and biliary tract (18%). DNA samples from 192
(95% Confidence Intervals 1.13, 1.16) for the total cohort. We looked at the adjusted patients passed quality control, of which 156 (62%) received fluoropyrimidines during
interaction of being diagnosed in KFH for each quintile: Highest SES 1.03 (1.00, 1.07), follow-up. Only 2 patients (1%) were heterozygous for actionable DPYD intermediate
metabolizer risk variant alleles: one with c.2846A . T (rs67376798, D949V) and one with
Upper Middle SES 1.11 (1.08, 1.14), Middle SES 1.21 (1.18, 1.25), Lower-Middle SES 1.25
c.1129–5923C . G [rs75017182; HapB3 SNP c.1236G . A; rs56038477]. No patients were
(1.21, 1.29) and Lowest 1.34 (1.29, 1.38). Compared to patients not diagnosed in Kaiser
found to have other CPIC-listed DPYD risk variants. Additionally, we investigated the allele
Foundation hospitals (non KFH), patients diagnosed in Kaiser Foundation hospitals
frequencies of other 30 DPYD variants and observed low-frequency variation (between
(KFH) were more likely to be diagnosed with local disease in the CBC cohort 1.12 (1.09,
0.260 and 0.0032) in rs56038477, rs1801160, rs17376848, rs1801159, rs1801158,
1.16) and in the non CBC cohort 1.16 (1.14, 1.17). Conclusions: Vertically integrated rs45589337, rs2297595, rs200562975, and rs1801265. Several of these may be related to
health care systems have shown advantages in preventive care. We show that insured decreased DPYD activity and warrant further analysis regarding their impact on adverse
patients diagnosed in Kaiser Foundation hospitals present with more localized and less drug reactions. Conclusions: In contrast with reports from Caucasic populations, we
advanced disease than patients diagnosed in non KF hospitals. This advantage was seen found a very low allele frequency of DPYD actionable variants. Our findings highlight the
in a group of patients with established screening program and was also seen in diseases limitation of current pharmacogenomic testing recommendations and panels, which may
without screening programs. A sub analysis showed that the advantage more pro- not be appropriate for admixed ethnic populations such as Hispanics/Latinos due to
nounced the lower the SES. How a vertically integrated care delivery system provides disparities in representation. There is a need to study the role of other DPYD variants in
these advantages deserves further study. Research Sponsor: None. larger patient samples to understand their role in the toxicity risk of admixed populations in
Mexico and Latin America, to explore the use of novel techniques such as Next Generation
Sequencing, and to investigate the effect of other related genes on toxicity risk. Research
Sponsor: AGA Research Foundation.

1619 Poster Session 1620 Poster Session

State-level trends and associated disparities in melanoma burden in the Association of allostatic load (AL) and residential segregation with breast
United States. First Author: Furkan Bahar, Mount Auburn Hospital/Harvard Medical biopsy outcomes after screening mammography. First Author: Braelyn Wek-
School, Cambridge, MA werth, Massachusetts General Hospital, Boston, MA
Background: Melanoma, an aggressive skin cancer, poses a significant public health Background: Allostatic Load (AL) and residential segregation have been associated
challenge in the United States despite advancements in detection and treatment due to with the risk of breast cancer (BC). However, the independent effects of AL and
its high mortality. This study analyzes trends in melanoma incidence, mortality, and measures of residential segregation (MRSs) on cancer detection and the false positive
disease burden in the US from 1990 to 2021. Methods: Data on incidence rates (IR), (FP) biopsy rate in a screening mammography population have not yet been assessed.
mortality rates (MR), disability-adjusted life years (DALYs), and estimated annual Methods: We retrospectively identified women aged ³40 who underwent screening
percentage changes (EAPCs) from 1990 to 2021 were extracted from the Global Burden mammography between 1/1/2021-12/31/2021 and subsequent breast biopsy from the
of Disease 2021 database. Regional trends within the US were evaluated to identify Mass General Brigham Biobank. We collected age and self-reported race/ethnicity. Each
state-level patterns and disparities. Results: In 2021, melanoma IR in the US was 27.2 participant’s zip code was geocoded to the corresponding census tract. We computed
per 100,000, far exceeding the global average of 3.8. Between 1990 and 2021, IR in- five MRS indices: Dissimilarity (DD), Isolation (BI), Delta (D), Absolute Centralization
creased by 64.8%, compared to a global rise of 28.8%. Among US states, Maine had the (AC), Spatial Proximity (SP). We collected the following biomarkers obtained within two
highest IR at 37.8 per 100,000, while the District of Columbia reported the lowest IR at years before the index screen: cardiovascular, metabolic, immunologic, renal lab values.
8.3. Alaska showed the largest relative increase (EAPC of 81.5%), while New Jersey was AL was assigned one point for each lab value in the worst quartile and summed
the only state to report a decline in incidence, with an EAPC of -1.6%. The US melanoma (continuous). We collected diagnostic breast imaging and biopsy encounters within
mortality rate was 3.0 per 100,000 in 2021, compared to the global rate of 0.78. US 12 months after the index screen. Multiple imputation accounted for missing data.
mortality increased by 5.5%, while the global mortality rate rose by 25.8%. West Virginia Multivariable logistic regression assessed age, race, AL and each of our MRSs asso-
recorded the highest mortality rate at 4.49 per 100,000, whereas the District of Columbia ciation with cancer detection and FP rates. We applied Rubin’s rules to estimate overall
had the lowest rate at 1.20 and the most significant improvement, with an EAPC of odds ratios (OR), confidence intervals (CI), and p-values for all covariates. Results: Of
-43.5%. In terms of DALYs, the US reported a rate of 79.2 per 100,000 in 2021, sig- the 418 eligible women, 59.6% (N=249) had an FP biopsy, and 66.3% (N=277) had breast
nificantly higher than the global average of 21.27. Between 1990 and 2021, DALYs in the cancer, including cases of ductal carcinoma in situ. On average, women were 62 years
US declined by 14.4%, contrasting with an 8.5% increase globally. The DALY/incidence old (SD=13); 85.6% White. DD was associated with a reduced risk of benign high-risk
ratio in 2021 was 2.9 in the US, compared to 5.6 globally, indicating notable differences lesions (OR=0.69, 95% CI:[0.49-0.95]; p=0.025), and homogeneous, affluent census
in disease burden across populations. Melanoma disproportionately affected males, with tracts—whether predominantly Black or White—were similarly protective (OR=0.85, 95%
an IR of 33.9 per 100,000 compared to 20.7 in females (male-to-female ratio: 1.6). Cl:[0.72, 0.99]; p=0.041). MRS indices were linked to lower benign high-risk outcomes
Mortality rates followed a similar pattern, with males experiencing a rate of 4.0 per (e.g. SP, OR=0.53, 95% CI:[0.31, 0.91]; p=0.021). Age and race significantly predicted
100,000 vs. 2.1 for females, consistent with global trends. Conclusions: Despite the adverse events (AEs). Older age was consistently associated with increased AEs across
decrease in DALYs, melanoma incidence and mortality continue to rise in the US, all models (OR = 1.09, 95% CI:[1.00, 1.18]). Cancer detection also increased with age (OR
exceeding global averages. Significant disparities persist across states and genders, = 1.20, 95% CI: [1.10, 1.30]; p , 0.001). AL was significantly linked to cancer detection
reflecting the complex interplay of risk factors and behavioral patterns. The lower DALY/ (OR = 1.13, 95% CI:[1.00, 1.28]). Conclusions: Some MRSs are associated with cancer
incidence ratio in the United States highlights the likely effectiveness of US treatment detection and high-risk FP. AL remains associated with these cancer and high-risk FP,
options in mitigating disease burden per case. Variations among states may also be even accounting for these segregation measures. These factors may contribute to an
attributed to differences in ethnic distribution, which influence genetic susceptibility, increased risk of cancer, highlighting the significance of spatial and socioeconomic
healthcare access, and prevention efforts. Research Sponsor: None. influences on screening outcomes. Clinical Relevance Statement: AL may serve as a
biomarker to enhance biopsy selection following screen-detected mammographic
abnormalities, potentially improving cancer detection rates. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
102s CARE DELIVERY/MODELS OF CARE

1621 Poster Session 1622 Poster Session

Racial differences in serious immune-related adverse events among cancer Survival disparities between patients with breast cancer with and without
patients receiving immune checkpoint inhibitors. First Author: Cho Han Chiang, HIV at Jaramogi Oginga Odinga Teaching and Referral Hospital (JOOTRH).
Department of Medicine, Mount Auburn Hospital, Harvard Medical School, Cambridge, First Author: Harriet Fridah Adhiambo, Kenya Medical Research Institute, Nairobi, Kenya
MA Background: Breast cancer and HIV/AIDS pose significant public health challenges.
Background: Immune checkpoint inhibitors (ICIs) are associated with an increased risk of adverse Women living with HIV face higher mortality rates when diagnosed with breast cancer
events (irAEs). We aimed to evaluate disparities in serious irAEs among patients from different than HIV-negative women. Although advancements in treatment have improved survival
racial backgrounds. Methods: We performed a propensity score-matched study using the TriNetX outcomes, limited evidence exists on the impact of HIV on breast cancer outcomes in low-
Analytics Network database, which includes de-identified data from over 140 healthcare insti- resource settings. This study examines survival disparities between breast cancer pa-
tutions. We included adult cancer patients treated with ICIs and grouped patients into White, Black,
tients with and without HIV. Methods: We conducted a retrospective cohort study of
Asian, and Hispanic cohorts. The outcomes were incident composite irAEs, which included
pneumonitis, colitis, thyroiditis, hypophysitis, adrenal insufficiency, Stevens-Johnson syndrome
breast cancer patients diagnosed at JOOTRH between January 2013 and September
(SJS) / toxic epidermal necrolysis (TEN), and hepatitis within 12 months of ICI. Patients were 2024. Data from paper-based records included demographics, clinical data and outcomes
matched using variables: age, sex, ICI type, cancer type, metastatic disease, and underlying (survival status). Survival, defined as time from diagnosis to death or last follow-up,
comorbidities. Results: We identified 72,501 cancer patients who received ICIs, including 56,937 accounted for transfer out, death, being alive, or lost to follow-up. Variables with .20%
White, 7,027 Black, 5,623 Asian, and 2,914 Hispanic patients. Cohorts were adequately balanced missingness were excluded. Survival disparities by HIV status were estimated using
across covariates after matching. White and Hispanic patients showed similar risks of irAEs, both Kaplan-Meier, with mortality relationships analyzed via Cox Proportional Hazards Model.
having approximately 30% higher risk of developing serious irAEs compared with Black and Asian Results: Out of 494 breast cancer patients, 101(20%) were HIV+, 219 (44%) had unknown
patients. Compared with Black patients, White and Hispanic patients had higher risks of colitis and HIV status, and 174(36%) were HIV-. At diagnosis, HIV+ patients were younger (median:
adrenal insufficiency. Compared with Asian patients, White and Hispanic patients had higher risks 48, [IQR 40–56]) compared to HIV- patients (median: 51, [IQR 40–64], p=0.030) and had a
of pneumonitis, colitis, and thyroiditis. Conclusions: White and Hispanic patients have the highest lower median BMI (23.2 vs. 25.4, p=0.008). HIV positive patients had a longer median time
risks of developing serious irAEs. Further research is needed to explore the underlying causes and
to treatment initiation (56 days, IQR 25–127) compared to HIV-negative patients (44 days,
develop targeted interventions to mitigate these disparities. Research Sponsor: None.
IQR 20–94), although the difference was not statistically significant (p=0.4). In this
Hazard ratio for the effects of race on irAEs. cohort, 12% (60) of patients had died, with a higher mortality rate among HIV+ patients
White vs. White vs. White vs. Black vs. Hispanic vs. Hispanic vs. (17%, 17 out of 101) compared to HIV- patients (14%, 24 out of 174), while loss to follow-
Black Asian Hispanic Asian Black Asian
up was substantial in both groups (43% HIV+ vs. 37% HIV-, p,0.001). The crude 5-year
Outcomes
Composite irAE
n=7,207 each n=5,562 each
1.28 1.30
n=3,314 each n=3,839 each n=2,680 each
1.03 0.95 1.30
n=2,224 each
1.39
survival probability was 14% lower in HIV+ patients (59%, [95% CI 38 – 91]) than HIV-
(1.16-1.41) (1.16-1.45) (0.91-1.18) (0.82-1.10) (1.11-1.52) (1.16-1.67) patients (73%, [61–87]). Survival, adjusted for age, smoking, employment, and cancer
Pneumonitis 1.55 1.93 1.01 0.68 2.86 4.35 stage, did not vary significantly by HIV status (HR for HIV+ vs. HIV-: 1.13, 95% CI:
(0.77-3.12) (0.93-4.01) (0.42-2.44) (0.24-1.91) (0.78-11.1) (0.96-20.0)
Colitis 1.44 1.72 1.04 1.12 1.22 1.69 0.51–2.50, p = 0.8). However, survival was significantly lower among patients with health
(1.27-1.63) (1.47-2.00) (0.88-1.23) (0.91-1.34) (0.99-1.52) (1.32-2.17) insurance (HR = 0.35, 95% CI: 0.12–0.97, p = 0.044) and those with . primary/elementary
Thyroiditis 1.01 1.61 1.09 1.31 1.41 1.85 school education (HR: 0.13, 95% CI: 0.02, 0.88, p = 0.037) compared to those with primary
(0.81-1.25) (1.19-2.17) (0.81-1.47) (0.91-1.89) (0.98-2.04) (1.19-2.86)
Hypophysitis 0.34 0.50 0.68 2.00 0.65 1.47 education only. Patients who did not receive treatment had a significantly higher mortality
(0.07-1.67) (0.05-5.56) (0.11-4.04) (0.37-10.9) (0.11-3.85) (0.24-9.09) risk (HR = 3.52, 95% CI: 1.54–8.04, p = 0.003. Conclusions: In this study, breast cancer
Adrenal 1.70 1.18 1.10 0.64 2.50 0.89
insufficiency (1.13-2.55) (0.79-1.75) (0.69-1.77) (0.36-1.11) (1.20-5.26) (0.48-1.67) patients living with HIV had poorer crude 5-year survival probabilities compared to their
SJS/TEN 0.61 0.05 0.76 0.51 1.43 0.33 HIV-negative counterparts, although the adjusted survival did not differ significantly by
(0.15-2.54) (0.01-0.40) (0.17-3.40) (0.15-1.68) (0.24-8.33) (0.07-1.61) HIV status. Factors associated with significantly lower survival included lack of treat-
Hepatitis 1.11 0.74 0.88 0.68 1.41 0.97
(0.88-1.40) (0.59-0.93) (0.67-1.16) (0.50-0.91) (0.99-2.00) (0.69-1.37) ment, lower education levels, and absence of treatment. These findings underscore the
need for targeted interventions to improve breast cancer outcomes among HIV-positive
patients, particularly in low-resource settings. Research Sponsor: None.

1623 Poster Session 1624 Poster Session

Association between Geriatric 8 frailty, guideline treatment, treatment ad- Prognostic awareness in older adults with metastatic cancer: Secondary
herence, and overall survival in older patients with cancer (PROGNOSIS-G8). analysis of a randomized controlled trial. First Author: Cristiane Decat Bergerot,
First Author: Helena Møgelbjerg Ditzel, Department of Oncology, Odense University Oncoclı́nicas&Co, Sao Paulo, Brazil
Hospital, Odense, Denmark Background: Prognostic awareness plays a key role in patient outcomes, particularly among
Background: Frailty is frequent among older adults with cancer and may affect oncologic older adults with metastatic cancer. This secondary analysis of a randomized controlled trial
treatment tolerance. Frailty screening, with tools such as the Geriatric 8 (G8), is recom- (RCT) evaluated the effect of a Geriatric Assessment-guided Intervention (GAIN-S) on patient
mended to help guide clinical decision-making. While the G8 has been strongly associated responses to prognosis items over time between two arms. Methods: Eligible participants,
with survival, its relationship with treatment adherence remains less clear. This study aimed aged 65+, diagnosed with metastatic solid cancers, and undergoing treatment across multiple
to evaluate the association between G8-identified frailty and treatment outcomes in a large Brazilian states, were randomized 1:1 into two arms. The GAIN-S included a geriatric as-
cohort of older adults with diverse cancer types. Methods: This single-center prospective sessment (GA), which devised tailored treatment based on identified impairments. Patients in
cohort included adults, age $70 years, with solid cancers who underwent G8 screening at the usual care (UC) arm received standard care. Both arms completed the Illness and
their initial oncology consultation. Treatment-related outcomes included one-year overall Prognostic Awareness Impact Questionnaire (PAIS) at baseline (T1) and 12 weeks (T2),
survival, first-line oncologic treatment adherence within 9 months, and whether patients assessing Emotional (10 items; range 0-30) and Adaptive (12 items; range 0-36) domains; each
were offered guideline treatment. Guideline treatment was defined as regimens consistent rated on a 4-point Likert scale The change in PAIS (T2 - T1) was calculated for each participant
with recommendations from national guidelines for first-line oncologic treatment, allowing in both the GAIN-S and UC arms, and then the mean changes in PAIS between the two arms
add-on protocol treatment, while less-than-guideline treatment referred to regimens not were compared via independent t-test. Results: Eighty-six patients were approached; 80
among first choices, often deemed inferior. Adherence to the doctor-patient selected provided consent (93% enrollment rate). At 12 weeks, the analytic sample included 77 patients.
treatment plan was defined as the absence of discontinuations, dose reductions after Demographic characteristics were well-balanced between arms, with a mean age of 74.5 years
(SD=6.1), primarily female (55.8%), self-identified as White (71.4%), and 50.4% had at least a
treatment initiation, or un-administered treatments (i.e., excluding delays). Data on de-
college education. The most common cancer types were genitourinary (29.9%), breast (24.7%),
mographics, comorbidity, cancer diagnosis, treatment, and survival were extracted from
and gastrointestinal (22.1%). At T1, no significant differences were noted in PAIS between
medical records. Associations between G8 frailty (#14/17 points) and outcomes were
arms. However, there was significant improvement in PAIS Emotional (UC: mean change=-0.26,
analyzed using multivariate logistic regression and Cox proportional hazards regression,
SD=1.6 vs GAIN-S: mean change=0.87, SD=1.4; P=0.002) and Adaptive (UC: mean change=-
adjusting (adj.) for confounders. Results: Among the 1,398 patients screened, 65% were
0.07, SD=0.6 vs GAIN-S: mean change=0.74, SD=1.7; P=0.008) domains between in the GAIN-S
frail. Frailty doubled the risk of death at one year (adj. HR 2.0, 95% CI 1.7-2.4, p , 0.001). Frail arm (Table). Conclusions: This secondary analysis highlights the impact of GA-guided care on
patients who adhered to less-than-guideline treatment had a 69% lower mortality risk improving prognostic awareness in older adults with metastatic cancer. GAIN-S resulted in
compared to frail patients unable to adhere to guideline treatment (adj. HR 0.31, 95% CI 0.21- significant improvements in Emotional and Adaptive domains compared to UC. Tailored in-
0.47, p , 0.001). Non-frail patients were more likely to adhere to treatment (adj. OR 2.38, terventions addressing the specific needs of older adults with metastatic cancer may enhance
95% CI 1.49-3.81, p , 0.001) and were more often offered guideline treatment (adj. OR 1.98, understanding of prognosis and improve adaptive responses. Studies are needed to determine
95% CI 1.28-3.06, p = 0.002) compared to frail patients. Lastly, when receiving guideline whether these differences translate into meaningful improvements in outcomes. Research
treatment, non-frail patients had significantly better adherence than frail patients (adj. OR Sponsor: None.
3.08, 95% CI 1.72-5.52, p , 0.001). Conclusions: G8 frailty screening effectively identifies
Impact of GAIN-S intervention on prognostic awareness.
older adults at a higher risk of treatment non-adherence and mortality, facilitating tailored
treatment approaches. Our findings suggest that frail patients may benefit from initial less- Change over time
intensive treatments with potential escalation to improve adherence and survival. Imple- Arm 1 (T2-T1) Arm 2 (T2-T1) Difference
menting G8 screening in routine practice addresses the unique challenges associated with Mean (SD) Mean (SD) Arm 2 - Arm 1 (SE) p-value
frailty, ensuring more effective, equitable care for at-risk older adults. Research Sponsor: The Emotional Domain -0.263 (1.62) 0.872 (1.44) 1.135 (0.35) 0.002
Danish Cancer Society; Odense University Hospital; University of Southern Denmark; Agnes 10 items (range: 0-30)
and Poul Friis Fond; Dagmar Marshalls Fond; Academy of Geriatric Cancer Research Adaptive Domain -0.079 (0.67) 0.744 (1.73) 0.823 (0.30) 0.008
12 items (range: 0-36)
(AgeCare).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 103s

1625 Poster Session 1626 Poster Session

National cancer system characteristics and global pan-cancer outcomes. Validating Navya Earthshot: An AI-enabled point-of-care solution for
First Author: Edward Christopher Dee, Memorial Sloan Kettering Cancer Center, New guideline-adherent treatment planning in a decentralized cancer care
York, NY model. First Author: Umesh Mahantshetty, Tata Memorial Hospital, Homi Bhabha
Background: Approximately 29.9 million cancer cases and 15.3 million deaths are antic- National Institute, Homi Bhabha Cancer Hospital & Research Centre, Visakhapatnam,
ipated by 2040 globally. Health systems must invest in cancer system strengthening. A India
greater understanding of health system factors that can be leveraged to improve cancer Background: Adherence to guidelines increases overall survival, globally. In resource-
control may guide health system planning. Therefore, we conducted a pan-cancer ecological constrained settings, ~ 30% of patients receive undertreatment or overtreatment. Despite
study making use of most recent available national health system metrics for cancer significant investment in decentralized cancer care—with tertiary hub centers providing
outcomes and health system metrics, spanning the breadth of global income levels across support to non-specialized spoke centers—shortage of oncologists creates a knowledge
185 countries. Methods: Estimates of age-standardized mortality-to-incidence ratios were gap, which may be partially addressed by clinically validated AI solutions. Navya is a
derived from GLOBOCAN 2022 for patients with cancer of all ages. Health spending (% of clinically validated AI solution for cancer patients in use since 2014 which matches
gross domestic product [GDP]), physicians/1000population, nurses and midwives/ patient-specific data to evidence and generates treatment recommendations vetted via
1000population, surgical workforce/1000population, GDP per capita, Universal Health asynchronous expert review. Navya Earthshot is a new, provider facing solution for point
Coverage Service Coverage Index (UHC index), availability of pathology services, human of care cancer treatment planning for non-specialized providers, and is built as an AI
development index, gender inequality index, radiotherapy centers/1000population, and out- driven search interface on Navya’s validated domain model supporting subspecialized
of-pocket expenditure as percentage of current health expenditure were collected. The expert opinions in oncology. Methods: This multicenter, prospective validation took
association between MIR and each metric was evaluated using univariable linear regres-
place at 25 hospitals across India participating in a decentralized cancer care model. All
sions. Metrics with P , 0.0045 (Bonferroni corrected) were included in multivariable models.
patients with breast, oral and lung cancerbetween January and June 2024 with all
Variation inflation factor allowed exclusion of variables with significant multicollinearity. R2
decisions (curative and palliative; local and systemic therapies) were included. Navya
defined goodness of fit. Results: On univariable analysis, all metrics were significantly
associated with MIR of cancer (P , 0.001 for all). After including metrics significant on
Earthshot matched input patient data available in the patient medical record with National
univariable analysis and correcting for multicollinearity, the final multivariable model had R2 Cancer Grid (NCG) guidelines, and output evidence based treatment plans at the point of
of 0.8729. Therefore, the following variables were associated with lower (improved) MIR for care. The output was shared at each center, and concordance was scored by the tumor
cancer: 1) nurses/midwives per 1000 population (b = —0.0049, P , 0.057), 2) UHC index (b = board/treating oncologists, as well as by a group of domain experts experienced in
—0.0042, P , 0.001), 3) radiotherapy centers per 1000 population (b = —11.21, P = 0.072), analyzing NCG guidelines. Results: Navya Earthshot processed 1787 decisions in a
and 4) GDP per capita (b = —1.7x10-6, P , 0.001). On analysis stratified by sex, the following decentralized cancer care system, pertaining to 40% (725) breast, 20% (351) oral, 40%
were associated with improved MIR for all cancers among females: 1) UHC index (b = (711) lung cancer diagnoses respectively. Patients were well represented with respect to
—0.0042, P , 0.001), 2) GDP per capita (b = —9.9x10-7, P = 0.02), and 3) gender inequality age ( , 45 years (23%) and . 45years (77%), and early stage (24%); advanced stage
index (b = 0.13, P = 0.084) (R2 0.8699). The following were associated with improved MIR for (58%) and incomplete diagnostic workup (18%)). Of the 1787 decisions, Navya Earthshot
all cancers among males: 1) nurses/midwives per 1000 population (b = —0.0053, P = 0.066), output referred 27% (478) to hub center tumor boards due to presence of uncommon
2) UHC index (b = —0.0042, P , 0.001), 3) radiotherapy centers per 1000 population (b = histologies or scenarios not covered by the NCG guidelines (3rd line therapy etc.). In the
—12.37, P = 0.076), 4) GDP per capita (b = —2.31x10-6, P , 0.001) (R2 0.8485). remainder 73% (1309) decisions, Navya Earthshot output diagnostic or treatment plans.
Conclusions: This comprehensive pan-cancer analysis of health system metrics suggests Of these, 85% (1114/1309) were scored concordant with NCG guidelines, and adopted by
progress towards UHC, strengthening the nursing/midwifery workforce, facilitating access to the local treating oncologist. The remaining 15% (195) decisions were referred to a hub
services such as radiotherapy, and mitigating gender inequality are key priorities in cancer center for treatment planning. Conclusions: Navya Earthshot can improve capacity of
control. These generalizable findings may guide efforts to strengthen cancer systems oncologists in resource-constrained settings and enhance adherence to guideline-driven
throughout the world. Research Sponsor: National Cancer Institute/U.S. National Institutes care in a decentralized cancer care model. In a majority of cases, this point-of-care
of Health; P30 CA008748; Prostate Cancer Foundation; National, Heart, Lung, and Blood solution can improve access to care locally, reduce reliance on tertiary hub centers, and
Institute; 1R38HL167238-01 grant. improve patient outcomes, globally. Research Sponsor: None.

1627 Poster Session 1628 Poster Session

Advancing global equity in cancer care: Comparative environmental impacts Achieving global breast cancer initiative key performance indices for breast
of radiotherapy in Brazil and the U.S. First Author: Katie Lichter, Department of cancer patients in Botswana by HIV status. First Author: Tara Friebel-Klingner,
Radiation Oncology, University of California, San Francisco, San Francisco, CA Johns Hopkins University, Baltimore, MD
Background: Climate change poses a significant threat to global health, necessitating Background: Low- to middle-income countries have disproportionately higher breast
efforts to address the environmental impacts of oncology care, particularly in low- and cancer (BC) mortality rates, partly due to late-stage diagnosis. In people with HIV (PWH), BC
middle-income countries (LMICs). Radiation therapy is a cornerstone of cancer treatment, mortality is worse compared to those without HIV. The WHO’s Global Breast Cancer Initiative
yet its delivery often involves high energy consumption and resource use, contributing to (GBCI) proposes to reduce mortality through 3 identified pillars: (1) health promotion for
environmental degradation and inequities in health outcomes. Despite these challenges, early detection with at least 60% of invasive BC diagnosed at stage 1 or 2; (2) timely
the environmental footprint of radiotherapy in the Global South remains largely unex- diagnosis, where evaluation, imaging, and pathology are completed within 60 days from first
plored. This study quantifies the environmental impacts of radiotherapy in Brazil, doctor’s appointment; and (3) $80% of patients completing treatment. Few real-world data
compares these findings to external beam radiation therapy (EBRT) delivery in the U.S., exist on these key performance indices in PWH. Understanding the influence of clinical and
and explores how sustainable practices may promote equity by reducing operational demographic factors associated with achieving the GBCI pillars can inform tailored in-
costs and expanding access to radiation therapy in underserved regions. Methods: A life terventions to increase BC survival. We aimed to assess pillars 1 and 2 in BC patients
cycle assessment (LCA) of EBRT for ten cancer disease sites was conducted at a radiation presenting to a referral hospital in Botswana, by HIV status. Methods: This prospective BC
oncology clinic (Vitta) in Brası́lia, Brazil, following ISO 14040 and 14044 standardized cohort included patients .18 years, presenting for BC care at Princess Marina Hospital
methodology. Data on medical supplies, equipment usage, building energy consumption, between 2015 and 2023. Patients with unknown HIV status and/or unknown stage were
and staff and patient travel from 2018-2023 was analyzed to assess environmental excluded. Pillar 1 was assessed using the cancer stage documented in the medical chart.
impacts across nine categories, including greenhouse gas emissions, air pollution, and Pillar 2 was assessed using patient recall of first-contact with health facility and pathology
carcinogenic potential. These results were compared with a previously published LCA of report date. We characterized socioeconomically disadvantaged districts as those with
EBRT across four U.S. healthcare centers. Results: Radiotherapy at Vitta had a lower poverty rate $20% and , 100% of the population living within 5km from a health facility.
Descriptive statistics and logistic regression were used. Results were stratified by HIV
environmental impact across all categories compared to U.S. centers. Transit-related
status. All p-values were two-sided. Data was analyzed using STATA 18.5. Results: 655
emissions were the largest contributor at Vitta, though they remained lower than those in
patients (median age: 51.2, IQR 42.4, 63.4) met eligibility criteria. 212 (31.8%) were PWH and
the U.S. due to shorter travel distances (median 15 miles/week by public transit at Vitta
11 (1.7%) men. 180 (27.1%) attained pillar 1, and 59 (9.3%) attained pillar 2. PWH were
vs. 48–90 miles/week by car in the U.S.). Vitta’s reliance on hydroelectric energy
younger (48.9 vs. 58.6; p , 0.001) and more likely to be single (70.3% vs. 47.3%; , 0.001).
eliminated emissions from building heating and reduced cooling-related emissions to
There was no significant difference between rates of PWH and those without HIV achieving
9.3% of the clinic’s total footprint, compared to 74.0% at U.S. sites using mixed-grid both pillar 1 (23.1% vs 28.9%, p = 0.12) and pillar 2 (11.4% vs 8.3%, p = 0.2). However, PWH
electricity and natural gas. However, impacts from medical supplies at Vitta were higher had a significantly shorter interval from first contact with the health facility to completed
across all categories, reflecting opportunities for resource optimization. pathology (11.9 months vs. 20.1 months, p = 0.013). There was no significant difference
Conclusions: This study provides novel insights into the environmental impact of ra- between residents of socioeconomically disadvantaged districts versus non-residents in
diotherapy in an LMIC context, underscoring the importance of regional differences in achieving pillar 1 (29.7% vs. 27.0%, p = 0.72) or pillar 2 (8.3% vs. 9.3%, p = 0.84).
care delivery. The reduced environmental footprint at Vitta highlights the value of Conclusions: Majority of BC patients presenting to the referral hospital in Botswana did not
sustainable practices, such as renewable energy and public transit, in mitigating the achieve the GBCI pillars 1 and 2. However, PWH had significantly less diagnostic interval,
health sector’s climate impact while reducing operational costs. These findings support which may be reflective of frequent contact for PWH in established care and may present an
the development of adaptable, scalable models for global radiation oncology, particularly opportunity for care integration for PWH. Early detection and patient navigation inter-
in publicly funded and rural clinics, to expand access and promote equity in cancer care. ventions may potentially help Botswana achieve the WHO’s GBCI goals and reduce BC
Future research should prioritize environmentally sustainable strategies that align with mortality. Research Sponsor: Fogarty International Center K01TW011481 Award; Doris Duke
the unique needs of LMICs and underserved populations. Research Sponsor: None. Charitable Foundation Fund to Retain Clinical Scientists at Penn/CHOP.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
104s CARE DELIVERY/MODELS OF CARE

1629 Poster Session 1630 Poster Session

Disparities in exercise referral practices and barriers among oncologists in Investigating gender demographics and equity among hematologists. First
public and private institutions in Latin America. First Author: Paulo Gustavo Author: Florence Broussais, Institut Carnot CALYM, Lyon, France
Bergerot, Oncoclı́nicas&Co, Medica Scientia Innovation Research (MEDSIR), Sao Paulo, Background: Despite a substantial number of women in the field of hematology,
Brazil representation in leadership roles remains inconsistent with ~50% of women among
Background: Exercise offers significant benefit for patients with cancer, improving medical graduates in many countries. This survey of hematologists sought to examine
physical function, quality of life, and treatment-related outcomes. However, barriers gender demographics and explore experiences with professional development including
such as limited referral practices and insufficient knowledge hinder its integration into mentorship, involvement in and leadership of clinical trials, and opportunities for career
care. This study compared referral practices, exercise assessment, and physicians’ advancement. Methods: An international online survey of hematologists was developed
perceptions in public (PUB) and private institutions (PRI). Methods: A cross-sectional collaboratively with the Women in Lymphoma (WiL) global organization in partnership
survey of 454 physicians from 21 Latin American countries was conducted using a 25- with the HERmatology initiative (AstraZeneca). It was distributed to and by the
item questionnaire on referral practices, patients’ exercise habits, and perceived barriers membership of WiL. All participants were licensed medical practitioners, self-
and facilitators to implementing exercise programs. Descriptive statistics summarized characterized on gender, age, years of practice, country of practice, and seniority of
respondent characteristics and adherence to exercise-related practices. Chi-square role. Statistical differences between males and females were calculated by pairwise z-
tests were used to compare differences in referral practices and perceived barriers/ tests with a significance level of 90%. Results: From October to December 2024, 237
facilitators between physicians working in PUB versus PRI institutions. Results: Out of hematologists from 34 countries were surveyed: 182 female and 55 male. The proportion
454 participants, most were from PUB (52%), mainly from Mexico (17%), Brazil (12%), of females to males in hematology practices significantly favored females (53% [51, 56]
and Colombia (10%). In the PRI (48%), Brazil led with 51%, followed by Argentina (18%), vs. 47% [44, 49]). Heads of Department (60% [55, 66] vs. 40% [35, 45]) and direct
and Peru (8%). Female representation was higher in PUB compared to PRI (57% vs. 43%, supervisors (59% [54, 64] vs. 41% [36, 46]) were significantly more likely to be male.
P = 0.01). Physicians in PUB were less likely than those in PRI to assess exercise habits Females reported serving as Principal Investigator significantly less frequently for
(53% vs. 18%, P = 0.001), refer patients (72% vs. 36%, P = 0.001), or provide guidance industry-sponsored clinical trials over the past 12 months than males (2.1 vs 0.9, p ,
(56% vs. 12%, P = 0.001). Resource limitations were more common in PUB (e.g., no 0.1), but no significant differences as a Site Investigator (3.4 vs 2.4, p = 0.2). No
referral location: 86% vs. 70%, P = 0.04). Barriers included treatment side effects (PUB: differences were observed in female vs. male participation or leadership in cooperative
66% vs. PRI: 40%, P = 0.001) and lack of knowledge on prescribing exercise (PUB: 63% group / academic clinical trials. There was a trend for fewer female physicians invited to
vs. PRI: 27%, P = 0.001). Physicians in PUB emphasized facilitators like access to write an article or publication in the past 12 months compared to males (2.7 vs. 4.2, p =
qualified professionals (90% vs. 66%, P = 0.001) and personal experience (90% vs. 80%, P 0.15), as well as fewer opportunities to receive public recognition from a senior col-
= 0.01). Conclusions: This study reveals significant disparities in cancer exercise league (1.1 vs. 1.9, p = 0.15). Females overall were significantly less likely to have clinical
practices between oncologists in public and private institutions across Latin America. (p , 0.1) or career mentors (p , 0.1), but earlier-career females were more likely than
Oncologists in public institutions were less likely to assess exercise, refer patients to later-career females to have both (p , 0.1). ‘Family or caregiving responsibilities’ was
exercise programs, and provide guidance. They also reported greater barriers, such as the top factor discouraging women from pursuing careers in hematology (86% agreed),
treatment side effects and lack of knowledge on exercise prescription and resources. followed by a ‘Lack of representation of women in leadership positions’ (59% agreed).
Facilitators such as access to qualified professionals were less prominent in public Conclusions: Differences were noted between males and females across several career
institutions. These findings highlight the need for targeted interventions to improve dimensions, including representation in leadership roles, involvement in industry-
exercise integration in cancer care, particularly in resource-limited settings. Research sponsored clinical trials, and access to mentorship. These results underscore the
Sponsor: None. continued need for systematic efforts to reduce unconscious bias and promote greater
diversity within hematology. Providing additional structured support, particularly to
boost female leadership in industry-sponsored trials, will help advance gender equity,
ultimately benefiting both medical progress and patient outcomes. Research Sponsor:
None.

1631 Poster Session 1632 Poster Session

Clinical profiles, patient expectations and outcomes from an integrative How do health concerns present in cancer survivors who require inter-
oncology clinic in India: A novel integrated model of care in oncology. First preters? Database analysis of a structured survivorship clinic for early-
Author: Kanakavalli K. Kundury, JSS Academy of Higher Education & Research, Mysuru, stage cancer. First Author: Lawrence Kasherman, Concord Clinical School, Faculty of
India Medicine and Health, University of Sydney, Concord, NSW, Australia
Background: Integrative approaches are used in Oncology care, often as an auxiliary Background: Cancer survivors (CS) of Culturally and Linguistically Diverse (CALD) back-
measure to the Standard of Care (SoC). There is paucity of data regarding Integrated grounds face disparities in care. We aimed to compare demographics and health concerns of
oncology model approach combining alternate systems of medicines like Ayurveda & CALD CS with non-CALD CS following completion of primary treatment. Methods: The Sydney
Siddha, Yoga and Dietary modification recommendations along with SoC . In this study, we Cancer Survivorship Centre (SCSC) database was analyzed to compare baseline differences in
present an audit of this novel Integrated oncology model of care provided by online and in- demographics and health concerns between early-stage CS of solid tumor or hematologic
patient consultations at Isha Integrative Oncology Clinic (IIOC), Isha Health Solutions (IH), cancers requiring interpreters (CALD CS) during initial consultations to those who did not (non-
Coimbatore, India. Methods: A clinical audit was conducted for 514 patients who have CALD CS). Descriptive statistics were used to illustrate distribution by age, gender and tumour
received care through Integrated oncology consultations ( in-person or online) at IIOC from type. Survivors completed questionnaires on symptoms, quality of life (QoL), distress, exercise
January 2016 to July 2024. This abstract focuses on the statistical analysis of data of the time and were assessed by a psychologist for fear of cancer recurrence (FCR). Univariate
initial 196 consecutive patients based on descriptive data from clinical proformas and analyses were used to determine differences at presentation to the survivorship clinic between
follow-up visits for symptomatic response outcomes. Updated data will be presented at the CALD and non-CALD groups. Results: From September 2013 to April 2024, 939 initial con-
sultations (median 10.9 months from diagnosis) with consenting CS were conducted at SCSC.
conference. Results: Among 196 patients analysed, 99% of patients had online con-
15% (n = 137) required interpreters in 21 different languages (Mandarin (n = 48, 35%), Korean (n
sultations. The median age was 52yrs (7-81yrs) and Male: Female ratio was 1: 1.3. 50%
= 26, 19%) and Cantonese (n = 19, 14%)): 83 (61%) used professional interpreters, 25 (18%)
belonged to age group 40-60 yrs whereas 29% were between 60-85 yrs. Most common
family/friends. CALD CS were more likely to be female (58%), aged 40-64 (51%) and have
cancers in males were hematological cancers (20%), prostate cancers (11%), GI cancers colorectal cancer (55%). Common symptoms in CALD CS of at least moderate severity included
(11%) and in females were breast (26%), ovary (10%) and colon (5%). The most common fatigue (39%), numbness (32%), and pain (31%), not significantly different to non-CALD CS.
stage was Stage 4 (58%) followed by Stage 3 (40%). Most common symptoms were pain Significantly lower proportions of CALD CS reported trouble concentrating (19 vs 29%, p =
(49%), anorexia (43%), fatigue (42%), lack of sleep and anxiety (39%). Most common side 0.023), hot flashes (14 vs 23%, p = 0.017) and problems with sexual function (12 vs 20%, p =
effects of chemotherapy were fatigue (32%), constipation (25%), anorexia/weight loss 0.032). CALD CS were significantly more likely than non-CALD CS to report less minutes/week
(25%), pain (16%) etc. Most common expectations were cure/avoid relapse (32%), doing vigorous (13.2 vs 32.8; p , 0.001), moderate (32.3 vs 73.0; p , 0.001) and resistance
symptomatic relief (15%), reduction from chemotherapy side effects (11%), integration of exercise (4.6 vs 18.7; p , 0.001), but more light exercise (219.6 vs 115.5; p = 0.041). Mean
Yoga (10%) etc. Only 6% of patients wished to avoid chemotherapy. Integrated oncology global FACT-G QoL score for CALD CS was 77.4 (SD 20), with physical (mean 21.4, SD 6) and
model based on complementary systems of medicine (Ayurveda & Siddha), dietary emotional (mean 17.6, SD 5) domains most impacted; these were not significantly different
changes and yoga was provided to all patients (100%). After using this model of care, from non-CALD CS scores. There were no significant differences between groups in mean
improvement in cancer-related symptoms was reported by 90% of patients and compliance distress thermometer scores (3.09 vs 3.39/10; p = 0.288), rates of moderate-to-severe distress
seen in 73% of patients. Conclusions: Our study is one of the largest clinical audits in (36 vs 41%; p = 0.952) or rates of psychologist-assessed moderate-to-severe FCR (19 vs 28%; p
Integrative oncology in published literature. Younger patients and advanced cancer pa- = 0.228). Conclusions: CALD CS experience similar physical and psychosocial health concerns
tients more often seek integrative oncology care and main expectation is to achieve better to non-CALD CS at initial survivorship clinic consultations but are less likely to report issues
cure rates and symptomatic relief. In this study, all patients were provided a novel in- with sexual function, concentration and hot flashes, and are more likely to exercise less. Future
tegrated oncology model with alternative medicines (Ayurveda & Siddha), yoga, and dietary work should focus on longitudinal effects and comparisons over time, with particular focus on
modifications in addition to their ongoing SoC resulting in good symptomatic relief and addressing cultural sensitivities and increasing exercise intensity. Research Sponsor: Aus-
high compliance rates. Integrative oncology model incorporating alternative medicine, tralian Government, National Health and Medical Research Council; 2021964; Cancer Institute
yoga, and dietary changes can be effectively offered to cancer patients alongside standard New South Wales; 2021/ CBG0002; Australian Government, National Health and Medical
treatment. Further prospective studies are warranted. Research Sponsor: None. Research Council; APP1176221.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 105s

1633 Poster Session 1634 Poster Session

Living beyond cancer: The long-term impact of breast cancer diagnosis on Blended survivorship and palliative care for patients with advanced lung
cognitive function. First Author: Cheng Peng, Brigham and Women’s Hospital, Boston, cancer receiving targeted therapy: An open pilot. First Author: Laura A. Petrillo,
MA Massachusetts General Hospital, Boston, MA
Background: While some clinical studies report greater cognitive difficulties in middle- Background: Targeted therapy improves survival and quality of life for patients with
aged women diagnosed with and treated for breast cancer over the short-term, obser- advanced non-small cell lung cancer (NSCLC) with driver alterations. However, advanced
vational studies of older persons, with longer follow-up, found that a history of cancer was NSCLC remains incurable, and the timing of progression on targeted therapy is un-
associated with lower Alzheimer’s disease risk. We estimated the relation of breast cancer predictable. Thus, many patients live with an abiding, distressing sense of uncertainty.
and treatment history with cognitive status and rate of decline among older women. We To better support patients with advanced NSCLC receiving targeted therapy, we 1)
further divided breast cancer survivors by: (1) time since cancer diagnosis (assessing developed and refined a blended early palliative care and survivorship intervention and 2)
recency), (2) age of cancer onset (reflecting body aging at diagnosis), and (3) stage conducted an open pilot to further refine the intervention. Methods: We conducted a
(capturing disease aggressiveness). To evaluate any overlap in shared or opposing genetic review of evidence and prior palliative care and behavioral health interventions to inform
risk, we estimated cognitive status by breast cancer polygenic risk score (PRS). the development of POISE, a structured, supportive care intervention for patients with
Methods: A cognitive sub-study was initiated in 1995-2001 in the Nurses’ Health Study, advanced NSCLC receiving targeted therapy, in which trained palliative care clinicians
including 1,378 breast cancer survivors and 14,196 cancer-free women. Breast cancer aim to enhance patient coping with uncertainty, setting lifestyle goals, and prognostic
diagnoses were self-reported and confirmed by medical records; treatment was self- awareness. To refine POISE, we first conducted qualitative interviews with 20 com-
reported. Cognitive function was assessed up to 4 times (over a mean of 6.6 years) and munity partners, including patient/caregiver advocates, palliative care clinicians, psy-
combined into 4 outcomes: global composite score, Telephone Interview for Cognitive chologists, and oncologists. We elicited feedback on POISE intervention content and
Status (TICS), verbal memory, and working memory. We used linear models to assess delivery. We used rapid analysis to analyze interview transcripts and identified themes,
breast cancer history and treatment with cognitive status (averaged across follow-ups). with which we finalized POISE (4 monthly one-hour palliative care visits). We then
We used mixed-effects models to assess breast cancer history and treatment with rate of conducted an open pilot study of POISE among 10 patients diagnosed with advanced
cognitive decline. We computed a breast cancer PRS and evaluated cognitive function NSCLC with a targetable mutation (i.e., EGFR, ALK, ROS1 or RET) in the past 6 months
across quartiles of PRS. Results: The mean age of breast cancer diagnosis was 65.4 years,
who were receiving care at a single academic cancer center. All participants received
and the mean time between cancer diagnosis and baseline cognitive assessment was 8.6
POISE and completed self-report surveys at baseline and at 12 and 20 weeks. Par-
years. We observed similar distributions of key risk factors for AD between women with
ticipants rated their satisfaction with POISE and completed exit interviews at 20 weeks.
and without history of breast cancer, including age, education, depression, and physical
We used a framework approach to analyze exit interviews and identify themes for further
activity. No significant differences in global cognitive status were noted comparing women
with a history of breast cancer with those who were cancer-free. Associations did not differ
intervention refinement. Results: Qualitative interviews highlighted a need to
by age of cancer onset ( , 65 vs. $65 years), time since diagnosis ( , 5 vs. $5 years), or strengthen the POISE clinician training and supervision plan to increase palliative care
stage. Genetically predicted breast cancer risk was not associated with the global clinicians’ behavior modification therapy skills. We revised the POISE manual to be more
cognitive status. Women with breast cancer and treated with hormone therapy, che- flexible to accommodate patient choice in session focus and added a list of community
motherapy, and/or radiation therapy had similar global cognitive status compared to organizations to support ongoing behavior change. With the revised intervention, we
cancer-free women. No significant differences by breast cancer history were observed for then initiated the POISE open pilot. We approached 13 eligible patients, of whom 10
TICS, verbal memory, or working memory. Over the modest follow-up time, we observed no (mean age = 67 years, 5 female, 5 male) consented to participate. Most (8/10) patients
significant differences in cognitive decline between women with a history of breast cancer completed all 4 sessions and all surveys. All (8/8) patients reported that POISE was
and cancer-free women. Conclusions: We observed no association of history of breast helpful and would recommend it to others. In exit interviews, patients suggested in-
cancer or breast cancer treatment with cognitive function status or rate of decline, corporating caregivers in sessions and adding a patient-facing workbook to assist with
suggesting there is neither harm nor benefit of breast cancer diagnosis on long-term skill acquisition. Conclusions: POISE warrants further study in a feasibility pilot ran-
cognition. Research Sponsor: National Institute on Aging. domized controlled trial. Clinical trial information: NCT04900935. Research Sponsor:
National Cancer Institute.

1635 Poster Session 1636 Poster Session

Experiences and preferences of cancer survivors across the immunotherapy Health related social needs in an urban academic breast cancer survivorship
journey. First Author: Shelley Fuld Nasso, National Coalition for Cancer Survivorship, program. First Author: Mumtu Lalla, Montefiore Einstein Comprehensive Cancer Center,
Silver Spring, MD Bronx, NY
Background: Immuno-oncology (IO) drugs are recommended by guidelines for several Background: Health-related social needs (HRSN) impact cancer care and are associated
tumor types and can significantly improve survival in patients with cancer. As patients with late stage at diagnosis (dx), prolonged time to therapy initiation, and poor outcomes.
experience long-term care due to extended survival, it is important to understand the Little data exists regarding long-term impact of cancer dx and therapy on HRSN in late
challenges of survivorship and experiences receiving IO. Here, we present IO treatment survivorship. We evaluated associations of cancer therapies with HRSN in a diverse cohort
experiences and preferences for IO based on modes of administration among cancer of breast cancer (BC) survivors seen in an urban academic medical center.
survivors. Methods: A web-based survey was administered to US cancer survivors in the Methods: HRSN screening was completed by BC survivors seen in the cancer center
fall of 2024. Patients were recruited via physician referral and were eligible to participate if survivorship program in 2023. Survivors were 4 or more years (yrs) from dx of Stage 0 to III
they were $ 18 years of age, diagnosed $ 1 year prior with any solid tumor cancer (any BC, with all therapy complete (including endocrine therapy) and no evidence of metastasis.
stage), and received IO within the past 5 years. The survey included the Quality of Life- HRSN screens were compared with those completed by newly diagnosed (dxed) BC
Cancer Survivor (QoL-CS) scale (range 0–10; higher scores indicate better QoL) and a patients (pts) and adult general medical pts. Charts of survivors were reviewed for details
direct preference exercise. The preference exercise asked patients to indicate whether they of BC dx and treatment [stage, hormone receptor (HR) and HER2 status, type of surgery,
would prefer IO be administered via subcutaneous (SC) injection or intravenous (IV) in- and use of chemotherapy, radiation, and endocrine therapy]; as well as demographic data
fusion and asked patients to rank the influence of common characteristics of each ad- (age, race, ethnicity, insurance status, ZIP code). Neighborhood distress was categorized
ministration on preferences. Study variables were analyzed descriptively. Results: The using Distressed Communities Index (DCI). Associations of clinical and demographic
mean (standard deviation [SD]) age of patients (N = 100) was 57.8 (7.2) years; 46% were factors with HRSN in BC survivors were evaluated by Chi-Square and ANOVA tests.
White and 51% were male. The most common tumor types reported were lung (26%), Results: 465 BC survivors [177 (38%) Black, 228 (49%) Hispanic] completed HRSN
melanoma (14%), kidney (12%), and colon (10%). Mean (SD) time since cancer diagnosis screening. Median age at assessment was 68 (range 43-98) and at dx was 57 (range 27-89);
was 4.1 (2.7) years and mean time since starting IO was 2.4 (1.5) years. Most (72%) median time from dx was 11.25 yrs (range 4-35). 63.5% resided in distressed ZIP codes;
patients reported that a typical visit for receiving an infusion lasted between 1–2 hours and 26.4% in at-risk ZIP codes. Most common HRSN in survivors completing their first HRSN
most (56%) traveled 30–60 minutes to receive their infusion. Some patients experienced screen (n = 395) were housing quality (5.3%) and food insecurity (4.8%). BC survivors were
interrupted access to their most recent IO treatment due to transportation delays (24%) less likely to endorse HRSN than newly dxed BC pts (16.7% vs 24.0%, p = 0.03), but had
and not having someone to accompany them to treatment (24%). The average QoL-CS similar HSRN as general medical pts (16.7% vs 14.6%, NS). Among survivors, younger age
score was moderate (mean [SD], 5.1 [1.3]) and 57% of patients agreed that IO improved at BC dx (p = 0.006) and at HRSN screen (p = 0.004) was associated with greater risk of
their QoL. Yet, 42% and 48% noted that daily activities and physical health, respectively, HRSN, with pts dxed prior to 57 almost twice as likely to endorse HRSN as those dxed after
were negatively affected by their most recent IO treatment. Most patients (92%) preferred a 57 (22.7% v 12.6%, p = 0.004). Younger survivors with HRSN were more likely to reside in
hypothetical SC injection over IV infusion; the top SC injection characteristics influencing distressed ZIP codes (p = 0.03). Hispanics seen in new survivorship visits were more likely
this preference were no need to access the vein with an IV catheter, the amount of time it to endorse HRSN than Hispanics seen in follow-up and than non-Hispanics seen in new or
takes to be administered (described as 5 vs 30 minutes), and the potential for more follow-up visits (p = 0.03). Race, insurance status, time from BC dx, stage, HR and HER2
flexibility in scheduling/location of receiving IO treatment. Conclusions: Our study found status, extent of surgery, and use of chemotherapy, radiation, or endocrine therapy were
that IO treatment improved the QoL of most cancer survivors. However, challenges to not significantly associated with HRSN in survivors. Conclusions: In this diverse cohort,
receiving IO, including the time required for travel and administration, were common. Most late BC survivors had similar HRSN prevalence as general medical pts. Younger age was
patients in our sample would prefer a SC administration of IO over an IV infusion, sug- associated with greater HRSN in survivors, while BC stage, receptor status, and extent of
gesting that less-invasive modes of administration requiring less time and greater flex- treatment were not. Hispanics presenting for first survivorship visit endorsed greatest
ibility may present opportunities to improve the patient treatment experience. Research HRSN. These findings have implications for interventions targeted to young survivors and
Sponsor: Bristol Myers Squibb. for culturally sensitive survivorship care. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
106s CARE DELIVERY/MODELS OF CARE

1637 Poster Session 1638 Poster Session

Feasibility trial of health coaching-based navigation after breast cancer Predicting overall survival in adults with cancer in the US using machine
treatment. First Author: Ruvarashe Rumano, The Ohio State University, Columbus, OH learning approaches integrating comprehensive social risk factors. First
Background: Following breast cancer diagnosis, 25% of survivors experience psy- Author: Yiwang Zhou, St. Jude Children’s Research Hospital, Memphis, TN
chosocial needs like depression, anxiety, and fear of recurrence. Integrative survivorship Background: Adults with cancer in the U.S. face an elevated mortality risk compared to
support services address these needs, but participation is lower among Black women, the general population, with social risk factors playing a critical role – particularly among
who report higher psychosocial distress. Patient navigation has emerged as a strategy to those with comorbidities. However, traditional mortality risk prediction models often
reduce these barriers. This study piloted a navigation-based intervention aimed at focus on treatment exposures and basic demographic factors, overlooking social risk
improving psychosocial symptom management over 6 months. Methods: A single-arm factors. We aim to develop a machine learning (ML) model that integrates compre-
feasibility trial was conducted with Black breast cancer survivors from November 2022- hensive social risk factors with traditional predictors to predict overall survival for adults
June 2024. Women aged 18+ with non-metastatic breast cancer were recruited. A with cancer in the U.S. Methods: We analyzed data from 6,181 nationally representative
trained lay navigator provided personalized support and resource facilitation to address adults diagnosed with cancer from the National Health Interview Survey (NHIS; 2013-
psychosocial and healthcare challenges. Interactions were analyzed qualitatively to 2014). A total of 74 risk factors, including basic demographics (e.g., age at the survey,
assess preliminary impact. Surveys administered at baseline and post-intervention sex, marital status, body mass index [BMI]), personal and household socioeconomic
included the Life and Longevity after Cancer, Patient Reported Outcomes Measurement status (SES; e.g., education, food insecurity), lifestyle, social support, and health status
Information System, and Breast Cancer Survivorship Experience Survey, with responses (e.g., chronic health conditions [CHCs], disability), were included in modeling. The
measured on a Likert scale. Quantitative data were analyzed descriptively. Results: Of primary endpoint was 5-year overall survival from the survey completion date, with
21 Black women who consented (mean age 64), 18 began the study while 3 did not secondary endpoints of 1- and 2-year survival. Death from any cause after the survey
proceed beyond consent. Of these, 44.4% completed all six months, while 61.1% was defined as an event, and subjects were censored 5 years post-survey. The sample
completed three months. Eight participants completed post-surveys. All participants was randomly split into 70% training and 30% testing. A random survival forest (RSF)
found the study easy to join, 5/8 found its length appropriate, and 6/8 were satisfied with model predicted survival. The time-dependent area under the receiver operating
session intervals. All would recommend the program, and 4 preferred in-person ses- characteristic (AUROC) curve and the Brier score (BS) assessed the model performance.
sions. Participants reported confidence in accessing future support programs. Key Both AUROC and BS range from 0 to 1, with higher AUROC for higher accuracy (dis-
benefits included managing stress, family stressors, and improving community support. crimination) and lower BS for better alignment between predicted and observed risk
Qualitative analysis of 108 transcripts from 17 participants identified six themes: (calibration). The Shapley additive explanations (SHAP) values were used to interpret
managing health challenges, communication, emotional well-being, resilience in daily variable importance in the established RSF model. Results: The mean age of subjects
life, program support, and future planning. Survey results showed improvements in well- during the survey was 65.6613.8 years, and 40.2% were male. For the established RSF
being and self-efficacy. LILAC scores increased from 52.7 to 63.9, PROMIS physical model, the AUROC (mean 6 standard deviation) for predicting 1-, 2-, and 5-year survival
scores rose from 12.4 to 13.3, and mental scores improved from 13.3 to 15.6. Self- was 0.795 6 0.026, 0.810 6 0.018, and 0.831 6 0.011, respectively, reflecting high and
efficacy decreased slightly in BCSES scores from 53.6 to 52.5. Conclusions: Among improved predictive accuracy over time. The BS for 1-, 2-, and 5-year survival was 0.039
Black breast cancer survivors, 61% completed biweekly sessions during the first 6 0.004, 0.065 6 0.005, and 0.119 6 0.005, respectively, indicating excellent cali-
3 months, with a drop in participation after transitioning to monthly sessions. Par- bration. The top five variables ranked by SHAP values include age at the survey (0.048),
ticipants reported benefits in addressing distress and accessing support programs. use of special equipment due to health problems (0.029), employment status (0.020),
Future efforts should focus on tailored strategies to enhance engagement and retention. number of CHCs (0.016), and BMI (0.015). Conclusions: By integrating social risk
Research Sponsor: OSUCCC Intramural Research Program; Speilman Fund. factors with traditional risk predictors, we developed an ML model that predicts overall
survival with high accuracy and excellent calibration for adults with cancer in the U.S.
Identifying key risk social factors enables targeted interventions, potentially improving
health outcomes and management for the adult cancer population. Research Sponsor:
None.

1640 Poster Session 1641 Poster Session

Pioneering cancer survivorship care in Latin America: Early results from Oncology primary care clinics for comprehensive care of high-risk adoles-
the OC sobre VIVER program in Brazil. First Author: Luciana Landeiro, cent and young adult (AYA) cancer survivors. First Author: Alique Gabrielle
Oncoclı́nicas&Co, Salvador, Brazil Topalian, University of Cincinnati, Cincinnati, OH
Background: Cancer survivorship poses a growing challenge, especially in low- and Background: Adolescent and young adult (AYA) cancer survivors experience early
middle-income settings. The OC SobreVIVER program, launched in October, 2020 by the development of chronic medical conditions compared to healthy peers. Due to their
Oncoclı́nicas Group, addresses survivorship care gaps by providing structured, multi- young age at diagnosis and living decades beyond treatment, they are also at higher risk
disciplinary care. Initially piloted for breast cancer survivors in two Brazilian units, the for second primary malignancies (SPM) and late effects than older adult-onset cancer
program expanded in 2023 to include colorectal and prostate cancer survivors across survivors. Primary care providers are responsible for most long-term care of survivors
eight states. This study presents preliminary data. Methods: This descriptive, obser- and many are unfamiliar with the effects of cancer treatment in younger populations.
vational study assessed the OC SobreVIVER program’s impact on survivorship care. Oncology primary care providers are uniquely positioned to address increased needs of
Multidisciplinary consultations involving oncologists, nurses, psychologists, and nu- AYA patients because of their additional survivorship expertise. Methods: In 2020, the
tritionists addressed late toxicities, recurrence risk, and quality of life, following in- University of Cincinnati Cancer Center established an oncology primary care clinic. An
ternational survivorship guidelines. Telehealth and digital tools improved accessibility. accompanying clinical registry was developed to track patient outcomes longitudinally
Data from July, 2023 to December, 2024 included clinical characteristics, program Electronic medical records of all patients seen between 1/2021 and 1/2025 (n = 901)
metrics, and patient-reported satisfaction via EORTC PATSAT 33 and Net Promoter were extracted and entered in REDCap. Records of AYA cancer survivors, defined as
Score (NPS). Descriptive statistics summarized key outcomes. Results: From July 2023 cancer diagnosis between the ages of 18-39, were queried and analyzed (9%, n = 85).
to December 2024, the program supported 577 survivors (518 breast, 47 colorectal, 12 Results: The patient population’s mean age was 36 yrs (range 20-74; std dev = 11.3).
prostate) across 11 practices in eight states, with 490 initial consultations, 91 follow- Hematologic cancers (37%) were most common followed by breast (13%) and brain (9%).
ups, and over 20 support groups [Link] consultations were available to all Additionally, 14% were diagnosed with a SPM. Comorbid conditions were prevalent with
patients, enhancing accessibility for underserved regions. Patient-reported satisfaction 60% of patients having cardiovascular disease such as hypertension. Neurologic (46%),
was high, with 95% of respondents classified as Promoters (scores 9–10) in the NPS, endocrine (44%), and psychologic (71%) co-morbidities were also common. Over half of
demonstrating strong program loyalty. Based on EORTC PATSAT 33 evaluations, over patients were overweight/obese (68%) and many patients were former (19%) or current
80% of respondents rated their medical care experience in the highest satisfaction range (7%) smokers. Eligible patients received their recommended cancer screening including
(41–50), and over 80% rated their nursing care experience in the highest category breast (82%), colon (60%), and cervical (40%). Due to treatment exposures, 53% of
(29–35). Minimal dissatisfaction was reported, with less than 2% of responses falling patients were eligible for cardiomyopathy screening and 73% received recommended
into the lowest satisfaction ranges (, 20 for medical care and , 14 for nursing care). echocardiograms. A reduced ejection fraction was found in 26% of patients screened.
Conclusions: The OC SobreVIVER program demonstrates the feasibility of a structured, Conclusions: Comprehensive primary care services and longitudinal monitoring are
multidisciplinary survivorship model in a private network across Brazil. Preliminary data imperative in this high-risk population. Oncology primary care provides necessary
highlight its effectiveness in addressing complex survivorship needs, enhancing ac- survivorship-informed care and longitudinal monitoring for early onset comorbidities
cessibility through telehealth, and maintaining high [Link] integrated, patient- and SPMs. Tailored education and outreach efforts for providers and patients should
centric approach ensures continuity of care across physical, social, psychological, address preventative health services needed in this high-risk population. Research
spiritual, and nutritional dimensions, fostering a more comprehensive and holistic Sponsor: None.
survivorship experience. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 107s

1642 Poster Session 1643 Poster Session

Risk and predictors of late second primary malignancies in long-term breast, Subtype-specific trends in lung cancer incidence and survival: A SEER
prostate, colon, and rectal cancer survivors. First Author: Tendai Kwaramba, analysis (2004–2021). First Author: Shubhank Goyal, University of Texas Rio
Department of Medical Oncology, Yale Cancer Center, Yale School of Medicine, New Grande Valley, Mcallen, TX
Haven, CT Background: Lung cancer remains a leading cause of cancer mortality, with adeno-
Background: In older, long-term (5-year) cancer survivors, mortality risks from aging and carcinoma, squamous cell carcinoma (SCC), and small cell lung cancer (SCLC) as key
treatment-related effects may surpass those of their index cancer. Second primary ma- subtypes. Advances in early detection (e.g., low-dose CT screening post-2013) and
lignancies (SPMs) occurring 5-10 years post-diagnosis are understudied in older patients. systemic therapies (e.g., targeted agents and immunotherapies [IO], approved 2015 for
This study aims to quantify SPM risk, identify predictors, and describe prevalent SPM sites in adenocarcinoma/SCC, 2018 for SCLC) have transformed care. However, links between
older survivors of breast, prostate, colon, and rectal cancer to guide survivorship care. incidence trends and survival outcomes remain [Link] study leverages SEER data
Methods: This retrospective cohort study analyzed patients aged 66+ with stage I-III cancer (2010–2021) to extend prior analyses (Howlader et al. 2000-2017, NEJM 2020) and
diagnosed between 2003-2011 using the SEER-Medicare database. Eligible patients evaluate subtype-specific trends, focusing on the post-2015 era of IO adoption and
survived $5 years post-diagnosis and had continuous Medicare Parts A & B enrollment from updated screening eligibility. Methods: Age-adjusted incidence and 3-year relative
1 year pre-diagnosis to 1 year post-diagnosis. The primary outcome was late SPMs oc- survival rates for adenocarcinoma, SCC, and SCLC were analyzed using SEER data
curring 5-10 years after index cancer diagnosis. Covariates included demographics, (2004–2021), stratified by stage. Joinpoint regression identified significant trend
comorbidities, index cancer characteristics, treatment, and early SPMs (diagnosed within 5 changes. Survival shifts .2% annually were temporally linked to key milestones: low-
years). Least absolute shrinkage and selection operator for variable selection and 5-year
dose CT screening (2013), EGFR/ALK inhibitors (2013), and IO (2015). Results: For
restricted mean survival time regression models were used. Cumulative late SPM incidence
adenocarcinoma, distant-stage survival rose from 12.8% (2011) to 23.0% (2018), with a
was calculated with mortality as a competing risk. The prevalence of specific SPMs was
sharp +3.1% annual increase in 2018 aligning with immunotherapy adoption. Localized
calculated as a proportion relative to the total number of SPMs within each cohort, and
categorized as hematologic, predominantly screen-detected (breast, prostate, colorectal), or
and regional survival improved by 5.3% and 6.5%, respectively, over the study period.
other solid tumors. Results: Of the 88,227 long-term survivors included with median age of These trends probably reflect the combined impact of early detection and systemic
73.3 (IQR 69.5-78), 6.2% developed early SPMs and 8.2% (7,231) developed late SPMs. The therapies. In SCC, localized-stage survival increased from ~56% to ~66% after 2015,
5-year cumulative incidence of late SPMs was 8.6%, highest in prostate (9.2%) and lowest in driven by earlier detection and reduced late-stage incidence. However, regional and
breast (6.7%) cancer survivors. Non-screenable cancers had the highest 5-year risk (6.2%), distant-stage survival saw only modest gains (~7%), highlighting limited advancements
followed by screen-detected (1.3%) and hematologic malignancies (1.1%). Lung was the for advanced disease. SCLC showed the least survival improvements despite declining
most common SPM overall (18.4% of SPMs), including in survivors of breast (21%), rectal incidence, largely attributed to reduced smoking. Distant-stage survival stagnated at
(19.2%) and colon (16.5%) cancers, while prostate was most common in rectal cancer ~4–6%, while localized-stage survival improved transiently by 17% from 2010 to 2012,
survivors (17.0%). Diagnosis of SPM in the early ( , 5 years) survivorship cohort was likely reflecting temporary improvements in early-stage management. However, overall
associated with shorter time to a new late SPM, particularly in prostate cancer survivors outcomes for SCLC remain poor, underscoring an urgent need for new therapies.
(RMST Ratio 0.97, 95% CI 0.96-0.98). Treatments and high-risk disease features showed no Conclusions: Advancements in lung cancer treatment and cancer screening have led to
significant associations with occurrence of late SPMs. Conclusions: Late SPMs were significant survival gains. Adenocarcinoma’s distant-stage survival improvements
diagnosed in 8.6% of older, long-term cancer survivors. Lung cancer was the most common highlight the transformative impact of immunotherapy, while localized SCC gains
SPM overall. Some screen-detected SPMs, such as prostate were also common which is emphasize the role of early detection and screening. SCLC remains a major challenge,
notable in a population aging out of screening guidelines. Prediction of SPMs was limited by with survival rates stagnating despite a steady decline in incidence. Further innovation in
the absence of modifiable risk factors, genetic data, and family history in SEER-Medicare SCLC therapies, along with equitable access to both screening and novel treatments, is
data. Early SPMs were the sole predictor of late SPMs while treatment and index cancer crucial to improving outcomes across all lung cancer subtypes. Survival data beyond
features showed no effect, suggesting other drivers of late SPM development in older 2018 for metastatic disease and post-2024 for localized disease (following IO approvals
survivors. Research Sponsor: American Cancer Society. for respective indications) will provide deeper insights into the real-world impact of
these therapeutic advancements. Research Sponsor: None.

1644 Poster Session 1647 Poster Session

Telehealth and cancer care delivery: An updated real-world analysis of the Pilot study of an enhanced telehealth program for patients with prostate
impact of telehealth on patient access and treatment utilization at a com- cancer. First Author: Erin Mary Bange, Memorial Sloan Kettering Cancer Center, New
prehensive cancer center. First Author: Kelsey H. Natsuhara, University of California, York, NY
San Francisco Helen Diller Family Comprehensive Cancer Center, San Francisco, CA Background: The time required for cancer treatment is considerable. Telehealth (TH) with home
Background: Telehealth in oncology (onc) has persisted due to its convenience and monitoring can minimize travel burden but patients’ perceptions regarding tradeoffs of clinic visits
potential to improve equitable access to care. Data on telehealth’s impact on treatment versus enhanced telehealth (ET) are uncertain. We piloted an ET intervention offering patients on
(tx) access are needed to guide long-term telehealth policies. Methods: We identified androgen deprivation therapy (ADT) for prostate cancer up to four components of care at home.
Methods: Prostate cancer patients on ADT were offered participation in 4 at home services: 1) TH
adult patients (pts) who completed $1 medical, surgical, or radiation onc visit with an
oncologist visits; 2) remote BP monitoring (BP); 3) home phlebotomy (HP); and 4) self-injection of
associated cancer diagnosis at our center from 2017-2019 (pre-telehealth) vs 2021-2023 ADT or denosumab at home with RN support via TH as needed. Clinicians referred eligible patients
(post-telehealth). Data from 2020 was excluded, given COVID irregularities. We selected 3 and specified frequency of each service. Completion rates were evaluated for each component.
disease groups with varying telehealth use – breast (low), GI (medium), and GU (high) – Patient perspectives, preferences for subsequent care, and assessments of the feasibility, ac-
allowing us to better control for external factors (eg COVID). We compared changes in visit ceptability, and appropriateness of each component were collected after participation in eligible
distribution, sociodemographic, and tx patterns within and between disease groups pre- services at least once over the course of the pilot. Feasibility, acceptability, and appropriateness
vs post-telehealth using Chi-square and ANOVA tests. Logistic regression analyses were measured by validated measures (scores from 1-5). Patient satisfaction with ET was
identified post-telehealth predictors of receiving tx, including the percentage (%) of in- measured using a net promoter score. All results are based on the last completed patient survey.
person (IP) visits per pt. Results: We analyzed 109,200 encounters and 26,907 pts pre- Results: Between 6/2023 and 6/2024, 39 patients enrolled in the pilot, the median age was 70
telehealth vs 143,159 encounters and 50,168 pts post-telehealth. Pt volume increased in (IQR: 61.5-76.7), 62% were white, and 69% (N = 27) lived with family or a partner. Table 1 shows the
all groups post-telehealth (+99% breast, +55% GI, +104% GU). Post-telehealth, the % of %/(N) of patients who chose to participate in each at home component with completion rates at the
patient and visit level. 73% (145/198) of remote monitoring failures were due to patient inability to
video visits (VVs) increased in breast (1.5% to 28.8%), GI (2.0% to 55.2%), and GU (3.0% to
execute the at home task. . 75% of patients reported that each service was convenient and saved
80.9%). In all groups, the % of pts seen from outside the SF Bay Area decreased (-7.5 time. Most patients reported no problems with completing the visit (TH 89%, BP 89%, HP 91%,
breast, -1.4 GI, -6.1 GU). In breast and GU, the % of pts receiving cancer tx decreased (-5.2, injections 67%). Patients were either very likely or somewhat likely to participate again (TH 97%, BP
-16.0, p,0.01), while the % of pts receiving GI tx was stable (-0.9). For infusion tx, 100%, HP 96%, injections 100%) and the majority preferred ET for future care (TH 79%, BP 79%, HP
predictors of tx receipt included metastatic disease (OR 3.9, 95% CI 3.6-4.2) and higher % 87%, injections 100%). 62% (24/39) of patients found at-home visits to be equally or less stressful
of IP medical onc visits (OR 3.6, 95% CI 3.3-3.9); while living outside the Bay Area was than in-person care. Patients found each service to be feasible, acceptable, and appropriate (mean
negatively associated with infusion tx (OR 0.5, 95% CI 0.4-0.5). For surgery, the % of IP score . 4.7 for all services) and gave ET a net promotor score of 79. Conclusions: Patients found
surgical onc visits (OR 0.7, 95% CI 0.6-0.7) and living outside the Bay Area (OR 0.8, 95% CI ET for the management of prostate cancer on ADT to be feasible, acceptable, and appropriate,
0.7-0.9) were negatively associated with tx receipt. For radiation tx, metastatic disease providing a more patient-centered and convenient alternative to traditional clinic-based care.
(OR 2.8, 95% CI 2.5-3.1), but not living outside the Bay Area (OR 0.6, 95% CI 0.5-0.7), was Future studies will explore scalability and applicability across diverse patient populations and
associated with tx receipt. The % of IP radiation onc visits was not significant (OR 1.1, 95% treatment regimens. Research Sponsor: National Cancer Institute/U.S. National Institutes of
Health; P30 CA008748; National Cancer Institute/U.S. National Institutes of Health; P50 CA271357.
CI 1.0-1.4). Conclusions: Among 3 disease groups with varying telehealth use, pt volume
increased in all groups, while the % of pts receiving tx at our center decreased. This % (N) Patients % (N) Patients completing
suggests that providers may be providing collaborative care while pts receive care locally. opting in for each component % (N) Total
each component (at least once) episodes completed
In regression analyses, higher % of IP medical onc visits was positively associated with
receiving infusions, highlighting the importance of IP visits during active medical onc tx. TH visits 100% (39) 59% (23) 90% (53/59)
Remote BP monitoring 92% (36) 94% (34) 60% (296/494)
However, this was not seen for surgery and radiation tx. For these time-limited inter- Home phlebotomy 69% (27) 100% (27) 97% (143/148)
ventions, pts seen virtually are still likely to access tx at our center. Further analyses are Home injections 28% (11) 73% (8) 90% (9/10)
needed to identify pts and visit types best suited for telehealth. Research Sponsor: All 4 components 18% (7) 29% (2) 70% (501/771)
Conquer Cancer Foundation Young Investigator Award.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
108s CARE DELIVERY/MODELS OF CARE

1648 Poster Session 1649 Poster Session

Impact of telehealth utilization on adherence to endocrine therapies in Comparative effectiveness of delivering early palliative care via video versus
privately insured breast cancer patients: A claims-based cohort study. in-person on end-of-life outcomes in patients with advanced lung cancer.
First Author: Shaimaa Elshafie, Clinical and Administrative Pharmacy, College of First Author: Jessica R. Bauman, Fox Chase Cancer Center, Philadelphia, PA
Pharmacy, University of Georgia, Athens, GA Background: Findings from our prior large-scale comparative effectiveness trial showed
Background: Adherence to endocrine therapies is essential for reducing recurrence and the equivalent effect of delivering early palliative care via video versus in-person on quality
improving survival in breast cancer patients. However, nonadherence remains a sig- of life among patients with advanced non-small cell lung cancer (NSCLC). We now report
nificant challenge, particularly among young women. Telehealth has emerged as a on whether the two care delivery modalities were equivalent with respect to patient-
promising tool to address barriers to care, yet its impact on long-term adherence to reported communication with clinicians about their end-of-life (EOL) care preferences and
endocrine therapies is poorly understood. This study evaluated the association between hospice utilization. Methods: Between 6/14/2018 and 5/4/2023, we enrolled 1250 pa-
telehealth utilization and adherence to endocrine therapies among privately insured tients with newly diagnosed advanced NSCLC in a randomized trial of early palliative care
women with nonmetastatic breast cancer. Methods: Using medical and pharmacy across 22 US cancer centers. Patients were randomly assigned to meet with a palliative
claims data from the Merative MarketScan database, we identified women under 65 care clinician every 4 weeks from enrollment through the course of the disease, either via
years old who were newly diagnosed with nonmetastatic breast cancer in 2018 and were video or in the outpatient clinic. Participants completed self-report surveys at baseline and
commercially insured for at least one year before diagnosis and five years after initiating weeks 12, 24, 36, and 48, including an item asking if they had discussed with their clinicians
endocrine therapy. Telehealth utilization was identified through billing codes, and the care they would want to receive if dying (yes/no); patients’ final assessments prior to
adherence was measured by the proportion of days covered (PDC) with adherence death or last follow up were analyzed. We reviewed patients’ health records to collect data
defined as $80% prescription coverage. Generalized linear mixed models were used to on hospice referral and length of stay. To test the equivalence in these outcomes, we
estimate odds ratios (ORs) and 95% confidence intervals (CIs) while adjusting for used a binomial generalized linear model with the identity link function (pre-specified
relevant patient-level covariates. Results: The study cohort included 1,141 patients with equivalence margin of 68% for patient-reported communication about EOL care) and
the majority (53%) aged 45-54 years and 92% exhibited moderate severity of comor- linear regression (pre-specified equivalence margin of 66 days for mean length of stay in
bidities. All patients initiated endocrine therapy in 2018 with 34% receiving tamoxifen, hospice). P-values were adjusted for multiplicity using a Bonferroni correction. Results: Of
the 1250 enrolled participants, 888 (71.0%) completed at least one survey post baseline
33% receiving an aromatase inhibitor, and 33% switching between agents during follow-
regarding whether they communicated with clinicians about EOL care preferences. Among
up. Telehealth utilization increased sevenfold during the second year of treatment
those, 29.1% of the video group and 26.0% of the in-person group reported “yes,” indicating
(coinciding with the COVID-19 pandemic) and peaked in the third year when 48% of
that they recalled such EOL care discussions (difference = 3.1%, 95% CI: -1.8%, 8.1%; p =
patients used telehealth and 2,178 visits were recorded. Adherence rates declined
0.26 for equivalence). During the course of the trial, 733/1250 (58.6%) patients died, of
steadily over time: from 75% in the first year to 36% by the fifth year. Patients who whom 537/733 (73.3%) were referred to hospice. Mean lengths of hospice stay were
utilized telehealth were 58% more likely to be adherent compared to those who did not 25.3 days for the video group versus 25.1 days for the in-person group (difference = 0.2,
utilize telehealth (OR = 1.58; 95% CI: 1.31-1.91; p , 0.0001). Geographic region, in- 95% CI: -7.0, 7.4; p = 0.46 for equivalence). When excluding outlying patients receiving
surance plan type, and endocrine therapy agent were also significant predictors of hospice services . 180 days (n = 13), the mean lengths of hospice stay were 19.1 (video
adherence. Conclusions: Telehealth utilization was significantly associated with im- group) versus 19.7 (in-person group) days (difference = -0.6, 95% CI: -4.6, 3.3; p = 0.06 for
proved adherence to endocrine therapies among privately insured breast cancer pa- equivalence). Conclusions: Although thresholds were not met to confirm equivalence
tients. These findings highlight the potential of telehealth to mitigate barriers to long- statistically, the two modalities for delivering early palliative care demonstrate very similar
term treatment adherence. Future research should explore strategies to sustain ad- outcomes with respect to patient-clinician communication about EOL care and hospice
herence and evaluate the broader clinical and economic implications of telehealth in this utilization. These findings provide further evidence of the utility of video visits for providing
population. Research Sponsor: None. high quality palliative and EOL care. Clinical trial information: [Link] Identifier
(NCT03375489). Research Sponsor: Patient-Centered Outcomes Research Institute; PLC-
1609-35995.

1650 Poster Session 1651 Poster Session

Assessing circadian rhythms and chemotherapy safety in remote patients Remote physiologic and behavioral monitoring to predict early treatment
with pancreatic ductal adenocarcinoma (PDAC) using a multidimensional response in metastatic cancer: High-Definition Oncology study (HDOs)
digital platform (MultiDom, NCT04263948). First Author: Francis Albert Lévi, UPR preliminary results. First Author: Leire Paz-Arbaizar, Department of Signal Theory
Chronotherapie, Cancers et Transplantation, Université Paris Saclay, Hôpital Paul and Communications, University Carlos III, Leganés, Spain
Brousse ID Isco 13918, Villejuif, France Background: Emerging evidence suggests that behavioral, physiologic or emotional
Background: The disruption of circadian clocks is associated with reduced survival and factors may act as real-time indicators of treatment response, with potential as modifiable
treatment tolerability in cancer patients (pts). Here, real-time analyses of telemonitored factors. Advances in remote monitoring technologies provide passive (e.g., heart rate,
circadian rhythms and electronic Pt-Reported Outcome (ePRO) are integrated within a sleep patterns) and active (e.g., self-reported emotions) data. HDOs collects such data and
multidimensional telemonitoring-telecare digital platform that triggers proactive telecare serial -omics from 300 women with metastatic cancer to identify novel markers, un-
toward improved quality of life (QoL) and treatment safety. Methods: The multicentre, derstand disease trajectories and develop a digital twin for individualized care. We present
interventional, prospective, longitudinal, single-arm study recruited pts receiving mFOLFIR- data from 25% accrual. Methods: Women receiving first-line treatment for metastatic
INOX chemotherapy (CT) q2-weeks for pancreatic ductal adenocarcinoma (PDAC). Early colorectal, lung, or hormone-positive breast cancer were eligible. Patients continuously
warning signals of circadian disruption, body weight loss, and ePROs severity are extracted wore a smartwatch and used the EB2 App to capture step count (SC), sleep duration (SL),
from real time analysis and graphical displays of (i) continuously telemonitored rest-activity phone usage (PU), time at home (TH), location clusters (LC), mean (MHR) and minimum
and chest surface temperature (chestemp) rhythms using a chest sensor, and (ii) daily body (mHR) heart rate, mean (MSHR) and minimum (mSHR) sleeping heart rate and sleeping
weight (e-balance), and (iii) daily self-rating of 23 symptoms using a GPRS tablet (MD oxygen saturation (SOS). Emotional valence was self-reported from a list of 20 emotions
Anderson Symptoms Inventory, MDASI). Pts participate for 1 week before (baseline) and and classified as negative (-1), neutral (0) or positive (+1). Aim 1: to explore the relationship
6 weeks after 1st CT course. Circadian disruption is defined by an (I , O) value , 96%. (I , O) is
between the variables and response (CB: CR+PR+SD vs. PD) at the first CT scan at day +90
the % accelerations per min In-Bed that are below median accelerations per min Out-of-Bed for
analyzing data from days 1-15 and 75-90 (Mann-Whitney U). Aim 2: to find Response-
3 days; (I , O) range in controls, 97-100%). Automatic alerts are sent via internet for decision
Associated Behavioral Patterns (RABPs) associated with CB or PD. First, Daily Behavioral
of proactive intervention by the oncology team, in case of circadian disruption, chestemp
increase by 1.5°C, weight loss . 5%, or MDASI symptom $ 7. Results: From 6/2021 to 7/ Profiles (DBPs) are obtained using unsupervised learning models from . 2 million days of
2024, 58 pts with advanced PDAC were selected (male, 52%), median age, 58 y.o. (range, 33- smartwatch and App data (external set). After identifying 256 DBPs with the VQ-VAE
82); WHO performance status (PS) 0/1/2, 36%/53%/10% of the pts; metastatic sites 0/1/ . 2, model, Latent Dirichlet Allocation defined RABPS based on the frequency and abundance
38%/24%/8%; liver metastases, 50%). Early PDAC-related complications prevented platform of DBPs per patient. RABPs were compared among classes (response type, age group)
use in 5 pts. The platform was used by 53 pts during a median of 45.5 days (IQR, 38-50], with . using X2. Bilateral P values , 0.01 were deemed significant. Results: from May 2023 to
85% compliance. At baseline, large between-pts differences in circadian rhythms were found April 2024, 77 female patients (median age 61; 28-80) were accrued (46 Breast, 23 Lung, 8
for both rest-activity I , O (median [IQR]), 98.4 accelerations/min [95.8-99.6]); range, 75 to Colorectal). At first CT, 72 (93.5%) achieved CB while 5 (6.5%) had PD. During days 1-15, CB
100), and chestemp circadian amplitude (median, 1.1°C [IQR, 0.7-1.4], range, 0.3°C to 2.7°C). patients showed lower PU (2.4 vs. 3.9 hours; P = 0.002), TH (18.5 vs. 22 hours; P = 2* 10^-7),
Baseline circadian disruption was larger in male pts (56% vs 31%; p = 0.07) and in those with PS MHR (78 vs. 88 bpm), mHR (59 vs 70 bpm), MSHR (75 vs 88 bpm) mSHR (66 vs 78 bpm) (all
1-2 (47% vs 14%; p = 0.02). Consistently, median chestemp circadian amplitude was less in Ps , 10^-10) and reported more negative EV. The trends persisted in days 76-90 in addition
males compared to females (0.8°C vs 1.3°C, p , 0.01) and in pts with PS = 1-2 compared to PS to SC (7235 vs 4038 steps/day; P = 0.00001) and decreased SOS (90.1% vs. 92.6%; P =
= 0 (0.8°C vs 1.3°C; p , 0.01). Maximum toxicities of CT were circadian disruption (100% of the 1.5*10^-8). Six RABPS were identified. Patients , 60 yo displayed more often RABPs 1, 2
pts), body weight loss . 5% (59%), and MDASI symptom $7 (46%), without any influence of and 5 (84% vs. 16%; P = 0.02). RABP1 breast cancer and RABP5 lung cancer patients were
baseline pt characteristics. Conclusions: The use of this multidimensional digital platform more likely to experience PD vs. CB (75% vs. 24%; P = 0.08; and 69% vs.19%, P = 0.07,
combining circadian and other physiology metrics with ePROs was feasible and accepted by respectively). Conclusions: Behavioral and physiologic data in days 1-15 and 76-90 were
the pts. Its implementation seemed to be clinically relevant toward improving the care of strongly associated with treatment response, independent of tumor type, age or treatment.
remote pts at risk of adverse events. Clinical trial information: 04263948. Research Sponsor: RABPS identifying patients at high risk of PD can be detected, highlighting their value as
Education and Research Direction of Ramsay-Santé. markers for early intervention. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 109s

1652 Poster Session 1653 Poster Session

Administration of tarlatamab in an outpatient setting utilizing remote pa- Teledermatology-dermoscopy: Expanding access to skin cancer screening
tient monitoring: Mayo Clinic experience. First Author: Ashley Potter, Mayo Clinic to reduce healthcare disparities. First Author: Brenda Santellano, Georgia Cancer
Rochester, Rochester, MN Center, Augusta University, Augusta, GA
Background: Tarlatamab is a bispecific T-cell engager targeting DLL3 that has shown Background: Skin cancer, the most common malignancy in the United States, continues
promise as a therapy in small cell lung cancer. Administering tarlatamab requires careful to rise in incidence. Underserved populations face significant barriers to care, including
consideration of immune-related adverse events (AE), including cytokine release syn- lack of insurance, language differences, and limited access to specialists. The Augusta
drome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). Free Dermatology Clinic, a student-run clinic, collaborates with the Teledermatology in
Prescribing information recommends patients to be monitored for 22-24 hours in an Rural Georgia program to address these challenges through store-and-forward (SAF)
appropriate healthcare setting for cycle 1 (C1), days 1 and 8. At Mayo Clinic in Rochester teledermatology-dermoscopy. This approach involves transmitting dermoscopic images
MN (MCR), tarlatamab is initially administered in a hospital-based outpatient (HBO) to a dermatologist for remote analysis. Implemented during Community Health Fairs
setting utilizing an innovative remote patient monitoring system (RPM). In this study, we (CHFs), this initiative aims to enhance access to skin cancer screening in underserved
describe our RPM process and patient outcomes thus far. Methods: We retrospectively communities. Methods: Volunteer medical students (MS) from the Medical College of
reviewed records of patients treated with tarlatamab from August through December Georgia (MCG) and general physicians underwent training to use a dermatoscope ef-
2024. All patients received tarlatamab in the HBO setting and were enrolled into RPM for fectively. Trained MS collected brief medical histories and dermoscopic images for SAF
C1, days 1–10. Patients attended daily HBO visits on days 1–3 and 8–10 of C1 and were referrals during CHFs serving individuals with incomes below 200% of the federal poverty
required to stay within 30 minutes of the hospital with a 24-hour caregiver for 48 hours line. Polarized dermoscopic images were captured using iPhones equipped with magnetic
following infusion days in C1-2. Data on patient characteristics, frequency of RPM alerts, dermatoscope attachments and were securely transmitted to a board-certified derma-
escalations, rate of CRS & ICANS, and need for hospitalizations were reported using tologist via a teledermatology platform. Dermatological recommendations were provided
descriptive metrics. Results: As part of RPM, patients are provided with a kit containing within one hour and communicated to patients, with Spanish translators facilitating
preconnected Bluetooth-enabled devices to measure vital signs (VS), including blood communication. Results: Across two CHFs (~8 hours each), 10 MS volunteered in shifts
pressure (BP), heart rate (HR), temperature (T), and pulse oximetry (SpO2). The kit also under the supervision of a general physician (n = 1) or dermatology resident (n = 1). A total
includes a cellular-enabled tablet for electronic ICANS questionnaires, which upload of 141 patients presented with dermatological concerns, of whom 24 (17.02%) were
directly into our electronic medical record system. Patients are required to log VS four referred for SAF consultations due to suspicious skin lesions. Among these, 17 (70.83%)
times daily, monitored in real-time by a centralized virtual RPM nursing (RN) team. had benign lesions, while 7 (29.17%) were identified as potentially malignant and referred
Embedded decision trees trigger alerts, prompting the RN team to contact patients and for in-person follow-up at the Augusta Free Dermatology Clinic for further evaluation or
follow care pathways for escalations as needed. Among the 16 patients treated, the biopsy. Comprehensive data were available for 112 patients (79.43%), most of whom were
median age was 68 years, and the median ECOG PS was 1. During the 10-day RPM female (58.04%, n = 65) and Latino/Hispanic (98.21%, n = 110). The majority were
period, a total of 2,233 VS entries were recorded. Of these, 88% (14/16) of patients had uninsured (73.21%, n = 82), Spanish-speaking (98.21%, n = 110), and required translation
alerts triggered, with a median of 2 alerts per patient. Alerts were mostly for out-of-range services (98.21%, n = 110). Nearly all participants worked in skilled agricultural roles
systolic BP (56%), HR (56%), and T (25%), with none for SpO2. The RN team escalated (98.21%, n = 110) and reported a median of two household dependents (range: 0–9).
alerts in 50% (8/16) of the patients. Overall, 38% (6/16) of patients were managed entirely Common diagnoses among follow-up patients included healthy skin (45.54%, n = 51),
in the HBO setting. Among the 63% hospitalized, 56% had CRS limited to grade (G) 2, and acne (5.36%, n = 6), melasma (5.36%, n = 6), benign nevi (5.36%, n = 6), dermatitis (4.46%,
31% had ICANS limited to G3. Two patients were hospitalized for other indications. n = 5), and folliculitis (3.57%, n = 4). Other less common conditions were diagnosed in
The mean length of stay was 2.6 days, with no ICU admissions. Conclusions: Our ex- 30.36% (n = 34) of patients. Conclusions: The implementation of teledermatology-
perience shows that frequent monitoring required with tarlatamab can be safely and dermoscopy at CHFs effectively addressed barriers to dermatological care in underserved
effectively executed in an outpatient setting with the utilization of an RPM system. It can populations. This approach demonstrated the potential to improve early detection of skin
potentially minimize the burden of hospitalizations for patients and the healthcare cancer, facilitate timely care, and reduce healthcare disparities. Research Sponsor: United
system. Research Sponsor: None. States Department of Agriculture Rural Utilities Service (USDA).

1654 Poster Session 1655 Poster Session

Patient-reported experience with an immunotherapy telehealth platform. Outcomes of germline expedited point of care (POC) genetic testing through
First Author: Robert Michael Daly, Memorial Sloan Kettering Cancer Center, New York, telehealth in the Veterans Health Administration (VA). First Author: Akiko Chiba,
NY Department of Surgery, Duke University Medical Center, DUMC, Durham, NC
Background: The Making Teleheath Delivery of Cancer Care at Home Effective and Safe Background: Germline genetic testing is standard of care for treatment planning for several
for Immunotherapy (MATCHES-IO) intervention seeks to improve the efficiency and patient malignancies. To increase access to genetic testing for Veterans, VA developed and disseminated
experience for those treated with single agent pembrolizumab. Because pembrolizumab is educational materials, laboratory portal access, and templates for ordering and documenting
administered as an outpatient infusion every 3-weeks, patients require up to 18 clinic visits consent and testing in the electronic health record to facilitate POC testing by oncology providers.
per year to receive treatment, which is arduous. During the COVID-19 pandemic, the FDA Here, we describe the outcomes of POC testing. Methods: POC tests ordered between 2/24/23-
granted accelerated approval for an extended interval dosing administered every 6 weeks, 11/18/24 by oncology providers at VA sites or with National TeleOncology (NTO) were identified
through the VHA’s Corporate Data Warehouse and the VINCI (VA Informatics and Computing
but despite this approval every 3-week dosing remains the standard (65% of prescriptions)
Infrastructure) research environment. Pathogenicity of variants was determined by the classi-
as clinicians perceive this schedule enables them to identify and manage toxicity early.
fication provided by the laboratory. Providers are recommended to use the POC testing for
Telehealth may be the impetus to change the standard to the more convenient 6-week in patients actively being treated for the following cancers: metastatic/high risk prostate, breast,
person schedule, but evidence is needed. Methods: We conducted a single arm pragmatic ovarian, exocrine pancreatic/ampullary, colon , 50 years of age, medullary thyroid, and
trial to evaluate the efficiency and patient experience of a telehealth immunotherapy pheochromocytoma/paraganglioma. Results: POC tests were ordered at 45 different VAs for
platform (MATCHES-IO) in patients with non-small cell lung, genitourinary, or melanoma 1293 patients. Total of 1364 tests were ordered, and 1195 (87.5%) tests were . The tests ordered
cancers receiving single agent pembrolizumab. MATCHES-IO evaluates whether in-person included 854 (62.6%) curated multigene panel and 510 (37.4%) targeted cancer panels. De-
visits for pembrolizumab therapy q6wk (rather than q3wk) with interim telehealth toxicity mographics are summarized in Table. Among 1382 diagnoses listed as the indication for testing
checks between in-person treatments for the first six months of therapy is more efficient (some patients had multiple), the most common cancer diagnoses were prostate (831, 60.1%),
and enhances patient experience relative to the standard q3wk infusion visits. The breast (206, 14.9%), and pancreatic (94, 6.8%). Most tests were ordered for patients who met POC
components of the platform include clinician-patient virtual visits, labs at home, biometric indications, but 13.6% of orders were for other indications. A total of 77 (6.4%) patients were
devices at home for vital sign monitoring, and electronic patient-reported outcomes to found to have pathogenic/likely pathogenic variants (PV) in dominantly inherited cancer pre-
monitor for common IO-related toxicities. Patient experience was assessed after each disposition genes. Seventy-one (5.9%) patients had PVs in high/moderate penetrance cancer
MATCHES-IO televisit for up to two televisits. We measured experience with a patient predisposition genes. Conclusions: POC testing is feasible and being widely adopted across the
experience survey that included how likely are they to recommend this intervention to VA. Further work is needed to determine if patients found to have actionable PVs through the POC
similar patients (scale 0-10). Results: Between July 2023 and January 2025, 59 patients mechanism have changes in treatment or are referred for follow-up genetic counseling. A
significant, minority of tests were ordered for patients with diagnoses not eligible for POC testing.
were enrolled, median age 69 (range 25–85), 81.5% White, 11.1% Black, 7.4% Asian, and
Continued education and support is planned to increase utilization of POC testing in oncology
61.1% male. Cancer types included thoracic (50.9% of patients), genitourinary (30.5%), and across the VA. Research Sponsor: None.
melanoma (18.6%). 45 patients (76.3%) have completed a MATCHES-IO televisit and
completed a patient experience survey. The median score for likelihood to recommend was Demographics of veterans with a point of care genetic testing order (n=1293).
10 (range 4-10). 97.8% patients perceived a benefit to the MATCH-IO televisit including White/Other races 701 (54.2%)/ (45.8%)
saved time (82.2% of respondents), patient convenience (71.1%), convenience for Non-Hispanic/Hispanic 1225 (94.7%)/29 (2.2%)
caregiver/family (44.4%), saved money (44.4%), and better monitoring of cancer and Male/Female 1034 (80%)/259 (20%)
treatment (24.4%). 57.8% of patients found the at home visit less stressful than the in- Rural/Urban 369 (29.5%)/915(70.8%)
POC test orders
person visit. Conclusions: Patients endorsed an enhanced experience with an immu- Testing completed (n=1364) 1195 (87.6%)
notherapy telehealth platform for extended dosing of pembrolizumab. Further follow-up is Comprehensive panel/targeted panel (n=1364) 854 (62.6%)/510(37.4%)
needed to confirm these experience findings and determine whether this platform im- POC eligible cancer diagnosis (n=1382) 1194 (86.4%)
Pathogenic/likely pathogenic variant in a dominant 71 (6.4%)
proved efficiency through fewer in-person visits. Research Sponsor: National Cancer high/moderate risk gene (n=1195)
Institute; Emerson Collective Digital Oncology Care.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
110s CARE DELIVERY/MODELS OF CARE

TPS1656 Poster Session TPS1657 Poster Session

DIPCAN, a multidimensional approach to precision oncology: Harnessing A pilot single-arm, pragmatic trial in progress of in-home versus in-clinic
genomic, clinical, pathological and radiographic data to advance personal- subcutaneous nivolumab administration through Cancer Care (connected
ized cancer treatment. First Author: Enrique Grande, Department of Medical On- access and remote expertise) Beyond Walls (CCBW) program. First Author:
cology, MD Anderson Cancer Center Madrid, Madrid, Spain Dina Elantably, Mayo Clinic, Jacksonville, FL
Background: Advances in big data analytics and artificial intelligence (AI) are enabling Background: Cancer treatments are traditionally administered in clinical settings, which
novel approaches to patient classification in oncology. While existing studies often can isolate patients from their familiar environments and exacerbate physical, psy-
correlate only a few data types, the DIPCAN Study (Digitalisation and Integral Man- chosocial, and financial burdens. Travel requirements further amplify these challenges,
agement of Personalised Medicine in CANcer) seeks a comprehensive, integrated particularly for underserved populations. Studies indicate that patients prefer receiving
analysis combining phenotypic, clinical, pathological, radiomic, and genomic data from care at home, and international models have demonstrated the safety of home-delivered
patients with metastatic cancer in Spain. DIPCAN aims to deepen insights into cancer’s chemotherapy since the 1990s; however, no U.S. clinical trial data exists. In response,
multifactorial nature, driving personalized care and more precise therapeutic strategies. Mayo Clinic has developed the Cancer CARE (Connected Access and Remote Expertise)
Methods: DIPCAN was initiated through a consortium comprising five technology and Beyond Walls (CCBW) program, a distributed cancer care delivery model that expands
healthcare SMEs—Genomcore, Quibim, Pangaea Oncology, Artelnics, and Atrys access to quality cancer care by bringing it to the home environment, providing in-home
Health—alongside Eurofins Megalab and the non-profit MD Anderson International cancer treatment, lab testing, telemedicine, and community paramedic support. This
Foundation Spain. Funding was secured via the Spanish Ministry of Economic Affairs trial evaluates the safety, acceptability, and impact of home-based subcutaneous (SC)
and Digital Transformation under the EU-funded Recovery, Transformation, and nivolumab (Nivo) administration compared to in-clinic treatment within the CCBW
Resilience Plan (R&D Missions Program in Artificial Intelligence, File No. initiative. Methods: This open-label, single-arm trial evaluates the impact of SC Nivo
MIA.2021.M02.0006). DIPCAN’s primary objective is to characterize and map clinical, administration location—home versus infusion center—on patient reported cancer care
phenotypic, genomic, and radiomic profiles of metastatic cancer patients across Spain. experience, patient-preferred treatment location, acceptability, safety, and patient-
Secondary goals involve developing Big Data, AI, and machine learning tools to enable reported outcomes. Eligible adult patients (ECOG 0-1) receiving IV Nivo for an FDA-
multidimensional analysis of these patients. Eligible patients are 18 years or older, have approved indication and residing within 75 miles of Mayo Clinic Florida will transition to
histologically confirmed metastatic solid tumors, a life expectancy exceeding three SC Nivo, receiving two initial in-clinic cycles. If tolerated without injection reactions, four
months, and available tumor material for histological and molecular analyses. Par- cycles will be administered at home before resuming in-clinic treatment. Exclusion
ticipants consent to undergo a comprehensive set of diagnostic and imaging procedures criteria include concurrent investigational/standard treatments or contraindications to
outlined in the study protocol. If recent tumor tissue (,3 years) is unavailable, patients immunotherapy. Fifty patients will be enrolled, with an estimated 75% (n=38) providing
may opt for a current biopsy or liquid biopsy. At no cost, participants receive con- cancer care ratings after 8 weeks in-clinic and 8 weeks at-home, offering 85% power to
sultations with oncology and drug development specialists, who document baseline detect a mean difference in ratings of 0.50 standard deviations. The primary endpoint is
characteristics and compile structured medical histories. Additional diagnostics include within-patient change in cancer care rating (0–10 scale, CAHPS Cancer Care Survey)
bloodwork emphasizing lipid metabolism, digital pathology, extensive NGS sequencing comparing 8 weeks of in-clinic vs. at-home care. Secondary endpoints include patient-
on tissue or blood, and a full-body MRI. All participants receive digital access to their preferred treatment location, comfort with home injections, safety (grade 3+ adverse
data and a clinical report with tailored recommendations for their physicians. With ethics events), function (EORTC QLQ-F17), symptoms (PRO-CTCAE), side effect impact (GP5),
approval in place, DIPCAN has enrolled 1,500 patients since June 14, 2022. Data and healthcare utilization (ER visits, hospitalizations). Cost will be assessed as a tertiary
collection is ongoing, with anticipated advancements in AI-driven analysis aimed at endpoint. The trial has FDA approval (IND #170079), Mayo Clinic IRB approval
refining precision oncology approaches for metastatic cancer in Spain. Clinical trial (#23009663), and [Link] registration (NCT06265285). Enrollment began in
information: 2021.M02.0006. Research Sponsor: European Union; MIA.2021.M02.0006. April 2024, with an expected accrual period of 24 months. To date, 10 patients have been
enrolled, and final analysis is expected 2.5 years after trial activation. Clinical trial
information: NCT06265285. Research Sponsor: BMS; Mayo Clinic.

TPS1658 Poster Session TPS1659 Poster Session

Trial in progress: Evaluating the effectiveness of Blue-button—A tool for DISCO App: A patient intervention to reduce the financial burden of cancer
institution-agnostic, EHR-integrated regional automated clinical trial pre- in a diverse patient population. First Author: Lauren M. Hamel, Karmanos Cancer
screening and matching. First Author: Waddah Arafat, Simmons Comprehensive Center, Wayne State University, Detroit, MI
Cancer Center, University of Texas Southwestern Medical Center, Dallas, TX Background: Financial toxicity, the material and psychological burden of treatment cost,
Background: Clinical trial enrollment is essential for advancing cancer treatment and affects up to half of people with cancer and can affect adherence and survival. Financial
improving patient outcomes. Despite the benefits, only 7% of adult cancer patients in the toxicity is a health equity issue, disproportionately affecting Black patients. Patient ed-
US enroll in clinical trials due to barriers such as limited awareness, the time-intensive ucation about cost and patient-provider cost discussions are recommended to mitigate
nature of manual prescreening and lack of relevant on-site clinical trials. This contributes financial toxicity but occur infrequently. Our goal is to mitigate financial toxicity through a
to insufficient enrollment, causing approximately 20% of trials to fail. Automated tailorable education and communication intervention, the DISCO App. The DISCO App was
prescreening using electronic health records (EHRs) offers a promising solution to shown to be feasible, acceptable, and preliminarily effective at prompting cost discussions
streamline trial identification and improve access. This study builds on our published and improving related outcomes in a pilot trial. The aim of the ongoing trial is to test the
feasibility study (Cancer 2023; doi: 10.1002/cncr.35022) that demonstrated the potential effectiveness of the DISCO App on short- and longer-term outcomes for Black and White
of an open-source clinical trial matching tool, developed in collaboration with ACS CAN patients with cancer. Methods: This study is a longitudinal RCT. Oncologists are eligible if
and MITRE Corporation, to improve locoregional trial identification for patients. they treat patients with solid tumors at the trial site. Patients of participating oncologists
Methods: Trial Design: This prospective, randomized, two-arm pilot study is being are eligible if they are $ 18 years of age; identify as Black or White; can read and write in
conducted at two sites: University of Texas Southwestern Medical Center (UTSW), an English; have an email address; and were diagnosed with a solid tumor for which systemic
academic center, and Tampa General Hospital (TGH), a community hospital. Patients are therapy is a likely recommended treatment. Strata were created to balance arms by patient
randomized to usual care or an intervention arm utilizing Blue-button, an automated race, income, age, and sex. Upon consent, patients are randomized to one usual care arm
prescreening tool. This SMART-on-FHIR tool automatically extracts deidentified patient (1) or one of two intervention arms (2 and 3). All patients are asked to allow one treatment
data (e.g., cancer type, stage, biomarkers) from the site EHR system and uses them to discussion with their oncologist to be video recorded for analysis. Prior to the recording,
query external matching services via the FHIR mCODE standard. Using the standard FHIR intervention patients utilize the DISCO App on an iPad. The DISCO App includes a video
ResearchStudy resource format, research coordinators review potential trial matches about treatment costs, ways to manage costs, and the importance of discussing costs with
returned within a specified radius of the practice for eligibility and discuss them with oncologists. Once patients enter their socio-demographic information (e.g., employment,
patients. At UTSW, this includes prostate, bladder, breast cancer and colon cancer insurance) and any financial concerns, they receive a tailored list of questions to ask their
cohorts. At TGH, the trial includes breast, prostate, and colorectal cancers, as well as oncologist. Arm 3 patients receive an intervention booster via email two months after the
recording. Patients complete measures at baseline, right after the recording, and at 1, 3, 6
glioblastoma and multiple myeloma. Statistical Methods:The primary endpoint is the
and 12 months after the recording. Measures assess outcomes including cost discussions,
proportion of patients enrolling in trials, comparing intervention and usual care arms.
communication quality, cost knowledge, self-efficacy for cost management, referrals for
Secondary objectives include evaluating usability, barriers to enrollment, and participant
support, short- and longer-term financial toxicity, and treatment adherence. The patient-
diversity. A total of 1200 patients (600 per arm) will be enrolled, with 81% power to
oncologist interaction is the unit of analysis and we will use multi-linear models to compare
detect a 75% relative increase in enrollment from 9.0% (control) to 15.8% (intervention)
outcomes by arm. We anticipate recruiting up to 15 oncologists and 240 patients. Data
at a one-sided alpha of 0.05. Trial Progress:The first phase of the trial addressed in-
collection began in March 2021 and will continue until July 2025. Participants to date
stitutional approvals, security compliance, and complex server and EHR integrations. The include 13 oncologists and 192 patients (116 Black, 76 White). Most patients completed
later phase addressed integration of the automated tool into diverse clinical workflows, the baseline assessment (n=164), the post-interaction assessment (n=137), and at least 1
engagement with site staff and participating patients. Both sites have enrolled ap- follow-up assessment (n=132); 125 treatment discussions have been recorded. The IRB
proximately 200 patients to date, half on the intervention arm. The final phase will reviewed the trial in December 2024 and approved continuation. Clinical trial information:
address primary and secondary outcomes as per trial design. Clinical trial information: NCT04766190. Research Sponsor: American Cancer Society; ACS RSG-20-026-01-CPHPS.
NCT05885880. Research Sponsor: American Cancer Society Cancer Action Network.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CARE DELIVERY/MODELS OF CARE 111s

TPS1660 Poster Session TPS1661 Poster Session

RACED (Reduction of Cervical Cancer Disparities): The impact of navigators Practical geriatric assessment (PGA) implementation strategies and cor-
and racial literacy training. First Author: Abna Faustina Sousa Vieira, Instituto do relative evaluations (PACE-70): A hybrid implementation-effectiveness
Câncer do Estado de S~ ao Paulo, Faculdade Medicina da USP, S~ ao Paulo, Brazil study in 3 community practices. First Author: Gabriel Aleixo, University of
Background: Cervical cancer is the third most prevalent cancer in Brazilian women. Pennsylvania, Philadelphia, PA
Approximately 17,000 new cases are expected for Brazil’s 2023-2025 triennium. The Background: The use of a geriatric assessment to inform oncologic care for older
complex multimodality treatment of locally advanced cervical cancer (LACC), which persons with cancer is an evidence-based practice that improves patient-clinician
relies on platinum-based chemoradiotherapy (CRT) and brachytherapy (BT), in addition communication, reduces treatment-related toxicity, and is recommended by national
due to the significant healthcare demands of patients with cervical cancer, creates guidelines. However, the implementation of a geriatric assessment can be time-
challenges for a public-funded health system. The Black population experiences the consuming and burdensome, leading to suboptimal use in clinical practice. Devel-
highest cancer mortality rates compared to the general population, partly due to in- oped and endorsed by the American Society for Clinical Oncology (ASCO), the Practical
equalities in social, economic, political, and health areas spheres. Data showed that, Geriatric Assessment (PGA) is designed to improve clinical usability and adoption, but
compared to White women, the mean age-adjusted mortality rates according to race/ its implementation in real-world settings has not been evaluated. The PACE-70 study
skin color were 27% higher in Black women. Around 60% of Black patients have a aims to evaluate PGA implementation and resultant chemotherapy dose modification
cervical cancer diagnosis in locally advanced or advanced stages. The incidence rate among older adults with advanced cancer treated in a community setting. An exploratory
among Black women was found to be significantly higher than that of their White aim will evaluate how the PGA, body composition and step count monitoring correlate
counterparts, with a relative risk of incidence nearly 50% higher. This disparity cannot be with chemotherapy toxicity and other clinical outcomes. Methods: The PACE-70 study
ignored. Methods: Our study is based on Critical Racial Praxis for Public Health. It is is a Type III hybrid implementation-effectiveness study enrolling at three community
inspired by the ACCURE (Accountability for Cancer Care Through Undoing Racism and sites within a large academic health system. Eligible participants will be 70 years or
Equity) initiative trial, composed of 3 anti-racist actions: 1- oncology navigation with older, have a diagnosis of advanced or metastatic solid malignancy, and be starting a
racial literacy, 2- real-time medical record alert system, and 3- race-specific feedback. new line of palliative-intent systemic therapy, where the expected prevalence of grade 3
Our intervention, in turn, consists of oncology navigation with racial inequities training toxicity exceeds 50 percent. The PGA will be administered via the electronic health
and improving interprofessional team knowledge about race and diversity through race- record (EHR), available for patients to complete independently prior to an initial medical
specific feedback. This prospective, single-center, non-randomized clinical trial of anti- oncology visit, or during the visit with staff assistance. Results from the PGA will be
racist actions and treatment support will compare prospective patients with a historical shared automatically with clinical teams via the EHR, including a Best Practice Alert
control from the same hospital. The primary endpoint is to increase the completion rate highlighting any identified geriatric impairment(s) and ASCO’s recommendation for PGA-
of definitive treatment with CRT+BT for 100 patients with IB2 to IVA cervical cancer adapted care. The primary outcome will be the PGA completion rate. The secondary
(convenience sample). The secondary endpoints are to analyze the implementation outcome will be the rate of chemotherapy dose modification among those with any
policy of this strategy and to make an economic assessment of the use of this identified geriatric impairment. Clinician perspectives on PGA implementation will be
implementation (we hypothesize that such measures reduce both visits to the emer- assessed via structured interviews among a sub-sample of participating clinicians. In a
gency room due to toxicity, as well as admissions to wards and ICU). Patient inclusion is subsample of patients consenting to additional data collection, exploratory analyses will
expected to begin in March 2025. Nurses are receiving training in oncology navigation examine correlations between the PGA, step counts (measured via FitBit) and body
and racial literacy in healthcare. As this is a race-conscious trial, the researchers plan to composition (measured via standard abdominal CT scans) with clinical outcomes,
prospectively compare outcomes between the intervention group’s Black/Brown and including toxicity, hospitalization, and death. PACE-70 will be the first study to report on
non-Black/Brown populations. In addition, given critical race theory, the research team real-world implementation of the PGA in a multisite community oncology setting. It will
comprises Black women in creation, design, and throughout the entire study continuum. provide insights on the facilitators and barriers of the PGA to inform chemotherapy dose
Clinical trial information: 85819325.0.0000.0068. Research Sponsor: Bristol Myers modification, as well as its potential predictive value for clinical outcomes. It will lay the
Squibb Foundation. foundation for larger trials of effectiveness seeking to encourage PGA implementation
and PGA-adapted care. Clinical trial information: pending. Research Sponsor: None.

TPS1662 Poster Session TPS1663 Poster Session

Patient Priorities Care for older breast cancer survivors: A patient-centered Impacting quality of life and pancreatic cancer survivorship through a
approach to improve quality of breast cancer survivorship. First Author: telehealth intervention. First Author: Vincent Chung, City of Hope, Duarte, CA
Dana Elena Giza, The University of Texas Health Science Center at Houston, Houston, TX Background: Pancreatic cancer patients experience significant debilitating symptoms
Background: Older adult breast cancer survivors have higher rates of chronic conditions as a direct or indirect result of disease, treatment, and co-morbidities resulting in higher
compared with other cancer survivors and may have a higher treatment burden. High symptom burden compared to other cancer types. The presence of comorbidities,
treatment burden is associated with poor quality of life and increased healthcare uti- declines in organ function, and increased need for assistance with daily function
lization. Aligning care with patient priorities can reduce the treatment burden during complicates the care of older adults with pancreatic cancer. Cancer not only affects the
survivorship care. Patient Priorities Care (PPC) is an approach designed to align care patient but also the entire family, especially the one that assumes the role of the family
around each patient’s goals to help decrease treatment burden. Although the PPC ap- caregiver (FCG). Robust evidence suggests that survivors with pancreatic cancer and
proach was previously designed for patients with multiple chronic conditions, we sought their FCGs experience high symptom burden and reduced quality-of-life (QOL). Despite
feedback from stakeholders, including older patients with a history of breast cancer, to robust evidence pointing to potential benefits of palliative care, many pancreatic cancer
adapt the PPC approach to breast cancer survivorship context. This quality improvement patients never receive any due to the workforce shortage. A scalable method of pro-
project aims to use PPC and patient’s self-defined goals to improve the quality of viding palliative care is needed. Methods: We are conducting a randomized pilot study
survivorship care for older adults breast cancer survivors. Methods: We are conducting a to determine the feasibility, acceptability, and preliminary efficacy of a centrally ad-
multicenter, randomized quality improvement project with a hybrid implementation- ministered telehealth, self-management survivorship care intervention in patients with
effectiveness design to evaluate the impact of the adapted PPC approach compared to pancreatic cancer and their FCGs. Patients within 8 weeks of initiating first line
standard survivorship care on treatment burden and quality of life (NCT06478589). We are treatment for metastatic disease are eligible. Prior to initiating the intervention sessions,
recruiting 120 older adult breast cancer survivors from outpatient oncology and primary an advanced practice nurse (APRN) will complete separate comprehensive survivor/FCG
care clinics. Eligible patients are: (1) $65 years of age; (2) stage I-III breast cancer, who QOL assessments using baseline surveys. The assessments will focus on QOL needs for
had finished active breast cancer treatment and are in the first year of survivorship care; survivors/FCGs and will include geriatric assessment for all regardless of age (activities
(3) have evidence of burdensome care; (4) English-speaking; and (5) able to provide of daily living, physical mobility, falls, social activity limitations, social support). Based
informed consent. Enrollment started in December 2024. Patients are stratified based on on the QOL and geriatric screenings, the Intervention APRN will complete a personalized
treatment burden at baseline; patients with a score of . 15 on the treatment burden care plan for the patient, and a personalized self-care plan for the FCG. This provides for
questionnaire (TBQ) undergo simple randomization with a 2:1 ratio to either the PPC or tailoring of the care plan to the participant’s needs and preferences and will be shared
standard survivorship care group. The PPC for Breast Cancer Survivorship intervention with each participant’s oncology care team after session completion. Both care plans
consists of: (1) a 30-minute priorities identification visit with a facilitator, 2) delivery of a will be organized around the four QOL aspects of care (physical, psychological, social,
structured patient priorities report to the survivorship care team, 3) survivorship care spiritual). Cultural aspects of care are taken into account and integrated appropriately
alignment using the patient’s priorities. The primary outcome measures are differences in within the four QOL aspects of care. A fully developed intervention resource manual with
treatment burden (TBQ) and quality of life (FACT-B) from baseline at 3 to 6 months. support reference materials and intervention content are provided to each patient and
Secondary outcomes include goal attainment achievement at 3,6 months for patients in FCG. The patient and FCG coaching sessions are bi-weekly, centrally administered and
the intervention group and adherence to standard and priorities-based breast cancer separate but parallel to allow participants to freely discuss their QOL needs. Clinical trial
survivorship recommendations at 12 months for both groups. Descriptive statistics will be information: NCT06524973. Research Sponsor: U.S. Department of Defense.
used to report patients’ baseline and clinical characteristics. All analysis will be intention
to treat. The effect of the intervention on changes (baseline to 3, 6 months) in the TBQ and
FAST-B scores using Bayesian analysis. For the goal attainment differences, we will
compare the differences in goal ratings from baseline at 3, 6 months within the Patient
Priorities Group. Clinical trial information: NCT04513977. Research Sponsor: NIA.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
112s CENTRAL NERVOUS SYSTEM TUMORS

LBA2000 Oral Abstract Session 2001 Oral Abstract Session

Efficacy and safety of STUPP regimen with or without anlotinib for newly A prognostic classification system for extent of resection in IDH-mutant
diagnosed glioblastoma: Results of a multicenter, double-blind, randomized grade 2 glioma: A report by the RANO resect group. First Author: Philipp
phase II trial. First Author: Yuanyuan Chen, State Key Laboratory of Oncology in South Karschnia, Department of Neurosurgery, Uniklinikum Erlangen, Friedrich-Alexander-
China, Sun Yat-Sen Univercity Cancer Center, United Laboratory of Frontier Radiotherapy University Erlangen-Nuremberg, Erlangen, Germany
Technology of Sun Yat-sen University and Chinese Academy of Sciences Ion Medical Background: The effects of resection in IDH-mutant grade 2 gliomas remain controversial
Technology Co., Ltd., Guangzhou, China since terminology for extent of resection was inconsistently applied across trials. We
aimed to (I) establish a standardized classification system for extent of resection and (II)
assess the impact of supramaximal resection on survival in IDH-mutant astrocytomas and
1p19q-codeleted oligodendrogliomas. Methods: Patients with newly diagnosed grade 2
IDH-mutant glioma meeting the WHO 2021 criteria were identified across sixteen centers in
the USA, Europe, and Asia as part of the RANO resect effort. Additional patients from UCSF
served for validation. Kaplan-Meier analyses and log-rank tests were applied to calculate
survival, and Cox’s proportional hazard regression model to adjust for multiple variables
(significance level: p # 0.05). Results: We identified 1391 newly diagnosed IDH-mutant
gliomas grade 2 between 1993-2024, of which 728 patients (379 astrocytoma, 349 oli-
godendroglioma) received no adjuvant treatment and allowed to study the effects of
resection. Smaller post-operative T2/FLAIR tumor remnants were favorably associated
The full, final text of this abstract will be available at with outcome. We classified those patients according to residual T2/FLAIR tumor volumes:
[Link] on the day of presentation and in the patients with ‘maximal T2/FLAIR resection’ (class 2; 0-5 cm3 remnant) had superior
online supplement to the June 10, 2025, issue of the Journal progression-free and overall survival compared to ‘submaximal T2/FLAIR resection’ (class
3; 5-25 cm3 remnant) or ‘minimal T2/FLAIR resection’ (class 4; .25 cm3 remnant), with 10-
of Clinical Oncology. year survival rates of 82.2% vs. 75.0% vs. 45.6% (respectively; p = 0.001). Resection of non-
infiltrated structures beyond T2/FLAIR borders provided an additional survival benefit as
characterized by a 10-year survival rate of 97.5%; thus defining class 1 ‘supramaximal T2/
FLAIR resection’ (HR for OS vs. class 2: 0.24, CI 0.1-0.5 / in astrocytoma: 0.26, CI 0.1-0.7 /
in oligodendroglioma: 0.21, CI 0.1-0.9). Effects of extensive resection on survival unfolded
after 3 years in astrocytomas, whereas survival curves separated after 6-8 years in oli-
godendrogliomas. The prognostic relevance of the four-tier classification was conserved
in a multivariate analysis controlling for clinical markers including pre-operative tumor and
1p19q-codeletion, in subgroups of either astrocytomas or oligodendrogliomas, and in a
separate cohort of 586 patients who received adjuvant chemo-/radiotherapy. The prog-
nostic value of the classification was further validated in the external UCSF cohort of 381
grade 2 IDH-mutant gliomas (p = 0.001). Conclusions: The proposed ‘RANO classification
for extent of resection’ serves as prognostic tool for patient stratification in grade 2 IDH-
mutant gliomas. While effects of extensive surgery are evident earlier in astrocytomas,
‘supramaximal’ resection translates into a survival benefit for both astrocytomas and
oligodendrogliomas and should be characterized in clinical trials. Research Sponsor: None.

2002 Oral Abstract Session 2003 Oral Abstract Session

Final clinical and molecular analysis of the EORTC randomized phase III A phase 2 study of pemigatinib for pre-treated glioblastoma or other gliomas
intergroup CATNON trial on concurrent and adjuvant temozolomide in an- with activating FGFR1-3 alterations: Results from FIGHT-209. First Author:
aplastic glioma without 1p/19q codeletion: NCT00626990. First Author: Enrico Franceschi, IRCCS Istituto delle Scienze Neurologiche di Bologna, Bologna, Italy
Martin J. Van Den Bent, Erasmus MC Cancer Institute, Rotterdam, Netherlands Background: FGFR genomic alterations occur in approximately 8% of gliomas. Inhibition of
Background: The 1st and 2nd interim analyses of the CATNON trial on anaplastic glioma FGFR1-3 with pemigatinib showed antitumor activity in a multihistology basket trial (FIGHT-
(NCT00626990) showed benefit from adjuvant (adj) temozolomide (TMZ) on overall survival 207) in which approximately 10% of participants (pts) had recurrent/progressive FGFR-
(OS) in patients with IDH mutant (mt) tumors, but no benefit of concurrent (conc) TMZ altered glioblastoma (GBM). We further investigated pemigatinib activity in primary brain
regardless of Isocitrate dehydrogenase 1 and 2 (IDH) mutation (mt) status. We now present tumors by conducting an international, multicenter, single-arm, 2-cohort, phase 2 study
the final analysis and the exploratory molecular marker analysis of the study. Methods: The specifically in adults with FGFR-altered pretreated gliomas (FIGHT-209; NCT05267106).
2x2 factorial design phase III CATNON trial randomized 751 adult patients with newly Methods: Pts were enrolled in 2 cohorts: A, histologically or molecularly defined GBM; or B,
diagnosed non-codeleted anaplastic glioma to either 59.4 Gy radiotherapy (RT) alone; the other gliomas, glioneuronal tumors, and neuronal tumors. Eligible pts had tumors
same RT with concTMZ; the same RT and 12 cycles of adjTMZ or the same RT with both harboring a FGFR1-3 fusion/rearrangement or mutation detected by an accredited labo-
concTMZ and adjTMZ. Methylation status including MGMT promoter methylation status ratory that had recurred/progressed after $1 prior therapy. Pemigatinib (oral, 13.5 mg on
were assessed with the Infinium MethylationEPIC Beadchip. IDH mutation (mt) status and days 1-14/21) was intended to continue until progression by Response Assessment in
glioma specific alterations were assessed with a glioma targeted panel using Agilent Neuro-Oncology (RANO) criteria determined by an independent review committee (IRC) or
SureSelect baits. Results: After a median follow-up of 10.9 years and with 499 events unacceptable toxicity. Efficacy of each cohort was evaluated independently. The primary
observed, in the intent-to-treat population the hazard ratio (HR) for OS adjusted for endpoint was objective response rate (ORR; partial plus complete) per RANO (cohort A),
stratification factors after concTMZ was 0.906 (95%CI 0.760, 1.082; p=0.28) and after with a goal of . 28%. Key secondary and exploratory endpoints were ORR in cohort B, ORR
adjTMZ 0.647 (95%CI 0.541, 0.773; p ,0.0001). In 660 patients IDH status could be de- by investigator assessment, progression-free survival (PFS) by IRC, overall survival (OS),
termined: IDH was mt in 444 tumors and wild type (wt) in 216 tumors. Median OS was 1.7 yrs safety, neurologic function by Neurologic Assessment in Neuro-Oncology (NANO), and
in patients with IDHwt tumors and 8.5 years in patients with IDHmt tumors. Benefit to TMZ efficacy correlations with diagnosis and specific FGFR-alterations. Results: Between May
was limited to patients with anaplastic glioma IDHmt of which 199 were still alive (45%). For 2022 and December 2023, 74 pts were enrolled in cohort A and 9 in cohort B. FGFR1-3
patients with IDHmt tumors the HR for concTMZ was 0.81 (95% CI 0.63-1.04; p=0.09) and for fusions/rearrangements were the most common genomic alterations in cohort A (n = 65
adjTMZ 0.54 (95% CI 0.42-0.69,p , 0.0001). No benefit was observed of concTMZ in IDHmt [88%]) and in cohort B, FGFR1 mutations (n = 8 [89%]). Pts had a median (range) age of 56
glioma patients that also received adjTMZ (HR 0.92 95% CI 0.63-1.36; p=0.69). In patients (20-79) years; 60% were male. On September 27, 2024 (data cutoff), 16 pts remained on
with IDHmt tumors that had received any TMZ median OS was 10.3 years, the median OS in treatment (cohort A, n = 11 [15%]; cohort B, n = 5 [56%]); 67 discontinued, primarily due to
patients treated with adjTMZ was 12.5 years (95% CI 9.4-15.0; p,0.0001). In exploratory progressive disease (n = 59 [71%]). In cohort A, ORR was 8% (6 partial responses [PR], 0
analysis, high-copy number Amplification of PDGFR and CDK4; Homozygous deletion of the complete responses [CR]); 21 pts (28%) had stable disease (SD); estimated 6-month PFS
CDKN2A/B locus, total copy number alterations, methylation subtype (A_IDH vs A_IDH_HG, rate was 17% (95% CI, 8.7-27.8) and 12-month OS rate 48% (95% CI, 35.6-60.2). In cohort B,
G-CIMP high versus low, MGMT-promoter methylation as determined by methylation arrays) the ORR was 22% (1 CR, 1 PR); 3 (33%) SD. Most treatment-emergent adverse events (AEs)
were all associated with outcome but none was predictive for benefit to TMZ. were low grade in severity (grade $3, 36.1%). Hyperphosphatemia, a class effect of FGFR
Conclusions: Despite more follow-up, concTMZ did not improve OS regardless of IDH inhibitors, was the most common AE (75%); 6 pts (7%) required dose reduction and 4 pts
status. AdjTMZ increased OS in patients with IDHmt tumors but not in patients with IDHwt (5%) discontinued due to AEs. Conclusions: ORR did not meet the pre-specified goal of .
tumors. Molecular factors of known prognostic significance for IDHmt 1p/19q intact an- 28% among pts with GBM harboring pemigatinib-sensitizing FGFR alterations. However,
aplastic glioma did not predict benefit to TMZ. Median OS in patients with IDHmt glioma durable disease stabilization was observed, notably in pts with CNS tumors other than GBM,
having received adjTMZ after RT was 12.5 years. Standard of post-operative care in patients and toxicities were manageable. More mature PFS and OS data will be presented with
with high grade IDHmt astrocytoma should be RT followed by 12 cycles adjTMZ. Funding exploratory molecular correlations. Clinical trial information: NCT05267106. Research
Source: MSD. Clinical trial information: NCT00626990. Research Sponsor: MSD; Dutch Sponsor: Incyte Corporation.
Cancer Socierty; 10685; Brain Tumor Charity; GN-000577; Stijd van Salland.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 113s

2004 Oral Abstract Session 2005 Oral Abstract Session

A phase II study of asandeutertinib (TY-9591) in advanced NSCLC patients Patritumab deruxtecan (HER3-DXd) in active brain metastases (BM) from
with EGFR-positive mutations and brain metastases. First Author: Pu-Yuan metastatic breast (mBC) and non–small cell lung cancers (aNSCLC), and
Xing, National Cancer Center/National Clinical Research Center for Cancer/Cancer leptomeningeal disease (LMD) from advanced solid tumors: Results from
Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing the TUXEDO-3 phase II trial. First Author: Matthias Preusser, Division of Oncology,
Key Laboratory of Clinical Study on Anticancer Molecular Targeted Drugs, Beijing, China Department of Internal Medicine I, Medical University of Vienna, Vienna, Austria
Background: Asandeutertinib (TY-9591), a deuterated osimertinib derivative, is a new central Background: BM and LMD are common and severe complications of solid cancers with high
nervous system-active 3rd generation epidermal growth factor receptor tyrosine kinase in- morbidity, poor prognosis, and limited treatment options. Antibody drug conjugates (ADCs)
hibitor (EGFR-TKI), that can potently and selectively inhibit EGFR- sensitizing mutations have shown high intracranial overall response rates (IC-ORR) in HER2-positive mBC and
(EGFRm+) and T790M resistance mutation. The phase I study for asandeutertinib EGFR-mutated NSCLC patients (pts). HER3-DXd, an ADC combining an anti-HER3 antibody
(NCT04204473) showed had a very superior clinical efficacy on the NSCLC with EGFR mu- with a topoisomerase (topo) I inhibitor, has shown promising results in mBC and aNSCLC
tations. This phase II study (NCT05146219) aimed to further evaluate the efficacy and safety of pts. Since HER3 is highly expressed in aNSCLC and mBC CNS metastases, we hypothesized
asandeutertinib in patients with locally advanced or metastatic EGFRm+ NSCLC with brain HER3-DXd may have clinical activity in BM from mBC and aNSCLC pts, and LMD from any
metastases (BM). Methods: 29 patients were enrolled and received asandeutertinib treatment solid tumor. Methods: TUXEDO-3 (NCT05865990) is an international, multicenter, multi-
at a dose of 160 mg once daily. The 27 patients with EGFR-sensitizing mutations (19 Del or cohort, single-arm, phase II trial enrolling pts with BM from mBC (cohort 1), aNSCLC (cohort
L858R) did not take any EGFR-TKI previously, while the 2 patients with EGFR T790M resistance 2), and LMD from any solid tumor (cohort 3). Key inclusion criteria were: Pts $18 years old,
mutation previously received 1st or 2nd-generation EGFR-TKIs therapy. The primary endpoints histologically documented disease, ECOG PS 0-2, and left ventricular ejection fraction $50%
were the intracranial objective response rate (iORR) assessed by investigator (INV) per RANO- in all cohorts; newly diagnosed/progressing BM with $1 brain lesions $10mm by MRI,
BM and the extracranial objective response rate (eORR) assessed by INV per the RECIST v1.1. and $1 line of prior systemic treatment, in cohorts 1 and 2; LMD per EANO-ESMO in cohort
Results: At the time of data cutoff at March 21,2024, the median follow-up time was 3. Pts received HER3-DXd 5.6 mg/kg IV Q3W until disease progression, unacceptable
16.4 months. The confirmed INV-iORR was 93.1% (95% CI: 77.2%-99.2%) (n = 29). The toxicity or withdrawal for any reason. Primary endpoint was IC-ORR per local investigator
confirmed INV-iORR for those who were treated with asandeutertinib as first line was 92.6% according to RANO-BM in cohorts 1 and 2, and 3-month OS in cohort 3. Sample size was
(95% CI：75.7%-99.1%) (n = 27). The 2 patients with previous EGFR-TKI therapy were in- based on Simon’s two-stage design. Primary endpoint was met if $3 IC responses (H0:
tracranial partial response (iPR). The median intracranial duration of response (iDoR) and
#5%; H1: $25%) in cohort 1 and 2; and if $3 pts with 3-month OS (H0: #5%; H1: $25%) in
intracranial progression-free survival (iPFS) were not reached. The 12-month iDoR was 82.8%,
cohort 3. Overall sample size was 60 pts with a target population of 20 pts per cohort.
and the 12-month iPFS was 96.6%. The median PFS was 13.5 months (95% CI: 12.5-NA) (n =
Results: Between December 2023 and July 2024, 61 evaluable pts were enrolled from 8
29) and 15.1 months (95%CI：12.5 - NA) (n = 27) for those without previous EGFR-TKI therapy.
Austrian and Spanish sites. Median age (min; max) was 57.0 (35.0; 75.0), 59.5 (37.0; 72.0)
Any intracranial or extracranial progression was evaluated as systemic progression, which may
lead to underestimation of the systemic PFS. The mean treatment was 402.9 days (n = 29). 27
and 51.5 (40.0; 66.0) years in cohorts 1, 2 and 3, respectively. At data cut-off, median follow-
(93.1%) patients experienced treatment-related adverse events (TRAEs). The most common up (min; max) was 4.4 (1.4; 10.1), 4.3 (0.2; 11.0) and 3.5 (0.8; 8.6) months in cohorts 1, 2 and
TRAEs ($10%) included decreased white blood cell count, decreased absolute neutrophil 3, respectively. Primary endpoints were met in all three cohorts. In cohort 1, 5/21 (23.8%) pts
count, decreased platelet count, elevated serum creatine phosphokinase, diarrhea, etc had IC response irrespective of BC subtype; 2 (40.0%) responders had received previous topo
(majority grade 1/2). Grade 3 TRAEs occurred in 27.6% patients while no grade 4/5 adverse I based ADCs. In cohort 2, 5/20 (25.0%) pts had IC response. In cohort 3, 11/20 (55.0%) pts
event. Six serious adverse events were reported by five patients (17.2%), of which two patients achieved 3-month OS irrespective of the LMD type. No new signals of toxicity were observed
(6.9%) were study drug-related. The interstitial lung disease, cardiomyopathy and keratitis and neurological symptoms, QoL and neurocognitive function remained stable or improved
were not reported. Conclusions: Asandeutertinib is highly effective and well-tolerated in over the treatment period. Tumoral HER3 expression did not correlate with treatment
locally advanced or metastatic EGFRm+ NSCLC patients with BM. Pivotal phase II study response. Conclusions: TUXEDO-3 is the first trial evaluating efficacy and safety of HER3-
(NCT05948813) and phase II trials (NCT05382728) are onging. Keywords: TY-9591; Deuterated DXd in pts with BM or LMD. HER3-DXd showed substantial CNS activity in parenchymal
osimertinib derivative; Brain metastases. Clinical trial information: NCT05146219. Research metastases and LMD, and may be a potential novel treatment for CNS disease in cancer pts.
Sponsor: None. Clinical trial information: NCT05865990. Research Sponsor: Daiichi Sankyo Company,
Limited; Merck Sharp & Dohme LLC, a subsidiary of Merck & Co.

2006 Oral Abstract Session 2007 Oral Abstract Session

A phase II study of an anti-telomerase CD4+ T-helper vaccine (UCPVax) with INB-200: Phase 1 study of gene-modified autologous gamma-delta (gd)
or without temozolomide in newly diagnosed glioblastoma. First Author: t cells in newly diagnosed glioblastoma multiforme (GBM) patients receiving
Antoine Carpentier, Hôpital Saint-Louis, Paris, France maintenance temozolomide (TMZ). First Author: Mina Lobbous, Cleveland Clinic,
Background: UCPVax is a therapeutic vaccine designed to stimulate CD4+ helper T cell Cleveland, OH
responses against telomerase (TERT), a protein highly expressed in glioblastoma (GBM). Background: Recent cell therapy and CAR-T initiatives for GBM have shown initial responses but du-
Temozolomide (TMZ), a standard chemotherapeutic agent in the treatment of GBM, has rability has been disappointing. We developed a novel approach to treat newly diagnosed GBM using
innate gd T cells following forced upregulation of tumor stress-associated targets. Methods: We
been shown to induce CD4+ T-cell lymphopenia, which could potentially impair the leveraged the TMZ-induced activation of the DNA damage response (DDR) pathway to transiently
immune response to the vaccine. We conducted a multicenter, 2-cohort, phase IIa study upregulate NKG2D-L targets on GBM. Co-administration of TMZ chemotherapy with gd T cells engineered
to evaluate the immunogenicity and efficacy of UCPVax, with or without TMZ, as for TMZ resistance by insertion of a methylguanine-DNA methyltransferase (MGMT)-expressing lenti-
adjuvant therapy in patients with newly diagnosed GBM following chemoradiation. vector (DeltEx Drug Resistant Immunotherapy – DRI) enables the targeting of residual GBM cells during
Methods: Patients with non-mutated IDH1 glioblastoma (GBM) were enrolled one the standard-of-care Stupp regimen. A total of 23 patients were enrolled, with 13 treated and (62% male;
month after completing concurrent radiotherapy and temozolomide (TMZ). Cohort A median age 66 (range: 21-74); 92% IDH-WT, 54% MGMT-unmethylated). Cohorts (C) 1, 2 and 3 received 1,
3 or 6 doses (1 x 107 DRI cells/dose) into the resection cavity with 150 mg/m2 of IV TMZ on Day (D) 1 of
received the vaccine alone, without additional TMZ, while Cohort B was treated with both
each Stupp regimen maintenance cycle. Results: No Dose limiting toxicities (DLTs) were seen nor were
the vaccine and six monthly cycles of TMZ. The primary endpoint was the induction of occurrences of cytokine release syndrome (CRS) or neurotoxicity (ICANS). Most common adverse events
TERT-specific CD4+ T cell responses, assessed ex vivo using the INF-g ELISpot assay. were related to underlying TMZ and Stupp regimen. As of January 24, 2025, median follow-up is
Secondary endpoints included epitope spreading, clinical outcomes, and safety. 16.9 months (m). The median PFS for patients is 8.3m for those who received a single dose of INB-200,
Results: Thirty-one GBM patients with unmethylated MGMT status were included in 9.9m for all patients (a 44% increase over the 6.9m mPFS of the Stupp) and 14.0m for patients who
cohort A, and 30 patients (50% with unmethylated MGMT status) were included in cohort received repeated doses, an 102.4% improvement over Stupp and 69% over single dose patients. A patient
with IDH mutant tumor remains progression free for almost 44 months and one with MGMT-unmethylated
B. The vaccine was well tolerated, with no vaccine-related serious adverse events.
tumor for 18 months. Biopsy specimens from three patients are available with general immune activation
Vaccine-expanded TERT-specific CD4+ T cells were detectable ex vivo in 25/30 (83%) of having been demonstrated. Conclusions: To date all patients had manageable toxicity with outpatient
patients in cohort A (no additional TMZ) and in 18/26 69% of patients in cohort B (treated treatment and a continued encouraging trend in longer PFS from treatment with DRI gd T cells. Clinical
with additional TMZ). Epitope spreading was induced in 29 out of 55 evaluable patients trial information: NCT04165941. Research Sponsor: IN8Bio, Inc.
(52.7%), corresponding to 15/26 (57.7%) in cohort A and 14/29 (48%) in cohort B. Median
TMZ Maint.
overall survival (OS) was significantly improved in patients who developed an epitope Age/ IDH/ Dose Cycles PFS
spread response compared to those who did not (19.3 vs. 12.8 months, P = 0.03). In the Subject Sex Methylation Resection level Received Response (mos) OS (mos)
44 patients with measurable disease at the time of inclusion, the radiological response 001 69/M IDH-WT, MGMT-unmethylated Total 1 5 SD 8.3 15.6
rate (RR) was 34%, including minor responses. In patients who developed epitope 003 75/F IDH-WT, MGMT-methylated Total 1 6 SD 11.9 17.7
004 21/F IDH-WT, MGMT-unmethylated Total 1 3 SD 7.4 9.6
spreading after vaccination (n = 22), the RR was 50%, compared to 18.7% in patients 007 75/M IDH-WT, MGMT-unmethylated Total 2 2 Un- - 5.1
without epitope spreading (P = 0.05). Furthermore, tumor-infiltrating lymphocytes evaluable
009 32/M IDH-mutant, MGMT- Total 2 12 SD 43.7+
against TERT were detected in 3 vaccinated patients who underwent surgery at re- methylated
currence. Conclusions: UCPVax demonstrated robust immunogenicity, even when co- 011 56/F IDH-WT, MGMT-methylated Total 2 6 SD 22.2 28.6
administered with TMZ, and was associated with improved overall survival (OS) in GBM 014 73/F IDH-WT, MGMT-unmethylated Subtotal 2 6 SD 8.7+ 8.7 without
progression
patients who developed an epitope spreading response. These findings support further 015 73/M IDH-WT, MGMT-methylated Subtotal 3 5 SD 7.1 11.8
clinical investigation of TERT-derived CD4+ helper vaccine in GBM patients. Clinical trial 017 74/F IDH-WT, MGMT-methylated Subtotal 3 3 SD 21.5+
020 66/M IDH-WT, MGMT-methylated Subtotal 3 3 SD 19.6+
information: NCT04280848. Research Sponsor: French Eastern Interregional Group of 021 57/M IDH-WT, MGMT-unmethylated Total 3 6 SD 18.1+
Clinical Research and Innovation (GIRCI-Est - APJ2017); Oligocyte. 022 53/M IDH-WT, MGMT-unmethylated Subtotal 3 6 SD 10.0 13.6
023 52/M IDH-WT, MGMT-unmethylated Subtotal 3 1 4.2 5.4

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
114s CENTRAL NERVOUS SYSTEM TUMORS

2008 Oral Abstract Session 2009 Clinical Science Symposium

Immunological correlates from phase I study of CARv3-TEAM-E in patients Multicenter trial of microbubble-enhanced transcranial focused ultrasound
with recurrent glioblastoma (GBM): INCIPIENT trial. First Author: Bryan D. Choi, (MB-FUS) with monthly adjuvant temozolomide for patients with high-grade
Massachusetts General Hospital, Boston, MA gliomas. First Author: Graeme Woodworth, University of Maryland School of Medicine,
Background: Chimeric Antigen Receptor (CAR) T cells for glioblastoma (GBM) have Baltimore, MD
been limited by the challenge of targeting a single tumor antigen in a heterogeneous Background: High-grade gliomas (HGGs) have few effective therapies targeting tumor cell recur-
disease. To address this barrier, we generated a novel engineered T-cell product (CARv3- rence, which remain shielded by blood-brain barrier (BBB). MB-FUS allows for controlled BBB opening
TEAM-E) that targets the EGFRvIII antigen while also secreting T-cell-Engaging Antibody (BBBO) enabling localized drug delivery and increased tumor biomarker release into systemic cir-
culation. Methods: MR-guided MB-FUS with real-time feedback was evaluated for HGG patients in
Molecules (TEAMs) against wild-type EGFR. Methods: The INCIPIENT clinical trial is a multicenter phase 1/2 trial (BT008: NCT03551249, NCT03616860) for adverse events (AEs) and
first-in-human study of CARv3-TEAM-E in patients with recurrent GBM (NCT05660369). feasibility [primary endpoints], efficacy [secondary endpoint], and plasma cell-free DNA (cfDNA) post-
Patients were treated with intraventricular CARv3-TEAM-E T cells (10E6 cells per in- procedure [exploratory]. After resection and 6 weeks of chemoradiation, peri-resectional infiltrative
fusion). A subset of patients were conditioned with lymphodepleting chemotherapy regions were targeted with MB-FUS during monthly adjuvant temozolomide cycles (MB-FUS+TMZ).
(LDC) consisting of cyclophosphamide and fludarabine. Immune cells were profiled in For efficacy, overall survival (OS) and progression-free survival (PFS) were compared with an external
the cerebrospinal fluid (CSF) and peripheral blood of patients by flow cytometry. cohort, created using restriction and coarsened exact matching (CEM). Results: Trial cohort had 34
Results: CAR T cells were detected in the CSF of all patients for an average of 33.6 days patients enrolled and evaluated from 5 sites in North America. No serious procedure-related AEs were
seen, with the most common AEs being mild, self-resolving. BBBO was seen in 100% of treatments,
(SD = 10.33). Granulocytes, NK cells, B cells, and monocytes appeared in the CSF
covering 82% targeted volumes with #3mm accuracy. Trial cohort had longer mPFS (univariate 13.5
immediately after infusion, decreasing to low levels over the course of several weeks. vs. 9.6 months, multivariate HR 0.62, 95%CI: 0.39-0.99, p=0.048) and mOS (36.4 vs. 19.1 months,
TEAM-positive T cells persisted in CSF until (median) day 33.6 (SD = 10.8) with a range of multivariate HR 0.50, 95%CI: 0.26-0.95, p=0.036), with treatment effect robust in sensitivity analyses.
21-56 days. CAR T cells were transiently detected in the peripheral blood of 9/10 patients Disease state correlated closely with longitudinal plasma cfDNA changes. Conclusions: MB-
at an average of 14 days (SD = 3.5) after infusion. Prior to infusion, CAR T cells were FUS+TMZ is a safe and feasible therapeutic approach for HGG, potentially improving survival and
predominantly CD4-positive and remained as such in the CSF over time. Those in the enabling longitudinal non-invasive monitoring. Clinical trial information: NCT03551249,
periphery exhibited CD4-to-CD8 polarization. Of patients who received multiple infu- NCT03616860. Research Sponsor: Insightec Inc; U.S. National Institutes of Health; R21NS113016.
sions, 3 out of 6 had CAR-positive T cells in the CSF after a second infusion, although Variables Trial cohort Matched Cohort †
their persistence was short-lived and was not detected in the periphery following repeat Patients (N) 34 158
infusions. LDC increased engraftment of CAR T cells in CSF but not in peripheral blood. Baseline characteristics used in CEM SMD†
Age, years, mean6SD 51.5 613.0 51.6 613.0 0.0
Patients with poor CAR persistence demonstrated the development of anti-CARv3- MGMT, Unmethylated, N (%) 16 (47.1%) 74 (47.1%) 0.0
TEAM-E antibodies in the CSF and serum, which increased with reinfusion. IDH, Wild type, N (%) 29 (85.3%) 135 (85.3%) 0.0
Characteristics tackled through restriction SMD†
Conclusions: Following initial infusion, intraventricularly delivered CARv3-TEAM-E Received resection & 6 weeks of chemoradiotherapy 34 (100%) 158 (100%) 0.0
T cells were detected in the CSF and peripheral blood in patients with recurrent GBM. Non-Hispanic, N (%)
Complete resection, N (%)
34
34
(100%)
(100%)
158
158
(100%)
(100%)
0.0
0.0
Reduced persistence was observed with subsequent infusions. This corresponded with KPS ‡70 – N (%) 34 (100%) 158 (100%) 0.0
the emergence of anti-CARv3-TEAM-E antibodies in treated patients. Clinical trial in- Other characteristics not used in CEM wSMD†
Sex, male – N (%) 16 (47.1%) 101 (63.9%) 0.34
formation: NCT05660369. Research Sponsor: Gateway for Cancer Research; National Race, White – N (%) 28 (82.4%) 137 (86.7%) 0.13
Preoperative tumor size, median cm3 (IQR) 19.8 (6.9, 42.9) 47.0 [30.0, 53.0) 0.61
Cancer Institute/U.S. National Institutes of Health; 1R01CA294071-01A1. Clinical Outcomes P
Unadjusted mOS, months (95%CI) 36.4 (21.1, NR) 19.1 (16.2, 22.8) ,0.001
OS HR for treatment adjusted for tumor size, IDH, & 0.50 (0.26, 0.95) 0.036
MGMT (95%CI)
Unadjusted mPFS, months (95% CI) 13.5 (9.9, 25.4) 9.6 (7.8, 11.9) 0.032
PFS HR for treatment adjusted for tumor size, IDH, & 0.62 (0.39, 0.99) 0.048
MGMT (95%CI)

SMD, weighted standardized mean difference.

†
Estimated using CEM weights.

2010 Clinical Science Symposium 2011 Clinical Science Symposium

Leveraging stimulated Raman histology-based cellularity for random forest Stereotactic radiation versus hippocampal avoidance whole brain radiation
prediction of glioblastoma recurrence. First Author: Sanjeev Herr, Drexel Uni- in patients with 5-20 brain metastases: A multicenter, phase 3 randomized
versity College of Medicine, Philadelphia, PA trial. First Author: Ayal Aizer, Brigham and Women’s Hospital/Dana-Farber Cancer
Background: Glioblastoma is a universally fatal diagnosis with extent of resection being Institute, Boston, MA
one of the most significant predictors of overall and progression-free survival. Most Background: Radiation therapy forms the mainstay of management for patients with brain
patients eventually experience recurrence, with sixty percent recurring along the re- metastases. Published randomized trials have found improved quality of life with ste-
section cavity. Recent work leveraging Stimulated Raman histology (SRH) and artificial reotactic radiation (SRS/SRT) over whole brain radiation (WBRT) in patients with # 4 brain
intelligence (AI) has approximated glioma cellularity within the infiltrative margins. It metastases; comparative trials in patients with .4 brain metastases are lacking. In addition,
remains unknown if these estimates of glioma burden at the infiltrative margins in- prior randomized trials have demonstrated the superiority of hippocampal avoidance WBRT
fluence glioblastoma recurrence. This study aims to evaluate a predictive model of focal (HA-WBRT) over traditional WBRT, but no study has compared SRS/SRT to HA-WBRT.
recurrence in patients with glioblastoma using SRH and AI-generated cellularity scores Accordingly, we conducted a multicenter, phase 3 randomized trial comparing SRS/SRT to
from tissue samples taken at the resection cavity margins. Methods: A multi-center, HA-WBRT in patients with 5-20 brain metastases. Methods: Eligible patients were age 18-
retrospective cohort study was conducted on patients diagnosed with glioblastoma who 80 with 5-20 brain metastases secondary to a solid primary other than small cell lung cancer,
underwent resection followed by spatial annotated tissues acquired from the resection were naı̈ve to prior brain-directed radiation, and lacked leptomeningeal disease. The primary
cavity margins. Tissues were analyzed using SRH optical imaging, and histopathology endpoint was the average of patient-reported symptom severity and interference over the
analysis was performed using confocal microscopy. Tissue cellularity was measured first six months post-baseline relative to baseline, using the MD Anderson Symptom
Inventory–Brain Tumor (MDASI-BT) module, a validated instrument assessing 22 symptoms
histologically and by optical imaging. Results: Over 400 patients and 2,200 specimens
and 6 interference measures integral to quality of life, each scored 0-10 with higher scores
were analyzed, of which a nested subset of 60 patients were selected based on selection
indicating greater symptomatology/interference in function. The target effect size was a
criteria. Using preoperative and postoperative imaging, margin samples were deter- symptom severity of 0.70, corresponding to 50% of the observed difference between
mined to be in an area of recurrence (n=58) or nonrecurrence (n=220). Cellularity was patients with a good (90-100) versus poor (#80) Karnofsky performance status; with 80%
significantly higher in the recurrent margin sample group when compared to the power and a two-sided alpha of 0.05, 196 patients were required. Results: Between 4/2017-
nonrecurrent group (p = 0.026), which was further confirmed by a pathologist- 5/2024, 196 patients enrolled, 98 in each arm. The median number of brain metastases was
determined cellularity score (0-3) that demonstrated similar findings (p = 0.026). Re- 14 (IQR 11-18); 25% of patients underwent prior neurosurgical resection. Baseline mean
sults were validated across three medical centers. Six classifiers were then trained for MDASI-BT symptom severity scores were 2.2 (SRS/SRT arm) and 1.9 (HA-WBRT arm),
recurrence prediction. Using nineteen of the most predictive variables, random forests p=0.20; respective interference scores were 3.5 and 3.2 (p=0.40). The average of weighted
(RF) performed best with an AUC of 0.848. RF screening for the minimum practical post-baseline severity and interference scores relative to baseline indicated lower
number of variables demonstrated an AUC of 0.805 using only FastGlioma, age and symptomatology/inference in the SRS/SRT arm, meeting the primary endpoint of the study
extent of resection as variables. Conclusions: AI-generated cellularity scores have the (difference between SRS/SRT and HA-WBRT: -1.06, p,0.001). Averaged post-baseline
potential to predict focal recurrence of glioblastoma, allowing for more tailored ap- symptom severity scores minus baseline were -0.03 and 0.59 in the SRS/SRT and HA-WBRT
proaches to surgical resection and radiotherapy to increase progression-free survival. arms, respectively (difference -0.62, with lower symptom severity in the SRS/SRT arm,
Research Sponsor: None. p,0.001); respective interference estimates were -0.62 and 0.89 (difference -1.50, with
lower interference in the SRS/SRT arm, p,0.001). Median survival was 8.3 and 8.5 months
in the SRS/SRT and HA-WBRT arms, respectively (p=0.30). Conclusions: This phase 3
randomized trial indicates that patients with 5-20 brain metastases experience fewer
symptoms and less interference in function after SRS/SRT as opposed to HA-WBRT, without
compromise of survival, supporting SRS/SRT as the standard of care in this population.
Clinical trial information: NCT03075072. Research Sponsor: Varian.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 115s

2012 Rapid Oral Abstract Session 2013 Rapid Oral Abstract Session
Utidelone in combination with etoposide and bevacizumab in HER2- Phase 1 study of HMPL-306, an inhibitor of mutant IDH1/IDH2 (mIDH1/2), in
negative breast cancer patients with brain metastasis: A prospective, western patients (pts) with advanced mIDH solid tumor, including glioma.
single-arm, phase II trial. First Author: Yehui Shi, Tianjin Medical University Can- First Author: Jordi Rodon Ahnert, The University of Texas MD Anderson Cancer Center,
cer Institute and Hospital, Tianjin, China Houston, TX
Background: For advanced HER2 negative breast cancer patients with brain metastasis, Background: Isocitrate dehydrogenase (IDH) 1 or IDH2 mutations or co-mutations have
systematic therapy has failed to yield satisfied efficacy, although bevacizumab and been associated with various tumors, including glioma. HMPL-306 (’306) is a novel,
etoposide have shown some effectiveness as mono- or combination therapy. Novel small-molecule, orally available, highly selective, and potent dual inhibitor of both mIDH1
microtubule inhibitor utidelone demonstrated good efficacy in advanced breast cancer and mIDH2. This is a phase 1 study of ’306 in pts with locally advanced or metastatic
patients in several clinical trials, and was also suggested a capability of blood-brain solid tumors with mIDH. Here, we report the results of the dose escalation stage.
barrier penetration. Therefore, utidelone in combination with bevacizumab and eto- Methods: Pts with locally advanced or metastatic solid tumors with any mIDH were
poside would be a promising regimen for HER2 negative breast cancer patients with enrolled to receive ’306 once daily (QD) for 28-day cycles. The mTPI-2 design was used
brain metastasis. Methods: Breast cancer patients with brain metastasis were enrolled for dose escalation, having explored in 8 successive cohorts (50-400 mg). The study
and Simon’s two-stage optimal trial design was used for this trial. If more than 3 out of aims to determine the maximum tolerated dose (MTD)/recommended phase 2 dose
13 patients showed central nervous system (CNS) response, 30 more patients would be (RP2D), evaluate safety, tolerability, preliminary efficacy and pharmacokinetics/
further enrolled. The acceptable ORR was set to be 40% for the trial. Utidelone (30mg/ pharmacodynamics (PK/PD). Results: As of Aug 9, 2024, 42 pts were administered ’306
m2/day, iv, d1-5), and etoposide (100mg/m2, iv, d1-3) were concurrently administered across 8 doses (n = 3, 3, 5, 12, 6, 4, 4, 5 in 50, 100, 150, 200, 250, 300, 350, 400 mg QD
with bevacizumab (10mg/kg iv, d1) every 21 days for 6 cycles, followed by maintenance cohorts, respectively), with 17 (40.5%) lower-grade glioma (LGG, grade 2 and grade 3
treatment with utidelone and bevacizumab until disease progression or unacceptable glioma) pts, 3 (7.1%) grade 4 glioma pts and 22 (52.4%) non glioma pts. The median age
toxicity. The primary endpoint is CNS-ORR. Secondary endpoints include CNS-clinical was 55 years, and 25 (59.5%) pts were male. During the dose escalation from 50 mg to
benefit rate (CNS-CBR), CNS-PFS, PFS, and safety. Results: 34 female HER2 negative 400 mg QD cohort, 1 pt given 250 mg QD experienced a dose-limiting toxicity (DLT) of
patients were enrolled, including 11 triple negative breast cancer patients and 23 grade 3 lipase increased. MTD was not reached. 12 (28.6%) pts reported grade $3
patients of luminal subtype, with a median age of 52 years (range 34-74) and a median adverse events (AEs), which reported in $ 2 pts was abdominal pain. Efficacy signals
treatment lines of three. Five patients had prior brain radiotherapy and 2 patients were were observed especially in LGG pts, in the efficacy evaluated set (N = 14), objective
previously treated with brain surgery. As of December 2, 2024, the median follow up response rate (ORR) was 7.1%, disease control rate was 100%; in the safety analysis set
duration was 11.5 months. The CNS-ORR was 67.6% (23/34), and the CNS-CBR was (N = 17), median progression-free survival (PFS) was 20.5 months (95% confidence
88.2% (30/34). The median PFS was 6 months (95% confidence interval [CI], 4.265- interval [CI]; 5.5-not estimable). One grade 2 glioma pt with multiple previous treatment
7.735). The median CNS-PFS was 15 months (95% CI, 6.760-23.240), with supportive on the 200 mg QD achieved minor response lasting 16.8 months. The ORR of grade 4
treatment for some of the patients after extracranial progression which included glioma pts and non glioma pts were not reached, the disease control rate were 33.3% and
etoposide re-administration, or abraxane, endocrine therapy, radiotherapy and immu- 25%, respectively. Drug exposures were dose-proportional from 50 mg to 400 mg.
notherapy. The major AE was peripheral neuropathy with 8.8% (3/34) of Grade 3, Steady-state with ~5-fold accumulation was reached after ~28 days of repeated daily
primarily classified as sensory. Most of the treatment-related AEs were grade 1 or 2 and dosing. In non-glioma pts, 2-HG inhibition plateaued after ~28 days, increasing with
were considered manageable and reversible. Conclusions: Utidelone in combination dose, reaching ~90% at $150 mg at C2D1. Conclusions: ’306 was well-tolerated in pts
with etoposide and bevacizumab has shown promising anti-tumor activity and man- with mIDH1/2 solid tumors, showing target inhibition and durable responses in LGG.
ageable toxicity in HER2 negative breast cancer patients with brain metastasis, and a Clinical trial information: NCT04762602. Research Sponsor: HUTCHMED Limited.
randomized control trial is warrantied. Clinical trial information: NCT05781633.
Research Sponsor: Tianjin Science and Technology Funding; 18ZXXYSY00070; Tianjin
Municipal Education Commission Funding; 2016YD03; Beijing Biostar Pharmaceuticals;
“358”Project, TJMUCH; 358-2023-06.

2014 Rapid Oral Abstract Session 2015 Rapid Oral Abstract Session
Vaccination by homologous antigenic loading with DOC1021 as adjuvant Results from phase 1 study of mycophenolate mofetil with chemoradiation
therapy for glioblastoma: Phase I clinical trial results. First Author: Joseph in newly diagnosed glioblastoma to target de-novo purine metabolism to
Georges, Banner University Medical Center, Phoenix, AZ overcome treatment resistance. First Author: Yoshie Umemura, Ivy Brain Tumor
Background: Glioblastoma is a devastating tumor for which median overall survival Center at Barrow Neurological Institute, Phoenix, AZ
(mOS) remains 14-18 months despite aggressive standard of care (SOC) treatment. Background: Mycophenolate mofetil (MMF) inhibits IMPDH and disrupts de novo purine
Clinical studies of dendritic cell (DC) vaccination for GBM have shown promise but have synthesis which is preferred by glioblastoma (GBM) whilst normal brain prefers resource ef-
been largely inconclusive. DC homologous antigenic loading leverages p38MAPK and ficient salvage pathway. A phase 0 study demonstrated the active drug metabolite reaching both
mTORC1 signaling cascades to initiate cDC1-like skewing of monocyte-derived DC, enhancing and non-enhancing GBM tissues in humans, and effective target engagement, noted
leading to potent downstream induction of tissue-homing cytolytic memory effector by reduced GTP/IMP ratio. This phase 1 trial assessed the tolerability of MMF with chemo-
T cells. Here we report results of a completed phase I study for glioblastoma (IDH-wt). radiation in newly diagnosed GBM patients (NCT04477200). Methods: Thirty adult patients with
Methods: This clinical trial evaluated autologous DC vaccine DOC1021 prepared from newly diagnosed GBM were given MMF, dosed BID, 1 week prior to and concurrently with
standard of care (SOC) radiotherapy (RT) of 60 Gy in 30 fractions with concomitant temo-
mobilized peripheral blood mononuclear cells (PBMC), loaded with autologous tumor
zolomide (TMZ) 75mg/m2, followed by MMF 1 day before + 5 days of each SOC TMZ 150-200mg/
lysate and amplified tumor mRNA, and administered bilaterally near deep cervical lymph m2 x 5/28-day cycle up 12 cycles. Optune was optional. Primary endpoint was dose limiting
nodes. Three courses of vaccine every 2 weeks plus weekly peg-IFN were administered toxicity (DLT) and maximally tolerated dose (MTD) of MMF combined with SOC GBM che-
after completion of chemoradiation. Four dose levels from 3.5 x 106 to 3.6 x 107 total moradiation. Time-to-event continual reassessment method was used to determine MMF
vaccine cells were tested. Patients with subtotal resection or tumor progression prior to dosing, with MTD defined as estimated rate of dose-limiting toxicity (DLT) closest to but not
vaccination were not excluded. Results: Sixteen newly diagnosed patients completed exceeding 30%. DLT periods were during and up to 4 weeks after concurrent chemoradiation
treatment, median age 61 years (range 47-73), 94% MGMT unmethylated, 25% subtotal (DLT1), and first two 28-day cycles of MMF with temozolomide (DLT2). Transient grade 4
resected. OS at 12-months was 88% compared to expected ~60% for SOC and 5 patients neutropenia x , 7 days and asymptomatic grade 4 lymphopenia were excluded from DLT.
are still alive at 19-30 months of follow-up. Two recurrent glioblastoma patients were also Kaplan Meier method was used to estimate overall survival (OS). Results: The median age was
treated and survived for 10-12 months. Most common AEs were mild flu-like symptoms 57 (range 20-75). The majority had KPS . 80 (67%) at baseline, and unmethylated MGMT (70%).
and injection-site reactions, and there were no dose limiting toxicities. Analysis of post- During DLT1 period, 5 DLT1 was noted out of 16 subjects on 2000mg BID (grade 3 hemiparesis,
vaccination PBMC indicated expansion of CD4+ (13/13 patients) and CD8+ (11/13) central cognitive disturbance, fatigue, and grade 4 thrombocytopenia x2), and none at 1500mg (N = 10)
memory T-cell compartments (p , 0.00006 and p , 0.003, respectively) as well as and 1000mg (N = 4). During DLT2 period, 1/6 subjects at 1500mg BID experienced DLT of grade 3
expansion of CD8+CD127+ MPECs (12/13; p , 0.002). Among 3/3 patients analyzed by fatigue, and none at 1000mg (N = 4) and 2000mg (N = 16). All DLTs were reversible. Four patients
spatial transcriptomics, intense CD25+ foci correlating with co-expression of effector did not receive MMF during DLT2 period due to withdrawal from the study (N = 2) and pro-
gression of disease (N = 2). The most common treatment related adverse events were fatigue
memory T-cell and migratory microglial markers were observed in post-vaccination but
(77%), leukopenia (67%), and nausea (53%). Of the dose levels studied, the MTD for DLT1 and
not pre-vaccination samples. For 8 patients who were observed rather than re-operated
DLT2 were both 2000mg BID (posterior probability of DLT1: 18.5%, posterior probability of DLT2:
for worsening T1-weighted signal on MRI in the 23 weeks after vaccination, signal 7.5%), however, due to frequent fatigue and nausea, DLT1 period starting dose was lowered to
gradually resolved and GBM-specific mOS is not yet reached compared to 15.1 months for 1500mg BID for the last 7 subjects. The recommended phase 2 dose is 1500mg BID combined
8 patients who received reoperation despite comparable clinical characteristics, sug- with concurrent RT+TMZ followed by TMZ. Median OS was 16.8 months with 25.5 months
gesting an immune-reactive microenvironment manifesting as pseudo-progression. median follow up duration (NR & 25.5 months in MGMT methylated, 14.2 & 24.9 months in
Conclusions: DOC1021 combined with SOC is safe and potentially efficacious in this MGMT unmethylated respectively). Conclusions: MMF can penetrate enhancing and non-
challenging population that included subtotal resections, pre-treatment progression and enhancing GBM with evidence of successful inhibition of de-novo purine synthesis in
15/16 MGMT unmethylated. A randomized phase II trial is being launched including humans, and is reasonably well tolerated when combined with chemoradiation newly diagnosed
criteria to avoid early re-operation for enhancing T1-weighted signal that may be pseudo- GBM patients. These promising results have led to a planned phase 2/3 randomized controlled
progression. Clinical trial information: NCT04552886. Research Sponsor: Cancer Cures 4 trial through Alliance for Clinical Trials in Oncology. Clinical trial information: NCT04477200.
Kids; N/A; Diakonos Oncology; N/A. Research Sponsor: Gateway for Cancer Research.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
116s CENTRAL NERVOUS SYSTEM TUMORS

2016 Rapid Oral Abstract Session 2017 Rapid Oral Abstract Session
Use of lucicebtide (ST101) in glioblastoma patients by antagonism of Safety and tolerability of intraventricular CARv3-TEAM-E T cells following
C/EBPb-dependent mesenchymal cell transition and immunosuppressive lymphodepleting chemotherapy in recurrent glioblastoma: INCIPIENT trial.
M2 macrophage polarization. First Author: Fabio Massaiti Iwamoto, Columbia First Author: Elizabeth R. Gerstner, Massachusetts General Hospital, Boston, MA
University Irving Medical Center, New York, NY Background: CAR T therapy is a novel, promising approach in glioblastoma (GBM) but
Background: C/EBPb is a master regulator of the mesenchymal phenotype in GBM and has tumor heterogeneity can limit efficacy when a single antigen is targeted. We designed a
an essential role in the maintenance of immunosuppressive M2 tumor-associated mac- second-generation CAR T molecule that targets epidermal growth factor receptor vIII
rophages (TAMs). Lucicebtide is a first-in-class antagonist of C/EBPb that has shown direct (EGFRvIII) and also secretes a T-cell–engaging antibody molecule (TEAM) against wild-
anti-tumor activity in GBM as well as the ability to reprogram TAMs in the TME toward type EGFR. Methods: In a phase 1, first-in-human study (INCIPIENT, NCT05660369),
immunostimulatory M1 macrophages. In a recent recurrent GBM (rGBM) P2 study, luci- patients with recurrent GBM with EGFRvIII mutation and/or EGFR amplification were
cebtide was well-tolerated and resulted in disease control in 9/30 patients, including two eligible to receive up to 6 intraventricular doses of 10x106 CAR T cells via Ommaya
PRs lasting . 1 year. With strong rationale for targeting C/EBPb in GBM, additional cohorts catheter after lymphodepleting chemotherapy (LDC) with fludarabine and cyclophos-
we explored in a window-of-opportunity (WoO) study (NCT04478279). Methods: The WoO phamide. Primary objective was safety and tolerability and secondary objective was
study enrolled 2 cohorts; 9 pts with rGBM that received 2-4 doses of lucicebtide 500mg QW preliminary tumor response determined by iRANO criteria. Results: CAR T
prior to surgery and resumed lucicebtide after surgery to progression and 9 ndGBM pts that manufacturing was successful for all patients. Seven patients (5 male) received at least
received 2-3 doses of lucicebtide 500mg QW prior to surgery and resumed lucicebtide + 1 intraventricular infusion. Two patients received 2 infusions (1 for progressive disease
chemoradiation after surgery until progression. Endpoints include efficacy parameters of
(PD) and 1 without PD). One patient received 3 infusions after experiencing initial PD. No
PFS and OS, safety as a single agent and in combination with chemoradiation, and
DLTs occurred. All patients experienced cytokine release syndrome (CRS) grade 1 lasting
pharmacodynamic analyses including spatial transcriptomics and TME characterization.
0-9 days with only 1 patient experiencing CRS grade 2 for 1 day. One patient experienced
Results: Lucicebtide was well-tolerated as a single agent and in combination with che-
moradiation. Tissue analysis indicates penetration past the BBB and tumor uptake, as well
ICANS grade 1 that lasted 2 days. All patients experienced tumor inflammation-
as C/EBPb target engagement. Lucicebtide + chemoradiation in ndGBM extended PFS associated neurotoxicity grade 1 with a duration of 2-9 days. Adverse events (grade
beyond historic benchmarks, with the majority of patients remaining on study without 3-4) at least possibly related to CAR T were febrile neutropenia (N = 1) and neutrophil
progression (7-22+ months). As of January 25, 2025, mOS could not be evaluated, with 8/9 count decrease (N = 1). Toxicity was managed with supportive care without need for ICU
patients alive. In rGBM, lucicebtide improved mPFS to 3.4 months and mOS to at least monitoring and 3 patients received at least 1 dose of anakinra (max duration = 4 days,
11.8 months, exceeding historical data with chemotherapy (historic mPFS ~ 2 months and median = 1 day). Best response was stable disease (SD) in 5 patients with 1 patient
mOS 5.6-9.8 months). Pathologic evidence of treatment effect, i.e. geographic necrosis, was achieving SD for 6 months after a single infusion and another experiencing a 33%
observed in 5/6 pts including otherwise treatment naı̈ve ndGBM patients. Spatial tran- decrease in tumor diameter after 2 infusions. All patients are alive 3-8 months after first
scriptomics analysis revealed a significant reduction in the mesenchymal gene signature infusion. From the preceding safety run-in arm of the study (without LDC), one patient
following lucicebtide, consistent with on-target antagonism of C/EBPb. Further, immune survived 12 months and another is still alive . 20 months after infusion.
activation in the TME, as indicated by increased M1/M2 ratio and CD8+ T cell infiltration, was Conclusions: Intraventricular CARv3-TEAM-E infusions were well tolerated, even with
associated with disease control. Conclusions: Lucicebtide is well-tolerated as monotherapy multiple doses, and no DLTs were noted. Toxicity was manageable in all patients with
and in combination with SoC. Improvements in PFS and OS in GBM patients following supportive care and anakinra was administered to 3 patients. Steroids were not required
lucicebtide exposure demonstrated penetration across the BBB and target engagement, to manage toxicity. A subset of patients experienced SD for several months. Clinical trial
resulting in on-target pharmacodynamic activity including a dramatic reduction in mes- information: NCT05660369. Research Sponsor: Gateway for Cancer Research; National
enchymal gene signature in tumor cells and a remodeling towards a more permissive Gene Vector Biorepository at Indiana University, which is funded under National Cancer
immune TME. These data provide the mechanistic rationale for continued clinical evaluation Institute contract HSN261201500003I Task Order No. HHSN26100077; Philanthropic
of lucicebtide as a novel approach for patients with GBM. Clinical trial information: support to the Cellular Therapy Program at MGH.
NCT04478279. Research Sponsor: None.

2018 Rapid Oral Abstract Session 2019 Rapid Oral Abstract Session
A phase 1 study of B7H3 CAR-T cells administered intracranially in recurrent Tirabrutinib for the treatment of relapsed or refractory primary central
glioblastoma. First Author: Gordon Li, Stanford University, Stanford, CA nervous system lymphoma: Efficacy and safety from the phase II PROS-
Background: Glioblastoma (GBM) is an aggressive malignancy with median survival of PECT study. First Author: Lakshmi Nayak, Dana-Farber Cancer Institute, Boston, MA
approximately 2 years from initial diagnosis and 9 months after first progression. Effective Background: Primary central nervous system lymphoma (PCNSL) is a rare, aggressive
treatments in the recurrent setting following upfront chemoradiation and adjuvant form of non-Hodgkin lymphoma localized to the brain, cerebrospinal fluid, or eyes. For
temozolomide are limited. The transmembrane glycoprotein B7H3 is over-expressed in patients with PCNSL, treatment options are limited, standard of care is not well estab-
GBM and chimeric antigen receptor T cells targeting B7H3 (B7H3-CART) have shown lished, and prognosis is poor, particularly in the relapsed or refractory (r/r) setting.
activity in several preclinical cancer models. Intracranial delivery of B7H3-CART may Tirabrutinib, a highly potent selective second-generation Bruton’s tyrosine kinase inhibitor,
optimize targeting the immune response to the tumor microenvironment while limiting is approved in Japan, Taiwan, and South Korea based on a phase I/II study that dem-
systemic toxicity. Methods: We conducted a single-arm phase 1 study in patients with onstrated clinical activity in Japanese patients with r/r PCNSL. There are no currently
recurrent GBM undergoing repeat resection. B7H3-CART was administered via intra- approved drug therapies for PCNSL in the US or Europe. Here we report results from the
tumoral and intraventricular Ommaya reservoirs monthly for a planned 6 months or until PROSPECT study (NCT04947319) conducted in the US. Methods: In this open-label phase
confirmed disease progression. When possible per investigator discretion, the dose was II study, patients with r/r PCNSL received oral tirabrutinib 480 mg as monotherapy once
divided evenly between the two reservoirs. The primary endpoints were safety and daily until disease progression or unacceptable toxicity. The primary endpoint was overall
manufacturing feasibility, with secondary endpoints focused on preliminary efficacy. Dose response rate (ORR) assessed by Independent Review Committee. Secondary endpoints
escalation was planned according to a standard 3+3 design (dose level 1: 10x106 cells; included duration of response (DOR), time to response (TTR), best overall response (BOR),
level 4 (max): 100x106). Adverse events within 28 days of first dose and at least possibly and safety. Overall survival (OS) and progression-free survival (PFS) were exploratory
related to B7H3-CART were considered dose-limiting toxicities (DLTs) if meeting additional endpoints. Results: Forty-eight patients were enrolled. Median age was 65.5 y (range, 34-
criteria: any grade 5 toxicity, grade 4 cytokine release syndrome, neutropenia, or 87). With a median follow-up of 11.2 mo as of November 1, 2024 (data cut-off), ORR was
thrombocytopenia lasting . 14 days, or any non-hematologic grade 3 toxicity lasting . 72 66.7% (n = 32), with a complete response rate (CRR), confirmed (CR) + unconfirmed (CRu),
hours. Neurotoxicity was considered a DLT if grade 4 for . 96 hours or new-onset grade 3 of 43.8% (n = 21) and a partial response rate of 22.9% (n = 11). Median DOR was 9.3 mo
for . 28 days. Serial CSF and serum samples were collected for translational studies to (range, 0.0-23.5), and median TTR was 0.95 mo (range, 0.9-3.7). Median OS was not
determine immune cell kinetics and the mechanisms of activity and resistance. reached (range, 1.0-33.0); median PFS was 6.0 mo (range, 0.0-26.0). Overall incidence of
Results: Eleven patients were enrolled, underwent apheresis, and had B7H3-CART suc- any-grade treatment-emergent adverse events (TEAEs) was 97.9% (n = 47) and grade $3
cessfully manufactured. Nine received at least one dose of B7H3-CART and were evaluable was 56.3% (n = 27). Any-grade treatment-related adverse events (TRAEs) were experienced
in the dose escalation cohort. One patient in dose level 2 (25x106 cells) experienced a DLT by 75.0% (n = 36), most frequently anemia (18.8%), fatigue (14.6%), neutrophil count
(grade 3 hypertension). No additional DLTs were observed in this dose level after expansion decreased (14.6%), pruritus (14.6%), rash (14.6%), and maculo-papular rash (14.6%).
to 6 patients, and the recommended phase 2 dose was established at 25x106 cells. Toxicity Grade $3 TRAEs were experienced by 27.1% (n = 13), most frequently neutrophil count
otherwise has been primarily related to tumor inflammation-associated neurotoxicity decreased (8.3%) and rash maculo-papular (4.2%). Deaths related to TEAEs occurred in 2
(TIAN), observed after 29 of 36 infusions (81%), and managed acutely with anakinra and (4.2%) patients: 1 patient died from seizure and pneumonia, and the other from a fall; these
dexamethasone. The median overall survival (mOS) from date of enrollment for patients grade 5 TEAEs were considered unrelated to study treatment. At data cutoff, 27.1% (n = 13)
receiving at least one dose of B7H3-CART is 14.6 months (95% CI: 2.3 - 26.8 months). One of patients remain on tirabrutinib treatment. Main reasons for discontinuation were
patient is currently receiving B7H3-CART and 4 others are being clinically followed up to disease progression (54.2%, n = 26) and death (8.3%, n = 4), and 1 (2.1%) patient dis-
22 months from enrollment. Conclusions: Intracranial administration of B7H3-CART in continued due to an AE; deaths included the 2 patients with grade 5 TEAEs.
recurrent GBM is technically feasible and safe. TIAN was common but manageable and Conclusions: With an ORR of 66.7%, CR/CRu rate of 43.8%, median DOR of 9.3 mo, and a
reversible with immunomodulators. Correlative analyses on surgical tissue, CSF, and manageable safety profile, the PROSPECT trial supports tirabrutinib monotherapy as a
serum are ongoing. Clinical trial information: NCT05474378. Research Sponsor: California potentially effective treatment option for patients with r/r PCNSL. Clinical trial information:
Institute for Regenerative Medicine; CLIN2-15094. NCT04947319. Research Sponsor: ONO Pharmaceutical Co., Ltd.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 117s

2020 Rapid Oral Abstract Session 2021 Poster Session

Using single-cell transcriptomics to reveal CD226 upregulation and en- Brain metastasis: Incidence, trend analysis, and impact on survival using
hancement of CD19-CAR-T function in the inhibitory CNS microenvironment SEER database (2010-2020). First Author: Shaimaa Fadel, Faculty of Medicine,
of refractory CNS lymphoma. First Author: Ulrike Gerdemann, Dana-Farber/Boston Suez Canal University, Ismailia, Egypt
Children’s Cancer and Blood Disorders Center, Boston, MA Background: Brain metastasis has a poor prognosis in cancer patients with high
Background: Refractory CNS lymphoma (CNSL) has a poor prognosis, with 5-year survival morbidity and mortality rates. An updated comprehensive analysis of patients with brain
of ~30%. We conducted a trial of axi-cel for CNSL, where we achieved a CR of 67%. To metastasis across all primary cancer sites is lacking. So, the study aims to provide the
deeply interrogate the immune mediators of response, we collected daily paired CSF and literature with updated evidence about recent trends and survival analysis of brain
peripheral blood (PB) samples post-CAR T infusion. This enabled single-cell transcriptional metastasis. Methods: Data of 75,797 patients with brain metastasis, diagnosed in 2010-
profiling at an unprecedented depth, identifying key drivers of CD19 CAR T responses and 2020, were extracted using the Surveillance, Epidemiology, and End Results (SEER)
compartment-specific mechanisms of CAR T function. Methods: CNSL patients were software. We used a rate session to calculate the incidence, percent change (PC), and
enrolled in the ‘Axi-cel in CNS Lymphoma’ Trial, NCT04608487. PB and CSF samples were annual percentage change (APC). Rates are per 100,000 and age-adjusted to the 2000
collected daily from Day 0–14 post-infusion, and 5’10x scRNA /TCR-Seq was performed. US Std population. Confidence intervals (CI) are 95% per rate with statistical significance
This analysis focused on peak CAR T expansion (Day 5-10) and included 1,224,178 T cells at P . 0.05. We used SPSS version 23 for data analysis and Kaplan Meier Curve and log-
from 17 patients. Samples were tested for compartment (CSF vs PB) and response-specific rank test for survival analysis. Results: Brain metastasis represented 1.9% of all cancer
(CR vs PD) transcriptomic differences using a mixed-effects model and GSEA. Functional cases with a mean age of 64.4 (Sd = 11.2). The age-adjusted incidence rate of brain
assays were conducted with CD19 CAR Ts and CD19 CAR+ Jurkat-NFAT reporter cells. metastasis was 7.1 with a PC of -9.6 from 2010 to 2020 (APC = -0.60; 95% CI: -1.2-0.001,
Results: Transcriptomic analysis identified distinct compartmental differences in CAR Ts, P , 0.05). The APC was significantly declining in Caucasians (-0.70; P . 0.05) and
with PB CAR Ts displaying a robust proliferation signature, while CSF CAR Ts were enriched African Americans (-1.2; P , 0.05) with a significant decrease in males (-1, P , 0.05)
for type I interferon and T-cell dysfunction signatures, including upregulation of inhibitory while the Asian or Pacific islanders (API) race had PC of 11.7 and APC of 1.30 (P , 0.05).
genes PD-1, TIGIT, TIM3, LAG3. In vitro, CSF-exposed CD19-CAR Ts showed increased Lung, breast, skin melanoma, and kidneys were the most common primary sites for brain
expression vs culture-media controls for TIGIT (up 40.1%, SEM 7.4), PD-1 (18.9%, SEM 3.8), metastasis (78.5%, 3.8%, 3.7%, and 3.2%). The 5-year relative survival of patients with
and Tim3 (60.6%, SEM 7.5). CAR T NFAT expression was reduced from 18.4, 0.1 SEM brain metastasis was 6.1% compared to the non-metastatic group 71.5%. The 5-year age
(media) to 7.2, 0.2 SEM (CSF) relative to unstimulated controls. Differential expression
standardized relative survival was 5.7% for the metastatic group. The 5-year relative
analysis comparing CSF CD8+ CAR Ts from patients achieving CR (n = 11) or PD (n = 4)
survival of brain metastasis was higher in the API race compared to Caucasians and
showed upregulation of Type I interferon signaling (IFIT1, IFIT3) in PD patients. In contrast,
African Americans (10.1%, 5.9%, and 5.3%). Conclusions: The results of this study
CR patients exhibited increased expression of counter-inhibitory genes (TCF7, CD226) in
show a very poor survival outcome for brain metastasis. However, there was a significant
CSF CAR Ts, suggesting a functional advantage of these CAR Ts in the inhibitory CNS
environment. Functional assays of CD19-CAR Ts overexpressing the costimulatory
decline in brain metastasis trends over the years, which highlights promising im-
molecule CD226, demonstrated higher lymphoma-cell killing vs WT-CAR Ts (48.6%, SEM provements in the early detection of primary cancers. Further stratifications showed
4.1 vs 24%, SEM 2.2, p = 0.01) and greater IFNg production (63.2%, SEM 1.6 vs 49.4%, SEM disparities according to race and primary cancer site. These data may have clinical-
1.7, P = 0.006). CD226 acts as a costimulatory receptor and counteracts TIGIT signaling by directed variations in screening and counseling for subpopulations with cancer.
competing for its shared ligands, CD112 and CD155. Notably, scRNA-Seq receptor-ligand Research Sponsor: None.
analysis identified CD112 exclusively expressed in myeloid cells, highlighting a critical
myeloid-CAR T interaction that enhances CD226high CAR T efficacy. Conclusions: ScRNA-
Seq suggests tissue-specific CAR T dysfunction in the CNS microenvironment, with CR
patients demonstrating upregulation of counter-inhibitory genes, including CD226. This
study offers novel insights into axi-cel’s mechanism of efficacy and identifies targets to
improve CNSL CAR T therapy. Research Sponsor: Kite/Gilead.

2022 Poster Session 2023 Poster Session

A phase I/II study to assess safety and preliminary evidence of a therapeutic Efficacy of immune checkpoint inhibitors (ICI) in patients (pts) with central
effect of azeliragon combined with stereotactic radiation therapy in patients nervous system (CNS) metastases (mets) from solid tumors: A systematic
with brain metastases (ADORATION). First Author: Rupesh Kotecha, Miami Cancer review and meta-analysis. First Author: Soraia Martins, Université libre de Bruxelles
Institute, Baptist Health South Florida, Miami, FL (ULB), Hôpital Universitaire de Bruxelles (H.U.B), Institut Jules Bordet, Brussels, Belgium
Background: Azeliragon is an oral, brain penetrating small molecule inhibitor of the Background: Brain metastases, a common complication of solid tumors, are associated with
receptor for advanced glycation end-products (RAGE), reducing neuroinflammation by poor outcomes. The role of ICIs in this setting remains unclear. This meta-analysis aims to assess
inhibiting peritumoral edema/vascular leakage and overcoming radiation resistance. intracranial efficacy of ICI-based systemic treatment in pts with CNS mets from solid tumors.
The primary objective of this study is to evaluate the safety and tolerability of azeliragon Methods: A systematic literature search of PubMed, Embase, CENTRAL, and conference pro-
ceedings (ESMO and ASCO) up to 15-Mar-24 (PROSPERO: CRD42021242755), was conducted to
plus stereotactic radiosurgery (SRS) for patients with brain metastasis substituting for
identify single-arm phase II or III, or randomized controlled trials of pts with CNS mets from solid
peri-procedural corticosteroids (loading dose [LD] and corticosteroid taper [CT]) and tumors at baseline treated with ICI-based systemic treatment. The primary objective, CNS ef-
secondarily to assess the potential efficacy of this novel therapeutic combination. ficacy, was measured by pooled CNS objective response rate (CNS-ORR) and weighted median
Methods: ADORATION (NCT05789589) is a single center, open-label, phase I/II trial. CNS progression-free survival (mCNS-PFS). Subgroup analyses evaluated the impact of disease
Eligible adults have a confirmed cancer diagnosis within 5 years, maximum brain and treatment characteristics. Overall effects were pooled using random-effects models.
metastasis diameter of #2 cm, and have discontinued corticosteroids at least 5 days Results: Out of 1 690 records screened, a total of 1 224 pts enrolled in 32 clinical trials were
prior to SRS. In phase I, participants were enrolled into sequential cohorts, starting with included. Overall, ICI-based systemic therapy led to a CNS-ORR of 38.0% (95% confidence interval
azeliragon + SRS + LD; depending on dose-limiting toxicities (DLTs), the next cohorts [CI] 31.8-45.5) and a mCNS-PFS of 9.1 months (mos). CNS efficacy was numerically higher in pts
could be either azeliragon + SRS or azeliragon + SRS + LD + CT. A DLT was defined as any with non-small cell lung cancer (NSCLC, n=383, CNS-ORR 45.8% [34.4-60.9]; mCNS-PFS 9.6 mos)
CNS-specific Grade $ 2 toxicity requiring corticosteroid treatment or any Grade $ 3 and melanoma (n=554, CNS-ORR 37.7% [30.6-46.5]; mCNS-PFS 11.4 mos) vs multi tumors
events not clearly due to the underlying disease or extraneous causes. Results: In the (n=128, CNS-ORR 24.8% [7.7-80.1] and mCNS-PFS 2.8 mos). Efficacy was also greater in first-line
therapy (n=553, CNS-ORR 45.2% [36.7-55.7]; mCNS-PFS 8.6 mos) vs second or later lines (n=190,
completed phase 1 portion, 3 patients were initially treated with azeliragon at 30 mg
CNS-ORR 14.9% [7.6-29.2]; and mCNS-PFS 2.6 mos). Dual ICI (n=279, CNS-ORR 43.9% [35.5-54.2];
twice daily for 6 days followed by SRS+LD within 7 days of starting drug then a mCNS-PFS 17.8 mos) and ICI plus non-ICI agents (n=432, CNS-ORR 48.7% [39.6-59.9]; mCNS-PFS
continuous dose of 20 mg daily for at least 8 weeks. As no DLTs were observed, the 7.1 mos) were more effective than single ICI (n=419, CNS-ORR, 20.4% [11.9-34.9]; mCNS-PFS 4.8
second cohort of 3 patients was treated with azeliragon and SRS without any corti- mos). Subgroup analyses showed superior CNS outcomes in cerebral mets, treated lesions,
costeroids (LD or CT). Of the 6 evaluable patients treated to 46 brain metastases, the asymptomatic pts, and low/no steroid use (table). Conclusions: ICI-based regimens demonstrate
most common primary histology was lung adenocarcinoma (n = 4). At data cutoff (1/8/ CNS efficacy in pts with solid tumors, particularly in pts with NSCLC and melanoma treated with
2025), the median follow-up was 4.9 months (3.8-9.4 months) and no DLTs were first-line combination therapies. Research Sponsor: None.
observed. Early response rate (RR) to the combination therapy was assessed at week 8,
Subgroup analyses by CNS Category
with a per-patient RANO RR of 100% (partial response [PR] for all 100%), and a per-lesion characteristics (n of pts for CNS-ORR) CNS-ORR, % (95% CI) mCNS-PFS, mos
RANO RR for all RANO-defined target lesions (n = 18) of 100% (PR 95.5%, complete
Type of CNS mets Cerebral (1063) 38.3 (31.9-45.9) 9.6
response [CR] 4.5%). For all brain metastases treated (n = 46) the RR was 93.5% (PR for Leptomeningeal (51) 34.6 (15.9-75.2) 2.4
69.6% and CR for 23.9%). Neurocognitive function batteries, symptom inventories, and CNS local therapy Treated (197) 37.6 (26.8-52.8) 8.9
quality of life evaluations remained stable during the 8-week early assessment period. Untreated (307) 43.1 (35.0-53.0) 5.2
Symptoms Asymptomatic/mild (832) 40.4 (34.5-47.3) 10.3
Conclusions: Azeliragon was safely substituted for corticosteroids in this phase 1 study Symptomatic (80) 31.3 (18.3-53.4) 2.5
with no DLTs observed. The early response rate appears encouraging and accrual to the Concomitant steroids No (391) 30.6 (21.6-43.4) 15.1
phase II expansion cohort (n = 40) with a primary endpoint of objective response rate is Low dose (434) 40.7 (31.5-52.6) 7.6
Any dose steroids (144) 41.6 (31.7-54.7) 3.3
ongoing. Clinical trial information: NCT05789589. Research Sponsor: Cantex
Pharmaceuticals.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
118s CENTRAL NERVOUS SYSTEM TUMORS

2024 Poster Session 2025 Poster Session

Irradiated tumor volume as a predictor of local recurrence and radionecrosis Efficacy of systemic therapy in breast cancer with CNS metastases: “Real-
in lung cancer with brain metastases treated with stereotactic radiosurgery. world” experience. First Author: Bipin Ghimire, Henry Ford Health System, Detroit, MI
First Author: Andreas Koulouris, Karolinska University Hospital, Solna, Sweden Background: The incidence of CNS metastases in breast cancer is rising. While local
Background: Stereotactic radiosurgery (SRS) is a standard local treatment for brain therapies such as surgery and radiation remain standard, data on upfront systemic
metastases (BM), but it may result in local recurrence (LR) or radionecrosis (RN). This therapies for active brain metastases, especially HER2-negative patients, is limited. This
study evaluates irradiated tumor volume as a predictor of LR and RN in SRS-treated lung study examines upfront systemic therapy efficacy for CNS metastases in breast cancer
cancer patients with BM. Methods: We retrospectively analyzed 431 lung cancer pa- patients at a single institution, including a majority African American (AA) population.
tients with BM who underwent SRS at Karolinska University Hospital, Sweden, (2009- Methods: A retrospective chart review included breast cancer patients with CNS
2020), encompassing all-comers from the Stockholm region. Associations among ir- metastases treated at Henry Ford Health (January 2014–July 2024). Eligible patients
radiated tumor volume and risks of RN, symptomatic RN, as well as LR at 6 and had not received concurrent local therapy; prior local therapy was permitted if unrelated
12 months, were assessed using Cox regression models. Furthermore, we evaluated the to the studied lesions. CNS response was to be assessed using RANO-BM (defines
diagnostic performance of Methionine PET-CT in differentiating RN from LR. measurable disease as lesions . 10 mm) for parenchymal and modified RANO-LM
Results: 40 patients (9.2%) developed asymptomatic RN, 37 (8.3%) symptomatic RN, criteria for leptomeningeal disease (LMD). Results: Among 35 patients (20 AA, 13
and 67 (15.5%) LR. Larger tumor volumes significantly increased the risks of RN and LR. Caucasian), with a median age of 54 years, HER2-positive was the most common
At 6 months, a tumor volume of 4.75 cm³ was associated with an RN risk reaching the receptor type (49%), followed by HR-positive (37%) and triple-negative (14%); nearly half
upper limit of 20%. By 12 months, substantially smaller volumes, such as 1.13 cm³, were (46%) had HER2-low disease. Parenchymal metastases were predominant (86%); three
related to same risk levels. Symptomatic RN followed a similar trend, with a volume of had co-existing LMD, and two others had only LMD. Most metastases were multiple; 91%
13.58 cm³ presenting a risk of up to 20% at 6 months, while at 12 months, considerably had lesions , 10 mm. 43% had prior WBRT or SRS to unrelated lesions. HER2-positive
smaller volumes, such as 3.8 cm³, corresponded to a symptomatic RN risk as high as patients had the highest CNS overall response rate (ORR, 53%) and disease control rate
40%. 20% risk of LR was observed with volumes of 5.66 cm³ and 3.28 cm³ at 6 and (DCR, 94%), followed by HR-positive (ORR 31%, DCR 69%) and triple-negative (ORR and
12 months, respectively. The sensitivity and specificity of Methionine PET-CT are 0.909 DCR 20%). Median PFS did not significantly differ between receptor groups (p = 0.130).
and 0.600 when MRI was considered the gold standard. Conclusions: Larger irradiated Trastuzumab-deruxtecan (T-Dxd) was the most common regimen (10/35) and within
tumor volumes were positively correlated with an increased risk of both RN and LR. At HER2-positive and HR-positive groups. T-Dxd achieved CNS ORR of 60%, DCR of 90%,
12 months post-SRS, smaller tumor volumes were associated with higher RN and LR and median PFS of 16 months. Tucatinib-based regimens showed a 100% DCR with
risks in comparison with 6 months. Methionine PET-CT, when used alongside MRI, did median PFS of six months. Other therapies, including sacituzumab, abemaciclib, and
not demonstrate a clear advantage in differentiating LR from RN. Research Sponsor: trastuzumab-emtansine, showed stable disease as the best response. Among AAs, HR-
European Society for Medical Oncology (ESMO); Hellenic Society of Medical Oncology positive was the most common receptor type (50%). These patients had ORR of 35% and
(HeSMO); Elena Iliopoulou Giama (EIG) Cancer Research & Scholarship Foundation; DCR of 65%. T-Dxd maintained ORR of 60% and DCR of 80%. Of five patients with LMD,
Scholarship - Legacy “M. M. Manassaki” by the University of Crete; Region Stockholm three were HER2-positive, and two were HR-positive, with an ORR of 60%, and DCR of
(clinical postdoctoral appointment); Stockholm Cancer Society; 204053. 80%. Conclusions: This “real-world” experience highlights that, at our institution, most
patients with breast cancer and CNS metastases considered for upfront systemic
therapy lack measurable disease (91% having lesions , 10 mm) typically required for
clinical trials. Nonetheless, the response rate aligns with published experiences. In
addition, we included patients with LMD, who are often excluded in trials. Our data also
suggests impressive CNS responses with T-Dxd, both overall and in AA patients. The
management of brain metastases and LMD in these patients is best approached in a
multidisciplinary format. Research Sponsor: None.

2027 Poster Session 2028 Poster Session

Whole brain radiotherapy and intrathecal injection of thiotepa, combined Effect of CD4+PD-1+CXCR6+ T cells on the response of immune checkpoint
with systemic treatment for the primary tumor, to treat solid tumor lep- inhibitor therapy in brain metastases of NSCLC. First Author: Yang-Si Li,
tomeningeal metastasis: A prospective, single center, single arm, phase II Guangdong Lung Cancer Institute, Guangdong Provincial People’s Hospital, Guangzhou,
clinical study. First Author: Siyu Guo, Zhejiang University, Hangzhou, China China
Background: To evaluate the efficacy and toxicity of a triple therapy regimen consisting Background: Brain metastases (BrM) in non-small cell lung cancer (NSCLC) presented a
of Hippocampal-sparing whole-brain radiotherapy (HS-WBRT), intrathecal Thiotepa significant challenge due to poor prognosis. While immune checkpoint inhibitors (ICIs)
(ITT), and primary lesion treatment for solid tumor leptomeningeal metastasis (LM). have been standard treatments for NSCLC, their efficacy in BrM is variable, emphasizing
(NCT06376292). Methods: Based on the comprehensive results of MRI and cytology the urgent need for predictive biomarkers and fundamental mechanisms. Methods: We
evidence, patients diagnosed with LM according to the diagnostic criteria in the EANO- prospectively collected 20 cerebrospinal fluid (CSF) and 4 BrM tumors from 18 NSCLC
ESMO guidelines meet the criteria. Patient began ITT twice a week, and underwent HS- patients with BrM undergoing ICI therapy for single-cell RNA sequencing (scRNA-seq),
WBRT as soon as possible. Before each ITT, cerebrospinal fluid (CSF) pressure is complemented by integrating data from multiple published datasets. Three independent
measured and CSF is collected for testing, including protein, tumor markers, IgG, al- cohorts (8 and 25 CSF, and 31 BrM tumors) underwent flow cytometry, proteomics, and
bumin levels of CSF, and albumin ratio. MRI re-examination is conducted every three multiplex immunohistochemistry for validation, respectively. Results: Our study
months, and the RANO-LM criteria is used to evaluate the response of patients after provided a high-resolution atlas of cellular dynamics in the CSF and BrM during ICI
treatment. Besides,the LM-PROG SCORE we designed can be used as an indicator to therapy in NSCLC patients with BrM. Notably, we identified a key immune cell subset,
evaluate the treatment effect and adjust the medication frequency or switch the CD4+PD-1+CXCR6+ T cells, as a positive predictor of ICI intracranial tumor responses,
treatment line of intrathecal Pemetrexed (IP). The primary endpoint is overall survival. which presented highly functional and transcriptomic similarities in both CSF and BrM
Results: As of December 1, 2024, a total of 57 patients have been enrolled. 40 patients tumor environment. Moreover, CXCR6 could serve as a specific marker for CD4+PD-1+
were included in the statistics. Most cases are lung cancer (27, 67.5%), in addition to T cells linked to ICI response. Further, we revealed that the novel cluster of CD4+PD-
breast cancer(8, 20%), gastric cancer (3, 7.5%), rectal cancer (1, 2.5%) and cervical 1+CXCR6+ T cells was closely associated with lymphocyte activation and aggregation in
cancer (1, 2.5%). The mOS was 7.8 months (95% CI 2.06-13.54 months). The mPFS was CSF and BrM of ICI responders, and cDCs of ICI responders interacted with CD4+PD-
5.63 months (95% CI 0.76-10.51 months). The RANO-assessed ORR to treatment was 1+CXCR6+ T cells for enhanced antigen presentation and inflammatory activation.
62.5% (10/16), DCR was 87.5% (14/16), with 24 patients (60%) not reaching the follow- Conclusions: Our findings revealed critical insights into the immune landscape of
up time. The three longest survival among alive patients are 21.3 months, 14.9 months, NSCLC BrM under ICI therapy, highlighting CD4+PD-1+CXCR6+ T cells in CSF as a
and 12.7 months. The most significant effect of combination therapy is the rapid relief of promising biomarker and illuminating fundamental mechanisms underlying ICI efficacy.
symptoms. 25% of patients (10/40) had already experienced unconsciousness (RASS„0) Research Sponsor: None.
at the time of diagnosis. After our treatment, all patients regained consciousness (RASS
= 0). The long-term therapeutic effect also significantly reduces tumor markers, protein
content, albumin and IgG content, pressure drop, and albumin ratio in CSF, indicating the
recovery of the blood-brain barrier. 85% of patients experience varying degrees of bone
marrow suppression during treatment, but most are mild and can tolerate subsequent
ITT maintenance after treatment. Conclusions: Under the efficacy guidance of LM-
PROG SCORE we designed, our combination therapy can greatly improve patients’
overall survival, progression free survival, and quality of life. And it was found that in
addition to imaging, indicators such as protein content, IgG and albumin content, Al-
bumin Ratio and tumor markers in CSF can serve as efficacy evaluation indicators.
Clinical trial information: NCT06376292. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 119s

2029 Poster Session 2030 Poster Session

IT-IO: Intrathecal administration of nivolumab and ipilimumab in combina- Evaluating the utility of DNA methylation signatures in tissue and biofluids
tion with systemic combination of nivolumab and ipilimumab in patients for lung adenocarcinoma brain metastasis prediction and non-invasive
with non-small cell lung cancer or melanoma and newly diagnosed lepto- detection. First Author: Jeffrey Zuccato, Oklahoma University Health Sciences Cen-
meningeal metastasis, a multicentric phase I study. First Author: Emilie Le Rhun, ter, Oklahoma City, OK
University Hospital Zurich, Zurich, Switzerland Background: Brain metastases (BM) are common and arise in 30% of lung adenocarcinoma
Background: The optimal management of patients with leptomeningeal metastases (LUAD) patients. Patients with LUAD that develop BM have significantly poorer outcomes, with a
(LM) from non-small cell lung cancer (NSCLC) or melanoma remains controversial. IT-IO 10-16 month median overall survival. Unfortunately, current clinical practice for BM prediction is
(NCT05598853) is a prospective phase I, multicenter, open label, interventional clinical limited and so BM are typically detected after they develop and grow to cause neurological
symptoms. Once BM are detected, currently neurosurgical tumor biopsies are performed to enable
study aiming at determining the recommended phase 2 dose (RP2D) of intrathecal BM diagnosis via neuropathological evaluation. The aims of this study were to develop DNA
nivolumab and ipilimumab in patients with newly diagnosed LM from NSCLC or mel- methylation-based models that predict LUAD BM and non-invasively detect BM in blood to enable
anoma. Methods: The diagnosis of LM had to be confirmed or probable by EANO ESMO early diagnosis and treatment. Methods: DNA methylomes were acquired from 402 tumor tissue
criteria. Planned whole brain radiotherapy (WBRT) was not allowed. Planned or prior and plasma samples in a cohort of 346 LUAD and BM patients. Machine learning models were built
craniospinal irradiation were not allowed. The treatment regimen consisted of intra- using DNA methylation signatures that stratify BM risk in tissue and detect BM in plasma. Models
thecal nivolumab (fixed dose 50 mg) / ipilimumab (increasing doses) in combination with were evaluated in independent validation datasets. A predictive nomogram was developed using
systemic combined nivolumab/ipilimumab. Three dose levels of IT ipilimumab were the BM prediction model together with clinical factors to provide composite patient-specific scores
planned: 5 mg (dose level 1), 10 mg (dose level 2), and 20 mg (dose level 3). RP2D, the reflecting BM risk. Results: The methylation-based BM predictor accurately stratified BM risk in a
primary endpoint, was determined in a 3+3 design. Secondary endpoints included univariable Cox model using validation set data (HR = 5.65, 95%CI 1.85–17.2, p = 0.0023). Model
utility was independent of the predictive value of clinical factors in a multivariable Cox model using
compartmental efficacy and survival. Results: A total of 19 patients, 6 female and 13
validation set data (Table 1: HR = 8.92, 95%CI 1.97–40.5, p = 0.0046). The 5-year model accuracy
male patients, 12 with melanoma and 7 with NSCLC, were enrolled between February was 0.81 and significantly higher than a similarly built cancer stage-based model (0.65), dem-
2022 and August 2024. Median KPS at study entry was 80, 12 patients had a positive onstrating utility over current practice. The combinatorial clinical-methylomic predictive nomo-
CSF. The dose escalation phase (n = 12) was completed without dose-limiting toxicity gram had enhanced utility with an accuracy of 0.82 univariable Cox HR of 17.2 (95%CI 4.13–71.3,
until dose level 3. The RP2D is nivolumab 50 mg and ipilimumab 20 mg. Sixteen SAE p , 0.0001), demonstrating comprehensive patient-specificity. The plasma-based model accu-
were noted, all unrelated or unlikely related to intrathecal therapy. Three patients are still rately classified BM from gliomas and lymphomas (AUROC=0.80), as typical clinical differential
alive. For the whole cohort, median overall survival was 3 (range 0.6-10.3) months, for diagnoses, in validation set data. The models were validated further in additional external data.
patients with a diagnosis of melanoma 2.9 (range 0.6-8.2) and for patients with NSCLC Conclusions: DNA methylation-based modeling of BM can accurately predict LUAD patients at risk
4.5 (range 0.8-10.3) months. OS at 6 months was 20% (one patient ongoing at for BM development and can non-invasively detect BM that develop. Future treatment approaches
5.1 months). Translational research is ongoing. Conclusions: No safety issue was may tailor initial LUAD treatment and ongoing cancer surveillance to a patient’s BM risk, allowing
for the potential to prevent and treat BM early. Research Sponsor: None.
noted. Efficacy data are preliminary and need to be confirmed in larger trials. Clinical trial
information: NCT05598853. Research Sponsor: Bristol Myers Squibb. DNA methylation-based BM prediction is independent of clinical factors in a multivariable Cox
proportional hazards model.
Variable HR 95% CI p
Methylome risk score 8.92 1.97–40.5 0.005
Age Years 0.96 0.92–1.02 0.177
Smoking Pack-years 0.99 0.95–1.03 0.496
EGFR Mutant vs wildtype 0.92 0.25–3.34 0.895
T T2 vs T1 1.58 0.41–6.04 0.505
T324 vs T1 1.49 0.28–7.98 0.642
N N1 vs N0 1.05 0.31–3.58 0.943
N223 vs N0 1.00 0.27–3.69 0.995
M M1 vs M0 145 12.2–1730 ,0.001

2031 Poster Session 2032 Poster Session

Effect of early integrated neuropsychological care in patients with brain Intrathecal deferoxamine in patients with leptomeningeal metastases:
metastases: A phase 2 randomized controlled trial (ATHENA trial). First Phase 1a analysis. First Author: Jessica Wilcox, Memorial Sloan Kettering Cancer
Author: Haley Kopp Perlow, Department of Radiation Oncology, Seidman Cancer Center Center, New York, NY
University Hospitals/Case Western Reserve University, Cleveland, OH Background: Leptomeningeal metastases (LM), the spread of cancer to the cerebro-
Background: Advancements in radiotherapy delivery through both hippocampal sparing spinal fluid (CSF), is associated with high morbidity and mortality. LM employ the iron-
whole brain radiotherapy (WBRT) and stereotactic radiosurgery (SRS) can better preserve binding transporter and receptor system, lipocalin-2/SLC22A17, to scavenge iron from
QOL and reduce cognitive decline. However, even patients treated with advanced brain the CSF to sustain their metabolic needs. In preclinical models of LM, intrathecal
radiotherapy techniques have a reduction in their QOL and cognitive abilities either due to administration of deferoxamine (IT-DFO), an iron chelator, resulted in reduction of LM
their radiation treatment, systemic therapy, or progression of disease. This Phase 2 growth and improvement of survival. We evaluated this novel treatment strategy in this
Randomized Controlled Trial (NCT05503251) aims to evaluate the impact of a neuro- first-in-human clinical trial in patients with solid tumor LM. Methods: This is a phase 1a,
psychological evaluation and intervention with a certified neuropsychologist on QOL and single-institution, clinical trial to determine safety and maximum tolerated dose (MTD)
cognitive function for brain metastases patients treated with radiotherapy. Methods: Brain of IT-DFO in patients with LM. Eligibility criteria included LM from any solid tumor, age $
metastases patients were randomized 1:1 to either neuropsychology evaluation and in- 18 years, Karnofsky Performance Status $ 60, life expectancy $ 8 weeks, and Ommaya
tervention plus brain radiotherapy or brain radiotherapy alone. The intervention arm included reservoir. Patients were enrolled in an accelerated 3+3 dose escalation design with a
five appointments with the neuropsychology team for testing, evaluation, and counseling primary endpoint of dose-limiting toxicity (DLT), defined as a grade 3 non-hematologic or
over a three-month period. Patients with any number of brain metastases and an estimated grade 4 hematologic toxicity in the first cycle of treatment. All patients received IT-DFO
survival of $ 6 months were included. Exclusion criteria included prior WBRT and pre-
twice weekly (cycle 1), once weekly (cycle 2), then once every two weeks (cycle 3+) in 28-
existing mental disability. Stratification factors for randomization were Karnofsky per-
day cycles. Patients were monitored for LM progression by neurological examination,
formance status (KPS, . 70 vs. # 70) and radiation cohort ( . 15 brain metastases
received WBRT, #15 received SRS). All patients receiving WBRT were prescribed mem-
neuraxial magnetic resonance imaging, and CSF cytology as per modified Response
antine. The primary endpoint was deterioration of QOL at 3 months as measured by Fact-Br. Assessment in Neuro-Oncology LM criteria. Results: A total of 8 patients received
Repeated measures analysis of variance was used to measure QOL. Cognition was measured treatment with IT-DFO from May 2022 to January 2025 at the time of data cut-off. The
by Hopkins Verbal Learning Test-Revised, Controlled Oral Word Association Test, and Trail median age at enrollment was 50 years (range, 26-69). The primary malignancy included
Making Test A/B, with cognitive decline defined as decline on at least one assessment using breast (n = 4), lung (n = 2), colon (n = 1), and sarcoma (n = 1). Patients were treated with
reliable change index. Results: Between August 2022 and June 2024, 110 patients were IT-DFO at doses of 10 mg (level 1, n = 4) and 30 mg (level 2, n = 4). IT-DFO was well
randomized. Baseline characteristics were balanced between arms and included a median tolerated, and the majority of adverse events (AEs) were grade 1-2. The most common
KPS of 90 (IQR 80, 90), median age of 62.5 (IQR 54, 70), 53% female patients, 43% of patients any grade AEs were vomiting (50%), nausea (37.5%), chills (25%), myalgias (25%), and
with a primary lung cancer, and most patients (74%) with #15 brain metastases. The tremor (25%). Two patients experienced DLTs at 30 mg (grade 3 vomiting, grade 3
median overall survival or time to last follow-up was 8.5 months. The primary endpoint, syncope). No grade 4-5 AEs were observed. The MTD was determined to be 10 mg. In this
deterioration of QOL at 3 months, was not different between the control and intervention heterogenous heavily pretreated population, median overall survival for the evaluable
arms (p = 0.93). Cognitive decline differences at 3 months were not significant between the cohort (n = 7) was 10.0 months (95% CI, 6.5 – NA). Conclusions: IT-DFO is a novel, well
control and intervention arms (24.1% vs. 27.3%, p = 0.33). Additionally, there were no tolerated investigational treatment for LM. A phase 1b dose expansion study at a dose of
differences at 3 months with verbal fluency, executive function, immediate recall, delayed 10 mg is currently underway to better define safety and efficacy endpoints. Clinical trial
recall, or delayed recognition between arms. Conclusions: This study did not meet its information: NCT05184816. Research Sponsor: MSK Center for Experimental Thera-
primary endpoint, better preserved QOL at 3 months for patients receiving early integrated peutics; F. M. Kirby Foundation; ASCO Conquer Cancer Young Investigator Award 2021.
neuropsychological care. Further evaluation of the delayed impact ( . 6 months) of
neuropsychology intervention on QOL and cognitive function will be reported when data are
available. Clinical trial information: NCT05503251. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
120s CENTRAL NERVOUS SYSTEM TUMORS

2033 Poster Session 2034 Poster Session

Improving adherence to cancer care for socioeconomically disadvantaged Risk of intracranial hemorrhage with DOACs vs LMWH in patients with
patients with central nervous system tumors. First Author: Joshua Amit Budhu, cancer-associated thrombosis and brain metastases. First Author: Ali Mush-
Memorial Sloan Kettering Cancer Center, New York, NY taq, Cleveland Clinic Foundation, Cleveland, OH
Background: Socially disadvantaged patients often face significant barriers to adhering Background: Intracranial hemorrhage (ICH) is a major and often devastating complication in
to and completing cancer treatment. Patients with central nervous system (CNS) tumors patients with brain metastases requiring therapeutic anticoagulation for cancer-associated
experience cognitive, neuropsychiatric, speech, motor, sensory, and gait symptoms that thromboembolism (CAT). While direct oral anticoagulants (DOACs) provide a convenient al-
exacerbate socioeconomic barriers to care. The Integrated Cancer Care Access Network ternative to low-molecular-weight heparin (LMWH), their safety in this population remains
(ICCAN) is a multi-institutional program developed by Memorial Sloan Kettering Cancer unclear. Comparing ICH risk between DOACs and LMWH is crucial for optimizing anti-
Center’s Immigrant Health and Cancer Disparities Service (IHCD). Through ICCAN, coagulation in these high-risk patients. Methods: This retrospective cohort study utilized
patients are provided with patient navigation and resources to mitigate barriers to TriNetX, a multi-institutional database, to analyze adults with solid tumors who developed CAT
within six months of brain metastasis diagnosis. Patients receiving therapeutic-dose DOACs
treatment adherence and completion. While ICCAN has been shown to increase
(apixaban, rivaroxaban, edoxaban) or LMWH within ten days of venous thromboembolism
treatment adherence and completion for patients with other cancers, it has not his- diagnosis were compared. 1:1 propensity score matching for over 50 covariates, including age,
torically enrolled patients with CNS tumors. Methods: Under this pilot ICCAN-CNS sex, and cancer type (Table 1). We assessed ICH incidence, bleeding events, ICU admissions,
program, 58 patients with either primary brain tumors or brain metastases were referred and all-cause mortality using Kaplan-Meier survival analysis and Cox proportional hazards
to the program. Of these, 45 patients enrolled, 5 patients died before contact was made, models. Subgroup analyses examined ICH risk by cancer type. Results: After matching, 4,275
2 patients declined participation, and 4 patients could not be reached. The patient patients were included in each group. DOACs were associated with a statistically significant
population was evenly split between primary brain tumors (mainly glioblastomas and lower risk of ICH (HR 0.855, 95% CI 0.731-0.999, p=0.049). Additionally, significantly lower
low-grade gliomas) and brain metastases. Patients were eligible if they were 18 years or rates of ICU admission (16.7% vs. 20.3%; p,0.001) and all-cause mortality at 12 months
older and receiving active treatment or were under active surveillance. If a patient was (42.4% vs. 48.9%; p,0.001) were observed in the DOAC group. Subgroup analyses showed a
eligible, they were administered an extensive needs assessment survey that took 60 trend toward lower ICH with DOACs in lung cancer (5.9% vs. 6.1%, p=0.726), melanoma (13.7%
minutes to complete. The interviews consisted of basic demographic questions, Alliance vs. 15.9%, p=0.432), and renal cell carcinoma (5.7% vs. 9.0%, p=0.103), but these differences
Distress Screening Tool, Health Related Social Needs (HRSN) Assessment, Essential were not statistically significant. No significant differences were found for breast (3.6% vs.
Needs Assessment, Patient Satisfaction with Cancer Care questionnaire, and an ICCAN- 4.5%, p=0.375) or colorectal cancer (4.0% vs. 5.4%, p=0.331). Conclusions: DOACs were
CNS specific questionnaire for patients with brain tumors. Caregivers were allowed to associated with significantly lower ICH, ICU admission, and mortality compared to LMWH in
assist patients with neurocognitive or speech deficits in completing the survey patients with brain metastases requiring anticoagulation, supporting their role as a viable and
questions. Patients were then provided with both patient navigation and resources well-tolerated alternative. While subgroup analyses did not show significant differences in ICH
risk by cancer type, the overall findings indicate a favorable profile for DOACs. These results
depending on their needs. Follow-up assessments were conducted at the 2-, 4-, and 6-
highlight the need for individualized anticoagulation strategies and warrant further prospective
month marks. Results: Patients with CNS tumors completed the initial needs as- validation. Research Sponsor: None.
sessments and additional questionnaires. Trends of delayed responses and the need for
questions to be repeated were observed, however, patients were still able to express Baseline characteristics after matching.
socioeconomic needs. The main needs expressed among these patients were income, Characteristic DOAC (n=4275) LMWH (n=4275)
employment issues due to lack of ability to work from their cancer, food access, and Age (years), mean 6 SD 63.4 6 11.9 63.5 6 11.9
transportation to appointments. Overall, the response to the program was positive from Female, n (%) 1969 (46.1%) 1958 (45.8%)
Lung Cancer, n (%) 2218 (51.9%) 2229 (52.1%)
patients, with many patients accessing resources with the assistance of the access Breast Cancer, n (%) 692 (16.2%) 672 (15.7%)
facilitator. Conclusions: This pilot demonstrates the feasibility and value of including Melanoma, n (%) 280 (6.5%) 279 (6.5%)
CNS tumor patients in patient and resource navigation programs. Future plans include a Renal Cell, n (%) 241 (5.6%) 244 (5.7%)
Colorectal, n (%) 408 (9.5%) 413 (9.6%)
randomized controlled trial using the ICCAN intervention for patients with glioblastoma.
Research Sponsor: None.

2035 Poster Session 2036 Poster Session

Risk factors (RF) for brain metastases (BM) in patients (pts) with metastatic Memantine in radiation-induced cognitive dysfunction in brain metastases:
breast cancer (MBC): An analysis of US electronic health records (EHRs). A double-blinded, randomized, placebo-controlled trial (CTRI/2022/01/
First Author: Jose Pablo Leone, Dana-Farber Cancer Institute, Boston, MA 039599). First Author: Haripriya Parapparambil Surendran, Amrita Institute of Med-
Background: BM are a significant clinical challenge in pts with MBC yet risk stratification for early ical Sciences and Research, Kochi, India
identification is suboptimal. We leveraged a large, contemporary, real-world database to characterize Background: Prospective double-blinded, placebo-controlled randomized study to evaluate
RF associated with BM to inform strategies for enriched surveillance and early detection. the role of memantine in brain metastasis (BM) in preserving cognitive function.
Methods: Selected pts from the nationwide Flatiron Health de-identified EHR-derived database had Methods: Clinic-radiologically diagnosed of BM patients planned for radiation therapy (RT)
MBC, had initiated first-line treatment (1L tx) before March 2023 (allowing for . 1 year [yr] of (SRS or whole brain RT) were randomized to receive memantine or placebo (20 mg/day) over
potential follow-up [FU]), and were free of BM at tx initiation. Clinical characteristics were examined
24 weeks. Cognitive function assessed by Addenbrooke’s Cognitive Examination (ACE).
as potential RF for BM incidence at any time during FU, using univariate sub-distribution (sd) and
Secondary outcomes included QoL, white matter volume changes (MRI T2 FLAIR), and plasma
cause-specific (cs) hazard ratios. Additional analyses focused on BM risk at 3 yrs from tx initiation
and included longitudinal data sequential Cox models with landmarks at every 6 months of FU, and a memantine levels (by LC-MS/MS). Safety was assessed using CTCAE v5.0 criteria.
nested case-control (NCC) design for concurrent BM detection. Analyses were conducted on Results: 130 BM patients were enrolled after randomization [placebo 64 & memantine
complete cases, no imputation method was used for missing data, and regression methods using (experimental arm n = 66]. In placebo and memantine arm mean age was 56.5 & 56.7; female
multivariate analyses were used to mitigate confounding. Predictive, modeling-based machine 39 & 40; high school education status 39 (30%) & 37 (28%); SRS in 40 (30%) & 35 (27%);
learning (LASSO Cox regression, random survival forests) was conducted using cs hazard ratios to frontal lobe lesion 43 (33%) & 55 (42%); PS 0-1 55 (42%) & 54 (41%) respectively. In the
identify potential predictors amongst 90 candidates (data-driven approach using most of the placebo arm, ACE scores at baseline in placebo and memantine arm 83.0 6 10.1 and 77.7 6
dataset). Analyses were conducted in the overall cohort and stratified by subtype: HER2-negative/ 12.7 (p = 0.78) respectively. At 4 months, ACE score in placebo and memantine arm 76.2 6
hormone receptor-positive (HER2–/HR+), HER2-positive/HR-negative (HER2+/HR–), HER2+/HR+, 14.3 and 82.2 6 12.7 (p = 0.04). At 6 months in placebo and memantine arm were 72.9 6 20.2
and triple-negative breast cancer (TNBC). Results: The study included 21,368 female pts initiating and 83.9 6 10.8 (p = 0.005). At 24 weeks, ACE scores change was +4.0 in memantine & -9.5 in
1L tx (n = 14,898 HER2–/HR+, 1006 HER2+/HR–, 3468 HER2+/HR+, 1996 TNBC), with 2,530 BM placebo arm; p = 0.001. Memantine arm had better preservation of memory (-3 vs. -2.5, p ,
events. Younger age, HER2+ and TNBC subtypes, and more extensive metastasis ($2 organ sites, 0.001), delayed recall (-1 vs. -1, p , 0.001), and verbal fluency (-1 vs. 0, p = 0.007). In the SRS
particularly liver, lung, or lymph nodes) were associated with higher BM risk (Table); bone-only subgroup, ACE scores in placebo and memantine at baseline, 4 and 6 month was 83.5 (6 9.5)
metastases conferred a lower risk. sd and cs hazard ratios were largely concordant. NCC analyses & 79.7 (6 12.5); 76.6 (6 14.7) & 85.3 (6 10.6); 72.5 (6 23.1) & 86.4 (6 9.5) respectively. At
identified similar predictors for concurrent BM. LASSO Cox modeling yielded a C-index of 0.74 overall 24 weeks, memantine arm improved ACE scores by +4 (0 to 12) compared to placebo -8 (-15.5
(HER2+ 0.70; HER2–/HR+ 0.73; TNBC 0.62). Similar C-indices were seen with random survival
to -2.5) (p , 0.001). At 24 weeks in WBRT, memantine arm sustained cognitive improvement
forests. Conclusions: Clinical characteristics, including metastatic distribution and tumor subtype,
can help identify pts at higher BM risk within 3 yrs of initiating MBC tx. Competing risks (of BM and
(ACE score +3) compared to further decline in placebo (-9.5, p , 0.001). At 24 weeks,
death) did not appear to substantially affect results, except for recurrence time and ECOG PS. percentage change in global health status in placebo & memantine arms were -5.57%. &
Although these findings are encouraging, further refinement of predictive models is needed to +63.3% respectively. 21% required dose reductions due to adverse events. Loss of appetite
improve discrimination and guide targeted neuroimaging and early intervention strategies. Research (25.7% vs. 12.5%, p = 0.05); gastric irritation (0 vs 7.5%; p = 0.02) were higher in memantine
Sponsor: F. Hoffmann-La Roche Ltd. arm. White matter volume changes in the placebo group correlated negatively with cognitive
decline (r = -0.544, p = 0.055), suggesting a potential role of edema in radiation-induced
sd hazard ratio cognitive dysfunction. Memantine at a 5 mg BID dose achieved a median trough concen-
Age, yrs (reference [ref]: <45) 45–55 0.81 tration of 118.06 ng/mL (IQR: 68–211), within the desirable therapeutic range (70–150 ng/
56–65 0.62 mL). 10 mg BID dose trough was 172 (85-290) and peak concentration 397 (258-499 ng/mL)
66+ 0.31
Subtype (ref: HER2–/HR+) HER2+/HR+ 2.45
exceeded alert threshold of 300 ng/mL. Conclusions: Memantine preserved cognitive
HER2+/HR– 3.40 function and QoL in RT for BM. Cognitive benefits were more in SRS than HA/WBRT. White
TNBC 2.33 matter volume change was negatively correlated to cognitive outcomes. 5 mg BID dose
Metastatic sites, n (ref: 1) 2–3 1.55 optimally balances efficacy with tolerability. Clinical trial information: CTRI/2022/01/039599.
4+ 2.63
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 121s

2037 Poster Session 2038 Poster Session

Predictors of overall survival in patients with brain metastases from HER2+ Hippocampal-avoidance whole-brain radiotherapy with dose escalation on
breast cancer. First Author: Qinmei Xu, Stanford University School of Medicine, metastases: A prospective randomized trial (HIPPORAD). First Author:
Stanford, CA Anca-Ligia Grosu, Department of Radiation Oncology, Medical Center - University of
Background: Predictors of overall survival (OS) after brain metastasis (BM) in HER2- Freiburg, Faculty of Medicine, Freiburg, Germany
positive breast cancer (BC) are not well characterized. This study aimed to identify clinical Background: The HIPPORAD trial aimed to evaluate a new method of whole brain
and imaging-derived (radiomic) features that predict OS and develop a combined model for radiation therapy (WBRT) with simultaneous integrated boost (SIB) to the metastases,
better prognostic performance. Methods: Our retrospective study analyzed 289 patients with versus without hippocampal avoidance in patients with brain metastases.
initially diagnosed with non-metastatic HER2-positive BC who later developed BM. We Methods: We conducted a prospective, multicentre, randomised, double-blind trial
used 25 clinical characteristics and 12 treatment parameters to develop a Clinical model. (DRKS00004598). Patients with 4-10 brain metastases . = 5mm were randomised at 13
We developed an Imaging model using a subset of 120 patients, who possessed evaluable centres in Germany 1:1 between WBRT+SIB with hippocampus avoidance (HA-
pre-treatment brain MRI for delineating tumor segmentations on all brain metastatic WBRT+SIB) (arm A) and WBRT+SIB (arm B). Patients and assessors of outcome
lesions. We extracted 1078 radiomic features from each tumor segmentation using were blinded to the randomised arm. All patients received WBRT with 30 Gy and SIB with
PyRadiomics, generating 8 feature sets based on 2 segmentation strategies (largest tumor 51 Gy or 42 Gy in 12 fractions, 5x/week. The primary endpoint was the change in
per patient versus all tumors combined) and 4 tumor feature types (entire tumor, solid neurocognitive function (assessed by the Verbal Learning and Memory Test [VLMT])
component, necrotic component, combined solid and necrotic features with statistical 3 months after treatment. Secondary endpoints included neurocognitive changes at 9
transformations). Morphological features, including lesion number, total size/volume, and and 18 months, development of anxiety and depression, quality of life and measures of
necrotic-to-solid ratios, were also incorporated, along with tumor intracranial location. Cox oncological outcome. Results: Between August 2nd, 2016 and September 7th, 2021, 170
proportional hazards regression model with Coxnet, integrating LASSO and Elastic Net patients were recruited and 136 were randomised between HA-WBRT+SIB (n = 67) and
regularization, was used to predict OS. For fair comparison, we randomly selected 30% (n = WBRT+SIB (n = 69). Of these, 38 patients in arm A and 42 in arm B were known to be alive
31) of the smallest subset (n = 103, largest brain metastasis with both necrotic and solid 3 months after treatment and were included in the primary endpoint analysis. The
components), all of which overlap with other model subsets, as validation cohort. Three change in overall learning performance at 3 months was not significantly different
model types—Clinical, Imaging and Combined—were compared using the concordance
between arms (p = 0.83). VLMT-scores decreased after 3 months, but improved at 9 and
index (C-index) to assess performance based on validation cohort. Results: Clinical model,
18 months, with HA-WBRT+SIB showing an overall superior trend over WBRT+SIB. At
built on the whole cohort (286 women, 3 men; mean age 54.52 6 12.79 years), identified 3
18 months, VLMT-scores improved to values above baseline in both arms. Patients
predictors of OS. Imaging model, built on a subset of 120 patients with brain MRI data,
treated with HA-WBRT+SIB had significantly less depression compared to patients
identified a radiomic signature (RS) consisting of 4 radiomic features most predictive of
OS. Using the same subset, the Combined model (C-index: 0.728 [95% CI: 0.590–0.855])
treated with WBRT+SIB at 3 (p = 0.047) and 18 months (p = 0.048). The 12-month-tumor
outperformed Clinical (C-index: 0.62 [95% CI: 0.44–0.78]) and Imaging (C-index: 0.62 [95% control for boosted metastases was 96% in Arm A and 88% in Arm B, while for the WBRT
CI: 0.46–0.77]) models in the held-out validation cohort (n = 31). Significant features area it was 78% in both arms. Time to hippocampal tumour progression was comparable
associated with increased mortality risk in the Combined model included a higher RS, between arms (p = 0.98). After 12 months, 4% of patients in arm A and 12% in arm B had
absence of tucatinib treatment for the primary BC prior to BM development, elevated Ki-67 suffered neurological death. Conclusions: To our knowledge, this is the first prospective
expression, Black race, higher N stage, and brainstem metastases. Among these factors, trial to show that hippocampal avoidance during WBRT leads to significantly lower rates
RS, with the largest absolute coefficient in the Combined model (0.38), emerged as the of depression. The development of VLMT values after HA-WBRT+SIB and WBRT+SIB
most important predictor of OS (hazard ratio: 20.03 [95% CI: 4.92–81.48], p , 0.005). with 30 Gy in 2.5 Gy-fractions was comparable and a good recovery was observed at
Conclusions: A distinct RS from brain MRI is the strongest predictor of OS in patients with 18 months in both arms. The method showed a considerably higher intracerebral tumour
BM from HER2-positive BC, surpassing clinical factors. RS may refine risk stratification control with lower neurological mortality rates compared to historical cohorts. Clinical
and guide treatment or clinical trial prioritization. Research Sponsor: Susan G. Komen. trial information: DRKS00004598. Research Sponsor: German Cancer Aid.

2039 Poster Session 2040 Poster Session

Initial report of memory avoidance whole brain radiotherapy to treat brain Genomic predictors of brain metastases in breast cancer. First Author: Anton
metastases: A prospective phase 2 trial. First Author: Joshua David Palmer, Safonov, Memorial Sloan Kettering Cancer Center, New York, NY
Department of Radiation Oncology, Ohio State University, Columbus, OH Background: Despite therapeutic advances in metastatic breast cancer (MBC), the rising
Background: A common approach for patients with extensive brain metastases requiring incidence of brain metastases (BM) remains a major challenge, contributing to poor
radiation is hippocampal avoidance whole brain radiotherapy (HA-WBRT) prescribed with prognosis and significant morbidity. Due to the absence of consensus screening
memantine; this was proven to be efficacious based on NRG CC001. However, a subset of strategies for BM, they are often detected only after clinical symptoms emerge. There is
patients who receive HA-WBRT with memantine still experience cognitive decline. Other therefore a pressing need for predictive biomarkers to identify breast cancer patients at
brain structures with important roles in memory and cognition include the corpus cal- risk of BM. Methods: This study included 3908 patients who underwent sequencing of
losum, fornix, amygdala, hypothalamus, and pituitary; these structures all have a low primary tumor (n = 1885) or non-brain metastasis (n = 2023) with MSK-IMPACT, a custom
propensity for brain metastases and therefore can be safely spared in a radiotherapy plan tumor-normal next generation sequencing assay. First, we performed penalized logistic
without increasing the risk of relapse. A subset of patients enrolled on a Phase 2 regression on a gene level to identify alterations in extracranial metastases or primary
Randomized Controlled Trial (NCT05503251) received an advanced “memory-avoidance tumors associated with development of BM. We adjusted for multiple hypothesis testing
WBRT (MA-WBRT) approach that spared these substructures in addition to the hippo- using Benjamini-Hochberg. Lastly, we developed a lasso machine-learning (ML) model,
campus, with a primary endpoint of improved cognition compared to a historical control incorporating baseline genomic and clinicopathologic features, to predict onset and
(NRG CC001). Methods: All patients with . 15 brain metastases on a prospective clinical timing of BM from initial diagnosis (for early stage cases) or metastatic disease (for
trial, which randomized patients to either neuropsychology evaluation and intervention MBC). Each analysis was stratified by receptor status, and repeated to account for loss of
plus brain radiotherapy or brain radiotherapy alone, received MA-WBRT. Exclusion criteria heterozygosity (LOH) of tumor suppressor genes. Results: Our cohort included 528 BM
included prior WBRT, pre-existing mental disability, and metastases within the avoidance events over a median follow-up of 58 mos. Pathogenic variants in several genes were
neurocognitive substructures. All patients received 30 Gy in 10 fractions of MA-WBRT and associated with subsequent BM development. In the HR+/HER2- subset (n = 2624),
were prescribed memantine. Cognition was measured by Hopkins Verbal Learning Test- pathogenic variants in the following genes portended the onset of BM: RB1 (OR 2.59 [1.39
Revised, Controlled Oral Word Association Test, and Trail Making Test A/B, with cognitive - 4.81], q = 0.011), NF1 (OR 2.22 [1.21 - 4.06], q = 0.039), TP53 (OR 1.91 [1.43 - 2.54], q ,
decline defined as decline on at least one assessment using reliable change index (same 0.001), PIK3CA (OR 1.47 [1.21-4.06], q = 0.028). Pre-existing LOH of RB1, in the absence
tests and definition as NRG CC001). Results: Between August 2022 and May 2024, 29 of an RB1 functional alteration, was associated with BM development (OR 1.37 [1.03 -
patients received MA-WBRT. Baseline characteristics included a median KPS of 80 (IQR 1.83], q = 0.090). TP53 LoF .OR 5.14 [2.21 - 11.9], q , 0.001) was enriched in the BM
70, 90), median age of 64 (IQR 54, 69), 62% female patients, and a plurality of patients group in HER2+ tumors, while amplification of CDKN2A (OR 11.6 [2.44 - 55.5], q = 0.01) or
with a primary lung cancer (48%), The median overall survival or time to last follow up was EGFR (OR 4.60 [1.54 - 13.8], q = 0.03) were enriched in TNBC. TP53 emerged as an
7.9 months. The three-month decline in neurocognitive function comparing the control important feature across all receptor subtypes in our machine-learning model; RB1 LoF
and intervention groups for patients receiving MA-WBRT was 15.4% and 18.8%, re- was also selected as an important feature in the HR+/HER2- group. Validation of the ML
spectively (p = 0.39). There was one failure in the right fornix 10 months after enrollment, model in an external cohort will be presented at the meeting. Conclusions: In a large
but this was associated with concurrent distant intracranial failure outside the memory cohort of genomically profiled breast cancer samples, we found several biologically
avoidance zone. Conclusions: The cognitive decline rate of approximately 17% at three plausible candidates for molecular harbingers of BM. For instance, the recurrent in-
months for patients receiving MA-WBRT compares favorably to a 3-month cognitive volvement of genes involved in cell cycle regulation (RB1, CDKN2A, TP53) has been
decline rate of 50% seen on NRG CC001. Additionally, MA-WBRT does not appear to implicated as candidates for BM tropism in other cancer types. Our approach also
significantly increase the risk of intracranial failure. Further evaluation of the delayed uncovers several alterations for which targeted therapies exist or are actively in de-
impact ( . 6 months) of MA-WBRT on cognitive function will be reported when data are velopment (NF1, PIK3CA). Our clinically actionable multimodal model of BM risk is poised
available. A direct comparison of MA-WBRT plus memantine vs. HA-WBRT plus mem- to facilitate the development of early detection strategies and guide-high risk patient
antine is forthcoming with a randomized phase 3 trial. Clinical trial information: selection for novel clinical trials to intercept this devastating complication. Research
NCT05503251. Research Sponsor: None. Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
122s CENTRAL NERVOUS SYSTEM TUMORS

2041 Poster Session 2042 Poster Session

Characterizing functional connectivity in brain tumor patients. First Author: AI-driven transcriptomic classification of glioblastoma: Associations with
Alexander B. Remsik, University of Wisconsin System, Madison, WI survival and tumor microenvironment. First Author: Juan Manuel Fernandez-
Background: Brain tumors (affecting 25,500 individuals in the US alone) cause unique Mu~ noz, SphereBio, Mendoza, Argentina
changes to neural function and connectivity resulting in impairments in various Background: Glioblastoma (GBM) is the most lethal primary brain tumor in adults,
functional domains, including cognition. Methods: Between 2012 and 2018, as part of with a median survival of ~15 months despite current therapies (surgery, radiation,
their preoperative surgical planning, 91 brain tumor patients with tumors in right (R) or temozolomide). Advances like immune checkpoint inhibitors, anti-angiogenic agents,
left (L) temporal (T), parietal (P) or frontal (F) lobes received resting-state functional and tumor vaccines have shown suboptimal results. The 2021 WHO classification
magnetic resonance imaging (rs-fMRI) and the COWAT verbal fluency (VF) test. UW highlights molecular markers (e.g., IDH, MGMT) for better stratification, but these fail to
Hospital and Clinics uses 1.5 T (112 axial slices, 1.0 3 1.0 3 1.5 mm) and 3 T (136 axial fully capture tumor microenvironmental dynamics. Using an AI-driven transcriptomic
slices, 1.0 3 1.0 3 1.2 mm) GE MRI scanners and includes high-resolution 3D BRAVO approach, we identified novel prognostic subtypes in IDH-wildtype GBM, aiming to refine
T1-weighted imaging, and rs-fMRI scans (eyes closed, 28 axial slices, stratification, enhance understanding of tumor biology, and guide personalized ther-
3.7533.7535.0mm) were also acquired during this imaging protocol. These individuals apeutic strategies. Methods: We accessed microarray data from The Cancer Genome
were compared to each other and to 40 age-matched non-tumor controls (Cs). Atlas (TCGA) (n=353 newly diagnosed, IDH-WT GBM) for a training set and RNA-seq data
Results: Groups were similar in age, and sex (p.0.05), but different in education from the Chinese Glioma Genome Atlas (CGGA) (n=170 primary and n=106 recurrent
(p=0.028) and VF scores (p=0.001). There were significant differences in post-hoc p- tumors) for validation. A proprietary SphereBio machine learning–based algorithm was
values when comparing VF scores between Cs and RT tumor patients (p=0.039). used to derive transcriptomic signatures with prognostic relevance. Subtypes were
Qualitative observations indicate Cs are more integrated have greater network strength assessed via Kaplan–Meier analyses in the training cohort and tested in both primary
and connectivity higher transitivity, efficiency, and modularity index compared to pa- and recurrent validation cohorts. Immune/stromal infiltration was quantified using a
tients. There were significant differences in VF scores between Cs, LF (p=0.001), RF tumor deconvolution tool (DA_505), and pathway enrichment (GAGE) was performed on
(p=0.012), RT (p=0.021), RP (p=0.015). Patients show hypo-frontality of hubs and a the validation sets. Results: AI-driven clustering revealed three transcriptomic subtypes
unique cerebellar module. LF patients showed correlation to transitivity at 25% with significant survival differences in both the training (p,0.0001) and primary val-
(R=0.404, p=0.041), and patients with RP tumors had significant correlations to idation (p=0.0004) cohorts. In the recurrent cohort, a similar survival trend by subtype
transitivity (R=0.654, p=0.029) and global efficiency (R=0.607, p=0.048). Cs and L was observed, though significance was diminished (p=0.12), likely due to limited sample
tumors did not show any significant correlation to VF scores, R tumors showed cor- size and therapy-related changes. Immune/stromal deconvolution showed distinct
relation with global efficiency at 25% sparsity (R=0.374, p=0.025). All remained sig- infiltration patterns: subtypes enriched for CD4+ and CD8+ T cells correlated with
nificant after FDR correction. Conclusions: Tumor location impacts rs-fMRI-derived GT prolonged survival. Pathway enrichment analysis in both primary and recurrent tumors
network structure and VF scores, suggesting tumors in left hemisphere and frontal areas highlighted potential targets involving embryogenesis, immune modulation, cell cycle,
caused the greatest impact to cognition. These methods may be used in future research and stress response. The persistence of a consistent survival trend and comparable
and clinical care to map, track, and predict functional connectivity changes resulting microenvironment and pathway patterns suggest that these transcriptomic subtypes
from brain tumor and can help inform clinicians and care trajectories. Research Sponsor: remain biologically relevant even after standard treatment. Conclusions: Our integrated
NIH (NINDS); T32CA009206, R01NS117568, R01NS123378, TL1TR002375, transcriptomic and microenvironment-focused approach identified three prognostically
R01CA264017, R01CA277728, UL1TR002373, P30CA014520; WI Partnership Program; distinct GBM subtypes, validated across independent cohorts. These findings under-
VA; BX005842-01A2; Wisconsin Alumni Research Foundation; MSN281757. score the utility of AI-driven transcriptomic signatures for personalized stratification,
with the potential to guide targeted therapeutic strategies and inform clinical trial design
in GBM. Research Sponsor: None.

2043 Poster Session 2044 Poster Session

Impact of neuroradiologists’ input on peer review meetings for CNS radio- Identification and validation of potential diagnostic plasma biomarkers for
therapy treatment planning. First Author: Abhishek Mahajan, The Clatterbridge diffuse gliomas by multiplex immunoassays. First Author: Miyo K. Chatanaka,
Cancer Centre NHS Foundation Trust, Liverpool, Merseyside, United Kingdom University of Toronto, Toronto, ON, Canada
Background: Evidence shows radiologists’ involvement improves planning accuracy, especially in complex Background: Diffuse gliomas are aggressive malignant tumors with poor prognosis. The
anatomical areas. However, their participation remains limited. This prospective observational study evaluates current standard of care includes measurement of molecular biomarkers in biopsy samples.
the impact of neuroradiologists’ input on changes to radiotherapy (RT) plans during peer review meetings.
Methods: Data were collected from 205 patients with CNS tumours planned for radiotherapy between May 2022
One unmet clinical need is to identify non-invasive biomarkers that may be used for dif-
and October 2023. We recorded demographics, diagnostic and RT planning scans, and therapy received. All ferential diagnosis of gliomas from other brain tumors. Pre-clinical and clinical validation of
images were reviewed by a neuroradiologist, with RT changes classified as major (affecting cure or disease such biomarkers could eliminate the need for biopsy, and support the implementation of
control) or minor (affecting target volumes or organs at risk). Summary statistics were calculated, and Pearson more personalized and/or emerging treatments and the earlier enrolment of patients into
chi-squared tests assessed whether changes in RT plans varied by tumour type, time since diagnosis, and clinical trials. Our objective is to use multidimensional proteomics to identify and validate
neuroradiologist findings. Data were analyzed using STATA SE v.17. Results: Of 205 patients, 81 (40%) had
gliomas, 77 (38%) had brain metastases, 23 (11%) had meningiomas, 17 (8%) had schwannomas, and 7 (3%) had
potential plasma biomarkers for glioma management. Methods: We used the proximity
pituitary tumours. The mean age was 60 years (SD 14), 56% were male, and 36% were treatment-naı̈ve. All but extension assay from Olink Proteomics to analyze 3,000 proteins in plasma of patients with
one had MRI scans, 128 (62%) had CT scans, and 9 (4.4%) had PET scans. The median number of diagnostic diffuse gliomas and meningiomas (as controls). By data visualization, we identified several
scans per patient was 2 (IQR 2-3), and all had two planning scans. The median interval between diagnosis and RT plasma proteins that were increased or decreased in gliomas in comparison to meningi-
planning scans was 35 days (IQR 21-61). Disease progression was observed in 67 (33%) patients. Major and omas. Several candidate markers were selected for validation with an independent set of
minor changes to RT plans were reported in 35 (17%) and 78 (38%) patients, respectively. A higher proportion of
retrospectively collected samples by using quantitative research-use-only electro-
RT plans changed for brain metastasis (26% & 40%) and glioma (11% & 43%). Changes were not associated with
tumour type (p = 0.07) or time since diagnosis (p = 0.12), but were significantly associated with neuroradiologist chemiluminescence assays available from Meso Scale Discovery. In the validation set,
findings (p , 0.0001). Conclusions: Neuroradiologists’ assessments led to major and minor changes in RT which included longitudinal data from patients, patient information included biopsy-
plans, regardless of tumour classification or interval since diagnosis. This expertise can enhance RT plan requiring molecular tumor abnormalities such as IDH1 status, ATRX expression, MGMT
accuracy, improving patient outcomes. Research Sponsor: None. promoter methylation, CDKN2A/B/p16 status, V1p 19q co-deletion and NF1 status. In the
Changes to radiotherapy treatments as per RCR categories and by tumour type, time interval, and neuroradiologist validation stage, we focused on diffuse gliomas. Results: In the discovery phase, asso-
findings. ciations between proteins were plotted to determine potential predictive ability for dis-
Changes to radiotherapy plan criminating diffuse gliomas vs. meningiomas. A partitioning algorithm was fit to determine
Major change Minor change No change the optimal combination of GFAP (the strongest biochemical marker), age and sex, as well
(N= 35) (N= 78) (N=92) P value*
as with other candidate proteins. Differential expression was seen for a few other proteins
Total 17 % 38% 45%
Tumour classification 0.07
such as NEFL, PROK1, FABP4, MMP3 and LMOD1. In the cross-sectional validation phase,
Glioma 9 (11.1) 35 (43.2) 37 (45.7) we verified strong associations between GFAP and FABP4 plasma concentration and GBM,
Meningioma 3 (13.0) 7 (30.4) 13 (56.5)
Schwannoma 2 (11.8) 3 (17.7) 4 (57.1)
astrocytomas, oligodendrogliomas and meningiomas, where these markers could differ-
Pituitary tumour 1 (14.3) 2 (28.6) 4 (70.6) entiate between the groups. Within diffuse gliomas, NEFL, GFAP, FABP4 and IL13 were
Metastasis
Time interval, weeks
20 (26.0) 31 (40.3) 26 (33.8)
0.12
significantly different. Conclusions: This study highlights the potential of plasma bio-
Less than 8 27 (77.2) 61 (78.2) 59 (64.1) markers to revolutionize glioma patient management through liquid biopsy applications. The
8- 4 (11.4) 10 (12.8) 20 (21.7)
16- 1 (2.9) 3 (3.9) 7 (7.6)
strong associations observed between plasma protein concentrations and glioma subtypes
24- - 2 (2.6) 1 (1.1) support a diagnostic power that addresses a critical unmet need in neuro-oncology. More
30- - 2 (2.6) 3 (3.3)
38- 3 (8.6) - 2 (2.2)
specifically, these biomarkers can help with patient differential diagnosis at initial pre-
Neuroradiologist findings ,0.0001 sentation, with future aims to investigate the prognostic value and the possibility of acting
Stable disease 17 (48.6) 35 (44.9) 71 (77.2) as surrogates of molecular changes that are currently used for optimizing therapy. Research
Residual disease 3 (8.6) 6 (7.7) 6 (6.5)
Disease progression 15 (42.8) 37 (47.4) 15 (16.3) Sponsor: Canadian Institutes for Health Research, The Canadian Brain Foundation, Canadian
*Pearson chi-squared test.
Cancer Society Research Institute; CCS707057; National Cancer Institute; P50CA221747.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 123s

2045 Poster Session 2046 Poster Session

Association of plasma biomarkers with diagnostic molecular markers for A multicenter randomized phase III study for recurrent glioblastoma com-
potential diagnosis and prognosis of diffuse gliomas. First Author: Leonardo paring bevacizumab alone with dose-dense temozolomide followed by
Macedo Filho, Penn State College of Medicine, Hershey, PA bevacizumab: JCOG1308C. First Author: Motoo Nagane, Department of Neuro-
Background: Diffuse gliomas were recently reclassified based on the 2021 WHO surgery, Kyorin University Faculty of Medicine, Tokyo, Japan
Classification criteria. Several molecular changes which carry diagnostic and prognostic Background: Temozolomide (TMZ) is an alkylating agent commonly used as the standard
power have been added to the classification parameters, including IDH1 mutation and therapy for newly diagnosed glioblastoma (GBM), with the DNA repair enzyme O6-
MGMT promotor methylation. To characterize these molecular changes, however, an methylguanine-DNA methyltransferase (MGMT) serving as a key prognostic and predictive
invasive biopsy is required. Our goal was to examine the relationship between seven factor. Despite treatment, GBM almost always recurs with limited therapeutic options,
plasma biomarkers for diffuse glioma and the established molecular changes and leading to poor prognosis. Since MGMT is consumed during the repair of TMZ-induced O6-
delineate if these markers can be used as surrogates of these molecular changes. methylguanine lesions in DNA, dose-intensified TMZ regimens are designed to deplete
Methods: Seven candidate markers, namely glial fibrillary acidic protein (GFAP), MGMT, thereby enhancing tumor sensitivity to TMZ. Bevacizumab (BEV), an anti-VEGF
neurofilament light (NEFL), matrix metalloproteinase 1, 3, 9 (MMP1, MMP3, MMP9), total agent, has shown efficacy in recurrent GBM (re-GBM), but no effective therapies exist after
Tau (tTau) and fatty acid binding protein 4 (FABP4) were evaluated by quantitative BEV failure. Introducing an active agent before BEV may improve outcomes. To test this, we
research-use-only electrochemiluminescence assays available from Meso Scale Dis- conducted a multicenter, phase III study comparing BEV monotherapy with dose-dense TMZ
covery by comparing the protein concentration distribution with non-parametric Wil- (ddTMZ) followed by BEV in re-GBM. Methods: Patients (pts) aged 20-75 years with
KPS $60 and histologically confirmed GBM at first recurrence were enrolled from 31
coxon rank sum tests and multiple testing adjustment. The molecular markers tested
Japanese hospitals. Participants were randomized to BEV monotherapy (10 mg/kg every
were IDH1, MGMT promotor and ATRX. The discovery cohort consisted of 49 IDH1
2 weeks; arm A) or ddTMZ (120-150 mg/m2, 7 days on/7 days off) followed by BEV at
mutant (39%) and 77 IDH1 wildtype (61%) gliomas. Among this retrospective cohort
progression (arm B). Treatment continued until progression or unacceptable toxicity. The
were 103 primary samples (collected at diagnosis) and 23 recurrent samples (collected primary endpoint was overall survival (OS). A planned sample size of 146 pts provided 70%
at time of recurrence). The retrospective validation cohort consisted of 36 IDH1 mutant power to detect a hazard ratio (HR) of 0.73 (median OS (mOS): 8 vs. 11 months) at a one-
(22%) and 129 IDH1 wildtype (78%), with 64 primary samples and 76 recurrent samples. sided alpha of 10%. MGMT promoter methylation and IDH mutation status were analyzed.
Results: Several of the proteomic markers showed significant associations with genetic Results: From July 2016 to April 2022, 146 pts (73 per arm) were randomized. MGMT
markers at an adjusted significance level of P , 0.05. For IDH1 status, the strongest promoter methylation was observed in 78 pts, while 49 were unmethylated. IDH1 mutations
association was with NEFL, with IDH1 wildtype samples showing higher levels of the were identified in 8 of 129 pts to be tested. The mOS was 11.0 months (95% CI: 9.0-12.8) in
protein. For ATRX expression, high FABP4 was correlated with ATRX retention. As arm A and 10.8 months (95% CI: 8.6-12.5) in arm B, with no significant difference (HR 0.922,
expected, survival analysis based on molecular markers yielded that IDH1 status was 95% CI: 0.655-1.297, one-sided p = 0.320). No significant OS difference was observed
most predictive of survival both in primary tumors and recurrent tumors. MGMT pro- between arms based on MGMT methylation status. The median progression-free survival
motor methylation was predictive of survival in primary cases but not recurrent cases. (PFS) was 4.0 months (95% CI: 3.8-5.7) in arm A and 2.0 months (95% CI: 1.9-2.1) in arm B
When combining the genetic markers with protein concentrations, we were able to see (HR 1.632, 95% CI: 1.168-2.281). Most pts in arm B exhibited progression at their first MRI.
some improvement in survival prediction. Conclusions: We demonstrate that some From the start of BEV treatment, mOS was 10.8 months (95% CI: 8.8-12.6) in arm A, and
plasma biomarkers, particularly NEFL and FABP4, show significant associations with 8.0 months (95% CI: 6.1-9.1) in arm B. Grade 3-4 adverse events included hypertension
key molecular changes in diffuse gliomas, including IDH1 status and ATRX retention/ (19.4%) in arm A and lymphopenia (52.1%) and leukopenia (8.2%) during ddTMZ in arm B.
loss. Future research will determine whether these proteomic markers can serve as Grade 4 toxicities were rare. Conclusions: While ddTMZ was well-tolerated, this study failed
surrogates for molecular alterations and assist in potentially improved diagnosis and to demonstrate a survival benefit for ddTMZ followed by BEV in re-GBM. BEV remains the
monitoring of diffuse gliomas. Research Sponsor: None. preferred treatment at first recurrence. Further research is needed to develop effective
therapies beyond the current standard for re-GBM. Clinical trial information: NCT02761070.
Research Sponsor: Japan Agency for Medical Research and Development; 17824890.

2047 Poster Session 2048 Poster Session

Use of brain protein I3 (BRI3) to predict disease fate in glioblastoma. First Effect of armed oncolytic adenovirus on immunotherapy for primary and
Author: Ifeanyichukwu Ogobuiro, The University of Miami Sylvester Comprehensive metastatic brain tumors. First Author: Jiasen He, The University of Texas MD
Cancer Center, Miami, FL Anderson Cancer Center, Houston, TX
Background: Genome-wide characterization has illuminated the molecular complexity of Background: Oncolytic viruses have shown promise in clinical trials for solid tumors,
human gliomas. Genetic alterations help predict the clinical behavior of gliomas, but including glioma and melanoma, but only a subset of patients benefits. We previously
variability persists. Accordingly, there is a need to expanding molecular signatures that showed that arming the oncolytic adenovirus Delta-24-RGD with OX40L can enhance
refine prognostication. The Brain Protein I3 (BRI3) gene, localized on chromosome 7, is antitumor immunity. To further boost efficacy, we developed Delta-24-RGDOX-IL15, co-
associated with high-grade glioma, yielding the highest hazard ratio among all high-risk expressing OX40L and IL-15, and tested it in preclinical models of primary and met-
genes in an in-silico glioma model. Given GBM’s known chromosome 7 gain and BRI3’s astatic brain tumors. Methods: To evaluate IL-15 and its receptor (IL15RA) expression
chromosomal location, we hypothesized that gene dosage gains and overexpression in patients with glioma and melanoma, we conducted gene expression and survival
serve as prognostic biomarkers. Methods: We used the Rembrandt (n = 461 patients) and analysis using the GEPIA web server, integrating RNA sequencing data from TCGA and
TCGA low-grade glioma (LGG) and GBM (n = 1,148 patients) databases for multi-omic GTEx. Transgene expression in Delta-24-RGDOX-IL15 was assessed via flow cytometry
analyses. RNA sequencing (Illumina HiSeq) and DNA methylation profiles (Illumina 450K) and ELISA, while viral potency was evaluated using replication and cell viability assays.
were analyzed in a combined LGG and GBM cohort, with high/low expression and hyper/ Anti-tumor activity was tested in syngeneic intracranial models derived from mouse
hypomethylation defined by median values. CNV analysis (GISTIC 2.0) classified 2 copies diffuse midline (DMG) glioma and melanoma cell lines in C57BL/6 mice, both of which
as gene-copy neutral and . 2 as gene dosage gain. Somatic mutation data (SNPs/ expressed GD2 and luciferase. Tumor growth was monitored with bioluminescent
INDELs) were derived from whole-exome sequencing to decipher IDH-wt and IDH-mutant imaging, survival with Kaplan-Meier analysis, and immune profiling of the tumor mi-
gliomas. Univariate and adjusted Cox models, Kaplan-Meier estimates, and receiver croenvironment using flow cytometry. Results: GEPIA analysis showed that melanoma
operating characteristic (ROC) curve analysis were performed. Results: Of 461 Rem- had higher expression of IL-15 and IL-15RA compared to glioma. In melanoma patients,
brandt patients, 47.29% were GBM, 31.89% astrocytoma, 14.53% oligodendroglioma, and higher expression of IL-15RA or IL-15 was linked to better overall survival (P , 0.005),
6.29% normal brain. Among 1,148 TCGA LGG/GBM patients, 551 had complete molecular while no survival difference was found in patients with glioma. Delta-24-RGDOX-IL15
and clinical data, with 28% IDH-wt status, 49% MGMT hypermethylation, and 36.66% BRI3 infected and co-expressed OX40L and IL-15 effectively in mouse glioma and melanoma
gene dosage gains. BRI3 expression was significantly higher in GBM, grade IV tumors, cells, and induced potent oncolysis. Delta-24-RGDOX-IL15-infected tumor cells sig-
IDH-wt, and mesenchymal subtypes. Among the IDH-wt cases, 75% showed BRI3 gene nificantly enhanced the oncolysis activity of GD2 CAR T cells in culture. Intratumoral
dosage gains with significantly elevated mRNA expression. EGFR, co-amplified in 80% of injection of the virus also resulted in better tumor reduction and improved survival in
IDH-wt cases, did not affect survival (25.27 vs. 17.90 months, p = 0.346). Conversely, BRI3 C57BL6 mice with gliomas derived from mouse DMG cells while no significant toxicity
gene dosage gain correlated with worse survival (79.3 vs. 17.93 mo, p = 0.001), as did was observed. Additionally, locoregional therapy with Delta-24-RGDOX-IL15 induced a
BRI3 high vs. low expression (17.90 vs. 25.27 months, p = 0.015) and MGMT hyper- systemic inflammatory response in the tumor microenvironment, characterized by
methylation (25.27 vs. 18.63 mo, p = 0.021). Univariate analysis linked patient age (HR: increased frequency of T cells and reduced that of myeloid cells. Conclusions: Higher
2.858 [1.781–4.586], p , 0.001), MGMT hypermethylation (HR: 2.632 [1.307–5.301], expression of IL-15/IL-15RA is associated with better survival in patients with mela-
p = 0.007), BRI3 high expression (HR: 1.938 [1.131–3.320], p = 0.016) and BRI3 gene noma. Delta-24-RGDOX-IL15 demonstrates potent oncolytic activity in both glioma and
dosage gain (HR: 2.504 [1.425–4.398], p = 0.001) to worse OS. Adjusted multivariate Cox melanoma cell lines. In the intracranial brain tumor mouse model, it exhibited promising
regression confirmed BRI3 gene dosage gain, age, and MGMT methylation as independent anti-tumor effects with enhanced T cell stimulation and minimal toxicity. Delta-24-
OS predictors in IDH-wt cases. ROC analysis revealed stronger prognostic perfor- RGDOX-IL15 is a promising candidate for combination with cellular therapies in both
mance for BRI3 gene dosage gain than MGMT hypermethylation (AUC: 0.737 vs. 0.616, primary and metastatic brain tumors. Research Sponsor: ChadTough Defeat DIPG
p , 0.001) in IDH-wt cases. Conclusions: Elevated BRI3 gene dosage and expression Foundation.
portend poor prognosis and could be incorporated into models predicting disease fate in
GBM. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
124s CENTRAL NERVOUS SYSTEM TUMORS

2049 Poster Session 2050 Poster Session

A phase I clinical trial on combined (neo-)adjuvant intravenous plus intra- Effect of 18F-DOPA-PET and advanced MRI on treatment response assess-
cranial administration of ipilimumab and nivolumab in recurrent glioblas- ment in IDH1/2-mutant gliomas treated with IDH inhibitors. First Author:
toma (NEO-GLITIPNI). First Author: Iris Dirven, Department of Medical Oncology, Diego Martı́n Prost, Hopital La Pitié Salpetriere, Paris, France
Laboratory for Medical and Molecular Oncology (LMMO), Translational Oncology Background: Small-molecule inhibitors targeting IDH1/2-mutant proteins (IDHi) have
Research Center (TORC), Vrije Universiteit Brussel (VUB), Universitair Ziekenhuis Brussel shown promise as treatments for IDH1/2-mutant gliomas. However, accurate as-
(UZ Brussel), Brussels, Belgium sessment of response using morphological magnetic resonance imaging (MRI) mea-
Background: Intravenous (IV) administration of ipilimumab (IPI) and nivolumab (NIVO) has surements remains difficult, and the potential of PET imaging with radiolabeled amino
shown limited activity in recurrent glioblastoma (rGBM). Intracerebral (iCer; within the brain acids in this context is yet to be explored. Here, we investigated 3,4-Dihydroxy-6-[18F]-
tissue lining the resection cavity) and intracavitary (iCav; through an Ommaya reservoir) fluoro-L-phenylalanine PET (¹⁸F-DOPA-PET) and MRI responses in IDH1/2-mutant gli-
administration (admin) of IPI and NIVO was proven to be safe and resulted in promising oma patients receiving IDHi. Methods: IDH1/2-mutant glioma patients receiving IDHi as
survival outcomes (Duerinck et al. Neuro-Oncol 2024). Adding a neoadjuvant (NEOADJ) part of trials or expanded access programs were included. Patients had pre- and post-
treatment phase to iCer/iCav IPI/NIVO may further improve outcome. Methods: In the Neo- treatment MRI and ¹⁸F-DOPA-PET. Centralized evaluations included 2D/3D measure-
Glitipni trial (NCT06097975), a single center, phase I clinical trial, patients (pts) with resectable
ments on T2-weighted FLAIR images, T1-post contrast, perfusion, and diffusion imaging
rGBM (WHO grade 4, IDH wild type) who progressed after radiotherapy and temozolomide,
for MRI, and metabolic tumor volume (MTV), total lesion glycolysis (TLG), and tumor-to-
with a baseline ECOG performance status of 0-2 and #8 mg methylprednisolone daily, received
2 NEOADJ cycles of IV IPI 1 mg/kg + NIVO 3 mg/kg followed by maximal safe resection (MSR)
background ratios (TBRs) for ¹⁸F-DOPA-PET. Disease response evaluation using volu-
in week 5 with iCer admin of IPI 5 mg + NIVO 10 mg and iCav admin of IPI 1 mg + NIVO 10 mg. metric assessments, RANO 2.0 and PET RANO 1.0 criteria were compared and con-
The adjuvant phase consists of biweekly postoperative iCav admin of IPI 1 mg + NIVO 10 mg fronted to outcomes. Results: From 2021 to 2024, 10 patients with IDH1/2-mutant
and IV NIVO 240 mg for 12 cycles, followed by monthly NIVO 480 mg IV maintenance for up to glioma (3 astrocytoma, 7 oligodendroglioma) receiving IDHi (4 ivosidenib, 6 vorasidenib)
two years. Results: 5 pts (4 male, median age 57 years (44-65); 1st recurrence in 3 pts) were were analyzed. Significant reductions in 18F-DOPA-PET parameters including TBRmean,
enrolled. All pts received the 1st and 4 pts also the 2nd NEOADJ dose of IV IPI/NIVO. Out of the 5 TBRmax, and MTV were observed in 8/10 patients, aligning with observed changes in
pts, 3 were not amenable to MSR with iCer/iCav IPI/NIVO admin according to the protocol perfusion and diffusion imaging. Seven partial responses and one complete response
because of disease progression during the NEOADJ treatment phase and required cortico- were identified using ¹⁸F-DOPA-PET, while both volumetric and standard 2D morpho-
steroids (1 pt in week 2, 2 pts in week 4). Two pts successfully underwent MSR with iCer/iCav logical MRI assessments indicated stable disease as best response. PET response was
admin of IPI/NIVO per protocol. One pt initiated adjuvant treatment with iCav IPI/NIVO and IV correlated with prolonged tumor control. Conclusions: This study highlights the po-
NIVO. There were no unexpected adverse events (AE). Two pts experienced an immune-related tential of ¹⁸F-DOPA-PET and advanced MRI sequences as valuable complements to
AE that required corticosteroids and interruption of study treatment (grade 4 hepatitis in 1 pt, standard RANO 2.0 MRI evaluations for assessing treatment response in glioma patients
onset 8 days after MSR and grade 2 colitis in 1 pt, onset 28 days after MSR). One pt developed a undergoing IDHi therapy. Research Sponsor: None.
thyroiditis during the NEOADJ treatment phase and 2 pts experienced a grade 3 treatment
related AE that was not immune-related (seizure and Ommaya reservoir infection). None of the
rGBM were characterized by a high tumor mutational burden on next generation sequencing.
Gene expression profiling, and pharmacokinetic analysis of NIVO and IPI in the cerebrospinal
fluid and blood are ongoing. After a median follow-up of 15 weeks (9-35w) all pts are alive, one
pt remains free of progression (median progression free survival: 4.3 weeks).
Conclusions: Four weeks of NEOADJ IV IPI/NIVO (comprising 2 admin) is safe, but symp-
tomatic disease progression was observed in 3 out of 5 rGBM pts prior to the planned MSR with
iCer/iCav IPI/NIVO admin in week 5. Therefore, the trial is being amended by shortening the
NEOADJ treatment phase to 2 weeks (1 admin) and planned MSR with iCer/iCav IPI/NIVO
admin in week 3. Clinical trial information: NCT06097975. Research Sponsor: None.

2051 Poster Session 2052 Poster Session

Phase I/II study of maintenance therapy with metformin and temozolomide Modification of the novel RANO clinical risk score for low- and middle-
for newly diagnosed glioblastoma. First Author: Yoshitaka Narita, Department of income countries without access to MGMT methylation testing. First Author:
Neurosurgery and Neuro-Oncology, National Cancer Center Hospital, Tokyo, Japan Imdat Eroglu, Gazi University School of Medicine, Department of Medical Oncology,
Background: Glioblastoma (GBM) is an aggressive primary brain tumor with poor Ankara, Yenimahalle, Turkey
prognosis. A potential strategy for overcoming therapeutic resistance involves the Background: The RANO Resect Group developed a new risk score using simple vari-
development of novel therapies that target cancer stem/initiating cells. Our previous ables, including age, Karnofsky performance scale (KPS), RANO resection class (RRC),
research demonstrated that metformin (MF), an antidiabetic drug, induces the differ- and MGMT methylation (MGMTm), for patients with IDH-wildtype glioblastoma (GBM).
entiation of stem-like glioma-initiating cells and suppresses tumor formation via AMPK- Although this score is easy to apply and demonstrates high prognostic accuracy, routine
FOXO3 activation (Stem Cells Transl Med, 2012). We conducted a phase I/II study to testing for MGMTm remains inaccessible in many low- and middle-income countries. In
evaluate the clinical efficacy of MF combined with standard maintenance temozolomide this study, we aimed to modify the RANO risk score by excluding MGMTm.
(TMZ). Our phase I findings indicated that MF at doses of up to 2,250 mg/day combined Methods: This is a single-center, retrospective analysis of IDH-wildtype GBM patients.
with maintenance TMZ was well tolerated (Cancers, 2022). Here, we present the We applied the same scoring system established by the RANO Resect Group, excluding
complete results of the phase I/II study. Methods: Patients aged 20–74 years with MGMTm. The point (p) allocations were as follows: RRC1 = 0p, RRC2 = 1p, RRC3 = 2p,
supratentorial GBM, Karnofsky Performance Status $ 70, and a history of initial RRC4 = 5p; KPS . 80 = 0p, KPS , 80 = 3p; age , 65 = 0p, and age . 65 = 1p (age was
chemoradiotherapy with TMZ were eligible. During the phase II study, patients received not scored if RRC = 1p). The relationship between overall survival (OS) and the variables
MF monotherapy for 14 days, followed by six cycles of TMZ combined with daily MF (age, KPS, and RRC) was evaluated using univariate and multivariate Cox regression
(2,250 mg) for 365 days. The primary endpoint was the 1-year progression-free survival analyses. Three risk classes were defined as numerical scores derived through ROC
(PFS) rate from the initiation of chemoradiotherapy with TMZ (target; one-sided alpha analysis. The primary endpoint was overall survival, and the secondary endpoint was
10%, power 70%, threshold 1-year PFS, 27%; expected 1-year PFS, 50%, based on the progression-free survival (PFS). Results: A total of 119 patients were included in the
historical EORTC/NCIC study (Stupp et al, 2005)). Results: From 2021–2023, 22 pa- study. Of these, 100 patients received chemoradiotherapy with temozolomide followed
tients were enrolled in 5 hospitals and 21 patients received TMZ combined with daily MF. by adjuvant temozolomide (CRT-TMZ), while 13 received CRT only, 1 patient received
The cohort included 12 men and nine women, with a median age of 50 years (32–69 temozolomide only, and 1 received radiotherapy only. Four patients were unable to
years). According to the WHO 2016 classification, the initial histology revealed 18 IDH- undergo any treatment. RRC, age, and KPS classifications were all significantly as-
wild-type and 3 IDH-mutant GBMs. The 1-year PFS was 47.6 % (90% CI; 29.2–64.0), sociated with overall survival in both univariate and multivariate analyses (p ,
achieving the primary endpoint. The 2-year overall survival rate was 54.5%. Grade $ 3 0.001).Based on ROC curve analysis, three risk classes were identified: low risk (0–1
adverse events included lymphocytopenia (19%), thrombocytopenia (4.8%), appetite points, n:47, 39.5 %), intermediate risk (2–3 points, n:27, 22.7 %), and high risk ($4
loss (4.8%), body weight loss (4.8%), nausea (4.8%), and seizures (4.8%). points, n:45, 37.8 %). The median OS was 35.5 months (95% CI: 21.1–50) for the low-risk
Conclusions: Maintenance therapy with 2,250 mg/day of MF combined with TMZ for group, 16 months (95% CI: 9.9–22.1) for the intermediate-risk group, and 5 months (95%
newly diagnosed GBM is promising. A phase III study comparing MF combined with TMZ CI: 3.8–6.1) for the high-risk group (p , 0.001). Similarly, the median PFS was
vs. TMZ alone for the treatment of GBM is planned. Clinical trial information: 16.3 months (95% CI: 12.3–20.2) for the low-risk group, 9.9 months (95% CI: 7.2–12.6)
jRCTs031200326. Research Sponsor: Japan Agency for Medical Research and Devel- for the intermediate-risk group, and 4.1 months (95% CI: 3.4–4.8) for the high-risk group
opment; AMED 21ck0106623h0002. (p , 0.001). Conclusions: The modification of the novel RANO clinical risk score by
omitting MGMTm remains highly prognostic for patients with IDH-wildtype GBM. Since it
relies on very basic parameters and is easy to use, the modified RANO score can serve
as a practical tool in countries where MGMTm testing is inaccessible. Research Sponsor:
None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 125s

2053 Poster Session 2054 Poster Session

The impact of IDH mutation and 1p/19q codeletion on immune-checkpoint Factors influencing clinical trial enrollment in glioblastoma patients: A
inhibitor efficacy in recurrent gliomas. First Author: Shameel Shafqat, Mayo Clinic retrospective study at the University of Vermont Medical Center. First Au-
Comprehensive Cancer Center, Rochester, MN thor: Leena Ziane, Larner College of Medicine, Burlington, VT
Background: Recurrent gliomas are highly aggressive brain tumors, often resistant to Background: The standard of care for patients with glioblastoma (GBM) is a combi-
conventional treatments. Immune checkpoint inhibitors (ICI) have emerged as promising nation therapy of radiation and temozolomide. NCCN guidelines recommend clinical
therapeutic agents by targeting tumor cells through immune modulation. However, trials should be offered to glioblastoma patients when appropriate. This study aims to
clinical trials have demonstrated limited efficacy in recurrent gliomas. This study aimed investigate the number of patients enrolled in clinical trials at the University of Vermont
to identify potential factors influencing treatment efficacy of ICIs in recurrent gliomas. Medical Center (UVMMC) and/or referred to larger academic centers for trials. This study
Methods: This retrospective study, conducted across the Mayo Clinic following IRB sought to identify and analyze characteristics of patients enrolled in trials at UVMMC and
approval, included patients $ 18 years diagnosed with adult-type diffuse gliomas. other outside academic centers. Methods: A retrospective review was undertaken of the
Eligible patients received treatment with at least 2 cycles of ICI for recurrent glioma electronic health records of ninety patients with GBM from 2021 to 2023 who were
between 2014 – 2024. Patients treated with ICIs as initial therapy were excluded. followed at UVMMC. Patient age, gender, and educational and employment status were
Clinical, radiographic, histological, and molecular data were analyzed, with missing collected. We assessed all comers who were offered a clinical trial, patients who
information excluded. Responders to ICI were defined as patients who did not meet proceeded to enroll in a clinical trial, and factors that influenced enrollment. We also
iRANO criteria for progressive disease based on first radiographic response assessment assessed location where the trial was conducted (UVMMC vs Outside Center) and the
(and confirmatory follow up imaging as needed for possible pseudo-progression). trial interventions provided at the various centers. Results: We assessed 90 patients
Survival outcomes [Progression-Free Survival (PFS) and Overall Survival (OS)] and diagnosed with GBM at UVMMC from 2021-2023. 87% of all patients were offered the
potential predictive variables were analyzed using the Kaplan-Meier method and Cox- opportunity to enroll in a clinical trial. 17% of patients who were offered a trial suc-
Regression Analyses. Results: 67 patients met eligibility criteria (mean age: 45.1 6 cessfully enrolled. Amongst enrolled patients, 62% completed their clinical trial at
15.0 years; 64.2% male; 94% white). 64 (95.5%) patients received Pembrolizumab, 2 (3%) UVMMC while 38% were referred and treated at nearby academic centers in the New
Nivolumab, and 1(1.5%) combined Ipilimumab/Nivolumab, with a median treatment England Area including Mass General Hospital and Dana-Faber Cancer Institute (DFCI).
duration of 2.77 (1.39 – 19.4) months. All had prior alkylating chemotherapy. The OS Intrinsic barriers to enrollment included poor Karnofsky performance scale (KPS) scores,
(from diagnosis) for IDH wildtype (IDH-WT, n = 36), IDH mutant, 1p/19q non-co-deleted MGMT negative status, presence of leptomeningeal disease, and deep tumor locations
(IDH-MUT, n = 17) and IDH mutant, 1p/19q co-deleted (OLIGO, n = 14) gliomas were 3.1, precluding resection. Extrinsic factors included distance to academic centers, trial
9.2, and 18.6 years, respectively. The median PFS from time of ICI was 2.23 (0.69 – 27.3) closure to accrual, and socioeconomic status. There was a correlation between so-
months. 24 (36.9%) patients were identified as Responders. PFS was not significantly cioeconomic status and trial enrollment. Of the thirty-eight patients identified with
different between patients with IDH-MUT and IDH-WT gliomas (2.30 vs 2.07 months, higher educational attainment (college or higher), 87% were offered a clinical trial and
p = 0.593). However, patients with OLIGO gliomas had a significantly higher PFS 18% eventually enrolled. Among the non-college educated group of thirty-six, 92% were
compared to IDH-WT gliomas (5.16 vs 2.07 months, p = 0.021). The proportion of offered a trial. However, only 14% enrolled. Further analysis on age, gender, and clinical
responders was greatest in OLIGO gliomas, however, did not reach statistical signifi- trial intervention will be reported in future publications. Conclusions: Expanding access
cance (IDH-WT, 31.4%; IDH-MUT, 29.4%; OLIGO, 61.5%, p = 0.120). Overall PFS was not to clinical trials is critical to optimizing care for GBM patients. Our findings highlight
impacted by patient age, sex, and extent of initial resection. When analyses were limited access to academic centers is crucial to clinical trial enrollment. Further studies an-
to Responders, the PFS for IDH-WT, IDH-MUT and OLIGO gliomas were 5.75, 7.01 and alyzing modifiable barriers to clinical trial accrual are needed. Research Sponsor: None.
10.8 months, respectively (p = 0.434). Conclusions: Patients with recurrent OLIGO
gliomas may have a longer PFS with ICI therapy compared with recurrent IDH-WT and
IDH-MUT gliomas. However, there is significant variability in ICI treatment efficacy
between patients. Further molecular profiling is in progress to evaluate additional
predictive biomarkers of response. Research Sponsor: None.

2055 Poster Session 2057 Poster Session

Safety and tolerability of olaparib, temozolomide, and pembrolizumab in a Efficacy and safety of bevacizumab in combination with radiotherapy and
phase 2 trial in patients with progressive glioblastoma. First Author: temozolomide in patients with glioblastoma: A meta-analysis and meta-
Luis Nicolas Gonzalez Castro, Dana-Farber Cancer Institute, Boston, MA regression of randomized controlled trials. First Author: Ibrahim Khalil, Dhaka
Background: Glioblastoma is the most aggressive primary brain tumor of adults, with Medical College & Hospital, Dhaka, Bangladesh
patients obtaining limited survival benefit from standard-of-care therapies. We are Background: Glioblastoma (GBM), the most common primary brain tumor in adults, has a poor
conducting a phase 2, surgical window-of-opportunity study evaluating the combination prognosis despite standard treatment. This meta-analysis evaluates the efficacy and safety of
of pembrolizumab (anti-PD-1 immunotherapy), olaparib (PARP inhibitor), and temo- Bevacizumab, a VEGF inhibitor, when combined with radiotherapy and Temozolomide in terms of
zolomide (alkylating chemotherapy) in patients with progressive glioblastoma. Olaparib progression-free survival (PFS), overall survival (OS), and treatment-related adverse events.
and temozolomide have the potential of synergistically enhancing the tumor’s sus- Methods: A systematic search of PubMed, Cochrane Library, Embase, and [Link]
identified randomized controlled trials (RCTs) evaluating Bevacizumab with radiotherapy and
ceptibility to immune checkpoint immunotherapy through DNA damage and activation
Temozolomide. Ten RCTs involving 4,425 patients (2,249 in the Bevacizumab arm and 2,176 in
of immune pathways, potentially amplifying the efficacy of pembrolizumab. We the control arm) met the inclusion criteria. A random-effects model calculated mean differences
provide a preliminary report on the safety and tolerability of the olaparib, temozolomide, (MD) for continuous outcomes and risk ratios (RR) for dichotomous outcomes with 95%
and pembrolizumab combination. Methods: We enrolled patients with radiographically confidence intervals (CI). Results: Bevacizumab significantly improved PFS by a mean dif-
progressive glioblastoma, IDH-wildtype, MGMT promoter unmethylated, on 2mg daily or ference of 2.39 months (95% CI: 1.34 to 3.44; P = 0.0005), but no significant benefit was
less of dexamethasone. Patients participating in the safety lead-in (Cohort 1) did not observed in OS (MD: 0.46 months, 95% CI: -0.53 to 1.45; P = 0.318). The therapy increased the
require surgically-resectable disease, while those enrolled in the surgical arm (Cohort 2) risk of vascular adverse events (RR: 1.52, 95% CI: 1.10 to 2.11; P = 0.023). While trends towards
did. Treatment is provided in consecutive 42-day cycles with olaparib 200mg BID and increased hematologic adverse events (RR: 1.24, 95% CI: 0.95 to 1.60; P = 0.093) and hy-
temozolomide 50mg QD given on days 1-7 and 22-29, and pembrolizumab 400mg IV pertensive events (RR: 2.17, 95% CI: 0.91 to 5.16; P = 0.066) were observed, they did not reach
given on day 1. Adverse events (AEs), dose-limiting toxicities (DLTs), and high-frequency statistical significance. Other adverse events, including serious adverse events (RR: 1.21, 95%
toxicities ($50% occurrence) were assessed, along with dose modifications or delays CI: 0.84 to 1.76; P = 0.221), grade 3-4 thrombocytopenia (RR: 1.05, 95% CI: 0.44 to 2.52; P =
required to manage treatment-related toxicities. AEs were graded according to CTCAE 0.858), visceral perforation (RR: 1.92, 95% CI: 0.54 to 6.90; P = 0.202), and thromboembolic
v5.0 criteria, with pre-specified measures to address reversible toxicities through dose incidents (RR: 1.32, 95% CI: 0.88 to 2.00; P = 0.120), showed no significant increase. Meta-
regression analysis indicated that study-level covariates, including patient age, sex distribution,
adjustments. Results: Six patients were enrolled in the safety lead-in (Cohort 1), which
and histologic differences did not significantly influence the primary outcomes. However, MGMT
followed a 3+3 design. Grade 1–2 leukopenia, lymphopenia, and neutropenia were
methylation status demonstrated borderline significance (P , 0.1), suggesting potential
observed in 3 patients and resolved without intervention. Grade 4 neutropenia was prognostic relevance. Conclusions: Bevacizumab modestly improves PFS but not OS, with an
observed in two patients (starting in cycle 2 for one patient, and on cycle 5 for the other). increased risk of vascular toxicities. Personalized treatment strategies and further research are
This resolved with dose delays, reduction in temozolomide dose or discontinuation of essential to optimize its role in glioblastoma management. Research Sponsor: None.
temozolomide (1 patient). In Cohort 2, which is ongoing (25 patients currently enrolled),
Adverse event outcomes associated with bevacizumab in glioblastoma patients.
toxicity profiles remain consistent with those observed in Cohort 1 patients. Grade 4
Adverse Event Outcomes RR 95%CI P value
neutropenia has been observed in 1 (6%) of the enrolled patients, improving with dose Any Adverse events 1.18 [0.64,2.16] 0.364
delays and temozolomide dose reduction. Conclusions: The combination of olaparib,
Hematologic adverse events 1.24 [0.95,1.60] 0.093
temozolomide, and pembrolizumab demonstrates a tolerable safety profile in patients Any serious adverse events 1.21 [0.84,1.76] 0.221
with progressive glioblastoma. Observed hematologic toxicities are reversible with Any vascular adverse events 1.52 [1.10,2.11] 0.023
appropriate management, supporting the ongoing evaluation of this regimen to assess Any grade 3-4 thrombocytopenia 1.05 [0.44,2.52] 0.858
its efficacy and therapeutic potential. Clinical trial information: NCT05463848. Research Visceral perforation 1.92 [0.54,6.90] 0.202
Any arterial or venous thromboembolic incidents 1.32 [0.88,2.00] 0.120
Sponsor: Merck & Co. Any hypertensive events 2.17 [0.91,5.16] 0.066

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
126s CENTRAL NERVOUS SYSTEM TUMORS

2058 Poster Session 2059 Poster Session

Assessment of age in the clinical risk stratification of patients with IDH- The tumoral molecular landscape of long-term survivors with isocitrate
mutant gliomas. First Author: Michal Nisnboym Ziv, Duke University Medical Center, dehydrogenase wildtype glioblastoma: Lessons from ETERNITY (EORTC
Durham, NC 1419). First Author: Michael Weller, Department of Neurology, University Hospital
Background: Prognosis for mutant isocitrate dehydrogenase (mIDH) gliomas is Zurich, and Brain Tumor Centre, University Hospital and University of Zurich, University
influenced by tumor type, size, neurologic deficits, and age. Traditionally, patients over Hospital Zurich, Zurich, Switzerland
45 are considered high-risk, prompting consideration of early chemoradiation. Recent Background: Predictors of long-term survival in patients with isocitrate dehydrogenase
promising results with the mIDH inhibitor vorasidenib challenge traditional age-based (IDH)-wildtype glioblastoma remain incompletely understood. ETERNITY (EORTC 1419)
risk stratification, sparking debate over its role in treatment decisions. We evaluated is the largest registry study of glioblastoma patients surviving for 5 years or more
survival relative to age and molecular data obtained from next-generation sequencing worldwide. Methods: Here we characterized the DNA methylation and mutational
(NGS). Methods: Tumor specimens from 598 mIDH gliomas were analyzed using NGS landscapes of 142 tumors from ETERNITY patients and compared the findings with
and WTS at Caris Life Sciences (Phoenix, AZ). Samples were stratified by age at di- different reference cohorts. Results: The majority of tumors of the ETERNITY cohort
agnosis into four groups: 12-26y, 27-40y, 41-60y, and . 60y. Real-world overall survival showed molecular profiles corresponding to established methylation subclasses of IDH-
(calculated from initial diagnosis to last contact) was obtained from insurance claims wildtype glioblastoma. ETERNITY tumors were enriched for the mesenchymal subclass,
data and analyzed using Kaplan-Meier and Cox proportional hazards models. Covariates depleted of the receptor tyrosine kinase 1 subclass, and showed a high frequency of
in the multivariate regression analysis included radiation treatment, temozolomide MGMT promoter methylation. While large chromosomal alterations were remarkably
treatment, and mutation status of different biomarkers. Results: In mIDH astrocytoma similar in all cohorts, circumscribed homozygous deletions on chromosome 10q in-
group, age distribution was 12-26y, n = 74 (12.4%); 27-40y, n = 271 (45.3%); 41-60y, n = cluding the MGMT gene were enriched in ETERNITY tumors. Gene panel sequencing
205 (34.3%); and . 60y, n = 48 (8.0%). In mIDH oligodendroglioma group, age dis- showed similar types and frequencies of gene alterations as in the reference cohorts
tribution was 12-26y, n = 18 (5.5%); 27-40y, n = 76 (23.2%); 41-60y, n = 137 (41.8%); with a trend towards more frequent RB1 mutations. Deconvolution analyses of global
and . 60y, n = 57 (17.4%). For each subtype, comparisons in survival were made DNA methylation data revealed fewer monocytes in the MES methylation class in
between patients 27-40y vs. 41-60y given larger sample size, and patients with ETERNITY compared with the reference cohort. ETERNITY tumors from patients without
temozolomide treatment before biopsy were excluded (about 10%). Univariate analysis documented relapse showed no specific molecular profile. Small subgroups of tumors
showed that 27-40y patients had shorter survival in astrocytoma (HR = 1.63, 95% CI: 1.07 corresponded to rare incompletely defined tumor entities. Conclusions: The present
– 2.50, p = 0.022). However, after adjusting for confounding factors in multivariate study illustrates the profound association of MGMT gene alterations with outcome, but
analysis, age was not associated with survival. In contrast, TP53 (HR = 4.0, 95% CI: 1.43- also suggests as yet unidentified clinical or molecular pathways and potential host-
11.24, p = 0.008 – mutation rate = 95.4%) and TERT-promoter (HR = 10.36, 95% CI: 4.05- dependent features in long-term survival with glioblastoma. Clinical trial information:
26.45, p , 0.0001 – mutation rate = 9.0%) mutations were independently associated NCT03770468. Research Sponsor: Brain Tumor Funders’ Collaborative Consortium.
with poorer survival in astrocytoma patients. Univariate analysis showed that age was
not associated with survival in oligodendroglioma (HR = 1.07, 95% CI: 0.79-3.65, p =
0.168). KRAS mutations were independently associated with poorer survival in oligo-
dendroglioma patients (HR = 4.36, 95% CI: 1.12-16.92, p = 0.033 - mutation rate = 3%).
Conclusions: In this enriched dataset of mIDH low grade glioma patients, which in-
cluded NGS, age did not contribute to survival differences when comparing patients
between 27-40 years with those aged 41-60 years. Rather, selected genetic alterations
such as KRAS for oligodendroglioma and TP53 and TERT mutations for astrocytoma
were associated with poorer survival. The results suggest that NGS, rather than age, may
drive prognosis for mIDH glioma patients. Research Sponsor: None.

2060 Poster Session

Efficacy and safety of depatuxizumab mafodotin (ABT-414) in EGFR-
amplified glioblastoma: A systematic review and Bayesian network meta-
analysis. First Author: Sunjida Amin Promi, Chittagong Medical College Hospital,
Chittagong, Bangladesh
Background: Glioblastoma (GBM) is a highly aggressive brain tumor with a poor prognosis,
typically resulting in a median survival of 12–15 months. Epidermal growth factor receptor
(EGFR) alterations, present in half of GBM cases, are key therapeutic targets. Depatuxizumab
mafodotin (Depatux-M, ABT-414), an EGFR-targeting antibody-drug conjugate, represents a
novel therapeutic option. This Bayesian network meta-analysis assessed the efficacy and
safety of Depatux-M in EGFR-amplified GBM. Methods: Eight randomized controlled trials
(RCTs) involving 1,183 patients were analyzed. Trials evaluating Depatux-M as monotherapy or
combined with temozolomide (TMZ) and/or radiotherapy (RT) were included. Outcomes in-
cluded overall survival (OS), progression-free survival (PFS), and safety (grade 3/4 adverse
events and keratitis). Bayesian models estimated mean differences (MDs) and relative risks
(RRs) with 95% credible intervals (CrI), while SUCRA values ranked treatments.
Results: Depatux-M plus TMZ showed modest OS improvement over TMZ alone (MD:
0.91 months; 95% CrI: -11.83 to 13.86; SUCRA: 62.09%). Depatux-M monotherapy showed
minimal OS benefit (MD: 0.07 months; 95% CrI: -12.69 to 12.95; SUCRA: 51.2%), and the
combination of Depatux-M, TMZ, and RT had the lowest OS benefit (MD: -2.17 months; 95% CrI:
-19.83 to 15.74; SUCRA: 35.48%). For PFS, Depatux-M monotherapy performed best (MD:
1.46 months; 95% CrI: -4.92 to 7.78; SUCRA: 81.00%), while Depatux-M plus TMZ (MD:
-0.45 months; 95% CrI: -6.85 to 5.89; SUCRA: 40.03%) and Depatux-M, TMZ, and RT (MD:
-1.54 months; 95% CrI: -10.34 to 7.24; SUCRA: 28.32%) were less effective. Depatux-M
ABSTRACT
monotherapy had a lower RR for grade 3/4 adverse events (RR: 1.38; 95% CrI: 0.23 to 8.07) and
keratitis (RR: 2.62; 95% CrI: 0.43 to 15.63) compared to combination regimens, with the highest
keratitis risks observed in Depatux-M, TMZ, and RT. Conclusions: Depatuxizumab mafodotin
WITHDRAWN
offers limited survival benefits in EGFR-amplified GBM, with monotherapy showing the most
favorable PFS. However, significant safety concerns, particularly keratitis, warrant further
research to optimize its therapeutic potential and identify more tolerable regimens. Research
Sponsor: None.
Efficacy and safety outcomes of depatuxizumab mafodotin in EGFR-amplified glioblastoma.
Overall Progression-Free Grade 3/4 Adverse
Regimen Survival (OS) Survival (PFS) Events (RR) Keratitis (RR)
Depatux-M + 0.91 (-11.83 to 13.86); -0.45 (-6.85 to 5.89); 1.54 (0.20 to 13.53); 4.40 (0.51 to 31.10);
TMZ SUCRA 62.09 SUCRA 40.03 SUCRA 41.75 SUCRA 33.07
Depatux-M 0.07 (-12.69 to 12.95); 1.46 (-4.92 to 7.78); 1.38 (0.23 to 8.07); 2.62 (0.43 to 15.63);
SUCRA 51.20 SUCRA 81.00 SUCRA 49.06 SUCRA 62.34
Depatux-M + -2.17 (-19.83 to 15.74); -1.54 (-10.34 to 7.24); 0.98 (0.09 to 9.65); 6.63 (0.62 to 66.40);
TMZ + RT SUCRA 35.48 SUCRA 28.32 SUCRA 68.37 SUCRA 12.73

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 127s

2062 Poster Session 2063 Poster Session

Relationship between aperiodic dynamics and transcriptomic alterations Comparing ERK signaling and tumor microenvironment in BRAF-altered
and a neural signature of glioma-induced excitation-inhibition gliomas. First Author: Lucy Chen, Johns Hopkins School of Medicine, Baltimore, MD
dysregulation. First Author: Youssef Sibih, University of California, San Francisco, Background: Many patients with BRAF-altered glioma (V600E mutations, fusions) re-
San Francisco, CA spond to BRAF inhibitors (BRAFi), but some—particularly those with high-grade gliomas
Background: Diffuse gliomas disrupt neuronal dynamics, leading to excitation-inhibition (HGG)—progress on treatment. Here, we aim to assess MAPK/ERK activation associated
(E/I) imbalance and associated functional impairments. The aperiodic component of the with various BRAF alterations in a large dataset of adult and pediatric patients with low-
power spectral density (1/f slope) has emerged as a proxy for estimating E/I balance, grade (LGG) and HGG using previously validated transcriptomic signatures.
offering a novel framework for understanding glioma-induced neural dysregulation. This Methods: Samples underwent next-generation sequencing and whole transcriptome
study is the first to validate the relationship between 1/f slope and E/I dysregulation in sequencing at Caris Life Sciences (Phoenix, AZ). The MAPK Pathway Activity Score (MPAS,
glioma by integrating electrophysiological, genomic, and behavioral data. Wagle 2018, NPJPO) and two MEK inhibitor sensitivity signatures (Dry 2019, Cancer Res &
Methods: Resting-state intraoperative subdural electrocorticography (ECoG) data were Pratilas 2009, PNAS) were calculated from RNA-seq data. Tumor immune microenvi-
recorded from 13 glioma patients. Power spectral analysis at a frequency of 70–150Hz ronment was assessed using immune deconvolution (quanTIseq) and the Tumor In-
(high-gamma) computed 1/f slopes, and electrodes were classified as glioma-infiltrated flammation Signature (TIS). Results were compared for HGG in two age groups (AYA: 0-
or normal-appearing based on preoperative MRI T2-FLAIR. Linear mixed-effects models 39y; adult: . 39y) in V600E mutation, fusions, or controls (BRAF-WT/IDH-WT/NF1-WT).
assessed E/I balance across tissue and glioma subtypes. Single-nucleus RNA sequencing For LGG 0-39y, V600E were compared to fusion, but not to BRAF-WT due to near-universal
(snRNA-seq) was performed on 14 spatially annotated glioma tissue samples from re- enrichment of MAPK alterations in LGG. There were insufficient LGG . 39y for analysis.
gions classified as inhibitory or excitatory by 1/f slope. Behavioral analysis of language Mann-Whiney U tests were used at a = 0.05. Results: In adult HGG (V600E: n = 35, fusions:
tasks examined functional correlates of E/I imbalance. Results: The cohort included n = 11, WT: n = 3235), both V600E and fusions showed significantly higher MAPK/ERK
23.0% WHO grade 2 IDH-mutant oligodendrogliomas, 38.5% WHO grade 2-3 IDH-mutant signatures than WT (all p , 0.01), with no significant difference between V600E and
astrocytoma, and 38.5% glioblastoma (GBM). Glioma-infiltrated electrodes (n=142) fusions. In AYA HGG (V600E: n = 32, fusions: n = 11, WT: n = 235), all three MAPK/ERK
exhibited significantly lower 1/f slopes than normal-appearing electrodes (n=518;p , signatures were significantly higher in V600E compared to WT (all p , 0.01), while fusions
0.0001), reflecting an excitation-dominant state. Subtype analysis revealed hierarchical fell in between BRAF V600E and WT (p . 0.05). In both AYA and adult HGG, B cells were
E/I imbalance, with GBM showing the steepest reductions in 1/f slope compared to higher in WT compared to V600E, while among infiltrated cells only M1 and M2 mac-
astrocytoma and oligodendroglioma (glioma-infiltrated: p,0.0001; normal-appearing: rophages were elevated in fusions compared to WT (p , 0.05). Comparison of MAPK/ERK
GBM vs. oligodendroglioma, p=0.012; GBM vs. astrocytoma, p=0.019). SnRNA-seq signatures in LGG (V600E: n = 28, fusions: n = 54) revealed no significant difference
revealed elevated excitatory and reduced inhibitory signaling gene expression in glu- between V600E and fusions. TIS did not differ among BRAF alterations in any groups.
tamatergic and GABAergic neuronal populations across glioma-infiltrated (n=12; n=4 per Pathway analysis revealed RAS signaling and inflammation were enriched in BRAF V600E
subtype) and normal cortex (n=2) samples. Excitatory module scores were significantly compared to WT in both HGG and LGG (NES . 2; FDR , 0.005), regardless of age. When
higher in excitatory 1/f samples compared to inhibitory 1/f samples, validating the 1/f comparing all V600E HGG patients (0-90y) previously treated with BRAFi (n = 21) to those
slope as a genomic correlate of E/I imbalance in human cortical tissue. Behavioral who were not (n = 46), no difference in MAPK/ERK signatures were seen. Pathway analysis
analysis of language tasks demonstrated error-related reductions in 1/f slope (e . i), revealed samples with prior BRAFi had upregulated complement activation, B-cell acti-
emphasizing the functional impact of glioma-induced dysregulation. vation, and opsonization (NES . 3.5; FDR , 10e-5), while treatment-naı̈ve samples had
Conclusions: Diffuse gliomas are associated with a profound shift toward excitation higher BRAF/MAPK signaling (NES . 2, FDR , 0.03). Conclusions: These data confirm
dominance in both glioma-infiltrated and normal-appearing cortex. For the first time, the higher MAPK/ERK dependence signatures and RAS signaling in BRAF-altered HGG (V600E,
1/f slope is validated as a robust measure of E/I imbalance through electrophysiological, fusion) compared to BRAF-WT controls. Differences in immune cell infiltration were
genomic, and behavioral analyses. These findings position the 1/f slope as a physio- observed between BRAF alteration classes. Changes in humoral immunity may be cor-
related with acquired resistance to BRAFi, in line with previous reports of increased B-cell
logically relevant biomarker of glioma-induced neural dysregulation, offering significant
infiltration in BRAFi-resistant melanomas. Research Sponsor: None.
potential to inform therapeutic strategies. Research Sponsor: None.

2064 Poster Session 2065 Poster Session

Phase II propensity-matched controlled trial evaluating metformin as an Association of MGMT status with survival in low and high-grade IDH-mutant
adjunct to neo-adjuvant, concomitant, and adjuvant temozolomide and astrocytomas. First Author: Katherine E. Schwetye, Washington University, St. Louis,
hypofractionated-accelerated radiotherapy (M-HART) in glioblastoma pa- MO
tients (NCT02780024). First Author: George Shenouda, McGill University Health Background: The role of MGMT promoter methylation status on survival for IDH-mutant
Centre, Montréal, QC, Canada astrocytomas is less known than for glioblastoma, IDH-wildtype (GBM). Further, different
Background: Phase II, propensity-matched trial, to assess feasibility and toxicity of adding Metformin (MTF) to neo- laboratories utilize a variety of techniques to measure methylation of the MGMT gene
adjuvant, concomitant and adjuvant Temozolomide (TMZ) and hypofractionated accelerated radiotherapy (M-HART), promoter region, including pyrosequencing, methylation-specific polymerase chain re-
for patients with Glioblastoma (GBM). We compared median survival time (MST), and progression-free-survival (PFS)
of M-HART versus a contemporaneous cohort of propensity-score matched controls (PSMC) who received standard action (PCR), and direct Sanger sequencing; these techniques are limited by low quan-
of care (SOC). Methods: Eligible patients were $ 18 years with newly diagnosed GBM, ECOG score # 2, with known titative accuracy, short read length, and low sample throughput. In the current study, we
MGMT status, gross total or partial resection, and residual surgical cavity . 15 mm from brainstem, or optic used a large database of next-generation sequencing (NGS) and whole-transcriptome
apparatus. Four weeks from surgery, M-HART patients started 2 weeks of neo-adjuvant MTF/TMZ followed by sequencing (WTS) performed in a single laboratory to determine the role of MGMT status
concomitant MTF/TMZ + HART 60 Gy/20 daily fractions, and 6 cycles of adjuvant MTF/TMZ. The PSMC patients
received Stupp’s regimen. We used a nearest neighbor matching with a caliper width of 0.2 SD and compared on survival in IDH-mutant astrocytomas (CNS WHO grades 2-3 and 4), as well as in GBM.
patients’ characteristics using chi-square test (Table). Propensity scores were estimated using logistic regression Methods: 10,181 glioma samples were analyzed by NGS (592, NextSeq, or WES, NovaSeq)
model, with probability of M-HART treatment as dependent variable. Results: From April 2015 to November 2020, 50 and WTS (NovaSeq) at Caris Life Sciences (Phoenix, AZ), including determination of
patients participated in the M-HART trial and matched with 50 PSMC cohort treated during the same period, with a methylation status of the MGMT promoter region by pyrosequencing. Real-world overall
median follow up of 24.1 (M-HART) vs 17.6 months PSMC, respectively. M-HART patients had significantly longer
MST of 24.1 (95% CI, 15.2- 30.3) vs. 17.7 months for PSMC patients (95% CI, 12-20) (HR, 0.62 [95% CI, 0.40-0.93]; P =
survival was obtained from insurance claims data and calculated from initial diagnosis to
0.02), and significantly longer PFS of 13.7 (95% CI, 11.7 to 18.8) vs. 11.0 months (95% CI, 9-12) (HR, 0.63 [95% CI, last contact, while TMZ-OS was calculated from first dose of temozolomide to last of
0.42-0.95]; P = 0.02). M-HART treatment was an independent predictor of survival. M-HART patients with treatment. Hazard ratios (HRs) were analyzed using Cox proportional hazards model and p
methylated-MGMT and gross total resection had significant longer MST of 41.9 vs. 17.8 months for PSMC (95% CI, values (log-rank test). Multivariate regression analysis was performed on age, gender,
15.1-20.5 months) (HR 0.21 [95% CI, 0.09-0.49]; P = 0.001). Conclusions: M-HART protocol is novel, feasible, and
well-tolerated approach with significantly longer MST and PFS as compared to propensity-matched SOC controls.
radiation treatment, temozolomide treatment, and mutations in different biomarkers.
These results add to growing evidence for the use of Metformin as an adjunct to HART and TMZ especially in M- Fisher’s exact tests was used at a significance level of 0.05. Results: 693 IDH-mutant
MGMT GBM. Clinical trial information: NCT02780024. Research Sponsor: No funding was received. astrocytomas CNS WHO grades 2 or 3 (“g2/3”), 251 IDH-mutant astrocytoma CNS WHO
Characteristics of M-HART versus propensity-matched standard of care control patients. grade 4 (“g4”), and 4469 glioblastoma (“GBM”) met inclusion criteria. Univariate and
M-HART N=50 (%) CONTROLS N=50 (%) P-value multivariate survival analysis showed that MGMT promoter methylation (mMGMT vs.
Age (years) unmethylated, uMGMT) was associated with improved OS only in GBM (HR = 0.62, 95% CI:
£ 60
> 60
34 (68)
16 (32)
27 (54)
23 (46)
0.218 0.57 – 0.67, p , 0.0001), but not in astrocytoma-g2/3 and g4. Similarly, TMZ-OS was only
Sex significantly longer in mMGMT vs. uMGMT in GBM (HR = 0.53, 95% CI: 0.48 – 0.58, p ,
Male 22 (44) 30 (60) 0.161
Female 28 (56) 20 (40) 0.0001). In astrocytoma-g2/3, ATRX mutation was more prevalent in mMGMT than uMGMT
ECOG-score
0-1 43 (84) 47 (94) 0.318
(73.9% vs. 62.6%, p , 0.05), and SETD2 was more prevalent in uMGMT than mMGMT (4%
2 7 (14) 3 (6) vs. 1.1%, p , 0.05). Tumor mutational burden (TMB)-high was more prevalent in mMGMT
Surgery
Gross Total 41 (82) 39 (78) 0.803 than uMGMT in astrocytoma-g4 (14.8% vs. 2.7%, p , 0.05) and in GBM (5.4% vs. 1.9%, p ,
Subtotal
MGMT status
9 (18) 11 (22) 0.0001). In GBM, many genes had different mutational rates between mMGMT and uMGMT
Unmethylated 34 (68) 29 (58) 0.015 groups, including MSH6 (2% vs. 0.7%, p , 0.001), ATRX (3.1% vs. 1.6%, p , 0.01), and
Methylated 16 (32) 21 (42)
Re-operation CDKN2A (4.3% vs. 2.5%, p , 0.01). Conclusions: mMGMT was not associated with better
Yes
No
24 (84)
26 (52)
18 (36)
32 (64)
0.077 survival in IDH-mutant astrocytoma-g2/3 or g4with respect to OS or TMZ-OS, whereas
Chemotherapy at recurrence mMGMT conferred improved survival in GBM. These results, derived from a large database
Yes 14 (28) 27 (54) ,0.001
No 36 (74) 23 (46) using same platform (next-generation sequencing at a single laboratory), support similar
findings from recent, smaller cohort studies. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
128s CENTRAL NERVOUS SYSTEM TUMORS

2066 Poster Session 2067 Poster Session

Prognostic impact of DDR mutations (mt) in IDH mutant high-grade gliomas Prognostic value of inflammatory markers in glioblastoma: A meta-analysis
(HGG). First Author: John L. Villano, University of Kentucky Markey Cancer Center, of NLR and PLR stratified by cutoff values. First Author: Abril Carrillo, UT
Lexington, KY Southwestern Medical Center, Dallas, TX
Background: The oncometabolite 2-hydroxyglutarate (2HG) produced by IDH1/2 mt in Background: Glioblastoma (GBM) is the most aggressive primary brain tumor in adults,
HGG has profound effects on numerous pathways including DNA damage repair (DDR). with an unfavorable prognosis. Identifying prognostic markers is crucial to stratify
We investigated the prognostic effect of DDR mt in IDH mutant vs. wild type (wt) tumors patients and tailor therapeutic approaches. The neutrophil-to-lymphocyte ratio (NLR) and
in a large cohort using a real-world database. Methods: A total of 4894 HGG tumors platelet-to-lymphocyte ratio (PLR) are inflammatory markers that have gained attention
tested at Caris Life Sciences (Phoenix, AZ) with NextGen sequencing of DNA (592-gene as outcome predictors in various cancers. While numerous studies have demonstrated
panel or whole exome sequencing) were included in the study. DDR alteration was associations between elevated NLR, PLR, and poor GBM outcomes, consensus on
defined as a pathogenic mutation in one of . 20 DDR genes. Patient survival was standardized cutoff values remains elusive, limiting their clinical application. In this meta-
obtained by insurance claims data and calculated from the initiation of tissue collection analysis, we stratified the data based on preoperative NLR and PLR cutoff values, aiming
(rwOS). Cox proportional hazards model was used to calculate hazard ratios (HR) and to identify the most accurate cutoff thresholds to predict overall survival (OS) outcomes.
log-rank tests to calculate p values, which were adjusted for multiple comparisons. Methods: We performed a systematic search on PubMed, Medline, OVID, Embase, and
Significance was set at p,0.05. Results: In the 1121 HGG carrying either IDH1 or 2 Cochrane in November 2024 to identify cohort studies reporting hazard ratios (HR) for OS
mutations, 100 carried a DDR mutation (8.9%) . When comparing DDR mutant (mt) vs. associated with preoperative NLR and/or PLR in patients with histopathologically
wild type (wt), no difference was seen in patient age (median 39 vs. 38 yrs; p = 0.8); confirmed GBM. Two reviewers screened articles; discrepancies were resolved by
gender (female 45% vs. 42%, p = 0.9), race or ethnicity (p . 0.1). The most frequent consensus. Data analysis was conducted using a random-effects model. Pooled HRs with
mutations were seen in MSH6 (24% of the DDR mt), ATM (18%), MLH1 (15%), MSH2 95% confidence intervals (CI) were calculated and subgroup analysis were performed.
(13%), MSH3 (10%) and BRCA2 (10%). When comparing the rwOS of DDR mt vs. wt, a Results: From 227 studies initially identified, 22 studies met our inclusion criteria,
significantly shorter survival was seen (24m vs. 51m, HR = 1.87, 95% CI [1.41-2.48], p , encompassing a total population of 3,423 patients with GBM. In our general cohort, when
0.001); the effect persisted in the subset of tumors collected prior to temozolomide compared to lower PLR, a higher PLR yielded a pooled HR of 1.30 (95% CI: 1.12-1.50, p ,
treatment (26m vs. 64m, HR = 1.92 [1.35-2.74], p , 0.001). In contrast, in IDH wt tumors, 0.001). Specific cutoff subgroup analysis revealed that, the , 135 cutoff group had a HR
patients with (N = 223) or without DDR mutation (N = 3550) showed similar survival of 1.09 (95% CI: 0.62-1.93, p = 0.60), in the . 135 cutoff group, the HR was 1.42 (95% CI:
(17.5m vs. 20.6m, p = 0.1). In the IDH mutant cohort, DDR mt was associated with an 1.19-1.70, p = 0.0001). Regarding the NLR analysis, cutoff subgroup analysis showed that
increased tumor mutational burden (TMB) compared to DDR wt tumors (median = 6 vs. 4 for NLR with a cutoff of , 3, the HR was 1.39 (95% CI: 0.96–2.02, p = 0.08), for NLR with a
mutations/mb, by Wilcoxon). Multivariate analysis within the IDH mutant tumors in- cutoff between 3 and 4.9, the HR was 1.56 (95% CI: 0.98–2.51, p = 0.06). For studies with a
dicated that both TMB and DDR status were independently associated with poorer rwOS, NLR cutoff of 4, the HR was 1.40 (95% CI: 1.23–1.58, p = 0.01). For studies with a NLR
with TMB showing an adjusted HR of 1.01 per unit increase (p = 0.005) and DDR status cutoff . 4.9, the HR was 1.85 (95% CI: 1.37–2.50, p , 0.0001). The overall pooled HR for
with an adjusted HR of 1.59 (p = 0.028). Conclusions: In a large real-world database, we elevated NLR regardless of cutoff value was 1.40 (95% CI: 1.23–1.58, p , 0.00001).
demonstrate IDH mt HGG with a DDR mutation exhibit significantly poorer survival Conclusions: Elevated preoperative NLR and PLR are significant prognostic markers for
compared to DDR wt. This is not seen in IDH wt, where survivals of the two groups are worse OS in GBM patients. Stratifying data by cutoff values revealed that PLR . 135 and
similar. These results stand in sharp contrast to reported prognostic effect of DDR NLR . 4.9 were more consistently correlated with poor survival outcomes. These findings
mutation in many other solid tumors. The data suggest that DDR mutations in the suggest that higher cutoff values for these markers may better predict OS, particularly for
context of 2HG accumulation in IDH mt HGG may be an indicator of profound genomic NLR where values . 4.9 demonstrated a stronger association than the commonly used
instability that confers severe negative impact on patient survival. Clinicians managing cutoff of 4. The results highlight the potential utility of NLR and PLR as accessible, cost-
high-grade gliomas should consider the presence of DDR mutations in IDH mutant effective prognostic tools. Future prospective studies are warranted to validate these
patients as a poor prognostic category in this overall favorable prognostic group and findings, refine optimal cutoff thresholds, and explore their applicability. Research
consider therapeutic approaches accordingly. Research Sponsor: None. Sponsor: None.

2068 Poster Session 2069 Poster Session

A phase 1, first-in-human study of regorafenib plus temozolomide with or Azeliragon, a RAGE inhibitor, in combination with temozolomide and radio-
without radiotherapy in patients with newly diagnosed MGMT methylated, therapy in patients with newly diagnosed glioblastoma: Preliminary results
IDHwt glioblastoma: The REGOMA-2 trial. First Author: Marta Padovan, University of phase Ib/II CAN-201 NDG trial. First Author: Juan Manuel Sepulveda, Medical
of Padua, Padua, Italy Oncology Service, Hospital Universitario 12 de Octubre, Madrid, Spain
Background: Regorafenib (REG) is an oral multikinase inhibitor and in vivo studies Background: Azeliragon is an orally available inhibitor of the receptor for advanced
demonstrated a synergistic antitumor effect when combined with radiotherapy (RT) and glycation end-products (RAGE). RAGE pathway promotes cell proliferation and angio-
temozolomide (TMZ) against glioblastoma (GBM). We conducted a phase 1 study to genesis, contributing to glioblastoma (GBM) progression and resistance to temozolomide
evaluate the safety, dose limiting toxicity (DLT), maximum tolerated dose (MTD) of REG, (TMZ) and radiation (RT). Azeliragon has extensive clinical safety data in patients (pts) with
pharmacokinetics (PK), preliminary activity of this combination. Methods: This phase 1 Alzheimer’s disease. Our hypothesis was that azeliragon may enhance the efficacy of
multicenter academic study used a "3 + 3" design to evaluate REG doses of 80mg (Level 1), Stupp regimen in newly diagnosed GBM. Methods: CAN-201 NDG is an open-label, single
120mg (Level 2), 160mg (Level 3) in 2 different cohorts of pts with histologic diagnosis of arm, phase Ib/II trial in Spain. Newly diagnosed IDH wild-type pts with GBM, MGMT
MGMT-methylated, IDHwt GBM (WHO 2021) and ECOG PS 0-1. In cohort A, pts who methylation locally available and with tumor resection were recruited. Pts received
completed the concurrent chemoradiotherapy (CT-RT) regimen received REG in combi- azeliragon in combination with standard radiotherapy and TMZ followed by maintenance
nation with standard maintenance TMZ; cohort B received REG concurrently with standard with azeliragon. The trial consists of an initial dose finding phase in a 6 dose escalation
CT-RT and continued REG with maintenance TMZ. REG was administered according to the strategy with a subsequent expansion phase (up to 14 additional pts) at the recommended
standard schedule of 3 weeks on/1 week off. The DLT evaluation period for cohort A was phase 2 dose (RP2D). The dose levels were: 5 mg/day (L1), 10 mg/day (L2) and 20 mg/day
during the first two maintenance cycles and for cohort B during the concurrent CT-RT (L3). The primary objective is to determine the RP2D, defined as the dose for which , 33%
phase. During the DLT period, blood and clinical assessments were performed weekly. pts experience a dose limiting toxicity (DLT) within 28 days from initiation of dosing. Main
Toxicity was assessed by CTCAE v 5.0. RANO criteria were used for neuroradiologic secondary endpoints include progression-free survival (PFS), overall survival (OS) and
assessment. Pharmacokinetics (PK) was also evaluated. Results: In cohort A, none of the changes in corticosteroid requirements. Results: From Oct 2023 to Jul 2024, 20 pts were
9 pts enrolled (median age 52ys) had a DLT at any dose; 1 pt in Level 2 had REG delayed and included, 6 in L1, 8 in L2 and 6 in L3. The median age was 52 years (range: 40-69). Most pts
TMZ dose reduced due to grade (G) 2 thrombocytopenia. At Level 1 and 2, 1 G3 hae- were male (65%), ECOG 0-1 (95%) and MGMT unmethylated (60%). No DLTs were observed.
matological toxicity, respectively. At Level 3, 1 pt had a G3 gastrointestinal toxicity. In Serious adverse events, all considered unrelated to azeliragon, were reported in 4 pts
cohort B, 12 pts were enrolled (median age 53ys); at Level 3, 2 of 6 pts reported a DLT (n=1 (20%), being hemiplegia, pyrexia, infectious meningoencephalitis, epilepsy and neuro-
G3 hypertransaminasemia with a dose reduction of REG (51%) and TMZ (50%) and n=1 G4 logical decompensation. Non-serious Grade 3-4 adverse events (AE), also considered
thrombocytopenia at the last day of RT). One case of G3 hypertension and 1 case of G3 unrelated, were G3 hematological AEs in 33.3% in L1 and 37.5% in L2. G1-2 azeliragon-
hypertransaminasemia were also reported. REG was reduced in another pt due to G2 pain related AEs were reported in 33.3%, 25% and 66.7% of pts in L1, L2 and L3, respectively.
(no DLT); at Level 2 there was 1 case of G3 hyperbilirubinemia. There were no G3-4 AEs at Azeliragon treatment was ended due to progression in 83.3% and 62.5% of pts in L1 and L2,
Level 1. PK analysis of REG alone or in combination with TMZ showed a significant respectively. All pts on L3 are still on treatment. With a median follow-up time of
reduction (P=0.038) in the AUC, with a geometric mean ratio (GMR) of 80% (CI90 64 – 98%) 8.4 months, pts in L1 showed a median PFS of 5.2 months (95% CI, 4.4-Not Reached [NR])
when given together with TMZ. PK analysis of TMZ showed a slight but significant re- and 9.8 months (95% CI, 6.2-NR) in L2. No progression of disease was observed in L3 with a
duction in the Cmax and AUC (P = 0.003 and 0.015, respectively) when given with REG, with range of follow-up of 4.9-7.0 months. Median OS in L1 was 11.1 months (95% CI, 9.4-NR).
GMR of 72% (CI90 57 – 91%) and 86% (CI90 79 – 92%), respectively. These results suggest a Data was not mature enough to calculate OS in L2 and L3. Conclusions: Azeliragon in
weak PK interaction between the two drugs. Conclusions: The MTD of REG for cohort A combination with standard RT and TMZ is safe, with no dose-limiting toxicities reported so
was 160mg, for cohort B 120mg with a weak PK interaction between the two drugs. The far at the initial three dose levels. To further explore the safety and efficacy profile of
MTD of 120mg can be considered the recommended dose of REG in combination with azeliragon, we are now expanding the study to include two additional dose levels of 30 mg/
standard Stupp therapy for the phase 2 study. Preliminary activity analyses are ongoing. day (L4) and 50 mg/day (L5). Enrollment is currently open for level L4. Clinical trial in-
Clinical trial information: NCT06095375. Research Sponsor: None. formation: NCT05635734. Research Sponsor: CANTEX Pharmaceuticals, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 129s

2070 Poster Session 2071 Poster Session

From retrospective analysis to real-world impact: Mismatch repair defi- Immunohistochemical and gene expression profiles as predictors of survival
ciency detection in gliomas by tissue and liquid ctDNA NGS in glioma in recurrent high-grade glioma treated with intracranial nivolumab, ipilimu-
management. First Author: Suman Suryanarayana Karanth, FMRI, Gurugram, India mab, and autologous CD1c(BDCA-1)+/CD141(BDCA-3)+ myeloid dendritic
Background: Tissue biopsy remains the gold standard for Microsatellite Instability and cells (myDC). First Author: Cleo Bertels, Universitair Ziekenhuis Brussel (UZ Brussel),
MMR (mismatch repair) gene alteration assessment. The regulatory approval of immune Brussels, Belgium
checkpoint inhibitors is for mismatch repair-deficient cancers, regardless of tumor type. Background: Innovative treatments are needed for recurrent high-grade glioma (rHGG)
Surgical resection or biopsy is challenging when the glioma is located deep in the brain patients (pts) as current salvage therapies fail to improve overall survival (OS). Immune
or brainstem. Circulating tumor DNA (ctDNA) next-generation sequencing (NGS) offers a checkpoint inhibitors lack efficacy in rHGG when administered IV. This single center,
non-invasive alternative, garnering attention in extracranial cancers, however not so in multicohort phase I trial (Glitipni, NCT03233152) investigated intracerebral (iCer) ad-
gliomas due to the lower concentrations of tumor-derived biomarkers. While Cere- ministration of ipilimumab (IPI) +/- nivolumab (NIVO) +/- myDC after maximal safe re-
brospinal fluid (CSF) provides superior sensitivity and specificity for gliomas, it is not a section (MSR), followed by adjuvant intracavitary (iCav) IPI/NIVO through an Ommaya
standard test in gliomas with its own technical issues. Methods: A retrospective reservoir. Methods: Eligible pts (ECOG # 2, # 8mg/day methylprednisolone) with rHGG
analysis was conducted using databases to evaluate the prevalence of pathogenic (WHO 2021 grade 3/4, IDH-1/2 wild type (wt) or mutant) after standard postoperative
inactivating alterations in MMR genes in glioma tissue samples. Data were queried from radiotherapy (RT) and temozolomide (TMZ) were included. Pts underwent MSR or ste-
the MSK, Clin Cancer Res 2019 database for targeted sequencing on MSK-IMPACT and reotactic biopsy (if unresectable) , 24h after receiving NIVO IV (10mg), followed by iCer
FMI panels, comprising 1004 tissue samples (837 with matched normal) from 923 injection of varying doses of IPI +/- NIVO +/- myDC and Ommaya catheter placement
glioma patients through the cBioPortal platform. Frequencies of MMR gene alterations depending on the cohort (C). NIVO was administered IV (all cohorts) +/- iCav (C3-7) bi-
were assessed. Results: A total of 850 patients (out of 923) were retrospectively weekly up to 12 cycles. Baseline tumor microenvironment characteristics were assessed
analyzed for MMR gene alterations, with 40.3% (343) being female and 59.6% (507) by immunohistochemical (IHC) analysis and gene expression profiling (GEP).
male. Primary samples constituted 79.8% (679), and recurrent samples 20.2% (172). Results: Between 2016 and 2023, 110 pts (68% male, median age 57, 92% ECOG 0/1) were
OncoKb level alterations were categorized as Level 1 (0.5%), Level 2B (16.6%), Level 3B enrolled. At primary diagnosis, the majority (85%) were glioblastoma pts (WHO grade 4,
(13.6%), Level 4 (32.8%), and none (36.4%). MMR gene alterations were found in 35 IDH-wt), treated with the standard of care (MSR + RT + TMZ) (71%). All pts received 10mg
samples (4%), with MSH2 and MSH6 each detected in 2%, MLH1 in 1%, and PMS1, PMS2, NIVO IV preoperatively. Ninety percent of the pts who underwent the neurosurgical
procedure started the postoperative treatment. Early discontinuation of study treatment
and MSH3 in less than 1% of samples. In a specific case, ctDNA NGS was performed on a
occurred in 76% of pts, mainly due to tumor progression (86%). Treatment-related adverse
9-year-old boy diagnosed with diffuse intrinsic pontine glioma as tissue biopsy was not
events (TRAE) were mild (CTCAE grade 1/2), no grade 5 TRAE occurred. Most frequent
feasible. Survivals are 9 to 11 months despite multimodality treatment. ctDNA NGS
TRAE were fatigue, headache and fever. At database lock (Jan 1st, ‘25), 9 pts remained
identified a truncating MSH6 alteration at 100% Variant Allele Frequency, suggesting
progression-free. When including durable benefit from bevacizumab at first progression
biallelic inactivation of MSH6. Additionally, an IDH R132C activating mutation, a TP53 (13 pts), PFS and OS were significantly higher in C5/6 (+myDC) compared to other cohorts
splice site SNV and high tumor mutational burden (bTMB) at 132.33 Mut/Mb were (-myDC) of our trial with resectable rHGG, and to a historical control group treated with
detected. Post radiation resulted in no change in tumor size. Injection pembrolizumab VEGF(R)-inhibitors (descriptive p , 0.05 for each pairwise comparison). Absence of B7H3
3 weekly was initiated. Follow up MRI revealed further reduction in size and tumor has on resected tumor tissues as demonstrated by IHC (C4, 5, 7) showed longer median OS,
remained stable with ongoing therapy. Conclusions: Identifying MMR alterations po- which was consistent with GEP. PD-L1 expression and density of CD8, Granzyme B or
tentially broadens the therapeutic options for glioma. The compelling case of the 9-year- FOXP3 positive cells/mm2 did not correlate with survival. A proliferative gene signature on
old boy highlights the clinical utility of ctDNA NGS in identifying actionable MMR gene GEP was significantly correlated with shorter PFS and OS. Conclusions: Intracranial
alterations, leading to successful immunotherapy with pembrolizumab and continued administration of IPI/NIVO co-administered with myDC was feasible and safe, resulting in
stable disease beyond 13 months. While few studies exist on utility of ctDNA in gliomas, encouraging survival in pts with resectable rHGG. Baseline B7H3 levels and a proliferative
it is time for bigger studies in both primary and recurrent gliomas where biopsy is not gene signature correlated with survival. Clinical trial information: NCT03233152. Research
feasible. Research Sponsor: None. Sponsor: None.

2072 Poster Session 2073 Poster Session

Development and validation of a droplet digital PCR (ddPCR) assay to detect Combining abemaciclib, temozolomide, and radiation in DIPG PDOX models:
MGMT promoter methylation in FFPE tumors of glioblastoma (GBM) Insights from single-cell RNA-seq on cellular subtypes and genes critical for
patients. First Author: Mahrukh M Syeda, NYU Langone Medical Center, New York, NY responsiveness and resistance. First Author: Zilu Huang, Robert H. Lurie Com-
Background: Methylation of the MGMT gene promoter (MGMTp) is a critical biomarker prehensive Cancer Center, Feinberg School of Medicine, Northwestern University,
to inform GBM prognosis and guide treatment decisions, including clinical trial eligibility. Chicago, IL
Rapid reporting of MGMTp methylation can help stratify patients early and facilitate Background: Diffuse intrinsic pontine glioma (DIPG) is a highly aggressive pediatric brain
clinical trial referral during the crucial post-operative period where treatment options are tumor, with two-year survival less than 10%. Radiation therapy (XRT) offers limited survival
being considered by patients and physicians alike. Methods: We developed a probe- benefits, and effective therapies are urgently needed. This study investigates the FDA-
based ddPCR method to detect and quantify MGMTp methylation. Assay specificity was approved CDK4/6 inhibitors-Abemaciclib, XRT, and temozolomide (TMZ) as a therapeutic
assessed using bisulfite-converted, unmethylated DNA. Assay linearity and Limit of regimen for DIPG using organoids and patient-derived orthotopic xenograft (PDOX) models,
Detection (LoD) were determined using MGMTp methylated cell line DNA samples of aiming to improve survival outcomes and gain insights into the underlying cellular and
decreasing fractional abundances (35%, 5%, and 0.5%) and decreasing DNA inputs molecular mechanisms of DIPG treatment responsiveness and resistance to support rapid
(30ng, 10ng, 3ng,1ng and 0.5ng). The limit of Blank (LoB) was calculated using 16 translation into clinical trials. Methods: The efficacy of Abemaciclib, XRT, TMZ, and their
replicates of unmethylated bisulfite-converted PBMC DNA, and DNA from tonsil FFPE combinations was evaluated in DIPG organoids and PDOX models (IBs-A0317DIPG and IBs-
samples (n = 9). Reproducibility studies were conducted on two different days with two 9119DIPG, H3.3K27M mutation). In vitro, PDOX organoids were treated with Abemaciclib,
different operators. Accuracy and concordance were assessed using an in-house TMZ, with/without XRT. Synergy was assessed using the Bliss Independence model. In vivo,
MGMTp methylation pyrosequencing assay as an orthogonal method to analyze11 six treatment arms were tested: (1) control, (2) XRT (2 Gy/day 3 5), (3) Abemaciclib (75 mg/
melanoma and one glioblastoma cell line. Preliminary clinical validation was conducted kg, p.o., 14 days), (4) Abemaciclib + XRT, (5) TMZ (50 mg/kg, p.o., 5 days) + XRT, (6)
via analysis of FFPE tumor DNA from 34 GBM patients with clinical MGMTp methylation Abemaciclib + TMZ + XRT. Single-cell RNA sequencing and IHC were used to assess cellular
subtypes responses, gene expression changes, and resistance mechanisms. Results: In
pyrosequencing results. Results: The MGMTp methylation ddPCR assay demonstrated
DIPG organoids, the combination treatment yielded Over Bliss values . 0 (0.25 and 0.58 in
100% specificity to detect promoter methylation. The method linearly quantified both
A0317DIPG and 9119DIPG models, respectively) demonstrating synergistic activities.
total DNA and methylated DNA along a range of input DNAs with conserved fractional
In PDOX models, the triple therapy showed improved median survival compared to
abundances. The LoB was 0.036% and 0.034% using PBMC and tonsil FFPE DNA,
other treatment arms and significant survival advantage over control (p = 0.0157) and
respectively. The LoD was 0.075%. The bisulfite conversion and assay were highly Abemaciclib alone (p = 0.0461) in A0317DIPG, and control (p , 0.0001), XRT alone
reproducible, with a coefficient of variation , 20%. Among the 12 cell lines analyzed by (p = 0.0032), Abemaciclib alone (p = 0.0006), Abemaciclib + XRT (p = 0.0046), and TMZ +
both pyrosequencing and ddPCR the concordance was 100%. Ten of 11 GBM tumor XRT (p = 0.0001) in 9119DIPG models. Single-cell RNA sequencing revealed six tumor
samples identified as MGMTp methylated by the clinical pyrosequencing assay were subtypes: AC-like, NPC-like, OPC-like, MES-like, mitotic, and radiation-resistant cells. The
also identified as methylated by the ddPCR assay. Five of 23 clinically unmethylated triple therapy increased NPC-like and mitotic cell populations while decreasing AC-like and
tumors were positive in the ddPCR assay with generally very low fractional abundances OPC-like cells in both models. Additionally, we identified a novel radiation-resistant
(0.08%, 0.12% 0.3%, 0.97% and 10.11%). Conclusions: We report preliminary validation subpopulation that expanded after XRT treatment. Dynamic gene expression analysis in
of a highly sensitive and specific ddPCR assay to detect MGMTp methylation. Given the different cell types identified key target genes and cell-type specific pathways that mediate
minimal sample requirements and rapid turnaround time for ddPCR assays, this test therapy responsiveness and resistance. Conclusions: Our study demonstrates that the
could eventually be utilized in clinical laboratories to quickly report MGMTp methylation combination of Abemaciclib, TMZ, and XRT offers a novel, synergistic approach for DIPG,
status for GBM patient management. Research Sponsor: None. significantly improving survival in preclinical PDOX models. Single-cell RNA sequencing
reveals the roles of different cell types and molecular changes underlying resistance,
highlighting potential targets for future anti-resistance strategies in DIPG management.
Research Sponsor: The Lou and Jean Malnati Brain Tumor Institute (MBTI).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
130s CENTRAL NERVOUS SYSTEM TUMORS

2074 Poster Session 2075 Poster Session

MRI-based radiomics for prediction of isocitrate dehydrogenase subtype in Correlative and spatial transcriptomic analysis of olaparib and durvalumab
glioblastoma multiforme through artificial intelligence models: A systematic in patients with recurrent/refractory IDH-mutant gliomas. First Author: Xin
review and meta analysis. First Author: Sonali Belankar, M. S. Ramaiah Medical Wang, Department of Medical Oncology, University of Toronto, Sunnybrook Health-
College, Bangalore, India Sciences Centre, Toronto, ON, Canada
Background: Gliomas are aggressive tumours with poor prognosis. Isocitrate Dehy- Background: Combination of immune checkpoint and PARP inhibition has potential
drogenase (IDH) mutations are present in approximately 12% of all Glioma tumours and synergistic effects in IDHmt gliomas in pre-clinical models. Durvalumab and olaparib
are considered biomarkers for prognosis and response to chemotherapeutic agents. IDH demonstrated objective responses in a subset of patients (pts) with IDHmt gliomas
mutant gliomas have better prognosis in comparison to IDH wild type Gliomas. IDH (NCT03991832). We report mutational, transcriptomic, and spatial correlative analysis of
mutant Gliomas exhibit features like T2-Flair mismatch sign, reduced blood flow seen on pts samples from baseline and at time of progression. Methods: Pts with recurrent/
perfusion-weighted images and reduced enhancement on MRI, which aid in identifi- refractory IDHmt gliomas received olaparib 300 mg twice daily and durvalumab 1500 mg IV
cation of IDH mutation. Radiomic imaging techniques extract quantitative features from every 4 weeks until disease progression as determined by RANO 2.0 criteria. Whole exome
medical images like MRI and CT scans with the help of advanced algorithms and the sequencing (WES, n = 28) and total RNA sequencing (RNA-seq, n = 21) were performed on
extracted data can be utilized in the development of specific artificial intelligence (AI) baseline archival formalin-fixed, paraffin-embedded tumor samples. Baseline tumor mi-
models like Neural Networks for the prediction of IDH mutation. Thus, MRI based croenvironment was characterized with multiplex-immunohistochemistry (n = 29).
Radiomics is an emerging non invasive technique in comparison to conventional biopsy, Matched responders (n = 4) and non-responders (n = 6) were further profiled using 10X
for the determination of IDH mutation. The meta-analysis conducted aims to analyse the Visium HD for spatial transcriptomics. An unsupervised deconvolution method was applied
diagnostic potential of Radiomic imaging in predicting IDH mutations in Gliomas. using consensus non-negative matrix factorization for de novo discovery of expression
Methods: A systematic search was conducted in PubMed, Google Scholar and Scopus. programs corresponding to cell types and cell states. Associations with objective response
PRISMA guidelines were followed. A boolean expression was constructed to retrieve and (OR) to therapy were determined using either Fisher’s exact test or rank-sum test.
select articles from major medical databases. The R Studio package was used to Results: In the 29 pts enrolled between January 2020–February 2023, median age was
evaluate the potential of the diagnostic test. The Meta, Metadata and Mada packages 40.5 (range 23–66) and 41% were female. The initial tumor grade was 2 (n = 9), 3 (n = 8),
were utilised to evaluate Pooled accuracy, sensitivity and specificity. Results: A total of and 4 (n = 12). The OR rate was 14% (95% CI 3.9–32%), 1 complete response and 3 partial
responses. All cases were mismatch repair proficient. The median tumor mutation burden
35 studies and 7522 radiomic features were assessed through this meta analysis. The
(TMB) was 16.5, with TMB . 10 in 21 pts (75%). Baseline TMB was not associated with
Pooled Sensitivity and Specificity were estimated to be 86.70% ([74.85; 87.51], 95% CI,
response. The most common co-mutations were TP53 (n = 21, 75%), ATRX (n = 20, 71%),
p, 0.0001, I^2 = 92.7% [90.9%; 94.1%]) and 82.75% ([0.7912; 0.8587], 95% CI, p,0.0001,
ARID1A (n = 7, 25%), CIC (n = 4, 14%), and NF1 (n = 3, 11%), none were associated with
I^2 = 82.8% [77.4%; 86.9%]) utilising the random effects model. The pooled Accuracy was
response. There were no canonical mutations in BRCA1, BRCA2, or PALB2. Pathway
found to be 81.28% ([0.6037; 0.9253], 95% CI, p.0.01, I^2 = 0.0% [0.0%; 45.4%]). analysis on differentially expressed genes between responders and non-responders
Conclusions: Through the compilation of previously conducted studies, MRI based showed convergence on interferon signaling and inflammation among responders (p ,
Radiomics show High Pooled Sensitivity of 86.70% and High Pooled Specificity of 0.001). Lower pre-existing M2-polarized tumor associated macrophages/microglia (high
82.75% in the detection of Isocitrate Dehydrogenase mutations in Gliomas. Pooled expression of CD68, PDL1, CD163) was associated with response (p , 0.01). These
Accuracy rate of 81.8% indicates steady reliability of Radiomics in the prediction of IDH findings were supported by metaprograms in the HD spatial data, which showed higher
mutations. MRI based Radiomics is a dependable and consistent non invasive technique levels of CD8+ cytotoxic T-cells at baseline in responders. Conversely, M2-polarized
in the detection of IDH mutations in GBM and can be utilized for the generation of macrophage/microglia were enriched in non-responders. Paired progression samples will
predictive models, enhancing clinical diagnosis and tailored management based on IDH additionally be presented. Conclusions: Responders to olaparib and durvalumab had
mutation. Research Sponsor: None. decreased baseline M2-polarized macrophages/microglia and increased pre-existing
immunogenicity (interferon signaling). Several spatially conserved expression metapro-
grams targeting baseline immune infiltration were associated with response. Clinical trial
information: NCT03991832. Research Sponsor: None.

2076 Poster Session 2077 Poster Session

The utilization of palliative care services by patients with glioblastoma: A Identification of novel electrophysiologic biomarkers of cognition in glioma-
cross-sectional study with care partners of patients with GBM recruited from infiltrated cortex. First Author: Vardhaan Ambati, University of California, San
Facebook support groups. First Author: Lauren Robbins, UNC Wilmington, Wil- Francisco, San Francisco, CA
mington, NC Background: Diffuse gliomas, the most common primary brain cancers, often invade
Background: Glioblastoma(GBM) is a terminal brain cancer that has a rapid onset and speech-critical areas. Maximal resection improves survival, but damage to functional
results in symptoms such as headaches, vomiting, seizures, anxiety, depression, agi- cortex may cause permanent impairments. Direct cortical stimulation (DCS) differen-
tation, speech impairment, memory impairment, infections, brain bleeds, mobility tiates functional (DCS+) from nonfunctional (DCS-) cortex by temporarily disrupting
changes, changes in sleep patterns and cognitive changes. Palliative care helps patients neuronal activity, yet it remains unknown how DCS+ sites elicit transient impairments.
diagnosed with an incurable or chronic disease manage physical, social and psycho- DCS is technically challenging and resource-intensive. As a result, fewer than 50% of
logical symptoms while undergoing treatment, and has been shown to increase survival glioma patients receive optimal surgical care. This translational study aims to identify
for patients with cancer. Previous data of former care partners of patients with GBM electrophysiologic biomarkers of DCS+ cortex to 1) aid in safe resection by avoiding
indicated only 30% of patients used palliative care during the disease trajectory, whereas functional cortex and 2) to elucidate causal relations behind these transient impair-
90% engaged hospice during the end of life stage. Methods: To better understand the ments. Methods: Local field potentials of subdural array data from glioma infiltrated
reasons for underutilization of palliative care, primary caregivers of patients with GBM cortex was annotated as DCS+ or DCS- prospectively. We compared spectral elec-
recruited from a Facebook support group in February 2024 completed a 38-question trophysiologic variations (mean Theta [4-8 Hz], Alpha [8-13 Hz], Beta [13-30 Hz], and Full
survey about where care was received, who was on their care team, and whether they had Gamma [30-150 Hz] ranges) at resting state between DCS+ and DCS- sites using linear
discussions around palliative care. Inclusion criteria: current primary care partner of a mixed-effects models (to account for patient-level differences). Results: 1421 cortical
patient with GBM over the age of 18, and willing to participate (IRB exempt;H24-0393). sites of language were studied in 91 patients including 21 Oligodendroglioma WHO grade
The care partner was excluded if their patient was no longer living. Results: Of the 77 2-3, 19 Astrocytoma WHO 2-3, 3 Astrocytoma WHO 4, 48 IDHwt glioblastoma [GBM] WHO
care partners who participated in the study, the median age of the caregivers was 56 4). 115 (8.0%) were DCS+. After alignment to ECoG electrode arrays, 512 cortical sites
years (98% female) and the median age of the patients with GBM was 60 years (87% (49 DCS+) were assigned to electrodes. In oligodendrogliomas, DCS+ (N=16) vs DCS-
female). Patients with GBM were initially diagnosed in community hospitals, major (N=132) sites had higher alpha (77.7 6 111.9 vs 38.6 6 39.8, p=0.018), beta (23.1 6
medical centers, university medical centers, brain tumor centers, and as part of inci- 18.1 vs 11.9 6 17.9, p=0.033), and full gamma (0.6 6 0.6 vs 0.3 6 0.3, p,0.001) power.
dental findings. Approximately 1/4 of patients pursued second opinions. Of the 21 Similarly, in astrocytoma, DCS+ (N=13) vs DCS- (N=147) sites had significantly higher
patients initially diagnosed in community medical centers, over half received treatment theta (98.5 6 94.9 vs 57.1 6 67.2, p=0.020), alpha (83.5 6 76.3 vs 39.9 6 42.4,
in other facilities. Medical care team members differed by facility type with a marked p=0.003), beta (55.9 6 58.4 vs 16.3 6 17.8, p,0.001), and gamma (0.9 6 0.9 vs 0.4 6
difference in palliative care utilization. Only 27% of care partners reported palliative care 0.4, p=0.006) power. Interestingly, when comparing DCS+ (N=20) and DCS- (N=237) sites
was part of the care team. Notably,when care team members discussed palliative care in GBM patients, no significant differences were found in any studied ranges (all
with patients and their care partners (n = 21), 71% of the dyads utilized palliative care. In p.0.05). Conclusions: This study is the first of its kind to identify unique electro-
contrast, when palliative care was not discussed (n = 56), only 7% of the care partners physiological biomarker differences (at resting state) for oligodendroglioma and as-
used these services (p , 001). Participants reported they would have benefited from trocytoma speech cortex. It has two key implications. First, clinically, the identification
additional supportive services provided by patient navigation and palliative care. of electrophysiologic biomarkers may improve direct cortical stimulation (DCS) map-
Conclusions: This data highlights the value of discussions about palliative care in its ping: It can make surgeries faster by identifying cortex critical for cognition (speech)
utilization among GBM patients and their care partners. These conversations should based on these biomarkers, safer by helping neurosurgeons avoid resecting critical
take place early in the disease’s progression to ensure that care partners receive the regions, and more accessible. Second, this research suggests that different tumor types
necessary education and resources in a timely manner. Research Sponsor: None. (low-grade gliomas vs. GBM) remodel speech areas differently, prompting further in-
vestigation into tumor-specific effects on neural circuits. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 131s

2078 Poster Session 2079 Poster Session

A single-institution retrospective study of multicentric gliomas stratified by Diagnostic accuracy of machine learning models in glioma classification: A
IDH mutational status. First Author: Chuyin Yang, University of California, Los meta-analysis. First Author: Maya Gowda, Cornell University, New York, NY
Angeles, Los Angeles, CA Background: Machine learning (ML) is promising in IDH-based glioma classification using
Background: Multicentric glioma (MCG) is a subset of diffuse glioma that can be syn- magnetic resonance imaging (MRI), but variability in methods and algorithms necessitates a
chronous or metachronous and is defined as the occurrence of two or more tumor foci, with comprehensive evaluation. This meta-analysis assesses the pooled diagnostic performance of
separation of FLAIR (Fluid-attenuated Inversion Recovery) hyperintensity on MRI. MCG has ML-based approaches. Methods: A literature search was conducted in January 2025 across
not been extensively studied in studies stratifying IDH wild-type and mutant gliomas. This PubMed, MEDLINE, and Cochrane. MICCAI, RSNA, and SNO meeting abstracts were addi-
large single-institution study investigates the prevalence of MCG, examines the prognostic tionally reviewed. Eligible studies evaluating ML models for IDH-based glioma classification
implications of MCG, and characterizes metachronous MCG (mMCG) in a cohort that has using MRI were included. Data were pooled using a random-effects model, analyzing sen-
been stratified by IDH mutational status. Methods: In this IRB approved UCLA study, we sitivity, specificity, heterogeneity, and publication bias via Egger’s test and funnel plots. Leave-
identified diffuse glioma patients with known IDH mutational status and adequate MRI one-out analysis was conducted. Results: A total of 5982 cases were analyzed. Gliomas were
studies. Patients with multiple lesions on the MRI study pre-surgery or up to 3 months post- classified as WHO Grades II (11.5%), III (23.1%), and IV (65.4%). Histopathology-based ref-
surgery were considered synchronous MCG (sMCG). Patients who developed a new in- erence standards, including genetic and molecular testing, were used in 73.1% of studies, while
dependent lesion at least 6 months after initial surgery were considered mMCG. To qualify as immunohistochemistry, pathology, biopsy-proven markers, and immunohistopathologic di-
MCG, we identified additional tumors on MRI that had no overlapping FLAIR borders and met agnosis were each used in 3.8–7.7%. Deep learning models, including CNNs and ResNet, were
one or more of the following: pathologically confirmed with biopsy, exhibited growth and the most used classifiers (30.8%), followed by Support Vector Machines (26.9%). Ensemble
thickening over time, and developed or increased in enhancement. Difference in prevalence methods, such as Random Forest accounted for 19.2%, regression-based approaches (LASSO,
was compared using Student’s t-test. Kaplan-Meier and Cox-multivariate analyses were logistic regression) for 15.3%, and other techniques like multilayer perceptron and AdaBoost
for 7.7%. This meta-analysis included 25 studies for sensitivity and 26 for specificity, using a
used to analyze OS and time to metachronous (TtM) appearance. Results: We identified 911
random-effects model with DerSimonian-Laird estimation. Pooled sensitivity was 83.0% (95%
consecutive IDH wild-type, high-grade diffuse glioma patients from 2013-2023 and 515
CI: 79.5–86.5%) and specificity was 78.6% (95% CI: 73.7-83.4%), both statistically significant
consecutive IDH mutant patients from 2007-2024 with pre-surgical MRI or MRI within three
(p , 0.0001). Substantial heterogeneity was found (I² = 100% for both), with Cochran’s Q
months of initial surgery. From the examined cohort, we found 39 IDH mutants with 21 sMCG
values of 1.9e+06 for sensitivity and 4.2e+06 for specificity (p , 0.001). Leave-one-out
and 18 mMCG and 153 IDH wild types with 95 sMCG and 63 mMCG. In eight IDH wild-type analysis showed minimal variation in pooled estimates (sensitivity: 82.6–83.8%, specificity:
cases but no IDH mutant cases, mMCG arose from sMCG patients. We found that MCG had 77.6–79.4%). Egger’s test revealed significant small-study effects (p = 0.0001 for both),
higher prevalence in IDH wild-types than in mutants (WT = 16%, Mut = 7%, p , 0.0001), and suggesting potential publication bias. Conclusions: ML models demonstrated moderate di-
IDH mutant MCG showed more male predominance than IDH wild-type MCG (Mut = 73%, WT agnostic performance in IDH-based glioma classification, achieving a sensitivity of 83.1% and
= 58%, p , 0.0001). In IDH mutant patients, mMCG, but not sMCG, was associated with lower specificity of 78.6%. However, substantial heterogeneity and potential biases pose significant
OS (mMCG: HR = 2.476, p = 0.0115; sMCG: HR = 0.6437, p = 0.5027). However, in IDH wild challenges to their clinical implementation. To enhance the reliability and broader applicability
types both sMCG and mMCG and were associated with lower OS (mMCG: HR = 1.589, p = of ML models in IDH-based glioma diagnosis, standardization of imaging protocols and
0.0025; sMCG: HR = 1.347, p = 0.0332). There was no difference in TtM between the two external validation are imperative. Research Sponsor: None.
groups (HR = 0.5738, p = 0.5318). Amongst patients with multiple biopsied lesions, IDH wild
Meta-analytical findings.
types had consistent pathologies between lesions in all examined patients (29/29), but 71%
of IDH mutants exhibited different pathologies between lesions (5/7). Conclusions: Our Metric Sensitivity (%) Specificity (%)
study examined a cohort of adult diffuse gliomas stratified by IDH mutational status and Pooled Estimate 83.0 (79.5–86.6) 78.6 (73.7–83.4)
shows MCG is less common in IDH mutant gliomas and sMCG is not associated with worse Z-Value 46.4 31.76
P-Value ,0.0001 ,0.0001
prognosis. Further studies to identify molecular features underlying MCG will be valuable. I² (%) 100 100
Notes: For mMCG, FLAIR overlap might have occurred had the new lesion been observed T² 80.0 159.0
synchronously. Research Sponsor: Bradley Zankel Foundation; NIH/NCI P50 CA211015- Cochran’s Q 1.9e+06 (df=24, p,0.001) 4.2e+06 (df=25, p,0.001)
01A1 (UCLA SPORE in Brain Cancer). Leave-One-Out Range 82.6–83.8 77.6–79.4
Egger’s Test (P-value) 0.001 0.001

2080 Poster Session 2081 Poster Session

Identifying key molecular drivers of survival and therapeutic targets in Effect of ivosidenib and vorasidenib on 2-hydroxyglutarate levels in low
glioblastoma through integrated transcriptomic analysis. First Author: grade glioma: An in vivo MR spectroscopy study. First Author: Max Saint-
Mohammad Kashkooli, Harvard-MIT Health Sciences and Technology, Boston, MA Germain, Johns Hopkins University School of Medicine, Baltimore, MD
Background: Glioblastoma multiforme (GBM) remains one of the most aggressive brain Background: IDH-mutant gliomas are slow-growing infiltrating tumors of astrocytic
tumors, characterized by poor survival rates and limited therapeutic success. Advancing (AS) or oligodendroglial (OG) origin (WHO grade 2/3). The mutations occur in genes that
treatment requires identifying molecular drivers of tumor progression and therapy re- encode the metabolic enzyme IDH1 or, more rarely, IDH2 and lead to production of 2-
sistance. Integrated transcriptomic analyses offer a powerful means to uncover path- hydroxyglutarate (2-HG) that can be measured with optimized in-vivo MRS. Small-
ogenic pathways, therapeutic targets, and refine prognostic tools. This study aimed to molecule IDH inhibitors (IDHi) ivosidenib (inhibits mIDH1 enzyme) and vorasidenib
explore GBM’s molecular landscape to identify survival-relevant genes and actionable (inhibits mIDH1 and mIDH2 enzymes) showed good tumor penetrance and ~95% 2-HG
pathways, ultimately paving the way for precision therapies to improve outcomes. reduction measured in tumor biopsies. Both ivosidenib and vorasidenib have evidence of
Methods: RNA-sequencing data from TCGA, CGGA, CPTAC, GLASS, GSE121720, and responses in tumor growth rate. As volume reductions are often observed, but after a
GSE147352 datasets, encompassing 783 samples, were analyzed. Only primary, untreated delay of several months, thus an early response biomarker highly desired. The aim of this
GBM tumors with survival data were included. Data normalization was performed using study was to explore whether MRS measurements of 2-HG can be used to non-invasively
the voom function in limma, and batch effects were corrected using ComBat while monitor response to treatment with IDHmut-inhibiting drugs and compare this response
preserving survival-related and gender-specific variations. Differential expression anal- to changes in tumor volume. Methods: 14 patients (Age $18y) with a histomolecularly
ysis (DEA) was used with thresholds of |logFC| . 1 (two-fold expression change) and confirmed IDH1 mutated diffuse glioma (AS or OG) received ivosidenib (n = 12) or
adjusted p , 0.001, accounting for age, gender, race, and IDH1 mutation. Protein–protein vorasidenib (n = 2) therapy as part of their routine clinical care. MRI was performed
interaction (PPI) networks were constructed using STRING, and Cox regression identified before treatment (baseline) and repeated (follow-up) with a median on-drug follow-up of
survival-related genes. DrugBank was used to link survival-associated genes to potential 6 months [4, 12 IQR]. Tumors were segmented from FLAIR images in 3D Slicer and
therapeutics. Results: Of the 783 samples, 488 met inclusion criteria (473 tumor, 15 non- volumes were calculated in cm3. All MRS data were processed in Osprey with the built-in
tumor). DEA identified 1,453 differentially expressed genes (DEGs) with significant LCModel fitting module. Comparisons between baseline and follow-up measured me-
differences between tumor and non-tumor samples. PPI analysis highlighted 270 hub tabolite levels were conducted using paired t-tests or (in cases where the normality
genes, of which 47 were significantly associated with survival. Notably, CDK1, RRM2, and assumption was not met) non-parametric Wilcoxon tests. Results: We analyzed spectra
BIRC5 showed negative prognostic effects (hazard ratio [HR] . 1.7, adjusted p , 0.05), from 11 patients with both baseline and follow-up sessions. All spectra exhibited
while RPL3L, RPL21, and RPL9 exhibited protective effects (HR , 0.5, adjusted p , characteristic tumor features: reduced total N-acetylaspartate (tNAA) and elevated
0.001). DrugBank analysis identified gallium nitrate (targeting RRM2) and Alsterpaullone levels of total choline (tCho), lactate (Lac), and myo-inositol (mI), along with a 2-HG peak
(inhibiting CDK1) as promising therapeutic candidates. Conclusions: This study at 2.25 ppm that is visibly smaller after treatment. The decrease in 2-HG levels was
provides a comprehensive bioinformatics analysis of GBM, integrating multiple tran- highly significant (p , 0.001) across all included patients undergoing ivosidenib/
scriptomic datasets to identify critical survival-related genes, including CDK1, RRM2, and vorasidenib therapy, regardless of reference standard. Volumetric assessment revealed
BIRC5, as key therapeutic targets. Notably, it highlights the protective roles of ribosomal tumor growth arrest and a subtle reduction in tumor growth in some individuals.
proteins (RPL7, RPL9, RPL21), challenging the traditional view that ribosomal upregu- Conclusions: This is the first in-vivo evidence using MRS that ivosidenib/vorasidenib
lation solely drives tumorigenesis. These findings emphasize the dual nature of ribosomal reduces 2-HG. We found that 2-HG levels respond specifically and rapidly to treatment,
dysfunction, which has been implicated in both oncogenic and tumor-suppressive while volumetric changes manifest slowly and more gradually, consistent with previous
pathways. Through rigorous data normalization, batch correction, and PPI analysis, studies. This preliminary study suggests that in vivo MRS-derived 2-HG estimates could
the findings offer robust insights into GBM biology and its molecular drivers. These results serve as sensitive and specific biomarkers for monitoring low grade gliomas in vivo in
lay a strong foundation for future studies to validate their clinical relevance, refine response to small-molecule IDH inhibitor therapy after initiation of treatment. Longi-
prognostic models, and advance precision therapy strategies for GBM patients. Research tudinal volumetric and radiomic analyses are underway. Research Sponsor: None.
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
132s CENTRAL NERVOUS SYSTEM TUMORS

2082 Poster Session 2083 Poster Session

Vault proteins as prognostic biomarkers and therapeutic targets in lower- Accelerator-based boron neutron capture therapy, a randomized controlled
grade gliomas. First Author: Sebawe Syaj, Division of Hematology and Oncology, trial for refractory recurrent high-grade meningiomas. First Author: Shin-Ichi
Department of Medicine, University of Pittsburg Medical Center, and UPMC Hillman Miyatake, Kansai BNCT Medical Center, Osaka Medical and Pharmaceutical University,
Cancer Center, Pittsburgh, PA Takatsuki, Japan
Background: Vault proteins, including MVP, VPARP, and TEP1, are components of the Background: High-grade meningiomas (HGMs) recurred after X-ray treatment showed pes-
vault complex, and are involved in drug resistance, DNA repair, and cell survival. simistic prognosis. We conducted “A phase II investigator-lead RCT using accelerator-based
However, their role in low-grade gliomas (LGG) remains unclear. This study explored the BNCT system for refractory recurrent high-grade meningioma”. Methods: We prepared 2 study
expression and prognostic significance of LGG with the aim of uncovering their potential arms, BNCT test treatment arm (12 subjects) and control best supportive care arm (6 subjects)
as biomarkers. Methods: Using the TCGA-LGG cohort, Kaplan Meier (KM) and uni/ in RCT fashion. PFS judged by the third-party committee was primary endpoint and PFS judged
by investigators themselves and OS of BNCT arm and so on were secondary endpoints. Rescue
multivariate Cox proportional hazards regression (CPH) analyses were performed using
BNCT was permitted for control group patients, if they were judged as PD by investigators. First
R 4.3.3, and hazard ratios (HR) with 95% confidence intervals (CI) were reported. MVP, patient-in and last patient-in were August 2019 and August 2021, respectively. Last patient visit
TEP1, and VPARP (PARP4) expression levels were further stratified according to the LGG was February 2024 and final OS survey and final observation for effectiveness and safety in
subtype. Multi-gene KM plots were generated to stratify overall survival (OS), BNCT arm were performed in July 2024. These results were compared to EORTC’s RCT of
progression-free survival (PFS), and disease-specific survival (DSS) using gene set trabectedin. Results: Three and 2 grade 3 subjects were included in BNCT and control arm,
variation analysis (GSVA) scores. “ImmuCellAI” algorithm was utilized for immune respectively. Others were grade 2 subjects. All cases were confirmed relapse after some ra-
infiltration analysis. The GDSC and CTRP databases were used for drug sensitivity diotherapy in follow-up images. One subject allocated in BNCT arms was excluded after en-
analyses. Results: As indicated by KM and univariate CPH analyses, MVP (hazard ratio rollment due to protocol violation. At the end of the observation, as primary endpoint, PFS of
[HR] = 1.5, 95% CI: 1.3-1.8), TEP1 (HR = 1.7, 95% CI: 1.4-2.0), and VPARP (HR = 1.6, 95% each arm judged by committees showed statistical significance (p=0.0157, Log-rank). As one of
CI: 1.3-1.9) predicted poor OS. MVP, in multivariate CPH model adjusted to tumor grade the secondary endpoints, PFS of each arm judged by investigators also showed statistical
and histology, remained significant (HR = 1.37, 95% CI: 1.14-1.70). The expression of significance (p=0.0002). Median PFS judged by committees were 14.4 (95% CI:7.9-26.4) and 1.4
MVP, TEP1, and VPARP was the highest in astrocytomas and the lowest in oligo- (1.0-9.0) months for BNCT and control arm, respectively. Median PFS judged by investigators
dendrogliomas (p , 0.05), with oligoastrocytomas in between. 3-gene KM signature showed 14.7 (7.6-22.8) and 1.5 (1.0-9.0) months, respectively. Five out of 6 cases in control arm
received rescue BNCT after PD assessments. Other endpoints are listed in the table.
analysis revealed a negative association between higher GSVA scores and OS, PFS, and
Conclusions: As primary endpoint, PFS judged by committee showed statistical significance
DSS (log-rank p , 0.01) (Table). Immune infiltration analysis indicated positive
between treatment and control arms. Recently, the results of RCT of “Trabectedin” for recurrent
macrophage, Th1, Th2, and dendritic cell infiltration with higher GSVA (derived from HGMs, organized by EORTC, (EORTC-1320-BTG) was reported. Unfortunately, there was no
MVP, TEP1, and VPARP) (r . 0.40, FDR , 0.05) and negative infiltration of naive CD8+ effect of Trabectedin not only in PFS but in OS. Therefore, EORTC’s report seems to be natural
and neutrophils (r , -0.30, FDR , 0.05). Drug analysis revealed vincristine resistance course of recurrent refractory HGMs. Our current BNCT shows extremely excellent results in
with higher MVP expression (r = 0.44, FDR , 0.001). Conclusions: MVP, TEP1, and comparison with EORTC’s RCT in mPFS, PFS-6 months, mOS, OS-1 year and ORR (Table). Clinical
VPARP were associated with poor survival outcomes and distinct immune infiltration trial information: 2051190044. Research Sponsor: AMED.
patterns in LGG. These findings highlight the potential of vault proteins as biomarkers Comparison of both RCT (current study and EORTC trabectedin).
and therapeutic targets for LGG. Research Sponsor: None.
Current study (n=18) EORTC trabectedin (n=90)
Univariate Cox proportional hazards ratios for the different survival types for BNCT Control Trabectedin Control
GSVA scores of the three major proteins of the vault (MVP, TEP1, and VPARP).
mPFS (months) 14.4 1.4 2.43 4.17
Survival type Hazard Ratio Cox P value Logrank P value Higher risk of death PFS-6 months (%) 100 44.4 21.1 29.1
mOS (months) 46.9 - 11.4 10.6
OS 1.7 ,0.001 24E-2.03 Higher GSVA OS-1 year (%) 100.0 - 48.1 43.0
PFS 1.43 0.01 9.09E-03 Higher GSVA OS-2 year (%) 90.9 - - -
DSS 1.75 ,0.001 2.18E-03 Higher GSVA ORR (CR+PR) (%) 27.3 0 1.6 0

2084 Poster Session 2085 Poster Session

Phase 2 study of nivolumab for patients with meningiomas refractory to Analysis of genetic mutation profile and CNS pharmacokinetics in relapsed/
surgery and radiotherapy with immune-related response criteria. First Author: refractory primary CNS lymphoma patients responding to novel emavusertib
Hikaru Sasaki, Tokyo Dental College Ichikawa General Hospital, Ichikawa, Japan (IRAK4i) and BTKi combination. First Author: Christian Grommes, Department of
Background: The majority of meningiomas, which are the most common central nervous Neuro-oncology, Memorial Sloan Kettering Cancer Center, New York, NY
system (CNS) tumors, are benign and often cured by surgical resection alone. However, Background: Primary Central Nervous System Lymphoma (PCNSL) is a rare and ag-
20%–30% of meningiomas can be malignant tumors of CNS WHO grade II or III that are gressive non-Hodgkin lymphoma with no approved treatments for relapsed/refractory
refractory to repetitive resection and radiotherapy. Moreover, a proportion of grade I (R/R) patients, representing a critical unmet need. MyD88 mutations in ~70% of PCNSL
meningiomas is associated with an aggressive clinical course reminiscent of grade II patients drive Interleukin-1 receptor associated kinase 4 (IRAK4) activation, promoting
tumors. Reports of effective medical therapy for those tumors are extremely rare. NF-kB signaling, inflammation, and tumor progression. Emavusertib, a potent oral IRAK4
Methods: A single-arm, open-label, phase 2 study was conducted to evaluate the efficacy inhibitor, crosses the blood-brain barrier and shows preclinical synergy with Bruton
and safety of nivolumab for meningiomas refractory to surgery and radiotherapy. tyrosine kinase inhibitors (BTKi), re-sensitizing BTKi-resistant cell lines. This study
Nivolumab (480 mg) was administered intravenously every 4 weeks and continued until evaluates the molecular and pharmacokinetic (PK) data associated with responses to
tumor progression or unacceptable toxicity for up to 365 days. The primary endpoint was emavusertib + ibrutinib combination therapy in R/R PCNSL patients. Methods: The
the objective response rate (ORR) determined by a central independent review committee. safety, clinical activity, and potential biomarkers of emavusertib in R/R PCNSL are being
With a one-sided significance level of 5%, a power of 80%, a threshold response rate of 5%, investigated in the ongoing open-label, Phase 1/2 TakeAim Lymphoma trial
and an expected response rate of 20%, the required sample size was calculated to be 27 (NCT03328078). Pre-dose and 1.5-hour post-dose plasma samples were collected on
patients using the exact binomial test. Considering a 10% attrition rate, the target sample Cycle 3 Day 1, Cycle 5 Day 1 and Cycle 7 Day 1. Cerebrospinal fluid (CSF) samples were
size was set at 29. To avoid premature discontinuation of potentially effective immu- obtained via a lumbar puncture within 1.5 hrs of collection of the post-dose plasma PK
notherapy, response was evaluated based on the iRANO (meningioma) criteria, which is sample on Cycle 3 Day 1. Mutation analysis was based on patients’ molecular pathology
based on the RANO (meningioma) criteria (Neuro Oncol 21(1):26-36, 2019) with the reports provided by trial sites. Sequencing of archival tissues, CSF and plasma are in
integration of the immune-related response criteria outlined previously (Lancet Oncol progress. Results: As of 06 December 2024, CSF concentration data were available for 7
16(15):e534-e542, 2015). Archival tumor specimens from all 29 cases were obtained for PCNSL patients. The mean emavusertib concentration in CSF was 81.3 ng/ml (54.7-
biomarker analyses. Results: A total of 29 patients started the study therapy. Response 104.0) in patients receiving 100 mg emavusertib BID (n = 4). In patients receiving 200 mg
assessment by the central review committee was performed for 28 patients: grade I emavusertib BID (n = 3), the mean emavusertib concentration in CSF was higher at 175.7
meningioma in 5, grade II in 19, and grade III in 4 by definition of the 2016 WHO criteria. ng/ml (114.8-209.4), which is 2.2X the mean value in patients who received 100 mg
The best overall response was PR in 1, SD in 13, and uPD/cPD in 14. The ORR was 3.6% emavusertib BID (p-value = 0.02). All 7 patients received 560 mg ibrutinib QD, and the
and progression-free survival at 6 months was 23.9%. Biallelic inactivation of the NF2 ibrutinib concentrations in the CSF were consistent with findings from previously
gene was detected in 20/27 cases (74%), whereas biallelic inactivation of the CDKN2A published clinical studies. MyD88 mutation status was available for 7 patients of which
gene was identified in 7/27 cases (26%). One patient who had multiple grade I me- all had prior exposure to BTKi regimens. Among these, 6 patients had MyD88 mutation of
ningiomas with biopsy-proven lung metastases showed near CR following initial ra- which 4 patients had responded (3 complete responses and 1 partial response) to
diological progression. The TMB of the tumor was 8.1/MB. Next-generation sequencing emavusertib + ibrutinib combination with duration of response (DOR) up to 18.9 months
found that none of the tumors had mutations of the DNA mismatch repair genes. with data collection ongoing. Conclusions: Preliminary CNS pharmacokinetic data
Nivolumab was well tolerated. Conclusions: Although nivolumab monotherapy failed to demonstrates that emavusertib concentration in CSF increases with increasing ema-
meet the prespecified primary endpoint, our study demonstrated that a subset of patients vusertib dose. Patients with MyD88 mutations showed expected promising preliminary
could benefit from the therapy and that immune-related response criteria are necessary to efficacy to emavusertib + ibrutinib combination and may overcome BTKi resistance.
evaluate immunotherapy for meningiomas. Clinical trial information: jRCT2031190074. Enrollment in this trial is ongoing. Clinical trial information: NCT03328078. Research
Research Sponsor: Ono pharmaceuticals. Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 133s

2086 Poster Session 2087 Poster Session

Unraveling survival disparities in primary central nervous system (CNS) Predicting survival in malignant meningiomas: A machine learning
lymphoma: An analysis of race, socioeconomic factors, and treatment approach. First Author: Mustafa Alshwayyat, Jordan University of Science and
outcomes using the Surveillance, Epidemiology, and End Results program Technology, Irbid, Jordan
(2000–2021). First Author: Imran Khan, NYC Health and Hospitals/Woodhull, Background: Intracerebral meningiomas account for over 90% of all meningioma cases, with
Brooklyn, NY only 1–3% classified as malignant. Malignant meningiomas remain understudied compared
Background: Primary central nervous system lymphoma (PCNSL) is a rare B-cell non- with other brain tumors. This study is the first to apply machine learning (ML) to identify
Hodgkin lymphoma with survival outcomes influenced by treatment, demographic, and prognostic factors and improve outcomes of malignant intracerebral meningiomas. Methods:
socioeconomic factors (Villano JL et al., Br J Cancer, 2011). This study evaluated Data were obtained from the SEER database (2004-2021). Patients who met any of the
survival disparities associated with race, socioeconomic status (SES), and treatment following criteria were excluded: diagnosis not confirmed by histology; previous history of
cancer or other concurrent malignancies; or unknown data. To identify prognostic variables, we
modalities in PCNSL patients using a large U.S. population database. Methods: This
conducted Cox regression analysis and constructed prognostic models using ML algorithms to
retrospective cohort study used the SEER-17 database to analyze data from 7,068
predict the 5-year survival. Patient records were randomly divided into training (70%) and
patients diagnosed with PCNSL between 2000 and 2021. Demographic, socioeconomic, validation (30%) sets. A validation method incorporating the area under the curve (AUC) of the
and treatment data were collected. Kaplan-Meier analysis was used to compare survival receiver operating characteristic curve was used to validate the accuracy and reliability of the
across groups, and Cox proportional hazards models identified independent prognostic ML models. We also investigated the role of multiple therapeutic options using Kaplan-Meier
factors. Results: The cohort included 7,068 patients (52.3% male; mean age: 63 years, survival analysis. Results: A total of 1,363 patients were included. Most patients were White
SD 6 15). Racial distribution was 63.9% Caucasians, 16.0% Hispanics, 12.2% Asian/ (71.8%) or female (56.7%). The median patient age was 62 years, and the median tumor size
Pacific Islanders, 7.3% African Americans, and 0.4% American Indian/Alaskan Natives. was 4.8 cm. Most of the tumors were localized (67.8%). Adjuvant radiation therapy was
Among these, 27.3% received radiation, and 64.3% received chemotherapy. During the administered to 50.2% of the patients. Patients aged , 62 years exhibited better 5-year
study period, 73.5% of patients died from PCNSL. Survival analysis revealed that Asian/ survival rates, with an overall survival (OS) of 79.4% and cancer-specific survival (CSS) of 82%,
Pacific Islanders had the longest median overall survival (OS) at 22 months (95% CI: compared to those aged $ 62 years, who had an OS of 40.9% and CSS of 50.6%. Tumors
16.5–27.5), followed by Hispanics (16 months; 95% CI: 11.8–20.2), Caucasians smaller than 4.5 cm were associated with higher survival rates (OS: 67.7%, CSS: 72.4%) than
(11 months; 95% CI: 9.8–12.2), and American Indian/Alaskan Natives with the shortest larger tumors (OS: 53.6%, CSS: 61.9%). The impact of adjuvant radiation therapy showed an OS
survival at 5 months (95% CI: 0–11.2) (p,0.001). Socioeconomic analysis showed a of 59.9% and 64.5%, respectively, compared with those who did not receive radiation, with an
direct association between higher income and improved OS: patients with household OS of 59.5% and CSS of 68.7%. Multivariate Cox regression analysis identified older age (HR:
incomes $75k had a median OS of 13 months (95% CI: 11.2–14.8), compared to 3.6, 95% CI: 3.03–4.4) and large tumor size (HR: 1.4, 95% CI: 1.22–1.7) as poor prognostic
factors. The Random Forest and MLP classifiers were the most accurate models. The ML
6 months (95% CI: 4.2–7.8) in those earning ,50k (p,0.001). Multivariable Cox re-
models identified age as the most significant prognostic factor. The performance metrics for all
gression identified male sex (HR 1.21, p,0.001) and older age (HR 1.027, p,0.001) as
the ML algorithms are summarized in Table. Conclusions: This study underscores the
adverse prognostic factors, while chemotherapy significantly improved survival (HR transformative potential of ML in enhancing personalized medical approaches for malignant
0.43, p,0.001). Radiation provided a modest benefit (HR 0.913, p=0.005). intracerebral meningiomas. Furthermore, whether the benefits of adjuvant radiotherapy
Conclusions: This large study demonstrates that lower income levels and racial dis- outweigh the risks remains unclear, indicating the need for further targeted research to in-
parities are associated with reduced survival in PCNSL. Findings underscore the need for vestigate its therapeutic impact on these rare tumors. Research Sponsor: None.
equitable healthcare access and tailored therapeutic strategies to address these in-
equities. Keywords: CNS lymphoma, survival disparities, socioeconomic status, race, ML Model Accuracy Precision Recall F1 score AUC
treatment outcomes, public health oncology. Research Sponsor: None. LR 63% 50.8% 61.2% 55.6% 0.696
KNN 63.3% 51.2% 45.1% 48% 0.660
RFC 69.1% 58.9% 60.2% 59.5% 0.743
GBC 66.2% 55.6% 52.6% 54.1% 0.723
MLP Classifier 67.48% 57.47% 53.76% 55.56% 0.716

TPS2088 Poster Session TPS2089 Poster Session

Phase IIa study of aDC1 vaccines targeting HER2/HER3 combined with A multicenter, randomized, controlled, pivotal trial of microbubble-
pembrolizumab in patients with asymptomatic brain metastasis from breast enhanced transcranial focused ultrasound for patients with NSCLC brain
cancer. First Author: Shipra Gandhi, Roswell Park Comprehensive Cancer Center, metastases (LIMITLESS). First Author: Manmeet Singh Ahluwalia, Miami Cancer
Buffalo, NY Institute, Baptist Health South Florida, Miami, FL
Background: Brain metastases develop in up to 50% of patients (pts) with metastatic Background: The efficacy of systemic therapies for brain metastases (BM) is hindered
breast cancer. Overexpression of HER3 in brain metastatic breast cancer (BMBC) is a by the blood-brain barrier (BBB) and brain-tumor barrier. Transcranial low-intensity
resistance factor to HER2-targeted therapies and a driver of brain metastasis. Disease focused ultrasound combined with IV microbubble oscillators (MB-FUS), allows for
progression is associated with loss of anti-HER2 and anti-HER3 immunity. Previously, localized, controlled, non-invasive and temporary BBB opening, which has been shown to
we have demonstrated that glioma-specific peptide-loaded aDC1 which produces enhance tumor drug delivery of systemic therapies, as well as impro efficacy of im-
CXCL9, CXCL10, CXCL11, and CCL5, the chemokines that attract CXCR3- and CCR5- munotherapies. Non-small cell lung cancer (NSCLC) is the most common cause of BM,
expressing cytotoxic T-lymphocytes (CTLs) and T-helper 1 (Th1) cells, induce clinical and this randomized controlled trial (RCT) aims to evaluate the safety and efficacy of
responses and long-term disease stabilization in pts with aggressive recurrent primary MB-FUS-mediated BBB opening combined with standard of care (SOC) systemic therapy
brain tumors (Okada et al. JCO 2011. PMID: 21149657). We hypothesized that anti- versus systemic therapy alone for patients with NSCLC BM. Methods: LIMITLESS is
HER2/3-loaded aDC1 combination with PD1 blockade will result in a strong Th1/CTL prospective, multicenter, parallel-arm, RCT, ongoing at up to 30 centers, that randomizes
response against HER2/3 epitopes (Basu A et al. Cancer Immunol Res. 2022 PMID: patients with NSCLC BM, in a 2:1 ratio to either: (i) Arm 1: MR-guided MB-FUS plus all
34785506) that will translate into anti-cancer benefit in the central nervous system FDA approved on-label use of immune checkpoint inhibitors (ICIs) with or without
(CNS) and systemically. Methods: This is a phase II single-arm, non-randomized chemotherapy regimen (SOC systemic therapy), or (ii) Arm 2: SOC systemic therapy
multicenter study (NCT04348747). Eligibility includes pts with BMBC $18 years, alone. Included patients are $18 years aged, with normal organ function, KPS $70, and
ECOG PS #1, normal marrow and organ function with asymptomatic untreated brain have $0.5 cm size BM meeting measurable disease criteria as per RANO-BM. Patients
metastases $ 5 mm. The study subjects receive aDC1 q3 weeks x 3 along with on both arms receive standard-of-care therapy, while those on arm 1 also undergo MB-
pembrolizumab every 3 weeks. Thereafter, aDC1 booster doses can be administered FUS. Patients undergo pre-treatment brain MRI, followed by IV administration of
every 3 months until disease progression, intolerable side effects, or withdrawal from microbubbles for enhanced sonication effects. BBB opening is performed using a
study, up to 24 months. Baseline and 9-week post-aDC1 peripheral biopsies (non-CNS) transcranial 220 kHz device with 1024-element phased array transducer with real-time
are required for six pts. The primary endpoint is CNS response rate (RR) by RANO-BM acoustic feedback-based power control for maintaining effective microbubble activa-
criteria. If no CNS response is observed after 12 pts, the study will be terminated. If $ 1 tion. The primary study outcome is the overall objective response rate (ORR) at 6 months
response is observed, then 9 more pts will be enrolled, for a total of 21 pts. If $ 3 CR are as assessed using RANO-BM criteria. Using a Bayesian design for power analysis, a
observed, the proposed therapy will be considered promising for further evaluation. superior ORR of 60% is assumed for MB-FUS arm versus 30% in the control arm for a
Secondary endpoints include non-CNS RR per RECIST v1.1, median CNS, non-CNS and total sample size of N = 96, 64 participants in MB-FUS and 32 in control arm, for 80%
overall progression-free survival, overall survival, and safety. Exploratory endpoints power using a two-sided chi-square test with an alpha of 0.05. For the upper-bound
include changes in intratumoral biomarkers (CTLs, PDL1, chemokines) in pre- and post- estimate, ORR of 45% in MB-FUS arm and 30% in the control arm, the study needs N =
treatment peripheral tumor biopsies and immune changes in the blood. So far, 7 of the 369 participants: 246 in LIFU arm and 123 in control arm. The secondary outcomes are
planned 21 pts have been enrolled. Clinical trial information: NCT04348747. Research best objective response rate and median time-to-response per treatment arm. Ex-
Sponsor: U.S. Department of Defense. ploratory outcomes are median progression-free survival (PFS), overall survival (OS),
median intracranial PFS, median extracranial PFS, and quality of life. Patient enrollment
commenced in 2022 and is ongoing ([Link] Registration: NCT05317858).
Clinical trial information: NCT05317858. Research Sponsor: Insightec Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
134s CENTRAL NERVOUS SYSTEM TUMORS

TPS2090 Poster Session TPS2091 Poster Session

Delayed or upfront brain radiotherapy in treatment-naı̈ve lung cancer pa- FORTE: A phase 2 master protocol assessing plixorafenib for BRAF-altered
tients with asymptomatic or minimally symptomatic brain metastases and cancers. First Author: Karisa C. Schreck, Johns Hopkins School of Medicine, Baltimore,
ALK rearrangements (DURABLE). First Author: Joshua David Palmer, Department MD
of Radiation Oncology, Ohio State University, Columbus, OH Background: Plixorafenib (FORE8394; PLX8394) is a novel, oral, small-molecule BRAF in-
Background: Patients with non-small cell lung cancer (NSCLC) with ALK rearrange- hibitor highly selective for BRAF V600 monomers and BRAF-containing dimers. Plixorafenib
ments have a high frequency of brain metastases. Alectinib was shown to be superior to binding disrupts RAF dimerization, targeting both BRAF V600 mutations and fusions, thereby
crizotinib in the first-line treatment of patients with ALK-positive NSCLC in the ALEX trial, preventing paradoxical activation and avoiding the need for combination with a MEK inhibitor.
and the intracranial response rate (CNS ORR) was 85.7% with alectinib versus 71.4% with In a phase 1/2a study, plixorafenib demonstrated promising safety and clinical activity across a
crizotinib in patients who received prior radiotherapy and 78.6% versus 40.0%, re- range of doses tested in tumors with BRAF V600 mutations or fusions. The most common
spectively, in those who had not. Alectinib has also shown benefit in earlier stages of adverse events (AEs) included predominantly low-grade liver function test changes and grade 1
fatigue, nausea, diarrhea, and vomiting. Methods: The FORTE Phase 2 basket study is
NSCLC. Given the high intracranial efficacy rate demonstrated by alectinib, as well as the
currently enrolling patients $10 years of age into 4 sub-protocols. Study details are shown in
known toxicities of cranial irradiation, the role of early irradiation of CNS disease vs theTable. Eligible patients have received prior therapy for advanced disease, have measurable
delaying radiation in favor of treatment with alectinib needs to be defined to inform disease, and have a Karnofsky ($16 years) or Lansky (,16 years) Performance Score of $60
clinical practice. Methods: NCT05987644 is a multi-center, multi-cohort study con- at study entry. All patients receive plixorafenib continuous dosing, in some cohorts coad-
sisting of a Phase 1b and Phase 2 portion. The Phase 1b portion of the study is a single- ministered with cobicistat, a pharmacokinetic (PK) booster. Prior MAPK inhibitor therapy is
arm, open label study of alectinib in patients with CNS disease. Twelve subjects will be excluded unless otherwise specified below. As of January 2025, the trial is recruiting par-
enrolled in the Phase 1b portion of the study and treated with alectinib alone; patients ticipants in 9 countries globally, with 54 sites activated. Clinical trial information:
with PD will come off study treatment and move on to standard of care treatment per NCT05503797. Research Sponsor: Fore Biotherapeutics.
national guidelines. The phase 2 portion will be a randomized, non-blinded, open-label Sub-Protocol
study. Forty-four subjects will be enrolled and randomized 1:1 to either alectinib upfront Sub-Protocol A Sub-Protocol B C Sub-Protocol D
(Arm A) or alectinib + SRS (arm B). A group sequential design will be implemented with Patient Population Advanced solid and BRAF V600- Rare1 BRAF BRAF V600-mutated
one interim analysis for futility and, and one final analysis using the composite outcome. primary CNS tumors mutated V600-mutated melanoma2 or thyroid cancer
harboring BRAF recurrent primary advanced solid without anaplastic or
The primary objective of phase 1b is to determine the safety and feasibility of delayed fusions CNS tumors tumors undifferentiated components
brain radiation in patients with ALK fusion positive NSCLC and CNS metastases. The Planned Enrollment ~100 ~50 ~75 ~12
Design Single-arm, open-label, Bayesian optimal phase 2 design 1:1 randomized, open-label crossover
primary objective of the phase 2 study is to determine whether treatment with alectinib design to compare plixorafenib
results in preserved neurological status and control of CNS disease at 12 months administered alone and with
PK booster
compared to alectinib plus SRS. Secondary endpoint will be intracranial progression free Planned Efficacy Interim N=25 N=25 N=25 None
Analyses N=50 N=50
survival at 12 months (icPFS12), response rate and icPFS, OS, and safety and tolerability. Primary Endpoint ORR 3
Intra-patient PK
The study is open and accruing at 4 sites. Clinical trial information: NCT05987644. Key Secondary DOR, DCR, PFS, OS, PK, Safety Safety, ORR, DOR, DCR, PFS, OS, Safety
Endpoints
Research Sponsor: Genentech. Key Exploratory Longitudinal ctDNA assessments4
Endpoint
1
BRAF V600-mutated tumors occurring in #40,000 US patients annually (eg, ovarian/gynecologic cancers, cholangiocarcinoma,
small intestinal/gastrointestinal cancers other than colorectal adenocarcinoma, neuroendocrine cancers).
2
Patients with melanoma should have received and not tolerated a prior BRAF inhibitor.
3
Response assessed by BICR using RECIST v1.1 for solid tumors or RANO HGG or LGG for primary CNS tumors. ORR for primary CNS
tumors using RANO 2.0 is an exploratory endpoint. Tumors assessed at cycle 1 day 1, every 9 weeks for 48 weeks, then every
12 weeks.
4
Plasma for all patients; plasma and CSF for patients with primary CNS tumors.

TPS2092 Poster Session TPS2093 Poster Session

Retifanlimab with bevacizumab and hypofractionated radiotherapy to treat PEAR-GLIO: Clinical evaluation of an AI-driven functional precision medi-
recurrent glioblastoma. First Author: Nishika Karbhari, Mayo Clinic, Rochester, MN cine platform for therapeutic efficacy in gliomas. First Author: Matthew Williams,
Background: Glioblastoma (GBM) is the most common primary brain malignancy in Ourotech Ltd t/a Pear Bio, London, United Kingdom
adults. GBM is universally recurrent and associated with dismal outcomes. Re- Background: Gliomas and other primary brain tumors remain a leading cause of cancer-
irradiation (reRT) is ideal for evaluating combination therapy for recurrent GBM related mortality, with limited predictive biomarkers to guide therapy selection. The
(rGBM) due to its multifactorial mechanism of action, including downstream immu- PEAR-GLIO trial investigates the use of Pear Bio’s AI-driven ex vivo platform to assess
nomodulatory activity. RT (especially multi-fraction) increases immunogenicity in the therapeutic sensitivity of FDA-approved and experimental treatments on patient-
preclinical models by promoting immune activation, immune migration, and antigen derived 3D immune-microtumors. This observational study seeks to validate whether
uptake. Additionally, a recent study demonstrated enhanced PD-L1 expression in the this platform can provide actionable insights for patient stratification and treatment
glioma tumor microenvironment (TME) following RT, and combining stereotactic RT optimization in subsequent trials. The trial also incorporates patient and public in-
with a PD-1 inhibitor improved survival in murine models. Retifanlimab is a humanized volvement and engagement (PPIE) to understand perspectives and enhance study
monoclonal anti-PD1 IgG4 antibody that received FDA approval for adults with met- design and accessibility. Methods: PEAR-GLIO (NCT06038760) is a UK-based, obser-
astatic or recurrent locally advanced Merkel cell carcinoma. Bevacizumab, an anti-VEGF vational study enrolling 50 patients diagnosed with operable primary brain tumors,
antibody, is a treatment for radiation necrosis/cerebral edema with less immune including grades 2–4 gliomas. Inclusion criteria require histologically confirmed ma-
suppression than corticosteroids. In a previous Phase 2 study, hypofractionated RT lignancy, the ability to provide $0.4g of tumor tissue and 40mL of whole blood, and
(HFRT), retifanlimab, and bevacizumab was associated with a 9-month overall survival consent for data and sample use. Exclusion criteria include pre-surgical chemotherapy
(OS) rate of 71.4%. To demonstrate the efficacy of this regimen compared to HFRT and or radiotherapy within 30 days and inoperable disease. Tumour-extracted and immune
bevacizumab, we have designed a new randomized controlled Phase 2 trial. We hy- patient cells are cultured as physiologically-relevant 3D immune-microtumors and
pothesize that combination reRT with retifanlimab will produce a more robust anti-tumor exposed to FDA-approved and experimental treatments. Phenotypic and molecular
immune response and improve OS compared to reRT without retifanlimab. responses, including changes in tumour viability, cell death, migration, immune cell
Methods: This is a multicenter, open-label, randomized, controlled Phase 2 study of infiltration are assessed using live imaging and computer vision. The study uniquely
retifanlimab, bevacizumab, and HFRT for adult patients with rGBM. Patients are ran- integrates real-time confocal imaging and omics analyses to evaluate drug mechanisms
domized 1:1 to the experimental (bevacizumab + retifanlimab + HFRT) or control cohort of action. This includes correlation of ex vivo responses with biomarkers such as MGMT
(bevacizumab + HFRT). Key eligibility criteria include age $ 18 years, Karnofsky methylation, IDH mutation, and 1p/19q co-deletion, alongside exploratory analysis of
performance status $ 60, $ 4 months since administration of any prior bevacizumab, experimental therapies. Recruitment began in October 2023, with 12 patients of the
and dexamethasone dose # 4 mg at the time of randomization. The primary endpoint is target 50 enrolled thus far. Data from the first cohort will inform platform optimization
9-month OS. Secondary endpoints include OS, progression-free survival, objective re- and scalability. All biological samples are anonymized, with outcomes tracked per RANO
sponse rate, neurologic assessment by NANO criteria, and adverse events profile. guidelines. Even at this early stage, PPIE has helped improve trial design. We are
Protocol treatment will continue up to two years, or until progression or intolerable concurrently validating the platform in other high-unmet-need indications including
toxicity. Survival follow up will continue every two months, up to four years. Seven of the early-stage breast cancer (NCT05435352), metastatic breast cancer (NCT06182306)
planned 94 patients have been enrolled as of submission on 1/28/25. Clinical trial #: and metastatic kidney cancer (NCT06264479) hoping to shift the paradigm in precision
NCT06160206. Funding provided by Incyte. Clinical trial information: NCT06160206. treatment selection. Clinical trial information: NCT06038760. Research Sponsor: Our-
Research Sponsor: Incyte Corporation. otech (t/a Pear Bio).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 135s

TPS2094 Poster Session TPS2095 Poster Session

A multicenter, pivotal trial of microbubble-enhanced transcranial focused Liposomal curcumin and standard radiation and temozolomide for newly
ultrasound (MB-FUS) for plasma-based liquid biopsy in patients with glio- diagnosed high-grade gliomas: A phase 1/2 study. First Author: Matthias
blastoma (LIBERATE). First Author: Manmeet Singh Ahluwalia, Miami Cancer In- Holdhoff, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore,
stitute, Baptist Health South Florida, Miami, FL MD
Background: Liquid biopsy in glioblastoma (GBM) is hindered by a lack of requisite Background: Curcumin, derived from turmeric (Curcuma spp.), exhibits anti-
circulating tumor (ct) and cell-free (cf) DNA levels in blood due to the blood-brain barrier inflammatory and antitumoral activity in preclinical studies, including inducing cell
(BBB). This limits the identification of blood-based tumor biomarkers along with the cycle arrest, apoptosis, autophagy and disrupting key cancer signaling pathways (e.g.,
development and use of biomarker-driven systemic therapies. Low intensity focused STAT-3, AKT, VEGF, NF-kB, and IDO). Despite its promise, oral curcumin has limited
ultrasound combined with intravenously administered microbubble oscillators (MB-FUS), bioavailability. Liposomal curcumin (LC), a novel intravenous formulation, achieves
leads to non-invasive BBB opening. This trial aims to evaluate the utility of LIFU for plasma curcumin levels over 1000 times higher than oral administration and prefer-
bolstering blood ctDNA and cfDNA for enhance liquid biopsy in patients with GBM. entially accumulates in tumor cells. In preclinical glioma models, LC has antitumoral
Methods: LIBERATE is an ongoing, prospective, multi-center, self-controlled, pivotal trial efficacy, particularly when combined with cytotoxic therapies. Previous trials in healthy
evaluating safety and technical efficacy of transcranial MR-guided MB-FUS for increasing volunteers and cancer patients demonstrated LC’s safety, pharmacokinetics, and
blood ctDNA and cfDNA levels in adults, aged 18-80 years with GBM. Patients with manageable adverse effects, with doses up to 300 mg/m² being well-tolerated.
suspected GBM planned for tumor biopsy or resection at 17 centers in US and Canada are However, a case of hemolytic anemia was observed in a prior study at this dose
being enrolled. Patients with multifocal tumors or tumors arising from deep midline, in a patient who was also taking several known hemolytic drugs, suggesting the need for
thalamus, cerebellum, or brainstem are excluded. Patients are administered IV micro- further safety evaluation at this and potentially higher doses. Methods: This Phase I/II
bubbles for enhanced sonication, after which MR-guided BBB opening using a 220 kHz open-label, study evaluates LC combined with standard radiation (RT) and concomitant
device, with 1024-element phased array transducer, is performed with real-time acoustic and adjuvant temozolomide (TMZ) in newly diagnosed HGG patients (NCT05768919).
feedback control for effective cavitation. Pre- and post-procedure, phlebotomy and MRI The primary endpoints are MTD, RP2D and safety. Secondary endpoints include
brain are done. Patients are offered optional 2nd procedure during adjuvant chemotherapy treatment feasibility ($80% adherence to LC, RT, and $60% to TMZ), and exploratory
phase if willing. Primary efficacy endpoint is correlation between biomarker patterns in efficacy measures (PFS, OS by RANO criteria). The study has two phases: (1) dose-
tumor tissue collected during surgery/biopsy and blood collected following MB-FUS escalation using the TITE-BOIN method to determine MTD, and (2) dose-extension to
procedure. Confirmatory secondary efficacy endpoint is ratio between greatest yield of evaluate RP2D safety and feasibility. Up to 50 patients will be screened to enroll 30. LC is
cfDNA in blood post-MB-FUS compared to cfDNA level in blood pre-MB-FUS. The primary given weekly at 4 dose levels (240, 300, 350, and 400 mg/m²) alongside standard
study hypothesis is that agreement rate on biomarker pattern between resected/biopsied adjuvant TMZ (150–200 mg/m² x 5 days every 28 days) and RT. Treatment continues for
tumor tissue and blood is . 70%. The secondary hypothesis is that MB-FUS BBBO leads up to 6 TMZ cycles, with LC monotherapy possible afterward until progression or toxicity.
to a $2-fold rise in blood cfDNA. Assuming the true agreement rate expected is 89%, a MRI is done before and 4 weeks post-chemoradiation, then every 2 cycles of TMZ, as per
sample of N = 50 patients will provide 90% power to meet the primary endpoint (Exact test, standard of care. DLTs are evaluated over 10 weeks to determine the MTD which will be
Binomial Proportion, one-sided Alpha = 0.025). Exploratory endpoints include (1) sen- determined by TITE-BOIN dose escalation rule and Safety Review Committee’s guidance.
sitivity of detection of known specific somatic mutations in ctDNA from blood samples A separate exploratory protocol is offered to patients interested in additional imaging,
collected before and after MB-FUS, (2) estimation of ctDNA levels in samples collected at which uses chemical exchange saturation transfer (CEST) MRI to visualize liposome
30-minutes, 1-hour, 2-hour, and 3-hour post-MB-FUS to determine time of greatest yield, accumulation in tumor tissue non-invasively. As of 1/24/2025, 14 patients have been
(3) correlation of MRI parameters related to grading of BBB opening and ctDNA-based enrolled in the dose-escalation part of this study. Clinical trial information:
biomarkers from post-MB-FUS blood samples, (4) biomarker correlation between plasma NCT05768919. Research Sponsor: SignPath Pharma.
cfDNA sampled during adjuvant chemotherapy phase and tumor tissue harvested at
surgery. Patient enrollment commenced in 2022 and is ongoing (NCT05383872). Clinical
trial information: NCT05383872. Research Sponsor: Insightec Inc.

TPS2096 Poster Session TPS2097 Poster Session

A global phase 3, open-label, randomized 2-arm study comparing the clin- Personalized targeted glioblastoma therapies by ex vivo drug screening:
ical efficacy and safety of niraparib with temozolomide in adult participants Advanced brain tumor therapy clinical trial (ATTRACT). First Author:
with newly-diagnosed, MGMT unmethylated glioblastoma. First Author: Nader Anna Sophie Berghoff, Division of Oncology, Department of Medicine I, Medical Uni-
Sanai, Ivy Brain Tumor Center at Barrow Neurological Institute, Phoenix, AZ versity of Vienna, Vienna, Austria
Background: Glioblastoma (GBM) is associated with dismal prognosis and poor quality Background: Targeted therapies used in a personalized treatment concept have rev-
of life. In approximately 60% of tumors, the O6-methylguanine methyltransferase olutionized the management of several solid cancers. So far, various clinical trials
(MGMT) promoter is unmethylated and the prognosis is even more dire, with a median aiming to introduce the concept of personalized targeted therapies in glioblastoma have
overall survival (OS) of 12.7 months following surgical resection, temozolomide (TMZ), failed, as no clinically meaningful responses were observed. Importantly, most clinical
and fractionated radiotherapy (RT). Poly (ADP-ribose) polymerase (PARP) mediates trials investigating molecular targeted therapies included all-comers and concentrated
DNA damage response in GBM and niraparib is an investigational PARP1/2-selective on genetic biomarkers to predict treatment response. Given the biological complexity of
inhibitor. At ASCO 2024, we reported on a Phase 0/2 study of niraparib plus radiotherapy glioblastoma, genetic biomarkers might give only an insufficient insight into the re-
in newly-diagnosed, MGMT-unmethylated glioblastoma (GBM), demonstrating superior sponse of a given patient, and more personalized approaches are warranted. As novel
tumor pharmacokinetic and pharmacodynamic performance compared to other studied approaches to guide personalized treatment in glioblastoma are urgently needed, we
PARP inhibitors and a median overall survival (OS) of 21.7 months. Based on the proof- designed a prospective clinical trial to investigate the novel approach of cultivated
of-concept data, a global registrational Phase 3 study (Gliofocus) was initiated. patient-derived tumor cells (PDCs) for ex vivo drug screening. Methods: In this ran-
Methods: This Phase 3, open-label, randomized 2-arm study (NCT06388733) will domized phase 2 study, we are testing the ability of PDC-based ex vivo drug screening to
compare niraparib versus TMZ in 450 adult participants with newly-diagnosed, MGMT- formulate a personalized recommendation for maintenance treatment in patients with
unmethylated GBM. Participants must have a biopsied or resected GBM, per 2021 World newly diagnosed glioblastoma with unmethylated MGMT promoter after neurosurgical
Health Organization classification. MGMT promoter methylation status is determined resection followed by combined radio-chemotherapy. Based on overall survival as the
locally by validated pyrosequencing or quantitative methylation-specific polymerase primary endpoint, we plan to include 240 patients (120 per group) to show with a power
chain reaction assays. Other key inclusion/exclusion criteria include: (1) Karnofsky of 80% that we can increase the median survival from 12 to 17 months (hazard ratio 0.7).
performance status of $70, (2) no prior treatment for GBM (including brachytherapy or Patients are randomized 1:1 to either the standard group (no drug screening) or the
BCNU wafers), (3) no tumor-treating field therapy, and (4) suitability for RT of 60 Gy in 30 intervention group (drug screening and personalized recommendation for maintenance
fractions using ESTRO-EANO ‘single phase’ targeting approach. Following 1:1 ran- treatment). In the intervention group, automated drug screening is performed on PDCs
domization, niraparib (Arm A) or TMZ (Arm B) is administered concomitantly with RT and with 28 drugs used for treatment of solid tumors and hematological malignancies. Based
then adjuvantly until disease progression by Blinded Independent Central Review (BICR) on the cytotoxic/cytostatic activity of these drugs, as quantified by relative viability
or until completion of 6 cycles of TMZ. . The primary endpoints of the study are based on adenosine triphosphate levels, a molecular tumor board recommends a
progression-free survival (PFS) (per RANO 2.0; HR = 0.612, 90% power, 1-sided alpha = personalized treatment regimen. The first patient was enrolled in July 2024. Interim
0.001) and overall survival (OS) (HR = 0.698, 90% power, 1-sided alpha = 0.0239). analysis of the ATTRACT study (NCT06512311) is expected in late 2027, and final results
Secondary endpoints include overall response rate, health-related quality of life, neu- in 2030. Moreover, the clinical trial is accompanied by a comprehensive translational
rocognitive function, and the safety and tolerability of niraparib compared to TMZ. The research program to gain insights into the biological underpinnings of treatment re-
first patient was accrued in June 2024 and an interim futility analysis is planned in 2025. sponse in glioblastoma. Clinical trial information: NCT06512311. Research Sponsor:
This study, sponsored by the Ivy Brain Tumor Center and with drug and funding provided Ludwig Boltzmann Society.
by GSK, is expected to enroll in a minimum of 115 clinical sites across 11 countries.
Clinical trial information: NCT06388733. Research Sponsor: GSK.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
136s CENTRAL NERVOUS SYSTEM TUMORS

TPS2098 Poster Session TPS2099 Poster Session

Trial in progress: Feasibility of CSF and plasma ctDNA in BRAF-altered Neuro-oncology anywhere 242: Pilot study evaluating telehealth and in-
glioma during treatment with plixorafenib. First Author: Karisa C. Schreck, Sidney person assessments in patients with glioma receiving oral chemotherapy—
Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore, MD Clinical trial in progress. First Author: Ugur Sener, Division of Neurology, Mayo Clinic,
Background: Gliomas with BRAF alterations are often difficult to treat in the recurrent Rochester, MN
setting due to emergent resistance to FDA-approved targeted therapies. Additionally, it Background: Gliomas are the most common primary central nervous system (CNS)
can be difficult to assess response to treatment given the limitations of radiographic malignancy in adults, accounting for 26.3% of all brain tumors. Care at high volume
techniques and the infeasibility of serial tissue sampling. This protocol serves as a centers is associated with an overall survival benefit, but access to in-person evaluations
prototype for determining the feasibility of using CSF and plasma circulating tumor DNA can be challenging due to disease-related neurological disability and loss of income.
(ctDNA) as biomarkers for response to a novel-BRAF inhibitor, plixorafenib. Plixorafenib Telehealth represents a convenient and efficient alternative to in-person evaluations, but
is a small-molecule selective inhibitor of BRAF-V600E and BRAF-fusion alterations that acceptability and comparative safety of this care delivery modality has not been pro-
does not induce paradoxical reactivation of MAPK signaling. Methods: This study is a spectively evaluated among glioma patients undergoing chemotherapy. Methods: This
single institution trial of plixorafenib in patients (18+ years of age) with BRAF-V600E single-arm non-randomized pragmatic clinical trial evaluates patient satisfaction with,
mutant glioma following progression on prior BRAF-targeted therapy who are rec- and safety of video-enabled telehealth assessments compared to in-person evaluations
ommended for a clinically-indicated diagnostic or debulking surgery. Eligible patients for patients with glioma undergoing temozolomide chemotherapy. The study includes
have recurrent BRAF-V600E mutant glioma (any grade) with measurable disease (by adult patients with a diagnosis of glioma requiring adjuvant temozolomide chemotherapy.
RANO 2.0), have a Karnofsky performance status . 70, and are able to undergo surgery. Participants act as their own controls, alternating between in-person and telehealth
Leptomeningeal disease is allowed. A total of 12 evaluable patients will be enrolled. assessments while undergoing chemotherapy dosed per standard of care. Monitoring
Enrolled patients undergo clinically-indicated resection or biopsy for confirmation of labs are completed locally and transmitted electronically. For participants without access
disease progression and characterization of putative resistance alterations. All patients to Wi-Fi or a device (e.g. mobile phone or computer), cellular-enabled tablet devices are
have a ventricular reservoir placed at time of surgery with CSF and plasma sampling. provided to facilitate appointments and completion of electronic study components. All
Patients will initiate oral plixorafenib 900mg daily with cobicistat, a CYP3A inhibitor and participants who travel to in-person appointments are reimbursed for travel expenses.
PK enhancer, when clinically recovered from surgery. Patients will take the drug The primary outcome measure is patient satisfaction with care delivered, as measured by
continuously under fasting conditions. MRI, CSF, and plasma assessments will occur institutional Press-Ganey survey scores obtained following telehealth and in-person
approximately every two months to evaluate disease status. The primary endpoint is assessments. A key secondary outcome measure is completion rate of planned oral
proportion of samples with detectable tumor ctDNA baseline and after one month of chemotherapy, tracked using a digital pill diary incorporated into our institutional
treatment with plixorafenib. Secondary endpoints include the correlation of ctDNA with electronic health record. The digital diary allows real-time tracking of chemotherapy
disease status over time and response rate to plixorafenib. The trial is IRB approved and adherence and adverse events experienced by participants. Other secondary outcomes
currently open to enrollment. Clinical trial identifier NCT06610682. Clinical trial infor- include acute care utilization days following telehealth and in-person visits (defined as
mation: NCT06610682. Research Sponsor: Ivy Brain Tumor Foundation; Fore emergency department evaluations and days of inpatient stay), neurologic disability as
Biotherapeutics. measured by the Neurologic Assessment in Neuro-Oncology (NANO) scale, and disease
related quality of life measured by the EORTC QLQ-C30. All participant surveys are self-
reported and completed electronically. This decentralized pragmatic clinical trial provides
unprecedented, prospective real-world data on utilization of telehealth services compared
to in-person visits for patients undergoing chemotherapy for glioma. We expect the data
generated to inform the design and conduct of future decentralized interventional neuro-
oncology trials. NCT06625047 opened for enrollment in October 2024, 16 of 30 intended
participants were accrued as of January 2025. Clinical trial information: NCT06625047.
Research Sponsor: Mayo Clinic.

TPS2100 Poster Session TPS2101 Poster Session

Update on GBM AGILE: A global, phase 2/3 adaptive platform trial to Dual targeting of VEGF and PD-1: A phase I/II trial of ivonescimab, a novel
evaluate multiple regimens in newly diagnosed and recurrent bispecific antibody, in recurrent glioblastoma. First Author: Anuj Dilip Patel,
glioblastoma. First Author: Emma Maria Viktoria Hyddmark, Global Coalition for University of Texas MD Anderson Cancer Center, Houston, TX
Adaptive Research, Larkspur, CA Background: Patients with recurrent glioblastoma have limited effective treatment op-
Background: GBM AGILE (Glioblastoma Adaptive, Global, Innovative Learning Envi- tions due to the highly immunosuppressive microenvironment and rapid proliferation
ronment) is a biomarker based, multi-arm, international, seamless Phase 2/3 response fueled by neoangiogenesis. Anti-angiogenic therapy, including targeting vascular endo-
adaptive randomization platform trial designed to efficiently identify investigational thelial growth factor (VEGF) with bevacizumab, and immune checkpoint inhibition with
therapies that improve overall survival and confirm efficacious therapies and associated programmed cell death protein 1 (PD-1) inhibitors, have independently had limited efficacy
biomarker signatures to support drug approvals and registration. GBM AGILE is a in these tumors. Ivonescimab is a humanized tetravalent bispecific antibody against PD-1
collaboration between academic investigators, patient organizations, and industry to and VEGF, which has demonstrated cooperative binding in vitro leading to increased
support new drug applications for newly diagnosed and recurrent glioblastoma. binding of PD-1 in the presence of VEGF and vice-versa1. Ivonescimab has shown activity
Methods: The primary objective of GBM AGILE is to identify therapies that improve in multiple phase 3 trials conducted in China in non-small cell lung cancer, including one
overall survival in patients with newly diagnosed or recurrent glioblastoma. Operating trial which demonstrated activity in patients with brain metastases, but has not yet been
under a Master Protocol, GBM AGILE allows multiple drugs from different evaluated in patients with primary brain tumors. This trial evaluates ivonescimab in
pharmaceutical/biotech companies to be evaluated simultaneously and/or over time patients with recurrent glioblastoma. Methods: This investigator-initiated study consists
against a common control. New investigational therapies are added as new information of a phase I and II component; the primary objectives are safety and tolerability for phase I
about promising drugs is identified, while other therapies are removed as they complete and determining progression-free survival for phase II. The phase I component evaluates 3
evaluation. Bayesian response adaptive randomization is used within subtypes of the dose levels of ivonescimab (7.5, 10, and 20 mg/kg every 3 weeks), employing a Bayesian
disease to assign participants to investigational arms based on their performance. GBM optimal interval (BOIN) design for assessing toxicity. Once the recommended phase II dose
AGILE has screened over 2300 patients and enrollment continues to be robust. An is determined, the phase II portion will follow a Bayesian optimal phase II (BOP2) design,
estimated 25% of all US glioblastoma patients enrolled in clinical trials participate in with interim analyses at pre-specified enrollment points allowing for monitoring of efficacy
as well as ongoing evaluation of toxicity. The maximum accumulative sample size at the
GBM AGILE. The trial is open at select sites in the United States, Canada, Switzerland,
target dose will be 30 patients. Radiographic assessment will utilize the Response As-
France, Germany, and Australia. In addition to the efficient evaluation of investigational
sessment in Neuro-Oncology 2.0 criteria. Key eligibility criteria include adults with re-
arms, a primary goal of GBM AGILE is to expand knowledge of glioblastoma to support
current glioblastoma, IDH-wildtype (by WHO CNS 2021 classification) at first or second
advancements in treatment using the data collected within the trial (learning envi-
recurrence with Karnofsky Performance Scale $60 and normal blood counts and organ
ronment). Over 7 million data points are currently available for inclusion in the de- function. Prior therapy with anti-angiogenic agents (including bevacizumab) or check-
velopment of a longitudinal model. Such a model may be able to inform randomization point inhibitors is excluded, as well as concurrent corticosteroids $ 2 mg/day dexa-
by providing earlier and continuous information regarding patient and arm performance. methasone or equivalent. Samples of archival tumor, blood and stool microbiome will be
In addition, serial magnetic resonance imaging scans and biospecimens from baseline collected for correlative studies as an exploratory evaluation of predictive biomarkers of
through patient progression are being collected for further analysis. An initial 500 response or resistance to ivonescimab. The study has been approved by the institutional
baseline tissue samples are being characterized using whole genome sequencing and review board and accrual to phase I will commence in the first quarter of 2025. 1. Zhong T,
whole transcriptome analysis. Clinical trial information: NCT03970447. Research Huang Z, Pang X, et al. 1194 Mechanism of action of ivonescimab (AK112/SMT112): a first-
Sponsor: None. in-class tetravalent Fc-silent bispecific antibody with dual blockade of PD-1 and VEGF that
promotes cooperative biological effects. Journal for ImmunoTherapy of Cancer 2023;11:
doi: 10.1136/jitc-2023-SITC2023.1194. Clinical trial information: NCT06672575. Research
Sponsor: Summit Therapeutics.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 CENTRAL NERVOUS SYSTEM TUMORS 137s

TPS2102 Poster Session

Regorafenib versus local standard of care in patients with grade 2-3 me-
ningioma no longer eligible for loco-regional treatments: The MIRAGE trial.
First Author: Alberto Bosio, Veneto Institute of Oncology IOV, IRCCS, Padua, Italy
Background: Meningiomas are the most common intracranial tumors. Standard
treatment involves surgical resection with curative intent. When gross total resection is
not achievable, or in case of recurrence, RT is frequently utilized. On the other hand, the
role of systemic treatments remains poorly supported by [Link] is an oral
multi-tyrosine kinase (RTK) inhibitor. It exhibits high selectivity for VEGFR1/2/3, while
also inhibiting PDGFRb, FGFR1, and c-RAF/RAF1 and BRAF pathways, highly expressed
in high-grade meningiomas. Methods: TheMIRAGE Trial (NCT06275919) is a multi-
center, open-label, randomized phase 2 clinical trial evaluating grade 2/3 meningioma
pts who have progressed following surgery and RT. A total of 94 pts are being ran-
domized (1:1) to receive either Regorafenib (160 mg orally for 3 weeks on, 1 week off) or
local standard-of-care therapies (e.g., bevacizumab, hydroxyurea, somatostatin ana-
logues).Major inclusion criteria include histological confirmation of WHO 2021 grade 2-3
meningioma, radiologically documented progression at least 24 weeks from RT (esti-
mated planar growth . 25% in two dimensional tumor areas within the prior 12 months
or development of a new lesion) with at least 1 measurable lesion (minimum 10 x 10 mm)
on baseline MRI, ineligibility for further surgery and/or radiotherapy, absence of ex-
tracranial lesions and a WHO performance status of [Link] primary endpoint is 6-month
PFS (6m-PFS). Assuming a 6m-PFS of 20% in the control arm and 40% in the regorafenib
arm (corresponding to a HR = 0.57) with a = 5%, b = 85%, 104 patients are needed to
assess the targeted efficacy. Response to treatment will be assessed by using RANO
criteria. Secondary endpoints include OS, ORR, DCR, volumetric analysis of the target
lesions, safety and health-related quality of life. Multi-omics exploratory analysis will
also be performed to investigate possible prognostic and predictive biomarkers.
Radiomics analysis will also be performed. MIRAGE, initiated in September 2024, is an
academic trial promoted by the Istituto Oncologico Veneto, IOV-IRCCS and will recruit
patients across 15 neuro-oncology Centers in Italy with an estimated study duration of
18 months. Discussion: MIRAGE is the first randomized phase 2 trial analyzing the role
of a RTK inhibitor (regorafenib) in prolonging PFS in pts with grade 2-3 meningioma who
are ineligible for further surgery and/or radiotherapy. Clinical trial information:
NCT06275919. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
138s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2500 Oral Abstract Session 2501 Oral Abstract Session

Assessment of efficacy of LBL-024, a novel and uniquely designed bispe- First-in-human phase I/II trial evaluating BNT142, a first-in-class mRNA
cific antibody against PD-L1 and 4-1BB, combined with etoposide/ encoded, bispecific antibody targeting Claudin 6 (CLDN6) and CD3, in
platinum-based chemotherapy in treatment-naive advanced extrapulmo- patients (pts) with CLDN6-positive advanced solid tumors. First Author:
nary neuroendocrine carcinoma (EP-NEC): A multicenter phase Ib/II trial. Timothy A. Yap, The University of Texas MD Anderson Cancer Center, Houston, TX
First Author: Ming Lu, Peking University Cancer Hospital /Beijing GoBroad Hospital, Background: CLDN6 is an oncofetal cell surface protein silenced in normal adult tissues
Beijing, China but aberrantly activated in testicular, ovarian, non-small cell lung (NSCLC) and other
Background: The prognosis for patients with EP-NEC is very poor. A recognized 1L treatment for cancers. The investigational therapeutic BNT142 is a novel lipid nanoparticle (LNP)-
advanced disease is etoposide/platinum-based chemotherapy with no standard 2L/3L treatment. LBL- encapsulated mRNA encoding the anti-CLDN6/CD3 bispecific antibody RiboMab02.1. After
024 blocks the immunosuppressive pathway of tumor cells by targeting PD-L1 and effectively co- intravenous administration, BNT142 RNA-LNPs are taken up by liver cells and are
stimulates T cells by targeting 4-1BB, to improve the anti-tumor immune response. Here we report the translated into RiboMab02.1. The first results of the dose escalation part of the BNT142-01
safety and efficacy of LBL-024 combined with etoposide and cisplatin or carboplatin (EP/EC) as first
trial testing 7 dose levels (DL) are presented here. Methods: BNT142-01 (NCT05262530)
line treatment in patients with advanced NEC. (NCT06157827). Methods: This is a phase Ib dose
escalation and phase II dose optimization/expansion clinical trial. Phase Ⅰb enrolled previously is a Phase I/II, open-label, multi-center trial to evaluate weekly BNT142 treatment with
untreated advanced EP-NEC and SCLC patients, phase Ⅱ enrolled previously untreated advanced EP- premedication (antipyretics, antihistamines, fluids) at the investigators’ discretion in pts
NEC patients. Three dose levels of LBL-024 (6, 10 and 15 mg/kg, i.v. Q3W) plus EP/EC in phase Ib were with CLDN6+ ($10% of cells with at least weak membrane positivity) advanced solid
evaluated, 2 dose levels (6 and 15 mg/kg, i.v. Q3W) of LBL-024 plus EP/EC were evaluated in a tumors. Primary objectives include safety, tolerability and identifying the recommended
randomized dose optimization. The primary endpoints were tolerability, safety, efficacy (RECIST 1.1) Phase 2 dose (RP2D), secondary and exploratory objectives include pharmacokinetics,
and RP2D, the secondary endpoints were PK, PD and ADA. Results: As of December 26, 2024, a total pharmacodynamics and preliminary efficacy (RECIST 1.1). Results: As of 02 Dec 2024, 65
of 53 patients were enrolled, with 13 patients in Phase Ib and 40 patients in dose optimization stage of pts (median age 57 years [range 18 – 79]; 75% female; 60% ECOG 1; 44 ovarian, 10
Phase II. Phase Ib included 2 patients with SCLC, 1 with MiNEN, and 10 with EP-NEC. All 40 patients in testicular, 5 NSCLC, 6 rare cancers) received $1 dose (median 7, range 1 – 38) of BNT142.
Phase II were EP-NEC. During the Dose escalation stage, no DLTs were observed. During the Dose
Of 65 pts, 46 (71%) had $4 prior lines of systemic therapy. Mostly mild to moderate
optimization stage, 15 mg/kg of LBL-024 was selected as RP2D based on PK/PD, efficacy, safety and
ER analysis. Out of 49 patients, the ORR across all dose levels is 77.6% and the DCR is 93.9% among
treatment-related adverse events (TRAEs) occurred in 41 (63%) pts, including 15 (23%) pts
which 9 patients were unconfirmed. The ORR in 21 EP-NEC patients at RP2D dose is 81.0% and DCR is with $G3 TRAEs. Most common ($10%) TRAEs were cytokine release syndrome (CRS) in
95.2% among which 3 patients were unconfirmed. Additionally, 2 patients with SCLC achieved 100% 14 (22%) pts (1 pt [2%] $G3), aspartate or alanine aminotransferase (AST, ALT) increased
ORR. LBL-024 TRAEs of all-grade occurred in 53 patients (100%), with grade $3 TRAEs in 17/53 in 12 (19%) pts (8 pts [12%] $G3), and pyrexia, chills or fatigue in 8 (12%) pts (0/0/2 pts
patients (32.1%). Conclusions: LBL-024 combined with chemotherapy was well-tolerated. The ex- [0%/0%/3%] $G3, respectively). TRAEs leading to dose reduction, treatment interruption
tremely improved response observed in EP-NECs is significantly higher than the historic reports (about or discontinuation occurred in 1 (2%), 12 (19%) or 2 pts (3%), respectively (mostly G3; most
30%~55%). Data including ER analysis of this ongoing study will be updated by a follow-up submission common related terms AST or ALT increased and infusion related reaction). Two (3%) pts
to ASCO. Clinical trial information: NCT06157827. Research Sponsor: None. had a dose limiting toxicity (G4 ALT increased [DL5], leading to dose reduction, and G5 CRS
Clinical benefits of first line treatment in evaluable patients bin phase Ib/II. [DL6]). BNT142 led to transient, dose-dependent increases in inflammatory cytokines.
15 mg/kg Translated RiboMab02.1 was detected in serum in a dose-dependent manner, peaking 24 –
Phase Ib Phase II
(Dose escalation) (Dose optimization) (N=21) Total 72 h post-dose. Across all DLs, the disease control rate (DCR) was 58% with a tendency of
EP-NECs (N=49) higher efficacy in the higher DLs. In ovarian cancer, there were 7 RECIST 1.1 partial
6 mg/kg 10 mg/kg 15 mg/kg 6 mg/kg 15 mg/kg
(N=3) (N=4) (N=6*) (N=18) (N=18) responses (PRs) and the DCR was 75%. Conclusions: BNT142 demonstrated a man-
ageable safety profile and promising anti-tumor activity at the higher DLs, with 7 RECIST
ORR, 2 (66.7%) 3 (75.0%) 4 (66.7%) 14 (77.8%) 15 (83.3%) 17 (81.0%) 38 (77.6%)
N (%) 1.1 PRs in CLDN6+ ovarian cancer, a tumor usually refractory to immunotherapy. We
DCR, 3 (100.0%) 4 (100.0%) 4 (66.7%) 17 (94.4%) 18 (100.0%) 20 (95.2%) 46 (93.9%) provide the first clinical proof-of-concept for an mRNA encoded bispecific antibody. Dose
N (%) optimization is ongoing. Clinical trial information: NCT05262530. Research Sponsor:
*2 patients with SCLC, 1 with MiNEN and 3 with EP-NEC. BioNTech SE.

2502 Oral Abstract Session 2503 Oral Abstract Session

a-bias
Efficacy and safety results of a first-in-class PD-1/IL-2 bispecific A therapeutic vaccine for fibrolamellar hepatocellular carcinoma. First Author:
antibody fusion protein IBI363 in patients (pts) with immunotherapy- Marina Baretti, Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, Baltimore,
treated, advanced acral and mucosal melanoma. First Author: Bin Lian, MD
Peking University Cancer Hospital & Institute, Beijing, China Background: Fibrolamellar hepatocellular carcinoma (FLC) is a rare form of liver cancer
Background: Despite great success of immunotherapy (IO) in advanced melanoma, there affecting children and young adults that is driven by a chimeric protein, DNAJ-PKAc. The
remains an unmet clinical need for resistant tumors. Pts with acral and mucosal mela- development of molecular inhibitors of DNAJ-PKAc has been hampered by unacceptable
nomas show limited benefit from current therapies. IBI363, a first-in-class PD-1/IL-2a-bias on-target toxicity, but the chimera results in a tumor-specific antigen (neoantigen) that
bispecific antibody fusion protein that blocks PD-1 and activates a-bias IL-2 to rejuvenate may be targeted immunologically. Methods: We conducted a phase 1 clinical trial of a
exhausted tumor-specific T cells, has shown encouraging efficacy in pts with advanced therapeutic vaccine targeting DNAJ-PKAc (FLC-Vac), in combination with nivolumab and
melanoma. Here, we present results of IBI363 from a phase 1 study (NCT05460767) and a ipilimumab, in children and adults with advanced FLC. The primary objectives were safety
phase 2 study (NCT06081920) of pts with IO-treated, advanced acral and mucosal mel- and T cell responses, defined as 2.5-fold increase of interferon gamma (IFN-g)-producing
anoma. Methods: Eligible pts with IO-treated advanced acral and mucosal melanoma were DNAJB1-PRKACA chimera-specific T cells in the peripheral blood after week 10 (priming
enrolled. IBI363 was administered intravenously at 0.1 mg/kg every week, 0.3/0.6/1 mg/kg phase). The study was planned with 12 evaluable patients. FLC-Vac, consisting of a
every 2 weeks (Q2W), or 1/1.5/2/3 mg/kg every 3 weeks (Q3W). Primary endpoints for the peptide encoding the DNAJB1-PRACA fusion plus poly-ICLC adjuvant, was administered
phase 1 study were dose-limiting toxicity (DLT) and safety, and for the phase 2 study were on weeks 0, 1, 2, 3, 6, 9 during the priming phase of the study. Nivolumab, 3 mg/kg,
safety and investigator-assessed objective response rate (ORR), disease control rate (DCR), followed by ipilimumab, 1 mg/kg, was administered every 3 weeks for 4 doses during the
duration of response (DoR) and progression-free survival (PFS) according to RECIST v1.1. priming phase. After completion of the priming phase, FLC-Vac and nivolumab were
Results: As of December 6, 2024, 91 pts were enrolled across the phase 1 (n = 76) and continued in maintenance. Key exclusion criteria include age , 12 years and prior
phase 2 (n = 15) studies (male: 47%; median age: 57 years; Asian: 100%; ECOG PS 1: 66%; treatment with immune checkpoint inhibitors. The trial incorporated a safety lead-in
stage IV: 89%); 47 pts had acral melanoma and 44 had mucosal melanoma. Median follow- portion in which the first 3 patients received vaccine monotherapy for 3 weeks prior to
up time was 8.2 months. Median treatment duration was 13.4 weeks (range: 2.0-72.4). receiving combination therapy. Results: Among 16 patients enrolled, 12 completed the
Treatment-emergent adverse events (TEAEs) occurred in 90/91 (98.9%) pts including 27 vaccine priming phase and were evaluable for both immunological and clinical endpoints.
(29.7%) pts with grade $3 ($G3) TEAEs. TEAEs led to treatment discontinuation in 3 The median age was 24 years (range: 12-47). Grade 3 treatment-related adverse events
(3.3%) pts, and 1 (1.1%) pt had a TEAE leading to death which was considered to be were reported by six patients (37.5%). DNAJ-PKAc-specific T cell responses were de-
treatment-related (sepsis). Most common TEAEs were arthralgia (59.3%, with 4.4% $G3), tected in 9/12 patients after treatment. In the subset of patients who completed the initial
rash (42.9%, with 3.3% $G3), and anemia (42.9%, with 2.2% $G3). Among all pts with at priming phase the disease control rate (DCR) was 75% (9/12), with three partial responses
least one post-baseline tumor assessment (n = 87), 1 pt had a complete response, 22 had (25%). All 3 responding patients are without evidence of active cancer after undergoing
partial responses, 33 had stable disease, 31 had progressive disease. ORR was 26.4% (95% surgical debulking of residual disease. All patients with clinical responses also had DNAJ-
CI: 17.6-37.0) with 16 responses confirmed and 2 pts still waiting for confirmation; DCR PKAc-specific T cell responses, from whom we identified multiple class II-restricted T cell
was 64.4% (95%CI: 53.4-74.4). Among pts treated at 1mg/kg and above (n = 74), the ORR receptors (TCRs) with specificity for DNAJ-PKAc. Correlates of response included both
was 28.4% (95%CI: 18.5-40.1) and DCR was 68.9% (95%CI: 57.1-79.2). Patients treated at 1 functional neoantigen reactivity and changes in TCR repertoire features over time. In two
mg/kg Q2W (n = 30) had median DOR 14.0 months with a median follow-up of 9.1 months patients who experienced eventual progression after initial clinical response, we found
and 50.0% events; the median PFS was 5.7 (95% CI, 3.6-6.7) months with a median follow- evidence that the loss of efficacy was likely due to T cell exhaustion, and in one case was
up of 11.0 months and 73.3% events. Conclusions: IBI363 showed encouraging efficacy in restored with checkpoint rechallenge. Conclusions: Our findings demonstrate the po-
pts with IO-treated advanced acral and mucosal melanoma. The safety profile was ac- tential for therapeutic vaccines targeting DNAJ-PKAc in FLC and suggest a rubric for
ceptable and manageable. Further global clinical development of IBI363 in melanoma is
evaluating effective anti-neoantigen immunity. Clinical trial information: NCT04248569.
ongoing. Clinical trial information: NCT05460767 and NCT06081920. Research Sponsor:
Research Sponsor: ASCO CDA (Dr Yarchoan); Fibrolamellar Cancer Foundation; BMS; R01-
Innovent Biologics (Suzhou) Co., Ltd.
CA265009.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 139s

2504 Oral Abstract Session 2505 Oral Abstract Session

Clinical responses to SYNC-T therapy: In situ personalized cancer vaccina- Phase 1 study of B440, an oral Bifidobacterium-engineered WT1 cancer
tion with intratumoral immunotherapy in patients with metastatic vaccine, in patients with metastatic urothelial cancer. First Author: Toshiro
castration-resistant prostate cancer (mCRPC). First Author: Charles J. Link Shirakawa, Kobe University, Kobe, Japan
Jr., Lankenau Institute for Medical Research, Wynnewood, PA Background: B440 is an innovative oral cancer vaccine comprised of recombinant Bifi-
Background: Metastatic castration-resistant prostate cancer (mCRPC) has a poor dobacterium engineered to express WT1 tumor-associated antigen. By delivering the WT1
response to immunotherapy limited by both a low ORR and high frequency of severe protein to dendritic cells in gut-associated lymphoid tissue, B440 is designed to induce a
immune-related adverse events. SYNC-T is a novel in situ therapy that synchronizes the tumor-specific cellular immunity. Preclinical data demonstrated effective WT1-specific T-
presence of tumor antigens, an immune therapy drug, and immune cells in the tumor and cell induction and anti-tumor activity in murine models of urothelial, prostate, and renal
locoregional lymph nodes. SYNC-T Therapy combines device-induced partial cryolysis cancers. Methods: This open-label, single-arm, phase 1 study evaluated the safety and
of a targeted tumor to create a personalized multi-antigen vaccine, followed immediately preliminary efficacy of B440 in patients with metastatic urothelial cancer who had pro-
by intratumoral infusion of the multitarget novel drug candidate SV-102, leading to T-cell gressed after all standard therapies, including cytotoxic chemotherapy, PD-1/PD-L1 in-
activation and an effective systemic immune response. Methods: 15 subjects, 13 with hibitors, and antibody-drug conjugates. Twelve patients were enrolled in two dose cohorts
bone metastases, and documented failure to prior hormonal therapy (n = 10) or refused (800 mg or 1,600 mg, n = 6 each), administered once daily for five consecutive days per
therapy (n = 5) were recruited to a single-arm study (NCT05544227). Image-guided week over four weeks (20 total doses). The primary endpoint was dose-limiting toxicity
partial cryolysis of a tumor was followed by intratumoral infusion of SV-102, comprised (DLT), assessed during the treatment. Secondary endpoints included safety (adverse
of fixed low dose of anti-PD-1 mAb, anti-CTLA4 mAb, CD40 agonist mAb, and TLR9 events [AEs] graded by CTCAE v5.0), best overall response (BOR), and progression-free
agonist CpG-ODN. All subjects received the same dose of SV-102. Subjects received survival (PFS) by RECIST v1.1. WT1-specific immune responses were measured via
SYNC-T Therapy q4 weeks for up to 12 cycles (median = 6). One site of primary prostate ELISPOT assays detecting interferon-g-producing T cells. Results: All 12 patients
or soft tissue metastasis was targeted at each cycle. Primary objective was to evaluate completed the treatment: (median age: 74.5 years [range: 39–81]; primary tumors in
safety and tolerability with a secondary objective to assess tumor response by PCWG3 bladder [n = 5], renal pelvis [n = 4], ureter [n = 3]). No DLTs were observed in either dose
and RECIST 1.1. Results: 15 subjects were treated and evaluable. Median age was 61 cohort. Treatment-related AEs were generally mild (Grade 1), with the most common
events being transient IL-6 elevations and cold-like symptoms (n = 3 each). The disease
(48-74). Prior treatments included one or more of 1st, 2nd generation hormonal blockade,
control rate (DCR) was 50%, as six patients achieved stable disease (SD) as their BOR. Six
chemotherapy, immunotherapy, or radiation therapy. Within 15 evaluable subjects there
patients also demonstrated WT1-specific T-cell induction confirmed by ELISPOT. ELISPOT-
were 8 radiographic CRs (53%, the two-sided 95% CI is 29.4% to 78.7%, rejecting 20% CR
positive patients had a significantly longer PFS compared to ELISPOT-negative patients
null hypothesis; p = 0.0085) with complete resolution of primary, bone, and soft tissue
(median PFS: 113 days vs. 57 days; P = 0.0033). Although not included in the study
metastases and 5 PRs with an ORR of 87%. Median time to response was 3 months protocol, six patients subsequently underwent pembrolizumab rechallenge at the dis-
with a median duration of 12 months to date (range 1.2 -14.6). Among the 15 subjects, 3 cretion of their physicians. Of these, three achieved clinical responses (one complete
have died resulting in 80% survival with 14 months median follow-up. SYNC-T Therapy response [CR] and two partial responses [PR]). Spider plot analyses indicated early tumor
was well-tolerated with 41 TEAEs in 13 subjects. The majority (95%) of TEAEs were shrinkage among ELISPOT-positive patients, with maximum reductions of 2100%, 249%,
Grade 1 or 2, most commonly fever and hematuria. There were 2 Grade 2 irAEs of and 232.7% from baseline. Notably, three of the four ELISPOT-positive patients achieved
hepatitis and hypothyroidism and 2 Grade 3 TEAEs of urinary retention and spinal cord objective responses upon rechallenge. Conclusions: B440 exhibited a favorable safety
compression. PSA analysis during and post SYNC-T Therapy will be presented. PK profile and no DLTs up to 1,600 mg. The induction of WT1-specific immunity correlated
analysis revealed minimal systemic exposure to SV-102 components. PD analysis with improved PFS during B440 therapy and enhanced responses upon pembrolizumab
showed induction of inflammatory cytokines and the emergence of multiple, novel T-cell rechallenge. These data support further investigation of B440 in larger, randomized trials
clones. Conclusions: SYNC-T Therapy was well-tolerated achieving an 87% ORR in and potential combination with other immunotherapies in WT1-expressing malignancies.
subjects with mCRPC or who refused ADT. These encouraging clinical results have led to Clinical trial information: jRCT2051220143. Research Sponsor: Japan Agency for Medical
further study of SYNC-T SV-102 in a US, multicenter, Phase 2a trial for subjects with Research and Development (AMED); 23ym0126081h0002; Immunorock Co., Ltd.
mCRPC. Clinical trial information: NCT05544227. Research Sponsor: Syncromune, Inc.

2506 Oral Abstract Session 2507 Oral Abstract Session

Effect of erythrocyte-antibody conjugates on cancers resistant to check- RETRACTED: Phase 1 clinical trial of EpCAM CAR-T cell therapy in patients
point blockade immunotherapy: A phase I trial. First Author: Xiaoqian Nie, with gastrointestinal cancers. First Author: Tianhang Luo, Shanghai Changhai
Westlake University, Hangzhou, Zhejiang, China Hospital, Shanghai, China
Background: Despite the clinical success of immune checkpoint blockade therapy, the
majority of patients do not benefit due to inadequate efficacy as well as immune-related
adverse toxicities. We have previously developed WTX-212, an erythrocyte-antibody
conjugate that covalently links anti-PD-1 antibodies to erythrocyte membranes. Unlike
conventional antibodies, WTX-212 accumulates in the spleen, where it effectively remodels
splenic immune landscape by expanding effector T cells and reducing the reservoir of
immunosuppressive myeloid cells. These changes further reprogram the tumor micro-
environment and suppress tumor growth in syngeneic mouse models. Based on promising
D
TE
preclinical results, we have investigated WTX-212 in cancer patients resistant to
checkpoint blockade therapy (NCT06026605). Methods: This is an investigator-initiated
trial designed to assess the safety, tolerability, and preliminary efficacy of autologous
WTX-212 monotherapy in patients with advanced malignancies. The primary outcome
measures include safety and tolerability according to NCI-CTCAE v.5.0. Secondary out-
AC

come measures preliminary efficacy based on RECIST 1.1 criteria. As of January 15, 2025,
14 heavily treated patients with 11 types of solid tumors, who had received PD-1/PD-L1
antibody-containing regimens as their last line of treatment but developed resistance, were
enrolled. These patients received WTX-212 monotherapy in two dose cohorts (2 3 10¹¹ or
3 3 10¹¹ cells, with 6-10 mg of conjugated antibody). Results: Repeated WTX-212
treatment showed no DLTs or TRAEs $3. No patient discontinued treatment due to
R

AEs. WTX-212 monotherapy demonstrated promising anti-tumor activity, with a DCR of

78.6% (11/14) and an ORR of 42.9% (6/14), including 1 CR and 5 PR. In the higher-dose
cohort (3 3 10¹¹ cells), DCR and ORR increased to 85.7% (6/7) and 57.1% (4/7), re-
ET

spectively, suggesting dose-dependent efficacy. Additionally, responders (CR+PR)

exhibited higher baseline levels of circulating polymorphonuclear myeloid-derived sup-
pressor cells (PMN-MDSCs) compared to non-responders (PD+SD), indicating that patients
with elevated PMN-MDSCs may benefit more from treatment. Importantly, WTX-212
treatment rapidly reduced PMN-MDSCs in the peripheral blood of responders compared
R

with non-responders, consistent with preclinical data. These preliminary results suggest
that WTX-212 is safe, well-tolerated, and effective at low doses, supporting further in-
vestigation into WTX-212 monotherapy and combination therapies. Conclusions: Our
study suggests that PD-1 blockade in the spleen using erythrocyte-antibody conjugates
triggers systemic anti-tumor responses while maintaining a favorable safety profile.
Erythrocyte-drug conjugates represent a novel approach for targeting immune cells in the
spleen, with broad implications for cancer treatment and drug development. Clinical trial
information: NCT06026605. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
140s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2508 Oral Abstract Session 2509 Clinical Science Symposium

Phase 1 clinical update of IMA203, an autologous TCR-T targeting PRAME in Phase III randomized study comparing ultra-low dose immunotherapy to
patients with PD1 refractory metastatic melanoma. First Author: Martin standard cytotoxic chemotherapy for solid tumors in second line and beyond
Wermke, University Hospital Carl Gustav Carus, Dresden, Germany setting (DELII: Development of Low dose Immunotherapy in India). First
Background: Frequent recurrence and limited long-term survival in unresected or met- Author: Vanita Noronha, Tata Memorial Hospital, Mumbai, India
astatic melanoma after relapse from 1L checkpoint inhibitor treatment highlight the critical Background: Although immunotherapy (IO) is approved in the second line and beyond setting
need for new therapies that deliver deeper, more durable responses. ACTengine IMA203 is for most solid tumors, cost limits its accessibility. Lower doses of IO have been shown to
an autologous TCR-T targeting PRAME, an intracellular protein displayed as peptide achieve adequate target occupancy, persisting for 3 months post administration. We hy-
antigen at high density on the surface of multiple solid tumors, including melanoma. pothesized that nivolumab would retain efficacy at one-twelfth the approved dose.
Methods: Patients treated in this ongoing Ph1a/b trial (NCT03686124) are $18yo, HLA- Methods: Open-label, phase III randomized superiority study in 500 patients with solid
A*02:01+, PRAME+, have recurrent and/or refractory solid tumors with no additional tumors, whose disease had progressed on at least one line of systemic treatment, with
standard of care treatments available, measurable disease (RECIST1.1) and ECOG PS 0-1. performance status 0-1. Patients were randomized 1:1 to ultra-low-dose nivolumab (20 mg
Patients receive Cy/Flu (500 mg/m2 & 30 mg/m2 x4 d) lymphodepletion prior to infusion, intravenously every 2 weeks) or standard chemotherapy. Standard chemotherapy options for
followed by low-dose IL-2 for 10 days. Results: As of Aug 23, 2024: 70 heavily pretreated lung and head-and-neck cancers were docetaxel 75 mg/m2 every 3 weeks or paclitaxel
patients with solid tumors (median 3 prior systemic therapies) across all dose levels 175 mg/m2 every 3 weeks or paclitaxel 80 mg/m2 once-a-week; esophageal and urothelial
(median total infused dose 2.09x109 TCR-T cells (0.08-10.02x109)) were enrolled and cancers: paclitaxel 175 mg/m2 every 3 weeks or 80 mg/m2 once-a-week. Therapy continued
assessed for safety. Baseline tumor burden (median sum of diameter): 11.78 cm; LDH . 1 until progression or intolerable toxicity. Primary endpoint was overall survival (OS).
x ULN: 64% of patients. IMA203 had an overall favorable tolerability profile. Most common Results: Between Jun 2020 and Feb 2024, we enrolled 500 patients: 250 to each arm. Median
TEAEs: chemotherapy-related cytopenias (100%), mild to moderate CRS (G1-2: 83%, G3: age was 49.5 years (IQR, 42-58), 408 (81.6%) patients were male. Primary cancers included
11%), infrequent ICANS (G1: 6%, G2: 4%, G3: 4%), no G5 events. Objective responses were head and neck (259, 51.8%), lung (182, 36.4%), esophagogastric (31, 6.2%), urothelial (14,
observed in melanoma, ovarian cancer, synovial sarcoma, and other tumor types. Suc- 2.8%), and microsatellite instability-high colorectal cancers (14, 2.8%). Patients had
cessful trafficking of IMA203 cells to various organs was evidenced by their ability to received a median of one prior line of systemic therapy (range: 1-8); 144 (28.8%) patients had
shrink metastatic tumor lesions in the lung, liver, pleura, peritoneum, skin, lymph nodes, received at least two prior lines of systemic therapy. PD-L1 positivity (TPS or CPS . 0) was
noted in 66.2% patients. Radiologic response was 7.7% and 8.1% in IO and chemotherapy
adrenal gland, bladder, kidney, spleen, and muscle. Across patients treated in dose es-
arms, respectively; P = 0.882. Disease stabilization rate was 37.7% and 39.3% in IO and
calation and dose expansion, higher doses of IMA203 TCR-T cells were associated with a
chemotherapy arms, respectively; P = 0.761. Median PFS was similar between the two arms:
higher rate of confirmed responses (p = 0.018), whereas tolerability profile remained
2.04 months (95% CI, 2.0-2.1) in IO arm, and 2.09 months (95% CI, 2.04-2.17) in chemotherapy
favorable. Exposure data (Cmax, AUC) demonstrated a clear dose-dependent improvement
arm (HR, 1.03; 95% CI, 0.86-1.23; P = 0.77). Median OS was 5.88 months (95% CI, 4.99-7.13) in
in clinical efficacy: Patients with confirmed PR had a higher concentration of IMA203 TCR-
IO arm, versus 4.70 months (95% CI, 3.91-5.65) in chemotherapy arm; P = 0.022; HR, 0.80
T cells in the periphery, compared to patients with unconfirmed PR, SD, and PD. In heavily (95% CI, 0.66-0.97). One-yr OS in IO and chemotherapy arms was 27.28% (22.19-33.54) and
pretreated patients (median 2 prior systemic therapies) with melanoma at RP2D (1-10x109) 16.88% (12.75-22.34), respectively; 2-yr OS was 11.19% (7.59-16.50) and 6.55% (3.95-10.89),
in Ph1b, cORR was 54% (14/26), with tumor shrinkage in 88% (23/26) of patients. Median respectively. Grade 3 and higher treatment-related adverse events were significantly lower in
DOR was 12.1 months with 7/14 confirmed responses ongoing (longest . 2 years). Median IO arm (42%) than chemotherapy arm (60.3%); P , 0.001. Conclusions: Ultra-low dose
PFS was 6 months and median OS not reached at 8.6 months mFU. Updated data with immunotherapy dosed at one-twelfth the standard approved dose is efficacious and sig-
longer follow-up will be presented. Conclusions: IMA203 TCR-T was well tolerated and nificantly prolongs survival in patients with solid tumors in the second line and beyond setting,
showed durable objective responses in patients with advanced melanoma. Given its as compared to standard cytotoxic chemotherapy. Low dose IO should be tested in various
promising risk/benefit profile and high PRAME prevalence in melanoma, a registration- settings and multiple malignancies. This will substantially increase global accessibility to IO.
directed Phase 3 trial (SUPRAME; NCT06743126) is underway to further evaluate its Clinical trial information: CTRI/2020/02/023441. Research Sponsor: R G Manudhane Foundation
efficacy in patients with previously treated (2L) advanced cutaneous melanoma. Clinical for Excellence; Trilokchand Papriwal Trust; Tata Memorial Hospital, Mumbai, India.
trial information: NCT03686124. Research Sponsor: None.

2510 Clinical Science Symposium 2511 Clinical Science Symposium

Overall survival according to time-of-day of combined immuno- Safety and efficacy of immune checkpoint inhibitors in solid organ trans-
chemotherapy for advanced non-small cell lung cancer: A bicentric bicon- plant recipients: A systematic review and individual patient data meta-
tinental study. First Author: Francis Albert Lévi, UPR Chronotherapie, Cancers et analysis. First Author: Muntaser Al Zyoud, University of Jordan, Amman, Jordan
Transplantation, Université Paris Saclay, Hôpital Paul Brousse ID Isco 13918, Villejuif, Background: Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment but pose
France unique challenges in solid organ transplant (SOT) recipients. Transplant rejection remains the
Background: Circadian rhythms moderate immune cells trafficking and function over the 24 predominant safety concern. We systematically evaluated the safety (focusing on allograft re-
hours. This could account for the near doubling of overall survival (OS) in patients (pts) jection) and efficacy of ICIs across all organ transplant types and ICI classes, providing updated
receiving immune checkpoint inhibitors (ICIs) as single agents at early times-of-day of evidence-based insights for clinical decision-making. Methods: A systematic review of PubMed,
EMBASE, and SCOPUS databases was conducted in accordance with PRISMA guidelines. Studies
administration (ToDA) in retrospective studies. Yet, (i) the cut-off time that differentiates ICI
reporting rejection or efficacy outcomes in SOT recipients treated with any class of ICI were
efficacy according to ToDA ranges from 11:30 to 16:30, and (ii) the relevance of ICI timing for included. The primary endpoints were the incidence of transplant rejection and survival following
OS is unknown in pts receiving immunochemotherapy (ICI-chemo). Methods: These issues ICI therapy. Secondary endpoints included objective response rate (ORR) and progression-free
are addressed using OS as the primary endpoint in retrospectively-included pts receiving survival (PFS) for malignancies. Analysis was performed using SPSS (V26.0) and R (V4.3.0).
1st-line ICI-chemo for stage IIIc-IV non-small cell cancer (NSCLC) in France (Cohort 1) or in Results: Of 2682 screened abstracts, 198 studies involving 331 SOT recipients met inclusion
China (Cohort 2). The median ToDA of the initial four ICI-chemo infusions was computed for criteria. The transplanted organs were liver (n=175), kidney (n=136), and heart (n=15). Rejection
each patient. Hazard ratio (HR) functions of an earlier death or an earlier progression were rates were highest in Kidney at 46.3% (63/136), followed by heart (40.0%, 6/15) and liver (26.9%,
computed for each cohort and for the pooled one, using ToDA cut-off times ranging from 10: 47/175). Across ICI classes, rejection rates were: Anti-CTLA4 (25%) Anti-PD1 (40.6%) and Anti-
30 to 13:00, with 30-minutes increments. Median ToDA of ICI-chemo determined the al- PDL1 (0%). Rejection rates were lower in patients receiving ICI pre-transplant (25.9%) compared to
location of patients to “Before” or “After” treatment groups. The temporal relations between post-transplant (40.9%). ORR varied by ICI class: Anti-CTLA4 (25%), Anti-PD1 (41.8%), Anti-PD1 +
HRs and ToDA as a continuous variable were further determined, using Cox models in- CTLA4 (28%), and Anti-PDL1 (72.7%). Cutaneous squamous cell carcinoma (cSCC) showed the
corporating periodic restricted cubic splines. Patients were dichotomized according to the highest ORR (49.1%), followed by hepatocellular carcinoma (40.8%) and melanoma (25.3%). Post-
best cut off ToDA candidate, with OS and PFS being estimated using Kaplan-Meier and transplant rejection risk was lower with Anti-CTLA4 (OR 0.22), 3rd-line ICI therapy (OR 0.24), and
compared using log-rank. The association between ToDA and OS, PFS and response rates corticosteroids (OR 0.46). Pre-transplant rejection risk decreased with .60-day washout periods
were evaluated using the Cox and logistic models controlling for main patient charac- (OR 0.10). Multivariate analysis identified key factors influencing rejection risk (Table 1).
Conclusions: ICI therapy in SOT recipients is high-risk yet promising. Key strategies include
teristics. Results: A total of 713 pts started treatment between 01/2018 and 10/2023
prolonged washout periods, Anti-CTLA4 therapy, and late-line ICI use. Prospective studies are
(Cohort 1, 165 pts; Cohort 2, 548 pts; median age, 62 y.o., male sex, 84%; pembrolizumab as
needed to refine protocols and identify predictive markers to improve outcomes in this population.
ICI, 51%; pemetrexed-carboplatin/cisplatin, 49%; paclitaxel-carboplatin, 51%). HR functions Research Sponsor: None.
in each cohort and in the pooled one, and the fitted curve using ToDA as a continuous
variable identified 11:30 as a likely best cut off time. Median OS was 33.0 months (mo.) [95% Multivariate analysis results.
CI, 27.5 - 41.0] in the 345 patients, who received 2-4 immunochemotherapy courses before Pre-Transplant ICI Post-transplant ICI
Factor Total (OR) P-Value administration (OR) P-Value administration (OR) P-Value
11:30) compared to 19.5 mo. [18.0 - 22.5] in those, who received 2-4 courses after 11:30 (N =
368) (p,0.0001). In the multivariable analysis, a median ToDA before 11:30 was associated Age (<60 vs > 60) 0.79 0.42 0.85 0.82 0.62 0.18
Sex (M vs F) 1.12 0.75 1.36 0.7 1.23 0.63
with prolonged OS with an adjusted HR of 0.47 [0.37-0.60]. Statistically significant dif- CTLA-4 0.21 0.03 - - 0.22 0.04
ferences in ToDA effects were found for OS, PFS in each cohort, and for response rate in Third line and later 0.26 0.01 0.17 0.44 0.24 0.01
each cohort and in the pooled data. Conclusions: In this large bi-continental study, ToDA of (vs 1st line)
immunochemotherapy administration before 11:30 was associated with improved OS, PFS NSCLC (vs HCC) 6.87 0.03 - - 3.59 0.23
RCC (vs HCC) 9.9 0.04 - - 5.88 0.11
and response rates, compared to later ToDA in pts receiving standard first line immu- Multiple tumors 19.18 0.01 - - 9.3 0.06
nochemotherapy for NSCLC. Randomized trials are needed to confirm this important finding Washout period - - 0.1 ,0.001 - -
and inform recommendations for clinical practice. Research Sponsor: None. (<60 vs>=60)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 141s

LBA2512 Rapid Oral Abstract Session 2513 Rapid Oral Abstract Session
The phase II NIBIT-ML1 study of nivolumab plus ipilimumab and ASTX727 or Durable responses in ICI-refractory or acquired resistance: Phase 2 study of
nivolumab plus ipilimumab in PD-1 resistant metastatic melanoma: Tumor NP-G2-044 combined with anti-PD-1 therapy. First Author: Anup Kasi, University
methylation landscape and correlation with clinical outcomes. First Author: of Kansas Cancer Center, Fairway, KS
Anna Maria Di Giacomo, University of Siena, Center for Immuno-Oncology, University Background: Although immune checkpoint inhibitors (ICIs) have transformed cancer
Hospital of Siena, NIBIT Foundation Onlus, Siena, Italy treatment, many patients still develop primary or acquired resistance. NP-G2-044 is a
first-in-class, oral fascin inhibitor that disrupts cancer cell motility, invasion, and
metastasis while promoting intratumoral dendritic cell (DC) activation and CD8+ T-cell
proliferation. Preclinical studies indicate NP-G2-044 synergizes with anti-PD-1 therapy
to convert nonresponsive tumors into responsive ones. Early-phase clinical data support
the feasibility of NP-G2-044 at pharmacologically active doses and its potential to
prevent metastasis when used as monotherapy. Methods: In this open-label Phase 2
trial (NCT05023486), patients with advanced or metastatic solid tumors and docu-
mented primary or acquired resistance to anti-PD-(L)1 therapy received NP-G2-044 plus
standard-of-care anti-PD-1. Efficacy was assessed using RECIST, with the primary
endpoint being objective response rate (ORR). Secondary endpoints included
progression-free survival (PFS), duration of response, disease control rate (DCR), and
The full, final text of this abstract will be available at safety. Results: Forty-five patients were enrolled, with 33 evaluable for efficacy. No
dose-limiting toxicities were observed. Objective responses occurred in 7/33 patients
[Link] on the day of presentation and in the (21%) [95 % CI 9 - 38.9%], including 4 complete responses (CRs)—2 by RECIST in cervical
online supplement to the June 10, 2025, issue of the Journal and endometrial cancers, and 2 pathological CRs in pancreatic and gastroesophageal
of Clinical Oncology. junction adenocarcinomas—and 3 partial responses (cutaneous squamous cell carci-
noma, non-small cell lung cancer, and cholangiocarcinoma). Three patients have been
cancer-free for over 7 months, and 5 have remained on therapy for more than 15 months.
The DCR was 76%, and 55% of patients showed no new metastases during the study.
One-year PFS is projected at 30%. The most common adverse events were diarrhea,
fatigue, nausea, and transaminitis (~30%), which was transient, reversible, and preceded
tumor response. Mechanistic analyses using multiplex immunofluorescence and
immunophenotyping revealed enhanced intratumoral cytotoxic T-cell infiltration, pro-
liferation, and granzyme B expression, along with an increase in activated DCs—
consistent with a strong immunomodulatory effect. Conclusions: NP-G2-044, in
combination with anti-PD-1 therapy, appears to have clinical activity across multiple
cancer types, overcoming both primary and acquired ICI resistance while producing
durable responses. Ongoing expansion cohorts and biomarker analyses aim to refine
patient selection. These findings underscore NP-G2-044’s potential to address meta-
static disease and improve cancer immunotherapy outcomes, offering a promising
therapeutic option for patients with limited alternatives. Clinical trial information:
NCT05023486. Research Sponsor: Novita Pharmaceuticals.

2514 Rapid Oral Abstract Session 2515 Rapid Oral Abstract Session
Preliminary results from the dose-escalation stage of a phase I trial of an An open-label, phase I trial of the SIRPa monoclonal antibody, BI 770371,
anti-CCR8 antibody in patients with relapsed/refractory cutaneous T-cell alone and in combination with the PD-1 inhibitor ezabenlimab in patients
lymphoma (R/R CTCL). First Author: Zhiming Li, Sun Yat-sen University Cancer with advanced solid tumors. First Author: Judy S. Wang, Florida Cancer Specialists
Center, Guangzhou, China & Research Institute, Sarasota, FL
Background: Mycosis fungoides (MF) and Sézary syndrome (SS) are the most common cu- Background: The signal regulatory protein alpha (SIRPa)/CD47 axis is a critical regulator of
taneous T-cell lymphomas (CTCL). CCR8 is expressed in the skin resident memory T cells. ICP- myeloid cell activation and serves as a myeloid-specific immune checkpoint, making it a
B05 (CM369) is a humanized monoclonal antibody against CCR8 with potent ADCC activity. Here potential therapeutic target. The pan-specific SIRPa monoclonal antibody, BI 770371,
we report safety, efficacy and PK/PD findings for ICP-B05 during the dose-escalation stage of a blocks the SIRPa/CD47 interaction, leading to reactivation of innate antitumor immune
Phase I study. Methods: Patients with R/R CTCL received ICP-B05 at 150 mg, 300 mg, 450 mg responses. This Phase I trial (NCT05327946) aimed to determine the maximum tolerated
and 600 mg I.V. Q2W. Patients with R/R CTCL who failed at least 1 prior standard systemic dose (MTD) and recommended dose for expansion of BI 770371 6 ezabenlimab in patients
regimen. Primary objectives included safety and tolerability of ICP-B05, MTD and RP2D. (pts) with advanced solid tumors. Methods: Pts with $1 measurable lesion and ECOG PS of
Secondary Objectives included the PK/PD and objective response per investigator. Results: By 0/1 were enrolled. Pts received escalating doses ofBI 770371 alone or in combination with
the cutoff date of 6th Jan, 2025, a total of 13 patients with R/R CTCL were treated, with 4 patients ezabenlimab 240 mg once every 3 weeks. Treatment continued until progressive disease,
in 150 mg, and 3 patients each in 300 mg, 450 mg and 600mg, respectively. Eleven patients had a unacceptable toxicity, or pt withdrawal. BI 770731 dose escalation was guided by a Bayesian
diagnosis of MF, one had SS and one had pcALCL. The median age was 46 years, and the median Logistic Regression Model with overdose control. Primary endpoint was dose-limiting
prior lines of therapy were 3 (2-6). There were 10/13 (76.9%) patients had lymph nodes in-
toxicities (DLTs) in the MTD evaluation period (Days 1–21). Secondary endpoints were
volvement and 1/13 (7.7%) patient with SS had above 90% Sézary cells in peripheral blood at
adverse events (AEs) and DLTs in the on-treatment period. Results: At data cut-off (Nov 22,
baseline. TEAEs occurred in 12 (92.3%) patients, and $Grade 3 TEAEs occurred in 6(46.2%)
2024), 21 pts had received BI 770371 monotherapy across 6 dose levels, and 15 pts had
patients. The most common$Grade 3 TEAEs is hematological AEs, including lymphopenia
(8.3%), neutropenia (8.3%) and thrombocytopenia (8.3%). Two patients (16.7%) reported serious received BI 770371 in combination with ezabenlimab (combination group) across 5 dose
TEAEs, including edema and cardiac failure reported by the SS patient which was assessed as levels. In the monotherapy group, median age was 63 years (range: 26–77) and 95% of pts
not related to ICP-B05, and thrombocytopenia and anemia reported by a MF patient. There was had received $3 prior lines of therapy. In the combination group, median age was 61 years
no fatal TEAE reported. There were 12 patients received at least one skin lesion assessment (range: 27–78), and 80% had received $3 prior therapies. No DLTs were reported during the
followed the mSWAT. 4/12 patients (33.3%) achieved PR, and 7patients (7/12, 58.3%) were MTD evaluation period with BI 770371 monotherapy or with the combination; 1 pt in the
assessed as SD with reduction (medium: - 27%) in skin lesion. The 6-month PFS rate was 82.5% monotherapy group had a DLT (grade 2 encephalitis, which resolved within 1 week) during
(95% CI: 46.1%-95.3%). At baseline, CCR8+ in skin lesions (medium: 8.38%, range: 3.22–49.6%) the on-treatment period (likely due to prior nivolumab and ipilimumab treatment). In total, 14
was assessed in 11 out of 13 patients. Among the five patients with CCR8+ levels exceeding (67%) and 10 (67%) pts in the monotherapy and combination groups, respectively, had a
10%, four (80%) achieved PR. PK analysis showed that serum exposure (Cmax and AUC0-14D) treatment-related AE (TRAE). Most common TRAEs with BI 770371 monotherapy were
increased with dose escalation. PD analysis demontrated significant depletion of CCR8- pruritus (24%) and fatigue (19%). Most common TRAEs with the combination were fatigue
expressing cells in CTCL skin lesions. Significant reduction of CCR8+ malignant T cells and decreased appetite (each 20%). Most TRAEs were grade 1/2, one pt in the combination
(-80%) and CCR8+ regulatory T cells (Treg, -68%) were observed at C3D1 compared with baseline group had two grade 3 TRAEs (diarrhea and fatigue); there were no grade 4/5 TRAEs. Two
in the skin lesion. Similar PD effects were observed in the peripheral blood as well with an suspected unexpected serious adverse reactions were seen: grade 2 encephalitis (mon-
average decrease of 91% in CCR8+ malignant T cells and 16% in Treg at C3D1 when compared otherapy) and grade 3 diarrhea (combination). One pt had an AE leading to discontinuation
with baseline. Conclusions: The current study is the first and only report on the preliminary (grade 2 encephalitis). One pt in the combination group had a partial response; 13 (62%) and
efficacy data of anti-CCR8 targeted therapy for CTCL patients. The effectiveness of ICP-B05 was 8 (53%) pts in the monotherapy and combination groups, respectively, had stable disease.
supported by the PD effects in both skin lesions and peripheral blood in the depletion of CCR8+ Conclusions: These preliminary data indicate that BI 770371 is well tolerated alone and in
cells. ICP-B05 is safe and well tolerated and its safety profile made it a good candidate for combination with ezabenlimab, with promising antitumor activity seen in heavily pretreated
combo therapies for CTCL patients with lymph node and other organ involvement. Clinical trial pts with advanced solid tumors. The MTD of BI 770371 was not reached in either group.
information: NCT05690581. Research Sponsor: None. Clinical trial information: NCT05327946. Research Sponsor: Boehringer Ingelheim.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
142s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2516 Rapid Oral Abstract Session 2517 Rapid Oral Abstract Session
A novel application of deep learning (DL)-based MRI with liquid biomarkers Role of autoimmune reactivity in neurotoxicities (N-Tox) in melanoma
for immune effector cell-associated neurotoxicity syndrome (ICANS) after patients treated with immune-checkpoint inhibitors (ICI). First Author:
chimeric antigen receptor (CAR) T-cell therapy. First Author: Kathryn Ries Agrima Dutt, New York University Grossman School of Medicine, New York, NY
Tringale, UC San Diego, La Jolla, CA Background: N-Tox is a grossly understudied immune-related adverse event (irAE),
Background: ICANS is a complication of CAR T-cell therapy, yet risk factors and despite its association with mortality (e.g. encephalitis) and morbidities (e.g. peripheral
quantitative diagnostic criteria, particularly neuroimaging criteria, remain incompletely neuropathy). We reported that pre-treatment sera autoantibodies (auto-Abs) are im-
characterized. We implemented a novel application of a deep learning (DL)-based MRI plicated in the pathogenesis of irAEs (Johannet et al. CCR 2022). We here examined the
approach alongside clinical and liquid biomarkers to better characterize neurotoxicity rate and patterns of N-Tox in melanoma patients who received ICI in the adjuvant setting
after CAR T-cell therapy. Methods: We analyzed all patients with non-Hodgkin lym- and whether baseline specific serum auto-Abs are associated with N-Tox. Methods: We
phoma (NHL) or acute lymphoblastic leukemia (ALL) who underwent CAR T-cell therapy examined clinicopathological features and baseline auto-Abs of 965melanoma patients
at UCSD with a commercial product from 2018-2024. ICANS was graded as per American (551 male and 414 female) enrolled in two phase III clinical trials: Checkmate 238 and
Society for Transplantation and Cellular Therapy (gr1-4). Variables included stage, Checkmate 915 (797 resected stage III, 166 resected stage IV, and 2 unknown). Patients
performance status, and prior receipt of high-dose methotrexate (HD MTX), intrathecal received ipilimumab (n = 423), nivolumab (n = 347), or both (n = 195). We compared pre-
(IT) chemotherapy, central nervous system (CNS) involvement, CNS-directed radio- treatment serum auto-Ab profiles using the HuProt Human Proteome Microarray v4.0
therapy (CNS RT), and extracranial RT. Labs obtained pre-infusion, 3 days post-infusion, (CDI Laboratories, Mayaguez, PR) that has 21,000+ individually purified full-length
and during ICANS (or 7 days post-infusion for those without ICANS) were evaluated. human proteins and protein isoforms in duplicate, in patients who developed at least a
Available post-infusion brain MRIs were processed with a 3D U-Net convolutional neural single incidence of N-Tox grade $2 to those who developed only other types of irAEs
network to quantify T2 FLAIR hyperintensity volumetrics. Linear mixed regression grade $2. We used a threshold of logFC . 0.3 and false discovery rate (FDR) adjusted P
models accounting for zero inflation assessed longitudinal DL-derived FLAIR. Multi- value , 0.05 to determine differentially expressed auto-Abs in patients with N-Tox.
variable regression models assessed factors associated with ICANS. Results: Of 163 Results: 329/965 (34%) patients developed N-Tox. There were 426 total incidences of
patients (89% NHL, 11% ALL), 52 had IT chemotherapy, 27 had HD MTX, 24 had prior CNS N-tox (grade 1 n = 258, grade 2 n = 132, grade 3 n = 34, grade 4 n = 2). 97/329 patients
disease, and 22 had prior CNS RT. Most (106) received axicabtagene ciloleucel (34 developed more than one grade of N-Tox. Patients who received ipilimumab were more
tisagenlecleucel, 23 brexucabtagene autoleucel) and most had CRS (133, 82%). ICANS likely to experience any grade N-Tox (P = 0.002) in a multivariate model. Any grade N-tox
occurred in 73 (45%) at a median of 7 days post-infusion (39 gr1-2, 34 gr3-4). Post- was also associated with a lower recurrence rate (P = 0.004). Gender and melanoma
infusion, 21 patients had ³1 brain MRI (93 MRIs total). Baseline factors associated with stage were not associated with N-Tox (P . 0.05). A signature of 160 auto-Abs, including
ICANS were lactate dehydrogenase (LDH; odds ratio [OR] 1.03 p = 0.002) and prior IT those targeting mitochondrial proteins (ATP5PO, COX6C, NDUFA3, NDUFB6), calcineurin
chemotherapy (OR 2.5 p = 0.01). There was a trend toward association of gr3-4 ICANS (PPP3CC, PPP3R1), cellular architecture (RAC1), and inflammation/apoptosis (TRAF2)
with HD MTX (OR 2.8 p = 0.07). Post-infusion, CRS grade was associated with ICANS (OR were significantly overexpressed in the N-Tox grade $2 cohort (n = 143) compared to
2.8 p , 0.001). LDH (1.02 p = 0.004) and C-reactive protein (OR 1.2 p , 0.001) were non-N-Tox grade $2 (n = 569). Pathway analysis revealed these auto-Abs were enriched
elevated during ICANS. Patients with ICANS had significantly greater FLAIR (intercept in several pathways involved in neuroinflammation and neurodegeneration, including
23.8 cm³ p , 0.001) and there was increased FLAIR over time across all patients (b = TNF-a signaling, B cell receptor signaling, interleukin-2 production, natural killer cell
3.3 cm³ p = 0.05). There was a trend toward association between higher ICANS grade and mediated cytotoxicity, and cellular senescence. Conclusions: Our data demonstrate
DL-derived FLAIR (p = 0.09). Conclusions: Here, we demonstrate a novel application of that the incidence of N-Tox is higher than previously reported, possibly due to stringent
DL-based MRI quantification of ICANS post-CAR T-cell therapy. This metric, along with assessment and follow up in clinical trial settings. The multiplicity of pathways involved,
clinical features, emerged as potential quantitative biomarkers of ICANS. These findings some of them directly involved in neurodegeneration and neuroinflammation, suggests a
warrant further investigation and have informed a prospective study, including stan- complex N-tox pathogenesis that requires further clinical and pre-clinical investigations.
dardized brain MRI pre- and post-infusion, to develop a comprehensive phenotype of Research Sponsor: National Cancer Institute; P50CA225450.
neurotoxicity following CAR T-cell therapy. Research Sponsor: None.

2518 Rapid Oral Abstract Session 2519 Rapid Oral Abstract Session
Natural killer cell transcriptomic expression and prediction of survival after Tumor-wide RNA splicing aberrations and their potential as therapeutic
immune checkpoint blockade across cancers. First Author: Hirotaka Miyashita, neoantigen targets. First Author: Darwin Kwok, University of California, San Fran-
Dartmouth Cancer Center, Lebanon, NH cisco, San Francisco, CA
Background: Preclinical and clinical evidence has suggested the role of natural killer Background: Tumor heterogeneity and low mutational burden limits the availability of
(NK) cells in tumor immunity and prognosis across various cancer types, but their effective immunotherapy targets. Aberrant RNA-splicing (neojunctions) represents an
significance during immune checkpoint blockade (ICB) treatment is poorly understood. underexplored yet promising source of neoantigens. To address this, we developed a
This study investigated the impact of tumor-infiltrating NK cells, surrogated by the RNA neoantigen discovery platform (SNIPP) that characterizes a novel class of clonally-
expression of genes related to NK cells in the tumor microenvironment, on the outcomes expressed, splicing-derived neoantigens. Furthermore, we validated the immunogenicity
of the patients who undergo ICB, using real-world, pan-cancer data. Methods: We an- of these neoantigens by identifying specific TCRs that drive CD8+ T-cell-mediated tumor
alyzed RNA sequencing data of 395 immune-related genes from 514 patients with various killing. Methods: SNIPP identified public neojunctions by analyzing TCGA RNA-seq data,
cancers included in the Study of Personalized Cancer Therapy to Determine Response and selecting neojunctions with a positive sample rate (PSR) . 10% and filtering out those
Toxicity (NCT02478931). After excluding 25 patients ineligible for survival analysis, we found in GTEx normal tissue RNA-seq data (PSR , 1%) across 12 cancer types. To
defined two distinctive cohorts: patients who received ICB (ICB cohort, N = 217) and those characterize intratumorally conserved neojunctions, we performed maximally-distanced
who did not (non-ICB cohort, N = 272). Among the 395 immune-related genes, 43 were multi-site biopsies (n = 535) within glioma patients (n = 56) and generated RNA-seq data for
selected as NK-related genes according to the Human Protein Atlas. Patients in each each intratumoral site. Two independent algorithms were utilized to predict peptide
cohort were clustered into two groups based on the NK-related gene expression. The processing likelihood and HLA-binding affinity of splicing-derived neoantigen candidates.
associations between the clusters and the clinical outcomes, including overall survival Neoantigen-specific TCR sequences were identified via in vitro sensitization of PBMCs and
(OS) and progression-free survival (PFS), were analyzed using univariate and multivariate subsequent 10x V(D)J scRNA-seq. These TCRs were transduced into CD8+ T-cells, which
analyses. In the multivariate analysis, cancer types, line of immunotherapy, positive were tested downstream for immunogenicity and cytotoxicity against glioma cell lines.
programmed-death ligand 1 immunohistochemistry (PD-L1 IHC, $ 1%), high tumor Results: Our pipeline identified 789 public neojunctions, including 32 neojunctions con-
mutational burden (TMB, $ 10/Mb), and microsatellite instability (MSI) were adjusted. currently detected in transcriptomic and proteomic glioma datasets and confidently
Results: The ICB cohort (N = 217) was divided into two clusters (hot vs. cold), char- predicted to be presented by HLA-A*02:01. IVS and subsequent 10x V(D)J scRNA-seq
acterized by general abundance and paucity of NK-related gene transcripts (N = 101 and identified TCR clonotypes reactive against neojunctions in RPL22 (n = 7) and GNAS (n = 1),
116, respectively). The clusters were not significantly associated with histology, positive with the latter exhibiting high intratumoral conservation (detected in . 90% of spatially-
mapped biopsies across 17/56 patients (26.78%)). TCR-transduced CD8+ T-cells recognized
PD-L1 IHC, high TMB, or MSI. Those in the hot cluster demonstrated significantly longer
and were immunogenically activated and demonstrated cytotoxicity against endogenously
overall survival (OS) after starting ICB compared to those in the cold clusters in univariate
processed and presented neoantigens in GBM and melanoma lines. Additionally, IDH1-
analysis (hazard ratio [HR] and 95% confidence interval [CI]: 0.65 [0.45-0.92], p = 0.015)
mutant oligodendrogliomas exhibited significantly higher neojunction expression compared
and multivariate analysis (HR and 95% CI: 0.57 [0.34-0.87], p = 0.010). The cluster was not
to IDH1-mutant astrocytomas and IDH1wt subtypes. Differential gene expression analysis
significantly associated with PFS. The non-ICB cohort (N = 272) was similarly divided into
(DESeq2) revealed reduced expression of splicing factors in oligodendrogliomas, attributed
two clusters (hot vs. cold), with the characteristics of generally high and low NK-related to their specific co-deletion of chromosomes 1p and 19q. CRISPRi-mediated knockdown of
gene RNA expressions. (N = 114 and 158, respectively). However, in the non-ICB cohort, these splicing factors (e.g. SF3A3, SNRPD2) in IDH1wt glioma cells resulted in significantly
patients in the hot clusters did not demonstrate significantly prolonged OS compared with increased expression of corresponding neojunctions. Conclusions: Our study highlights a
those in the cold cluster either with univariate or multivariate analysis (HR and 95% CI: novel class of neoantigens derived from tumor-wide aberrant RNA splicing. The SNIPP
0.93 [0.65-1.32], p = 0.67 and 0.97 [0.76-2.01], p = 0.90 respectively). platform effectively identifies public intratumorally-conserved neojunctions with strong
Conclusions: Transcriptomic expression of NK-related genes in tumor tissue indepen- therapeutic potential. Furthermore, elevated neojunction expression in oligodendroglioma
dently and significantly predicted longer survival after ICB treatment, which implies a role underscores the mechanistic link between dysregulated splicing factor expression and RNA
of tumor infiltrating NK cells in immunotherapy outcome. Research Sponsor: U.S. Na- splicing abnormalities. Research Sponsor: None.
tional Institutes of Health; CA023100.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 143s

2520 Rapid Oral Abstract Session 2521 Poster Session

Perturbational single-cell RNA sequencing of patient tumors in Merkel cell Phase 1/2, open-label, first-in-human study of the anti-GPC3 T-cell engager
and small cell lung carcinomas. First Author: Curtis J. Perry, Yale School of SAR444200 in patients with advanced solid tumors: Updated efficacy and
Medicine, New Haven, CT biomarker analysis. First Author: Ecaterina Elena Dumbrava, Department of Inves-
Background: Despite the transformational impact of immune checkpoint blockade, many cancer tigational Cancer Therapeutics, The University of Texas MD Anderson Cancer Center,
patients do not experience long-term survival. T cells with innate immune signatures can secrete Houston, TX
inflammatory cytokines/chemokines and deliver potent cytotoxic signals potentially ideal for Background: SAR444200 is a novel NANOBODY T cell engager that simultaneously binds TCRab and
tumor immunity. The novel double-stranded RNA sensor RIG-I agonist SLR14 improved the control glypican-3 (GPC3) to co-engage T cells with GPC3+ tumor cells, resulting in T cell-dependent cellular
of murine melanoma. We tested the hypothesis that SLR14 transforms T cells to a cytotoxic state in cytotoxicity. We present updated safety, efficacy, and biomarker data from the dose escalation cohort
immunologically “cold” human tumor specimens. Methods: We developed an approach, called (Part 1A) of a multicenter, first-in-human, Phase 1/2 trial (NCT05450562). Methods: Patients (Pts)
PERCEPT, to directly test the response of patient tumor and immune samples to novel and with GPC3+ refractory solid tumors received SAR444200 intravenously (IV) weekly with lead-in doses
established therapies ex vivo using perturbational single-cell RNA sequencing (Table). We obtained at dose levels (DLs) 1 (3 mg), 1A (1 mg), 2A (2.5 mg), 3A (4.5 mg), 4A (18 mg), 5A (36 mg), 6A and 7A
9 surgical resections from primary or metastatic melanoma and Merkel Cell Carcinoma (MCC) (different lead-in doses, target dose 70 mg). Pts with ECOG PS #1, and $1 measurable lesion per
tumors and lymph node metastases and made suspension replicates of tumor and infiltrating RECIST 1.1, were eligible. Primary objective for Part 1A was safety. Key secondary objectives included
immune cell co-cultures. We stimulated for 42-48 hours (Table). We Flourescently Activated Cell efficacy and PK/PD. Levels of interleukin-6 (IL-6) and interferon gamma (IFNg) were evaluated with a
Sorted live cells and then barcoded for multiplexed single-cell sequencing using 10x scRNAseq. We multiplex ECL-based assay. Study imaging was performed every 9 weeks following first infusion.
used CINEMA-OT to identify factors associated with response and resistance to the perturbations Circulating tumor (ct) DNA was analyzed in blood samples using a mutational profiling NGS approach.
tested. We developed and validated CRISPR-KO MCC and small cell lung cancer (SCLC) cell lines, Results: As of October 15, 2024, 33 pts (23 pts with hepatocellular carcinoma [HCC]) were treated with
and co-cultured with CD14+ monocytes or monocyte-derived DCs. Results: Stimulation with RIG-I SAR444200 (premedicated with dexamethasone 15 mg IV or equivalent) for a median of 23 cycles
agonist SLR14 induced expression beyond canonical IFN-stimulated genes in tumor cells, NK cells, (range, 1–32). Median lines of prior therapies were 3–4. Most pts (32 [97%]) experienced $1 AE of any
and T cells. SLR14 stimulates tumor-infiltrating T cells into antiviral states in tumor-immune co- Grade. Grade $3 TEAEs in 16 (48.5%) pts and serious TRAEs in 8 (24.2%) pts were reported. 2 Grade 3
cultures and primes in vitro T-cell production of IFNg. However, MCC immune infiltrate re- cytokine release syndrome events were reported as DLTs at DL6A and 1 at DL7A during the lead-in
sponsiveness to IFN or SLR14 was notably decreased compared to the melanoma samples, and dosing. Key efficacy data are summarized in Table 1. An increase in IL-6 and IFNg (maximum of 1326
perturbational computational analyses with CINEMA-OT identified the cytokine midkine (MDK) pg and 461 pg on average per DL, respectively) was observed during lead-in-doses in pts from DL1 to
associated with nonresponse in MCC. Knockout of MDK restored response to IFN and SLR14 by DL5A, supporting CRS diagnosis. Cytokine levels declined after Cycle 1. Of 18 HCC pts with baseline
MCC and SCLC tumor cell lines, as well as co-cultured CD14+ monocytes or monocyte-derived DCs. alpha fetoprotein (AFP) $20%, 5 (27%) pts showed $50% AFP reduction. Median time of observing
Conclusions: Our approach revealed that midkine, a multifunctional cytokine, suppresses innate any AFP decrease was 4 weeks post-treatment. Among these, 3 pts had sustained decrease over 13
immune sensing of IFN and SLR14 in both tumor and immune cells, disrupting the tumor immunity cycles. Stable disease (SD) was reported in 10 (30.3%) pts including 2 who were on study drug for 12
cycle at multiple points. We show that this effect, while comparatively infrequent in melanoma, is and 22 months. Of the 18 pts with measurable ctDNA, 4 (including 3 pts at DL5 and above) had
pronounced in MCC and SCLC. Our study thus uses a direct assessment of patient tumor and reductions in ctDNA (18%–48%) from baseline. Conclusions: SAR444200 was tolerated at the in-
vestigated DLs in pts with GPC3+ advanced solid tumors. Decrease in AFP post treatment along with
immune samples to identify a novel resistance mechanism enriched in neuroendocrine tumors
SD in a subset of pts is suggestive of preliminary anti-tumor activity. Clinical trial information:
MCC and SCLC. Research Sponsor: Conquer Cancer/ASCO Young Investigator Award 2023;
NCT05450562. Research Sponsor: Sanofi.
Astrazeneca; U.S. National Institutes of Health; 1R37CA279834-01A1.
Efficacy analysis.
Therapeutic class Stimulation Target Clinical development DL1 DL1A DL2A DL3A DL4A DL5A DL6A DL7A
3 mg 2W 1 mg 2W 2.5 mg 2W 4.5 mg 2W 18 mg 3W 36 mg 3W 70 mg 3W 70 mg 3W All
Immune checkpoint inhibitor aPD-1 PD-1 Standard of care n (%) n=4 n=4 n=4 n=4 n=4 n=6 n=3 n=4 (N = 33)
Cytokine IFNg IFNGR Early-phase clinical trials
IFNb IFNAR or BOR 1 (25.0) 1 (25.0) 1 (25.0) 1 (25.0) 2 (50.0) 3 (50.0) 1 (33.3) 0 10 (30.3)
Innate immune agonist Poly(I:C)-NT TLR3 Pre-clinical Stable Disease 3 (75.0) 3 (75.0) 3 (75.0) 2 (50.0) 2 (50.0) 2 (33.3) 1 (33.3) 2 (50.0) 18 (54.5)
Progressive
Poly(I:C)-T MDA5/TIG-I/TLR3 disease
ADU-S100 STING
SLR14 RIG-I BOR, best overall response; DL, dose level; W, weeks.
Combination aPD-1 + IFNb PD-1/IFNAR

2522 Poster Session 2523 Poster Session

Lacutamab in patients with relapsed and refractory Sézary syndrome: Long Lacutamab in patients with relapsed and/or refractory mycosis fungoides:
term follow-up from the TELLOMAK phase 2 trial. First Author: Pierluigi Porcu, Long-term follow-up and translational data from the TELLOMAK phase 2
Sidney Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA trial. First Author: Pierluigi Porcu, Sidney Kimmel Cancer Center, Thomas Jefferson
Background: Sézary syndrome (SS) is a rare and aggressive cutaneous T-cell lymphoma, University, Philadelphia, PA
which commonly expresses KIR3DL2, a killer immunoglobulin-like receptor, reported in $ Background: The most common type of cutaneous T-cell lymphoma is Mycosis Fungoides
85% of patients. SS is characterized by erythroderma, significant blood involvement, (MF) accounting for 50-60% of cases. Extracutaneous involvement occurs mainly in lymph
lymphadenopathy and poor prognosis (10-20% 5-year survival). Lacutamab is a first-in- nodes or blood; 25% of patients are diagnosed at advanced stage with a 5-year survival of 15-
class monoclonal antibody designed to specifically deplete KIR3DL2-expressing cells via 25%. Lacutamab is a first-in-class monoclonal antibody designed to specifically deplete
antibody-dependent cell-cytotoxicity and phagocytosis. Methods: TELLOMAK is an in- KIR3DL2-expressing cells via antibody-dependent cell-cytotoxicity and phagocytosis.
ternational, Phase 2 trial with multiple cohorts (NCT03902184). We report here long term KIR3DL2 is a killer immunoglobulin-like receptor expressed in MF patients.
follow-up results from Cohort 1, evaluating lacutamab in patients with relapsed/refractory Methods: TELLOMAK is an international, multi-cohort phase 2 trial (NCT03902184). MF
(R/R) SS after at least 2 prior systemic therapies including mogamulizumab. Lacutamab patients who had received at least 2 prior systemic therapies were treated with lacutamab
750 mg is administered until progression or unacceptable toxicity. Primary endpoint was 750 mg until disease progression or unacceptable toxicity. Primary endpoint was Objective
Objective Response Rate (ORR) based on the evaluation of 4 compartments: skin, blood, Response Rate (ORR) by global response score based on the evaluation of 4 compartments:
lymph nodes and viscera according to the International Consensus criteria Olsen 2011. skin, blood, lymph nodes and viscera according to the International Consensus criteria Olsen
Secondary endpoints included but were not limited to duration of response (DOR), pro- 2011. Key secondary endpoints included duration of response (DoR), progression free
gression free survival (PFS), safety, and quality of life assessments. Results: As of October survival (PFS), safety, and quality of life. Here we report long term follow-up data of MF
17, 2024, recruitment was completed with 63 SS patients enrolled. Median age was 69 years patients. Results: As of October 17, 2024, recruitment was completed, with 107 MF patients
(range: 42-86), the median prior lines of systemic therapies were 5.0 (range: 2-13), 65.1% enrolled. The median age was 62 years. The median number of previous systemic lines was 4
and 34.9 % patients had stage IVA1 and stage IVA2 at baseline respectively, all patients had (range: 1-14). Median follow-up was 22.1 months (m) (95% CI 19.4, 23.6). Global confirmed
blood involvement (B2), 63.5% had confluence of erythema covering $ 80% body surface ORR was 19.6% (CI 13.2, 28.1; Olsen 2011), and response in skin was 29.0% (CI 21.2, 38.2).
area (T4), 34.9% had lymph node lymphoma involvement (N3). Median follow-up was Median time to response was 2.8 m (min, max 1-37) and median DoR was 13.8 m (7.4, NE),
25.1 months (95% CI 21.0-29.4). Global confirmed ORR was 42.9% (CI 31.4-55.1) including 6 median PFS was 10.2 m (CI 8.0, 15.4). Among the KIR3DL2 $1% pts (N = 48), ORR was 20.8%
(9.5%) CRs who are all still in CR; with a median time to response of 2.8 months (range 1-10) (CI 11.7, 34.3; Olsen 2011), and response in skin was 33.3% (CI 21.7, 47.5), median DoR was
and a median duration of response of 25.6 months (CI 11.0, NE). According to each 13.8 m (CI 4.6, NE) and median PFS 11.8 m (CI 5.6, 16.8). Among the KIR3DL2 , 1% pts (N =
compartment, ORR in skin was 52.4% (CI 40.3-64.2) including 9 (14.3%) CRs, ORR in blood 59), ORR was 18.6% (CI 10.7;30.4; Olsen 2011), and response in skin was 25.4 (CI 16.1, 37.8),
was 50.8% (CI 38.8-62.7) including 21 (33.3) CRs, and ORR in lymph nodes was 28.8% (CI median DoR was 15.7 m (CI 5.1, NE) and median PFS 9.5 m (CI 6.5;16.6). Grade $ 3 related
18.3-42.3) including 9 (17.3) CRs. Median PFS was 8.3 months (CI 5.1-18.7). Grade $ 3 Treatment-Emergent Adverse events (TEAEs) were observed in 5/107 (4.7%) patients, se-
related Treatment-Emergent Adverse Events (TEAEs) were observed in 20.6% patients. rious related TEAEs in 4/107 (3.7%) patients and related TEAEs leading to study drug
Serious related TEAEs were observed in 9.5% patients and related TEAEs leading to study discontinuation in 3/107 (2.8%) patients. The most common ( . 10%) related TEAEs were
drug discontinuation in 6.3% patients. Data from additional key endpoints will be presented. fatigue (12.1%), nausea (13.1%), asthenia (11.2%) and arthralgia (11.2%). Data from ad-
Conclusions: The long term follow-up data from TELLOMAK study in a R/R SS population ditional key endpoints and translational data will also be presented. Conclusions: The long-
previously treated with 2 or more prior systemic therapies including mogamulizumab, term follow-up data from the heavily pre-treated MF population enrolled to the TELLOMAK
confirm that lacutamab shows promising clinical activity with ORR 42.9% (95% CI 31.4- study confirms promising clinical activity of lacutamab regardless of KIR3DL2 expression,
55.1) and median duration of response of 25.6 months (11.0, NE) and an overall favourable with ORR 20.8%, a median duration of response of 13.8 m, a median PFS of 10.2 m and a
safety profile. These data support the further development of lacutamab in an effort to bring favorable safety and tolerability profile. These data support the further development of
improved treatments to patients with SS. Clinical trial information: NCT03902184 // EU CT lacutamab in an effort to bring improved treatments to patients with MF. Clinical trial
number: 2023-507777-18-00. Research Sponsor: Innate Pharma. information: NCT03902184 // EU CT number: 2023-507777-18-00. Research Sponsor: IN-
NATE PHARMA.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
144s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2524 Poster Session 2525 Poster Session

Safety and efficacy of OR502, an antibody targeting leukocyte A phase 1 study of the OX40 agonist BGB-A445, with or without tislelizumab,
immunoglobulin-like receptor B2 (LILRB2), 6 cemiplimab in patients with an anti-PD-1 monoclonal antibody, in patients with advanced NSCLC,
advanced solid tumors from a phase 1 study. First Author: Shiraj Sen, NEXT HNSCC, or NPC. First Author: Min Hee Hong, Yonsei Cancer Center, Yonsei Uni-
Oncology, Dallas, TX versity College of Medicine, Seoul, South Korea
Background: OR502 is a humanized IgG1 antibody that targets LILRB2, blocking its binding to HLA Background: BGB-A445 is a monoclonal antibody OX40 agonist that does not compete with the natural
ligands A, B and G. OR502 prevents and reverses myeloid cell-mediated immune suppression and rescues OX40 ligand, reducing the likelihood of a hook effect and distinguishing it from other OX40-targeting
T cell effector functions. Preclinical data demonstrate best-in-class properties. We report on the therapies. Here, we present results from the dose expansion portion of a ph 1, open-label, dose
completed monotherapy and combination dose escalation cohorts from the ongoing, first-in-human, escalation/expansion trial of BGB-A445 in pts with advanced solid tumors (NCT04215978). Ph 1a
phase 1-2 study of this novel antibody. Methods: Patients had progressive, histologically confirmed, results were previously presented (Desai et al. J Clin Oncol. 2023). Methods: Previously treated pts with
metastatic/unresectable solid tumors with $ 1 prior systemic standard of care treatments. Primary NSCLC (Part A1), HNSCC (Part A2), or NSCLC with PD-L1 $50% (Part C) received BGB-A445 mon-
objectives were OR502 safety/tolerability and identifying a dose for future study supported by LILRB2 otherapy, while pts with treatment-naı̈ve recurrent/metastatic NPC (Part B) received BGB-A445
receptor occupancy (RO) and pharmacokinetics (PK). Secondary objectives included assessment of anti- combined with tislelizumab and chemotherapy. Primary endpoints included ORR per investigator
tumor activity. We used a modified toxicity probability interval-2 design with a 25% dose-limiting toxicity (RECIST v1.1); secondary endpoints were to assess PFS, DOR and DCR, safety/tolerability, PK, and host
(DLT) rate and a 20–30% equivalence interval. Patients received OR502 IV (100–1600 mg) over 30 immunogenicity. Results: As of Sep 25, 2024, 54 pts were enrolled in Part A1, 19 in Part A2, 12 in Part B,
minutes, every 3 weeks (Q3W) as monotherapy (n = 19) or with cemiplimab (350 mg) (n = 20). Results: In and 7 in Part C. In the efficacy evaluable analysis set, ORR was 0% in Parts A1, A2, and C, and 70% (7/10;
dose escalation (n = 39), there were no DLTs, treatment-related deaths, related SAEs, grade $ 3 all confirmed PRs, one unconfirmed CR) in Part B. In Parts A1, A2, B, and C, confirmed DCR was 49.0%,
treatment-related AEs or signals from vital signs, ECGs or laboratory results. One patient (monotherapy, 33.3%, 100.0%, and 57.1%, respectively. TEAEs occurred in the majority of pts (Table). The most
400 mg) discontinued due to grade 2 AEs. Infusion-related reactions (IRRs) occurred in 6 patients (3 common treatment-related TEAEs were pyrexia (10.0% [8/80]), chills (5.0% [4/80]), and anemia (5.0% [4/
monotherapy [1 at 800 mg and 2 at 1600 mg] and 3 combination [400, 800 and 1600 mg]). All IRRs were 80]) in the monotherapy cohorts, and anemia (75.0% [9/12]), decreased WBC (66.7% [8/12]), decreased
grade # 2 and were mitigated by extending infusion duration to 60 minutes, with secondary prophylaxis if neutrophils, and decreased platelets (58.3% [7/12], each) in the combination cohort. Treatment-related
necessary (acetaminophen, diphenhydramine). All but 4 patients were evaluable for efficacy, see table. serious TEAEs occurred in 2.5% (2/80; pyrexia and asthenia in a single pt each) of pts in the mon-
Monotherapy responses were seen at 200 and 800 mg in melanoma and non-small cell lung cancer otherapy cohorts and 8.3% (1/12; febrile neutropenia) in the combination cohort. There were no BGB-
(NSCLC), respectively. In combination, 1 patient with soft tissue sarcoma (1600 mg) had a cPR. There A445 or tislelizumab-related TEAEs leading to treatment discontinuation or death. The most common
were 13 deaths due to progressive disease. Durable stable disease (SD) was seen in: sarcomas, cutaneous imAE was rash (2.5% [2/80] in the monotherapy cohort; 33.3% [4/12] in the combination cohort). No
squamous cell carcinoma, thymoma, thyroid, melanoma, hepatocellular carcinoma and colorectal cancer. Gr $3 imAEs or IRRs were reported. Conclusions: BGB-A445 alone or in combination with tislelizumab
OR502 RO was near-complete at $ 200 mg and PK was roughly dose-proportional. Combination with and chemotherapy was generally well tolerated across all doses in pts with advanced NSCLC, HNSCC,
cemiplimab did not affect RO or PK. Conclusions: OR502 has excellent safety and tolerability 6 and NPC, and showed preliminary antitumor activity. Clinical trial information: NCT04215978. Research
cemiplimab. Based on efficacy, predictable PK and near-complete RO, two mini-expansion cohorts are Sponsor: BeOne Medicines Ltd.
evaluating OR502 800 mg Q3W 6 cemiplimab in patients with cutaneous melanoma or NSCLC who have
failed or progressed after $ 12 weeks of anti-PD-(L)1. Clinical trial information: NCT06090266. Research Safety.
Sponsor: OncoResponse, Inc.; The Cancer Prevention and Research Institute of Texas. Part C
Part A1 Part A2 Part B NSCLC and
Best objective response (RECIST 1.1), PK and RO. NSCLC HNSCC NPC PD-L1 ‡50%
OR502 (n=17) OR502 + cemiplimab (n=18) (N=54) (N=19) (N=12) (N=7)

PR 2 1 Any treatment-emergent AE 47 (87.0) 16 (84.2) 12 (100.0) 7 (100.0)

cPR 1 1 Gr ‡3 17 (31.5) 5 (26.3) 11 (91.7) 3 (42.9)
SD 9 8 Serious 21 (38.9) 4 (21.1) 2 (16.7) 4 (57.1)
Durable SD (‡ Week 12) 7 4 Leading to death 4 (7.4) 2 (10.5) 0 (0) 0 (0)
Best overall response rate % 12 6 Leading to treatment discontinuation 8 (14.8) 3 (15.8) 2 (16.7) 0 (0)
Disease control rate (CR+PR+SD)% 65 50 Any treatment-related treatment- 28 (51.9) 7 (36.8) 12 (100.0) 3 (42.9)
PK (100–1600 mg) 7.6–15.9 8.4–12.4 emergent AE
t1/2 (day) Gr ‡3 1 (1.9) 0 (0) 11 (91.7) 0 (0)
Peripheral RO% (100–1600 mg) 91–101 88–99 Any immune-mediated AE 6 (11.1) 1 (5.3) 6 (50.0) 1 (14.3)
Classical monocytes Infusion-related reactions 6 (11.1) 3 (15.8) 3 (25.0) 1 (14.3)
Neutrophils 89–100 84–100
Pts with multiple adverse events (AEs) are counted once. All AEs are listed as n (%).

2526 Poster Session 2527 Poster Session

Phase 1b dose extension study of a next-generation anti-CD47 monoclonal Safety and efficacy of QLS31905 in patients with advanced solid tumors:
antibody IMC-002 combined with lenvatinib in patients with advanced Updated data from phase 1 study. First Author: Yakun Wang, Key Laboratory of
hepatocellular carcinoma (HCC). First Author: Jung Yong Hong, Division of He- Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department
matology-Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan of Gastrointestinal Oncology, Peking University Cancer Hospital and Institute, Beijing,
University School of Medicine, Seoul, South Korea China
Background: IMC-002 has shown significant preclinical efficacy and safety, which are attributed to Background: QLS31905 is a Claudin18.2/CD3 bispecific antibody. Here we report the updated
its unique binding site and distinct mechanism of action. These preclinical findings strongly data of a phase 1 study of QLS31905. Methods: This multicenter phase 1 trial (NCT05278832)
support its clinical development as a cancer therapeutic. Phase 1a trial confirmed its superior recruited patients (pts) with advanced solid tumors who had progressive disease or were
safety and tolerability. Here we present initial results from the phase Ib trial, focusing on safety, intolerable to or inapplicable of standard therapy. In dose-escalation stage adopting accelerated
efficacy, PK, and biomarker. Methods: Eligible pts had advanced HCC that progressed following at titration and interval 3+3 design, pts regardless of Claudin18.2 status were administered
least 1 prior systemic therapy and ECOG PS #1. IMC-002 was administered at 20 mg/kg Q3W in QLS31905 via intravenous infusion in 11 sequential single doses (0.5, 1.5, 5, 15, 45, 100, 200,
combination with Lenvatinib, continuing until disease progression. Tumor assessments were 350, 500, 800, 1200 mg/kg qw or q2w) with priming dose from 350 mg/kg. In dose-expansion
conducted every 6 weeks using RECIST 1.1 and iRECIST. A target-mediated drug disposition stage, Claudin18.2-positive ($1% tumor cells) pts were recruited. The primary endpoint was
(TMDD) PK model incorporating FcRn recycling was developed to predict PK values for Q3W dosing dose limiting toxicities (DLT) and maximum tolerated dose (MTD) in dose-escalation stage, and
and evaluated for consistency with observed data. Immunohistochemistry (IHC) images of CD47 was objective response rate (ORR) in dose-expansion stage. Results: As of Jul 26, 2024, 31 pts
expression were analyzed using Lunit SCOPE uIHC, an AI-based platform capable of distinguishing were included from 0.5 mg/kg qw to 1200 mg/kg q2w in dose-escalation stage, and 48 pts were
staining positivity and cell types at the single cell level. Results: A total of 13 pts with refractory included in five cohorts (100~200 mg/kg qw and 350~800 mg/kg q2w) in dose-expansion stage.
HCC received IMC-002 in combination with Lenvatinib. Most patients had received prior anti-PD-(L) The 1200 mg/kg q2w cohort is ongoing. There were 43 (54.4%) pts with gastric or gastro-
1 therapy (11 pts) and had an ECOG PS of 1 (9 pts). Among the 10 pts evaluable for efficacy, the esophageal junction (G/GEJ) cancer and 26 (32.9%) with pancreatic adenocarcinoma (PAC).
ORR was 30%, and the DCR was 70%. The median TTP was 8.3 months. AI-driven analysis of CD47 Over half of (61.8%) pts had received $2 lines of prior treatment. No DLT occurred. MTD was not
membrane specificity, using a subcellular model, revealed that samples with a high proportion of reached. Treatment-related adverse events (TRAEs) occurred in 79 (100%) pts, of whom 34
non-membrane-specific cells were associated with poor clinical outcomes (ORR 0%, DCR 33%). In
(43.04%) were $grade 3. The most common $grade 3 TRAEs ($3%) were lymphocyte count
contrast, samples with a low proportion demonstrated improved responses (ORR 60%, DCR 80%).
decreased (21.5%), g-glutamyl transferase increased (3.8%), neutrophil count decreased (3.8%),
We confirmed that 96.3% of the observed concentrations of IMC-002 in Phase 1b not only fell within
cytokine release syndrome (CRS [3.8%]), and anemia (3.8%). CRS occurred in 17 (21.52%) pts
the 90% prediction percentiles of PK model developed for Q3W dosing schedule but also dem-
including two pts with grade 3 and one with grade 4, and all recovered. Two pts (2.53%)
onstrated steady-state achievement (after cycle 2 of Q3W) and consistent Ctrough exposure above
discontinued treatment due to TRAEs of abdominal pain and CRS, respectively. No TRAE leading
the MEC (. 24 mg/mL). All TRAEs were grade 1-2 (100%), with 92% occurring during cycle 1. TRAEs
reported in more than one patient included skin rash and transient vitreous floaters. Anemia was to death occurred. In 33 Claudin18.2-positive pts in 350~1200 mg/kg q2w cohorts, six pts (three
observed in only one patient, while no cases of neutropenia, thrombocytopenia, or treatment- with G/GEJ cancer and three with PAC) had partial response. ORR was 18.18% (95% confidence
related SAEs were reported. Conclusions: IMC-002, when combined with Lenvatinib at a dose of 20 interval [CI]: 6.98%, 35.46%), disease control rate (DCR) was 87.88% (95% CI: 71.80%, 96.60%),
mg/kg Q3W, demonstrated a promising efficacy and safety profile. AI-driven biomarker analysis median progression-free survival (PFS) was 4.21 months (95% CI: 2.99, 5.55), and median overall
identified potential predictive value, supporting the need for further investigation in larger clinical survival (OS) was 9.53 months (95% CI: 7.69, not evaluable). Among the Claudin18.2-positive pts
trials. Clinical trial information: NCT05276310. Research Sponsor: ImmuneOncia Therapeutics Inc. in 350~1200 mg/kg q2w cohorts, ORR, DCR, median PFS, median OS was 15.79%, 89.47%,
4.40 months, 9.20 months in 19 pts with G/GEJ cancer, and was 25.00%, 91.67%, 3.94 months,
Tumor response by CD47 non-specific cell proportion.
not reached in 12 pts with PAC, respectively. QLS31905 exposure was generally linear with the
Cohort Non-specific cell proportion Mean (6SD) ORR DCR administered dosage. There was no tendency of accumulation after multiple administrations.
A 0.24 (60.05) 0% 33% Conclusions: QLS31905 was safe and tolerable, and showed encouraging efficacy in Clau-
B 0.05 (60.05) 60% 80% din18.2-positive pts with gastrointestinal tumors. QLS31905 is worthy of further exploration in
p-value 0.00 0.03 0.14
combined therapy in phase 2 trials. Clinical trial information: NCT05278832. Research Sponsor:
Cohort A: High proportion ($15%) of ’non-specific’ cells; Cohort B: Low proportion (,15%) of ’non-specific’ cells. None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 145s

2528 Poster Session 2529 Poster Session

Epigenetic and phenotypic signatures of T-cell response to blinatumomab in PD-1 blockade in combination with bevacizumab and nab-paclitaxel for
pediatric relapsed and refractory B-ALL. First Author: Tyler G. Bruno, St. Jude second-line treatment in cancer of unknown primary (Fudan CUP-002): A
Children’s Research Hospital, Memphis, TN prospective, single-arm phase II study. First Author: Zhiguo Luo, Department of
Background: By forming an immunological synapse between T cells and tumor antigen, Medical Oncology, Fudan University Shanghai Cancer Center, Shanghai, China
bispecific T cell engagers (BiTEs) like blinatumomab have shown great promise in treating B- Background: Cancer of unknown primary (CUP), a heterogeneous tumor characterized by
cell acute lymphoblastic leukemia (B-ALL). However, many relapsed and refractory (R/R) histologically confirmed metastases with undefined primary, accounts for 2-5% of all
patients fail to achieve long-term survival, with 40% not surviving past 24 months. Prolonged malignancies. Our previous study, Fudan CUP-001, confirmed that site-specific first-line
T cell activation with blinatumomab therapy may lead to changes in differentiation that treatment improves progression-free survival (PFS) in patients with CUP compared to
leave the T cell population unable to elicit a sustained anti-tumor response. A deeper empirical treatment. However, no evidence-based standard of care currently exists for
understanding of the dynamics of the T cell compartment in R/R B-ALL patients will lead to second-line treatment of CUP. We conducted the Fudan CUP-002 study by Simon’s two-
improved treatment strategies and optimized patient selection for blinatumomab therapy. stage design to evaluate the efficacy and safety of co-administration of F520 injection (anti-
Methods: To characterize T cell persistence and response in this context, we assessed PD-1 antibody), bevacizumab, and nab-paclitaxel in patients with CUP who have progressed
memory and exhaustion phenotypes in blinatumomab-treated T cells isolated from 10 R/R after first-line treatment. Methods: In this prospective, single-arm phase II study (Clin-
pediatric B-ALL patients treated with blinatumomab. CD8+ T cells were isolated from [Link], NCT04848597), patients with previously treated CUP received intravenous
peripheral blood and bone marrow samples and analyzed for memory and exhaustion F520 injection at a dose of 200 mg and bevacizumab 7.5 mg/kg every 3 weeks for up to 2
phenotypes via flow cytometry. Absolute lymphocyte counts were measured and linked to years, and intravenous nab-paclitaxel 125 mg/m2 administered on day 1 and day 8 every
the sample flow cytometry data to assess expansion and contraction of T cell memory 3 weeks for up to 8 cycles. The primary endpoint was confirmed objective response rate
subsets throughout the course of therapy. Whole genome enzymatic methyl sequencing was (ORR) by blinded independent central review (BICR) per Response Evaluation Criteria in Solid
performed on post-treatment PD-1 High and PD-1 Low CD8 T cells to determine the Tumors (RECIST) version 1.1. The secondary endpoints included PFS, overall survival (OS),
multipotency of the patient T cell compartment after blinatumomab treatment. disease control rate (DCR), and safety. Results: Between June 2, 2021, and January 10,
Results: After 7 days of continued blinatumomab infusion, patient T cells demonstrated a 2025, a total of 48 eligible subjects were enrolled in the study. In the overall population, the
significant expansion of terminally differentiated and effector memory T cells. Notably, we median age was 60 (range: 29 to 72) years, with 31 males (64.6%) and 17 females (35.4%). At
observed that non-responders had a high tumor burden at the start of the therapy and the data cutoff on January 10, 2025, the median follow-up was 27.1 months (95% CI, 20.2 to
possessed a large population of naı̈ve CD8 T cells that failed to expand. These CD8 T cells 37.2) and 3 (6.3%) cases continued treatment. The ORR was 54.2% (95% CI, 40.3 to 67.4),
exhibited a significant increase in expression of TIM-3 and PD-1 compared to responders and the DCR was 95.8% (95% CI, 86.0 to 98.9), as assessed by BICR per RECIST version 1.1.
after the 7-day infusion. Methylation analysis of post-treatment CD8 T cells showed de- The median PFS was 16.7 months (95% CI, 12.6 to not available (NA)), with the 12- and 24-
creased methylation of exhaustion regulators IKZF1 and CD300a in non-responders month PFS rates at 68.6% and 38.5%, respectively. The median OS was 24.6 months (95% CI,
compared to the responders. Additionally, in vitro treatment of T cells with blinatumo- 14.6 to 29.5), and the 12- and 24-month OS rates were 72.5% and 53.0%, respectively. The
mab induced T-cell exhaustion in a target-dependent manner. Conclusions: Blinatumomab median duration of response (DoR) was 22.5 months (95% CI, 12.5 to NA), with the 12- and
therapy in pediatric B-ALL patients induced variable epigenetic and phenotypic changes to 24-month DoR rates at 78.9% and 47.5%, respectively. Treatment-related adverse events
the T cell compartment indicative of exhaustion, corresponding to differences in T cell (TRAEs) of any grade were reported by 46 (95.8%) patients. Hematologic toxicity (43, 89.6%)
expansion and persistence between patients. Our study is the first to link epigenetic changes and liver injury (25, 52.1%) of any grade were the most frequently reported TRAEs, and grade
in exhaustion regulators with response variability in blinatumomab-treated patients. Fur- 3-4 TRAEs were observed in 25 (52.1%) patients. Grade 3-4 immune-related adverse events
thermore, our findings highlight a potential role of baseline T cell composition and tumor (irAEs) occurred in 8 (16.7%) participants, with pneumonitis (2, 4.2%) and endocrine dis-
burden in determining therapeutic outcomes. These insights provide a novel framework for orders (2, 4.2%) being the most common. Conclusions: Second-line PD-1 blockade in
improving patient stratification and treatment strategies to mitigate T cell exhaustion in combination with bevacizumab and nab-paclitaxel is an effective and well-tolerated
blinatumomab therapy. Research Sponsor: American Lebanese Syrian Associated Charities treatment regimen for patients with CUP. Clinical trial information: NCT04848597.
(ALSAC), St. Jude Children’s Research Hospital. Research Sponsor: Clinical Research Plan of SHDC.

2530 Poster Session 2531 Poster Session

Comprehensive analysis of NSAIDs use and oncological outcomes in non- Preliminary monotherapy efficacy of novel immune checkpoint blockade
small cell lung cancer patients treated with immune checkpoint inhibitors. GV20-0251 (anti-IGSF8) in advanced melanoma patients with primary re-
First Author: Yanlin Li, Second Affiliated Hospital of Xi’an Jiaotong University, Xi’an, sistance to anti-PD1. First Author: Kristopher Wentzel, The Angeles Clinic and
China Research Institute, A Cedars-Sinai Affiliate, Los Angeles, CA
Background: The effect of non-steroidal anti-inflammatory drugs (NSAIDs) on immune Background: GV20-0251 is an AI-designed, first-in-class, cross-species reactive, Fc-
checkpoint inhibitors (ICIs) efficacy in non-small cell lung cancer (NSCLC) remains contro- attenuated IgG1 antibody that targets the novel cancer immune checkpoint IGSF8
versial. Although the COX-2/PGE2 pathway, a primary target of NSAIDs, has been implicated in which is broadly expressed across solid tumors. In syngeneic tumor models, anti-IGSF8
diminished immunotherapy response, direct clinical association with NSAIDs and ICIs in real alone or with anti-PD1 inhibits tumor growth by increasing cytotoxicity and infiltration of
world has yet to be established. This study aims to evaluate the impact of NSAIDs use— natural killer cells (NK) and antigen cross-priming by dendritic cells which in turn activates
considering types, duration, and timing—on ICI efficacy, alongside its effects on PGE2 and T cells. Methods: The phase 1 portion of this first-in-human, phase I/IIa study
immune cell profiles. Methods: We included stage III-IV NSCLC patients receiving PD-1/PD-L1 (NCT05669430) was conducted across multiple U.S. centers. The study utilized a standard
antibodies in 5 centers. Blood and tumor samples were collected in perspective cohort. NSAIDs 3+3 design to evaluate the safety, pharmacokinetics (PK), pharmacodynamics, immu-
were categorized based on selectivity (non-selective COX inhibitors, selective COX-2 inhibitors) nogenicity, and preliminary efficacy of GV20-0251, and to establish a preliminary rec-
and chemical structure (salicylates, propionate derivatives, others). PGE2 and cytokines were ommended phase 2 dose (RP2D). Results: Forty-two patients with advanced solid tumors
measured in blood by ELISA. RNA sequencing data were obtained from databases. Tumor (median age 61 years, median 4 prior treatment lines) were enrolled across six dose levels
tissues were collected for immunohistochemical staining of immune cells. Multivariate Cox
(0.5, 1, 3, 6, 10, and 20 mg/kg) and two schedules (D1/D8 Q3W and D1 Q3W). GV20-0251
and logistic regression were used in analyses of progression-free survival (PFS) and objective
demonstrated favorable safety and tolerability across all doses and schedules with no
response rate (ORR). Results: 883 patients were included, with 140 NSAIDs users and 743
dose-limiting toxicities, and 10 and 20 mg/kg D1 Q3W were selected as the preliminary
non-users. 196 patients were enrolled prospectively with samples. Multivariate analysis
showed that NSAIDs use was significantly associated with improved PFS (HR 0.67, 95% CI RP2D. Treatment-related adverse events occurred in 55% of patients, predominately grade
0.51-0.88, P = 0.005) and ORR (OR 1.87, 95% CI 1.29-2.72, P = 0.001). Subgroup analyses 1/2, with a single grade 3 event of pneumonitis. The most common treatment-related AEs
indicated that non-selective COX inhibitors, salicylates, long-term use, and pre-ICI initiation were fatigue and rash (12% each), with no dose-dependent trends. Full target occupancy
were correlated with better outcomes. In contrast, selective COX-2 inhibitors, propionate and half-life of 26 days with linear PK were observed at $10 mg/kg, without significant
derivatives, others, short-term use, and post-ICI initiation showed no effect on PFS or ORR. serum cytokine elevation or anti-drug antibody signals. Among 38 efficacy-evaluable
Blood analyses indicated that NSAIDs significantly lowered PGE2 levels, particularly salicy- patients, 17 had cutaneous melanoma, all of whom progressed on prior anti-PD1 therapy
lates and long-term use. Higher PGE2 was associated with worse outcomes. For immune cells, and 16 progressed on prior anti-CTLA4 therapy. Among the 9 melanoma patients with
RNA sequencing revealed that COX-2 and mPGES-1 were significantly correlated with neu- primary resistance to anti-PD1, confirmed partial response (PR) was achieved in 3 (33%)
trophil enrichment and neutrophil-related cytokines. Single-cell RNA-seq showed high ex- patients and tumor shrinkage was observed in an additional 3 patients. Notably, responses
pression of COX-2 and mPGES-1 in neutrophils. Analysis of samples confirmed that NSAIDs were observed in 2 patients with liver metastases, which are typically refractory to im-
use was associated with reduced neutrophils and neutrophil-related cytokines in blood and munotherapy. Although no responses were seen in the melanoma patients with acquired
less neutrophil infiltration in tumor. Conclusions: NSAID use is an independent predictor of resistance to anti-PD1 (n = 8) or in patients with other tumor types (n = 21), potentially due
improved PFS and ORR in NSCLC patients receiving ICIs. Specifically, non-selective COX to the lower frequency of IGSF8 protein expression in these tumors, tumor shrinkage was
inhibitors, salicylates, long-term use, and pre-ICI initiation are associated with better clinical observed in one non-small cell lung cancer (n = 4) and one cervical cancer (n = 1) patient.
outcomes. NSAID use may enhance ICIs efficacy by reducing serum PGE2, which could serve Preliminary immunohistochemistry analyses of trial patient biopsies suggest IGSF8 high
as a predictive biomarker. Furthermore, NSAIDs decrease neutrophils in both blood and tumor, tumors have low anti-PDL1 at baseline, and GV20-0251 treatment increases tumor-
potentially contributing to the improvement in ICI efficacy. Research Sponsor: Basic Research infiltrating NK and T cells. Conclusions: GV20-0251 demonstrated a favorable safety
Funds for Central Universities; National Natural Science Foundation of China; Shaanxi Province profile in heavily pretreated patients with advanced solid tumors and showed promising
"Sanqin Scholars" Innovation Team Support Program; Shaanxi Province Health and Medical monotherapy efficacy in cutaneous melanoma patients with primary resistance to anti-
Research Innovation Team Support Program. PD1. Clinical trial information: NCT05669430. Research Sponsor: GV20 Therapeutics.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
146s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2532 Poster Session 2533 Poster Session

Prophylactic infusion of allogeneic double-negative T cells as immune Targeting cancer leptomeningeal metastasis with allogeneic chimeric anti-
modulators to prevent relapse in high-risk AML patients after allo-HSCT: gen receptor gd T-cell therapy. First Author: Peiwen Ma, National GCP Center for
A phase I trial. First Author: Xiaoyu Zhu, The First Affiliated Hospital of University of Anticancer Drugs, National Cancer Center/Cancer Hospital, Chinese Academy of Medical
Science and Technology of China, Hefei, China Sciences & Peking Union Medical College, Beijing, China
Background: Our previous study demonstrated that double-negative T cells (DNTs) hold potential Background: Leptomeningeal metastasis (LM) occurs in 1–10% of patients with advanced solid
for treating relapsed or refractory acute myeloid leukemia (r/r AML) following allogeneic he- tumors during disease progression. LM significantly worsens prognosis due to the rapid onset and
matopoietic stem cell transplantation (allo-HSCT). In a first-in-human Phase I trial (ChiCTR-IPR- progression of symptoms associated with elevated intracranial pressure. Currently, no therapies
1900022795), we reported a complete response (CR) rate of 50% (5/10) with a favorable safety specifically targeting LM have been approved. Here, we report our clinical observations from two
profile. This Phase I/II study aims to evaluate the safety and efficacy of off-the-shelf allo-DNTs in patients treated with intrathecal infusion of allogeneic B7H3-targeted CAR-gdT cells(QH104).
preventing relapse in AML patients following allo-HSCT. Methods: Six high-risk AML patients Methods: This is an open-label, single-arm clinical study designed to evaluate the safety and efficacy
undergoing allo-HSCT were enrolled and assigned to two dosage groups: 1310^8 DNTs/kg and of QH104 in patients with LM originating from B7H3-positive solid tumors(NCT06592092). Eligibility
1.5310^8 DNTs/kg. Each patient received three infusions at one-month intervals without prior criteria included a diagnosis of LM from any B7H3-positive solid malignancy. QH104 was administered
lymphodepleting chemotherapy. The median time from transplantation to the first infusion was as a single dose of 3310^7 cells via lumbar puncture or Ommaya reservoir infusion. Treatment-
3.1 months. Primary endpoint was the occurrence of adverse events and dose-limiting toxicities, emergent adverse events were graded using CTCAE v5.0 and ASTCT criteria. Efficacy was assessed
while the secondary endpoint was cumulative incidence of relapse (CIR). GMP-grade DNTs were using the RANO-LM criteria. Results: As of January 2025, two female lung adenocarcinoma patients
expanded ex vivo from healthy donor PBMCs and cryopreserved in liquid nitrogen until infusion. were enrolled. They had previously received treatments targeting LM, including intrathecal chemo-
Results: As of January 20, 2025, with a median follow-up of 17.55 months post-HSCT, four of six therapy and oral EGFR tyrosine kinase inhibitors. No adverse events higher than grade3 were reported.
patients (66.7%) remained in MRD-negative CR, with the longest recurrence-free survival exceeding One patient experienced a transient episode of absence seizures on Day 1 after cell infusion, which was
17 months. The two relapsed patients both carried high-risk genetic mutations (TP53 mutation) considered treatment-related. At the Day 30 assessment post-infusion, both patients had stable
and were MRD-positive prior to transplantation. They succumbed at 11.4 and 14.2 months post- disease, with a reduction or complete elimination of tumor cells in the CSF and improvement in clinical
symptoms associated with LM. CSF component analysis demonstrated the persistence of CAR-gd
HSCT respectively. Donor-derived DNTs were detectable in peripheral blood shortly after each
T cells for one week post-infusion. CSF cytokine analysis revealed increased levels of interleukin-5, -6,
infusion, peaking at 1–4 days and persisting for up to 28 days. In two patients with MRD-negative
-9, -13, and -22, TNF-a and IFN-g post-infusion compared to baseline. No significant increases in CAR-
CR, infused DNTs remained detectable for up to 360 days post-infusion. Elevated levels of IFN-g, IL-
gd T cells or cytokines were detected in peripheral blood. Conclusions: Our initial clinical experience
6, and IL-10 post-infusion indicated immune activation. Importantly, no dose-limiting toxicities,
with the first two patients with leptomeningeal metastasis (LM) provides preliminary evidence
neurotoxicity, cytokine release syndrome greater than Grade 2, or graft-versus-host disease were supporting the safety and efficacy of B7H3-targeted CAR-gdT cell immunotherapy in this patient
observed. In contrast to the two relapsed patients, MRD-negative CR patients showed expanded group. A longer follow-up period and a larger patient cohort are necessary for a comprehensive
levels of CD4+, CD8+, and DNT cells, particularly those with the effector memory T cell phenotype. evaluation of therapeutic efficacy and response durability. Clinical trial information: NCT06592092.
Both the infused DNTs and the recipient’s CD4+ and CD8+ T cells in these patients secreted higher Research Sponsor: National Natural Science Foundation of China; 82272953; The National Key
levels of granzymes A and K. To investigate the interaction between CD8+ T cells and allo-DNTs in Research and Development Program of China.
MRD-negative CR patients, co-culture experiments were conducted. CD8+ T cells exhibited an
Patients’ baseline character, administration routes and treatment evaluation.
increase in the secretion of granzyme B and IFN-g within 3–4 days. Transcriptome sequencing and
B7H3+
multi-cytokine analyses revealed strong immune activation. Conclusions: The dual ability of DNTs CAR -gdT
to suppress GvHD while preserving the graft-versus-leukemia effect, along with its potential for off- B7H3 cells Administration Treatment
Diagnosis Sex Age Genetic mutation Score infused route response at Day 30
the-shelf availability, makes it a transformative therapy in the post-transplant setting. Clinical trial
Patient Lung F 53 EGFR 21 exon 70 33107 Lumbar Stable disease
information: NCT05858814. Research Sponsor: National Natural Science Foundation of China; # 01 adenocarcinoma L858R mutation puncture (CSF cytology: remain positive
U23A20453, 82270223 and 82170209; Anhui Provincial Key Research and Development Project; # with LM CNS imaging: Stable
Symptoms assessment score:6 to 4)
2022e07020015; Anhui Health Research Project; # AHWJ2022a011; Anhui Provincial Department Patient Lung F 58 EGFR 21 exon 40 33107 Ommaya Stable disease
of Education Scientific Research Project; 2023AH010079; Anhui Provincial Natural Science 02 adenocarcinoma L858R mutation reservoir (CSF cytology: turned negative
with LM CNS imaging: Stable
Foundation; 2308085J09; the Fundamental Research Fund for the Central Universities; Symptoms assessment score:4 to 2)
YD9110002047.

2534 Poster Session 2535 Poster Session

Safety and efficacy of non-viral aPD1-MSLN JL-lightning–CAR-T in ad- Phase I trial of personalized AI-identified TCR-transduced T cell therapy in
vanced malignant mesothelioma in a phase I trial. First Author: Yan Sun, advanced solid tumors. First Author: Shuhang Wang, Department of Clinical Trial
Shanghai Cell Therapy Group Co., Ltd, Shanghai, Shanghai, China Center, National Cancer Center/National Clinical Research Center for Cancer/Cancer
Background: CAR-T cells face challenges in solid tumors, including weak in vivo pro- Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College,
liferation, immunosuppressive tumor microenvironments (TME), and limited tumor infil- Beijing, China
tration. We firstly developed an innovative non-viral JL-Lightning-CAR-T fast process to Background: TCR-T cell therapy shows promise in treating solid tumors but is limited by the
enhance CAR-T stemness, in vivo expansion, and persistence. The autologous non-viral need for personalized TCR identification. We developed TCR-XFinder, a deep learning model
aPD1-MSLN JL-Lightning-CAR-T cells were manufactured in just 30 hours, targeting using a 3-stage transfer-learning strategy, to rapidly identify personalized tumor-reactive TCRs
mesothelin (MSLN) and secreting anti-PD-1 antibodies to counteract the immunosup- within 10 days after tumor tissue acquisition. This study reports the first-in-human phase I trial
pressive TME and improve the efficacy of solid tumor treatment. Here, we report the safety of KSX01, a TCR-transduced T cell therapy identified by TCR-XFinder. Methods: We
and preliminary efficacy of this novel CAR-T therapy in advanced malignant pleural conducted a phase I, dose-escalation study (NCT06150365) to evaluate the safety and efficacy
mesothelioma (MPM) in a first-in-human phase I pilot study ([Link]: of KSX01 in patients with advanced solid tumors. Tumor tissues were subjected to single-cell
NCT06249256). Methods: A single-arm, open-label, dose-escalation study was designed RNA sequencing and TCR sequencing to identify tumor-reactive TCRs using TCR-XFinder.
and enrolled MPM patients who had failed standard therapies and had confirmed MSLN These TCRs were validated and transduced into autologous T cells, which were expanded and
and PD-L1 expression on tumors by IHC. Patients received a single dose of non-viral aPD1- infused back into patients. Patients received preconditioning with cyclophosphamide (500 mg/
MSLN JL-Lightning-CAR-T cells following lymphodepletion (Flu 30 mg/m²/day, Cy 300 mg/ m²/day) and fludarabine (30 mg/m²/day) for 3 days, followed by intravenous infusion of KSX01
m²/day) for 2-3 days. The dose escalation was designed as DL1 (0.5-0.6310⁶/kg) and DL2 TCR-T cells at two dose levels (5310⁹ 6 30% and 1310¹⁰ 6 30% cells). Safety was assessed
(0.8-1.0310⁶/kg). Adverse events were evaluated using CTCAE v5.0, and clinical re- by monitoring adverse events and cytokine release syndrome (CRS). Efficacy was evaluated by
RECIST v1.1 criteria. Results: Four patients with advanced solid tumors (alveolar soft part
sponses were assessed by mRECIST 1.1 or RECIST 1.1. CAR expression was analyzed by
sarcoma, epithelioid sarcoma, colon cancer, and clear cell renal cell carcinoma) were enrolled.
qPCR, and anti-PD-1 antibodies were detected by MSD. Results: Patients: Seven ad-
KSX01 TCR-T cells were well-tolerated at both dose levels, with no dose-limiting toxicities
vanced MPM patients were enrolled and received single dose CAR-T cell infusion. Efficacy: (DLTs) observed. All patients experienced Grade 3–4 pancytopenia, which was expected
In DL1 (0.5-0.6310⁶/kg), one patient achieved partial response (PR) with a disease control following lymphodepletion. One patient developed Grade 2 CRS, resolved with tocilizumab. No
rate (DCR) of 75% (3/4). In DL2 (0.8-1.0310⁶/kg), all of three patients achieved objective Grade 3 or higher AEs related to KSX01 were noted. At the first tumor assessment (Day 28), all
response (ORR 100%, 3/3), with one patient achieving complete response (CR) at 3 months patients showed disease control, with one patient achieving a partial response (PR) and a 46%
and maintaining it for over 9 months. Pharmacokinetics: Anticipated CAR-T cell expansion reduction in target lesion size. Another patient achieved PR with second infusion at higher
and anti-PD-1 antibodies increase detected in circulation. CAR-T Cmax reached up to dose. qPCR analyses confirmed the infiltration and long-term anti-tumor effect of infused TCR-
47,307 copies/mg, detectable for over 3 months. Anti-PD-1 antibody Cmax reached up to T cells. A transient post-infusion increase in interferon-g (IFN-g), interleukin-6(IL-6), IL-10, IL-
376,938 pg/ml, detectable for over 6 months. Tmax for MSLN-CAR-T and anti-PD1 4, tumor necrosis factor-a (TNF-a), and CRP levels was observed in all patients. While the Cmax
nanobody occurred between Day 7 and Day 14 post infusion. IFN-g and IL-6 levels and Tmax values varied among cytokines, the first Tmax for 80% of cytokines and CRP occurred
also increased during this period. Safety: In DL1, CRS was observed in 1 of 4 patients within the first week post-infusion. Re-biopsy of tumor lesions showed infiltration of infused
(Grade 1), with no ICANS or DLT. In DL2, CRS was observed in 2 of 3 patients (Grade 3-4), TCR-T with persistent cytotoxic function and ameliorate the microenvironment for the en-
with no ICANS. Grade 3 immune-mediated pneumonia occurred in 2 of 3 patients in DL2, dogenous tumor-reactive T cells. Conclusions: The first-in-human phase I trial of KSX01 TCR-
managed by clinical intervention strategies. All patients experienced Grade 3-4 hema- T cell therapy demonstrated promising safety and efficacy in patients with advanced solid
tologic toxicity, reversible with supportive care. Conclusions: Non-viral aPD1-MSLN JL- tumors. TCR-XFinder enabled rapid identification of personalized tumor-reactive TCRs,
Lightning-CAR-T cells demonstrated robust proliferative capacity, manageable safety supporting the clinical feasibility of this approach. Further studies are warranted to explore
profile, and significant anti-tumor potential, offering a promising therapeutic approach for optimal dosing and combination strategies to maximize clinical benefit. Clinical trial infor-
advanced MPM patients. Clinical trial information: NCT06249256. Research Sponsor: mation: NCT06150365. Research Sponsor: National Natural Science Foundation of China;
None. 82272953; The National Key Research and Development Program of China.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 147s

2536 Poster Session 2537 Poster Session

Artificial intelligence (XGBoost) in predicting outcomes among CAR-T ther- A novel cellular immunotherapy using vaccine generated neoantigen-
apy patients: The impact of malnutrition and comorbidities using the Na- specific effector T cells. First Author: Andrew Edward Sloan, Piedmont Health-
tional Inpatient Sample (2020-2022). First Author: Tong Ren, University of South care and Case Western Reserve University School of Medicine, Atlanta, GA
Florida (USF) Morsani College of Medicine/HCA Florida Oak Hill Hospital, Brooksville, FL Background: Cellular immunotherapy languished in obscurity until genetically engi-
Background: Chimeric Antigen Receptor T-cell (CAR-T) therapy has revolutionized neered chimeric antigen receptor T cells were shown to effectively treat B lymphocyte
hematologic malignancy treatment but remains costly, with limited access and com- cancers. Genetic studies also revealed that some cancer cell mutations produce
plications like prolonged hospitalization, sepsis, and mortality. Malnutrition, common in neoantigens, which are consistent with historical studies demonstrating that cancer cell
cancer patients, worsens these outcomes. Despite AI’s growing role in oncology, its use vaccination generates neoantigen specific immune responses in rodents and humans.
in risk stratification for malnourished CAR-T recipients is underexplored. This study Vaccination leads to an increase in neoantigen-specific T cells in lymphoid tissue that
leverages the National Inpatient Sample (NIS) 2020-2022 to develop AI-driven models are released into the blood, which carries them to sites of disease activity, e.g., cancer
predicting length of stay (LOS), mortality, and sepsis, incorporating the Charlson tissue. TVAX Biomedical hypothesized that the natural power of the patient’s immune
Comorbidity Index and other factors. Methods: Using the NIS database, adult CAR-T system could be exploited using a novel neoantigen-specific cellular immunotherapy.
therapy patients were identified with ICD-10 codes. Key variables included demo- Methods: Patients are vaccinated with their own attenuated cancer cells plus an
graphics (age, gender, race/ethnicity, income), clinical factors (Charlson Comorbidity immunologic adjuvant, e.g. GM-CSF, to increase the number of circulating neoantigen
Index, sepsis, admission type), and hospital characteristics (size, teaching status). AI primed T cells. Patients are leukapheresed to collect the T cells. The collected T cells are
models (XGBoost, Random Forest, Neural networks) were trained on the 2020 dataset exposed to activation and proliferation stimulating agents to generate the neoantigen-
and validated on 2020-2022 data. Hyperparameter tuning via grid search was performed specific effector T cells that are used for treatment. TVAX is currently testing this
to optimize model performance. LOS was modeled as a continuous outcome, while treatment paradigm for efficacy and safety in newly diagnosed (MGMT-negative)
mortality and sepsis were classified as binary outcomes. Data preprocessing included glioblastoma patients when they have minimized immunosuppression and minimal
handling missing values, one-hot encoding of categorical variables, and standardizing residual disease, TVI-AST-008. Results: For this novel cellular immunotherapy to be
continuous variables. SHapley Additive exPlanations (SHAP) were used to interpret effective, T cell mediated immune responses must be generated in vaccinated patients.
feature importance. Results: The study analyzed 1,912 CAR-T hospitalizations over Delayed type hypersensitivity skin testing, a method for detecting T cell mediated
three years, with 11.5% identified as malnourished. AI models demonstrated strong immunity in humans, showed reactions in patients with leukemia, brain, breast, colon,
predictive performance, with XGBoost (RMSE: 3.5 days, R² = 0.82) for LOS, Random lung, kidney, melanoma, ovarian, prostate and sarcoma (data to be presented). Multiple
Forest (AUC: 0.91) for mortality, and Neural Networks (AUC: 0.87) for sepsis. Malnu- autologous vaccinations led to detectable responses in all patients. The combination of
trition significantly worsened outcomes, increasing LOS by 14.2 days (p , 0.001) and cancer cell/immunologic adjuvant vaccination plus neoantigen-specific T cellular im-
mortality risk by 3.2-fold (p , 0.001). Patients with Charlson Comorbidity Index munotherapy has been shown to be highly effective against a wide range of cancer types
scores $3 had 9.8-day longer LOS and 2.9-fold higher mortality risk (p , 0.001). Racial in preclinical studies and to be effective against the least immunogenic cancers.
disparities were evident, with Black patients at 25% higher risk of prolonged LOS and Conclusions: The possibility that neoantigen-specific T cells could effectively treat
Hispanic patients at increased risk of sepsis (p , 0.05). Malnourished patients in non- some human cancers has been documented through studies with tumor
teaching hospitals with high comorbidity burdens had the worst outcomes, emphasizing infiltrating lymphocytes (TILs). However, TIL efficacy is limited to a small number of
the need for targeted interventions in high-risk populations. Conclusions: AI-driven (hot) cancers. Preclinical model studies demonstrated that neoantigen-specific effector
models incorporating malnutrition and Charlson Comorbidity Index accurately predict T cells enter cancer tissue, initiating a cascade of T cell mediated immunologic events
LOS, mortality, and sepsis in CAR-T patients. Early identification and management of that ultimately leads to killing of cancer cells by cytotoxic T cells and cytokine activated
malnutrition and comorbidities, particularly in racially diverse populations, are critical to accessory cells. The benefit of the vaccine enhanced neoantigen-specific effector T cell
improving outcomes. Future research should focus on prospective validation and AI therapy (TVAX Immunotherapy) is that it expands the range of human cancers that could
integration into clinical workflows to mitigate disparities. Research Sponsor: None. be safely and effectively treated. Clinical trial information: 05685004. Research Sponsor:
NIH Grant, Office of Orphan Products.

2538 Poster Session 2539 Poster Session

Engineering iPSC-derived mesenchymal stem cells (i MSCs) to secrete IL-7/ Universal solid tumor therapy with CD5-deleted, DSG2-directed CAR-T cells.
IL-15 for modulation of the tumor microenvironment in a "cold" ovarian First Author: Robert D. Carlson, Thomas Jefferson Universiy, Philadelphia, PA
tumor model. First Author: Sandeep Singh, University of Texas MD Anderson Cancer Background: CAR-T cell therapy has been curative for many patients with refractory,
Center, Houston, TX progressive hematologic cancers, resulting in several FDA approvals. However, this
Background: We previously discovered that bone marrow derived Mesenchymal stromal therapy has not been successful for solid cancers, reflecting the need for suitable
cells (BM-MSCs) migrate to the stroma of numerous cancers and their metastases, antigen targets for each disease and solutions to immunological barriers in solid tumors.
forming tumor-associated fibroblasts (TAFs) and can be modified to secrete proteins Here, we have identified the desmosomal cadherin, desmoglein 2 (DSG2), as an effective
within the tumor microenvironment (TME). MSCs have not been utilized extensively in CAR-T cell therapy target in epithelia-derived solid tumors. DSG2 contributes to cell
cancer therapy due to their immunosuppressive properties, limited replicative capacity, proliferation, migration, and other emerging tumor-promoting pathways, resulting in its
and variable quality depending on the source. Methods: Here, we report the charac- upregulation in nearly all solid cancers and correlating with poor prognosis. Moreover,
terization of induced mesenchymal stromal cells (iMSCs) derived from pluripotent stem we explored CRISPR-Cas9-mediated elimination of the inhibitory receptor CD5 to en-
cells (iPSCs), which were uniquely generated from adult skin fibroblasts using a transient hance in vivo CAR-T cell expansion and solid tumor efficacy. Methods: DSG2-directed
mRNA transfection technique. Notably, iMSCs demonstrated superior proliferative ca- CAR-T cells were generated from human T cells using a scFv derived from a murine
pacity under both normoxic and hypoxic conditions, while preserving their trilineage hybridoma targeting the extracellular domain of DSG2 in a 3rd generation CAR design
differentiation potential. Comprehensive molecular profiling, including RNA sequencing, with CD28, 4-1BB, and CD3z signaling domains. CD5 elimination employed electro-
single-cell mass cytometry (CyTOF), and Luminex assays, revealed strong phenotypic and poration of complexed gRNA-Cas9 ribonucleoprotein (RNP). DSG2 expression was
functional similarities between iMSCs and BM-MSCs. Crucially, no evidence of sarcoma characterized in human cancers and cell lines and CAR-T cell activity was examined
formation was observed in NSGS mice following intraperitoneal, subcutaneous, or in- in vitro by cytokine production and target cell cytolysis. In vivo efficacy studies
travenous administration of iMSCs, highlighting their robust safety profile. We employed cancer xenografts in NSG mice treated with CAR-T cells. Safety studies
engineered a DNA cassette into these cells to enable constitutive superphysiological employed a human DSG2 transgenic mouse treated with syngeneic murine CAR-T cells
expression of interleukin(IL)-7 and IL-15, expressed as either individual molecules (P2A) for clinical, serum biomarkers, and histopathological evaluation. Results: In vitro
or a single fused molecule (FUS). Both, P2A and FUS iMSCs demonstrate the capacity to studies revealed recognition and lysis of solid cancer cell lines and effector cytokine
drive T cell proliferation autonomously in co-culture experiments. Results: IL7/IL15- production. Administration of DSG2-directed CAR-T cells eliminated metastatic cell-
modified iMSCs induced tumor cell death in a triple co-culture system comprising iMSCs, derived xenografts, patient-derived xenografts, and orthotopic tumors derived from
the ovarian cancer cell line ID8, and human PBMCs. In a syngeneic mouse model of various solid cancers, including colorectal, pancreatic, lung, prostate, breast, and liver.
ovarian cancer (ID8 cells in C57BL/6 mice), intraperitoneal administration of P2A or FUS- Moreover, elimination of CD5 enhanced the expansion of DSG2-directed CAR-T cells
iMSCs resulted in reduced tumor burden and extended survival. Immunohistochemical in vivo, resulting in curative efficacy at sub-therapeutic CAR-T cell doses. Safety studies
and flow cytometric analyses revealed massive infiltration of activated T cells, mac- revealed no toxicity in any human DSG2 transgenic mouse tissues. Conclusions: These
rophages, and other immune cells into the tumor microenvironment (TME) in both FUS or studies reveal the robust antitumor activity of DSG2-directed CAR-T cells in solid tumors,
P2A groups, but not in unmodified iMSC controls, or in PBS injected animals. The TME in which is enhanced by CD5 deletion, without toxicity in a human transgenic mouse model.
P2A- and FUS-treated mice showed enrichment in tumoricidal M1-type macrophages, Thus, CD5-deleted DSG2-directed CAR-T cells are a promising therapeutic approach that
with no detection of exhausted or regulatory T cells, in contrast to controls. may be safe and effective for all solid cancers. Research Sponsor: Kleberg Foundation;
Conclusions: IL7-IL15-secreting iMSCs migrate into solid tumors, induce massive im- U.S. Department of Defense; W81XWH-19-1-0263; U.S. Department of Defense;
mune cell infiltration into the TME and enhance antitumor immunity in a syngeneic mouse W81XWH-22-1-0207; DeGregorio Family Foundation; U.S. National Institutes of Health;
model of cancer. These cytokine-producing iMSCs represent a potentially promising 1R21 CA267087; U.S. National Institutes of Health; 1R21 CA286339; The Courtney Ann
anticancer immunotherapy by converting “cold” into “hot” tumor microenvironments. Diacont Memorial Foundation and Lorraine and David Swoyer; U.S. National Institutes of
Research Sponsor: Eterna Therapeutics, Inc. Health; T32 GM008562; U.S. National Institutes of Health; T32 CA236736.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
148s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2540 Poster Session 2541 Poster Session

Mayhem under the microscope: T cell cytotoxicity and serial killers captured Impact of stromal-targeting antitumor CAR T cells in solid tumors. First
in situ. First Author: Greg Sawyer, Moffitt Cancer Center, Tampa, FL Author: Abdul Khan, Roswell Park Cancer Institute, Buffalo, NY
Background: Developing a deeper understanding of the dynamics of immune cell- Background: While chimeric antigen receptor (CAR) T cells have shown tremendous
mediated cytotoxicity is critical to advancing immunotherapy and cell therapy. The success in hematological malignancies, but such efficacy has not been achieved in the
results from the multidisciplinary effort reported here include numerous measurements setting of solid tumors. One of the hurdles to CAR T cells therapy in solid tumors is the
and movies of immune cell-mediated cytotoxicity with striking examples of serial killing, presence of physical stroma and cancer associated fibroblasts (CAFs), which inhibit the
foraging, path-tracking, triple killing events, measurements of cytokine gradients at entry of activated T cells to the tumor sites. Membrane bound protein Leucine-rich repeat
tumor margins, and other dynamics. Some cytotoxic events revealed peak apoptotic containing 15 (LRRC15) has been shown to be highly expressed on CAFs in many solid
signatures just minutes after T Cell engagement. Methods: In vitro studies of immune tumors including pancreatic cancer as well as directly expressed on tumors of mes-
cell killing are traditionally performed using time-lapse imaging and biochemical assays, enchymal origin including sarcomas, glioblastomas and melanomas. LRRC15 has very
but these methods are often limited by spatial and temporal resolution, throughput, and limited expression in normal tissues. The goal of the current study was to explore the
the ability to extract the dynamics of cellular interactions. This study integrates high- impact of stromal targeting antitumor (STAT) in solid tumors using LRRC15-directed CAR
resolution and high-speed laser scanning confocal microscopy with artificial intelligence T cells. Herein we demonstrate that CAR T cells directed to tumor stroma can eradicate
(AI), and machine learning (ML) approaches to provide a high-resolution data-driven solid tumor. Methods: LRRC15-directed CAR T cells were validated in in vitro assays that
analysis of immune cell killing dynamics in vitro. In these studies, we use human CAR included specific lysis, cytokine secretion, and proliferation. STAT CAR T cells were
T cells with an anti-cd19.28z and Burkitt Lymphoma. Results: We have engineered a administered intravenously into NSG mice after engrafted with osteosarcoma SaOS2 and
perfusion-enabled 3D culture system integrated microscopy to assess cellular dynamics pancreatic PANC1 tumor cell lines as well as patient derived tissues (PDXs). The efficacy
for extended periods of time. Perfusion culture maintains the interstitial flow of liquid of STAT CAR T cells was also assessed in syngeneic mouse model engrafted with murine
culture media, clearing the microenvironment of toxic metabolites and reactive oxygen OS F420 and pancreatic KPC tumor cell lines. Tumor was harvested from mice at various
species. This platform uses a Liquid-Like Solids (LLS) to mimic the transport dynamics timepoints and analyzed for LRRC15 expression as well as for the presence of T cells.
of a capillary bed. Integrated microscopy allows in situ quantification of spatiotemporal Results: LRRC15-directed CAR T cells specifically lysed SaOS2 cells. Upon stimulation
cytokine concentrations, immune cell tracking, immune cell killing dynamics, and in- with SaOS2 cell line, LRRC15-directed CAR T cells resulted in robust expansion and
vasion dynamics. Cytokine on and off-rates were referenced alongside measured bead secreted cytokines including IL2, IFN-Y, GM-CSF and TNFa. The LRRC15-directed CAR
fluorescence intensities and positions to fit spatiotemporal reaction-diffusion models T cells were able to eliminate tumor in NSG mice xenografted with SaOS2 cell line as well
out to a 1,600 um radius. Fast-scanning confocal microscopy facilitated in-situ ob- as OS PDX. LRRC15-directed CAR T cells significantly increased the survival of mice. Next,
servation of the evolutionary dynamics of tumor progression. In-situ cytokine mea- we sought to test the efficacy of LRRC15-directed CAR T cells in NSG mice engrafted with
surements revealed local IL-8 concentrations reached a maximum value of 2 ng ml-1 LRRC15- tumor/CAFs+ cell line and PDX. We showed that the NSG mice injected with the
after 10 hours. A cellular production rate was estimated at 2 molecules cell-1 s-1. LRRC15- PANC1 cell line acquired stroma with CAFs positive for LRRC15 within 2-
Conclusions: T Cell cytotoxicity is shown to be incredibly heterogeneous spanning from 3 weeks. LRRC15-directed CAR T cells were able to significantly inhibit the progression of
minutes to hours. The platform developed in this study demonstrates a powerful method both PANC1 tumors as well as PDAC PDX in NSG mice. In syngeneic mouse model,
for real-time, high-resolution imaging of cancer-immune interactions within a controlled LRRC15-directed CAR T cells resulted in tumor regression of both LRRC15+ sarcoma F420
3D environment. By leveraging in situ fast-scanning fluorescence microscopy, the and LRRC15- pancreatic KPC tumors. Conclusions: To our knowledge this is the first
platform enables precise quantification of spatiotemporal cytokine concentrations, study demonstrating targeting stroma in solid tumors using STAT CAR T cells. LRRC15-
T cell motility, proliferation, cytotoxic activity, and tumor invasion patterns. Research directed CAR T cells showed antitumor efficacy in mouse models engrafted with LRRC15+
Sponsor: None. as well as LRRC15- tumors. Collectively, we show that targeting LRRC15+ CAFs in the
tumor with CAR T cells has the potential to inhibit solid tumor progression as well to
circumvent the challenge of limited penetration of T cells into the tumor site by disrupting
the stroma. Research Sponsor: None.

2542 Poster Session 2543 Poster Session

Association of lymphopenia rescue and CA19-9 levels with overall survival Predictors and clinical outcomes of CMV reactivation in CAR-T therapy: A
following IL-15 superagonist N-803 and PD-L1 t-haNK chemo- systematic review. First Author: Faiza Humayun Khan, Montefiore [Link]’s Cornwall,
immunotherapy for 3rd line or greater metastatic pancreatic cancer. First Collaborative Opportunities for Research, Training, And Excellence in Innovation
Author: Tara Elisabeth Seery, Chan Soon-Shiong Institute for Medicine, El Segundo, CA (CORTEX), Newburgh, NY
Background: Lymphopenia and high CA19-9 levels are associated with poor prognosis Background: Cytomegalovirus (CMV) reactivation is a common complication in im-
in pancreatic cancer patients. N-803 (ANKTIVA), an IL-15 superagonist is the first FDA munocompromised patients, particularly those undergoing chimeric antigen receptor T-
approved molecule with a mechanism of action of rescuing lymphopenia by cell (CAR-T) therapy. CMV reactivation has been linked to increased morbidity and
proliferating lymphocytes (NK and T cells). In QUILT-88, a Phase 2 multi-center study mortality due to immune dysregulation, relapses, and treatment-related toxicity. This
(NCT04390399), participants with 2nd line or greater locally advanced or metastatic systematic review investigates the predictors and outcomes of CMV reactivation in CAR-
pancreatic cancer (mPC) received N-803 and PD-L1-targeted high-affinity natural killer T recipients, focusing on survival, relapses, and non-relapse mortality (NRM).
(PD-L1 t-haNK) cell therapy in combination with low-dose chemotherapy as $ 3rd line Methods: A systematic review was conducted following PRISMA guidelines to compare
therapy. The absolute lymphocyte count (ALC), CA19-9 level, and correlation with overall characteristics and outcomes between CMV reactivation (R) and non-reactivation (NR)
survival (OS) was assessed. Methods: Patients (n = 84) received low-dose SBRT and groups. A comprehensive search of PUBMED, EMBASE, and CENTRAL identified 172
low-dose chemotherapy in combination with N-803 and PD-L1 t-haNK cells to or- studies, of which only four met the inclusion criteria after screening. A descriptive
chestrate responses of the innate and adaptive immune system, a paradigm change in statistical analysis was performed to calculate frequencies and percentages.
the treatment of mPC. The association between OS and ALC , or $ median 1.045 x 109 Results: Among 462 patients with CAR-T therapy, 114 (24.7%) experienced CMV
cells/L and CA19-9 , or $ median 4079.6 U/mL at baseline was assessed. reactivation. The median time from CAR-T therapy to CMV reactivation was 20 days
Results: Median OS for 3rd line patients (n = 43) was 6.2 months (95% CI 5.0 - 7.1;) and (Range: -1 to 73), with an incidence of CMV disease with end-organ damage at 1.73%.
for all patients $ 3rd to 6th line patients (n = 84) was 5.7 months (95% confidence interval Most patients received axicabtagene ciloleucel (71%) and had lymphoma (88%). Our
[CI] 4.3 – 6.4). OS was positively associated with both higher ALC and lower baseline analysis identified a higher proportion of patients receiving BCMA-targeted CAR-T
CA19-9 levels. OS was significantly higher for participants (median OS: 7.1 months) with therapy (7% vs. 2.9% in NR) and a greater prevalence of prior allogeneic hematopoietic
ALC $ 1.045 x 109 cells/L and CA19-9 , 4079.6 U/mL than for those participants stem cell transplantation (allo-HSCT) in the R group (35% vs. 7% in NR). Severe CRS
(median OS: 3.1 months) with ALC , 1.045 x 109 cells/L and CA19-9 $ 4079.6 U/mL (HR (Grade $3) was more common in the R group (11.4% vs. 8.6%), as was severe ICANS
3.6, p , 0.001). Higher ALC count was associated with prolonged OS over the course of of $3 (37.7% vs. 26.7%). Immunosuppressive therapy use, including steroids (56% vs.
the study. Grade 3 or higher TEAEs occurred in 95% of patients and were largely 42.8%) and combination therapy with tocilizumab and anakinra (12% vs. 5%), was
chemotherapy-associated. Conclusions: The multimodal chemo-immunotherapy pro- significantly higher in the R group. Outcomes varied across studies, with Lin et al. and
tocol to induce immunogenic cell death resulted in OS that exceeded 6 months for both Khawaja et al. reporting higher one-year mortality in R vs. NR groups (57% vs. 23%, P =
3rd and $ 5th line patients, exceeding OS achieved by other therapies in this setting by .001; 53% vs. 38%). Khawaja et al. also noted higher NRM (48% vs. 33%). Chen et al.
~2 months. It is notable that both favorable baseline ALC/CA19-9 and on-study higher identified CMV reactivation (HR 2.3, 95% CI: 1.2–4.5, P = .02) as an independent
ALC was associated with prolonged survival, given N-803’s ability to increase both NK mortality predictor, with relapse rates of 71.4% in R and 41.2% in NR. Conclusions: CMV
and CD8+/CD4+ T cells, the first FDA approved agent that proliferates lymphocytes in the reactivation is a significant complication in CAR-T therapy, linked to worse outcomes,
face of lymphopenia. These findings support further investigation of this novel ther- including increased mortality, relapse, and NRM. Predictors include BCMA-targeted
apeutic regimen that includes PD-L1 t-haNK, and N-803 that, as an IL-15 superagonist, CAR-T therapy, prior allogeneic HSCT, severe CRS/ICANS, and associated treatments.
may be able to overcome lymphopenia and improve prognosis. Clinical trial information: Targeted CMV monitoring, prophylaxis, and immunosuppressive strategies are essential
NCT04390399. Research Sponsor: ImmunityBio, Inc. for mitigating reactivation risks and improving outcomes in CAR-T recipients. Research
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 149s

2545 Poster Session 2547 Poster Session

Validation of an optimized tissue-agnostic genome-wide methylome en- Personalized tumor-informed circulating tumor DNA as predictor of pro-
richment assay to predict clinical outcomes in patients treated with gression risk after long-term responses to immunotherapy in advanced non-
pembrolizumab. First Author: Enrique Sanz Garcia, Princess Margaret Cancer Cen- small-cell lung cancer. First Author: Fang Wu, Department of Oncology, The Second
tre, University Health Network, University of Toronto, Toronto, ON, Canada Xiangya Hospital, Central South University, Changsha, China
Background: Recent work from the INSPIRE study (PMID38393391) suggests that ki- Background: Immune checkpoint inhibitors (ICIs) have remarkably improved survival in
netics of cell-free DNA (cfDNA) methylation profiles reflect immunotherapy treatment advanced non-small-cell lung cancer (NSCLC), with about 30%～40% of patients achieving
response in solid tumors. Here we provide validation data of a tissue-agnostic, genome- long-term responses. However, biomarkers for predicting progression remain undefined.
wide methylation enrichment assay based on cell free methylated DNA immunoprecipi- Circulating tumor DNA (ctDNA) has demonstrated its ability to predict recurrence in
tation and high throughput sequencing (cfMeDIP-seq) designed for clinical use, to de- resected NSCLC, but its potential to forecast progression following prolonged responses to
termine response to immunotherapy. Methods: This study utilizes samples and clinical ICIs requires investigation. Methods: CR1STAL study is a multicenter, prospective cohort
data from the INSPIRE study, a single-institution investigator-initiated phase II study of study investigating ctDNA surveillance to monitor progression risk in advanced NSCLC
pembrolizumab in multiple solid tumors given every 3 weeks (NCT02644369). A prior treated with first-line ICIs (NCT05198154). Patients with advanced NSCLC with long-term
published analysis of cfMeDIP used TCGA to develop a classifier and demonstrated an responses, defined as a PFS of about 1 year, were enrolled. Peripheral blood samples were
association of response to immunotherapy. In contrast, in this analysis, a novel quan- collected alongside radiographic evaluations. ctDNA was detected using a personalized
titative and highly specific measurement of ctDNA was estimated using a generative tumor-informed assay. Somatic variants were identified using a targeting 1,021 genes,
machine learning model trained on differentially methylated regions identified from a large followed by the design of individualized target-capture. ctDNA-positive was defined as the
cfMeDIP methylome atlas from individuals with and without cancer. In a blinded validation detection of ctDNA at any time during surveillance. The primary endpoint was PFS, defined
analysis, Firth’s logistic regressions were used to test differences in objective response as the time from enrollment until progression or death. Secondary endpoints included OS
(ORR) and clinical benefit rate (CBR) defined as complete or partial response or stable and ORR. Exploratory endpoints included the association between ctDNA features and
disease . / = 6 cycles between patients with a decrease in ctDNA from baseline to cycle 3 survival, and comparison to other biomarkers. Results: We analyzed 199 sample from 42
of treatment, and those with an increase in ctDNA. Sensitivity for no objective response, NSCLC patients. The median age was 60.5 years with 88.1% male, and 64.3% at stage IV.
specificity for objective response, and positive and negative predictive values (PPV and The median number of sample collections was 4, with a median follow-up time of
NPV) were summarized. Cox regressions and log-rank tests were used to evaluate dif- 24.7 months. ctDNA was detected in 54.8% of patients (23/42), with 82.7% of patients (19/
ferences in progression-free survival (PFS) and overall survival (OS) between the two 23) showing ctDNA-positive occurring within 2 years of ICIs treatment. A total of 23 PFS
groups. Results: The analysis included 64 unique patients with a median follow up of events were observed. The ctDNA-positive group showed significantly worse PFS com-
18.43 months (a total of 128 samples), including head & neck (n = 9), triple negative breast pared to the negative group (HR: 7.65, p , 0.001), with a positive predictive value of 90.0%
(n = 10), ovarian (n = 11), melanoma (n = 7), and other mixed solid tumor types (n = 27). A and a specificity of 88.2%. Additionally, ctDNA-positive provided a median lead time of
decrease in ctDNA was associated with significantly better objective response than an 6.6 months prior to radiological progression. ctDNA-positive significantly associated with
increase [odds ratio (OR) 33.89 (4.07, 44426.47), p = 0.0001], 58% sensitivity, 100% poorer OS (HR: 68.42, p = 0.003) and lower ORR (60.9% vs 89.5%, p = 0.036). 18 exhibited
specificity, 100% PPV and 35% NPV. Significantly better CBR [OR 10.17 (2.74, 55.74), p = clonal mutations. Compared to the ctDNA-negative group, the patients with clone had
0.0002] was also observed. A decrease in ctDNA was associated with significantly better significantly worse PFS (HR: 9.38, p , 0.001) than those with subclone (HR: 4.16, p =
PFS [hazard ratios (HR) 0.28 (0.15, 0.49) p , 0.0001] and OS [HR 0.42 (0.24, 0.76) p , 0.063). The ctDNA positivity rate was 84.6% in cases of local progression, 80.0% in distant
0.003]. Conclusions: A clinical tissue-agnostic, genome-wide methylome enrichment metastases with brain exhibiting lower positivity rates. Additionally, peripheral CEA
approach using cfMeDIP-seq accurately predicts clinical outcomes in patients treated with showed inferior predictive value for PFS (HR: 1.76, p = 0.303) than ctDNA.
pembrolizumab in multiple advanced solid tumors. This test provides relative quantifi- Conclusions: ctDNA has emerged as a promising biomarker for predicting progression risk
cation of methylated ctDNA to predict response to immmunotherapy and does not require of ICIs in advanced NSCLC patients with long-term responses. ctDNA surveillance enables
tumor tissue. This analysis highlights potential generalizability across tumor types in earlier detection of progression and supports treatment adjustments through adaptive
response monitoring. Clinical trial information: NCT02644369. Research Sponsor: Adela, therapy. Clinical trial information: NCT05198154. Research Sponsor: None.
Inc; Merck.

2548 Poster Session 2549 Poster Session

Novel dynamic circulating biomarkers for predicting therapeutic efficacy of An exploratory study to predict the efficacy and prognosis of immunother-
PRaG regimen in advanced refractory solid tumors. First Author: Yuehong Kong, apy for extensive-stage small cell lung cancer based on peripheral blood
The Second Affiliated Hospital of Soochow University, Suzhou, Jiangsu, China dynamic immune profiles. First Author: Lin Wu, Hunan Cancer Hospital, Changsha,
Background: Common biomarkers for predicting the efficacy of immune checkpoint China
inhibitors (ICIs), such as programmed death-ligand 1 (PD-L1) expression, face notable Background: The high-dimensional classification information of peripheral blood
challenges with tumor tissue sampling and the inability to enable dynamic monitoring. mononuclear cells can provide abundant efficacy and prognosis-related data. However, in
Circulating T lymphocyte subset classification and cytokines offers a promising al- the field of immunotherapy for extensive-stage small cell lung cancer (ES-SCLC), the
ternative, reflecting T cell functionality and predicting ICI responses. The PRaG regimen, biomarkers that can predict the efficacy and prognosis need to be explored and clarified.
combining PD-1 inhibitors, radiotherapy, and granulocyte-macrophage colony- Methods: Cytometry by Time-Of-Flight (CyTOF) was applied to the dynamic monitoring of
stimulating factor (GM-CSF), has shown efficacy in patients with metastatic or re- immunotherapy using clinical resources such as dynamic peripheral blood from ES-SCLC
fractory solid tumors unresponsive to standard therapies. This study seeks to develop an patients. By labeling the following proteins: CD45, CD3, CD4, CD8, CD25, CD127, CD45RA,
efficacy evaluation model by intergrating dynamic peripheral blood lymphocyte subsets CD45RO, CCR7, TCRgd, CD19, CD66b, CD14, CD56, CD16, CD11c, CD123, HLA-DR, CD38,
and cytokines, based on comprehensive analysis of clinical data from the PRaG studies. CD57, CXCR3, CCR6, CCR4, CXCR5, CD95/Fas, LAG-3, Tim-3, CTLA-4, PD-L1, PD-1, CD278/
Methods: Data from the PRaG 1.0 (ChiCTR1900026175), PRaG 2.0 (NCT04892498), and ICOS, and TIGIT, this study performed high-dimensional fine-phenotyping of peripheral
PRaG 3.0 (NCT05115500) studies were analyzed to evaluate the objective response rate blood immune cells from ES-SCLC patients. We further explored the dynamic immune
(ORR) by RECIST 1.1. Machine learning models, including linear, sequential, attention- profile of peripheral blood that could predict the efficacy and prognosis of immunotherapy
based, and hybrid models, were employed to predict disease progression. These models in combination with efficacy assessment and survival indicators. Results: 81 dynamic
utilized dynamic peripheral blood data from thirty-five lymphocyte subsets and seven peripheral blood samples (baseline, after two cycles of treatment[C2], and progressive
cytokines, collected across treatment cycles. Model efficacy was further validated using disease) were collected from ES-SCLC patients who received first-line immunotherapy
independent data from two additional PRaG studies (NCT05790447 and NCT06112041). combined with chemotherapy (n = 20) and chemotherapy alone (n = 7) in this study. In the
Results: As of November 30, 2023, 132 patients were included in the study, with a immunotherapy group, a high percentage of senescent CD4+TEM/CD4+TEM at baseline (P
median age of 63 years. Patients over 65 accounted for 41.7%, and 60.4% had more than = 0.029) was significantly associated with longer PFS. High TIGIT expression at baseline (P
= 0.016) was significantly associated with shorter PFS. In addition, PD-1 (CD4+TCM, P =
five metastatic sites. Patients with an ECOG score of 2-3 made up 59.7% of the cohort.
0.017; Naive CD4+T, P = 0.031; pDCs, P = 0.031; NK, P = 0.007; Early NK, P = 0.007; Late NK,
The ORR was 20.13%, and the disease control rate was 48.19%. Dynamic monitoring of
P = 0.02) and TIGIT (CD8+TEM, P = 0.046; NK, P = 0.038) expression levels at baseline in
peripheral blood features across treatment cycles facilitated the development of an
multiple cell subpopulations were significantly negatively correlated with OS. In contrast,
LSTM-HeterGNN model, which integrates long short-term memory (LSTM) networks
the above peripheral blood immune profile was not a predictor in the chemotherapy group.
with heterogeneous graph neural networks (heterGNN). This model outperformed ten In the immunotherapy group, peripheral blood dynamic monitoring showed that increased
other models, achieving a ROC AUC of 0.818. Independent validation further demon- gdT cell percentage after treatment was significantly associated with longer PFS and OS
strated robust performance, with a ROC AUC of 0.801. Conclusions: This study un- (PFS, P = 0.035; OS, P = 0.032). Increased CD4+ TEM and CD4+ TCM percentage after
derscores the potential of the PRaG regimen as an effective salvage therapy for treatment was significantly associated with shorter PFS and OS (CD4+ TEM: PFS, P =
advanced solid tumors after the failure of standard treatments. The LSTM-HeterGNN 0.021, OS, P = 0.036; CD4+ TCM: PFS, P = 0.01; OS, P = 0.014). Meanwhile, CTLA-4 and ICOS
model, leveraging dynamic peripheral blood biomarkers, provided precise efficacy expression in total cells at progressive disease was significantly higher than C2, suggesting
predictions, surpassing traditional models. These findings lay the groundwork for dy- that it might be related to immunotherapy resistance. In the chemotherapy group, the
namic treatment monitoring and optimization. Larger sample sizes are required to above peripheral blood dynamic immune profile did not predict the efficacy and prognosis
further validate the model’s generalizability. Research Sponsor: None. of chemotherapy. Conclusions: Dynamic peripheral blood immune profile can predict the
efficacy and prognosis of immunotherapy in ES-SCLC. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
150s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2550 Poster Session 2551 Poster Session

Longitudinal tumor-informed cfDNA whole genome sequencing to capture Plasma extracellular vesicles as biomarkers of primary versus acquired
residual disease during neoadjuvant immune checkpoint inhibition in re- resistance to immune checkpoint inhibitors (ICI) in patients (pts) with solid
sectable gastroesophageal cancer. First Author: Blair V. Landon, Sidney Kimmel tumors. First Author: Scott Strum, Princess Margaret Cancer Centre – University Health
Comprehensive Cancer Center, Johns Hopkins School of Medicine, Baltimore, MD Network, University of Toronto, Toronto, ON, Canada
Background: Although circulating tumor DNA (ctDNA) detection represents a promising Background: Plasma extracellular vesicles (pEVs) have emerged as promising biomarkers
approach to capture minimal residual disease (MRD), the clinical performance of ctDNA in the field of oncology. They can be obtained through minimally invasive methods, and
MRD during neoadjuvant immune checkpoint inhibition (ICI) remains understudied. Here hold the potential to help differentiate the clinically relevant subgroups of primary (PR) vs
we employ a tumor-informed whole genome sequencing (WGS) approach to capture acquired resistance (AR) to ICI treatments. We hypothesized that individual pEV-derived
residual disease and link ctDNA dynamics with pathologic response and clinical outcomes. protein cargo, or combinations thereof, associate with PR vs AR to ICI. Methods: A cohort
Methods: WGS was performed on tumor (n = 28), matched WBC (n = 28), and longitudinal of patients was derived from the Immune Resistance Interrogation Study (IRIS;
plasma samples (baseline, post-ICI cycle 1, post-ICI cycle 2 and pre-op; n = 97) from 28 NCT04243720), with plasma collected at the time of progression on ICI in advanced or
patients with resectable gastroesophageal cancer treated with neoadjuvant ICI and adjuvant settings (n = 69; n = 44 primary resistance [PR], n = 25 acquired resistance [AR]).
chemoradiation prior to surgical resection (NCT03044613). Tumor-specific single nu- Plasma-derived extracellular vesicles (pEVs) were isolated using serial ultracentrifugation
cleotide variants were identified from tumor and WBC datasets, from which a high and characterized per ISEV guidelines. All samples were analyzed using OLink Immuno-
confidence candidate variant set was used to determine the presence of ctDNA through a Oncology proteomics to evaluate 92 proteins. Statistical analyses included the Mann–
random forest machine learning model. ctDNA status and tumor fraction (TF) were de- Whitney U test, binary logistic regression, and log-rank tests. Primary and acquired re-
termined based on the level of signal compared to a reference population of noncancerous sistance were defined according to trial protocol. Results: A total of 57 out of 69 samples
donor plasma samples (n = 80). Serial ctDNA TF dynamics were correlated with overall (OS) (n = 37 PR, n = 20 AR) generated evaluable proteomics data. Of the 92 proteins analyzed, 11
and recurrence-free survival (RFS) in comparison to a tumor-naı̈ve targeted NGS gene were significantly overexpressed in AR compared to PR (ADGRG1, CD28, FGF2, IL10,
panel liquid biopsy approach. Results: Twenty-four of the 28 patients (86%) with evaluable IL12RB1, IL2, IL33, IL4, MCP3, PD-L2, PTN) (p , 0.05), with IL10 and IL33 showing the
specimens had ctDNA detected in a least one timepoint: 22 of 25 (88%) evaluable patients strongest associations (p, 0.01). When stratified by cancer type, 9/11 proteins were
had ctDNA detected at baseline, 20 of 25 (80%) evaluable patients had ctDNA detected overexpressed in AR vs PR among melanoma pts (n = 39; ADGRG1, CD28, FGF2, IL10, IL33,
post-ICI cycle 1, 18 of 26 (69%) evaluable patients had ctDNA detected post-ICI cycle 2, and IL4, MCP3, PD-L2, PTN) (p , 0.05), whereas only IL12RB1 (p , 0.01) was overexpressed in
5 of 21 (24%) evaluable patients had ctDNA detected at the pre-op timepoint. In contrast, HNSCC pts (n = 16). Analysis of 5 proteins most strongly associated with AR (IL10, IL33,
the tumor-naı̈ve targeted NGS approach detected 13 of 30 (43%), 12 of 30 (40%), 11 of 30 IL4, MCP3, CD28) yielded a sensitivity of 70% and specificity of 95% for AR vs PR, with a
(37%) and 5 of 25 (20%) patients at baseline, post-ICI cycle 1, post-ICI cycle 2 and pre-op positive and negative predictive value of 88% and 85%, respectively; AUC 0.853 (p , 0.001;
respectively. A ctDNA TF peak was detected at either the post-ICI cycle 1 or cycle 2 95% CI 0.742-0.963). Conclusions: In summary, 11 pEV-derived proteins from blood
timepoint for 50% of the patients. A 50% reduction in ctDNA TF at the post-ICI cycle 2 samples at progression on ICI independently statistically associated with AR vs PR, and a
timepoint showed a sensitivity of 80% and specificity of 69% for prediction of complete combination of 5 of them generated a highly accurate predictive model for AR. Immuno-
pathologic response, which was improved compared to the tumor-naı̈ve liquid biopsy modulatory cytokines IL10 and IL33 held the strongest associations, known to activate
approach and importantly showed a significantly higher evaluable rate (86% vs 62% for signaling cascades implicated in ICI resistance through the JAK-STAT and NF-Kappa-B/
tumor-informed and tumor-naı̈ve respectively). Similar trends were observed between MAPK pathways, respectively. Differentially expressed proteins may signify distinct
major pathologic response and ctDNA TF. A dramatic reduction of ctDNA TF ($65%) at the mechanisms of ICI escape. Despite requiring validation, our results highlight the potential
pre-op timepoint predicted longer OS and RFS (log-rank p = 0.0035 and p = 0.0032 re- of pEV-derived proteins as predictive biomarkers for ICI resistance in solid tumors. Future
spectively). Conclusions: Tumor-informed cfDNA whole genome sequencing analyses studies of pEV proteomics in the pre-treatment setting, as well as exploring other cargo
showed reliable and sensitive detection and quantification of ctDNA during neoadjuvant ICI such as RNA, may provide additional insights into the biology of resistance, and discover
and adds to the body of evidence supporting the clinical utility of ctDNA residual disease in minimally invasive clinically relevant biomarkers. Clinical trial information: NCT04243720.
interpreting clinical outcomes. Research Sponsor: Bristol-Myers Squibb; U.S. National Research Sponsor: BMO Chair in Precision Genomics, Dr. Lillian Siu.
Institutes of Health; CA121113; Cancer Research Institute.

2552 Poster Session 2553 Poster Session

Precision medicine research on chemo-immunotherapy combination treat- T-cell exhaustion and pre-existing T-cell immunity in circulation as predic-
ment for locally advanced or metastatic non-small cell lung cancer based on tive biomarkers for immunotherapy in NSCLC patients. First Author: Anastasia
deep plasma proteomics. First Author: Qiuchi Chen, Cancer Center, Union Hospital, Xagara, Laboratory of Oncology, School of Health Sciences, University of Thessaly,
Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China Larissa, Greece
Background: For locally advanced or metastatic non-small cell lung cancer (NSCLC) pa- Background: Pre-existing cancer-antigen specific T-cells describe the endogenous
tients lacking specific genetic mutations, chemotherapy combined with anti-programmed adaptive immunity before any treatment that may represent a valuable novel predictive
death-1/programmed death-ligand 1 immunotherapy has become standard first-line biomarker for ICI. In a recent publication we have shown a positive correlation of pre-
treatment with enhanced therapeutic efficacy and prolonged survival. However, 40-50% existing cancer-antigen specific CD8+ T-cells with the response to ICI. Here, we analyze
of patients do not benefit from chemo-immunotherapy and develop resistance. Currently, the major differences of exhausted T-cells between pre-existing positive (PreI+) and pre-
there is a lack of predictive biomarkers for the efficacy of combined therapy in NSCLC, and existing negative immunity (PreI-) NSCLC patients as well as, between different stages of
research on the regulatory mechanisms and drug targets is insufficient, either. We leveraged the disease. Methods: Blood was collected from 82 patients with NSCLC, 38 with stage
an advanced proteomics platform to profile serum in NSCLC patients, aiming to identify III and 44 with stage IV, before ICI therapy. PBMCs were isolated with Ficoll density
chemo-immunotherapy biomarkers and uncover resistance mechanisms. Methods: This gradient centrifugation from patients and 15 healthy donors (HD). PreI was calculated by
study collected pre-treatment plasma samples from 103 patients with locally advanced or detecting endogenous IFNg expressing cells with FACS after in-vivo co-cultures of
advanced NSCLC receiving chemo-immunotherapy. These samples were analyzed using a PBMCs with mixes of hTERT, MAGEA1, NY-ESO-1 kai Survivin cancer-associated
deep proteomics platform that integrates antibody arrays and mass spectrometry. Patients antigens. T-exhausted signatures were detected by multi-color flow cytometry using
were classified into "responders" (R, complete/partial response or stable disease .
antibodies against CD3, CD4, CD8, PD-1 and TCF1. Results: 47% (18/38) of patients with
6 months) and "non-responders" (NR, progressive disease or stable disease #6 months)
stage III disease and 41% (18/44) of stage IV had peptide specific T-cells (PreI+ patients).
based on treatment efficacy. Differentially expressed serum proteins were identified be-
tween the groups, and weighted gene co-expression network analysis (WGCNA) was applied.
Survival analysis revealed better OS only in stage III PreI+ compared to PreI– patients
Cox survival analysis was conducted on prognosis-related modules, leading to the iden- (Log-rank = 0.04), while for stage IV (p=0.081) there was only a trend. The percentages of
tification of key proteins associated with treatment efficacy and survival. Results: Through CD8 T-cells that were PD-1+TCF1+(p=0.030) and PD-1+TCF1-(p=0.041) were higher in
our high throughput blood proteomics platform, a total of 1,397 proteins were detected. The patients compared to HD, and additionally both T-cell populations harbored higher
median progression-free survival was 9 months, and the median overall survival was levels of PD-1 protein expression (p=0.003 and p=0.032 for stage III) as it was shown
32 months. A total of 175 differentially expressed proteins were identified between the R and with mean fluorescence intensity (MFI). Moreover, low percentages of PD-1+TCF1+
NR groups. WGCNA identified 12 distinct modules, with ME4 associated with poor prog- ware associated with longer survival (p= 0.037) only in stage III patients. By
nosis, enriched in inflammation, gene activation, and apoptosis suppression pathways, while subgrouping stage III patients, we observed that all patients with PreI+ harboring low
ME8 correlated with favorable prognosis and ERK1/ERK2 cascade regulation. In the NR percentages of exhausted PD-1+TCF1+ were alive at the end of the follow up.
group, upregulated proteins associated with poor prognosis included erythropoietin receptor Conclusions: Combinatorial analysis of Pre-existing tumor-antigen specific immunity
(HR: 1.41, p , 0.01), fibrinogen gamma chain (HR: 1.90, p: 0.03), Fc alpha receptor (HR: and the status of T-cells before initiation of ICI in stage III NSCLC could serve as a good
2.63, p , 0.01), and prion protein (HR: 1.30, p: 0.04). In contrast, upregulated proteins in the predictive factor of response. The study is ongoing. Research Sponsor: None.
R group linked to favorable prognosis were insulin-like growth factor-binding protein 2 (HR:
0.77, p: 0.02), keratin 19 (HR: 0.61, p: 0.02), and retinol-binding protein 4 (HR: 0.74, p: 0.03).
Conclusions: Through in-depth proteomics analysis, this study systematically character-
ized the plasma proteomic landscape of patients undergoing chemo-immunotherapy,
identifying potential novel biomarkers, and providing new insights to optimize clinical
decision-making. Research Sponsor: Beijing Xisike Clinical Oncology Research Foundation;
Y-Young2023-0125.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 151s

2554 Poster Session 2555 Poster Session

Effect of fusobacterium nucleatum on NF-kB/HIF-1a/CCL20 pathway and Monitoring PD-L1 expression in cancer-associated macrophage-like cells as
M2 macrophages infiltration in esophageal squamous cell carcinoma. First predictor of clinical outcomes in metastatic cancer patients treated with PD-
Author: Yu Su, The Fourth Hospital of Hebei Medical University, Shijiazhuang, China L1 immunotherapies. First Author: Dimpal M. Kasabwala, Creatv MicroTech, Inc.,
Background: Esophageal squamous cell carcinoma (ESCC) is one of the most common Monmouth Junction, NJ
digestive malignant tumor with the highest mortality rate in [Link] studies have Background: Studies have described the efficacy of immunotherapies (IMT) utilizing
shown that Fusobacterium nucleatum (F. nucleatum) can attenuate the efficacy of programmed death 1 receptor and its ligand (PD-L1) for treating solid tumors. However,
immunotherapy in ESCC patients through various [Link] of them is recriuting many patients (pts) fail to respond to IMT, necessitating better predictive biomarkers for
more M2-like macrophages through tumor-derived cytokines such as CCL20. improved stratification. Poor IMT responses are often attributed to the dynamic nature of
Methods: From January 2022 to June 2023, a total of 30 patients with ESCC from 4th PD-L1 likely changing after chemotherapy or radiation, but typically quantified by static
Hospital of HeBei Medical University were [Link] of the pateints did not apply any immunostaining. Recent studies have described PD-L1 upregulation in giant phagocytic
neoadjuvant therapy before and received radical [Link]-time reverse tran- stromal cells, i.e. Cancer associated macrophage-like cells (CAML), circulating macro-
scriptase - PCR(RT-PCR) were performed to examine the expressions of F. nucleatum in phages that engulf tumor before entering circulation and may predict IMT responses. We
tumor [Link] to the CT values,the level of the F. nucleatum infection could be conducted a pilot study to monitor the peripheral blood of n = 111 metastatic cancer pts
seperated in two groups - positive group and negative group. Immunohistochemistry undergoing systemic treatment with IMT in combination with other therapies, to evaluate
(IHC) staining were used to examine the expressions of CCL20、CD206 and HIF-1a in CAML PD-L1 prior to & post IMT induction with clinical correlation at 2 years. Methods: In a
prospective pilot study of n = 111 metastatic cancer pts, breast (n = 42), lung (n = 46), renal
both groups. Immunofluorescence(IF) was characterised CD206 on CD68+ macro-
cell (n = 10), prostate (n = 5), esophageal (n = 5) & colon (n = 3), all starting new lines of
[Link] addition,transwell assay was carried on to quantify macrophages migration
systemic chemotherapy in combination with IMT (pembrolizumab [n = 69], nivolumab [n =
ability in [Link] last,Western blot was used to determine the expression level of CCL20,
23], Durva [n = 4], or atezolizumab [n = 13]) for new recurrent metastasis (n = 45) or with
HIF-1a and p65/p-p65 on ESCC cells. Results: F. nucleatum DNA positivity was sig- previously treated progressive metastatic disease (n = 66). We isolated CAMLs from 7.5 ml
nificantly associated with higher expression of CCL20, HIF-1a and accumulation of baseline (T0) blood using the LifeTracDx PD-L1 test and scored PD-L1 as high or low. If
CD68+CD206+ [Link] showed that the expression of CCL20 and HIF-1a were possible, a follow-up sample (T1) was taken (~56 days) after IMT induction. Pts’ progressive
higher in F. nucleatum positive tumor tissue. After coculture with F. nucleatum and free survival (PFS) and overall survival (OS) hazard ratios (HRs) were analyzed by censored
Eca109 cells in vitro, CCL20 and HIF-1a production by ESCC cells were accel- univariate analysis based on RECIST v1.1 over 2 years. Results: T0 PD-L1 CAML data was
[Link] assay showed using siCCL20, CCL20-Nab or siCCR6 could decline the available for 78% (n = 86/111) of pts, with 34% (n = 29/88) having high CAML PD-L1 which
macrophages invasion level caused by CCL20. Otherwise, siHIF-1a could lowered was not correlated with improved PFS (HR = 0.99, p = 0.9416, CI = 0.6-1.6) or OS (HR = 1.1, p
the CCL20 expression level caused by F. [Link] addition,Western blot revealed = 0.9472, CI = 0.6-1.8). T1 PD-L1 CAML data was available for 74% (n = 82/111) of pts, with
NF-kB pathway was highly activated in Fn educated Eca109 [Link] BAY11-7082 44% (n = 36/82) having high CAML PD-L1, which significantly correlated with improved PFS
could decline both CCL20 and HIF-1a level triggered by F. nucleatum. (HR = 3.1, p = 0.0002, CI = 1.8-5.5) & OS (HR = 6.6, p , 0.0001, CI = 3.4-12.7). In comparing
Conclusions: Fusobacterium nucleatum promotes esophageal squamous cell carci- CAML PD-L1 change post IMT, it was found that consistently low PD-L1 at T0 & T1 had the
noma progression via NF-kB/HIF-1a/CCL20 pathway-mediated migration of M2-like poorest responses, median PFS (mPFS) = 4.9 months & median OS (mOS) = 7.2 months. In
macrophages into the tumour micro-environment and deteriorates the suppression of contrast, consistently high PD-L1 at T0 & T1 had better responses, mPFS = 8.4 months &
the local immune micro-environment within the [Link] bacteria may be a biomarker mOS = 18.3 months. Further, pts who increased in CAML PD-L1 after IMT had the best OS
to predict the efficacy of immunotherapy in ESCC. Research Sponsor: None. response rates, mPFS = 3.8 months & mOS = 20.9 months. Conclusions: We utilized a
cancer agnostic blood-based biopsy to monitor PD-L1 changes in circulating tumor immune
cells and predict clinical benefit to PD-L1 IMTs in several cancer types. While this initial pilot
study appears to stratify pts with better IMT responses, larger validation studies are needed.
Research Sponsor: Creatv MicroTech.

2557 Poster Session 2558 Poster Session

The classification of uncertain histologies based on cfDNA fragmentomic A molecular biomarker for longitudinal monitoring of therapeutic efficacy
analysis in patients with uncommon cancers screened for NCI-MATCH. First in a real-world cohort of advanced solid tumors treated with immune
Author: Chris Alan Karlovich, Molecular Characterization Laboratory, Frederick National checkpoint inhibitors. First Author: John Guittar, Tempus AI, Chicago, IL
Laboratory for Cancer Research, Frederick, MD Background: Clinical validation studies have shown that early dynamic changes in cir-
Background: The NCI Molecular Analysis for Therapy Choice (NCI-MATCH) was a precision culating tumor DNA (ctDNA) tumor fraction (TF) can predict clinical outcomes. Yet few
medicine trial that assigned targeted treatments to patients independent of histology. We studies have evaluated the clinical value of longitudinal monitoring throughout the course
sequenced . 2500 NCI-MATCH ctDNA samples of uncommon histology (excludes colon, of treatment. Here we evaluate longitudinal changes in ctDNA TF and clinical outcomes in
breast, non-small cell lung, and prostate) using the TruSight Oncology 500 (TSO500) v2 an advanced real-world pan-cancer cohort of patients (pts) treated with immune
ctDNA assay. We generated a probabilistic model from ctDNA fragment sizes to assign checkpoint inhibitors (ICIs). Methods: The cohort included deidentified pts from the
histology to patient samples designated as “not otherwise specified (NOS)” in pathology Tempus clinicogenomic database with stage IIIB or IV solid tumors who underwent ctDNA
reports. Methods: Fragmentomics data were computed by binning the segment-level data NGS and were treated with an FDA-approved ICI +/- chemotherapy (CT). Pts had a pre-
to each exon, followed by Shannon entropy calculation. The Least Absolute Shrinkage and treatment baseline liquid biopsy (T0) and $1 on-treatment sample Ti within 21-180 days of
Selection Operator (LASSO) algorithm was then used to build the histology classifier for ICI. xM for treatment response monitoring estimates TF via an ensemble algorithm that
each of 29 histologies (n = 1614 samples of known histology) using a 9:1 random training incorporates variant and copy number information. Response status was determined at
and validation separation for validation accuracy estimation. The 9:1 training/validation each on-treatment timepoint Ti relative to T0. Pts were classified as a Molecular Responder
was repeated 10 times. During each training/validation step, another 10-fold cross val- (MR) if TF , 1% at T0 and Ti or if TF decreased by $ 50% from T0 to Ti; otherwise, they
idation within the training set was used to find the best hyperparameter. The final model were classified as a Molecular Non-Responder (nMR). Pts with TF , 0.09% at T0 and Ti
was re-trained on total samples which combined 29 classifiers and final predicted his- were classified as ctDNA not detected. The longitudinal cohort included pts with $2 on-
tology was determined using a winner-take-all approach. Mutation data was generated by treatment timepoints, further classified based on the most frequent classification across
Illumina’s DRAGEN TSO500 ctDNA analysis pipeline and annotated using the OncoKB each Ti as longitudinal MR, longitudinal nMR and longitudinal ctDNA not detected. If no
knowledge base. Results: The fragmentomics-based classification model achieved 98.2 classification was dominant, the most recent classification was used. Real-world overall
validation accuracy across 1614 evaluable samples. The accuracy was increased to 99.6% survival (rwOS) was defined from T1 to death and assessed by log-rank test. Results: The
on a subset of samples (n = 1413) with high confidence prediction defined as prediction full cohort of 84 pts with . 10 solid tumors included 34.5% (n = 29) NSCLC and 16.7% (n =
probability . 0.5. The model was further used to classify 232 samples with NOS histology. 14) SCLC. The majority of pts (59.5%, n = 50) received ICI+CT, (64.0% first-line [1L]), and
Although there were no ground truth cases, those where a histology was predicted with 40.5% (n = 34) ICI-only (35.3% 1L). rwOS was longer for 53 MRs vs. 18 nMRs (median not
high confidence were closely aligned with their broader annotations. For example, among reached vs. 7.0 months, P , 0.005); 13 pts had no ctDNA detected. The longitudinal
37 pancreatic cancer (excluding Islets) NOS patient samples, 18/19 high-confidence subcohort of 35 pts was 31.4% (n = 11) NSCLC, 20.0% (n = 7) prostate cancer, and 14.3% (n
predictions were adenocarcinoma of the pancreas. The model was further explored on = 5) SCLC; 45.7% (n = 16) ICI+CT (56.3% 1L), 54.3% (n = 19) ICI-only (31.6% 1L). On average,
female reproductive system cancer NOS, which includes several rare histologies with pts had 3 on-treatment liquid biopsies, with a median time between on-treatment samples
limited or no representation in the prediction model. In this analysis, 17/23 high-confidence of 91 days. Twelve pts were longitudinal nMRs (10 MR or ctDNA not detected at T1), 18
predictions were designated as ovarian epithelial cancer (OEC) or related histologies. longitudinal MRs, and 5 longitudinal no ctDNA detected. There were 6 death events in the
Interestingly, one oncogenic BRCA2 mutation (p.T1388fs) was detected in a predicted longitudinal nMRs, 3 in the longitudinal MRs (all . 18 months after T1), and no death
ovarian epithelial cancer sample. Conclusions: The validation accuracy was high in this events in the longitudinal no ctDNA detected group. Longitudinal MRs had longer rwOS
exploratory analysis of uncommon histologies. High-confidence prediction was achieved than longitudinal nMRs (median 37.4 months vs 8.7 months, P , 0.005).
for adenocarcinoma of the pancreas although prediction confidence was lower in OEC or Conclusions: Longitudinal nMR was associated with worse survival compared to longi-
related histologies . This study may support an expanded role for large cfDNA targeted tudinal MR. Longitudinal, molecular biomarker dynamics may be a useful clinical treatment
panels beyond the identification of clinically actionable mutations to the classification of decision tool for monitoring treatment response to ICI therapy. Research Sponsor: Tempus
tumor histologies. Research Sponsor: National Cancer Institute. AI.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
152s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2559 Poster Session 2560 Poster Session

Predictive imaging of the immunotherapy and radioimmunotherapy re- Ultrasensitive ctDNA profiling to identify long-term survivors in phase I
sponse by immunoPET via a new target (CD103) and innovative protein immunotherapy trials. First Author: Oriol Mirallas, Vall d’Hebron University Hospital
formats in preclinical NSCLC. First Author: Léa Zimmermann, Université Paris- and Vall d’Hebron Institute of Oncology (VHIO), Barcelona, Spain
Saclay, CEA, CNRS, Inserm, BioMaps, SHFJ,, Orsay, France Background: Patients (pts) included in early clinical trials (ECT) typically show a median
Background: Immune checkpoint immunotherapies (ICI) have transformed cancer treat- overall survival of 8-10 months (PMID: 18042834 and 21975023), with long-term
ment, but patients have varied responses and potential risks of autoimmune disease. To survivors (LTS) rarely encountered. While prognostic scores have been developed to
improve ICI, we need to identify biomarkers to select responding patients and research new identify short-term survivors (STS) in ECT, predictors of LTS remain largely unexplored.
approaches. To this, resident memory T cells (TRM) LT CD8+ CD103+, have been identified Ultrasensitive circulating tumor DNA (uctDNA) serves as a reliable surrogate for tumor
as a promising tumor-specific biomarker for studying therapeutic efficacy involving ICI. burden (Toledo R et al, ASCO 2024), and may provide insights into LTS. This study aims
Internal radiotherapy appears to be a promising approach. Our objective is to develop new to identify key determinants of LTS. Methods: A case-control study was conducted on
therapeutic approaches combining radiotherapy and ICI (radioimmunotherapy) using pts treated at VHIO’s phase I unit between 2013 and 2023, as part of the institutional
CD8+CD103+ immunoPET imaging as a predictive biomarker of efficacy. Methods: After translational study RIO360. Pts with an overall survival (OS) . 3 years after C1D1 were
developing and characterizing a new CD103 radiotracer and validating the dual 18F PET-Scan classified as LTS and compared to STS with OS , 1 year, matched for age, sex, stage,
imaging for CD8+ and CD103+ in C57BL/6 mouse models, we implanted mice subcutaneously prior immunotherapy (IO), and ECOG. Clinical, histopathological, uctDNA, and treatment
with MC38 and othotopically syngenetically with LLC. We evaluated the efficacy of radi- outcomes were collected. uctDNA analysis was performed prior to C1D1 and throughout
oimmunotherapy vs ICI and the predictive effect of TRM in this syngeneic orthotopic model.
treatment (every 3 to 4 weeks) in the subset of patients treated with IO. Comparisons
To this, we implanted two syngeneic NSCLC cell lines, either LLC for a cold tumor or CMT167
between LTS and STS were performed using univariate binomial generalized linear
for a warmer tumor. We performed double imaging prior to treatment with [18F]-CD8 mutated
models. Results: Of 1282 pts, a total of 117 pts (9.1%) were classified as LTS (median
FcRn and [18F]-CD103 minibody on 2 consecutive days. We treated our mice 3 times, 3 days
apart, with the first dose of either cold Avelumab or [177Lu]-Avelumab (8MBq). The second
OS 5.2 years) and compared with 117 matched STS pts (median OS 0.6 year). The most
and third doses were cold Avelumab. We performed double post-treatment imaging with common tumor types among LTS were HNSCC (18%), melanoma (10%), and breast
[18F]-CD8 mutated FcRn and [18F]-CD103 minibody on 2 consecutive days, as well as ex vivo (9.1%), while in STS were colorectal (27%), breast (12%), and melanoma (11%).
analyses and a survival study. Results: There was a trend towards improved survival in Treatment with 3+ prior lines was more common in STS (45%) than LTS (24%) (p =
[177Lu]-Avelumab treated mice vs Avelumab treated mice but more markedly in the im- 0.002). Visceral metastases were present in 59% LTS and 77% STS (p = 0.010). Higher
munogenic model (30d vs 21d). The impact on tumor growth was assessed by comparing neutrophil count and dNLR were associated with STS (p , 0.05). LTS had lower mean
the two treatment groups with untreated mice. Radioimmunotherapy induced a significant uctDNA at baseline (7.5 vs 10.9 ppm; p , 0.001). Two consecutive uctDNA decreases
decrease in overall tumor growth compared with mice treated with ICI (0.45 ccm vs 0.68 were observed in 92% of LTS pts, while 80% of STS pts showed two consecutive uctDNA
ccm, *p , 0.05 turkey’s multiple comparison test) in the immunogenic model. In the cold increases (p , 0.001). Conclusions: In ECT, predictors of long-term survival include the
tumor model, there was a significant difference in the tumor size ratio before and after absence of visceral metastasis, less prior treatment exposure, low baseline uctDNA
treatment, for mice treated with radioimmunotherapy vs ICI (12.96 vs 26.08, ***p , 0.001 levels, and early decreases in uctDNA. This study further explores uctDNA dynamics
Uncorrected Fisher’s LSD). An increase in immune infiltration was validated by PET and flow during treatment, with detailed findings to be presented. Research Sponsor: BBVA.
cytometry for the immunogenic model (pre-treatment: 5.00%ID/cc max, versus post-
treatment 8.23%ID/cc max for CD8, *p , 0.05 two way-Anova). More heterogeneous re-
sults were observed in the cold tumor model. Conclusions: Radio-immunotherapy reduces
tumour growth and stimulates the immune system by circulating LTCD8. The dual 18F PET-
Scan imaging for CD8+ and CD103+ offer a promising non-invasive visualization of tumor-
infiltrating CD103+ TRMs. We need to correlate LTCD8+ /CD103+ infiltration with therapeutic
response. Research Sponsor: None.

2561 Poster Session 2562 Poster Session

Ultrasensitive ctDNA monitoring to reveal early predictors of immunother- Role of pelareorep in activating anti-tumor immunity in PDAC. First Author:
apy success in advanced cancer. First Author: Daisuke Nishizaki, UC San Diego Richard Trauger, Oncolytics Biotech, San Diego, CA
Moores Cancer Center, La Jolla, CA Background: Pancreatic ductal adenocarcinoma (PDAC) is highly lethal cancer with
Background: The potential of immune checkpoint inhibition (ICI) therapy is constrained limited immunotherapeutic options. Pelareorep (pela) is an intravenously delivered
by the inability to predict patient response. Circulating tumor DNA (ctDNA) has emerged unmodified reovirus containing a double stranded RNA genome that has been studied as
as a promising tool for real-time response tracking and early prediction of therapeutic an immunotherapeutic in multiple cancers including breast, anal, colorectal and pan-
outcomes. However, the clinical utility of ctDNA-based liquid biopsy faces a critical creatic. We previously reported high tumor response rates in first-line metastatic PDAC
challenge: reliable detection in low-shedding tumors and during dramatic therapeutic patients treated with pela combined with gemcitabine, nab-paclitaxel and atezolizumab.
responses when ctDNA levels approach the analytical threshold. We overcome this We report here the immunologic effects of pela in a cohort of first-line metastatic PDAC
technical limitation, achieving the precise longitudinal monitoring needed to optimize ICI subjects treated with pela plus chemotherapy and atezolizumab and the correlation of
therapy. Methods: We analyzed longitudinal plasma samples from 43 patients with these effects with tumor response. Methods: To examine its effects on pancreatic
treatment-refractory metastatic cancers spanning 8 distinct groups, composed primarily cancer, a phase 1/2 Simon 2-stage platform study (GOBLET) was performed that in-
of GI (n = 17) and gynecological cancers (n = 6). Patients underwent a median of 1 cluded patients with first-line locally advanced/metastatic unresectable PDAC. Anti-
previous line of therapy (range 0-8). Using NeXT Personal, an ultra-sensitive person- reovirus T cell activity was assessed by interferon gamma secretion (ELISPOT). Changes
alized liquid biopsy approach, we tracked up to 1,800 patient-specific somatic variants in the expression of plasma proteins were analyzed by Olink Response panels. T cell
per case across 250 plasma samples. This methodology achieves exceptional analytical receptor sequencing (TCR-seq; ImmunoSEQ Assay, Adaptive Biotechnologies) was
sensitivity, detecting circulating tumor DNA at levels as low as 1-3 parts per million performed on tissue collected prior to the start of therapy and on blood from baseline
(PPM). Results: ctDNA was detected across five orders of magnitude (2.0-239,315 through 3 treatment cycles to identify TILs expansion. Tumor responses were scored
PPM, median LOD: 1.76 PPM), with 31% of positive signals falling in the ultrasensitive according to the modified RECIST v1.1 criteria. Results: Increases in anti-reovirus T cell
range below 100 PPM. Early molecular response, measured by . 50% ctDNA reduction activation as determined by ELISPOT after cycle 3 of therapy were observed in 6/8
or sustained ctDNA negativity from baseline to first follow-up (median 23 days), strongly subjects. Three subjects with maximum responses (. 300 spots) showed .30% de-
predicted improved progression-free survival (PFS) (HR = 0.22, 95% CI 0.07-0.70, p = creases in tumor volume. Significant changes in plasma proteins as determined by Olink
0.006), representing a 3-fold increase in 1 year PFS rates. Achievement of durable included PD-L1, CXCL9, CXCL10, CXCL11 and IFN-g. Pre-treatment tumor tissue was
molecular complete response (dmCR), defined as sustained ctDNA clearance . used to identify the TIL clonal populations prior to therapy. Increased TIL clones in the
120 days, emerged as a powerful predictor of PFS, with dmCR patients maintaining 100% blood pre-treatment was associated with tumor responses. Pela treatment increased the
progression-free status at 12 months compared to 63% in non-dmCR patients (HR = expansion of pre-existing and new TIL clones in the blood after one cycle of treatment.
0.10, 95% CI 0.01-0.92, p = 0.017). This survival advantage persisted at 18 months with Sustained increases in pre-existing TIL clonal populations in the blood through cycle 3 of
80% PFS in dmCR patients versus 21% in non-dmCR patients. Conclusions: Early ctDNA therapy were observed in subjects who exhibited reductions in tumor volume.
kinetics predict long-term ICI outcomes across multiple advanced cancer types. The Conclusions: These findings, while preliminary, demonstrate that pela induces not only
ability to detect ultra-low ctDNA levels proved critical for accurate minimal residual anti-reovirus T cells but also activates innate and adaptive anti-tumor immunity in PDAC
disease assessment, even in this heavily pretreated cohort. These results establish high- subjects treated with chemotherapy and atezolizumab. Tumor responses are associated
sensitivity ctDNA monitoring as an essential tool for precise, real-time evaluation of with both the presence of TIL clones in the blood prior to treatment and the expansion of
immunotherapy response to guide clinical decision-making. Clinical trial information: pre-existing TILs in the blood on treatment. These findings provide additional insights
NCT02478931. Research Sponsor: Personalis. into the immunologic mechanisms by which pela-based therapy may provide clinical
benefit in patients with metastatic PDAC. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 153s

2563 Poster Session 2564 Poster Session

ctDNA features of acquired resistance to immunotherapy in advanced Mutation profiling of appendiceal cancer: Distinguishing tumor grades,
NSCLC. First Author: Sofiane Taleb, Gustave Roussy Cancer Center, Villejuif, France comprehensive mutation landscape, and ctDNA as a discovery tool. First
Background: Acquired resistance to systemic therapies, including immune checkpoint Author: Sefali Patel, AHN Cancer Institute, Pittsburgh, PA
inhibitors (ICIs) and tyrosine kinase inhibitors (TKIs), is a major clinical challenge in Background: Appendiceal cancer (AC) encompasses rare tumors with varying clinical
advanced non–small cell lung cancer (NSCLC). While mechanisms of resistance to behavior. Histologic grade is a key determinant of disease biology and prognosis. This
targeted therapies are well-documented, the genomic alterations associated with re- study utilized an in-house circulating tumor DNA (ctDNA) biomarker discovery pipeline
sistance to immunotherapy remain poorly understood. Circulating tumor DNA (ctDNA) to assess genetic determinants of histologic grade in AC, analyzing both peripheral blood
profiling offers a non-invasive approach to identify real-time genomic changes driving and tumor tissue. Methods: Paired peripheral blood and solid tumor samples were
resistance, providing novel insights into the underlying biology of immunotherapy collected from 52 patients undergoing surgery for AC (18 low-grade and 34 intermediate/
failure. Methods: We performed a prospective ctDNA sequencing study in 57 advanced high-grade). Comprehensive genomic profiling (CGP) using the TSO500 assay was
NSCLC patients from two cohorts: STING (n = 30, NCT04932525) and COPE (n = 27, performed on ctDNA, tissue-derived DNA and buffy-coat (germline)-derived DNA.
NCT04258137). Plasma samples were collected before treatment initiation and at Tumor-specific and germline mutations were analyzed using OncoKB, which classifies
disease progression following objective response to anti–PD-1 therapy or TKIs. ctDNA variants as oncogenic, likely oncogenic, or actionable (Level 1 therapeutic mutations
analysis was conducted to detect emergent genomic alterations, evaluate tumor mu- with an approved therapy). The concordance of mutations between solid tumor and
tation burden (TMB), and quantify circulating tumor fraction (TF). Results: The study plasma CGP assays was assessed, categorizing variants as detected in both tumor and
included 57 patients, with a median age of 62 years [IQR: 54.5–70], and 54.4% (31/57) ctDNA, ctDNA only, or tumor only. Results: ctDNA exhibited 82% concordance with
were male. At disease progression, 64.9% (37/57) of patients exhibited emergent ctDNA tumor tissue for known actionable mutations. Among 26 patients that were identified to
alterations associated with secondary resistance, independent of therapy type. Among have Level 1 therapeutic mutations, 88% (n = 23) had matching mutations in plasma.
these, 70% (26/37) harbored multiple newly arising mutations. In the non–oncogene- Frequently detected mutations included KRAS (40%), GNAS (30.8%), SMAD4 (28.8%),
addicted NSCLC cohort receiving anti–PD-1 therapy (n = 30), 56.6% (17/30) exhibited and TP53 (28.8%). Germline analysis revealed additional variants, including RUNX1
emergent resistance alterations, with 76.5% (13/17) harboring multiple mutations. (71.2%), NOTCH4 (50%), and BARD1 (48.1%). Tumor-specific TP53 and SMAD4 mu-
Frequently observed aberrations included mutations in NOTCH1/3 (n = 3), KEAP1 (n = 3), tations correlated with high-grade tumors; while GNAS was more prevalent in low-grade
KMT2B (n = 2), POLE (n = 2), SETD2 (n = 2), TYRO3 (n = 2), STK11 (n = 2), TSC2 (n = 1), tumors. Germline analysis identified NOTCH3 and SPEN mutations predominantly in
TGFBR2 (n = 1), PTPN11 (n = 1), SPEN (n = 1), STAG (n = 1), CDH1 (n = 1), and CTNNB1 (n high-grade tumors, suggesting that inherited determinants may determine tumor grade
= 1). The median progression-free survival (mPFS) was 7.0 months [95% CI: 5.0–10.3]. In (23.5% each). Plasma samples exhibited lower variant allele frequencies, limiting
the anti-EGFR TKI cohort (n = 13), 84.6% (11/13) displayed emergent ctDNA alterations, sensitivity for novel biomarker discovery. Concordance analysis revealed some mu-
including mutations in EGFR (n = 2), PIK3CA (n = 2), and MET (n = 1). The mPFS was tations were exclusive to solid tumors, while others were plasma-specific, highlighting
8.6 months [95% CI: 6.7–11.2]. TMB and circulating TF did not significantly change the need for a multi-modal genomic assessment. Conclusions: Tumor TP53, GNAS, and
between baseline and progression in both cohorts. Conclusions: This study highlights SMAD4 mutations serve as molecular classifiers for histologic grade differentiation in
the high prevalence of emergent genomic alterations detected by ctDNA in advanced AC, while germline NOTCH3, SPEN, RUNX1, NOTCH4, and BARD1 variants may influence
NSCLC patients developing resistance to ICIs and TKIs. In the ICI-treated cohort, histologic grade. ctDNA showed strong concordance for actionable mutations but had
mutations in NOTCH1/3, KEAP1, and STK11 emerged as significant resistance drivers, reduced efficacy for novel mutation discovery in the plasma samples. These findings
consistent with their roles in immune evasion, oxidative stress regulation, and impaired underscore the value of integrating tumor and germline profiling for classification and
immune cell infiltration, as supported by prior studies. Notably, this is the first study treatment stratification, while refining liquid biopsy methodologies to enhance sensi-
investigating features of ctDNA at acquired resistance to immunotherapy using an FDA- tivity in AC research. Research Sponsor: None.
approved assay. These findings underscore the heterogeneous and polyclonal nature of
resistance to ICIs. Research Sponsor: None.

2565 Poster Session 2566 Poster Session

Investigating the association of blood-to-tissue tumor mutation burden Identifying peripheral cancer-associated TCR signals for the early-detection
(TMB) ratio with overall survival and intratumor heterogeneity (ITH) in of lung cancer. First Author: Chen Huang, China-Japan Friendship Hospital, Beijing,
advanced NSCLC. First Author: Leeseul Kim, University of Chicago, Chicago, IL China
Background: High tumor mutation burden (TMB) measured via tissue-based next- Background: Tumor-associated antigens and neoantigens play a critical role in eliciting
generation sequencing (NGS) (tTMB) has been shown to predict better survival out- anti-tumor immune responses, leading to a significant amplification of tumor-specific T cell
comes in certain cancers treated with immune checkpoint inhibitors (ICIs). With the clones. Consequently, the detection of tumor-associated immune signaling in peripheral
increasing use and sensitivity of blood-based NGS, blood-based TMB (bTMB) is fre- blood is expected as a promising strategy for cancer screening. Notably, due to the am-
quently employed as an alternative. However, the prognostic significance of bTMB plification effect of the immune response, the identification of tumor-related immune
relative to tTMB and its correlation with intratumoral heterogeneity remain poorly signals demonstrates higher abundance compared to the direct detection of tumor-derived
understood. Methods: This study included advanced-stage NSCLC patients who un- molecules, such as circulating tumor DNA (ctDNA). This increased abundance allows for the
derwent both pre-treatment blood-based NGS (collected between October 2020 and extension of the screening time window, underscoring its potential for early cancer de-
February 2024) and tissue-based NGS. Intratumoral heterogeneity (ITH) was assessed tection. The present study aims to identify lung cancer-associated T cell receptor (TCR)
using the mutant-allele tumor heterogeneity (MATH) approach, with blood based NGS signatures, and develops a predictive model for to recognize lung cancer based on TCR
(bMATH). The highest allele frequency (HAF) was obtained from blood-based NGS. repertoire sequencing. Methods: Peripheral blood samples were collected from 2,699 lung
Survival analyses were conducted using Kaplan-Meier curves and Cox proportional cancer patients and 3,360 healthy individuals. The TCR repertoires were profiled using a
multiplex-PCR-based sequencing of the TCR-b chains. The frequency of TCR clonotypes
hazards models, focusing on the bTMB/tTMB ratio. Results: A total of 102 patients met
were calculated using MiXCR tools by aligning against human TCR-b gene segments. The
the inclusion criteria, 55 of whom had complete data for bTMB, tTMB, bMATH, and HAF.
cancer-enriched TCR sequences were identified by comprehensively considering the dis-
The median follow-up time was 12 months. Treatment regimens included ICI combined tribution and frequency of clonotypes in the comparison between the lung cancers and the
with chemotherapy (19 patients), ICIs alone (15), targeted therapy (16), and cytotoxic healthy controls. We proposed an lung cancer-enriched TCR score (LCS) to evaluate the risk
chemotherapy (5), with 33 receiving first-line treatment and 22 receiving second-line or for lung cancer based on the CDR3 sequences alignment. A robust machine learning model
beyond. The median interval from blood NGS to treatment initiation was 20 days (IQR was developed integrating multiple TCR repertoire characteristics and LCS. Results: The
8–28), and from tissue NGS to treatment initiation was 65 days (IQR 28–255). Mul- UMAP clustering based on TCR repertoire features revealed distinct TCR characteristics in
tivariable Cox proportional hazards analysis—adjusting for gender, smoking status, lung cancer patients compared to healthy individuals. 3,840 TCR clones were identified to be
stage, ECOG, regimen type, line of therapy, and numeric values of bMATH, bTMB, tTMB, significantly enriched in lung cancer cases. Among these, the CDR3b ’CATSRDTGGREKLFF’
and the bTMB/tTMB ratio—revealed that a higher bTMB/tTMB ratio was independently was identified as the most highly enriched clone specific to lung cancer patients.
associated with inferior overall survival (OS) (HR 1.16, 95% CI 1.03–1.30, p=0.01) but Furthermore, a significantly higher LCS was observed in lung cancers (p , 0.001). To
showed no significant difference in progression-free survival (PFS) (HR 1.10, 95% CI validate the utility of LCS, we applied the LCS measurement to an independent public TCR
0.94–1.29, p=0.23). The cutoff for the bTMB/tTMB ratio that best stratified OS was 0.81. dataset with 382 lung cancer patients, 195 healthy individuals, and 1,034 COVID-19 cases.
Patients below this cutoff experienced significantly longer OS (median OS 10 vs. The external validation demonstrated that lung cancers show a significantly increased LCS
49 months, HR 0.23, 95% CI 0.05–0.96, p=0.02). Additionally, a moderate positive compared with healthy controls (p , 0.001), while no significant difference in LCS between
correlation was observed between bMATH and the bTMB/tTMB ratio (Spearman’s COVID-19 cases and healthy controls (p = 0.25). The detection performance of model using
r=0.33, p=0.02). Conclusions: A higher bTMB/tTMB ratio was associated with poorer integrated TCR features and LCS demonstrated a sensitivity of 0.96 and a specificity of 0.95.
overall survival in advanced-stage NSCLC, highlighting its potential prognostic value. Conclusions: Cancer-related immune signals in peripheral blood can inform the anti-tumor
Moreover, the moderate correlation between bMATH and bTMB/tTMB suggests an responses. This study identified lung cancer-enriched TCR signatures, highlighting the
interplay between TMB and intratumor heterogeneity. Future studies are needed to potential as promising biomarkers for lung cancer detection. Clinical trial information:
confirm these findings and to explore their potential therapeutic implications. Research ChiCTR2200055761. Research Sponsor: None.
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
154s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2567 Poster Session 2568 Poster Session

Dietary compounds and patterns associated with immune checkpoint in- The role of lipid-laden Kupffer cells in immunosuppression and immuno-
hibitor (ICI) outcomes in advanced non-small cell lung cancer (NSCLC). First therapy response in MASLD-related hepatocellular carcinoma. First Author:
Author: Edmond Rafie, Research Center of the Centre Hospitalier de l’Université de Junzhe Jacky Zhao, Duke-NUS Medical School, Singapore, Singapore
Montreal, Montreal, QC, Canada Background: Hepatocellular carcinoma (HCC) related to metabolic dysfunction-associated
Background: The gut microbiome is a modulator of ICI activity. Diet is among the most steatotic liver disease (MASLD) is a rising global health burden. Despite systemic im-
important factors influencing the gut microbiome. We previously showed that high fiber was munotherapy being the first-line treatment for advanced HCC, clinical observation shows
not associated with outcome in NSCLC, in contrast to melanoma. However, the impact of that immune checkpoint inhibitors (ICI) offer lower benefit to patients with non-viral HCC,
dietary patterns and specific nutrients on ICI outcomes in NSCLC is unknown. Methods: At including MASLD-HCC. Macrophages, especially Kupffer cells (KCs), are the major PD-L1+
the CHUM Microbiome Centre, a nutritionist prospectively collected dietary history using a liver cells. With previous studies showing that more KCs are associated with poorer survival,
validated DHQ-II survey from 147 patients (pts) with advanced NSCLC treated with ICI alone we hypothesise that KCs in MASLD-related HCC show an immunosuppressive phenotype
or in combination with chemotherapy. Global dietary patterns and a systematic screen of 72 secondary to lipid accumulation, contributing to the poorer ICI responses in patients.
macro- and micronutrients (cut-offs defined by median) were examined for their association Methods: As proof-of-concept, we characterised macrophage phenotypes and lipid ac-
with progression-free survival (PFS) in univariable and multivariable cox-regression ana- cumulation in matched tumour and non-tumour tissues from HBV+ and non-viral HCC
lyses. Associations between diet and immune-related adverse events (irAE) were examined. patients (n = 6 / group). We next assessed the functional consequences of lipid loading using
In 69 pts, shotgun metagenomic sequencing (WMS) was performed on fecal samples to induced pluripotent stem cell (iPSC)-derived KCs in vitro, followed by validation in KC-
determine differential abundance of bacteria using linear discriminant analyses, heatmaps, containing iPSC- and patient-derived HCC organoids. We evaluated lipid accumulation,
and MaAsLin2. Results: Median age was 68, 46% were male. Median follow-up was paracrine signalling, transcriptomic changes, and cell-cell interactions in these organoids.
13 months. Total caloric intake adjusted for basal metabolic rate (Mifflin-St Jeor equation Results: In patient samples, macrophage lipid accumulation is associated with a PD-L1
using BMI and activity level) was not associated with PFS (p = 0.3). In univariable analyses, high, TREM2 high, KC-like phenotype in both non-tumour and tumour (Cohen’s d . 1.0,
the following nutrients were associated with improved PFS: vitamin K (HR 0.62, p = 0.03), fat power . 99%). In non-viral HCC samples, lipid-laden macrophages, with an M2-KC phe-
(HR 0.65, p = 0.04); while the following were associated with inferior PFS: starch (HR 1.61, p notype, are enriched twofold in non-tumour (Cohen’s d . 1.0, power = 72%) and tumour
= 0.03), carbohydrates (HR 1.56, p = 0.04), sucrose (HR 1.62, p = 0.03), and iron (HR 1.7, p = (Cohen’s d = 0.3, power = 6%). Light-sheet microscopy confirmed colocalization of lipids and
0.016). In a multivariable analysis examining all macro- and micronutrients and adjusting for PD-L1+ immune cells. Exposing iPSC-KCs to free fatty acids and IL-6 suppressed antigen
BMI, vitamin K intake was significantly associated with improved PFS (HR 0.60, 95% CI 0.37, presentation, while enhancing phagocytosis and T-cell exhaustion (p , 0.05) – effects not
0.97, p = 0.04). Fat-based diets such as keto-like diet (high fat, low starch) was associated reversed by ICI alone but alleviated when combined with tocilizumab or a CD36 inhibitor.
with improved PFS in univariable (HR, 0.47, p = 0.008) and multivariable analyses (HR 0.37, These changes are likely independent of lipophagy or FASN-mediated fatty acid synthesis.
95%CI 0.2, 0.68, p = 0.001). Compared to high starch diet, western diet (high fat, high starch) Our HCC organoid models preserve KCs and recapitulate their immunosuppressive phe-
was associated with increased risk of any grade irAE (24% vs 54%, respectively, p = 0.01). notype. scRNA-seq and CellPhoneDB analysis of HCC organoids highlighted significant KC-
WMS analyses revealed biologically relevant signals; fat-based diets were associated with hepatocyte crosstalk, mediated by IL-6, SPP1, and other cytokines, corroborated by Luminex
enrichment of favorable commensal bacteria such as Ruminococcus lactaris, Butyricimonas assay. Conclusions: These findings suggest that MASLD-associated lipid loading
faecihominis, Lachnospiraceae spp, with low fat associated with deleterious Veillonella promotes a distinctly immunosuppressive KC phenotype, contributing to diminished ICI
atypica. Starch-based diets were associated with high Prevotella spp. Sucrose-enriched responsiveness in HCC. Targeting the lipid-KC axis, for instance with IL-6 blockade or CD36
diets were enriched with Candidatus saccharibacteria, a known sucrose-fermenting bac- inhibitors, may help restore immune competence and improve ICI-based treatment out-
teria. Conclusions: Our results demonstrate the importance of diet on ICI outcomes in comes. Future studies, especially with a larger patient cohort and with our organoid
NSCLC and WMS results suggest this is mediated by the gut microbiome. Diet is a modifiable platform, should determine whether lipid-laden KCs can serve as a biomarker for ICI re-
lifestyle factor which may be targeted to improve ICI activity, meriting study in a randomized sponsiveness and further delineate the pathways driving their immunosuppressive be-
trial. Research Sponsor: Terry Fox Research Institue; Institut de Cancer de Montréal. haviour. Research Sponsor: National Medical Research Council, Singapore; MOH-
STaR21nov-0002; National Medical Research Council, Singapore; NMRC/OFLCG/003/2018;
National Research Foundation, Singapore; NRF-CRP26-2021RS-0001.

2569 Poster Session 2570 Poster Session

The economics of cancer immunotherapy: A five-year Medicare B expen- Phase I study of hV01, a recombinant human IL-21-expressing oncolytic
diture analysis of checkpoint inhibitors. First Author: Sharanya Tripathi, Saint vaccinia virus, as monotherapy in advanced solid tumors. First Author: Jian
Vincent Hospital, Worcester, MA Zhang, Fudan University Shanghai Cancer Center, Shanghai, China
Background: Immune checkpoint inhibitors (ICIs) have transformed cancer treatment, Background: Oncolytic viruses have shown an excellent safety profile and can initiate
prompting a comprehensive analysis of Medicare Part B spending trends from 2018 to antitumour immunity through in-situ tumor lysis and systemic immune response,
2022. Methods: We conducted a detailed examination of Medicare Part B claims data, thereby gaining significant attention in cancer immunotherapy. Here, we conducted a
focusing on spending per dosage unit, total expenditure, and beneficiary utilization phase I study of hV01, a genetically engineered oncolytic vaccinia virus expressing IL-21,
across multiple ICI drugs. Results: Our analysis revealed significant variations in ICI in patients with advanced malignant solid tumors. Methods: This open-label, single-
utilization and spending, with Keytruda (Pembrolizumab) emerging as the leading drug, dose escalation study evaluated four dosing levels (range, 1310^7 - 8310^8 PFU) of
totaling 4.94 billion and serving 67,022 beneficiaries in 2022. Notably, Opdivo (Nivo- intratumoral (i.t) injection of hV01, with a 28-day treatment cycle (NCT05914376). The
lumab) demonstrated substantial market presence with 1.85 billion in spending and primary aims were to determine the maximally tolerated dose (MTD), dose-limiting
26,957 beneficiaries. Price changes varied considerably, with Pembrolizumab toxicities (DLTs) and safety. Treatment response was evaluated by RECIST v1.1 criteria
experiencing a 12.9% price increase, Durvalumab rising 8.0%, and Atezolizumab in- on day 28 using a CT [Link] blood samples were analyzed for immune cell
creasing 10.1%. Interestingly, Libtayo showed a modest 1.5% price decrease. Other phenotypes with FACS and viral shedding with Q-PCR after the hV01 regimen.
significant drugs included Tecentriq (Atezolizumab) with 777.8 million in spending and Results: Thirteen patients with advanced solid tumors were enrolled; most had failed
12,812 beneficiaries, and Imfinzi (Durvalumab) with 562.7 million and 10,517 benefi- multiple prior therapies. Twelve patients completed at least one treatment cycle, with
ciaries. Conclusions: Medicare Part B spending on immune checkpoint inhibitors re- three in each dosing cohort. Six of the twelve patients (6/12, 50%) achieved stable
flects complex market dynamics, characterized by significant variations in drug pricing, disease (SD), while the rest had progressive disease (PD), evaluated at the end of the first
beneficiary utilization, and total expenditure, highlighting the evolving landscape of treatment cycle. Among the six patients with SD, one patient with pulmonary spindle cell
cancer immunotherapy. Research Sponsor: None. carcinoma in the 1310^7 PFU cohort had progression-free survival (PFS) for 138 days,
Cost trajectories and utilization patterns of eight immune checkpoint inhibitors (2018-2022). and one patient with dedifferentiated liposarcoma in the 6.0310^8 PFU cohort had a
Annual Growth
PFS for 206 days. One patient with nasopharynx cancer in the 8.0310^8 PFU cohort
Average Rate in Aver- Average achieved a partial response (PR) after four treatment cycles. No DLTs occurred during
Spending age Spending Average Sales the observation period of this study. The most commonly treatment-related adverse
per Dos- per Dosage Spending Price events (TRAE) were fever and anemia, which were reported as grade 1-2 events in eleven
Brand age Unit Unit (2018- Total Spending Total Benefi- per Benefi- (ASP)
Name Generic Name 2022 2022) 2022 ciaries 2022 ciary 2022 2022 patients (11/13, 84.6%) and six patients (6/13, 46.2%), respectively. The only grade 3
Bavencio Avelumab $85.86 2.10% $111,862,815 1,591 $70,310 $87.91
event was decreased lymphocyte count in one patient in the 6.0310^8 PFU cohort. FACS
Imfinzi Durvalumab $75.86 1.60% $562,741,221 10,517 $53,508 $77.53 analysis found a decrease in T cells and natural killer (NK) cells in the peripheral blood in
Jemperli Dostarlimab- $216.09 $4,502,157 79 $56,989 $217.87 all patients on the day following hV01 administration, and later, the levels of T cells and
Gxly NK cells gradually rebounded. Notably, four patients had significantly higher (more than
Keytruda Pembrolizumab $52.18 4.70% $4,935,971,049 67,022 $73,647 $53.42
Libtayo Cemiplimab- $27.02 0.10% $203,832,304 3,187 $63,957 $27.45 doubled or tripled) levels of NK cells compared to their baseline on day 15 after the hV01
Rwlc regimen. Conclusions: Single-dosing of hV01 per treatment cycle was well tolerated
Opdivo Nivolumab $28.78 4.10% $1,849,938,540 26,957 $68,626 $29.43 and safe. hV01 i.t. injection demonstrated primary efficacy in patients with advanced
Opdualag Nivolumab and $176.43 $42,513,402 680 $62,520 $180.64
relatlimab- tumors. The data support further study of multiple-dosing regimens and future trials
rmbw evaluating the benefits of combining hV01 with other antitumor therapies. Clinical trial
Tecentriq Atezolizumab $79.07 1.50% $777,758,575 12,812 $60,705 $80.79 information: NCT05914376. Research Sponsor: Hangzhou ConVerd Co., Ltd.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 155s

2571 Poster Session 2572 Poster Session

Influence of salmonella-IL2 in combination with FOLFIRINOX on overall and Initial safety and efficacy results from a first-in-human, phase 1/2 study of
progression-free survival in stage IV metastatic pancreatic cancer. First SAR445877, an anti-PD-1/IL-15 fusion protein, for patients with advanced
Author: Daniel Saltzman, Salspera Inc, Oakdale, MN solid tumors. First Author: Aung Naing, The University of Texas MD Anderson Cancer
Background: Salmonella-IL2 is an attenuated Salmonella Typhimurium strain carrying Center, Houston, TX
the human gene for IL-2. When orally administered, the bacterium colonizes tumors and Background: SAR445877 is a fusion protein of an PD-1 antibody combined with a detuned
locally releases IL-2, triggering immunologically-mediated tumor cell killing without IL-15, designed to selectively expand and activate CD8+ T and NK cells expressing both PD-1
untoward side effects. Salmonella-IL2 was tested in a phase 2 clinical trial in which and IL-2/15Rbg. Preclinical studies demonstrated the efficacy of SAR445877 in neoplastic
patients with stage IV metastatic pancreatic cancer were treated with standard of care models. Here, we present initial safety and efficacy observations from a first-in-human, dose
chemotherapy (SOC) combined with Salmonella-IL2. Methods: A Health Canada and escalation of SAR445877 monotherapy in patients (pts) with advanced solid tumors.
local IRB approved, non-randomized, two-arm human study evaluated the combination Methods: This open-label, multicenter, Phase 1/2 study in adult pts with any type of
of Salmonella-IL2 with standard of care (SOC) chemotherapy; Arm One patients received measurable, advanced unresectable or metastatic solid tumors (NCT05584670) comprised
Salmonella-IL2 plus FOLFIRINOX (FFX). Arm Two patients received Salmonella-IL2 plus two parts: dose escalation (Part 1) and dose expansion (Part 2). In Part 1, SAR445877 was
Gemcitabine/nab-Paclitaxel (GEM/nabP). Overall survival (OS), progression-free survival administered intravenously at two dosing schedules (Q2W and QW). Pts with advanced solid
(PFS), safety, and biomarker data in each arm were compared to corresponding values tumors that do not typically respond or were resistant/refractory to immune checkpoint
for reference patients (SOC Controls) receiving care at the study site in the four years inhibitors (ICI), and with at least 1 measurable lesion per RECIST 1.1, were eligible. Tumor
biopsy was performed at baseline and on treatment. The primary objective for Part 1 was
preceding the clinical trial (2016 to 2020). Four patients with stage IV pancreatic cancer
safety. Secondary objectives included efficacy, pharmacokinetics, pharmacodynamics, and
were treated via Salspera’s Expanded Access Program (EAP) using the Arm One regimen.
immunogenicity. Results: Thirty-two pts (Q2W) and 17 pts (QW), respectively, were en-
Results: In total, 34 patients (30 in the trial, 4 via EAP) were enrolled: 26 received
rolled. Median lines of prior therapy were 3 for both schedules. The Q2W schedule was
Salmonella-IL2 with FOLFIRINOX (average age 58.7 years, range 32-74, 54% born fe- tested at 6 dose levels (DLs) and QW schedule at 3 DLs. Median exposure to SAR445877 was
male) and eight received Salmonella-IL2 with GEM/NabP (average age 68 years, range around 9 weeks (Q2W range: 2–55 weeks; QW range: 1–46 weeks). All pts reported at least
56-82, 54% born female). SOC Control patients comprised 37 administered FOLFIRINOX one treatment-emergent adverse event (TEAE). Treatment-related AEs (TRAEs) were re-
(average age 58 years; range 33-75 years; 46% born female) and 31 given GEM/nabP ported in 47 pts (Grade $3: 12 pts [Q2W]; Grade $3: 5 pts [QW]). Most common TRAE was
(average age 65.1 years; range 45-84 years; 42% born female). Patients receiving more cytokine release syndrome (CRS), which was mainly Grade 1 or 2. Six pts discontinued
than five doses of Salmonella-IL2 with FOLFIRINOX (n = 20) had a mPFS of 15 months treatment due to any TEAEs. DLTs occurred in 4 pts in the Q2W cohort (n = 1 each metabolic
vs. 5.8 months in control patients who had received only FOLFIRNOX (p , 0.0001, 95% acidosis, pneumonia, hyperbilirubinemia, and GI hemorrhage) and in 2 pts in the QW cohort
CI, HR 0.3, Concordance Index 0.64). For those same compared cohorts, the mOS was (both CRS). Serious TEAEs were reported in 23 pts (Q2W) and 7 pts (QW). All toxicities were
20.3 months vs. 11.5 months (p = 0.07, 95% CI. HR 0.59, Concordance Index 0.59). Only manageable/reversible, and no TEAEs leading to death were observed. Confirmed partial
eight patients were enrolled in the GEM/NabP Arm thus limiting useful conclusions. response was reported in 5 pts (Q2W) and in 2 pts (QW) bearing melanoma, CRC, SCC of the
Overall, 41 serious adverse events were noted and attributed to SOC chemotherapy scalp, penile cancer, adnexal carcinoma, urothelial carcinoma, and myxofibrosarcoma, with
agents but none to Salmonella-IL2. Conclusions: Addition of Salmonella-IL2 to FOL- benefits lasting . 1 year. Five of the 7 pts had progressed on prior immunotherapy. Stable
FIRINOX for treating stage IV pancreatic cancer is associated with increased mPFS and disease $ 6 months was observed in 3 pts (Q2W) and 3 pts (QW). Antidrug antibodies
mOS. A multicenter, randomized, phase III trial is warranted. Clinical trial information: (ADAs), detected in 23 pts (Q2W) and 13 pts (QW), did not correlate with clinical benefit or
NCT04589234. Research Sponsor: None. toxicity. A trend of dose dependent increase of cytokine (IFNg, TNFa, IL-6, IL-8, IL-10) and
chemokine (CCL2, CXCL10, MIP1a, MIP1ß) release were detected at both dosing schedules.
Conclusions: SAR445877 monotherapy demonstrated a tolerable safety profile and
promising antitumor activity in pts with advanced solid tumors unresponsive or resistant to
ICI. Clinical trial information: NCT05584670. Research Sponsor: Sanofi.

2573 Poster Session 2574 Poster Session

SLAMF8 as a potential therapeutic target for modulating tumor-associated Preclinical development of GNTbm-38, a novel class I histone deacetylase
macrophages in colorectal cancer. First Author: Han Xingzhi, The Comprehensive inhibitor, while combined with anti-VEGFR TKI or anti-PD-1 Ab: Assessment
Cancer Center of Drum Tower Hospital, The Affiliated Hospital of Nanjing University of immune activation and immune memory in cancer immunotherapy. First
Medical School, Nanjing, China Author: Jia-Shiong Chen, New Drug Research and Development Center, Great Novel
Background: Reprogramming or repolarizing tumor-associated macrophages (TAMs) Therapeutics Biotech & Medicals Corporation (GNTbm), Taipei City, Taiwan
has emerged as a novel strategy in tumor immunotherapy, leveraging their remarkable Background: Several clinical trials explored ICI-based combinations in MSS mCRC patients,
plasticity. Our previous studies have demonstrated that SLAMF8, predominantly and the promising outcomes are lacking. Histone deacetylase inhibitors (HDACis) for cancer
expressed in macrophages, is a promising biomarker for predicting the efficacy of therapy may boost antitumor immune activity, reduce immunosuppressive cells, and play a
immune checkpoint inhibitor (ICI) therapy in colorectal cancer (CRC). This study aims to crucial role in controlling tumor progression. Therefore, rational drug combinations, con-
elucidate further the regulatory role of SLAMF8 in TAMs and its influence on the tumor taining class I HDACi or other immune-modulating drugs, may provide opportunities in
immune microenvironment. Methods: In vitro, M2 macrophage and TAM models were immunotherapy. Methods: The activities of GMTbm-38 were assessed in vitro, including H3
established to investigate the regulatory role of SLAMF8 in macrophage immuno- acetylation and cancer cell growth inhibition, etc. The murine colon cancer CT-26 model was
phenotypic transition and its impact on CD8+ T cell function using qRT-PCR, flow used to test antitumor efficacy in wild type and transgenic humanized PD1/PD-L1 mice.
cytometry, and co-culture assays. Additionally, pathway enrichment analysis of RNA-seq GNTbm-38 was combined with an anti-VEGFR TKI or murine/human PD1 antibody to test the
data and western blotting were conducted to elucidate the underlying molecular antitumor synergistic effect. RNA-seq, flow cytometry, and IHC were performed to illustrate
mechanisms. In vivo, SLAMF8-specific small interfering RNA (siSLAMF8) was utilized to the potential mechanisms. Results: GNTbm-38 induced histone 3 acetylation and inhibited
the cell growth of varieties of human cancer cells. By using CT-26 model, GNTbm-38
inhibit SLAMF8 expression in subcutaneous colorectal cancer (CRC) tumor models.
showed a superior efficacy profile in WT mice compared to immune-deficient mice. The
Subsequent observations included tumor growth monitoring and analysis of immune cell
antitumor activity related to induced immune activation and immune memory was de-
infiltration via flow cytometry. Furthermore, two subcutaneous tumor models, one
pendent on CD8+ T cell activation. Treatment with GNTbm-38 showed an increased number
sensitive and one resistant to PD-1 therapy, were constructed to explore potential of intratumoral CD8+ CTLs, a decreased number of MDSCs, and induced normalization of
synergistic effects between SLAMF8 inhibition and anti-PD-1 treatment. tumor vessels. GNTbm-38 substantially induced the expression of IFN-g response genes
Results: In vitro studies reveal that SLAMF8 facilitates the polarization of macrophages and enhanced antigen processing and presentation signatures. GNTbm-38 acts as a TME
toward the M2 phenotype, contributing to the immunosuppressive characteristics of reprogramming regulator in immunotherapy. When combined with TKI, GNTbm-38 sig-
TAMs and dysfunction of CD8+ T cells. Inhibiting SLAMF8 in vivo significantly delayed nificantly improved tumor response rate and survival rate through synergistic effect by
colorectal cancer (CRC) progression. Flow cytometry analysis confirmed that the in- normalizing tumor vessels, increasing tumor antigen presentation, increasing activated
filtration of anti-tumor TAMs (CD86+ F4/80+) and cytotoxic CD8+ T cells (IFNg+ GZMB+ CD8+ T cell infiltration into tumors, inducing memory T cell persistence, and inhibiting
CD8+ T cells) was augmented, while the infiltration of pro-tumor TAMs (CD206+ F4/80+) mobilization of immunosuppressive cells into tumors. Treatment with GNTbm-38 plus anti-
and exhausted CD8+ T cells (PD1+ LAG3+ CD8+ T cells) was reduced in tumors treated PD-1 Ab in the CT-26 model showed greatly improved tumor response rate and survival rate
with SLAMF8-specific small interfering RNA (siSLAMF8). Additionally, targeting SLAMF8 with a strong synergistic effect. Furthermore, in B-hPD-1/hPD-L1 mice (humanized model)
enhances the efficacy of anti-PD-1 therapy. Mechanistically, Western blotting experi- subcutaneously injected with B-hPD-L1 CT-26 cells, treatment of pembrolizumab and
ments confirmed that the phosphorylation levels of key molecules in the PI3K/AKT and GNTbm-38 resulted in a 46.5% inhibition on tumor growth. Therefore, our data provided a
JAK/STAT3 signaling pathways were significantly increased in SLAMF8-overexpressing strong rationale to explore the combination of GNTbm-38 with anti-VEGF TKI with or without
macrophages. Conclusions: Inhibition of SLAMF8 delays tumor growth, enhances the ICI. Conclusions: Collectively, our data show that GNTbm-38 exhibits markedly superior
sensitization to ICI therapy, and reverses the immunosuppressive microenvironment by pharmacokinetics, tolerability, and efficacy in animal models. GNTm-38 has been shown to
modulating TAMs via the PI3K/AKT and JAK/STAT3 pathways in CRC. These findings display powerful induction of immune activation and immune memory in combination
reveal the potential of SLAMF8 as a therapeutic target for immunotherapy focused on therapy with TKI/ICI against colon CT-26 cold tumor. Research Sponsor: Great Novel
TAMs. Research Sponsor: National Natural Science Foundation of China; 82303970. Therapeutics Biotech & Medicals Corporation (GNTbm).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
156s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2575 Poster Session 2576 Poster Session

Inhaled KB707, a novel HSV-based immunotherapy, as a monotherapy in A generative model for the design of novel inhibitors targeting the PD-1/PD-
patients with advanced solid tumor malignancies affecting the lungs: Effi- L1 pathway. First Author: Juan Velasco, Yale University, New Haven, CT
cacy and safety results from a phase 1/2 study. First Author: Wen Wee Ma, Background: The programmed cell death-1/programmed cell death ligand-1 (PD-1/PD-
Cleveland Clinic Taussig Cancer Institute, Cleveland, OH L1) signaling pathway plays a pivotal role in tumor immunosuppression. However, the
Background: The development of potent anti-tumor cytokines has been hindered by the design of de novo molecules with precise pharmacological and molecular properties
systemic toxicity of intravenous administration. KB707 is a novel gene therapy designed to remains a resource-intensive and financially demanding endeavor. It is hypothesized
deliver high doses of cytokines to the local tumor microenvironment. The agent is a that generative models trained on molecular graph encodings can design novel inhibitors
replication-defective herpes simplex virus type 1 (HSV-1)-based vector engineered to targeting the PD-1/PD-L1 pathway. Objective: This study aims to develop a generative
deliver human interleukin (IL)-12 and IL-2 with complementary anti-tumor effects. This model capable of designing novel, orally bioavailable inhibitors of the PD-1/PD-L1
replication-defective vector platform enables repeated dosing without significant toxicity pathway. Methods: A large language model was pre-trained on 1 million chemical
or clinically relevant immunogenicity while allowing for localized, sustained IL-12 and IL-2 structures derived from the ChEMBL database. Each structure was represented using the
delivery to induce both innate and adaptive anti-tumor immunity. This study evaluates Simplified Molecular Input Line Entry System (SMILES) strings, which were further
whether KB707 administered by inhalation will deliver efficacious dose to the lung while tokenized into discrete atomic and functional group-level tokens. The model employs an
minimizing systemic exposure in advanced solid tumor patients with predominantly lung Average-Stochastic Gradient Descent Weight-Dropped Long Short-Term Memory (AWD-
disease. Methods: KB707-02 is a Phase 1/2, open-label, multicenter, dose escalation (3+3 LSTM) architecture. Transfer learning was applied to fine-tune the pre-trained language
design) and expansion study of inhaled KB707 (NCT06228326). Eligible patients (pts) with model on the target chemical structures, enabling domain-specific adaptation for the
at least one measurable lung lesion at screening and histological confirmation of advanced desired application space. Results: The model demonstrated robust performance in
solid tumor malignancy in the lungs received nebulized KB707 once weekly for up to generating chemically valid, unique, and novel inhibitors targeting the PD-1/PD-L1
3 weeks followed by treatment every 3 weeks. The primary objective is to assess safety and pathway. It achieved a validity rate of 97%, a uniqueness rate of 96%, a novelty rate of
tolerability, with a secondary objective to evaluate preliminary efficacy per RECIST 1.1. 95%, and a diversity score of 76.04%. Additionally, the generated molecules exhibited
Results: As of 08 Jan 2025, a total of 39 pts were enrolled and received at least one dose favorable physicochemical properties, including a logarithm of the partition coefficient
of inhaled KB707. Monotherapy dose escalation and expansion was completed. The doses (LogP) of 4.52, atopological polar surface area (TPSA) of 113.06 Angstrom squared, an
evaluated were 108 and 109 PFU and the maximum tolerated dose (MTD) was not reached.
average of 9.62 rotatable bonds, 2.77 hydrogen bond donors, and 6.79 hydrogen bond
Treatment-emergent adverse events have been consistent with known adverse event
acceptors. Conclusions: A generative model was developed to design novel, orally
profiles of IL-2 and IL-12. The majority of treatment-related adverse events have been mild
bioavailable inhibitors of the PD-1/PD-L1 pathway. This approach provides an efficient
to moderate in severity and transient, with no Grade 4 or 5 adverse events observed. The 11
and automated tool for designing de novo molecules with precise molecular and
response-evaluable NSCLC pts were of advanced age (median 71 [54-77] years old) and
heavily treated (4 median lines of prior therapies; all received at least 1 line of prior
pharmacological properties, potentially accelerating drug discovery in immuno-oncol-
immunotherapy). The ORR was 27% (3/11) and DCR was 73% (8/11) with 7 out of 11 pts ogy. Research Sponsor: None.
remaining on study. The response rate of the target lesions in the lungs was 36%. The
median duration of response was not reached; treatment duration ranged from 10.3 to
33.3 weeks. Conclusions: KB707 administered by inhalation was safe and well tolerated.
The MTD was not reached, and the monotherapy recommended Phase 2 dose is 109 PFU.
Single agent anti-tumor effects were observed, including in heavily treated NSCLC patients.
The study has been expanded to evaluate the combination of inhaled KB707 plus
pembrolizumab, with or without chemotherapy, in advanced NSCLC pts. Enrollment in
these combination expansion cohorts is ongoing. Clinical trial information: NCT06228326.
Research Sponsor: Krystal Biotech Inc.

2577 Poster Session 2578 Poster Session

Completed phase 1a dose escalation study of the first oral ENPP1 inhibitor Deep learning–powered H&E whole-slide image analysis of endothelial cells
RBS2418 immunotherapy in subjects with metastatic solid tumors. First to characterize tumor vascular environment and correlate treatment out-
Author: Thomas Urban Marron, Division of Hematology and Medical Oncology, Tisch come to immunotherapy. First Author: Seungeun Lee, Lunit Inc., Seoul, South Korea
Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY Background: Beyond their vascular function, endothelial cells (ECs) regulate tumor
Background: ENPP1 clears cGAMP and ATP in the tumor microenvironment (TME). Its growth through recently described angiocrine signaling, influencing cancer progression
expression is associated with poor prognosis in cancer and development of metastases. and treatment outcomes. Here, we applied a deep learning model, which we validated
ENPP1 inhibition can protect cGAMP and ATP from hydrolysis and reduce adenosine levels in using spatial transcriptomics data, to a pan-cancer dataset to analyze the EC distribution
the TME. These immune modulators are known to activate APCs and increase T-cell infil- associated with the response to immuno-oncology (IO) treatments. Methods: An AI-
tration, promoting anticancer immunity. RBS2418 is a potent and selective oral first-in-class powered H&E analyzer, Lunit SCOPE IO, quantifies tumor microenvironment EC density
inhibitor of ENPP1. In this open-label, multi-site Phase 1a/b study, safety and efficacy of and tumor-infiltrating lymphocytes (TILs) density in cancer epithelium and stroma. We
RBS2418 is being evaluated as monotherapy and in combination with pembrolizumab in validated AI-powered cell type prediction by evaluating cell-specific gene expression
advanced/metastatic solid tumors. Methods: The phase 1a dose escalation part comprised through spatial transcriptomics (10x Xenium). 7,467 pan-carcinoma samples from The
100, 200, 400 and 800 mg BID dose levels of RBS2418 alone or in combination with Cancer Genome Atlas (TCGA) were analyzed for EC distribution and overall survival (OS).
pembrolizumab (200 mg IV q3w) in patients who have failed all approved treatments including From a previously described multi-center, multi-national pan-cancer cohort (Pan-IO, Shen
immunotherapy using a 3+3 study design. Study objectives were to evaluate safety, phar- et al JITC 2024), 1,654 patients were analyzed for IO treatment response. Results: We
macokinetics (PK), pharmacodynamics (PD), and clinical outcomes. Tumor and blood validated our AI prediction of cell types by demonstrating consistency with known cardinal
samples were collected to determine PK/PD and immune profiles using LC/MS, IF, IHC, TCR gene expressions from spatial transcriptomic results, where 77.5% of AI-predicted en-
and RNAseq analyses. Results: Dose escalation is complete, andRBS2418 was safe and well dothelial cells (ECs) expressed VEGFR2 (compared to 6.9% in tumor cells (TCs)), while
tolerated at all dose levels with no DLTs (n = 24). Treatment durations range from 1 to
VEGFA expression was 2.8 times higher in TCs. Consistent with previous studies, EC
15 months to date, and no treatment-related grade 3 adverse effects (TRAEs) or serious AEs
density was highest in RCC and HCC, while lowest in pancreatic adenocarcinoma,
(SAEs) have been observed. A total of 21 grade 1 or grade 2 TRAEs were reported in 9 (37.5%)
melanoma, and cholangiocarcinoma. While EC and TIL density were not correlated pan-
subjects; the most common TRAE was grade 1 fatigue. Median plasma concentrations of
cancer (r = 0.08), exceptions were head and neck cancer (r = 0.45, p , 0.001) and
RBS2418 increased in a dose-proportional manner. Plasma and tumor concentrations of
RBS2418 were maintained above the ENPP1 inhibition EC90 level in all patients and at all pancreatic cancer (r = 0.44, p , 0.001). In the TCGA cohort, the high EC density was
dose levels tested. Of 19 patients with adequate baseline tissue, pre-treatment ENPP1 and associated with prolonged OS (HR 0.84, p , 0.001). In contrast, within the Pan-IO cohort
cGAS co-expression (EG+ phenotype, n = 8) in tumors correlated with RBS2418 treatment- (HR 1.26, p , 0.001) and its lung cancer subgroup (HR 1.26, p , 0.001) high EC density
associated immune activation and significantly improved progression-free-survival (PFS) as was associated with shorter progression-free survival (PFS) on treatment with IO
compared to EG- phenotype at baseline (n = 11). A switch from “cold” tumor to “hot” tumor monotherapy. Moreover, the predictive impact of EC density varied by TIL status. In the
phenotype was consistently observed in EG+ subjects. Disease control rate (DCR) was 75% (6/ Pan-IO cohort, EC density predicted PFS in both TIL-high (HR 1.40, p , 0.001) and TIL-low
8) in EG+ and 9% (1/11) in EG- subjects. Conclusions: The Ph1a dose escalation study with cancers (HR 1.33, p , 0.001). Among four groups divided by median values, the EC-low
oral RBS2418 alone, and with pembrolizumab has been completed. All doses were safe and and TIL-high group showed the longest PFS (median PFS of EC/TIL high/low: 2.5m, high/
well tolerated with no grade 3 TRAEs, SAEs or DLTs. RBS2418 plasma concentrations en- high: 3.2m, low/low: 3.6m, low/high: 5.6m, p , 0.001). Conclusions: EC distribution
abling full cGAMP protection was observed in all patients at all dose levels. Immune activation varied among cancer types, and importantly high EC content strongly correlated with poor
and clinical benefits strongly correlated with EG+ phenotype. RBS2418 achieved Phase 1a IO monotherapy response, including NSCLC. These findings support exploring immu-
goals of safety, PK, PD, and target engagement and showed significant treatment benefits in notherapy combination strategies that include anti-endothelial approaches, which could
advanced metastatic cancer patients. The results support further development of RBS2418. encompass the emerging class of PD-1/VEGFR bispecifics, for tumors with high EC
Phase 1b dose expansion is in progress and the first Phase 2a study has been initiated for the content. Early evidence was observed for this for HCC (Chon et al ASCO GI 2024), with a
treatment of mCRC. Clinical trial information: NCT05270213. Research Sponsor: None. hazard ratio of 0.62 for high EC HCC for atezolizumab/bevacizumab. Research Sponsor:
None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 157s

2580 Poster Session 2581 Poster Session

SECN-15: A novel treatment option for patients with checkpoint inhibitor– Utilizing targeted intra-tumoral hyperthermia as an immunotherapy in
resistant tumors by targeting Neuropilin-1 with antisense oligonucleotides. immunogenically ‘cold’ tumor models. First Author: Carman Giacomantonio,
First Author: André Maaske, Secarna Pharmaceuticals, Planegg-Martinsried, Germany Dalhousie University, Halifax, NS, Canada
Background: SECN-15 is a high-affinity antisense oligonucleotide (ASO) targeting Background: Hyperthermia is an established adjunct in multimodal cancer treatments,
Neuropilin-1 (NRP1), a transmembrane protein that exerts a variety of protumorigenic with mechanisms including cell death, immune modulation, and vascular changes.
functions by interacting with various receptors and ligands. NRP1 contributes to an Traditional hyperthermia applications are resource-intensive and often associated with
immunosuppressive microenvironment, tumor growth, metastasis, and neoangio- patient morbidity, limiting their clinical accessibility. Gold nanorods (GNRs) offer a
genesis. The recent success of bispecific antibodies targeting PD-1/PD-L1 and VEGF precise, minimally invasive alternative by leveraging near-infrared (NIR) light to deliver
such as ivonescimab has highlighted the potential of combining checkpoint inhibitors targeted hyperthermia therapy (THT). THT induces controlled tumor heating, promoting
with anti-angiogenic approaches. Consequently, NRP1 represents a highly attractive immunogenic cell death (ICD) and modulating the tumor microenvironment (TME) to
target for treating patients with tumors that are resistant to or insufficiently responsive enhance immune engagement. This study explores the synergistic potential of GNR-
to checkpoint inhibitor therapies, such as gastric (GC) and breast cancer, where high mediated THT with immunotherapies in immunogenically ‘cold’ tumors to achieve
NRP1 expression correlates with poor prognosis. Methods: NRP1-specific locked durable anti-tumor immunity. Methods: GNRs from Sona Nanotech Inc.™ were intra-
nucleic acid (LNA)-modified ASOs were identified using our OligoCreator platform. We tumorally injected and activated using NIR light to induce mild hyperthermia (42–48°C)
assessed in vivo anti-tumor efficacy after systemic administration in various mouse for 5 minutes. Tumor responses were analyzed for cell death pathways and immune
tumor models as monotherapy and in combination with checkpoint inhibitors. Target modulation. The immunogenic effects of THT were assessed alone and in combination
downregulation was analyzed in tissues and in plasma by measuring soluble NRP1. Cell with intratumoral interleukin-2 (i.t. IL-2) or systemic PD-1 immune checkpoint blockade.
composition and transcriptome changes in tumors were analyzed using flow cytometry Immune cell infiltration, gene expression changes, and tumor growth kinetics were
and RNA sequencing. Exaggerated pharmacology was investigated in a 28 non-GLP evaluated. Results: THT reduced tumor burden through cell death mechanisms, in-
tolerability study in mice. In silico analyses of patient transcriptomics data were cluding upregulated ICD marked by calreticulin exposure within 48 hours. By 48 hours,
performed to prioritize indications for the upcoming Phase I/II clinical trial. CD45+ immune cell levels were increased, including increased levels of immunosup-
Results: Systemic administration of NRP1-specific ASOs resulted in robust knockdown pressive M2 macrophages. While THT led to innate immune cell stimulations highlighted
in tumors across various cell types, including macrophages and T cells. Soluble NRP1 by gene expression upregulation in the STING cGAS pathway and enhanced M1 and
levels were reduced in treated animals, serving as a target engagement biomarker. dendritic cell levels, tumor regrowth was observed within six days post-treatment. To
Tumor growth was delayed in the monotherapy setting, with several animals showing enhance THT’s immunogenic effects, the therapy was combined with intratumoral
complete responses. Combining NRP1-specific ASOs with checkpoint inhibitors en- interleukin-2 (i.t. IL-2) or systemic PD-1 immune checkpoint blockade. Sequential
hanced efficacy in models where checkpoint inhibitors alone had limited activity. administration of i.t. IL-2 post-THT induced robust CD8+ T-cell infiltration and led to
Transcriptomic analysis showed upregulation of inflammatory genes and down- sustained tumor regression in both treated and distant tumors, accompanied by the
regulation of extracellular matrix organization genes. No adverse effects were observed emergence of memory T cells. However, IL-2-induced immunosuppressive T-reg pop-
from persistent NRP1 downregulation in non-tumor-bearing mice. In silico analyses ulations were also sustained to tumor endpoint suggesting that therapy could be further
revealed that NRP1 expression is negatively associated with survival and increases in enhanced. Additionally, PD-1 expression, which was upregulated in CD8+ T cells by THT,
advanced GC stages. GC was selected as one of the priority indications for the upcoming was targeted with systemic PD-1 inhibition, further augmenting immune engagement
Phase I/II clinical trial to investigate SECN-15’s safety and efficacy as monotherapy and within the TME. Conclusions: These combinatory treatments demonstrated synergistic
in combination with PD-1 blocking antibodies. Conclusions: Targeting NRP1 with ASOs effects, promoting durable anti-tumor responses and immune memory. Collectively,
is a promising therapeutic strategy for solid cancers. Combining NRP1 ASOs with ICIs GNR-mediated THT effectively reduces tumor burden and remodels the TME, poten-
significantly enhances anti-tumor efficacy, potentially overcoming current ICI therapy tiating systemic immunity and enhancing the impact of complementary
limitations. IND-enabling studies are underway to advance SECN-15 into clinical immunotherapies. Research Sponsor: None.
development. Research Sponsor: None.

2582 Poster Session 2583 Poster Session

Employing novel pan-cancer targets for immunotherapy in leukemias and Prevalence of the HPV, EBV, and TTV viral RNA in the plasma of patients with
solid tumors. First Author: Ashley Varkey, Hackensack University Medical Center, solid and hematologic neoplasms and the detection of a specific immune
Hackensack, NJ signature. First Author: Gustavo Rivero, Tampa General Hospital, Tampa, FL
Background: Acute myeloid leukemia (AML) and many solid tumors are difficult to treat. Tumor- Background: Epstein2Barr virus (EBV) and human papillomavirus (HPV) are consid-
associated protein targets that are the focus of cancer immunotherapy research are prone to on- ered human oncoviruses. In contrast, the torque teno virus (TTV) is not associated with
target, off-tumor toxicity and antigen-negative relapse due to mutation or downregulation. Targeting any disease but its detection in circulation is associated with the status of the immune
cancer-specific markers less susceptible to resistance is key for safer therapies. Our research explores
high mannose (Man 9) oligosaccharides and phosphatidylserine (PS) as non-protein targets. Man9
system. In this study, we examine the prevalence of active EBV, HPV and TTV viral RNA
glycans are absent on healthy cells but are present in cancers like AML, breast, colon, and lung. PS, in patients treated for solid tumors or hematologic neoplasms. In addition, we compared
exposed during malignant transformation, is found on colon, prostate, and brain tumors. Methods: We differential expression of selected immune and inflammatory biomarkers between Virus
have engineered trispecific T cell engagers (Man9/PS/CD3) to target Man9 and/or PS-positive cancers. positive (V+) and Virus negative (V-) cases using peripheral blood cell-free RNA (cfRNA).
We tested solid tumor cell lines (pancreatic, lung, colorectal) via flow cytometry and found that the Methods: cfRNA was extracted from the peripheral blood of 581patients with a di-
dual affinity molecule (Man 9 x PS) had high binding to many solid tumors (Table 1). We assessed their agnosis of hematologic neoplasms and 558 patients with solid tumor. cfRNA was
efficacy in vivo and specificity using glycan microarray, and immunohistochemistry to confirm tumor sequenced by NGS using a targeted RNA panel of 1600 genes and the viral RNA of TTV,
specificity and predict favorable safety profiles. Results: Flow cytometry showed that our therapeutic
EBV and HPV. Two thirds of the samples were used for training and one third for testing
molecules specifically bind to AML cells and various solid tumors while sparing healthy tissues. Glycan
microarrays confirmed selective binding to abnormal glycans on cancer cells. Immunohistochemistry machine learning (ML) system (Bayesian/Random Forest) and exploring the presence of
of FFPE tissues indicated tumor specificity and enrichment on cancer stem cells. In vitro studies specific inflammatory profiles distinguishing V+ from V- patients. Results: RNA testing
(coculture of luciferase-transduced target cells with activated CD8+ T cells in the presence of absence was selected to ensure that only active and proliferating viruses were detected. We
of the T cell engager) demonstrated strong anti-leukemia activity against AML cell lines, with IC50 detected TTV in 52/1139 (4.6%), EBV in 251/1139 (22%), and HPV in 68/1139 (6.0%). TTV
values of 5–10 pM. In vivo studies in human CD3 transgenic mice treated with intravenous doses of with EBV codetection was observed in 11 samples (1%), and with HPV in 4 patients
VTRU200 (Man9 x PS x CD3) showed significant therapeutic responses, based on in vivo biolumi- (0.4%). Co-detection of EBV with HPV was observed in 13 patients (1.1%). Using 90
nescence imaging. Conclusions: Our data supports Man9 and PS as promising non-protein targets for biomarkers in ML algorithm can reliably distinguish V+ from V- with AUC of 0.725 (CI:
pan-cancer immunotherapy. The dual targeting approach with T cell engagers reduces on-target, off-
tumor toxicity and antigen-negative relapse, advancing a first-in-class Man9 x PS x CD3 trispecific
0.658-0.791) in the testing set. Significantly higher levels of B-cell markers are noted in
T cell engager. We have also designed and validated a bispecific (Man9 x PS) chimeric antigen receptor V+ patients. PD-L1 mRNA was significantly (P ,0.001) higher in V+ patients, which
(CAR) and research is ongoing for CAR-T therapy for pancreatic cancer. With IND-enabling studies suggests that these patients may be more responsive to checkpoint immune therapy.
underway, we aim to advance this breakthrough immunotherapy for AML and other cancers, targeting CD70 is also detected at high level in V+ patients (P,0.0001). Upon comparing between
an IND submission within 18 months. Research Sponsor: Vitruviae; Alex’s Lemonade Stand Foun- the V+ groups (TTV, EBV, and HPV), there was no statistical difference between the three
dation; Hyundai Hope on Wheels; National Cancer Institute. groups after adjusting for multiple testing. However, some difference in cytokine levels
Cell type Man9/PS positivity (%) Sample size was noted between TTV-positive patients and HPV-positive patients. CD36, IFNA2 and
Mouse AML (cell line) 75-80 2
IL17A were higher in TTV-positive cases as compared with HPV-positive cases (P-value
Human AML (cell line) 54-100 9 0.0003, 0.005 and 0.004, respectively). Conclusions: Globally, detectable active viruses
Human adult AML (primary) 35-98 8 in plasma of patients with cancer is relatively high (29%) and this detection is associated
Human pediatric AML (primary) 31-87 7
Human pediatric ALL (primary) 80-97 3 with a specific immune/inflammatory “activation” signature characterized by tran-
Human MM (cell line) 88-99.9 2 scriptomic upregulation of PD-L1 and CD70 and increase in B-cells. There is no specific
Human DLBCL (cell line) 56-93 3
Human pancreatic cancer (cell line) 46-95 3
signature that distinguishes between the V+ subgroups [TTV vs EBV vs HPV]. However,
Human colorectal cancer (cell line) 66-98 2 transcriptionally, CD36, IFNA2 and IL17A upregulation distinguished TTV+ and HPV+
Human lung cancer (cell line) 82 1 cases, a phenomenon that may indicate HPV ability to initiate an immunosuppressive
tumor microenvironment. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
158s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2584 Poster Session 2585 Poster Session

Phase 1 trial of HCB101, a novel Fc-based anti-SIRPa-CD47 fusion protein, Role of p57 in cGAS-STING-mediated innate sensing and immunotherapy
in subjects with advanced cancers. First Author: Lucy Yan, Hanchor Biopharma response in hepatocellular carcinoma. First Author: Shirong Zhang, The First
Inc., Taipei City 114, Taiwan Affiliated Hospital of Xi’an Jiaotong University, Xi’an, ShaanXi, China
Background: CD47-targeting agents face challenges, including “on-target, off-tumor” Background: Hyperactivation of cell cycle programs in cancer cells suppresses the
toxicities affecting red blood cells, limited efficacy, and manufacturing complexities. antitumor immune response. The endogenous cyclin-dependent kinase inhibitor p57 is
Clinical holds and discontinued trials highlight these difficulties. These issues have an important tumor suppressor and a potential therapeutic target for hepatocellular
constrained their therapeutic potential and broadened the need for better solutions. carcinoma (HCC). However, the role of p57 in modulating antitumor immunity to HCC
HCB101 is an engineered human SIRPa fused to human IgG4 crystallizable fragment remains unclear. Methods: We examined p57 expression in HCC patient samples prior
(Fc) protein developed using the proprietary FBDB platform, blocks the signal of the to treatment with immune checkpoint inhibitors (ICIs) through immunohistochemistry
SIRPa-CD47 pathway, enhancing macrophage-mediated phagocytosis. Preclinical (IHC). Multiple mice tumor models were constructed to explore the role of p57 on the
studies show HCB101’s potent antitumor activity across solid tumors and hematological recruitment of CD8+ T cells in the tumor immune microenvironment. Through performing
malignancies. HCB101’s safety profile in repeat-dose cynomolgus monkey toxicity transcriptome sequencing, we analyzed the differential genes and activation pathways
studies revealed acceptable red blood cell or platelet abnormalities, supporting its induced by p57 overexpression; through Real-Time Quantitative Polymerase Chain
potential as a best-of-the-kind SIRPa-CD47 directed immunotherapy. Methods: This Reaction (RT-qPCR), western blot (WB）, Enzyme-Linked Immunosorbent Assay
Phase 1, open-label, dose-escalation trial evaluates HCB101’s safety, tolerability, (ELISA), IHC, immunofluorescence（IF）, Flow Cytometry (FCM) and other molecular
pharmacokinetics, pharmacodynamics, and preliminary antitumor activity in advanced experimental methods, we verified the molecular mechanism of increasing CD8+ T
solid tumors or non-Hodgkin lymphoma (NHL) in the US, Taiwan, and mainland China. infiltration and elevating PD-L1 caused by p57 overexpression; Through constructing
Eligible adults have treatment-refractory cancers, an Eastern Cooperative Oncology mice model and giving different treatments, we explored the anti-tumor efficacy of p57
Group (ECOG) performance status of 0 or 1, and adequate organ function. The 3+3 dose- overexpression combining with ICIs. Results: We found that patients with p57 ex-
escalation Bayesian Optimal Interval (BOIN) design assesses dose-limiting toxicities pression had a higher disease control rate, correlating with the number of tumor in-
(DLTs) in the first 28-day cycle. Secondary endpoints include objective response rate filtration CD8+ T cells. Using mouse models, we discovered that p57 promoted CD8+
(ORR), duration of response, and progression-free survival. Exploratory endpoints in- T cells infiltration and that CD8+ T cells were required for p57 to function as a tumor
clude CD47 receptor occupancy (RO), correlating response, and immune cell infiltration growth suppressor. Furthermore, through RNA-sequencing analysis and the multiplex
(NCT05892718). Results: 32 participants (median age 61 years; 74% male) enrolled assay in vitro and in vivo, we found that p57 induced chromosomal instability and
across 7 escalating cohorts (0.08-5.12 mg/kg, QW). Patients had a median of 4 prior subsequently stimulates cGAS-STING-type I IFN signaling, leading to upregulation of the
regimens. 68% had solid tumors, and 32% had NHLs. HCB101 was well tolerated, only 1 chemokines CCL5 and CXCL10, which promoted CD8+ T cell infiltration into the tumor
DLT reported at 2.56 mg/kg dose level (G3 platelet decrease). The most common microenvironment. Meanwhile, p57 also elevated the expression of PD-L1 on the surface
treatment related AEs were anemia (17%), all grade 1or 2, that did not require blood of HCC cells. Moreover, combining p57 overexpression with anti-PD-1 treatment syn-
transfusion or other treatments. HCB101 systemic exposure increased in a dose- ergistically inhibited tumor growth in vivo. Conclusions: Our studies demonstrated that
dependent manner. Preliminary efficacy showed 29% stable disease (6 patients) in p57 may serve as a new biomarker for ICIs efficacy and that increasing p57 expression
21 evaluable patients, with 2 patients have SD . 16 wks, and 1 patient has SD . 23 wks. is a potential therapeutic strategy for improving the efficacy of immunotherapy in HCC
Doses $1.28 mg/kg achieved $90% CD47 RO in peripheral T cells. patients. Research Sponsor: National Natural Science Foundation of China.
Conclusions: HCB101 demonstrated an acceptable safety profile and preliminary an-
titumor activity in heavily pretreated advanced cancer patients. These findings support
its further clinical development, including expansion cohorts to evaluate efficacy in
specific tumor types or in combination with other agents. Clinical trial information:
NCT05892718. Research Sponsor: None.

2586 Poster Session 2587 Poster Session

Phase 1 study of LB1410, a bivalent TIM-3/PD-1 bispecific antibody, in Effect of KROS 101, a small molecule GITR ligand agonist, on T effector
patients with advanced solid tumors or lymphoma. First Author: Jiajian Liu, L&L cells, T reg cells and intratumoral CD8 T cell cytotoxicity. First Author: John S.
Biopharma Co. Ltd., Shanghai, China Yu, Cedars-Sinai Medical Center, Los Angeles, CA
Background: LB1410 is a recombinant humanized anti-PD-1/TIM-3 bispecific antibody Background: A small molecule was identified that stabilizes the trimerization of the
(BsAb) developed by L&L Biopharma Co., Ltd. for patients (pts) resistant to or refractory to glucacorticoid-induced tumor necrosis factor receptor (GITR) ligand which then leads to
PD-1/PD-L1 treatments, showing superior T/DC cell activity and in vivo anti-tumor ef- the trimerization of GITR and magnified signaling of GITR. GITR signaling of T cells
ficacy compared to a combination of TIM-3 and PD-1 antibodies in preclinical studies. results in T effector cell expansion and T reg reduction. An antagonist to the GITR ligand
Here, we report the dose escalation and dose expansion results of LB1410 as mono- was also indentified which stabilizes the GITR ligand dimer formation preventing tri-
therapy in patients with advanced solid tumors (Keyplus-001). Methods: Eligible patients merization. Methods: Binding of KROS 101 to the GITR ligand was assessed and the
were $18 years old with ECOG PS 0-1 and advanced solid tumors. Dose cohorts ranged binding region of GITR ligand to KROS was determined with targeted deletion of GITR.
from 0.001 mg/kg to 20 mg/kg IV Q2W: 0.001 mg/kg-1 mg/kg in an accelerated titration T cell suppression studies were performed with T cell proliferation assays and T cell
design, and 3 mg/kg - 20 mg/kg using a traditional 3+3 design. Selected dose levels were cytotoxicity assays with glioblastoma target cells using patient derived PBMCs. Double
expanded in patients with advanced clear cell renal cell carcinoma (ccRCC) and cervical humanized GITR/GITRL mice bearing B16-F10-LUC2 tumors were treated with KROS 101
cancer (CC). The primary objective was safety, including dose-limiting toxicities (DLTs). or controls. Tumor-infiltrating lymphocytes were analyzed by flow cytometry and tumors
Secondary/exploratory objectives included efficacy, pharmacokinetics (PK), and im- assessed. Results: KROS 101 agonist binds to GITRL with high affinity with a KD of
munogenicity. Results: As of January 15, 2025, a total of 79 patients received LB1410 at 340nM by Surface Plasmon resonance. T cell suppression assay showed KROS 101 had
doses ranging from 0.001 mg/kg to 20 mg/kg as of January 15, 2025. The median age was peak proliferative induction of T effector cells at 25 uM concentration to 52% increase in
59 years, and 70% of patients were male. All enrolled patients had multiple organ proliferation, and peak proliferation induction of T effector cells with 1:1 ratio of T reg
metastases or multiple metastases in a single organ and were heavily pretreated with cells at 50uM concentration to 80% increase in proliferation of T effectors. KROS 101
anti-tumor therapies. Of the patients, 76.9% (60/79) had solid tumors that had failed treated T cell show enhanced effetor function and selectively target glioblastoma and
standard therapies and were resistant or refractory to anti-PD-1/PD-L1 treatments. cancer stem cells in vitro. KROS 101 enhances tumor immune infiltration by Increasing
Treatment-related adverse events (TRAEs) occurred in 63.3% of patients. The most CD3+ T Cells, CD8+ T Cells, and M1 Macrophages While Reducing Tregs and Myeloid
common TRAEs ($10%) included anemia (24.1%), proteinuria (12.7%), increased alanine Cells in vivo. KROS 101 enhances cytotoxicity by increasing IFNg and TNFa While
aminotransferase (11.4%), increased aspartate aminotransferase (11.4%), hyponatremia Reducing TIGIT and TIM3 in CD4+ and CD8+ T Cells In Vivo. Conclusions: KROS 101 is a
(10.1%), increased blood lactate dehydrogenase (10.1%), and weight loss (10.2%). Grade 3 GITR ligand agonist that increases T cell proliferation and increased cytotoxicity and
TRAEs occurred in 7 patients, including 3 with hypokalemia, 2 with hypertension, 1 with reduces the T reg population more effectively than TRX 518 which is a therapeutic GITR
hyponatremia, 1 with proteinuria, and 1 with pulmonary embolism. No dose-limiting antibody that was in clinical trial. Research Sponsor: None.
toxicities (DLTs) were observed. On- treatment scan was available for 66 patients. The
observed overall response rate (ORR) per RECIST 1.1 was 3/66 (4.5%), with 3 confirmed
partial responses (PRs) in patients with ccRCC and CC. The disease control rate (DCR) was
45.5% (30/66). In patients with ccRCC, the ORR was 16.7% (1/6), and the DCR was 66.7%
(4/6). In patients with CC, the ORR and DCR were both 66.7% (2/3). Of the 27 patients with
stable disease, 3 had stable disease for nearly 12 months, and 2 of them are still receiving
ongoing treatment in the study. Conclusions: LB1410 has a manageable safety profile
and demonstrates potential efficacy at tolerable doses in heavily pretreated patients,
particularly those with immune-oncology (IO)-refractory or resistant ccRCC and CC.
Clinical trial information: NCT05357651. Research Sponsor: L&L Biopharma Co., Ltd.,
Shanghai, China.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 159s

2588 Poster Session 2589 Poster Session

10
Machine learning–driven approaches for predicting T-cell–mediated immu- Phase 1 study of DK2 (EGFR), a tumor-targeted IL2 x IL10 dual immuno-
nity and beyond. First Author: Chongming Jiang, Terasaki Institute for Biomedical cytokine, in advanced cancer patients: Dose escalation, immune activation,
Innovation, Los Angeles, CA and safety results. First Author: Alexander I. Spira, NEXT Oncology Virginia, Fairfax,
Background: Recognition of peptides presented by the major histocompatibility VA
complex (MHC) through the T cell receptor (TCR-pMHC) is crucial for T cell function, Background: IL-2 induces anti-tumor immunity and toxicity, predominantly vascular leak
influencing disease conditions such as cancer, infections, and autoimmune disorders. syndrome (VLS), leading to edema, hypotension, organ toxicity, and therapy-inhibiting
Despite previous attempts, predictive models of TCR-pMHC specificity remain chal- regulatory T cell (Treg) accumulation. DK210 (EGFR) couples wild-type IL-2 to a high
lenging. Methods: Inspired by recent breakthroughs in protein structure prediction affinity variant of EBV IL-10 via an scFv that binds to epidermal growth factor receptors.
achieved by deep neural networks, we explored structural modeling using AlphaFold 3 The IL-10 component was designed to block IL-2 mediated cytokine release syndrome
(AF3)-based AI-enabled computation as a potential avenue for predicting TCR epitope (CRS) and VLS while retaining T cell activation and proliferation and limiting Treg ex-
specificity. Results: We show that a specialized version of the neural network predictor pansion. We report the clinical and pharmacodynamic results from the dose escalation in
AlphaFold can generate models of TCR-pMHC interactions, effectively distinguishing the DEKA-1 phase 1 (NCT05704985) study. Methods: Eligible patients (pts) had advanced/
valid peptide epitopes from invalid ones with increasing accuracy. Strongly immuno- metastatic tumors known to express EGFR, progressive disease on $1 lines of systemic
genic epitopes could be identified and selected for vaccine development through in- treatment, and ECOG #1. DK210 (EGFR) (2-16 mg; 0.025-0.5 mg/kg for an 80 kg subject)
silico high-throughput processes. Higher-affinity and specificity T cells could also be was self-administered subcutaneously 3 times per week in 21-day cycles following a BOIN
computationally designed to achieve improved efficacy and safety profiles for T cell design. Adverse events (AEs) including serious (SAEs) were evaluated using CTCAE version
therapy. An accurate TCR-pMHC prediction model is expected to significantly benefit T- 5.0. Cytokines and anti-drug antibodies were monitored during the first cycle and every 3
cell-mediated immunotherapy and facilitate advanced drug design. cycles thereafter. RECIST 1.1 tumor responses were evaluated every 9 weeks. Results: 35
Conclusions: Overall, precise prediction of T-cell immunogenicity holds substantial pts (14 RCC, 6 NSCLC, 9 CRC, 5 PDAC, 1 SCC) were enrolled. Median age was 63 yrs (range
therapeutic potential, enabling the identification of peptide epitopes associated with 35-80). Treatment-related AEs (TRAEs; any grade) in $ 10% pts were injection site re-
tumors, infectious agents, and autoimmune diseases. Although much work remains actions (63%), fever (40%), fatigue (31%), nausea (23%), anemia (17%), chills (17%),
before these predictions could achieve widespread practical utility, deep learning-based eosinophilia (14%), CRS (14%), diarrhea (11%); the majority were G1-2. G3 TRAEs included
fatigue (n = 4), anemia (n = 3), syncope (n = 3) and single events of acute kidney injury,
structural modeling represents a promising path toward the generalizable predictions of
cellulitis, hypoalbuminemia, and lymphopenia. No DLTs were observed. While MTD was not
TCR-pMHC interactions and beyond. Research Sponsor: None.
exceeded, a dose proportional induction of IFNg was observed through 8 mg but not at
16 mg. Therefore, higher doses were not explored, and a 12 mg dose level was introduced
for dose optimization. No appreciable increase in IL-6, TNF-a, or IL-1b was seen other than
1 subject at 4 mg with G2 CRS and elevated IL-6. IFNg and IL-5 were concomitantly
induced proportional to dose level and reached saturation between levels 3 and 4,
consistent with pK exposure of DK210 (EGFR). Sustained IL-5 associated eosinophilia was
observed and correlated with drug concentration but did not require intervention. IL-2 and
IL-10 induction led to sustained elevation of IL-2Ra and IL-18, respectively. 33% of
evaluable patients had a best ORR of SD. Conclusions: DK210 (EGFR) demonstrates
strong IL-2 activity shown by IL-5 driven eosinophilia, shed IL-2Ra, T and NK cell pro-
liferation and expansion but not Treg accumulation. These effects have been decoupled
from IL-2 driven toxicity, confirming the hypothesis of the balancing effect of IL-10. These
data support further evaluation in combination with T cell engagers, T cell therapeutics,
and kinase inhibitors. Clinical trial information: NCT05704985. Research Sponsor: Deka
Biosciences.

2590 Poster Session 2591 Poster Session

PCT1:CO-STIM TCR T-cells to overcome the hostile tumor micro- First-in-human mRNA CAR therapy: Correlative biomarker analysis from the
environment and target triple-negative breast cancer. First Author: Dora MT-302 phase 1 study targeting TROP2 in patients with advanced epithelial
Hammerl, Pan Cancer T, Rotterdam, Netherlands tumors. First Author: Charlotte Rose Lemech, Scientia Clinical Research, Randwick,
Background: Adoptive T-cell therapy has demonstrated impressive efficacy in hema- Australia
toligical cancers but the solid tumor micro-environment presents a unique challenge. Background: Outcomes for patients with epithelial cancers, including breast, lung and
However, recently TCR-T cell therapy has shown benefit in difficult-to-treat solid tumors gastrointestinal tumors, remain poor particularly in advanced stages. MT-302, an mRNA-
and selection of specific tumor targets and control of the tumor micro-environment can based chimeric antigen receptor (CAR) therapy, seeks to address this unmet need by
unlock the broader potential of T cell therapy. Triple-negative breast cancer (TNBC) is a reprogramming myeloid cells in vivo to recognize and kill TROP2-expressing tumors,
difficult-to-treat tumor as it lacks classical targets for hormone and antibody-based recruit immune cells into tumor and induce systemic anti-tumor responses. Its CAR
therapies. It harbors a highly immune-suppressive microenvironment and rarely responds construct combines an anti-TROP2 scFv with truncated CD89 and becomes functionally
to immune-checkpoint inhibitors. We sought to identify a novel target to make TNBC active only upon association with FcRg-expressing myeloid cells, ensuring precise immune
amenable for adoptive T-cell therapy with T-cell receptor (TCR)–engineered cells and engagement. Delivered as an off-the-shelf, repeatable intravenous treatment without the
applied a unique next-generation gene-engineering approach to make T-cells overcome the need for preconditioning, MT-302 overcomes the logistical and technical challenges of
hostile microenvironment. Methods: (i) Discovery of TNBC-restricted target: We applied in traditional cell and CAR therapies. MT-302 is being evaluated in a Phase 1, multicenter,
silico analyses of .500 TNBC samples and .1,500 healthy tissues and validated findings open-label dose-escalation study (NCT05969041) in adults with advanced epithelial
with qRT-PCR and immune stainings of .300 TNBC samples as well as 40 healthy tissues. cancers expressing TROP2. Here, we present a correlative biomarker analysis from the
(ii) Discovery and selection of PCT1 TCR: We enriched ROPN1-specific TCRs from naı̈ve first-in-human MT-302 study. Methods: Tumor biopsies and peripheral blood samples
repertoires and assessed specificity, sensitivity and performed preclinical safety studies. were collected pre- and post-dose. Biomarkers were evaluated utilizing advanced tech-
(iii) Development of TCR:CO-STIM technology to overcome immune suppression: we nologies including immunohistochemistry (IHC), Xenium and Hyperion imaging, flow
designed a panel of murine TCRs harboring different intracellular co-stimulatory domains, cytometry, Chromium single-cell sequencing, T cell receptor sequencing and Meso Scale
thereby providing additional stimulation to T cells aimed to overcome immune suppression Discovery (MSD). These methods assessed TROP2 expression on cancer cells, TROP2 CAR
in solid cancer. We tested their ability to extend anti-tumor durability in a murine melanoma expression within immune cells, systemic pharmacodynamic effects, immune cell infil-
model, performed comprehensive permutations to enable stable expression of fully human tration and tumor microenvironment changes. Results: Results showed robust TROP2
TCR:CO:STIM and applied it to multiple TCR specificities. Results: For TNBC, we identified CAR expression in circulating myeloid cells within hours of dosing. In tumor biopsies, CAR-
that Ropporin (ROPN1), a protein expressed homogeneously in .90% of TNBC and positive myeloid cells co-localized with TROP2-expressing cancer cells. Post-dose tumor
persistent across disease stages but absent from healthy tissues, as an ideal target. We biopsies exhibited an increase in antigen presentation markers and pro-inflammatory
identified 13 clonal TCRs directed against 9 different ROPN1 epitopes. The lead TCR, signaling compared to baseline. MT-302 elicited systemic interferon-driven chemokine
termed PCT1 TCR, demonstrated high sensitivity and specificity towards ROPN1+/HLA-A2+ responses and reprogrammed the tumor immune microenvironment, promoting effector
cell lines and patient-derived organoids in 3D. Our TCR:CO-STIM technology significantly T cell recruitment. T cell receptor sequencing confirmed the emergence of novel T cell
improved duration of response in a murine model and improved T cell fitness. Notably, clones, consistent with adaptive immunity activation. Baseline IHC confirmed high TROP2
when repeatedly challenged with ROPN1+/HLA-A2+ TNBC cells, PCT1:CO-STIM, but not expression in enrolled patients, correlating with pharmacodynamic activity and immune
PCT1 TCR T-cells, could resist up-regulation of T-cell exhaustion markers and retained reprogramming. Conclusions: This Phase 1 study provides the first evidence of successful
tumor-killing capacity for 3-10 extra rounds of stimulation. Importantly, PCT1:CO-STIM did delivery of mRNA-CAR therapy in humans. These biomarker findings demonstrate that MT-
show any signs of tonic signaling nor crossreactivity nor alloreactivity towards any major 302 selectively engages myeloid cells and induces robust innate and adaptive anti-tumor
HLA-I allele. Conclusions: TCR:CO-STIM technology has shown enhanced activity of a pharmacodynamic responses, providing support for further investigation of MT-302’s
selective and specific TCR targeted at ROPN1 and we are progressing PCT1:CO-STIM to the potential as a transformative treatment for patients with TROP2-expressing epithelial
clinic for the treatment of TNBC. Research Sponsor: None. cancers. Clinical trial information: NCT05969041. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
160s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2592 Poster Session 2593 Poster Session

AI-driven biomarker prediction in oncology: Enhancing pathological image Systemic antitumor virotherapy: Pre-clinical evaluation of tumor targeting,
analysis with EXAONEPath. First Author: Hyung Kyung Kim, Samsung Medical efficacy, and safety of lead candidate (CLD-401). First Author: Duong Hoang
Center, Seoul, South Korea Nguyen, Calidi Biotherapeutics, San Diego, CA
Background: Hematoxylin and eosin (H&E)-stained whole-slide images (WSIs) are Background: Systemic antitumor virotherapies are a promising modality of cancer
fundamental in cancer diagnosis, providing critical insights into tumor morphology and immunotherapy. However, challenges include quick virus clearance from the blood-
the tumor microenvironment. Traditionally, biomarker assessment has relied on manual stream and potential off-target toxicity. To overcome these limitations, we have
pathological evaluations, which are prone to human error and limited in scalability. developed a novel strain called RT vaccinia virus which can be enveloped by an ex-
Subtle biomarker expressions that evade visual detection further challenge conventional tracellular membrane during the manufacturing process and become resistant to
methods. Methods: We developed EXAONEPath, an artificial intelligence (AI) model human-complement. Our RTNova program, using extracellular enveloped RT (envRT)
trained on approximately 73,000 pan-cancer H&E-stained WSIs, to predict key cancer vaccinia viruses, focuses on generating potent systemic virotherapies with improved
biomarkers. The model was evaluated across three major biomarker prediction tasks: survival in circulation, tumor-specific targeting and enhanced therapeutic efficacy.
Tumor Mutation Burden (TMB) Prediction in Lung Adenocarcinoma (LUAD): Using the Methods: The RT virus was genetically engineered to improve tumor selectivity and
TCGA-LUAD cohort, the model was trained (n=373), validated (n=47), and tested (n=47). increase resistance to complement-mediated inactivation. The virus’s ability to kill
Cross-institutional validation was conducted on Samsung Medical Center (SMC) (n=341) cancer cells was tested using the NCI-60 panel. The resistance of envRT vaccinia virus
and an in-house dataset (n=254). EGFR Mutation Prediction in LUAD: The TCGA-LUAD against human humoral immunity and its rapid spread were assessed ex-vivo. Targeting,
dataset was split into training (n=382), validation (n=48), and test (n=48) sets. Additional biodistribution, therapeutic efficacy, and safety profile of selected envRT was evaluated
validation was performed on the SMC LUAD cohort (n=341). Microsatellite Instability in multiple animal models. Results: Out of several genetically modified RT Vaccinia
(MSI) Prediction in Colorectal Adenocarcinoma (CRC): A combined TCGA-STAD/TCGA- viruses, we selected the one with three knockouts (3KO): TK (Thymidine kinase), A46R
READ dataset was used for training (n=432), validation (n=55), and testing (n=54). The (immunomodulator), and VGF (Vaccinia virus growth factor). These genetic modifi-
model was further validated on the SMC CRC cohort (n=974). Results: EXAONEPath cations significantly improved tumor-selective amplification and safety profile while
demonstrated a strong predictive performance: TMB in LUAD: AUROC scores of 0.77 maintaining therapeutic efficacy. The 3KO RT virus demonstrated strong oncolytic
(TCGA), 0.81 (SMC), and 0.76 (in-house). EGFR Mutation in LUAD: AUROC scores of 0.78 activity against more than 60 different human cancer cell lines NCI-60. Additionally, the
(TCGA) and 0.84 (SMC). MSI in CRC: AUROC scores of 0.92 (TCGA) and 0.86 (SMC). 3KO RT virus was genetically engineered with CD55-domain fused with viral envelope
Conclusions: EXAONEPath advances AI-driven pathological image analysis by auto- A33R. This chimeric protein is designed to be expressed specifically in the extracellular
mating biomarker prediction with high accuracy and cross-institutional robustness. Its envelope of the viral particle to robustly protect the envRT and viral progeny from
strong performance in predicting clinically relevant biomarkers, including TMB, EGFR inactivation by human complement. Targeting and biodistribution studies revealed that
mutations, and MSI, highlights its potential for integration into precision oncology RT virus targeted all tumors after intravenous administration followed by significant
workflows. Future research will focus on expanding biomarker applications and en- tumor-selective amplification and spreading. In multiple immunocompetent mouse
hancing cross-institutional generalizability for broader clinical impact. Research models, including metastatic lung cancer, RT virus demonstrated excellent tumor killing,
Sponsor: None. and expression of selected therapeutic payload. Conclusions: We have developed a new
scalable process to manufacture extracellular enveloped antitumor virotherapies and
identified the first lead candidate from RTNova Platform, designated as CLD-401. This
candidate, CLD-401, demonstrates promising therapeutic efficacy and safety in pre-
clinical models. It effectively addresses the challenge of targeting and treating met-
astatic lung cancer by delivering Immunotherapeutics directly to disseminated tumors.
Research Sponsor: None.

2594 Poster Session 2595 Poster Session

The predictive value of BRCA mutation on survival of cancer patients treated Monotherapy of envafolimab in patients with high tumor mutational burden
with immune checkpoint inhibitors: A systematic review and meta-analysis advanced solid tumors: Results from a phase II clinical trial. First Author: Jian
of phase III randomized clinical trials. First Author: Mus’ab Theeb Mustafa, The Li, Department of Gastrointestinal Oncology, Key laboratory of Carcinogenesis and
Hashemite University, Faculty of Medicine, Zarqa, Jordan Translational Research (Ministry of Education), Peking University Cancer Hospital &
Background: Immune checkpoint inhibitors (ICIs) have significantly enhanced survival Institute, Beijing, China
for various types of cancers; however, resistance has limited the number of patients who Background: Tumor mutational burden (TMB) has emerged as a predictive biomarker of immune
can benefit from these regimens. Therefore, additional biomarkers are necessary to checkpoint blockade response in cancers. Envafolimab, a humanized single-domain anti-PD-L1
hopefully overcome resistance. Currently, the role of BRCA Mutation in ICI therapy antibody subcutaneous administration (s.c.), has been approved in China for the treatment of
remains poorly understood and controversial. Methods: We systematically searched advanced solid tumors with MSI-H. The study is to explore the potential anti-tumor activity in
patients with TMB-high (TMB-H) in China. Methods: The study consists of two parts. Part 1 is to
PubMed, Web of Science, and Cochrane for phase III randomized clinical trials (RCTs) explore the association of single agent envafolimab activity with tissue TMB (tTMB) measured by
comparing ICI with placebo or standard-of-care cancer treatment stratified by BRCA Onco500 assay in patients with advanced solid tumors. Part 2 will further evaluate the efficacy of
mutation status as wildtype or mutant type up to 19 November 2024 regardless of envafolimab in advanced solid tumor patients base on the cutoff of TMB value identified from Part
cancer type or stage. The included phase III trials must report at least one of the 1. Envafolimab is administrated s.c. at 400 mg every 4 weeks until disease progression, adverse
following: Progression-free survival (PFS) or overall survival (OS); the meta-analysis was events, or other reasons causing treatment discontinuation. Efficacy and safety are assessed in all
conducted using RevMan 5.4 pooling hazard ratio (HR) with 95% confidence intervals patients who received at least one dose of envafolimab. tTMB is assessed by a central lab using
(CI) with a p-value of , 0.05 considered significant. Results: We conducted a meta- SimcereDx Onco500 assay (Jiangsu Simcere Medical Device Co., Ltd, China). The primary endpoint
analysis of six phase III RCTs involving 3,328 patients: three trials investigated ovarian was objective response rate (ORR) assessed by independent review committee per RECIST v1.1
cancer, two investigated prostate cancer, and one investigated breast cancer. The criteria. Results: As of Nov 15, 2024, a total of 70 patients with advanced cancers (colorectal
cancer [9,12.6%], cervical cancer and soft tissue sarcoma [8 each; 11.4%], and other 18 tumor
analysis revealed that ICIs significantly improved both OS and PFS for patients with
types) have received envafolimab in Part 1. 30 (42.9%) patients had received $3 systemic
BRCA mutation, with HR of 0.61 (95% CI, 0.46 – 0.81, p = 0.0008) and 0.64 (95% CI, 0.47 therapies (median 2; range 1-19). Median follow-up time was 31.2 months (range: 0.6-36.8). 49
– 0.89, p = 0.008), respectively. Furthermore, for patients with wildtype BRCA, the (70%) patients had at least one treatment-related adverse events (TRAEs), and 6 (8.6%) had grade 3
analysis revealed that using ICIs significantly improves PFS with HR of 0.81 (95% CI, 0.72 or 4 TRAEs. The most common TRAEs were anemia and alanine aminotransferase increased (9
– 0.90, p = 0.0001). However, ICIs did not significantly improve overall survival, with HR each; 12.9%). Grade 2 decreased appetite was the only TRAE resulting in treatment discontin-
of 0.94 (95% CI, 0.84 – 1.06, p = 0.33). Conclusions: The use of ICIs in cancer patients uation. No treatment-related death reported. TMB$13 mut/Mb with Onco500 panel was selected
with BRCA mutation is associated with significant improvement in PFS and OS; however, as the threshold of TMB-H based on the clinical data and previous comparison of platforms for
in patients with wildtype BRCA, the use of ICIs showed only significant improvement in determining TMB value in patients. Key efficacy outcomes are presented (Table). ORR and DOR
PFS with no significant improvement in OS. Research Sponsor: None. were higher in patients with tTMB $13mut/Mb (33.3% and 20.2m) than patients with tTMB ,
13mut/Mb (4.3% and 3.8m). Conclusions: tTMB could be a useful predictive biomarker for re-
sponse to envafolimab in patients with pre-treated advanced solid cancer. The Part 2 of this study
is ongoing (NCT04891198). Clinical trial information: NCT04891198. Research Sponsor: None.
tTMB‡13 mut/Mb tTMB<13 mut/Mb
(n = 24) (n = 46)
Objective response rate, n (%) [95% CI] 8 (33.3) [15.6-55.3] 2 (4.3) [0.5-14.8]
Complete / partial response 1 (4.2) / 7(29.2) 0 (0) / 2 (4.3)
Stable disease / progressive disease 2 (8.3) / 9 (37.5) 16 (34.8) / 24 (52.2)
Median DoR, months (95% CI) 20.2 (4.4-NE) 3.8 (NE-NE)
Median PFS, months (95% CI) 2.8 (1.8-8.7) 1.9 (1.8-3.6)
Median OS, months (95% CI) 13.2 (5.7-NE) 12.7 (7.4-18.2)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 161s

2596 Poster Session 2597 Poster Session

A phase 1 study of fixed-dose regimens of serplulimab, an anti-PD-1 an- Consolidative camrelizumab following definitive concurrent chemoradio-
tibody, in patients with advanced solid tumors. First Author: Ching-Liang Ho, therapy with involved-field irradiation in locally advanced esophageal squa-
Division of Hematology and Oncology, Tri-Service General Hospital, Taipei, Taiwan mous cell carcinoma: A single-arm phase 2 trial. First Author: Jun Wang,
Background: Serplulimab is a recombinant humanized IgG4 monoclonal antibody tar- Department of Radiation Oncology, the Fourth Hospital of Hebei Medical University,
geting PD-1. A two-cohort phase 1 study was conducted to evaluate the safety of ser- Shijiazhuang, China
plulimab monotherapy in patients with advanced solid tumors (NCT03468751). Findings Background: Definitive concurrent chemoradiotherapy (dCCRT) is considered the
from the dose-finding cohort has been previously reported at the 2022 ASCO Annual standard treatment for esophageal squamous cell carcinoma (ESCC). The PACIFIC study
Meeting (No. e14560). Here we present results from the dose expansion cohort, in which demonstrated that consolidation durvalumab significantly improves overall survival (OS)
fixed-dose regimens were evaluated. Methods: This multicenter phase 1 study enrolled in patients with stage III non-small cell lung cancer (NSCLC) after dCCRT. However, the
patients with locally advanced or metastatic solid tumors who have failed or are intolerant efficacy of consolidation immunotherapy in ESCC still remains unclear. We conducted a
to standard therapy or for whom no standard therapy is available. In the dose expansion clinical trial to evaluate the efficacy of camrelizumab in patients with unresectable,
cohort, patients received intravenous serplulimab at 200 mg Q2W, 300 mg Q3W, 400 mg locally advanced ESCC following dCCRT. Methods: This single-arm, phase 2 study
Q4W, or 600 mg Q6W. The primary endpoints were adverse event profile and maximum enrolled patients with locally advanced ESCC. All participants received dCCRT with
tolerated dose (MTD). Secondary endpoints included pharmacokinetic (PK), immunoge- involved-field irradiation (IFI). Patients were treated with camrelizumab within 1 to
nicity, pharmacodynamics (PD), and efficacy. Results: As of data cut-off on Jan 5, 2024, 42 days after completing dCCRT. Camrelizumab was administered intravenously over 30
37 patients received at least one dose of serplulimab at 200 mg Q2W (n = 9), 300 mg Q3W minutes every 2 weeks for up to 12 months. The primary endpoint was progression-free
(n = 9), 400 mg Q4W (n = 10), or 600 mg Q6W (n = 9). All patients were Asian, 70.3% male; survival (PFS). Secondary endpoints included disease control rate (DCR), objective
median age was 60.0 yrs (range 33–88). Patients had head and neck cancer (n = 10, response rate (ORR), duration of response (DoR), overall survival (OS), and safety.
27.0%), esophageal cancer (n = 6, 16.2%), colorectal cancer (n = 4, 10.8%) or other types of Results: Thirty-five patients were enrolled between April 2020 and November 2023.
tumor. Most patients had metastatic disease (64.9%). All patients had prior systemic Data from 32 patients were analyzed. As of December 22, 2024, the median follow-up
cancer treatment, including 4 (10.8%) with prior immunotherapy; 51.4% had $ 3 prior lines was 25.1 months (IQR 5.5–56.8). Twelve patients experienced disease progression, and
of therapy. All 37 patients were included in safety, PK, and PD analyses; 35 response-
seven patients died. The DCR was 59.4%. The median PFS and OS were not reached. The
evaluable patients were included in efficacy analysis. No dose-limiting toxicity was re-
1- and 2-year PFS rates were 81.3% and 60.6%, respectively. The 1- and 2-year OS rates
ported, and MTD has not been determined. Treatment-related adverse events (TRAEs) were
were 96.9% and 81.0%, respectively. The most common adverse events were grade 1-2.
observed in 19 patients (51.4%), including 7 (18.9%) reporting grade $ 3 TRAE. TRAE
No grade 4 or 5 adverse events were reported. Pneumonia occurred in 31.3% of patients,
incidence was similar across regimen groups. Following multiple infusions, the geometric
mean t1/2 , ss was from 341.1–751.3 h, and geometric mean CLss was 0.006–0.009 L/h.
all of whom experienced grade 1-2. Conclusions: Consolidative camrelizumab following
Treatment-emergent anti-drug antibody (ADA) was detected in 7 (18.9%) patients. No definitive concurrent chemoradiotherapy with IFI shows promising efficacy and man-
difference in safety or PK was noted between ADA-positive and -negative patients. Profiles ageable toxicity in patients with unresectable locally advanced ESCC. Clinical trial
of PD-1 receptor occupancy in circulating CD3+ T cells and interleukin-2 stimulation ratio information: NCT04286958. Research Sponsor: None.
were similar across dose groups, suggesting dose-independent functional blockade. Six
patients (300 mg Q3W, 4; 400 mg Q4W, 2) achieved partial response, resulting in an ORR of
17.1%. Among the responders, 12-month duration of response rate was 66.7% (95% CI
confidence interval, 19.5–90.4). Median progression-free survival was 2.3 months (95% CI,
1.9–5.1). Conclusions: Fixed-dose regimens of serplulimab showed favorable safety, PK,
and PD characteristics and preliminary anti-tumor activity, supporting its further inves-
tigation. Clinical trial information: NCT03468751. Research Sponsor: Shanghai Henlius
Biotech, Inc.

2598 Poster Session 2599 Poster Session

Efficacy of low-dose nivolumab in advanced cancers: A retrospective anal- Immune related liver toxicity, management, and outcomes in ICI treated
ysis from medical oncology clinic in Eastern India. First Author: Kiran patients with advanced or metastatic cancers. First Author: Zara Izadi, Bristol
Yidagur Gangadharaiah Lokesh Sr., All India Institute of Medical Sciences, Bhubaneswar, Myers Squibb, Princeton, NJ
Bhubaneswar, India Background: The impact of hepatic immune-related adverse events (HirAEs) and their
Background: Immunotherapy with PD-1/PDL1 blocking monoclonal antibodies has improved management (mgmt) on clinical outcomes in patients receiving immune checkpoint
survival across several malignancies at different stages of these malignancies. But in Low- & inhibitors (ICI) has not been fully examined. We aimed to evaluate the association
middle-income countries, only 1-3% of cancer patients can access the standard dose of Im- between HirAEs, their mgmt, and overall survival (OS) in ICI-treated cancer patients.
munotherapy. In this study, we aim to assess the response to low dose (LD) of Immunotherapy Methods: Data were drawn from the Flatiron Health Research Database, an EHR-based
(nivolumab) across a broad range of malignancies. Methods: The study is a retrospective database representing 280+ U.S. community oncology practices. Adults with advanced
descriptive study. A total of 104 patients with advanced cancers were included in the study.
non-small cell lung cancer (aNSCLC), advanced melanoma (aMel), or metastatic renal
Patients received a lower dose of Nivolumab (20/40 mg), ones in a 2-weekly-4weekly schedule,
with treatment continued until disease progression or intolerable toxicity Their demographics,
cell carcinoma (mRCC) who initiated ICI between 1/1/16 - 12/31/20 were included and
clinical profile, response to therapy, and adverse events were analyzed. Results: Male to female followed from ICI initiation to death, loss to follow-up, or end of the study period (12/31/
ratio was 5:1. The median age of patients was 49 years (range - 15 years to 78 years) and 70% 2021). CTCAE Grade 2 or higher HirAEs and mgmt actions (immunosuppression using
patients were ECOG-PS 1-2 30% were ECOG-PS 3-4. Overall 56 patients were diagnosed with corticosteroids or other immunosuppressants, ICI-regimen holds, ICI-regimen discon-
Squamous cell carcinoma (SCC) Head&neck, 15 had Renal cell carcinoma (RCC), 14 had tinuations) and hospitalizations were curated from unstructured data. Cox regression
Malignant melanoma, 5 had lung cancer, 4 with Hepatocellular carcinoma (HCC), 2 each of was used to evaluate the association between HirAEs, their mgmt (both as time-varying
Gynecological cancer, Gall bladder cancer & CUP and 1 each of stomach cancer, Urinary bladder covariates) and OS adjusting for baseline characteristics such as line of therapy and
malignancy (HGUC), & Lymphoma. The most common metastatic sites were Lungs (46%) . corticosteroid use. The earliest HirAE per patient was examined in OS analysis.
Bone(27%) . Liver(15%). A total of 73 patients were included for assessment (received $2 Results: The study included 529 aNSCLC, 557 aMel, and 431 mRCC patients. For
cycles of Nivolumab). The overall response rate (ORR) was 39.7% and the Disease control rate aNSCLC, aMel, and mRCC, respectively, 23.4%, 41.5%, and 30.9% experienced at least
(DCR) was 54.7%. Median PFS was 4 months (range - 1month to 26months) with Median OS one HirAE, with a median time to onset of 59, 60, and 63 days. Among all HirAEs, elevated
being 11 months (range - 3months to 30months). Grade 3-4 adverse events were seen in 21/ liver enzymes were the most common (72.9% in mRCC to 76.6% in aMel), followed by
104(20%), the most common being dermatological (8/21) followed by anemia (7/21) and
hepatitis (8.3% in aNSCLC to 12.6% in aMel). Immunosuppression was used to treat
endocrinal AEs (4/21). Conclusions: Low-dose Nivolumab showed good response rates in
HirAEs in 47.6%, 57.6%, and 38.3% of aNSCLC, aMel, and mRCC patients with HirAEs. In
advanced malignancies with manageable toxicities even in poor general condition. The cost of
therapy was 1/5th to 1/10th of the standard dose of Nivolumab, highlighting its potential as a aNSCLC, aMel, and mRCC, respectively, median survival was 12.7, 52.2, and 25.5 months
cost-effective alternative in resource-limited settings. Research Sponsor: None. and HirAEs were associated with a higher risk of all-cause mortality than no HirAEs [HR
(95%CI): 1.8 (1.3-2.2); 1.4 (1.0-1.8); 1.3 (1.0-1.8)]. In mRCC, ICI-regimen holds and
Response and survival analysis.
discontinuations were associated with a higher risk of all-cause mortality than im-
Charectiristics N=73 munosuppression alone (HR$4.0; P#0.05). In aNSCLC, HirAEs that led to hospitali-
Response rates zation were associated with a higher risk of all-cause mortality regardless of HirAE mgmt
CR 5 (HR: 6.2; P , 0.01). In aMel HirAE mgmt was not associated with OS. Conclusions: ICI-
PR 24
SD 11 related HirAEs were associated with higher mortality in aNSCLC, aMel, and mRCC. HirAE
PD 33 mgmt impacted OS differently across cancer types, highlighting the need for tailored,
ORR 39.7% timely, and multidisciplinary mgmt strategies in the ambulatory care setting, especially
DCR 54.7%
Survival analysis for cancers with poorer prognosis. Research Sponsor: Bristol Myers Squibb.
mPFS 4 months
mOS 11.5 months

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
162s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2600 Poster Session 2601 Poster Session

Baseline autoimmune diseases and characteristics of solid tumor patients Immune related kidney toxicity, management, and outcomes in ICI treated
on immune checkpoint inhibitor (ICI) therapy enrolled in a prospective study patients with advanced or metastatic cancers. First Author: Zara Izadi, Bristol
of immune-related adverse events (irAEs): SWOG S2013 (I-CHECKIT). First Myers Squibb, Princeton, NJ
Author: Krishna Soujanya Gunturu, Hartford HealthCare Cancer Institute, Hartford, CT Background: The impact of kidney immune-related adverse events (KirAEs) and their
Background: I-CHECKIT is a prospective observational study whose primary objective is to management (mgmt) on clinical outcomes in patients receiving immune checkpoint
develop and independently validate a risk prediction model for the development of Common inhibitors (ICI) has not been fully examined. We aimed to evaluate the association
Terminology Criteria for Adverse Events (CTCAE) Grade 3 or higher non-hematological irAEs in between KirAEs, their mgmt, and progression-free survival (PFS) and overall survival
patients with solid tumors during the first year of treatment with ICI. Methods: Any patient (OS) in ICI-treated cancer patients. Methods: Data were drawn from the Flatiron Health
initiating ICI per their treating oncologist and National Comprehensive Cancer Network Research Database, an EHR-based database representing 280+ U.S. oncology practices.
guidelines was eligible to participate in this study. Eligibility criteria were unrestrictive and Adults with advanced non-small cell lung cancer (aNSCLC), advanced melanoma (aMel),
included participants with active autoimmune disease, decreased performance status, and any
or metastatic renal cell carcinoma (mRCC) who initiated ICI between 1/1/16 - 12/31/20
stage of cancer. One important exclusion criterion was planned receipt of ICI with chemo,
were included and followed from ICI initiation to death, loss to follow-up, or end of the
biological or targeted therapy. Hormonal therapy and palliative radiation were allowed. The
study is close to accrual in May 2024. Here we describe baseline participant characteristics. study period (12/31/21). CTCAE Grade 2 or higher KirAEs and mgmt actions (immu-
Results: Of a total of 2084 enrolled participants, 62 were ineligible. Community based NCORP nosuppression using corticosteroids or other immunosuppressants, ICI-regimen holds,
sites enrolled the majority (n= 1,181, 56%) of participants. Participants were also enrolled from ICI-regimen discontinuations) and hospitalizations were curated from unstructured
SWOG Latin American sites and Veteran Affairs (VA). 31% (n=656) had skin cancer (melanoma data. Cox regression was used to evaluate the association between KirAEs, their mgmt
90%, squamous 5% and Merkel cell 2.9%), 29% (n=604) had lung cancer and 12% (n=256) had (both as time-varying covariates) and PFS and OS adjusting for baseline characteristics
kidney cancer. 17% (n=346) received combination ICI. The median age was 69.9 years. 64% such as line of therapy and corticosteroid use. The earliest KirAE per patient was
were male, 90% (n=1866) were white, 8% (n=165) Hispanic/Latino, 5% (110) black, and 1% (16) examined in survival analyses. Results: The study included 513 aNSCLC, 463 aMel, and
Asian. 12% had performance status (PS) 2 or greater. 67% of the participants were overweight 451 mRCC patients. For aNSCLC, aMel, and mRCC, respectively, 21.1%, 29.6%, and 33.9%
or obese. 9% (n=180) of the participants had active autoimmune disease such as rheumatoid experienced at least one KirAE, with a median time to onset of 70, 84, and 128 days.
arthritis, Type I diabetes, hypothyroidism, psoriasis, Crohn’s, ulcerative colitis. 3% (n=54) had a Nephritis ranged from 2.1% of KirAEs in aNSCLC to 4.6% in aMel. Elevated creatinine was
history of autoimmune disease not currently requiring treatment. Conclusions: The I-CHECKIT the most common KirAE (23.6% in mRCC to 29.1% in aNSCLC), followed by acute kidney
observational study enrolled participants representative of a real-world population, as most of injury (20.5% in mRCC to 28.4% in aNSCLC). Immunosuppression was used to treat
the participants were from community practices such as NCORP. Most participants had KirAEs in 67.6%, 80.3%, and 67.3% of aNSCLC, aMel, and mRCC patients with KirAEs. In
melanoma, resulting in a higher proportion of white participants. 9% of this population had aNSCLC, aMel, and mRCC, respectively, median OS was 17.1, 58.6, and 33.7 months,
baseline active autoimmune disease, a population excluded in initial clinical trials, highlighting
median PFS was 7.8, 8.9, and 9.5 months, and patients with KirAEs had longer PFS than
the broader use of ICI in daily clinical practice. Clinical trial information: NCT04871542.
those without KirAEs [HR (95%CI): 0.65 (0.51-0.83); 0.74 (0.57-0.97); 0.67 (0.53-0.84)].
Research Sponsor: NIH/NCI/NCORP grant UG1CA189974.
In aNSCLC, KirAEs were associated with shorter OS [1.31 (0.98-1.74); P = 0.06]. In all
Baseline characteristics. cancers, KirAEs that occurred during hospitalization or led to hospitalization were
Age All No (%) Single ICI No (%) Combo ICI No (%) Skin No (%) Lung No (%) associated with shorter OS (HR$2.84; P , 0.02). KirAE mgmt was not associated with
61-65 years 301 (14) 244 (14) 57 (16) 94 (14) 91 (15) OS or PFS. Conclusions: Results suggest that while KirAEs might indicate an intensified
66+ years 1324 (64) 1119 (64) 205 (59) 372 (57) 430 (71) immune response, their management and impact on survival vary across cancer types
PS
0-1 1822 (88) 1523 (88) 299 (87) 607(93) 500(83)
and call for cancer-specific strategies for early identification and management of KirAEs
2 and + 258 (12) 211 (12) 47 (14) 48 (7) 101 (17) in the ambulatory care setting. Research Sponsor: Bristol Myers Squibb.
Active Autoimmune 180 (9) 153 (9) 27 (8) 54 (8) 51 (8)
Hypothyroidism 113 (9) 97 (9) 16 (8) 28 (8) 34 (8)
Type I Diabetes 15 (1) 10 (1) 5 (1) 5 (1) 4 (1)
Psoriasis 11 (1) 9 (1) 2 (1) 3 (0) 3 (0)

2602 Poster Session 2603 Poster Session

Hepatotoxic adverse events with immune checkpoint inhibitors: Real world Impact of body mass index on immunotherapy outcomes and complications
pharmacovigilance study using FAERS database. First Author: Panah Tushar in solid tumor patients: A real-world evidence analysis. First Author: Moath
Parab, Saint Vincent Hospital, Worcester, MA Albliwi, Cleveland Clinic Foundation, Cleveland, OH
Background: Immunotherapy with immune checkpoint inhibitors (ICIs) has revolution- Background: Obesity alters immune function by modifying cytokine profiles and altering
ized cancer treatment. With their increasing use, it is important to track and manage immune cells. Body mass index (BMI) influences cancer outcomes, including response to
potential adverse events (AEs) . One such AE of ICI therapy is immune-mediated liver therapy. Several studies have shown that patients (pts) with a higher BMI respond better to
injury (ILICI). We aim to review the real-world data on ILICI using FDA Adverse Event immunotherapy (IT). This study assesses the impact of BMI on IT outcomes, admission risk,
Reporting System (FAERS) database. Methods: We queried FAERS using a search-by- and major complications in pts with solid tumors. Methods: We utilized TriNetX, a global
product strategy on 22nd January 2025 and retrieved 224889 adverse events from 2013- data platform from 104 healthcare institutions, to analyze outcomes in cancer pts on IT.
2024. We employed 5 ICIs in the analysis (Pembrolizumab, Nivolumab, Atezolizumab, Patients were categorized into two groups: BMI , 25 (n = 8,460) and BMI $ 25 (n = 13,631).
Durvalumab, and Ipilimumab). Descriptive statistics were carried out, and dis- Propensity Score Matching balanced groups for age, sex, race, comorbidities, smoking
proportionality analysis was done by calculating the reportable odds ratio (ROR) with 95% status, and alcohol use. Pts aged 18–65 years with solid tumors who received $1 IT dose
confidence intervals (CI). ROR was considered significant when the lower limit of the 95% were included with a 1-year follow-up. IT regimens included PD-L1, PD-1, or CTLA-4 an-
CI was . 1. RORs were calculated for all hepatic events in general and Autoimmune tibodies. Cancers analyzed: esophagus, bladder, stomach, endometrium, melanoma, lung,
Hepatitis (AIH), Drug-induced liver injury (DILI), Vanishing bile duct syndrome (VBDS), kidney, head and neck, and breast. Outcomes included: ICU admissions, hospital admissions,
Primary biliary cholangitis (PBC), and Venooccluisve disease (VOD). Results: Total AEs mortality, heart failure, ischemic stroke/transient ischemic attack (TIA), venous throm-
from all included ICIs were 224889 and Hepatobiliary AEs constitute 9.2% of all AEs boembolism (VTE), myocardial infarction, polyneuropathy, pneumonitis/pneumonia, and
across all ICIs. ROR for any hepatic event is highest with Durvalumab i.e., 17.97 acute kidney injury (AKI). Risks were assessed using 1-year event-free survival and survival
(16.08,20.08) in general as compared to the rest of the ICIs. AIH was seen highest with analysis. Results: After 1:1 PSM, the two groups each consisted of 8,460 pts, with balanced
Ipilimumab with a ROR of 48.0 (43.1, 53.5); DILI with Pembrolizumab with a ROR of 5.8 baseline variables. Post-matching, the mean age was ~52.5 6 9 years, 56.0% were White,
and 54.7% were female. A BMI , 25 was found to be a predictor of increased risk for
(5.3, 6.4) (Overlapping CI); VBDS and PBC with Pembrolizumab with a ROR of 7.93 (4.9,
multiple adverse outcomes. The 1-year risk-free survival was significantly lower in pts with
12.8) and 7.91 (4.46,14.03) respectively. Atezolizumab showed the highest ROR of 6.15
BMI , 25 compared to those with BMI $ 25 for ischemic stroke/TIA (94.6% vs. 95.9%, log-
(3.6, 10.4) for VOD. (Table) Conclusions: This is the largest real-world study demon-
rank P , 0.01), ICU admissions (83.3% vs. 89.05%, P , 0.01), hospital admissions (37.5%
strating specific hepatotoxic AEs with ICIs. Our results show varying patterns of hep-
vs. 48.1%, P , 0.01), mortality (64.01% vs. 80.62%, P , 0.01), heart failure (84.3% vs. 87.6%,
atotoxicity with ICIs. Knowing these patterns will help us make better decisions in treating
P , 0.01), and pneumonitis/pneumonia (79.1% vs. 84.9%, P , 0.01). However, there was no
patients with ICIs. Research Sponsor: None. significant difference in the incidence of VTE, myocardial infarction, polyneuropathy, or AKI
Baseline characteristics, hepatic AEs and outcomes. (P . 0.05). We performed several sensitivity analyses using different BMI cutoff groups and
Pembrolizumab Nivolumab Atezolizumab Durvalumab Ipilimumab compared outcomes to the BMI , 25 group. In these balanced comparisons, we found
Baseline characteristic (n= 67603) (n=79283) (n=28521) (n=13303) (n= 36179)
similar trends except for an increased incidence of polyneuropathy in BMI $ 35 compared to
ROR for any hepatic 8.24 (7.66,8.87) 7.46 (6.95, 8.01) 7.94 (7.08, 8.89) 17.97 (16.08, 20.08) 11.47 (10.54, 12.48) BMI , 25 (1-year risk-free: 89.2% vs. 91.6%, P , 0.01) and in BMI $ 40 compared to BMI ,
event (95% CI)
ROR for AIH 21.34 (18.9, 24.0) 27.3 (24.8, 30.1) 22.3 (18.7, 26.5) 22.3 (18.7, 26.5) 48.0 (43.1, 53.5) 25 (1-year risk-free: 89.0% vs. 91.4%, P , 0.01). Conclusions: Our study provides real-world
ROR for DILI 5.8 (5.3, 6.4) 3.74 (3.34, 4.19) 5.3 (4.61,6.29) 5.3 (4.61, 6.29) 5.04 (4.37, 5.82) evidence on the types of complications experienced by pts with BMI , 25 when treated with
ROR for VBDS 7.93 (4.9, 12.8) 4.37 (2.41, 7.94) 0.5 (0.03, 8.8) 1.1 (0.15, 7.86) 2.61 (0.84, 8.1) IT for different types of solid tumors. It also explains, at least partly, the improved outcomes
ROR for VOD 2.22 (1.2, 3.9) 3.79 (2.5, 5.6) 6.15 (3.6, 10.4) 0.44 (0.06, 3.12) 2.08 (0.9, 4.63)
ROR for PBC 7.91 (4.46,14.03) 3.37 (1.5, 7.5) 0.78 (0.04, 12.5) 1.51 (0.21, 11.1) 2.46 (0.6, 9.8) and tolerability observed in pts with higher BMI receiving IT. Based on our findings—such as
the increased risk of hospital admissions and pneumonitis—we postulate that pts with low
Hepatic AEs included in the ROR calculation with these agents are Drug-Induced Liver Injury (DILI), Autoimmune hepatitis
(AIH), Vanishing bile duct syndrome (VBDS), Veno-occlusive disease (VOD), Primary biliary cholangitis (PBC). BMI may have a stronger inflammatory reaction to IT, leading to a higher incidence of
complications. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 163s

2604 Poster Session 2605 Poster Session

Final analysis of a multicenter, open-label, phase 2 study evaluating the Effects of UCHL1 on tolerogenic DC maturation and promotion of mregDC-
efficacy and safety of tislelizumab (TIS) in combination with fruquintinib (F) Treg crosstalk to nullify anti–PD-L1 therapy. First Author: Yu-Fei Zhao,
in patients (pts) with selected solid tumors. First Author: Keun-Wook Lee, Division Zhongshan Hospital Fudan University, Shanghai, Shanghai, China
of Hematology and Medical Oncology, Department of Internal Medicine, Seoul National Background: Immune-checkpoint blockade (ICB) therapies have revolutionized cancer
University Bundang Hospital, Seoul National University College of Medicine, Seongnam, treatment, but such immunotherapy regimens fail in a subset of patients. Dendritic cells
South Korea (DCs) are a heterogeneous group of professional antigen-presenting innate immune cells
Background: Immunotherapy in combination with antiangiogenic agents has shown promising antitumor that activate adaptive immunity and determine the efficacy of immunotherapies. While
activity compared with either agent alone. We report efficacy and safety data from the final analysis of the they can also be hijacked by tumour-mediated factors to contribute to immune tolerance
phase 2 BGB-A317-Fruquintinib-201 trial evaluating the programmed cell death-1 antibody TIS combined
with the selective vascular endothelial growth factor receptor (VEGFR)-1, -2, and -3 inhibitor F in pts with
and tumor progression. However, little is known about the molecular mechanisms that
advanced solid tumors. Methods: This was an open-label, multicenter, two-part study with a safety run-in drive the tolerogenic maturation of DCs in the tumor microenvironment (TME).
followed by dose-expansion. Eligible pts were adults with advanced or metastatic unresectable gastric Methods: We enrolled 85 patients with advanced hepatocellular carcinoma (HCC)
cancer (GC), microsatellite stable colorectal cancer (MSS CRC), or locally advanced surgery-/radiotherapy- exhibiting varying response to immunotherapy, and profiled the tumor ecosystems
ineligible and programmed death ligand-1–positive (PD-L1+; defined as PD-L1 $1%) stage IIIB/IV non- using a single-cell transcriptomes sequencing (scRNA-seq) and mass cytometry by time
small cell lung cancer (NSCLC). F 5 mg daily (3 weeks on, 1 week off) plus TIS (300 mg IV Q4W) was
of flight (CyTOF) for 10 patients, and plasma protein level quantification, conducted both
administered as second-line therapy for pts with GC, third-line therapy for pts with MSS CRC, and first-line
therapy for pts with PD-L1+ NSCLC. The primary outcome measure was overall response rate (ORR) per pre-treatment and post-treatment across all patients. We integrated our in-house data
RECIST v1.1. Secondary endpoints included other efficacy measures and safety. Results: The median and 6 additional published scRNA-seq cohorts of 83 donors to generate a compre-
study follow-up was 11.6 months (mo; range, 0.4-32.8). A total of 84 pts were enrolled (GC, n=31; MSS CRC, hensive landscape of cellular dynamics underlying different responses to immuno-
n=31; PD-L1+ NSCLC, n=22). One study treatment component-related death was reported in the GC cohort therapy. We verified the prognostic value in our in-house tumor microarray (TMA) of 342
and 1 in the PD-L1+ NSCLC cohort. The recommended phase 2 dose was established at F 5 mg daily patients. Results: UCHL1 overexpression nullifies anti-PD-L1 therapy by driving con-
(3 weeks on, 1 week off) in combination with TIS with no observed dose-limiting toxicities. Efficacy and
safety are reported in the Table. Any-grade treatment-emergent adverse events (TEAEs) occurred in 83
ventional DC transformation into mature DC enriched in immunoregulatory molecule
(98.8%) pts; proteinuria (32.1%), hypoalbuminemia (27.4%), and hypothyroidism (25.0%) were most (mregDC) via tolerogenic maturation and promoting mregDC and regulatory T (Treg) cell
common. 9/32 (10.7%) pts had grade $3 immune-mediated AEs. Conclusions: Despite the limited sample crosstalk, thereby restrains CD8+ T anti-tumor immunity. Mechanistically, UCHL1 en-
size, TIS+F demonstrated moderate antitumor activity in pts with advanced solid tumors, with manageable hances glycolysis and lactate accumulation in TME by stabilizing HIF-1a, which further
safety observed in pts with GC and MSS CRC. Further investigation of TIS+F is warranted in the GC and MSS promotes SREBP2 activation and nuclear translocation in DC. We verified the positive
CRC settings. Clinical trial information: NCT04716634. Research Sponsor: BeOne Ltd. correlations of UCHL1 with HIF-1a/VEGFa/LAMP3/FOXP3 in 342 patients with HCC.
MSS CRC PD-L1+ NSCLC Genetic ablation or pharmacological inhibition of UCHL1 all reduce the mregDC and Treg
GC (N=31) (N=31) (N=22) accumulation, restore the immuno-surveillance of tumour-infiltrating lymphocytes, and
ORR, n (%) 4 (12.9) 3 (9.7) 9 (40.1) safeguard anti-tumour immunity and efficacy of anti-PD-L1 therapy in mouse models.
Disease control rate, n (%) 23 (74.2) 23 (74.2) 15 (68.2)
Clinical benefit rate, n (%) 10 (32.3) 12 (38.7) 13 (59.1)
Conclusions: UCHL1 hijacks tolerogenic DC maturation and promotes mregDC-Treg
Median progression-free 4.6 (3.4, 7.4) 4.6 (3.6, 7.2) 15.6 (1.8, NE) crosstalk to nullify anti-PD-L1 therapy. Genetic ablation or pharmacological inhibition of
survival, mo (95% CI) UCHL1 unleash the immuno-surveillance of tumour-infiltrating lymphocytes, and
Median overall survival, 10.5 (5.2, 14.6) 10.0 (4.7, 15.2) NR (6.0, NE)
mo (95% CI) safeguard the anti-tumor immunity. Plasma level of UCHL1 predicts the efficacy of anti-
Median duration of response, NR (5.6, NE) 11.9 (3.7, NE) NR (7.7, NE) PD-L1 therapy in patients with HCC. Clinical trial information: NCT04649489. Research
mo (95% CI)
Grade ‡3 TRAE, n (%) 10 (32.3) 12 (38.7) 14 (63.6) Sponsor: None.
Serious TRAE, n (%) 3 (9.7) 3 (9.7) 9 (40.9)
TEAE leading to discontinuation 5 (16.1) 3 (9.7) 7 (31.8)
of any study treatment, n (%)
CI, confidence interval; NE, not evaluable; NR, not reached.

2606 Poster Session 2607 Poster Session

A phase 1b/2, open-label study of selective Axl, Mer and CSF1R inhibitor Outcomes of conversion surgery after immune checkpoint inhibitor-based
adrixetinib (Q702) in combination with intravenous pembrolizumab in pa- combination therapy in initially unresectable hepatocellular carcinoma: A
tients with selected advanced solid tumors: Results of a phase 1 study retrospective cohort study. First Author: Mingjian Piao, Department of Liver Surgery,
(QRNT-008). First Author: Hong Jae Chon, CHA Bundang Medical Center, CHA Uni- State Key Laboratory of Complex Severe and Rare Diseases, Peking Union Medical
versity, Gyeonggi-Do, Korea, Republic of College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical
Background: Adrixetinib (Q702) is an orally administrated novel Axl/Mer/CSF1R ty- College, Beijing, China
rosine kinase inhibitor for which the primary mechanism of action of tumor regression is Background: Hepatocellular carcinoma (HCC) has a high incidence rate and is often
through immune-stimulating effects. The safety profile, pharmacokinetics (PK) and asymptomatic in its early stages. Combination therapies using immune checkpoint
efficacy data for Q702 in combination with pembrolizumab are presented. inhibitors (ICIs) have demonstrated survival benefits and high objective response rates,
Methods: QRNT-008 (NCT05438420) is an ongoing Phase 1b/2 multicenter, open-label, offering hope for conversion surgery in patients with initially unresectable HCC. We
dose escalation and expansion study in patients with advanced esophageal, gastric/GEJ, aimed to investigate the oncological outcomes of conversion surgery compared to those
hepatocellular, and cervical cancers who have progressed on prior anti-PD-1/PD-L1 with continuing systemic treatment alone in patients who responded well to ICIs-based
treatment. The Part 1 dose escalation was guided by a mTPI design to determine the therapy, as well as the surgical outcomes associated with conversion surgery.
Part 2 dose of Q702 in combination with pembrolizumab. Patients received Q702 (week Methods: We consecutively enrolled patients diagnosed with HCC between January 1,
on/off dosing regimen) orally at 100 mg or 120 mg doses in combination with pem- 2019 and April 1, 2024. These patients received treatment with ICIs combined with either
brolizumab (200 mg Q3W) intravenously in 42-day cycles. Results: As of the data cutoff anti-VEGF antibodies or tyrosine kinase inhibitors. Tumor response and resectability
(December 19th, 2024), 29 patients received Q702 plus pembrolizumab across 2 dose were assessed every 2 months. Patients who responded positively and met the criteria
levels: 7 patients at 100 mg and 22 patients at 120 mg. The median number of prior lines for conversion surgery were included. Results: Among 613 patients with initially
of systemic therapy was 4 (range 1-7). Of the 29 patients (3 esophageal; 11 gastric; 2 unresectable HCC, 136 achieved conversion and met the surgical resection criteria
GEJ; 9 hepatocellular; 4 cervical) who received Q702 across all doses, there were no during combination therapy. The median follow-up time was 26.9 and 42.5 months for
treatment discontinuations due to the treatment-related AEs (TRAEs). Most common the surgery and non-surgery groups, respectively. The median PFS was 29.1 months in
TRAEs $10% were AST increase (51.7%), ALT increase (41.3%), CPK increase (37.8%) the surgery group versus 11.2 months in the non-surgery group (P , 0.0001, hazard
and LDH increase (34.5%). One patient dosed at 120 mg experienced 1 DLT (G3 skin rash ratio [HR] = 0.40 [0.25–0.63]). The median OS was 50.8 months in the surgery group,
and G3 diarrhea). Adrixetinib PK analyses showed dose dependent increase of AUC0-last compared to 25.8 months in the non-surgery group (P , 0.0001, HR = 0.27 [0.15–0.47]).
and Cmax. Overall response assessment (RECIST 1.1) included 1 confirmed complete The median RFS was 18.7 months in the surgery group. Multivariate Cox regression
response (CR) in a patient with metastatic gastric cancer (GC) and 6 patients with stable analysis indicated that conversion surgery was independently associated with improved
disease (SD) across multiple tumor types. Among 6 SD patients, 1 GC and 1 hepato- OS and PFS (P , 0.001), and continuing the original treatment post-surgery significantly
cellular cancer (HCC) patient continued treatment for $24 weeks. influenced OS and RFS. Conclusions: Conversion surgery after meeting the surgical
Conclusions: Preliminary data from QRNT-008 study showed that selective Axl/Mer/ criteria during immunotherapy provides significant prognostic benefits for patients with
CSF1R inhibitor Q702 plus pembrolizumab has a manageable safety profile. The Part 2 initially unresectable HCC, demonstrating high safety and R0 resection rates. For those
dose of Adrixetinib is confirmed at 120 mg. Preliminary anti-tumor activity in patients specifically selected based on their response to immunotherapy and undergoing
previously treated with anti-PD-1 supports further development of the combination. conversion surgery, promptly resuming the original treatment after surgery is necessary.
Clinical trial information: NCT05438420. Research Sponsor: Qurient Co., Ltd. Our results emphasize the importance of continuing immunotherapy post-conversion
surgery to prevent recurrence in patients who respond to immunotherapy. Research
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
164s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2608 Poster Session 2609 Poster Session

Phase I/II study of the EP4 antagonist vorbipiprant combined with anti-PD-1 Bispecific innate cell engager (ICE) AFM24 in combination with atezolizu-
immunotherapy: Safety and efficacy results in metastatic gastrointestinal mab in patients with advanced/metastatic EGFR-expressing non-small cell
non-colorectal cancers. First Author: Filippo Pietrantonio, Department of Medical lung cancer (NSCLC) without driver mutations: Initial results from a phase 2a
Oncology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy study. First Author: Hye Ryun Kim, Division of Medical Oncology, Yonsei Cancer Center,
Background: Novel combination strategies are being explored to enhance the effec- Yonsei University College of Medicine, Seoul, South Korea
tiveness of immune checkpoint inhibitors (ICIs). Prostaglandin E2, through its receptor 4 Background: Novel treatments are needed for patients with advanced/metastatic NSCLC
(EP4), is a major contributor to immunosuppression in the tumor microenvironment. In a without actionable driver mutations who progress after prior therapies including checkpoint
dose-response phase I/II study, the EP4 antagonist vorbipiprant (CR6086) combined with inhibitors (CPI) and platinum-based chemotherapy. AFM24 is a tetravalent, bispecific ICE
PD-1 blockade was well tolerated and showed promising efficacy in refractory mismatch- that binds CD16A on NK cells and macrophages and EGFR on solid tumors, redirecting and
repair-proficient/microsatellite stable metastatic colorectal cancer (CRC) (Pietrantonio enhancing the innate and possibly the adaptive immune response. The EGFR-wildtype
et al, Clin Cancer Res 2024). Here we report the results from a study extension in non- (EGFR-WT) NSCLC expansion cohort of the Phase 1/2a study (NCT05109442) is evaluating
colorectal gastrointestinal (GI) cancers with the vorbipiprant dose selected for further the combination of AFM24 and atezolizumab. Methods: AFM24 is given weekly at 480 mg
development in combination with immunotherapy. Methods: Twenty-seven adult pa- intravenously (IV) in combination with 840 mg atezolizumab IV fortnightly to patients with
tients (pts) with metastatic non-colorectal GI cancers, ECOG PS #1, and $1 prior advanced or metastatic EGFR-WT NSCLC who progressed on $1 prior line of therapy,
treatment line were included in 3 cohorts (9 pts each): gastric cancer (GC) with PD-L1 including at least a platinum doublet and a CPI. The primary endpoint is overall response rate
Combined Positive Score (CPS) $5 (cohort A), GC with PD-L1 CPS , 5 (cohort B), and GI (ORR) by RECIST v1.1 by Investigator assessment. Secondary endpoints include safety,
cancers other than CRC and GC (cohort C). Pts receive oral vorbipiprant (90 mg twice pharmacokinetics, and immunogenicity. Treatment is given in 28-day cycles until disease
daily) plus iv balstilimab (3 mg/kg every 2 weeks) until disease progression, unacceptable progression, intolerable toxicity, investigator discretion, or patient withdrawal of consent.
toxicity or death. Primary endpoints are safety and disease control rate (DCR) per RECIST Results: As of 15 January 2025, 43 patients received AFM24 and atezolizumab for a mean
1.1. Secondary endpoints include objective response rate, progression-free and overall (range) duration of 19.6 (1–78) weeks. Median (range) age is 67 (40–79) years; 72% male; all
survival (ORR, PFS, OS). Exploratory endpoints include tissue and blood biomarkers. patients had an ECOG performance status of 0 (14%) or 1 (86%). Median (range) number of
Results: At a cutoff date of November 20, 2024, enrolment is completed. In cohort C, we prior lines is 2 (1–7). All patients had discontinued their previous CPI treatment due to
enrolled: 5 BTC, 2 pancreatic and 2 ampullary cancer patients. Overall, median age was 61 progressive disease. The combination was well tolerated with no unexpected toxicities;
infusion-related reactions, the most common adverse events (AE), were reported in 54% of
(interquartile range: 55-68) years, similar among cohorts; 70% were men, with a slightly
patients (28 Grade 1–2, 4 Grade 3). Most common $G3 treatment-related AEs were ALT/
higher prevalence in Cohort A; the median number of prior treatment lines was 3 (IQR: 2-4)
AST elevations in 2 patients, all fully resolved. The 35 response-evaluable patients showed
overall and in gastric cohorts, and 2 (IQR: 2-3) in other GI cancers cohort. Prior ICIs were
an ORR of 23% (8 responses: 1 complete response, 7 partial responses), tumor shrinkage in
administered in 44%, 22% and 11% in Cohort A, B and C, respectively. No treatment-related
46% (16/35) and a disease control rate (DCR) of 77%. Of the 8 responders, 6 had never
serious or grade . 3 adverse events were reported. Promising activity was observed. In
achieved an objective response on prior CPIs. Preliminary median progression-free survival
cohort A, 3 pts had a partial response (PR), 2 of them still ongoing and 2 lasting more than (PFS) is 5.5 months (95% CI 2.9–7.4), with 29% of patients still on treatment.
6 months; in addition, 1 pt had stable disease (SD). In cohort B, 4 pts had SD, 1 of them still Conclusions: AFM24 in combination with atezolizumab shows promising clinical efficacy in
ongoing and 2 lasting more than 6 months. In cohort C, 1 pt with pancreatic cancer had a patients who failed prior treatment including platinum-based chemotherapy and CPI.
PR, still ongoing for . 6 months; in addition, 1 BTC patient had SD. Median PFS and OS Patients showed a tolerable and well-managed safety profile. A considerable DCR of 77%,
were: 4,5 and 9,7 months in Cohort A, 1,8 and 6,8 months in Cohort B, 2,0 and 4,5 months in with some long, sustained responses was observed. Confirmed responses were achieved in
Cohort C. Responses occurred irrespective of MSI/MMR status and prior exposure to ICIs. patients who had not responded to prior CPI. This combination treatment approach could
Conclusions: Vorbipiprant combined with PD-1 blockade was well tolerated and showed offer a promising chemotherapy-free alternative to patients who have exhausted the
signs of activity in non-colorectal GI cancers, thus confirming a broader spectrum of available therapeutic options and could provide a strategy to overcome resistance to prior
activity on top of the results in MSS CRC. Clinical trial information: NCT05205330. CPI. Clinical trial information: NCT05109442. Research Sponsor: Affimed GmbH.
Research Sponsor: Rottapharm Biotech.

2610 Poster Session 2611 Poster Session

Combination of bispecific innate cell engager (ICE) AFM24 with atezolizu- Primary efficacy and safety results of BAT1308, a PD-1 inhibitor, + che-
mab in patients with advanced/metastatic non-small cell lung cancer motherapy 6 bevacizumab in phase 2 trial for persistent, recurrent, or
(NSCLC) with EGFR kinase domain mutations (EGFRmut): Initial results metastatic cervical cancer. First Author: QingLei Gao, Tongji Hospital, Tongji
from a phase 2a study. First Author: Omar Saavedra, Vall d’Hebron Institute of Medical College, Huazhong University of Science and Technology, Wuhan, China
Oncology (VHIO), Barcelona, Spain Background: BAT1308 is a fully humanized and high-affinity anti-PD-1 IgG4k antibody.
Background: Immune checkpoint inhibitor (ICI) monotherapy has shown limited activity Previous phase 1 study demonstrated BAT1308 had a promising efficacy in patients with
against advanced EGFRmut NSCLC. However, combinatorial approaches may enhance the advanced cervical cancer. Here we present the primary safety and efficacy results in phase 2
clinical outcomes and are under evaluation. AFM24 is a tetravalent, bispecific ICE that binds study for BAT1308 combined with platinum-based chemotherapy 6 bevacizumab as first-
CD16A on NK cells and macrophages and EGFR on solid tumors, redirecting and enhancing line therapy for PD-L1–positive persistent, recurrent, or metastatic cervical cancer.
immune responses towards EGFR-expressing tumors. Atezolizumab, an anti-PD-L1 anti- Methods: In this multicenter, single-arm, open-label, phase 2 study, eligible patients
body, has been approved in patients with various solid tumors. The EGFRmut NSCLC ex- were $18 to #75 years of age with PD-L1 CPS $1, FIGO Stage IVB cervical cancer, who did
pansion cohort of this Phase 1/2a study explores a possible synergistic effect of AFM24 in not receive prior systemic anti-tumor therapy for persistent, recurrent or metastatic cervical
combination with atezolizumab in heavily pretreated patients with NSCLC EGFRmut cancer and not amenable to curative treatment. Patients received BAT1308 (300 mg Q3W
(NCT05109442). Methods: AFM24 is given weekly at 480 mg intravenously (IV) in com- for up to 24 months) plus platinum-based chemotherapy (paclitaxel 175 mg/m2 + cisplatin
bination with 840 mg atezolizumab IV fortnightly to patients with advanced or metastatic 50 mg/ m2 or carboplatin AUC 5) and, per investigator discretion, bevacizumab (15 mg/kg).
EGFRmut NSCLC who progressed on $1 prior line of therapy, including $1 prior TKI. The The primary endpoint was safety. The major secondary endpoint was objective response
primary endpoint is overall response rate (ORR) by RECIST v1.1 by Investigator assessment. rate assessed by investigator according to RECIST 1.1. Results: As of January 7, 2025, a
Secondary endpoints include safety, pharmacokinetics, and immunogenicity. Treatment is total of 29 patients were enrolled, with a median age of 53 years (range 32-69), 20 (69.0%)
given in 28-day cycles until disease progression, intolerable toxicity, investigator discretion, patients had ECOG performance status of 1, 15 (50.7%) patients with PD-L1 CPS $ 10, 24
or patient withdrawal of consent. Results: As of 15 January 2025, 28 patients received (82.8%) patients had squamous-cell carcinoma, 17 (58.6%) patients received previous
AFM24 and atezolizumab for a mean (range) duration of 21.7 (2–65) weeks. Median (range) neoadjuvant or adjuvant chemotherapy or chemoradiotherapy with a paclitaxel + platinum
age is 65 years (32–83); 67.9% were female. All patients had received prior EGFR-specific regimen, 7 (24.1%) patients had previous untreated metastatic disease at trial entry.
TKI, 82% had received platinum-based chemotherapy and 75% 3rd gen TKIs. Patients Bevacizumab was used by 23 (79.3%) patients in this phase II study. All 29 subjects received
received a median (range) of 3 (1–8) prior lines of treatment. The combination was well combination therapy. 27 subjects completed at least one efficacy assessment. The ORR was
tolerated with no new or unexpected toxicities observed compared to each single agent. The 74.1%, with a confirmed ORR of 70.4%. The complete response rate was 11.1%, and the
most common treatment-related adverse events (TRAE) were infusion-related reactions in disease control rate was 100%. Currently, 16 subjects remain on treatment. Among those
64% of patients (19 Grade 1–2, 1 Grade 3). 9 patients had $G3 TRAEs, the most common who discontinued the study, 8 withdrew informed consent, 4 experienced disease pro-
being neutropenia/neutrophil count decrease, with no associated infections. No other gression, and 1 died. The 6-month, 9-month, and 12 -month PFS rates were 83.4%, 78.8%,
immune TRAEs were reported. The 22 response-evaluable patients achieved an ORR of 23% and 78.8% respectively. The median PFS has not yet been reached. The most common
(1 CR, 3 PRs, 1 unconfirmed PR), a DCR of 64% and tumor shrinkage in 50% of patients. adverse events were anemia (82.8%), white blood cell decreased (51.7%), alopecia (51.7%),
Responses were deepening over time in 3 patients. With a median follow-up of 9 months, the thrombocytopenia (48.3%), and neutropenia (44.8%). Grade 3 and above adverse events
median PFS was 5.5 months (95% CI 1.9–not-evaluable). 6 (27%) patients have received occurred in 72.4% of 29 patients, and $ Grade 3 irAEs observed in 3 (10.3%) patients.
treatment for over 10 months. Conclusions: AFM24 combined with atezolizumab dem- Serious adverse events occurred in 44.8% of the patients. Conclusions: BAT1308 combined
onstrated encouraging clinical efficacy in patients with EGFRmut NSCLC who had exhausted with platinum-based chemotherapy 6 Bevacizumab as first-line therapy showed durable
prior lines of therapy. Treatment showed a well-managed safety profile. This approach anti-tumor activity and manageable safety profile for PD-L1-positive (CPS $ 1) persistent,
potentially offers a feasible, chemotherapy-free therapeutic option for the EGFRmut NSCLC recurrent or metastatic cervical cancer. These data are consistent with the earlier results
patients who have progressed to prior TKIs and platinum-based chemotherapy and warrants and provide support for further studies. Clinical trial information: NCT06123884. Research
further evaluation. Clinical trial information: NCT05109442. Research Sponsor: None. Sponsor: Bio-Thera Solutions, Ltd.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 165s

2612 Poster Session 2613 Poster Session

Neoadjuvant serplulimab with concurrent chemoradiotherapy in resectable Niraparib plus PD-1 inhibitor for patients previously treated with immune
esophagogastric junction adenocarcinoma: Phase 2 updated results. First checkpoint inhibitor for solid tumors with homologous recombination repair
Author: Yuping Ge, Peking Union Medical College Hospital, Beijing, Beijing, China gene mutation (IMAGENE): A phase II basket study. First Author: Taigo Kato,
Background: This study evaluated the efficacy and safety of neoadjuvant Serplulimab Department of Urology, Osaka University Graduate School of Medicine, Suita, Japan
combined with concurrent chemoradiotherapy for locally advanced resectable Background: Prior clinical trials have established the effectiveness of poly (ADP-ribose)
esophagogastric junction (EGJ) adenocarcinoma. Methods: Eligible patients with re- polymerase (PARP) inhibitor (PARPi) or immune checkpoint inhibitor (ICI) monotherapy in
sectable EGJ（cT3-4 or N+M0）adenocarcinoma received neoadjuvant Serplulimab patients with cancer characterized by mutations in homologous recombination repair (HRR)
(300 mg) plus SOX (oxaliplatin 130 mg/㎡; TS1 40-60 mg) for the first cycle, followed by genes. This trial aims to evaluate the efficacy and safety of PARPi and PD-1 inhibitor in
Serplulimab with concurrent chemoradiotherapy (oxaliplatin 100 mg/㎡, TS1 40-60 mg; patients with HRR gene-mutated solid tumors previously treated with ICIs.
radiotherapy dose 45 Gy/ 25 fractions) during the second and third cycles. Surgery was Methods: IMAGENE is an open-label phase II basket study evaluating the efficacy and
performed 6-8 weeks after chemoradiotherapy. Tissue and blood samples were col- safety of niraparib and PD-1 inhibitor in patients with HRR gene-mutated cancers that have
lected for genetic analysis. Primary endpoints included pathological complete response shown resistance to one or more standard therapy including ICIs. HRR mutation status was
(pCR) and major pathological response (MPR). Results: From March 2023 to November evaluated by circulating tumor DNA (ctDNA) or tumor tissue DNA. The primary endpoint was
2024, 24 patients were enrolled and 19 patients underwent radical resection. The R0 rate confirmed objective response rate (cORR) per investigator assessment. Patients were
was 100%. pCR rate was 26.3%, MPR rate was 36.8%. T downstaging rate 78.9%, ypN0 treated on a 21-day cycle with niraparib (200 mg orally daily) and nivolumab/pembrolizumab
(240 mg/body intravenously every 2 weeks or 3 mg/kg every 3 weeks, respectively).
89.5%. The median DFS was not reached. Microsatellite stable status was 100%. PD-L1
Results: Of 47 enrolled patients, 22 were enrolled based on HRR gene alteration detected in
CPS expression: , 1 (5.3%), 1-5 (42%), . 5 (52.6%), . 10 (31.6%). PD-L1 expression
ctDNA. The most common tumor type was gastric cancer (36.2%), followed by bladder
was associated with pathological response, with MPR rates of 57.1% for CPS $5 and
cancer (BC, 25.5%), and renal cell carcinoma (RCC, 12.8%). As of data cutoff (September 30,
14.3% for CPS , 5. Minimal residual disease positivity (MDR+) before enrollment was 2024), the median duration of follow-up was 31.7 weeks. The median number of prior
68.7%, and 6 MRD+ patients converted to MRD- after neoadjuvant therapy. Grade $3 treatment line was 6.5 (1 to 20). In total, the cORR was 4.4% (90% CI, 0.8 to 13.3), with a
adverse events occurred in 33.3% of patients, with manageable treatment-related disease control rate (DCR) of 55.6% (90% CI, 42.3 to 68.3). Notable DCRs were observed in
adverse events. Conclusions: Neoadjuvant Serplulimab with concurrent chemo- patients with RCC (83.3%) and BC (80.0%). The median duration of response and
radiotherapy showed promising efficacy for locally advanced resectable EGJ adeno- progression-free survival (PFS) were 35.0 weeks (90% CI, 20.0 to 50.0) and 11.6 weeks (90%
carcinoma. Improved R0 and ypN0 rate may change the surgical procedure in the future. CI, 7.0 to 13.6), respectively. At data cutoff, 28 patients (62.2%) had died, with 12-month
Follow-up will assess correlations between biomarkers and outcomes. Clinical trial overall survival (OS) rate of 41.7% (90% CI, 29.1 to 53.8). Interestingly, BRCA1/2-mutated
information: NCT05918419. Research Sponsor: None. patients had significantly shorter OS compared to those with other HRR gene mutations (p =
0.0096). The three most common treatment-emergent adverse events (TEAEs) were nausea
(31.1%), vomiting (31.1%), and anaemia (26.7%). Grade $3 TEAEs were reported in 18
patients (40.0%), with the most common being anaemia (17.8%). TEAEs led to treatment
regimen discontinuation in one patient, and there were no deaths due to TEAEs.
Conclusions: Niraparib combined with PD-1 inhibitor showed modest activity even in
heavily pretreated patients with HRR gene-mutated cancers who progressed on ICI therapy.
Furthermore, patients with specific cancer type had promising benefit from this combination
therapy, warranting further investigation in specific populations. Clinical trial information:
jRCT2051210120. Research Sponsor: The Japan Agency for Medical Research and De-
velopment; 21ck0106656h0001; Takeda Pharmaceutical Co., Ltd.

2614 Poster Session 2615 Poster Session

Stereotactic radiotherapy plus immunotherapy and influence on prognosis A randomized controlled study of tislelizumab combined with concurrent
in driver-gene–negative non-small cell lung cancer patients with brain oligo- chemoradiotherapy in the treatment of locally advanced cervical cancer and
metastases. First Author: Xiaomei Gong, Tongji University Affiliated Shanghai Pul- the predictive value of T lymphocyte subsets for efficacy. First Author: Fang
monary Hospital, Shanghai, China Wu, Department of Radiation Oncology, the First Affiliated Hospital of Guangxi Medical
Background: This study seeks to elucidate the therapeutic benefits of integrating ste- University, Nanning, Guangxi, China
reotactic radiotherapy (SRT) with immunotherapy for treating brain oligo-metastases (BMs) Background: Concurrent chemoradiotherapy (CCRT) has been the standard of care for
in patients with non-small cell lung cancer (NSCLC). Methods: In this retrospective real- locally advanced cervical cancer (LACC) for over 20 years. However, 30-40% of treated
world study, patients with driver-gene-negative NSCLC and 1-3 BMs were enrolled to patients have recurrence or progression within 5 years. Methods: A total of 53 patients with
evaluate the therapeutic benefits of combining SRT with immune checkpoint inhibitors LACC who were treated in the First Affiliated Hospital of Guangxi Medical University from
(ICIs) and chemotherapy. The primary endpoint was overall survival (OS). Secondary May 2023 to November 2024 were prospectively collected and randomly divided into the
endpoints included intracranial progression-free survival (iPFS), progression-free survival CCRT group (N = 26) and the CCRT+T group (N = 27). The treatment plan was as follows:
(PFS), and the response of intracranial lesions. Results: Based on chemotherapy (CT), 65 200 mg of tislelizumab was intravenously infused on the first day of radiotherapy, once every
patients underwent SRT+ICIs therapy, 47 patients underwent SRT, and 44 patients un- 3 weeks, for 1 year or until disease progression or intolerable toxicity, whichever occurred
derwent ICIs. For patients with with . 500 mm3 BMs, SRT + ICIs + CT significantly improved first. Chemotherapy involved single-agent cisplatin at a dose of 40 mg/m2. External irra-
the OS (22.1 vs. 13.5 vs. 18.5 months, p = 0.012), iPFS (17.5 vs. 7.8 vs. 11.8 months, p , diation used 6MV-X-ray intensity-modulated radiotherapy with a dose of 45-50Gy/25f.
0.001), PFS (11.3 vs. 7.6 vs. 5.3 months, p = 0.019), and iORR (56.3% vs. 20.3% vs 28.9%, p = Simultaneously, peripheral blood T lymphocyte subsets were detected in the enrolled
0.001) compared to SRT+CT or ICIs + CT therapy. In the sub-group of patients of symp- patients before and at the end of radiotherapy. The primary endpoint of the study was the
tomatic BMs, SRT + ICIs + CT significantly improved the OS (24.7 vs. 14.7 vs. 17.5 months, p objective response rate, and secondary endpoints included toxicity, PFS and OS.
= 0.012), iPFS (13.7 vs. 9.8 vs. 11.8 months, p = 0.046), and iORR (34.5% vs. 13.8% vs 23.7%, Results: Follow-up was completed by January 2025, with a median follow-up time of
p = 0.027) compared to SRT + CT or ICIs + CT therapy as well. Concurrent SRT with ICIs (time 13.7 months (5.2-20.5 months) in the CCRT group and 11.2 months (5.9-20.6 months) in the
interval , 2 weeks) significantly improved the OS (28.2 vs. 15.4 months, p = 0.01), iPFS CCRT+T group. The CCRT+T group had higher complete response (CR) and objective re-
(25.8 vs. 12.1 months, p = 0.014), and iORR (63.2% vs. 37.0%, p = 0.017) when compared to sponse rates (ORR) compared to the CCRT group, with CR rates of 44.4% versus 19.2% (p =
sequential SRT with ICIs. Combined therapy did not increase the incidence of any grade of 0.035) and ORR rates of 100% versus 84.6% (p = 0.046). The patients with CR were subjected
central nervous system and immune-related adverse events. Conclusions: Based on to logistic regression analysis. The results of univariate and multivariate analysis showed
chemotherapy, the combination of concurrent SRT and ICIs improve the prognosis of driver- that CD4+T cell percentage (p = 0.034) and tislelizumab use (p = 0.038) were significantly
gene-negative NSCLC with BMs without increasing the occurrence of adverse events. associated with CR. However, no statistical difference was observed in the OS (p = 0.414)
Furthermore, it demonstrates increased effectiveness for treating larger and symptomatic and PFS (p = 0.716) between the two groups, as shown by the Kaplan-Meier survival curves.
intracranial lesions, specifically those . 500 mm3 in volume. Research Sponsor: None. After treatment, the CCRT+T group had increased levels of total T cell percentage,
Data on patient outcomes. CD4+T cell percentage, CD4/CD8, double positive T lymphocyte subset percentage, absolute
All Patients Sub-Group in Lesions > 500 mm3
Sub-Group in Symptomatic
Brain Metastases SRT+ICIs+CT Sub-Group
T lymphocyte count, absolute CD4+T lymphocyte count, and B lymphocyte (CD19) compared
SRT+ SRT+
to the CCRT group. Notably, the double positive T lymphocyte subset percent (p = 0.025) and
Median,
months
SRT+ICIs+CT SRT+CT ICIs+CT p-
(n=65) (n=47) (n=44) Value
ICIs+CT SRT+CT ICIs+CT p-
(n=44) (n=28) (n=17) Value
ICIs+CT SRT+CT ICIs+CT p-
(n=38) (n=38) (n=21) Value
Concurrent Sequential p-
(n=31) (n=34) Value
absolute CD4+T lymphocyte count (p = 0.047) showed significant increases. There was no
OS 23.0 14.2 18.7 0.033 22.1 13.5 18.5 0.012 24.7 14.7 17.5 0.012 28.5 15.4 0.01
significant difference in the incidence of acute adverse reactions between the two groups
iPFS 15.3 8.5 13.0 0.002 17.5 7.8 11.8 , 13.7 9.8 11.8 0.046 25.8 12.1 0.014 (P . 0.05). Conclusions: CCRT+T demonstrated superior short-term efficacy in treating
0.001
PFS 9.8 6.7 8.0 0.072 11.3 7.6 5.3 0.019 8.9 6.7 5.7 0.13 12.2 9.0 0.009 LACC compared to CCRT, although long-term efficacy necessitates further follow-up ob-
iORR 48.8% 24.5% 49.0% 0.01 56.3% 20.3% 28.9% 0.001 34.5% 13.8% 23.7% 0.027 63.2% 37.0% 0.017
iDCR 83.3% 75.5% 76.5% 0.46 85.4% 62.1% 72.2% 0.063 22.2% 17.2% 12.7% 0.61 92.1% 76.1% 0.05 servation. The combination of tislelizumab and CCRT in the treatment of LACC can improve
SRT, stereotactic radiotherapy; ICIs, immune checkpoint inhibitors; CT, chemotherapy; OS, overall survival; iPFS, intracranial progression free survival; PFS, progression free the levels of T cell subsets, with tolerable acute toxic and good safety. CD4+T cell per-
survival; iORR, intracranial overall response rate; iDCR, intracranial disease control rate.
centage and tislelizumab use may be associated with CR. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
166s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2616 Poster Session 2617 Poster Session

PD-1 inhibitors combined with radiotherapy and GM-CSF, sequentially Phase 1 study of recombinant interleukin 15 in combination with nivolumab
followed by IL-2 regimen in advanced refractory solid tumors: A prospective, and ipilimumab in subjects with refractory cancers. First Author: Jibran Ahmed,
multicenter clinical trial. First Author: Pengfei Xing, Center for Cancer Diagnosis and National Cancer Institute, Bethesda, MD
Treatment, The Second Affiliated Hospital of Soochow University, Suzhou, China Background: Combining immune checkpoint inhibitors with cytokine therapies holds
Background: Low frequency of durable responses in patients treated with immune promise in cancer immunotherapy. Recombinant human interleukin-15 (rhIL-15) in-
checkpoint inhibitors demands for taking complementary strategies in order to boost creases circulating CD8+ T cells and natural killer (NK) cells. This Phase 1 study
immune responses against cancer. Our previous PRaG1.0 trial also demonstrated that evaluated the safety (NCI-CTCAE v5.0), tolerability, and preliminary efficacy (RECIST
PD-1 inhibitors in combination with radiotherapy and granulocyte macrophage-colony v1.1 and iRECIST) of rhIL-15 with nivolumab and ipilimumab. Safety data for rhIL-15/
stimulating factor (GM-CSF) could improve clinical response in patients with advanced ipilimumab and rhIL-15/nivolumab doublets were reported earlier (O’Sullivan et al., AACR
refractory solid tumors (ChiCTR1900026175). In an effort to further enhance efficacy, we 2019). Here, we present updated triplet dose-escalation results and correlative analyses.
conducted this PRaG2.0 trial ([Link]: NCT04892498) and optimized the Methods: This open-label, non-randomized Phase 1 trial employed a 3+3 dose-
PRaG1.0 regimen by incorporating interleukin-2 (IL-2). Methods: The PRaG 2.0 regimen escalation design to determine the MTD/RP2D of rhIL-15 SQ (administered on days
was administered to patients with advanced refractory solid tumors who lacked or were 1-8 and 22-29, cycles (C) 1-4 only) combined with fixed doses of nivolumab (240 mg IV
unable to tolerate standard-of-care treatments. A treatment cycle consisted of ra- on days 8, 22, 36) and ipilimumab (1 mg/kg IV on day 8) in 42-day cycles in patients with
diotherapy (5 or 8Gy32-3f) delivered for one metastatic lesion, PD-1 inhibitor dosing advanced, refractory cancers. Correlative analyses assessed effects on circulating T cell
within one week after completion of radiotherapy, GM-CSF 200mg subcutaneous (SC) subsets, PD-1/PD-L1 expression, and immune cell activation in tumor tissue (using
injection once daily for 7 days, and then sequentially followed by IL-2 2million IU SC once multiplex immunofluorescence, immunohistochemistry and flow cytometry).
daily for 7 days. PRaG 2.0 regimen was repeated every 21 days for at least 2 cycles until Results: Thirty-one patients (median age: 56 years, range: 24-81) were enrolled and
no appropriate lesions for irradiation or reached the tolerance dose of normal tissues. evaluable for safety and response. The most prevalent cancer types were sarcoma,
Patients who could not continue radiotherapy and had not yet developed progression pancreatic, and colorectal cancers (n = 5 each). The MTD/RP2D was established at 1 mg/
disease (PD) allowed PD-1 inhibitors to be continued as maintenance therapy until PD or kg/day SQ rhIL-15, with a manageable safety profile. Common treatment-related adverse
unacceptable toxicity but no more than one year. The endpoints were Progression-Free events (TRAE) included injection site reactions (74%), fever (65%), and chills (65%).
Survival (PFS), objective response rate (ORR) and overall survival (OS). Results: As of Grade 3/4 lymphopenia was seen in 13% of patients (4 of 31). Confirmed partial response
31st October 2024, 66 patients were enrolled in the study. The median Progression-Free (cPR) was measured in 1/31 patients (3%); the patient had cholangiocarcinoma and was
Survival (PFS) was 4.3 months, and the median overall survival (OS) was 10.3 months. treated at dose level 1 (DL1, 0.5 mg/kg/day rhIL-15) and the response lasted through
The objective response rate (ORR) was 22.7%, and the disease control rate (DCR) was cycle 16. Stable disease (SD) occurred in 17/31 patients (55%, median: 2 cycles, range:
56.1% according to RECIST version 1.1. Treatment-related adverse events (TRAE) 1-10), including durable SD (10 cycles at DL1) in salivary gland squamous cell carci-
experienced in 57 (86.4%) patients, with 6 patients (9.1%) experiencing Grade $ 3 noma. Nine patients had progressive disease as a best response. Four patients did not
TRAEs. We found that in the period prior to disease progression, the absolute count of have tumor measurements after C1 or did not complete C1, including 3 with clinical
Treg cells increased compared to baseline, while the percentage of CD8+PD-1+/CD8+ disease progression and one with grade 3 TRAE. NK cells and gd-T cells increased in
cells decreased compared to baseline. Conclusions: The PRaG 2.0 trial demonstrates blood, but increases did not correlate with tumor infiltration measured on C1D42. CD8+
that PD-1 inhibitors in combination with radiotherapy, GM-CSF, and IL-2 could be a T cells increased modestly in blood and tumor without correlation to either clinical
potential treatment regimen for patients with advanced refractory solid tumors. The benefit, increased PD1+CD3+ lymphocytes, or PD-L1+ tumor cells on C1D42.
decrease in the CD8+PD-1+/CD8+% ratio and the increase in the absolute count of Treg Conclusions: These data suggest the addition of rhIL-15 to the combination of ipili-
cells may suggest potential tumor progression in patients. Clinical trial information: mumab and nivolumab is safe, elicited a pharmacodynamic response from the immune
NCT04892498. Research Sponsor: None. system in blood but not tumor, and did not improve overall response rate. Clinical trial
information: NCT03388632. Research Sponsor: U.S. National Institutes of Health.

2618 Poster Session 2619 Poster Session

Effect of AdAPT-001 on checkpoint inhibitor resistance in solid tumors. First Efficacy and toxicity of nivolumab and ipilimumab in rare cancer brain
Author: Anthony Paul Conley, Department of Sarcoma Medical Oncology, The University metastases: A multi-center basket trial analysis (NCI/SWOG S1609). First
of Texas MD Anderson Cancer Center, Houston, TX Author: Manmeet Singh Ahluwalia, Miami Cancer Institute, Baptist Health South Florida,
Background: Checkpoint inhibitors (CIs) have revolutionized cancer treatment, but most Miami, FL
patients do not respond to them because of primary or secondary resistance. Introduction or Background: Outcomes of patients with brain metastases (BM) treated with single or dual checkpoint
reintroduction of CIs to resistant patients is relatively contraindicated because of the likelihood inhibitors have been previously evaluated, showing similar intra-cranial and extra-cranial response rates
of an unfavorable harm-benefit profile. An intensive search is on for therapies that sensitize and overall survival (OS). However, prior research has primarily focused on common tumor types such as
resistant tumors to CI therapy. AdAPT-001 is an oncolytic adenovirus armed with a trans- melanoma and lung cancer. We report outcomes in patients with BM from the largest basket trial for rare
forming growth factor beta (TGFb) trap that eliminates the immunosuppressive cytokine, cancers (N=684 evaluable patients) to evaluate efficacy and toxicity. Methods: Patients were treated
with nivolumab (NIVO, 240 mg Q2W) and ipilimumab (IPI, 1 mg/kg Q6W) in the federally funded SWOG
TGFb. Clinical data demonstrate that AdAPT-001 reverses the tumor immune evasion phe-
S1609 DART trial (NCT02834013), conducted across .1000 sites. The protocol and consent were
notype and improves the efficacy of immune checkpoint therapy, increasing response rate and reviewed and approved by SWOG, the NCI, the NCI central institutional review board, and institutional
progression-free survival (PFS) in CI refractory angiosarcoma. Methods: Patients with CI- review boards of participating sites. Efficacy and toxicity were assessed in patients with and without BM
refractory solid cancers of any type including sarcomas, melanoma, and breast cancer who at enrollment. Progression-free survival (PFS), and OS were estimated using Kaplan-Meier methodology.
received AdAPT-001 with a concurrent checkpoint inhibitor. Response was assessed every Tumor response was evaluated per RECIST v1.1, toxicities were assessed using CTCAE v5.0. Hazard ratios
8 weeks using RECIST 1.1. Treatment beyond progression was allowed for patients clinically (HR) with 95% confidence intervals (CI) and P-values were calculated to compare outcomes.
benefiting, and if progression was confirmed on the next assessment then the first assessment Results: Similar response rates were observed in patients without BM (11%, n=707) compared to 10% in
meeting PD criteria was considered the date of progression. PFS on AdAPT-001 + CI was those with BM at enrollment (n=20). PFS and OS were comparable between patients with and without BM
compared to duration of treatment on the most recent previous regimen including a CI before at enrollment (HR=1.29 [0.81-2.07], P=0.28; HR=1.36 [0.81-2.27], P=0.24, respectively). Grade $3 CNS
study enrollment. Results: Among all patients who were treated with a CI in a previous line of treatment-related toxicity occurred in 3% of patients without BM versus 5% in those with BM (P=0.43).
Similarly, Grade 5 treatment-related toxicity was observed in 2% of patients without BM compared to 5%
therapy (Prior CI) before being treated with AdAPT-001 + CI, the 6-month PFS rate was 17% (3/
in patients with BM (P=0.31). Among 18 patients with BM with progression, intra-cranial progression only
18) with Prior CI and 33% (6/18) with AdAPT-001 + CI, and the 12-month PFS rate was 0% (0/ was seen in 1 (5.5%) patient, while extra-cranial disease progression only in 12 (66.7%); 5 (27.8%) patients
18) with Prior CI and 17% (3/18) with AdAPT-001 + CI. Particular activity was seen in experienced concurrent intra- and extra-cranial disease progression. Conclusions: In this unique cohort
angiosarcoma, where all patients progressed within 3 months on Prior CI and 75% (3/4) had of patients with rare tumors and BM receiving dual checkpoint inhibitor therapy, similar response rates
PFS . 11 months with AdAPT-001 + CI. Treatment with AdAPT-001 was well tolerated and no and survival outcomes were observed in patients with or without BM at enrollment. No significant
new safety signals emerged. Conclusions: CI treatment is not an option for many patients differences in CNS or non-CNS toxicity were noted. Funding: NIH/NCI/NCTN grants U10CA180888,
because of primary or secondary resistance to them and the potential for harm without benefit. U10CA180819. Clinical trial information: NCT02834013. Research Sponsor: None.
The data from this P2 clinical trial strongly suggests that AdAPT-001 circumvents resistance Clinical outcomes and cox regression analysis.
to CIs in multiple tumor types, improving PFS when compared to the patient’s prior CI regimen. Outcome No BM (n=707), n (%) BM (n=20), n (%) P-value
In addition, AdAPT-001 may prevent the development of CI-induced autoimmune toxicities. P3
Best RECIST Response 0.76
clinical trials in sarcoma and hepatocellular carcinoma are planned. Clinical trial information: Confirmed CR/PR 81 (11.5) 2 (10)
NCT04673942. Research Sponsor: None. Unconfirmed CR/PR 22 (3.1) 0 (0)
Clinical benefit (SD >6 mo) 97 (13.7) 3 (15)
Duration of treatment on a previous CI regimen compared to subsequent PFS with AdAPT-001 + SD <6 mo or censored 123 (17.4) 1 (5)
CI. Progression/Failure 384 (54.3) 14 (70)
Subject ID Months on prior CI regimen PFS (months) with AdAPT-001 + CI PFS Change (months) PFS, HR (95% CI)
Univariate 1.22 (0.77-1.93) 0.39
1 3.0 11.4 +8.4 Multivariate 1.29 (0.81-2.07) 0.28
2 3.0 2.0 -1.0 OS, HR (95% CI)
3 1.9 13.0 +11.1 Univariate 1.23 (0.75-2.02) 0.41
4 2.7 12.0 +9.3 Multivariate 1.36 (0.81-2.27) 0.24
Median 2.8 11.7 +8.9

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 167s

2620 Poster Session 2621 Poster Session

Clinical factors and prognostic outcomes of hyperthyroidism induced by Quantum mechanics–based multi-tensor AI/ML discovery and validation of
immune checkpoint inhibitor therapy. First Author: Baqir Jafry, Charleston Area actionable and mechanistically interpretable whole-transcriptome predic-
Medical Center, Charleston, WV tors of survival in response to immunotherapy from real-world clinical trial
Background: Hyperthyroidism is a recognized but less frequent immune-related ad- data. First Author: Orly Alter, University of Utah and Prism AI Therapeutics, Inc., Salt
verse event (iRAE) associated with Immune Checkpoint Inhibitor (ICI) use. While its Lake City, UT
occurrence has been documented, gaps remain in understanding the underlying risk Background: Prediction in cancer remains limited, and 90% of drugs continue to fail
factors, and its impact on patient outcomes. This study seeks to provide clarity on these trials and post-market validation. The entire multi-ome affects the disease. Previously,
aspects and guide improved management strategies. Methods: Data were obtained we developed quantum mechanics-based multi-tensor AI/ML to overcome the limita-
from the TriNetX research network for patients with cancers where ICI is used. Inclusion tions of typical AI/ML, e.g., neural networks and deep learning, in small-cohort, noisy,
criteria encompassed patients aged 18 years or older treated with ICIs (e.g., pem- high-dimensional, multi-omic clinical data [doi: 10.1073/pnas.0530258100, 10.1145/
brolizumab, nivolumab, atezolizumab, cemiplimab) between January 1, 2013, and De- 3624062.3624078]. We have demonstrated the algorithms in the discovery and vali-
cember 31, 2024. Patients with a history of thyroid disorders or Levothyroxine use were dation of whole-genome and -chromosome predictors of survival and response to
excluded. Competing risk analyses evaluated the likelihood of hyperthyroidism versus treatment in, e.g., brain, lung, ovarian, and uterine cancers [doi: 10.1063/1.5142559,
death up to 12 months after ICI use and the likelihood of beta-blocker usage after 10.1200/JCO.2024.42.16_suppl.10043]. Methods: Here, we use the algorithms to
diagnosis. Cox proportional hazards modeling was used to identify significant covariates discover two whole-transcriptome predictors of OS in response to atezolizumab PD-L1
associated with hyperthyroidism, including age, sex, cancer type, comorbidities, and ICI inhibitor immunotherapy in a 348-patient, multi-center, single-arm, bladder cancer
type. Backwards batchwise elimination was employed to retain only significant vari- clinical trial, and validate the predictors in the 401-patient bladder cancer cohort in the
ables. Results: Among 39,749 patients receiving ICIs, 2.3% developed hyperthyroidism Cancer Genome Atlas (TCGA). Results: The algorithms discovered the two predictors in
within 12 months after treatment. In patients with hyperthyroidism, 31.3% initiated the open-source, pre-atezolizumab, locally advanced or metastatic disease profiles of
metoprolol, 9.5% initiated propranolol, and 6.0% initiated atenolol for symptom man- the 348 patients alone. By incorporating the patient labels, both predictors were found to
agement. The mortality rate in the group was 32.1% based on the cumulative incidence. outperform the best indicator of response to the treatment to date, i.e., the tumor
An increased risk of hyperthyroidism was associated with endometrial cancer (HR: 1.44; mutation burden (TMB): The Cox proportional hazards model ratios, i.e., the corre-
95CI: 1.23-1.69; p , 0.0001), non-Hodgkin lymphoma (HR: 1.42; 95CI: 1.09-1.84; p = sponding relative risks, of 2.7 and 1.7, and concordance indices, i.e., accuracies, of 0.70
0.010), and kidney cancer (HR: 1.34; 95CI: 1.25-1.63; p = 0.001). Conversely, patients and 0.63, of each predictor, are greater than the ratio, of 1.3, and index, of 0.61, of TMB
with colon cancer (HR: 0.74; 95CI: 0.55-0.98; p = 0.038) had a lower chance of de- (Wald P-values=2.0310-7 and 8.2310-3 vs. 2.1310-5). The maximum Kaplan-Meier
veloping hyperthyroidism. Among all ICIs, atezolizumab (OR: 1.32; 95% CI: 0.82–2.13; p median OS difference of the two predictors together, of 22 months, is greater than that of
= 0.25) showed the strongest trend toward hyperthyroidism, followed by durvalumab TMB, of 16 months (log-rank P-values=3.6310-11 vs. 1.7310-4). One predictor is ad-
(OR: 1.24; 95% CI: 0.75–2.05; p = 0.39), pembrolizumab (OR: 1.23; 95% CI: 0.81–1.87; p = ditionally correlated with the objective response rate (ORR), and the other – with the
0.34), and nivolumab (OR: 1.22; 95% CI: 0.80–1.86; p = 0.37). In contrast, cemiplimab tissue of advanced or metastatic disease (Kruskal-Wallis P-values=9.8310-4 and
(OR: 0.61; 95% CI: 0.17–2.16; p = 0.44) showed a lower likelihood of causing hyper- 2.2310-3). Both predictors are similarly correlated with the OS of the 401 TCGA bladder
thyroidism. Conclusions: ICI-induced hyperthyroidism, while less common, has a cancer patients. Both are statistically independent of the imbalanced variations in the
significant impact on patient outcomes, including mortality. Identifying high-risk cancer patient demographics, e.g., race, or the tissue batches, e.g., pre- vs. post platinum-based
subtypes and implementing proactive management, including beta-blocker therapy for chemotherapy. By using the transcript labels, the predictors were interpreted in terms of
symptom control, are critical steps to mitigate adverse effects and improve patient care. known and new disease mechanisms and drug targets to sensitize the tumors to the
Personalized management and early intervention, particularly for high-risk groups, are treatment. Conclusions: Our multi-tensor AI/ML discovered and validated two whole-
essential to improving patient outcomes. Research Sponsor: None. transcriptome predictors of OS in response to atezolizumab that outperform TMB. This
further suggests that quantum mechanics-based algorithms can be used to derive
predictors that are consistent across studies and over time. Research Sponsor: None.

2622 Poster Session 2623 Poster Session

Phase II trial of neoadjuvant nivolumab and SOX in resectable gastric/ Delineation of immunotherapeutic predictive versus prognostic transcrip-
gastroesophageal junction cancer: Therapeutic response and biomarker tional programs to identify SLC22A5-centric carnitine metabolism-driven
correlations. First Author: Xiangdong Cheng, Zhejiang Cancer Hospital, Hangzhou, resistance to anti-PD-L1 treatment in advanced non–small-cell lung cancer.
Zhejiang Province, China First Author: Yuze Wang, The First Affiliated Hospital, Sun Yat-sen University,
Background: Neoadjuvant immunotherapy produces a major pathologic response (MPR) Guangzhou, China
in 40% of patients with locally advanced gastric cancer (LAGC). Herein, we conducted a Background: Prognostic factors indicate the natural course of a disease regardless of
phase prospective clinical study to investigate markers related to the response to treatment, whereas predictive factors determine the likelihood of response to specific
neoadjuvant immunotherapy. Methods: Patients with locally advanced gastric or gastro- therapies. Distinguishing between predictive and prognostic factors is essential for sepa-
esophageal junction cancer (cT3-4N+M0,CY0,P0) were enrolled and received either 3 rating treatment-specific outcomes from the inherent progression of cancer, thereby guiding
preoperative and 3 postoperative cycles of nivolumab (360 mg, IV, d1, Q21d) plus SOX clinical decision-making. We aim to dissect the predictive and prognostic transcriptional
regimen (oxaliplatin 130 mg/m2, IV, d1 with oral S-1 40-60mg, bid, d1-d14, Q21d ) therapy, programs underlying the efficacy of anti-PD-L1 versus chemotherapy in advanced non-small
followed by 11 cycles of nivolumab monotherapy. The primary endpoint was the pCR rate, cell lung cancer (NSCLC) to uncover mechanisms specific to immunotherapy resistance.
while the mPR, 3-year-DFS and 3-year-OS as the second endpoint. This clinical trial was Methods: Clinical and baseline tumor transcriptomic data were collected from two ran-
registered at [Link] (NCT05739045). In addition, tissue samples from patients domized controlled trials comparing atezolizumab with docetaxel: OAK (n=697, discovery
before and after preoperative therapy were performed scRNA-seq to explore the changes cohort) and POPLAR (n=192, validation cohort). Transcriptional program scores for each
of tumor microenvironment during the therapy and biomarkers associated with therapy biological process and metabolic pathway from the Reactome database were calculated
response. Results: Forty-six patients were enrolled from November 2022 to March 2023, using gene set variation analysis for each patient. Cox regression and P-value for interaction
tests were conducted to differentiate predictive versus prognostic effects of transcriptional
with a median age of 66 years (range, 34-74), and 38 (82.60%) were male. There were 28
programs. Tumor microenvironment and cell-cell communication underlying immunotherapy
(60.87%) patients with PD-L1 CPS $5. The study achieved its primary endpoints, with a
resistance were explored using bulk and single-cell transcriptomic data.
pCR rate of 21.74% and an MPR rate of 41.30%. Among the 46 patients who underwent D2 Results: Transcriptional programs in the OAK discovery cohort were divided into four
gastrectomy, the 1-year OS rate was 97.83% and 1-year DFS rate was 95.65%. Single-cell categories associated with different predictive effects specific to atezolizumab or docetaxel.
RNA sequencing of 131 tumor samples obtained from 46 patients with LAGC at multiple Carnitine metabolism was the most prominent process contributing to atezolizumab-specific
time points during neoadjuvant therapy showing that the expression of MHC-II was resistance, while porphyrin metabolism drove docetaxel-specific resistance. SLC22A5, the
upregulated in malignant cells in the pre-sensitive group. In a retrospective cohort of 226 only high-affinity carnitine transporter, was upregulated in atezolizumab-resistant patients.
patients treated with neoadjuvant immunotherapy, MHC-II-positive patients exhibited The predictive effect of SLC22A5-centric carnitine metabolism for resistance to atezolizumab
significantly higher rates of pCR and mPR compared to MHC-II-negative patients. rather than docetaxel was confirmed in the POPLAR validation cohort. Integrative analyses of
Furthermore, 30 MHC-II-positive GC patients were prospectively enrolled to receive bulk and single-cell transcriptomes revealed that cancer cell-specific SLC22A5 expression
neoadjuvant immunotherapy, showing pCR and mPR rates of 36.67% and 66.67%, re- induced M2 macrophage polarization and decreased CD8+ T cell infiltration via carnitine
spectively. Mechanistically, we observed that T cell-induced IFN-g signaling predomi- uptake, thus forming an immunosuppressive microenvironment. Conclusions: Our study
nated in the tumor microenvironment of the sensitive group before treatment. This elucidates the distinction between predictive and prognostic factors in advanced NSCLC
signaling pathway induces MHC-II expression in tumor cells, thereby enhancing the T cell- from a metabolic perspective. Cancer cells uptake of carnitine via SLC22A5 mediates re-
mediated antitumor immune response during neoadjuvant immunotherapy. In addition, sistance to anti-PD-L1 treatment. Combining inhibition of SLC22A5-centric carnitine
MHC-II expression in tumor cells can be detected via IHC and commercially available metabolism with anti-PD-L1 agents might be a promising strategy to reverse immune escape
antibodies in standard pathology laboratories, making it a potential biomarker to guide the in advanced NSCLC. Keywords: Predictive, Prognostic, Non-small cell lung cancer, Carnitine
selection of appropriate GC patients for neoadjuvant immunotherapy. [Link] metabolism, Resistance. Research Sponsor: National Natural Science Foundation of China;
registration: NCT05739045. Conclusions: Our study illuminates the role of MHC-II ex- 82373307; Natural Science Foundation of Guangdong Province; 2024A1515013214; the
pression in tumor cells in modulating the response to immunotherapy in gastric cancer. China Postdoctoral Science Foundation; 2024M753780; the institutional funding of The First
Clinical trial information: NCT05739045. Research Sponsor: None. Affiliated Hospital of Sun Yat-sen University.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
168s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2624 Poster Session 2625 Poster Session

Biomarkers associated with outcomes from OPTIMIZE-1: CD40 agonist Tumor mutational burden, PD-1, negative Wnt/b-catenin regulators, and
mitazalimab with mFOLFIRINOX in patients with untreated metastatic pan- positive MHC class II antigen presentation regulators as predictors of longer
creatic cancer. First Author: Philippe Alexandre Cassier, Centre Léon Bérard, Lyon, survival after immune checkpoint inhibitors across cancers: A comprehen-
France sive analysis of 400 immunity biomarkers. First Author: Yu Fujiwara, Department
Background: Metastatic pancreatic ductal adenocarcinoma (mPDAC) has a 5-year overall of Medicine, Roswell Park Comprehensive Cancer Center, Buffalo, NY
survival (OS) rate of less than 5% and remains a leading cause of cancer related mortality. Background: Immune checkpoint inhibitors (ICIs) have become a standard treatment, yet
Mitazalimab, a human CD40 agonistic IgG1 antibody, activates CD40 signaling in myeloid no universal biomarker consistently predicts prolonged survival across cancer types. While
cells, enhancing tumor sensitivity to chemotherapy and licensing dendritic cells to prime multiple biomarkers have been proposed, they have not been systematically compared or
and activate tumor-specific T cells. The OPTIMIZE-1 Phase 1b/2 trial evaluated the safety evaluated in a tumor-agnostic manner. This study comprehensively investigates the impact
and efficacy of mitazalimab combined with mFOLFIRINOX (mFFX) in treatment naı̈ve of 397 immunoregulatory transcripts and other factors on survival in patients with cancer
mPDAC patients (pts) (Van Laethem 2024). This combination therapy has shown promising treated with ICIs. Methods: The Profile-Related Evidence Determining Individualized Cancer
clinical efficacy compared to historical controls with a median OS of 14.9 months, median Therapy (PREDICT, NCT02478931) study enrolled 514 patients, including 217 treated with
duration of response of 12.6 months, median progression free survival of 7.7 months, and ICIs. The study analyzed the effect of bulk tumor transcriptomic expression of 397 im-
an overall response rate of 54.4 % (42.1% confirmed) (Geboes, 2024). Methods: Patients munoregulatory factors plus microsatellite instability [MSI], tumor mutational burden [TMB],
received mitazalimab on day 1 (priming dose), followed by a 2-week regimen starting with and PD-L1 immunohistochemistry (total = 400 biomarkers) along with cancer type on overall
mFFX on day 8 and mitazalimab on day 10. Associations between survival benefits and survival (OS) and progression-free survival (PFS) following ICI therapy. Transcriptome
both baseline and on treatment biomarkers (RNAseq data from tumor biopsies and expression was categorized into three groups based on percentile ranks compared to 735
longitudinal circulating tumor KRAS (ctKRAS) data) were assessed in patients from the full controls spanning 35 histologies: “High” (75–100th percentile), “Intermediate” (25–74th
analysis set (n = 57) treated with 900 mg/kg mitazalimab. Results: Differential gene percentile), and “Low” (0–24th percentile). Hazard ratios (HRs) for OS and PFS were es-
expression analysis (DGEA) identified a fibrosis-related gene signature (including genes timated using a Cox regression model. Storey’s q-value correction accounted for multiple
involved in extracellular matrix (ECM) remodeling) that correlated with improved OS (p = testing, and a multivariable analysis was performed to explore novel biomarkers and
0.002). Conversely, a distinct gene signature linked to chemoresistance mechanisms different TMB cutoffs (10 [mutations/mb], 16, 20, and continuous). Results: In the 217 ICI-
involved in the inactivation and secretion of mFFX components was associated with treated patients (median age: 61.2 years; women: 56.2%), the most common ICI was anti-
shorter OS. Comparing DGEA of three on-treatment biopsies from patients with partial PD-1 therapy (83.4%, N = 181), and the median TMB was 5.0 mut/mb. MSI was detected in 9
response to baseline samples revealed treatment-induced tumor changes. These analyses patients (4.1%). In multivariable analysis adjusting for age, sex, significant markers (q ,
identified upregulation of genes involved in myeloid cell biology and regulation of T cell 0.05) and co-inhibitory checkpoints (p , 0.1 in univariate analysis), and TMB (with different
responses, along with downregulation of immunosuppressive genes. Longitudinal data cutoffs and as a continuous variable), high expression of CIITA, KREMEN1, and PD-1 (but not
analyses revealed that ctKRAS clearance was reached by 72% of pts and was significantly PD-L1), as well as higher TMB ($10 mut/mb, $16 or $20 or continuous), were inde-
associated with longer OS. Molecular response was also associated with longer OS and pendently associated with longer OS (p#0.05). In the PFS analysis, high KREMEN1 ex-
predicted radiological response with 76.7% accuracy (sensitivity 72.7%, specificity 81.0%). pression and higher TMB ($16 or $20 or continuous but not $10 mut/mb) also
Further, molecular progression was significantly associated with OS and predicted ra- independently correlated with longer PFS. Cancer type was not independently correlated
diological response with 62.8% accuracy (sensitivity 71.4%, specificity 58.6%). with outcome. Conclusions: Our comprehensive biomarker analysis identified novel factors
Conclusions: A potentially predictive fibrosis-related gene signature, directly linked to associated with ICI efficacy. In addition to confirming the predictive role of TMB, the study
mitazalimab’s mode of action, was associated with improved OS. Biomarker correlations highlights CIITA, a regulator of MHC class II expression that enhances tumor antigen
further suggest a mitazalimab-driven contribution to clinical benefits in the OPTIMIZE-1 presentation, and KREMEN1, a suppressor of Wnt/b-catenin signaling that preserves an-
trial. These encouraging results will inform the planned randomized confirmatory trial of titumor immunity, as well as the PD-1 checkpoint, as predictive biomarkers for ICI therapy
mitazalimab in combination with mFFX in mPDAC. Clinical trial information: across cancers. Clinical trial information: NCT02478931. Research Sponsor: U.S. National
NCT04888312. Research Sponsor: Alligator Bioscience AB. Institutes of Health; P30 CA023100; U.S. National Institutes of Health; 5U01CA180888-08;
U.S. National Institutes of Health; 5UG1CA233198-05.

2626 Poster Session 2627 Poster Session

Impact of Helicobacter pylori infection on molecular alterations and immune Levels of immune responses in tertiary lymphoid structures of non-small
dynamics in gastric cancer. First Author: Daisuke Takayanagi, Showa University, cell lung cancer (NSCLC) and association with survival. First Author: Jiangping
Tokyo, Japan Li, Division of Thoracic Tumor Multimodality Treatment, Cancer Center, West China
Background: Helicobacter pylori (HP) infection, a major risk factor for gastric cancer Hospital, Sichuan University, Chengdu, Sichuan, China
(GC), modulates tumor immunity. Evidence indicates that H. pylori infection status Background: Tumor-infiltrating tertiary lymphoid structures (TLSs) are thought to have
correlates with the efficacy of immune checkpoint inhibitors (ICI), with varying outcomes anti-tumor activity and are believed to indicate a favorable prognosis in cancer patients.
in H. pylori-positive (HPP) and -negative (HPN) patients. This study investigated the However, the prognostic value of TLSs in non-small cell lung cancer (NSCLC) is un-
impact of HP infection on splicing alterations to elucidate its role in shaping immune known. Methods: RT-qPCR analysis of the reactive Th-cell subsets and fluorescence
responses and the tumor immune microenvironment (TIME), immune checkpoint mol- activated cell sorting (FACS) analysis of the cell composition in tumor and paired
ecule expression, transcriptional profiles in GC. Methods: Tumors and adjacent normal controls. Histological evaluation of the co-localization of tumor-associated CD20+
tissues were collected from 24 patients with GC, comprising 39 tumor samples from HPP B cells, CD4+ T cells and DC-LAMP+ mature dendritic cells (DCs) within TLSs, and
patients, 27 samples from HPN patients, and 13 and 10 normal samples of each, re- statistically analysis of the relationship between TLSs and overall survival. Results: The
spectively. RNA sequencing was performed on these tissues and whole blood RNA from results indicated high levels of immune responses in tumour microenvironment
six patients (four HPP and two HPN) to analyze the alternative splicing (AS) events, the ofNSCLC, which that high levels of Th-CXCL13, Th1, Tfh and Treg cell immune responses
transcriptional profiles, and immune-related gene expression. Differential gene expres- were the main reactions in tumor tissue of NSCLC, followed by weak Th2 and Th17,
sion (DGE) and enrichment analyses were conducted, and immune cell fractions were suggesting high levels of immune responses in tumor microenvironment of NSCLC.
evaluated using CIBERSORTx. Results: DGE analysis revealed that HPP tumors were Tumor-associated TLS is a complete and mature lymphoid follicle-like structure con-
enriched in genes related to cell cycle regulation, whereas HPN tumors were enriched in taining T cells, B cells and APCs in tumor of NSCLC. Six-color MIHC staining indicated
immune response pathways, including those involved in leukocyte activation, chemokine tumor-associated TLSs were complete and mature lymphoid follicle-like structures
signaling, and immune effector processes. Additionally, HPN tumors showed higher containing T cells, B cells and antigen presenting cells (APCs), and tumor-associated
expression of immune checkpoint molecules, such as CD160 (p = 0.016), PDCD1LG2 (p = TLSs. Architectural analysis showed that CD20+ B cell clusters were localized in the local
0.0082), and BTLA (p = 0.025). Immune cell profiling demonstrated increased proportions tumour microenvironment, and were mainly colocalized with CD4+ T cells, CD68+
of gamma-delta T cells (p = 0.0077), resting dendritic cells (p = 0.0002), and neutrophils (p macrophages, CD11c+ DCs and DC-LAMP+ mature DCs, which indicated the formation of
= 0.016), reflecting enhanced immune activation and a favorable ICI response. In contrast, mature TLSs. One of the primary functions of TLS is to support the survival of
HPP tumors were enriched in cell cycle-related pathways, suggesting a proliferative incoming lymphocytes. The vast majority of the evaluated tumour-associated TLSs in
phenotype. HPP tumors also exhibited higher levels of M0 macrophages (p = 0.0039) and NPC represented mature secondary-follicle-like TLSs, as indicated by the presence of
CD276 expression (p = 0.0082), indicative of an immunosuppressive TIME. AS analysis both CD21+ FDC and CD23+ GC-B cells. Understanding and analyzing the phenotypes of
identified increased intron retention (IR) events in HPP tumors, particularly in genes these TABs was important for interpreting the local immune response. We used six-color
associated with RNA processing and extracellular matrix remodeling. These alterations MIHC staining (CD10, CD20, CD38, CD138, CD27 and DAPI) to approximate six different
may contribute to immune evasion and tumor progression. In the peripheral blood, HPP TABs in whole tissue sections of NPC. These molecules were expressed to varying
samples exhibited upregulation of tripartite motif family genes, which are implicated in degrees in all paraffin sections, and MIHC showing that CD38+ plasmablast cells and
immune modulation. Conclusions: This study demonstrated that HP infection signifi- CD138+ plasma cells were primarily located at the tumor interstices or margins, which
cantly affects the TME and gene expression profiles of GC. HPP tumors are characterized indicated improved survival outcome. In conclusion, this is the first research of applying
by increased M0 macrophage populations, CD276 expression, and IR events that con- multiple complementary strategies to map the biological function and clinical relevance
tribute to immunosuppression and tumor progression. In contrast, HPN tumors exhibit of those cells in the TLSs of NSCLC. Conclusions: Our findings provide insights into the
greater immune activation and checkpoint molecule diversity. These findings highlight potential role of TLSs in the adaptive anti-tumor immune response, with implications for
the potential role of HP status in shaping the immune landscape of GC and influencing the development of biomarkers and therapeutic targets. Research Sponsor: None.
responsiveness to ICI. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 169s

2629 Poster Session 2630 Poster Session

Clinical significance of the CGRP pathway gene expression in advanced Predication of clinical outcomes of advanced cutaneous squamous cell
solid tumors: A sub-analysis of MONSTAR-SCREEN-2. First Author: Takao carcinoma to PD1 inhibition directly from histopathology slides using
Fujisawa, Department of the Promotion of Drug and Diagnostic Development, National inferred transcriptomics. First Author: Johnathan Arnon, Sharett Institute of On-
Cancer Center East, Kashiwa, Japan cology, Hadassah Hebrew Universty Medical Center, Jerusalem, Israel
Background: Calcitonin gene-related peptide (CGRP), a neuropeptide associated with Background: Metastatic or locally advanced cutaneous squamous cell carcinoma
pain perception, has emerged as a therapeutic target for migraine. Recent studies have (CSCC) not amenable to local therapy is treated with programmed death (PD)-1 in-
reported that the CGRP pathway is associated with suppression of anti-tumor immunity hibitors, namely Cemiplimab . While response rates are relatively high at approximately
and poor prognosis in patients (pts) with solid tumors via induction of CD8+ T cell 45%, no predictive biomarkers for PD-1 inhibition have been validated in advanced CSCC
exhaustion by sensory nerves. However, clinical significance of the CGRP pathway genes subjecting some patients, especially older, to unnecessary immune-related adverse
in oncology including the impact for the efficacy of immune checkpoint inhibitors (ICIs) events (irAE). We present a retrospective analysis of ENLIGHT-DP, a novel biomarker for
remains elusive. Herein, we evaluated the landscape of the CGRP pathway gene ex- response to PD-1 inhibition in advanced CSCC, calculated directly from histopathological
pression and their association with efficacy of ICIs by mRNA expression level of the slides. Methods: We retrospectively examined high resolution hematoxylin and eosin
CGRP pathway genes in advanced solid tumors from the SCRUM-Japan MONSTAR- (H&E) slide scans from archived tumor-tissue samples of advanced CSCC patients
SCREEN-2, a nationwide molecular profiling project. Methods: Pts with advanced solid treated with Cemiplimab to generate an individual prediction score to PD-1 inhibitors
tumors were enrolled; tumor tissues were profiled using whole exome/transcriptome using ENLIGHT-DP. This is composed of two main steps: (I) prediction of individual
sequencing (MI Profile, Caris Life Sciences, Phoenix, AZ, USA). The association between mRNA expression directly from H&E slides using the digital pathology-based DeepPT
the expression profiles of the CGRP pathway genes (CALCA encoding CGRP, CALCRL algorithm (II) use these values as input to ENLIGHT, a transcriptomics-based precision
and RAMP1 encoding CGRP receptors) and the efficacy of ICI monotherapy was an- oncology platform for prediction of response to cancer therapies. We then unblinded
alyzed. Results: Among 2,768 pts enrolled as of March 2024, mRNA expression data of clinical outcomes and assessed the predictive values of ENLIGHT-DP. Results: We
baseline tissue samples were available in 1,475 pts across 36 cancer subtypes: most evaluated 39 cases of advanced CSCC (tumors from various origins) at a median age of
common subtypes were colorectal adenocarcinoma (n = 352) and esophagogastric 81 years old (range 57-100). Of them, 32 cases (82%) were initially treated with surgery
adenocarcinoma (EGAC, n = 192). mRNA expression of the CGRP pathway genes were or radiotherapy with curative intent, but ultimately suffered disease progression. The
observed across diverse cancer types. One hundred and fifty-eight pts were treated with objective response rate (ORR) was 69%, median progression free survival (PFS) was
ICI monotherapies: most common subtypes were urothelial carcinoma (UC, n = 41), 11 months (CI 95% 10.4-17.6) and 6 patients (15%) suffered from severe irAE ne-
EGAC (n = 30) and head and neck squamous cell carcinoma (HNSCC, n = 30). The low cessitating treatment cessation. ENLIGHT-DP was predictive of response with ROC AUC
RAMP1 mRNA expression group tended to have a better objective response rate (ORR) = 0.67. Using a binary threshold for classification, calibrated on previous lung and head
and significantly better progression-free survival (PFS) than the high mRNA expression and neck cohorts, ENLIGHT-DP displays promising biomarker characteristics: 81.8%
group (ORR: 29.3% vs 15.2%, P = 0.056, PFS: 5.1 vs 2.7 months, hazard ratio [HR]: 0.68, PPV, 66.6% sensitivity and 4.0 OR (p = 0.04) for matched vs. unmatched patients.
95% CI: 0.47-0.98, P = 0.04). Among the patients with UC and HNSCC, the low RAMP1 ENLIGHT-DP successfully stratified PFS with HR 0.02 (CI 95% 0.0005-0.96, p = 0.05).
mRNA expression group tended to have a better PFS than the high mRNA expression Importantly, omitting cases in which ENLIGHT-DP was calculated on samples which
group (UC: 3.5 vs 6.0 months, HR: 0.69, 95% CI: 0.34-1.39, P = 0.3, HNSCC: 1.5 vs were taken more than 6 months prior to Cemiplimab therapy (i.e., during diagnostic
6.0 months, HR: 0.67, 95% CI: 0.30-1.51, P = 0.3). Conclusions: High RAMP1 mRNA excisions) did not influence the results. No other patient characteristics (e.g., age, stage,
levels were associated with worse therapeutic efficacy of ICI, highlighting the potential co-morbidities, previous treatments) were associated with outcomes.
of CGRP pathway as a resistance mechanism and a treatment target. Clinical trial Conclusions: ENLIGHT-DP demonstrates high predictive values for clinical outcomes of
information: UMIN000043899. Research Sponsor: SCRUM-Japan Funds. PD-1 inhibition in advanced CSCC, relying solely on easily accessible archived H&E
slides. Research Sponsor: Pangea Biomed.

2631 Poster Session 2632 Poster Session

Evaluation of AI-assisted PD-L1 CPS scoring in immunostained pan-organ Validation of ENLIGHT, an AI predictor of immune checkpoint blockade (ICB)
tumor whole-slide images. First Author: Céline Bossard, Pathology Department, IHP response and resistance, across the treatment span. First Author: Scott Strum,
Group, Nantes, France Princess Margaret Cancer Centre – University Health Network, University of Toronto,
Background: PD-L1 inhibitors have shown remarkable results in oncology, yet many Toronto, ON, Canada
patients fail to respond, underscoring the importance of reliable assessment of PD-L1 Background: Advanced computational AI algorithms, such as ENLIGHT and DeepPT (Med 2023, Nature
expression for patient selection. PD-L1 scoring, especially the Combined Positive Score Cancer 2024), represent a promising approach to identify predictive biomarkers for cancer therapeutics.
Evaluation of ICB response prediction via these algorithms through the full span of pre-treatment, on-
(CPS), is hindered by inter-observer variability, complex staining patterns, and technical treatment, and at progression time points provides a dynamic perspective of response prediction abilities.
discrepancies across platforms and antibody clones. These challenges may impact Methods: A post-hoc analysis of two pan-cancer clinical trials was performed: i) BIO2 is a biobanking
therapeutic decisions. Artificial intelligence (AI) offers a solution by standardizing PD-L1 protocol of ICB-naı̈ve patients (pts) treated with pembrolizumab (NCT02644369); and ii) The IRIS study
evaluation. This study evaluates Diadeep PD-L1 CPS AI solution designed to provide (NCT04243720) which enrolled pts who have progressed immediately post ICB. In BIO2, complete, partial
reproducible and robust PD-L1 scoring across diverse tumors and conditions. response or stable disease for .6 months was classified as responders (R), the rest as non-responders
Methods: AI performance was validated on 142 formalin-fixed, paraffin-embedded (NR). In IRIS, acquired and primary resistance were defined according to trial protocol. ENLIGHT matching
scores were calculated using either transcriptomics from NGS (EMS-NGS), or transcriptomics imputed
samples spanning multiple tumor types (GI, head and neck, breast, uterine cervix) and
directly from H&E slides using DeepPT (EMS-DP). The predictive value of EMS was compared to PD-L1
sourced from four centers, reflecting diverse staining protocols (22C3 and QR001 clones; IHC, tumor mutational burden (TMB) and tumor infiltrating lymphocytes (TILs) abundance by IHC, and its
BenchMark ULTRA and Omnis/Dako platforms). The routine scores were available for trajectory across timepoints was studied. Results: 76 pts from BIO2 (23:53, R:NR), and 37 pts from IRIS
these cases. A Gold Standard was established through independent retrospective scoring (18:19, AR:PR), comprising of 14 tumor types, were analyzed. We first established the value of ENLIGHT
by three blinded senior pathologists, which allowed to compute the intraclass correlation as a predictive biomarker using the BIO2 pre-treatment samples. EMS-NGS was a superior predictive
coefficient (ICC). The scoring was followed by collegial discussions to resolve discordant biomarker compared with PD-L1 IHC, TMB and TIL abundance, while EMS-DP was comparable (Table).
The EMS-NGS scores of responders were significantly higher than non-responders pre-treatment
cases and ensure medical consensus. After a washout period, pathologists re-evaluated
(medians: 0.92 vs. 0.62, p = 1.4e-4). Analyzing the trajectory of the EMS-NGS scores across two ad-
the cases with the AI assistance. AI-computed scores and routine manual scores were ditional timepoints reveals that while the scores of non-responding patients remained low (median: 0.62,
evaluated and compared by using the Gold Standard as a reference and the organ-specific 0.69, 0.67 for pre-, on–treatment and post-progression, respectively), it is higher among responders
recommended cut-offs. Results: The AI assistance improved interobserver agreement (median: 0.92, 0.78 for pre- and on–treatment, respectively). Finally, EMS-NGS was higher among pts with
among pathologists, with the ICC increasing from 0.62 to 0.74. This effect was particularly acquired vs primary resistance in IRIS (medians: 0.75 vs 0.59, p = 0.17). Conclusions: In two pan-cancer
pronounced for challenging cases with CPS , 20 (n = 91), where ICC improved from 0.19 cohorts, EMS-NGS outperformed conventional biomarkers in predicting ICB response. EMS-DP was
comparable to conventional biomarkers and could be calculated directly from H&E slides in a fast, low-
to 0.62, underscoring the AI’s value in reducing variability near clinical decision thresholds.
cost manner. EMS-NGS values were concordant with response or resistance throughout the ICB
Moreover, the AI-based scoring tool demonstrated superior accuracy (88%) compared to treatment course, reflecting the level of the tumor’s vulnerability to ICB inhibition. Further validation of
routine manual scoring (75%) in classifying PD-L1 expression based on clinical cutoffs. ENLIGHT in larger ICB-treated pts is warranted given these promising results. Clinical trial information:
Sensitivity was significantly higher with AI (96% vs. 78%, p , 0.001), while the positive NCT02644369, NCT04243720. Research Sponsor: BMO Chair in Precision Genomics, Dr. Lillian Siu.
predictive value was comparable (88% vs. 87%), indicating an improved ability to detect
ROC AUC (p) Sensitivity PPV (cf 30% baseline response rate) F1 Score
true positive cases. Conclusions: This study highlights the potential of an AI-driven tool to
EMS-NGS 0.74 (0.0003) 61 48 54
enhance PD-L1 scoring by significantly improving accuracy and reducing inter-observer EMS-DP 0.64 (0.02) 57 45 50
variability, particularly in cases near clinical decision thresholds where consistency is PD-L1 IHC 0.7 (0.003) 70 40 51
critical. By delivering reliable and reproducible results, the AI algorithm addresses key TMB 0.64 (0.03) 39 69 50
TILs 0.6 (0.065) 39 52 44
challenges in PD-L1 evaluation, ensuring more precise patient stratification for immu-
notherapy. Beyond accuracy, the integration of such tools into clinical workflows could
optimize patient selection and improve therapeutic outcomes, offering oncologists greater
confidence in treatment decisions. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
170s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2633 Poster Session 2634 Poster Session

Effect of elevated expression of LILRB4 and TSC22D3 on survival in lung Tertiary lymphoid structures and their association with immune checkpoint
cancer. First Author: Borys Hrinczenko, Michigan State University, East Lansing, MI inhibitor response and survival outcomes in patients with non-small cell
Background: HER2/neu mutations or amplifications, present in up to 23% of non-small lung cancer. First Author: Dmitrii Grachev, BostonGene, Corp., Waltham, MA
cell lung cancers (NSCLCs), activate STAT3, promoting inflammation and suppressing Background: Immune checkpoint inhibitor (ICI)-based therapy is currently the first-line
adaptive immune responses. Elevated systemic inflammation-immune index (SII) in small treatment for patients with lung adenocarcinoma (LUAD) and lung squamous cell
cell lung cancer (SCLC) correlates with reduced response to PD-1/L1 immune checkpoint carcinoma (LUSC) without actionable mutations. However, the commonly utilized
inhibitors (ICIs), leading to worse overall survival (OS) and progression-free survival (PFS). biomarkers, including PD-L1 protein expression and tumor mutation burden, are not
These findings highlight the need to mitigate inflammation in lung cancers to enhance ICI sufficiently accurate to predict the treatment response from ICI in this patient pop-
response and improve survival. Biomarker studies can identify genes that differentiate ulation. As tumor microenvironment (TME) and tertiary lymphoid structure (TLS) play a
inflammatory and immune pathways, offering therapeutic insights. Methods: To identify significant role in antitumor immunity, we explore these immunophenotypic factors to
genes regulating the HER2/neu oncogene, HER2/neu-overexpressing mice were crossed determine the potential biomarkers in patients with LUAD or LUSC. Methods: We
with genetically Diverse Outbred (DO) mice. The resulting DO F1 offspring were monitored evaluated all patients with LUAD or LUSC from three publicly available data and two
for HER2/neu-expressing tumor onset and growth. Tumor phenotypes were associated novel retrospective cohorts for transcriptomic-based immune TME subtype classifi-
with mouse haplotypes to identify quantitative trait loci (QTL) on chromosomes (chr) 2, 6, cation (immune-hot vs. immune-cold) and TLS signature, along with associated clinical
and X linked with aggressive HER2/neu tumors. Of the 35 candidate genes harboring and genomic data. Those with other histological subtypes of non-small cell lung cancer
protein-coding changes, homologous genes on human chr 1, 10, and X were analyzed using or those who harbored EGFR mutations or ALK rearrangements were excluded from our
Caris Life Sciences datasets of NSCLC and SCLC: 25,143 primary NSCLC; 12,365 meta- study. The cellular decomposition within tumor samples was calculated using the
static NSCLC; 621 primary SCLC; and 971 metastatic SCLC where primary biopsies are deconvolutional Kassandra algorithm. Survival analysis was evaluated using log-rank
those taken from the lung and metastatic biopsies from sites other than the lung. Patient test and multivariate Cox regression adjusted by PD-L1 status, KEAP1/STK11/KRAS/
cohorts stratified by gene expression (top vs. bottom 50%) were correlated with OS, ICI TP53 mutational status, immune TME subtype, and TLS signature. All statistical an-
response, and time on pembrolizumab or atezolizumab treatment (TOT). Results: Several
alyses were performed using Python. Results: A total of 514 patients were included
homologous candidate genes correlated with lung cancer survival in both NSCLC and
from five cohorts, with 272 and 505 having genomic and transcriptomic data, re-
SCLC. Notably, high expression of LILRB4, a macrophage-specific checkpoint molecule,
spectively. 59% of patients with LUAD or LUSC exhibited an immune-cold phenotype,
was associated with improved survival (HR 0.66–0.84, p , 0.001) across all lung cancer
which correlated with adverse overall survival (OS) and progression-free survival (PFS)
types and sites. Additionally, elevated TSC22D3, a glucocorticoid receptor-activated gene
regulating anti-inflammatory pathways, including LILRB4 expression, was positively as-
than immune-hot phenotype in LUAD. However, the ICI response rates were similar in
sociated with OS in metastatic NSCLC (HR 0.82, p , 0.00001) and SCLC (HR 0.77, p , both groups. Superior PFS and ICI response rates were observed in patients with high
0.001). Expression of LILRB4 and TSC22D3 further enhanced survival in ICI-treated pa- TLS signatures (. 88th percentile) in LUAD, even after multivariate adjustments.
tients (SCLC: HR 0.72; NSCLC: HR 0.82; p , 0.001). In NSCLC, high LILRB4 expression Immune signatures that were positively associated with ICI response included the
conferred an 18% improved TOT with pembrolizumab (HR 0.858, p , 0.0001). infiltration and trafficking of T and NK cells for LUAD and B-cell percentage for LUSC. In
Conclusions: LILRB4 and TSC22D3 are key genes linked to improved survival outcomes in contrast, CD8+ T-cell abundance did not correlate with ICI response. The presence of
both SCLC and NSCLC, likely through their roles in mitigating inflammation. Their as- KEAP1 or STK11 mutations also did not affect the response rates but were associated
sociation with enhanced ICI responses underscores their potential as therapeutic targets. with shorter OS and PFS. Conclusions: Transcriptomic-based immune-hot TME and
Future research will evaluate the relationship between LILRB4 and TSC22D3 expression high TLS signature may serve as novel predictive and prognostic biomarkers in patients
and HER2/neu status in NSCLC and validate protein-level correlations in positive cells. Our with LUAD, while the presence of KEAP1 or STK11 mutations only offered prognostic
long-term goal is to determine the therapeutic potential of LILRB4 and TSC22D3 in SCLC, values. Further prospective studies are warranted to expand to other treatment
NSCLC and HER2/neu-positive NSCLC. Research Sponsor: U.S. National Institutes of combinations with PD-(L)1 inhibitors. Research Sponsor: None.
Health; 7R01CA278818-02.

2635 Poster Session 2636 Poster Session

CCR8 positive Tregs and their correlation with immunotherapy response in Biological determinants of immune exclusion in non-small cell lung cancer:
advanced non-small cell lung cancer (NSCLC). First Author: Jean Philippe An analysis of the precision medicine BIP study. First Author: Jean-Philippe
Guégan, Explicyte, Bordeaux, France Guegan, ImmuSmol, Bordeaux, France
Background: Regulatory T cells (Tregs) expressing the chemokine receptor CCR8 are Background: Immune exclusion has been associated with resistance to immunotherapy in NSCLC.
pivotal modulators of the tumor immune microenvironment. CCR8 has recently emerged However, its biological determinants remain largely unknown. Instead of relying on preclinical models,
as a promising therapeutic target due to its selective expression on activated Tregs in the high-throughput profiling of patient samples using spatial transcriptomics (ST) and multiplex im-
munofluorescence (m-IF) offers a powerful approach to dissect immune profiles and uncover key
tumor microenvironment and its role in promoting immunosuppression. This study in-
drivers of immune response and resistance. Methods: Tumor samples collected from NSCLC patients
vestigates the prognostic and therapeutic implications of CCR8-positive Tregs in non- enrolled in the BIP precision medicine study (NCT02534649) prior to initiation of ICI therapy and
small cell lung cancer (NSCLC), focusing on their impact in relation to tertiary lymphoid divided into Discovery and Validation cohorts (n = 148 and 117, respectively). Response to treatment
structure (TLS) status. Methods: A validated 6-plex multiplex immunofluorescence (mIF) was assessed as per RECIST criteria. Multiplex immunohistochemistry (mIHC) with CD8 and panCK
panel was used to analyze tumor samples from NSCLC patients treated with immune markers was used to classify tumors as desert, excluded or inflamed through pathologist assessment
checkpoint blockers (ICB) in the BIP precision medicine study ([Link]: (PA) and image analysis ST using the NanoString GeoMx Whole Transcriptome Atlas compared gene
NCT04389143). Markers included CD4, CD8, CD20, FoxP3, PanCK, and CCR8, alongside expression profiles between inflamed and excluded tumors Spatially resolved T-cell receptor (TCR)
DAPI staining to assess immune contexture (infiltrated, excluded, desert), TLS status, and profiling assessed clonal diversity and repertoire to evaluate T-cell functionality. m-IF was used for
proteomic validation. Results: In both the training and validation cohorts, excluded tumors dem-
CCR8/FOXP3 double-positive Tregs. Clinical outcomes, including progression-free sur-
onstrated lower objective response rates (ORR), progression-free survival (PFS), and overall survival
vival (PFS) and objective response rate (ORR), were analyzed in 50 responders and 50 (OS) compared to inflamed tumors (Table 1), independent of PD-L1 expression in multivariate analysis.
non-responders. Findings were validated using transcriptomic data from the POPLAR ST identified marked overexpression of HLA-A/B (MHC class I) and CD74 (involved in MHC class II
(NCT01903993) and OAK (NCT02008227) studies, which evaluated atezolizumab versus processing) in inflamed tumors versus excluded tumors, underscoring their crucial roles in antigen
docetaxel in advanced NSCLC. Kaplan-Meier curves, hazard ratios, and Cox regression presentation. These results were validated by m-IF. Spatially resolved TCR profiling demonstrated
models were used for survival analyses. Results: CCR8-expressing Tregs were signifi- higher Gini coefficients and lower Shannon entropy in excluded tumors, indicating a more oligoclonal
cantly enriched in infiltrated tumors, showing a 1.5-fold increase compared to excluded TCR repertoire dominated by fewer T-cell clones. These findings suggest impaired antigen recognition
tumors (p = 0.057), a 3.3-fold increase compared to desert tumors (p = 0.001), and a 1.8- and restricted T-cell diversity in excluded tumors. Conclusions: Our classification approach using
fold increase in TLS-positive tumors compared to TLS-negative tumors (p = 0.003). These mIHC and IA offers a practical, and clinically actionable biomarker for predicting response to ICI
therapy. Immune exclusion, prevalent in NSCLC, is associated with resistance to ICI and characterized
findings highlight that activated Tregs co-infiltrate with CD8 T cells and other immune cell by reduced expression of key antigen presentation molecules such as HLA-A/B and CD74 and a
types. This enrichment was confirmed in samples from the POPLAR and OAK studies restricted TCR repertoire highlighting the need for novel strategies to overcome this immune barrier.
using transcriptomic analyses (3-fold increase, p = 2e-16). Due to their correlation with Research Sponsor: None.
overall immune cell infiltration, the presence of CCR8-positive Tregs was significantly
associated with better survival (HR 0.45, p , 0.001) across the entire patient cohort. Phenotype Objective Response Rate (ORR) PFS (Median, Months)
However, when stratified for TLS-positive tumors, the presence of activated Tregs was Discovery Inflamed 58% 12.8 (95% CI: 6.16-NA)
associated with diminished objective response rate and progression-free survival (n=32)
Excluded 38.7% 4.1 (95% CI: 2.4-10.3)
suggesting a negative impact of these immunosuppressive cells on response to ICB. (n=65)
Conclusions: This study provides the first evidence linking CCR8-positive Tregs with Desert 20% 2.8 (95% CI: 1.9-6.9)
immunotherapy resistance in NSCLC, particularly in TLS-positive tumors. These findings (n=51)
Validation Inflamed (n=40) 57.5% 11.3 (95% CI: 4.6-NA)
parallel observations in TLS-positive sarcomas, where Treg abundance predicted poor Excluded 43.3% 6.1 (95% CI: 3.4-14.9)
outcomes (Italiano et al., Nature Medicine, 2022). This study supports the exploration of (n=30)
CCR8-targeted therapies to deplete immunosuppressive Tregs and enhance the efficacy Desert 31.9% 4.4 (95% CI: 2.3-7.2)
(n=47)
of immunotherapy in TLS-positive NSCLC. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 171s

2637 Poster Session 2638 Poster Session

The predictive role of TRAIL gene expression in immune checkpoint inhibitor Randomized phase II trial evaluating the combination of TG4001, an HPV16
(ICI)-treated patients (pts). First Author: Obada Ehab Ababneh, The University of therapeutic vaccine, and avelumab (ave) in patients (pts) with immunother-
Texas MD Anderson Cancer Center, Houston, TX apy-naı̈ve recurrent and/or metastatic (R/M) HPV16-positive cervical or
Background: Despite FDA-approved molecular biomarkers such as PD-L1 levels, tumor anogenital cancer. First Author: Christophe Le Tourneau, Department of Drug De-
mutation burden (TMB), and microsatellite instability (MSI) status, only ~30% of velopment and Innovation (D3i), Institut Curie, Paris-Saclay University, Paris, France
matched cancer pts respond to ICI. TRAIL, a protein product of TNFSF10 gene, is a Background: Human papillomavirus (HPV) is a small DNA virus associated with cer-
member of the TNF superfamily involved in regulating immune responses and inducing vical, anogenital (AG) cancers and squamous cell carcinoma of the head and neck.
apoptosis when bound to either Death Receptor 4 or 5 (DR4/5) especially in cancer cells. TG4001 is a therapeutic vaccine based on modified vaccinia virus Ankara with insertion
While TRAIL has been studied for its prognostic roles in cancer, its predictive value for of modified non-oncogenic HPV-16 E6 and E7 antigens and interleukin-2 as adjuvant.
pts treated with ICI remains unclear. This study investigates the association between The phase I trial of TG4001 combined with ave showed a favorable safety profile
TRAIL expression and outcomes in ICI-treated pan-cancer pts. Methods: RNA ex- (Borcoman E. et al, 2023). Methods: Pts with R/M cervical and anogenital cancer and
pression levels of TRAIL were assessed in a cohort of 217 pan-cancer pts treated with who were checkpoint inhibitors naı̈ve were randomized independent of PD-L1 expression
ICIs at the University of California San Diego (UCSD) Moores Cancer Center. RNA between ave plus TG4001 or ave alone. Pts were required to have no more than one prior
transcripts were normalized using an internal housekeeping gene profile of 735 tumors line of therapy for R/M disease and no liver involvement. Primary endpoint was PFS.
and 35 histologies. Transcript abundances were percentile-ranked (0–100) and cate- Subgroup analysis (cervical, anal, other genital cancer) was preplanned in the protocol.
gorized as high ($75th percentile) or low ( , 75th percentile). Associations between Results: 90 pts were randomized between June 2021 and April 2024. 49 (54%), 27 (30%)
TRAIL expression and overall survival (OS) and progression-free survival (PFS) were and 14 (16%) pts had cervical, anal, and other genital cancers, respectively. Patients’
analyzed. Statistical significance was defined as p-value # 0.05. Results: Among the demographics were well balanced between the 2 arms. Median PFS (mPFS) was 3.0 and
217 ICI-treated pts, the median age was 61.9 years, and 56.2% were female. The most 2.8 months (mo) in the experimental and control arm, respectively (HR=0.87 [90%CI:
common cancer types were colorectal (24.9%), breast (8.8%), ovarian (8.3%), pancreatic 0.59-1.29], p=0.28). In the cervical cancer subgroup, mPFS was 4.3 and 2.1 mo in the
(7.4%), and lung (6.5%) cancers. FDA-approved ICI biomarkers favorable rates were PD- experimental and control arm, respectively (HR=0.58 [90%CI: 0.33-1.01], p=0.053).
L1 $1% in 40.1%, TMB-high ($10 mut/Mb) in 11.5%, and MSI-high in 4.8%. Based on the Overall Response Rate (ORR) in the whole population was 15.2% (7/46pts) in the ex-
ICI type used, 91.7% received anti-PD-(L)1 while 7.8% received anti-CTLA-4 with anti-PD- perimental arm and 13.6% (6/44pts) in the control arm. In the cervical cancer subgroup
1. Pts with high TRAIL expression (24%) had similar PD-L1, TMB, MSI profiles (p . 0.05). ORR was 20% (5/25pts) in the experimental arm and 8.3% (2/24pts) in the control arm.
Pts with high levels of TRAIL expression achieved better OS (HR = 0.41, 95%CI:0.25-0.69, There were no new safety signals. Three pts (6.5%) in the experimental arm and 2 pts
p = 0.0004) and PFS (HR = 0.67, 95%CI:0.47-0.96, p = 0.027). After adjusting for age, sex, (4.5%) in the control arm presented grade 3 or 4 treatment-related AEs. Translational
cancer type, PD-L1 IHC level ($1% vs. , 1%), TMB ($10 mut/Mb vs. , 10mut/Mb), MSI analysis including immunogenicity results will be presented. Conclusions: TG4001
status (stable vs. unstable), KRAS, TP53 and CDKN2A/B alteration status and immune combined with ave did not improve PFS over ave alone in the whole patient population.
checkpoints genes expression, overall survival remained significantly associated with Preplanned subgroup analysis in cervical cancer showed a positive efficacy signal in the
better survival in TRAIL high pts compared to TRAIL low pts (HR = 0.38, 95%CI:0.19-0.76, combined arm. Avelumab was provided by the healthcare business of Merck KGaA,
p = 0.006). However, no difference was found between both groups in regard to Darmstadt, Germany (CrossRef Funder ID: 10.13039/100009945). Clinical trial infor-
progression-free survival (HR = 0.68, 95%CI:0.42-1.10, p = 0.11). Conclusions: High mation: NCT03260023. Research Sponsor: None.
TRAIL expression is associated with improved overall survival in ICI-treated pan-cancer
pts, independent of cancer type or other predictive biomarkers. These findings suggest
TRAIL as a potential biomarker for ICI benefit. Larger studies in diverse and real-world
settings are warranted to validate these findings. Research Sponsor: None.

2639 Poster Session 2640 Poster Session

Induction of neoantigen-specific immune responses by [Link] in com- An open-label single-center investigator-initiated exploratory clinical study
bination with atezolizumab in heavily pretreated patients with advanced in patients with refractory or recurrent solid tumors: R-ISV-FOLactis trial.
solid tumors: Final analysis of the phase 1b VB N-02 trial. First Author: First Author: Ruojing Lv, Comprehensive Cancer Centre of Drum Tower Hospital, Medical
Sebastian Ochsenreither, Charité University of Medicine Berlin Comprehensive Cancer School of Nanjing University, Clinical Cancer Institute of Nanjing University, Nanjing,
Center, Berlin, Germany China
Background: [Link], a personalized DNA-based neoantigen vaccine, was evaluated with Background: Soft tissue sarcomas (STSs) are a highly complex group of tumors and the
atezolizumab in a Phase 1b trial to assess safety, clinical activity, and immune responses in treatment still remains a challenge. Immunotherapy has become a powerful clinical strategy,
heavily pretreated patients with advanced solid tumors. The NeoSELECT platform enriches for especially the application of therapeutic tumor vaccine. Hypofractionated radiotherapy
clonal neoantigens by analyzing RNA and circulating tumor DNA, incorporating frameshift (HFRT) can serve as an in situ vaccine and provide durable local control. We also develop a
antigens and single nucleotide variants. Methods: This open-label, dose-escalation trial bifunctional engineered Lactococcus lactis (FOLactis) which expresses an encoded fusion
investigated [Link] across three dose levels (3, 6 and 9 mg) combined with atezolizumab protein of Fms-like tyrosine kinase 3 ligand and co-stimulator OX40 ligand to conduct in situ
(1200 mg Q3W). Eligible patients had advanced or metastatic solid tumors, sufficient tumor vaccination (ISV). In this study, we establish a novel R-ISV-FOLactis strategy, which refers to
material for vaccine manufacturing, at least 10 identified tumor neoantigens, and measurable the combination of HFRT, intratumoral (IT) injection of FOLactis and synergetic anti-PD-1
disease. Immune responses were assessed using ELISpot assays (in vitro stimulation and therapy, to further enhance efficacy and realize the activation of the whole immunity cycle.
ex vivo), T cell receptor sequencing, and flow cytometry. Endpoints included safety, immune Methods: This study is an open-label, single-center trial aimed at patients with advanced
response, and antitumor activity per RECIST v1.1. Results: At study completion (October STSs who are unresponsive or intolerable to previous standard treatment. Patients will be
2024), 26 patients (median age 61 years, range 28–72; 62% female) received at least one dose treated with HFRT, the IT injection of FOLactis and PD-1 inhibitors. The primary endpoint is
of [Link]. Median prior therapy lines for advanced disease were three (range 1–6), and the objective response rate (ORR) of target lesions at 3 month and 6 month. The secondary
54% had prior immunotherapy, including checkpoint inhibitors. Tumor types included head endpoint includes the disease control rate (DCR) of target lesions, progression-free survival
and neck squamous cell carcinoma (15%), triple-negative breast cancer (15%), and others (PFS), overall survival (OS), etc. Results: This study started from July 2022 and ended in
(31%; most were tumors with low tumor mutational burden). The majority (69%) of evaluable December 2023, involving 30 eligible patients with solid tumors and 16 of them are patients
patients had low or negative PD-L1 expression. Injection site reactions (15%) and fatigue with STSs. The ORR and DCR of all target lesions after three months are 27.6% and 93.1%
(12%), mainly Grades 1–2, were the most common adverse events. A dose-limiting Grade 3 respectively, and in sarcomas, the ORR and DCR are 11.1% and 88.9%. We calculate the ORR
transient blood pressure increase occurred in the 9 mg cohort. No treatment-related serious and DCR of target lesions after six months, which are 56.3% and 100% respectively, and in
events or deaths occurred. Across all dose levels, [Link] induced robust and durable sarcomas, these are 41.7% and 100%. Systemic median PFS are 2.87 months. Median PFS of
neoantigen-specific immune responses. In vitro stimulated ELISpot assays detected vaccine- target lesions has not been reached. Among the evaluable target lesions, 6-month EFS is 50%
induced T cell responses in 85% (11/13) of evaluable patients and in 58% of evaluated in sarcomas (6/12) and 50% in all patients (8/16). We test the level of cytokines before and
neoantigens, while ex vivo ELISpot demonstrated responses in 22% (4/18) of patients and 5% after the first treatment and find that the changes in the percentage of CD8+ T cells,
of evaluated neoantigens. T cell receptor sequencing showed persistent T cell clone ex- CD103+CD8+ T cells and CD39+CD8+ T cells have significance. Moreover, in sarcomas, PFS is
pansion in 9/11 patients, indicating durable immune responses. Putative neoantigen-specific relevant to the level of CD103+CD8+ T cells before treatment, CD39+CD8+ T cells after
clones were detected in 6/7 analyzed patients, with persistent expansion in 4. All patients treatment, NK cells before treatment and immature DC cells after treatment. The most
achieving stable disease (34.8%, 8/23) exhibited neoantigen-specific immune responses. common treatment-related adverse events (TRAEs) are fever (83.3%), lymphocytopenia
Conclusions: [Link] combined with atezolizumab demonstrated a favorable safety (53.3%), hypocalcemia (30%), neutrophilia (26.7%) and nausea (26.7%). Grade$3 TRAEs
profile while eliciting robust, durable immune responses across all dose levels, even in heavily occur in 11 patients, including lymphocytopenia (30%), fever (6.7%), leukopenia (3.3%),
pretreated patients with advanced solid tumors. The potential correlation between immu- anemia (3.3%) and cardiac insufficiency (3.3%). Conclusions: The R-ISV-FOLactis strategy
nogenicity and clinical benefit supports further exploration in earlier treatment settings. demonstrates its efficacy among patients with advanced STSs and induces certain anti-tumor
Clinical trial information: NCT05018273. Research Sponsor: Nykode Therapeutics in col- immunity. The ISV of “FOLactis” may provide a promising option in the treatment of recurrent
laboration with Roche/Genentech. or refractory solid tumors. Clinical trial information: ChiCTR2200060660. Research Sponsor:
None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
172s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2641 Poster Session 2642 Poster Session

First-in-human study of ZGGS15, a dual-specific antibody targeting LAG-3 Adverse events profile of novel agents targeting immune checkpoints be-
and TIGIT, as monotherapy in patients with advanced solid tumors. First yond PD-1/PD-L1 and CTLA-4 in solid tumors: A meta-analysis. First Author:
Author: Ji Zhu, Cancer Hospital of the University of Chinese Academy of Sciences Yu Fujiwara, Department of Medicine, Roswell Park Comprehensive Cancer Center,
(Zhejiang Cancer Hospital), Hangzhou, China Buffalo, NY
Background: ZGGS15 is a novel humanized bispecific antibody of anti LAG-3 and TIGIT. Background: PD-1, PD-L1, and CTLA-4 blockade are standard therapies in multiple solid
It could reverse Treg inhibition of T cells and NK cells, and kills tumor cells by restoring tumors, and novel agents targeting alternative co-stimulatory and co-inhibitory immune
the function of T cells and NK cells. In non-clinical studies, ZGGS15 showed synergistic checkpoints are under development. Immune checkpoint inhibitors are well known to
anti-tumor effects with the anti-PD-1 antibody. We conducted a Phase 1 dose escalation cause immune-related adverse events (irAEs) and multiple studies reported the accurate
and expansion study to assess tolerability, safety, and efficacy of ZGGS15 as mono- incidence of irAEs from PD-1, PD-L1, and CTLA-4 blockade. With the increasing inves-
therapy in patients with advanced solid tumors. Methods: In the dose-escalation tigation and anticipated approval of novel immunotherapy agents, understanding their
phase, a standard "3+3" design, with an accelerated titration for the starting dose. toxicity profiles is critical. Methods: We systematically searched PubMed/MEDLINE,
Total of 6 dose levels, from 0.3 to 30 (0.3, 1, 3, 10, 20, 30) mg/kg administered by the EMBASE, and Web of Science for clinical trials published up to December 1st, 2024.
intravenous infusion, once every three weeks, in patients with advanced solid tumors Studies evaluating the safety of agents targeting co-inhibitory checkpoints (B7-H3, CD47,
who had failed to the available standard treatments. The first treatment cycle (21 days) TIGIT, LAG-3, and TIM-3) or co-stimulatory checkpoints (OX40, 4-1BB, CD27, ICOS, GITR,
was defined as the dose-limiting toxicity (DLT) observation period. The study as- CD70, and CD40) in solid tumors were included. Incidence rates of grade 1-5 (G1-5) and
sessments included tolerability, safety, preliminary efficacy, etc. and tumor responses grade 3-5 (G3-5) treatment-related adverse events (trAEs) and irAEs were extracted.
were assessed by RECIST1.1 and iRECIST criteria. Results: As January 8 2025, a total of Toxicity data were derived from phase 2 and 3 trials, as well as phase 1/2 trials with safety
22 patients (9 males and 13 females), with a median age of 59 years, participated in the information reported at the recommended phase 2 dose. Random-effects meta-analysis
dose escalation from 0.3 to 30 mg/kg and completed the DLT observation. Of the 22 was used to pool odds ratios (ORs) from two-arm studies evaluating the addition of LAG-3
patients, 11 (50.0%) received at least 3 prior lines of therapies, and eight (36.4%) had or TIGIT blockade to control-arm therapy, and proportional meta-analysis was conducted
previously treated with PD-1 or PD-L1 inhibitors. No DLT events were observed. TRAEs to analyze AE incidence across immunotherapy subtypes. Results: A systematic review
occurred in 20 (90.1%) patients, with only one patient (4.5%) experienced a Grade 3 TRAE identified 27 clinical trials with 40 cohorts comprising 3,946 patients and evaluating 10
of lymphocyte count decreased, and no Grades 4 or 5 TRAEs were reported. Among the immune checkpoints (B7-H3, LAG-3, TIGIT, CD47, OX40, CD137. TIM-3, CD40, CD27, CD40,
17 patients who had at least one post-baseline tumor scan, six had achieved stable ICOS). Meta-analyses showed that the addition of LAG-3 blockade to either PD-1
disease (SD) with a disease control rate (DCR) of 35.3%. In the subgroup of 8 patients blockade-based therapy or placebo was associated with increased G3-5 trAEs (OR
with lung adenocarcinoma, 5 (62.5%) had achieved SD, including two patients who 1.79, 95% confidence interval [CI]: 1.26–2.54, p = 0.001), G3-5 adrenal insufficiency (OR
had $ 2 prior lines of treatments and maintained SD over 36 weeks. Conclusions: The 8.43, 95% CI: 1.04 - 68.37, p = 0.046), G1-5 adrenal insufficiency (OR 4.81, 95% CI: 1.81-
results showed that ZGGS15 was well tolerated and had a very good safety profile. It is 12.78, p = 0.002) and arthralgia (OR 2.07, 95% CI: 1.29-3.30, p = 0.002). The addition of
anticipated that when in combination with other anti-cancer therapies, e.g., an anti-PD-1 TIGIT blockade to PD-L1 blockade-based therapy was associated with increased G1-5
or PD-L1 antibody, for advanced solid tumors, ZGGS15 may provide synergistic anti- rash (OR 2.32, 95% CI: 1.01–5.34, p = 0.048) (other outcomes will be shown). Proportional
tumor effects and further enhance treatment benefits. Clinical trial information: meta-analysis revealed varying irAE patterns across agents: G5 trAEs (0.9-2.9%), G3-5
NCT05864573. Research Sponsor: Suzhou Zelgen Biopharmaceuticals. Co., Ltd. pneumonitis (0.5-5.5%, highest in TIM-3 blockade), G3-5 colitis (0.2-5.4%, highest in LAG-
3 blockade), G3-5 hepatitis (1.5-5.5%, highest in TIM-3 blockade), G3-5 rash (0.8%-18.4%,
highest in CD40 agonists), G3-5 adrenal insufficiency (1.7-8.4%, highest in TIGIT
blockade) (details of all outcomes will be presented). Conclusions: This study highlights
the distinct toxicity profiles of novel immunotherapy agents, providing essential safety
data to support clinicians as these therapies move toward anticipated clinical approval.
Research Sponsor: None.

2643 Poster Session 2644 Poster Session

JAK inhibitor for the treatment of steroid refractory and life-threatening Albumin-myosteatosis gauge as a prognostic biomarker in patients treated
immune-related adverse events secondary to immune checkpoint inhibitors. with immune checkpoint inhibitors. First Author: Taha Koray Sahin, Hacettepe
First Author: Rami Habib, McGill University, Montreal, QC, Canada University, Department of Medical Oncology, Ankara, Turkey
Background: Immune checkpoint inhibitors (ICIs) boost anti-tumor immune responses Background: Although immune checkpoint inhibitors (ICIs) have heralded a new era in
but carry the risk of off-target effects, manifesting as immune-related adverse events cancer treatment, many patients do not respond, underscoring the need for biomarkers.
(irAEs). Approximately 20% of irAEs are refractory to steroids, requiring subsequent The albumin-myosteatosis gauge (AMG) is a recently developed integrated measure of
immunosuppressive therapies, while a smaller subset presents with fulminant reactions myosteatosis and serum albumin levels, reflecting systemic inflammation and mal-
requiring multiple agents simultaneously. Although biologics like TNF and IL-6 inhibitors nutrition. Herein, we investigate the prognostic value of AMG in patients with advanced
are often used, their targeting of single inflammatory pathways may be insufficient to cancer treated with ICIs. Methods: A total of 308 patients with advanced cancer treated
resolve irAEs in these critical scenarios. We conducted a prospective study to assess the with ICIs were included. Skeletal muscle index (SMI) and skeletal muscle radiodensity
safety and efficacy of oral JAK inhibitors, molecules capable of rapidly modulating (SMD) were measured from computed tomography images obtained at the level of the L3
multiple inflammatory cytokine signals, in managing steroid-refractory or life-threatening vertebra. The AMG was calculated by multiplying SMD by albumin and expressed as an
irAEs. Methods: The MIRAE Biobank is a prospective cohort study of cancer patients arbitrary unit (AU). Survival outcomes were assessed using Kaplan-Meier survival curves
treated with ICI at the Jewish General Hospital in Montreal, Canada. Patients with steroid and Cox regression models. Results: The median age (interquartile range) was 63 (55-
refractory subjects with grade . 3 irAE or persistent grade 2 toxicity despite optimal 70), and 198 (64.3%) were male. Non-small cell lung cancer (NSCLC) was the most
therapy, as well as life-threatening irAEs in the first-line setting treated with JAK inhibitors common primary cancer (28.2%), followed by renal cell carcinoma (RCC) (20.8%) and
were extracted from the database and their clinical data was summarized. Among those melanoma (20.2%). Regarding AMG, the cutoff values were determined to be 109.38 AU
who survived at least 30 days post-JAK inhibitor, we compared characteristics of re- for males and 102.11 AU for females. Multivariable analyses revealed that lower AMG
sponders and non-responders. Responders were defined as resolution of the irAE to values were independently associated with decreased OS (HR: 1.43; 95% CI: 1.08-1.90;
grade , 1 and , 10mg prednisone equivalents without any relapses during a 30-day p=0.012) and PFS (HR: 1.39; 95% CI: 1.07-1.79; p=0.011) compared to the AMG high-
period. Results: In this series, 29 patients were treated with JAK inhibitors for refractory group. Conclusions: Our findings suggest AMG, an easily accessible novel biomarker, is
or life-threatening irAEs. Mean age was 69 years, 34.5% were women, 82.8% received anti- an independent prognostic factor for survival in patients with advanced cancer treated
programmed cell death protein-1 (PD-1) antibodies alone, and 13.8% patients were with ICIs. Prospective studies are required to validate these findings and evaluate the
treated with the combination of anti-cytotoxic T lymphocyte antigen-4 and anti-PD-1 role of AMG measurement in aiding treatment choices. Research Sponsor: None.
antibodies. Cancer types were primarily melanoma (10, 34.5%) and lung cancer (6, 20.7%).
Primary irAEs for which JAK inhibitors were initiated included myocarditis (n=11), colitis
(n=4), arthritis (n=4), hepatitis (n=4), encephalitis (n=2), pneumonitis (n=2), myasthenia
gravis (n=1) and sicca (n=1). JAK inhibitors were used as second- (refractory to steroids
alone), third- or fourth- or more line in 9, 11 and 9 patients, respectively. Median duration
of JAK inhibitor exposure was 30.5 days. Among the 24 patients who survived at least
30 days, 17 (71%) responded after a median of 11 days from initiation of the JAK inhibitor.
Interestingly, this included 6/8 patients with myocarditis, 4/4 with arthritis and 2/3 with
colitis. Of those who responded to JAK inhibitor, 11/18 were steroid refractory and 6/6
were life-threatening cases requiring simultaneous treatment with steroids.
Conclusions: This preliminary data suggests that JAK inhibitors may be effective at
treating various types of steroid-refractory and life-threatening irAEs. Research Sponsor:
Arthritis Society Canada; #23-313.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 173s

2645 Poster Session 2646 Poster Session

Tumor flare reactions secondary to T-cell engaging immunotherapies: A Evaluating the role of exercise in modulating immunity and immunotherapy
study from the French REISAMIC registry. First Author: Alexandre Xu-Vuillard, outcomes in cancer: A systematic review. First Author: Samhitha Gundakaram,
Gustave Roussy, Villejuif, France Joan C. Edwards School of Medicine, Marshall University, Huntington, WV
Background: Tumour flare reactions (TFR) were recently reported in patients receiving T- Background: Immunotherapy has become a key cancer treatment, improving survival
cell engagers (TCE) and require further investigations. This study aims to investigate and reducing side effects. However, its effectiveness can be influenced by immune
incidence, predictive factors, and outcomes of pts with TFRs related to TCE. system function, overall health, and treatment-related side effects. Exercise, known for
Methods: This observational cohort study is nested in the French academic pharma- its health benefits, may also modulate immune responses and enhance immunotherapy
covigilance register, Registre des Effets Indésirables Sévères des Anticorps Monoclonaux outcomes. Despite promising evidence, the impact of exercise on immune function and
Immunomodulateurs en Cancérologie (REISAMIC, CNIL number 2098694v0). All patients cancer treatment remains insufficiently understood. This review aims to assess the role
treated at Gustave Roussy (France) with TCE, for all tumor indications except acute of exercise in modulating immunity and improving immunotherapy outcomes in cancer
leukemia, were included. The main objectives were to determine the incidence, predictive patients. Methods: A systematic review was conducted following PRISMA guidelines,
factors, associated clinical and biomarker profiles, as well as the outcomes of patients. with a comprehensive search of PubMed, Cochrane Library, EMBASE, and Clinical-
TFRs were clinically defined with transient worsening of tumor symptoms including tumor [Link] for studies published from 2010 to 2025. Eligible studies included randomized
pain, effusions, or compressive symptoms, mimicking tumor growth but without controlled trials (RCTs), non-randomized trials, pilot studies, and systematic reviews
reflecting disease progression. Statistical analyses included log-rank tests and Cox investigating exercise interventions on immune function and immunotherapy in cancer.
regression models. Results: Overall, 222 TCE-treated patients were included, median Key outcomes included changes in immune markers, immune function, and quality of
[range] age: 53 [5 - 87] years; male-to-female ratio: 1.44, median prior lines of therapies life. Data were extracted and analyzed using standardized protocols. Results: Eight
was 3 [2-9]. Of them, 147 (66.2%) had solid tumors, including prostate cancer (n=51), high- studies, involving 1,172 cancer patients across various types (lymphoma, CLL, mela-
grade serous ovarian carcinoma (n=28), and small-cell lung cancer (n=24), and 75 (33.8%) noma, NSCLC, breast, ovarian, prostate), were included. Exercise modalities studied
had hematologic malignancies, primarily multiple myeloma (MM, n=35) and diffuse large included aerobic exercise, resistance training, cycling, yoga, and mind-body practices
B-cell lymphoma (DLBCL, n=31). Common TCE targets included CD3/CD20 (n=40), CD3/ like qigong. Findings consistently showed positive effects of exercise on immune
KLK2 (n=36), CD3/BCMA (n=35), CD3/DLL3 (n=28), and CD3/B7-H4 (n=27). TFRs occurred function and treatment outcomes. Two studies reported that exercise mobilized T cells
in 54 pts (24.3%), with higher TFR frequency in solid tumors (34.0%) than lymphomas and NK cells, enhancing immune responses. Another two demonstrated improved ef-
(12.1%) and none in MM. Median TFR onset was 1 day [1–8]. Symptoms of TFRs were pain ficacy of immunotherapeutic agents such as rituximab. Additionally, three studies
(92.6%), compression syndrome (22.2%), and effusion (7.4%). Severity of TFRs was grade indicated that exercise improved physical fitness, body composition, and overall quality
3–4 in 68.5% of cases. No TFR-related deaths occurred. TFRs management included of life. Conclusions: This review provides evidence that exercise may enhance immune
corticosteroids (35.2%), opioids (61.1%), paracentesis (7.4%), JJ stenting (3.7%) and responses and improve outcomes in cancer immunotherapy. While exercise appears to
tocilizumab (3.7%). TFRs correlated with transient increases of CRP (182 vs. 69 mg/L; be a beneficial adjunctive therapy, the optimal type, intensity, and duration remain
p=0.0002) and LDH (316 vs. 225 U/L; p=0.04). Patients with TFRs were more frequently unclear. Further large-scale, high-quality trials are needed to define effective exercise
exposed to cytokine release syndrome (p=0.0001). Predictors included tumor serous regimens and explore their impact on immunotherapy across various cancer types. The
localization (p=0.052) and a higher CD4+/CD8+ ratio in blood (p=0.048). In solid tumors study is registered in PROSPERO: CRD42024627822. Research Sponsor: None.
pts; TFRs were associated with higher response rates (18.0% vs. 6.3%; p=0.013) and
disease control rates (72.0% vs. 49.5%); PFS (2.83 vs. 2.76 months; p=0.865) and OS (9.92
vs. 9.72 months; p=0.539) were comparable regardless of TFRs. Conclusions: TFRs
related to TCE are clinically significant adverse events, primarily observed in solid tumor
pts. TFRs may indicate a unique pattern of antitumor response distinct from progression.
Better recognition and management of TFRs should help to optimize tolerability of TCE
therapies. Research Sponsor: Gustave Roussy.

2647 Poster Session 2648 Poster Session

Safety, pharmacokinetics (PK), pharmacodynamics (PD), and preliminary Solid tumor–specific patterns of immune-related adverse events due to
efficacy of HLX301, a bispecific antibody targeting PD-L1 and TIGIT, in immune checkpoint inhibitor. First Author: Shabnam Eghbali, Vanderbilt University
patients with advanced solid tumors. First Author: Michelle Frances Morris, Medical Center, Nashville, TN
Sunshine Coast University Hospital, Birtinya, Australia Background: Immune checkpoint inhibitors (ICIs) are the backbone of therapy for several
Background: Immune checkpoint proteins PD-L1 and TIGIT are important components of solid tumors; however, they have a unique toxicity profile that may limit treatment. The
cancer-related T cell immunosuppression. HLX301 is a humanized, bispecific IgG1 an- objective of this systematic review was to identify differences in type and frequency of
tibody targeting PD-L1 and TIGIT that showed anti-tumor activity in preclinical studies. A immune-related adverse events (irAEs) across solid tumors. Methods: Using PubMed, we
phase 1/2 first-in-human study was conducted to evaluate HLX301 monotherapy in identified registrational phase 2 and 3 clinical trials of ICI-based therapy (i.e., single agent
patients with advanced solid tumors (NCT05102214). Here we report findings from the immunotherapy (single I/O), single I/O plus chemotherapy, single I/O plus kinase inhibitor,
dose escalation part (phase 1a). Methods: This multicenter study enrolled patients with double immunotherapy combination (double I/O), double I/O plus chemotherapy) for first-
locally advanced or metastatic solid tumors who had failed or were intolerant to standard line and second-line unresectable disease for which irAEs (dermatologic, endocrine,
therapy, or for whom no standard therapy was available. Phase 1a evaluated doses of gastrointestinal, hepatic, renal, pulmonary) were specified for the following tumor types:
0.25-15 mg/kg IV Q2W. Primary endpoints included safety, dose-limiting toxicity (DLT), melanoma, non-small cell lung (NSCLC), esophageal, colorectal (CRC), biliary tract (BTC),
and maximum tolerated dose (MTD). Secondary endpoints included PK, PD, and im- hepatic (HCC), renal (RCC), urothelial, endometrial, head and neck (H&N). Odds ratio (OR)
munogenicity. Results: As of Oct 27, 2023, 9 patients were enrolled (0.25 mg/kg, 3; 1 mg/ were used to analyze effect size. All analysis performed on Microsoft Excel. Results: 105
kg, 3; 2.5 mg/kg, 1; 5 mg/kg, 2). Patients were all White, 55.6% female, median age 72.0 trials (n = 32,896 patients) were identified with the most commonly studied regimens
yrs; 88.9% had metastatic disease; all had ECOG PS of 0 (44.4%) or 1 (55.6%). All patients being those that were PD-1 or PD-L1-based. While endocrinopathies were the most
had prior systemic cancer treatment, including 3 (33.3%) treated with PD-(L)1 blockade; 5 frequent irAE (~15-20%) with first-line single I/O, one tumor type was not more likely than
(55.6%) patients had $ 4 prior lines of therapy. All patients were included in DLT, safety, the other to develop endocrinopathies. Interestingly, patients with melanoma and RCC
and PK analyses. Median duration of HLX301 treatment was 10.3 weeks. One patient treated with first-line single I/O were significantly more likely to develop gastrointestinal
(11.1%) in the 5 mg/kg cohort reported DLT (grade 3 cytokine release syndrome [CRS]). irAE compared to those with NSCLC, CRC, HCC, urothelial, and H&N with OR of 3.41 –
MTD was not determined. All patients experienced at least one treatment-emergent 30.64 and 2.96 – 26.60, respectively. Second-line single I/O led to increased frequency of
adverse event (TEAE). TEAEs leading to death occurred in 3 (33.3%) patients, none of irAE and greater variation in the predominant irAE for a given tumor type – four tumor
these adverse events (AEs) were related to HLX301. Six (66.7%) patients experienced at types (melanoma, NSCLC, gastric, H&N) had dermatologic irAE as most frequent, five
least one treatment-related adverse event (TRAE). TRAE of grade $ 3 was reported in 1 (esophageal, CRC, HCC, urothelial, endometrial) had endocrine irAE, and two (BTC, RCC)
patient (11.1%; grade 3 CRS), who was also the only patient for whom TRAE led to had gastrointestinal irAE. Overall odds of developing irAE were greater with double I/O
treatment discontinuation. Treatment-related immune-related AEs occurred in 4 (44.4%) than with single I/O in the first-line setting and more pronounced in the second-line. For
patients and treatment-related infusion-related reactions (IRRs) in 2 (22.2%). TRAEs example, in melanoma, OR for endocrine irAE was 2.55 (95% CI 1.27 – 5.10) in first-line
occurring in $ 2 patients included IRR (22.2%) and arthralgia (22.2%). HLX301 exhibited and 17.03 (95% CI 8.04 – 36.05) in second-line and for hepatic irAE was 4.41 (95% CI 1.55
linear PK over 0.25-5 mg/kg after single infusion and very limited accumulation after – 12.50) in first-line and 7.63 (95% CI 2.45 – 23.78) in second-line. The addition of
multiple infusions. Mean PD-L1 and TIGIT receptor occupancy in peripheral CD3+CD8+ chemotherapy or kinase inhibitor did not significantly alter irAE frequency across tumor
cells reached saturation at 5 mg/kg. Anti-drug antibody was detected in 7 patients types. In fact, there were fewer irAEs in some tumor types with addition of kinase inhibitor
(77.8%). Among 8 efficacy-evaluable patients, 1 (5 mg/kg cohort) achieved partial re- in first-line unresectable disease compared to single I/O alone; for example, OR for
sponse and 2 achieved stable disease; objective response rate and disease control rate dermatologic irAE with kinase inhibitor compared to single I/O alone was 0.36 (95% CI
per RECIST 1.1 were 12.5% and 37.5%, respectively. Conclusions: HLX301 showed an 0.18 – 0.70) for melanoma and 0.22 (95% CI 0.08 – 0.59) for RCC. Conclusions: irAE
acceptable safety profile with preliminary anti-tumor activity. These findings could profiles vary across tumor type, treatment regimen, and line of therapy and do not
support further clinical investigation. Clinical trial information: NCT05102214. Research necessarily correlate with the primary tumor site. Research Sponsor: None.
Sponsor: Shanghai Henlius Biotech, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
174s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2649 Poster Session 2650 Poster Session

Impact of immune checkpoint inhibitor (ICI)-associated autoimmune he- Improved survival with sodium-glucose cotransporter-2 inhibitors and im-
molytic anemia (AIHA) on mortality in cancer patients: A retrospective mune checkpoint inhibitors in metastatic solid tumors. First Author: Sara
analysis. First Author: Haris Sohail, Charleston Area Medical Center, Charleston, WV Young, Division of Hematology and Oncology, University of Virginia, Charlottesville, VA
Background: Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment Background: Immune checkpoint inhibitors (ICI) are used in the first-line setting for the
but can cause serious immune-related adverse events; including rare yet severe au- treatment of many advanced solid tumor malignancies. Patients with type 2 diabetes
toimmune hemolytic anemia (AIHA) . This study evaluates the impact of ICI-associated mellitus (T2DM) have decreased response rates to ICI, and poor glycemic control is
AIHA on mortality in patients with solid cancers. Methods: A retrospective analysis associated with worse outcomes in patients with cancer. Further adjunctive therapies
using the TriNetX database examined patients with solid cancers treated with ICIs. increasing the efficacy of ICI in these patients are needed. Sodium-glucose
Patients were defined using ICD-10 codes and grouped into those who developed AIHA cotransporter-2 inhibitors (SGLT2i) have emerged as effective antihyperglycemic
after ICI use and those who did not. Baseline characteristics, including cancer diagnosis, medications with concomitant cardiovascular benefits. SGLT2i have been approved by
ICI medications, and comorbidities, were compared between groups using TriNetX’s the Food and Drug Administration for use in patients with DM and congestive heart
built-in t-tests and z-tests to calculate p-values. Confounding variables were adjusted failure (CHF). Pre-clinical studies have demonstrated the potential benefit of SGLT2i in
with 1:1 propensity score matching. Primary outcomes were 30- and 60-day mortality; slowing tumor growth in-vitro and in-vivo; however, there is a lack of clinical data
secondary outcomes included transfusions and hospitalizations. Outcomes were regarding SGLT2i use in patients with advanced malignancy receiving ICI. Methods: We
evaluated using measures of association, Kaplan-Meier log-rank tests, and a Cox performed a retrospective, matched cohort study of patients with stage IV malignancy
proportional hazards model. Results: Among 106,388 ICI-treated patients, 352 de- and T2DM or CHF, who were treated with ICI using the Epic Cosmos dataset
veloped AIHA. After matching (352 per group), the AIHA group had a significantly higher (2013–2024). Ten different solid tumor types, five ICI, and three SGLT2i were analyzed.
30-day mortality (15.25% vs. 5.14%; OR 3.493, 95% CI: 1.898-5.803, p , 0.0001) and 60- Among 4,808 patients treated with ICI, 282 had at least one overlapping cycle of ICI and
day mortality (22.87% vs. 6.86%; OR 4.195, 95% CI: 2.479-6.54, p , 0.001). Transfusions SGLT2i. 1:1 propensity score matching was conducted to balance baseline covariates
(OR 4.085, 95% CI: 2.897-6.525, p , 0.0001) and hospitalizations (OR 1.865, 95% CI: including cancer type, comorbidities, age at diagnosis, and sex. Kaplan-Meier curves
1.525-2.837, p , 0.0001) were also significantly higher in the AIHA group. Hazard ratios were generated to examine survival differences and Cox proportional hazards models
(HR) confirmed significant mortality risks in AIHA group at 30 days (HR: 3.16, 95% CI: were used to estimate hazard ratios (HR). The primary outcome was overall survival (OS)
1.849-5.402, p , 0.0001) and 60 days (HR: 3.673, 95% CI: 2.324-5.805, p , 0.0001) for the entire matched cohort and sub-groups by cancer, ICI, and SGLT2i types.
(table 1). HRs for transfusion and hospitalization were 4.309 (95% CI: 2.975-6.241, p , Results: Patients who received ICI and SGLT2i had significantly improved OS compared
0.0001) and 2.049 (95% CI: 1.692-2.48, p , 0.0001), respectively. Conclusions: This to those who received ICI alone (HR = 0.62, 95% CI: 0.49-0.79). Among cancer types,
study highlights the clinical impact of ICI-associated AIHA, which is linked to higher significant improvements in survival were observed in patients with renal cell carcinoma
mortality at 30 and 60 days, as well as increased transfusion and hospitalization rates. (RCC, HR = 0.48, 95% CI: 0.27–0.84) and non-small cell lung cancer (NSCLC, HR = 0.60,
Although rare, ICI-associated AIHA is a potentially fatal complication. Clinicians should 95% CI: 0.37–0.96). Among ICI types, ipilimumab + nivolumab (HR = 0.35, 95% CI:
maintain a high level of suspicion for AIHA in patients on ICIs, as early recognition and 0.15–0.79) and pembrolizumab (HR = 0.49, 95% CI: 0.32-0.74) showed a significant
intervention may improve outcomes. Research Sponsor: None. improvement in survival when used with SGLT2i. There was no statistically significant
difference in OS amongst the three SGLT2i types. Conclusions: In this retrospective,
Primary Outcomes HR 95 % Confidence Interval P-Value matched cohort study we observed encouraging improvements in OS in patients with
Mortality within 30-days* 3.16 (1.849,5.402) , 0.0001 T2DM or CHF receiving SGLT2i in addition to ICI across multiple solid tumor types.
Mortality within 60-days* 3.673 (2.324,5.805) , 0.0001 Patients with RCC and NSCLC derived the greatest benefit. Patients treated with either
Transfusion 4.309 (2.975,6.241) , 0.0001 ipilimumab + nivolumab or pembrolizumab had the best responses to therapy. SGLT2i
Hospitalization or 2.049 (1.692,2.48) , 0.0001 may be beneficial as adjunctive therapies in patients with advanced malignancy re-
emergency services
ceiving ICI. However, further prospective studies to validate our observations and
determine potential underlying mechanisms are needed. Research Sponsor: None.

2651 Poster Session 2652 Poster Session

Prognostic value of host genetic variants determining Bifidobacterium Early peripheral Treg expansion after SBRT combined with low-dose radio-
abundance in the lactose metabolism pathway for immunotherapy therapy to predict subsequent immune checkpoint inhibitor responses in
efficacy. First Author: Wenhui Liu, The Second Xiangya Hospital of Central South patients with metastatic lung or gastrointestinal cancers. First Author:
University, Changsha, Hunan, China Byoung Hyuck Kim, Seoul National University College of Medicine, Smg-Snu Boramae
Background: The potential predicative and therapeutic value of Bifidobacterium in immune checkpoint Medical Center, Seoul, South Korea
inhibitors (ICIs) treatment has been widely studied. However, the value of its genetic determinants on the Background: This study aims to explore the potential therapeutic advantages of
prognosis of ICIs treatment remains unclear. Methods: we examined the associations of 11 single nu-
cleotide polymorphisms (SNPs) located at host genes determining Bifidobacterium abundance with the combining stereotactic body radiation therapy (SBRT) and low-dose radiotherapy (LDRT)
outcomes of ICIs treatment in 370 eligible cancer patients. Results: Cox regression analysis revealed that prior to immune checkpoint inhibitor (ICI) treatment for metastatic lung or gastroin-
rs3739020 TT carriers experienced significantly extended OS (P-value = 0.003, adjusted HR = 0.46, 95%CI = testinal cancers, to induce an immune-favoring tumor microenvironment.
0.27-0.77) compared with GG+TG carriers. The LCT haplotype analysis showed that the lactose poor Methods: Patients with metastatic cancer and three or more measurable lesions
metabolizers exhibited significantly poorer OS (P-value = 0.002, adjusted HR = 0.45, 95%CI = 0.27-0.74) than scheduled for ICI therapy were enrolled in this study. Treatment consisted of three SBRT
the extensive or intermediate metabolizers. In polygenic SNP analysis, the high galactose level carriers
exhibited significantly prolonged OS (P-value = 0.007, adjusted HR = 0.32, 95%CI = 0.14-0.73) and doses of 8–10 Gy to the main target lesion and LDRT (2–3 Gy) for other lesions. Patients
progression-free survival (PFS, P-value = 0.001, adjusted HR = 0.45, 95%CI = 0.29-0.71). All the four SNPs without evidence of disease progression within 6 months after the first dose of ICI were
and the LCT metabolic phenotype were not associated with the occurrence of overall immune-related defined as responders, while others were classified as non-responders. Peripheral blood
adverse events (irAEs). Genetically predicted Bifidobacterium abundance was significantly associated with samples obtained before SBRT/LDRT (W0), 1 week after SBRT/LDRT but prior to ICI
an increased abundance of lactose metabolism pathway (P-value = 0.012, Beta coefficient = 0.549). initiation (W1), and 4 weeks after SBRT/LDRT (W4) were analyzed using multi-color flow
Conclusions: SNPs determining Bifidobacterium abundance in the lactose metabolism pathway have
prognostic value for immunotherapy efficacy, and the lactose extensive and intermediate metabolizers
cytometry. This trial has been registered at [Link] (registration number:
exhibited better immunotherapy efficacy. Research Sponsor: National Natural Science Foundation of China; KCT0005879). Results: Among the 13 enrolled patients (lung 9, gastrointestinal 4),
82204534. samples from 4 responders and 7 non-responders were analyzed initially, revealing a
The details of LCT metabolic phenotypes and the associations of LCT metabolic phenotypes with overall survival. median progression-free survival of 22.2 months for responders and 3.1 months for non-
LCT genotypes and their risk allele responders. The fold change in the proportion of regulatory (Foxp3+CD25+) CD4+ T cells
rs3739020G rs56263017C rs55809728A rs3739022A HR P (Tregs) among total CD4+ T cells at W1 compared to W0 (Treg/CD4–FCW1/W0) was lower
Lactose extensive metabolizer TT TT GG GG Refernces
in responders than in non-responders (0.66 vs. 1.22; P = 0.08). Furthermore, the fold
(EM, no risk allele exist) change in the proportion of suppressive Foxp3hiCD45RA+ Tregs among Tregs at W1
Lactose inermediate TG or GG TT GG GG 0.58,95%CI=0.33-0.99 0.049
metabolizer (IM, risk alleles TT TC or CC GG GG compared to W0 ([Link]/Treg–FCW1/W0) was significantly lower in responders than in non-
exist in 1-3 SNPs) TT TT AA or GA GG responders (0.80 vs. 1.18; P = 0.047). This difference was no longer evident after one
TG or GG TT GG AA or GA
TG or GG TC or CC GG GG cycle of ICI, as Treg/CD4–FCW4/W0 (1.15 vs. 0.86; P = 0.46) and [Link]/Treg–FCW4/W0 (1.06
TG or GG TT AA or GA GG vs. 1.33; P = 0.18) were not significantly different between responders and non-
TT TT GG AA or GA
TT TC or CC AA or GA GG responders. Conclusions: We investigated circulating T cell modulation and its po-
TT TC or CC GG AA or GA tential as a biomarker which revealed early expansion of Tregs in peripheral blood after
TG or GG TT AA or GA AA or GA
TG or GG TC or CC AA or GA GG SBRT/LDRT is associated with suboptimal response to ICIs in patients with metastatic
TG or GG TC or CC GG AA or GA cancers. Clinical trial information: KCT0005879. Research Sponsor: None.
TT TT AA or GA AA or GA
TG or GG TC or CC AA or GA AA or GA
Lactose poor metabolizer TG or GG TC or CC AA or GA AA or GA 0.33,95%CI=0.18-0.62 0.001
(PM, risk alleles exist
in 4 SNPs)

The patients were stratified into three LCT metabolic phenotypes (PMs, IMs, EMs) according to the existence of the risk alleles
(rs3739020 G, rs56263017 C, rs55809728 A, and rs3739022 A) in these four LCT/MCM6 [Link] lactose EMs and IMs exhibited
significantly extended OS than PMs.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 175s

2653 Poster Session 2654 Poster Session

The impact of immunotherapy versus chemotherapy on mortality and ad- Gut dysbiosis as a potential guide for immunotherapy (dis)continuation
verse events in cancer patients hospitalized with septic shock. First Author: after 2 years in non-small cell lung cancer: A mono-institutional, multi-omic
Saad Javaid, Charleston Area Medical Center, Charleston, WV assessment. First Author: Lorenzo Belluomini, Section of Innovation Biomedicine -
Background: Traditionally, chemotherapies have been associated with well-characterized tox- Oncology Area, Department of Engineering for Innovation Medicine (DIMI), University of
icity profiles and adverse events. Meanwhile, immune checkpoint inhibitors (ICIs) have revo- Verona, Verona, Italy, Italy
lutionized cancer treatment, but their secondary toxicities, particularly in patients with septic Background: Although most phase II and III clinical trials have set the duration of immune checkpoint blockers (ICB) for
shock, remain underexplored. Our study sought to evaluate the comparative risk of mortality and advanced non-small cell lung cancer (NSCLC) at two years, there remains uncertainty regarding the feasibility and safety
adverse events in patients treated with ICIs versus chemotherapy. Methods: Using the TriNetX of discontinuing treatment after this period. Of note, gut microbial taxonomic profiling prior to starting immunotherapy
shows promise as biomarker for predicting ICB response. Here, we recommend integrating multi-omics approaches over
Research Database, we identified cancer patients ($18 years) diagnosed with septic shock time (24 months -mo-) to inform clinical decision-making and guide personalized treatment strategies. Methods: Pts
between January 1, 2013, and December 16, 2024. Eligible cancers included malignancies of the completing 18 to 24 mo of ICB treatment between July 2016 and January 2023 were identified and enrolled
oral cavity, pharynx, larynx, stomach, kidney, bladder, skin, head, neck, and lung. Patients treated (NCT04567446) at Gustave Roussy. Clinical factors influencing treatment (dis)continuation were assessed at 24 mo.
with ICIs or chemotherapy within 6 weeks before or up to 2 weeks after septic shock diagnosis Multi-omic analyses, including gut-based biomarkers (TOPOSCORE by whole genome sequencing), PET-FDG imaging,
were stratified into two cohorts. Propensity score matching (1:1) adjusted for confounders. and ctDNA, were proposed at this timepoint. Key outcomes, including overall survival (OS) and progression-free survival
(PFS) rates, were analyzed. Results: Among 123 advanced NSCLC pts treated for $18 mo, 35 (28,5%) completed 24 mo,
Kaplan-Meier log-rank tests analyzed outcomes. Primary outcomes were mortality at 30, 60, and with 31 included in the analysis (4 excluded due to PD). Of these, 68% continued ICB, while 32% stopped between 23.5
90 days. Secondary outcomes included irAEs, organ failure, and hospitalization or emergency and 29.7 mo, mainly based on the physician or the patient decision. Clinical characteristics were comparable between the
services within 1 week to 1 month. Results: Through our study we identified 44,052 cancer 2 groups (Table 1). After a median follow-up of 59.1 mo, no significant OS and PFS differences were observed between
patients diagnosed with septic shock. A total of 4,284 patients were on chemotherapy within the pts who discontinued and those who continued (OS p=0.9012, PFS p=0.3715). Among the multi-omic assessments
time of their diagnosis, while 472 patients were on ICIs. ICI use was associated with higher 30-day performed at 24 mo, only gut-based biomarkers appeared to be conditionally associated with PFS24 rates. The pro-
portion of long-responders (progression-free at 24 mo) was higher among those with a favorable gut composition
(HR: 1.27, 95% CI: 1.03-1.57, p = 0.02), and 90-day (HR: 1.21, 95%CI: 1.009-1.45, p = 0.03) compared to those with harmful composition (81% vs 44%, respectively, p=0.0870). Conclusions: Our results suggest
mortality. But only a near significant difference was noted at 60-days (HR: 1.20, 95%CI: 0.99-1.45, that multi-omics approaches may help safely discontinue ICB after two years of treatment. In this context, multi-
p = 0.054). Additionally, the risk of organ failures (HR: 1.28, 95%CI: 1.11-1.48,p , 0.001) and institutional validation and the implementation of a translational multi-omic algorithm, including gut-based biomarkers,
hospitalization or usage of emergency services (HR: 1.35, 95%CI: 1.16-1.56, p , 0.001) was could provide insight into the optimal duration of ICB therapy beyond the predefined 24-mo period. Clinical trial in-
higher in patients treated with ICIs compared to chemotherapy. The frequency of irAEs was formation: NCT04567446. Research Sponsor: None.
slightly higher in patients with ICIs, though these results were only near significant (HR: 1.155, Clinical characteristics of the cohort (n=31).

95%CI: 0.977-1.364, p = 0.0532). Conclusions: Cancer patients treated with ICIs with septic shock Cessation group Pursuit group

exhibited higher short-term mortality, organ failure rates, and hospitalization or emergency Characteristics N = 10 N = 21 p-value*

service usage compared to those receiving chemotherapy. These findings are unexpected, given Gender - no. (%)
Male 5 (50) 12 (57) 0.7366
the perceived safety profile of immunotherapy compared to traditional chemotherapy. Further Female 5 (50) 9 (43)
Age years - median 61 (39-68) 62 (43-77)
research is warranted to better understand these risks and to develop strategies for mitigating (range)
adverse outcomes in this population. Research Sponsor: None. ECOG performance sta-
tus - no. (%)
0-1 9 (90) 15 (71) 0.2044
Log-Rank Test 2 1 (10) 6 (29)
PD-L1 expression - no.
HR 95% CI P-Value x2 (%)
<1% 0 2 (13) 0.5749
‡1%-<50% 3 (30) 2 (13)
Primary Outcomes ‡50% 7 (70) 11 (73)
30-day Mortality 1.272 (1.032,1.569) 0.0231 0.233 unknown - 6
60-day Mortality 1.203 (0.996,1.452) 0.054 3.714 Treatment regimen - no.
(%)
90-day Mortality 1.212 (1.009,1.455) 0.039 4.262 Chemoimmunotherapy 3 (30) 2 (10) 0.1907
Monoimmunotherapy 7 (70) 19 (90)
Line of treatment - no. (%)
Secondary Outcomes First 7 (70) 10 (48) 0.7332
Frequency of irAEs 1.155 (0.977,1.364) 0.0532 3.737 ‡ Second 3 (30) 11 (52)
Incidence of organ failure 1.285 (1.115,1.48) , 0.0001 23.556
*Chi-Square test.
Hospitalization or emergency services 1.349 (1.165,1.562) , 0.0001 22.937

2656 Poster Session 2657 Poster Session

Validation and refinement of Society of Immunotherapy of Cancer (SITC) PhaseX: Patient tumor avatars for evaluating anticancer therapeutics. First
definitions for PD-(L)1 resistance: An analysis of more than 1,300 partic- Author: Kanishka Fernando, National University of Singapore, Singapore, Singapore
ipants from SWOG. First Author: Megan Othus, Fred Hutchinson Cancer Center, Background: Current preclinical tumor platforms, such as in vitro 2D cell cultures and
Seattle, WA organoid models, fail to fully recapitulate the complexity of the tumor microenvironment,
Background: New immuno-oncology (IO) agents are commonly used in patients who previously including critical components like the extracellular matrix and immune interactions.
received PD-(L)1 inhibitors. In order to facilitate clinical trial interpretation and better delineate While humanized patient-derived xenograft models address some of these limitations,
therapeutic contributions of novel IO agents, consensus definitions of PD-(L)1 single-agent and they are costly, low-throughput, artificial, and technically challenging to establish.
combination immunotherapy resistance were published in 2020 (PMC7174063) and 2022
(PMC10016305) by the Society of Immunotherapy of Cancer (SITC); definitions outlined in Table.
PhaseX (Patient-derived hydrogel-assisted eXplants) presents a robust alternative,
Validation of these expert-derived definitions is currently lacking. Herein we analyze two SWOG trials preserving the native TME, including cellular diversity, gene expression, and immune
to evaluate the proposed definitions for the advanced and adjuvant settings. Methods: S1609/DART landscapes, for at least seven days. This study utilizes PhaseX to assess patient-specific
(NCT02834013) was a basket trial for patients with rare cancers treated with ipilimumab (1mg/kg responses to immune checkpoint blockade (ICB), chemotherapeutics, and targeted
intravenously [IV] every 6 weeks) plus nivolumab (240mg IV every 2 weeks). S1404 (NCT02506153) therapies while providing insights into their mechanisms of action. Methods: Fifteen
included an arm where patients with high-risk resectable Stage III melanoma received adjuvant fresh patient-derived tumor explants (PDTEs) from HNSCC patients were embedded in
pembrolizumab (200mg IV every 3 weeks for 1 year). In both trials, overall survival (OS) was measured bioengineered hydrogel and treated ex vivo with pembrolizumab. Supernatants were
from study registration to death from any cause with those last known to be alive censored. OS was
collected at 2 and 4 days post-treatment for immunoassay analysis. Tumor explants
evaluated with Kaplan-Meier and martingale residual plots and Cox regression models. Results: In
S1609 (advanced setting), 733 participants were analyzed: 127 (17%) were not evaluable for primary were either dissociated for high-dimensional flow cytometry or processed into FFPE
resistance due to death or off treatment before 6 weeks. Among the 570 evaluable, 366 met the SITC sections for immunofluorescence analysis at 2 and 5 days. Additionally, ten PDTEs
primary resistance definition and 204 did not. Martingale residuals plots indicated a positive as- representing various cancer types (peritoneal, colorectal, sarcoma, lung and ovarian)
sociation between time to progression and OS with no evidence of a threshold. With a 6-month were used to evaluate dose-dependent responses to commonly used chemothera-
landmark, participants with primary resistance had significantly shorter OS compared to those who did peutics, including cisplatin, doxorubicin, and erlotinib. Six PDTEs were treated with
not: hazard ratio (HR)=2.84, 95% confidence interval (CI) 2.28-3.55, p,0.001. In S1404 (adjuvant plasminogen activator inhibitor-1 (PAI-1), and their post-treatment metabolic activity
setting), 626 participants were analyzed with 12 (2%) meeting the definition of early recurrence/ was assessed using the resazurin cell viability assay. Results: This study highlights the
primary resistance and 138 (22%) meeting the definition of late recurrence/secondary resistance.
Using a 12-month landmark, there was no significant difference in OS between early and late re-
importance of capturing temporal dynamics in ex vivo tumor models to accurately
currences (HR=0.98, 95% CI:0.35-2.70, p=0.25), however late recurrences were associated with predict pembrolizumab responses, achieving 100% sensitivity and specificity in HNSCC
significantly shorter OS than no recurrence (HR=7.69, 95% CI:2.71-20.1,p,0.001). Conclusions: In patients. Increased IFN-g secretion and upregulation of chemokines (CXCL9, CXCL10,
the advanced cancer cohort, the SITC definitions were validated. In the adjuvant cohort, early re- CXCL11) distinguished responders, alongside elevated cytotoxicity markers (perforin,
currences were uncommon and there was no significant difference in OS between early and late granulysin, sFasL, sFas) contributing to cancer cell death. Responders exhibited reduced
recurrences suggesting a 12-month cutoff may be more appropriate than 12 weeks. Additional terminally exhausted CD8+ T cells (PD-1+TIM3+), allowing reinvigoration of functional
analyses in other patient cohorts are needed to further understand if the SITC definitions for advanced CD8+ T cells, while non-responders showed elevated Tox+CD38+ levels, indicating re-
cancers validate more broadly and if data-drive refinements are needed for the adjuvant setting.
sistance to PD-1 blockade. Spatial analysis revealed greater T cell infiltration in re-
Research Sponsor: US National Cancer Institute; Bristol-Myers Squibb Company.
sponders, facilitating tumor-cell interactions. Additionally, the PhaseX platform
Advanced demonstrated its utility in evaluating dose-dependent responses across multiple tumor
Primary Treatment . 6 weeks; no response or , 6 months types (sarcoma, colorectal, lung, cervical) and identified patient-specific responses to
Secondary Treatment . 6 months; response/stable . 6 months PAI-1 inhibition. Conclusions: The PhaseX platform accurately predicts patient-specific
Adjuvant responses for ICB, chemotherapeutics and targeted therapy across multiple tumor
Early/primary , 12 weeks last dose
Late/secondary 12 weeks types. These findings establish PhaseX as a valuable tumor platform to evaluate an-
ticancer therapeutics. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
176s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2658 Poster Session 2659 Poster Session

Examining the relationship between multi-agent immunosuppressive ther- Incidence and outcomes of immune checkpoint inhibitor (ICI) rechallenge
apy for immune-related adverse events (irAE) and infectious complications. after ICI pneumonitis: A single-center retrospective study. First Author:
First Author: Tristan Lee Lim, Mass General Cancer Center, Massachusetts General MacKenzie Adams, Brown University Health, Providence, RI
Hospital, Boston, MA Background: Limited evidence is available on the safety and efficacy of immune
Background: Treatment of severe irAEs with multiple immunosuppressive therapies (ISTs) checkpoint inhibitor (ICI) rechallenge following an immune-related adverse event (irAE).
decreases the morbidity and mortality of these conditions. Nevertheless, the rates of and risk Pneumonitis is a potentially life-threatening irAE, and minimal data is available with
factors for infectious complications in this population are not known. Methods: We regards to outcomes after rechallenge. In this single-center retrospective analysis we
conducted a retrospective study of patients (pts) who received an immune checkpoint in- analyzed patients who developed ICI pneumonitis and were later rechallenged with an ICI
hibitor (ICI) and experienced $1 irAE requiring treatment with corticosteroids along with at to further investigate incidence patterns and outcomes. Methods: We conducted a
least two lines of steroid-sparing ISTs administered either concurrently or within 90 days of manual chart review on a cohort of 69 patients with ICI pneumonitis at our institution
each other. We annotated all infections from ICI start until 90 days after IST. Opportunistic from 2015 to 2024. Rechallenged cases were defined as any patient who developed ICI
infections (OIs) were defined as herpesvirus (CMV, EBV, VZV, and HSV) and invasive fungal pneumonitis and were later trialed on the same or another ICI. Patient records were
infections. Infection density was reported as number of infections per 1000 patient-days and
reviewed to identify demographics, clinical features, treatment, and outcomes.
graded as mild (not requiring treatment), moderate (requiring oral treatment), severe (re-
Results: Of 69 patients with ICI pneumonitis, we identified 19 that were rechallenged
quiring hospitalization or parenteral treatment), life threatening, or fatal. Risk factors were
identified using univariable and multivariable Cox regression analysis adjusting for age at ICI
with an ICI. Of these, 10 were women and 9 were men. The average age was 64 years
initiation, sex, ICI regimen, ISTs, and steroid dose. Results: 175 pts (52% male, mean age: 66) (range: 42-82). Most patients had lung cancer (42%, 8/19) followed by melanoma (32%,
with 238 irAEs and 417 ISTs were analyzed with a median follow-up of 367 days. The as- 6/19). The treatment intent was palliative for 74% of patients (14/19). The most common
sociated ICI regimens included aPD-1 (n = 81, 46%) and aPD-1/aCTLA-4 (n = 67, 38%). The therapy was nivolumab (n = 15), however, pembrolizumab, atezolizumab, and durva-
most common irAEs were colitis (n = 67, 38%), hepatitis (n = 42, 24%), and myocarditis (n = 26, lumab were also used. The initial grade of ICI pneumonitis was as follows: grade 1 (n = 2),
15%). The most frequently used ISTs were mycophenolate mofetil (n = 92, 53%), infliximab (n grade 2 (n = 13), grade 3 (n = 3), and grade 4 (n = 1). Many of these patients were treated
= 77, 44%), and vedolizumab (n = 54, 31%). 103 pts (59%) developed 223 infections (median 2/ with steroids, with the average time on steroids being 136 days (range: 0-488). Ad-
pt, range: 1-8). 93 pts had 187 non-OIs. Of the 87 pts (50%) who had OI testing, 29 (33%) had ditionally, the only grade 4 patient was also treated with infliximab due to steroid-
36 OIs, most commonly EBV DNAemia (n = 14, 16%) and CMV reactivation (n = 12, 14%). 6 pts refractory pneumonitis. After recovery from their pneumonitis, all patients, except for
had .1 OI, including 1 pt with CMV, EBV, and HSV. OI density significantly increased after two, were rechallenged with the same ICI that they had originally been treated with. The
starting ISTs for irAEs, but non-OI density was unchanged (Table 1). aCD20 use was as- average time to rechallenge was 149 days (range: 12-714). Ultimately, 6/19 (32%)
sociated with increased non-OI risk (HR: 10.58, 95% CI: 3.64-30.72, p , 0.001), while there patients had recurrent ICI pneumonitis after rechallenge, but all six eventually recovered.
was a trend towards increased non-OI risk with a max prednisone dose .100mg (HR: 1.74, On grading ICI toxicity after recurrence, 60% (4/6) remained grade 2, one was down-
95% CI: 0.96-3.13, p = 0.07). In contrast, a max prednisone dose .100mg was associated with graded from grade 3 to 2, and one escalated from grade 2 to 3. Eleven percent (2/19) of
increased OI risk (HR: 2.89, 95% CI: 1.21-6.90, p = 0.017). 58 pts (33%) had severe or life- patients developed an irAE other than pneumonitis (e.g. colitis, arthritis) after
threatening infections, of whom 16 (9%) had OIs. 8 pts (5%) had fatal infections in this rechallenge. Finally, none of the 19 patients died from complications associated with ICI
population. Conclusions: Use of multiple ISTs for severe irAEs is associated with increased therapy. Conclusions: The recurrence rate of ICI pneumonitis after ICI rechallenge was
rates of opportunistic infections as well as a 5% infection-related mortality rate. Patients 32%. At initial presentation, most of these patients had lower grade (grade 1: 2/19, grade
requiring multiple lines of ISTs must be closely monitored for infectious complications, and 2: 13/19) ICI pneumonitis and most cases of recurrent pneumonitis remained at their
prophylaxis should be considered when appropriate. Research Sponsor: None.
initial grade 2. These results indicate that resuming ICI therapy could be considered in
Mean OI and non-OI density per 1000 patient-days while on ICI alone vs ISTs. select patients with mild to moderate pneumonitis. Further research is needed to in-
ICI ISTs p-value vestigate immunotherapy rechallenge as it remains a nuanced decision that involves
OI 0.03 1.70 0.002 assessing the potential benefits of continued tumor control against the risk of a life-
Non-OI 10.70 8.98 ns threatening toxicities. Research Sponsor: None.

2660 Poster Session 2661 Poster Session

Clinical outcomes of patients with or without DNA repair pathway alterations MicroRNA-based signatures of early and late immune-related adverse
by treatment type: The MD Anderson Cancer Center IMPACT 2 study. First events to anti-PD1 treatment. First Author: Joanne B. Weidhaas, University of
Author: Jacopo Venturini, Investigational Cancer Therapeutics, The University of Texas California, Los Angeles, Los Angeles, CA
MD Anderson Cancer Center, Houston, TX Background: Immune checkpoint inhibitors (ICIs) have transformed cancer treatment but are
Background: DNA repair deficiency is common among tumors, and emerging data suggest that genomic associated with toxicity in the form of immune-related adverse events (irAEs). We previously
instability is associated with response to immuno-oncology (IO) therapies (PMID 28630051). We evaluated reported a genetic signature predicting anti-PD1 irAEs in a retrospective analysis of a heavily
patients with advanced metastatic cancer across tumor types, who were treated on the IMPACT2 study pretreated melanoma cohort, which validated in a pan-cancer cohort. Here we investigate the
(NCT02152254) and analyzed their clinical outcomes by treatment type (anti-DNA damage repair [DDR] agents,
applicability of that signature in a prospectively collected cohort of GU, breast, and NSCLC
IO, or other [non-IO, non-anti-DDR]). Methods: Patients had tumor biopsies followed by molecular profiling in a
CLIA-certified lab. All cases were discussed at Molecular Tumor Board meetings. Patients were treated on cancer patients. We evaluated clinical and genetic differences between the original training set
early-phase clinical trials. Progression-free survival (PFS), overall response rate (ORR), and overall survival (OS) and this cohort and leveraged the expanded dataset to develop novel models for timing-
were compared by therapy type in patients with or without DDR/MMR mutations (DDR+). Results: Of 829 specific anti-PD1 toxicity. Methods: We analyzed clinical and genetic differences in the
enrolled patients, 510 had molecular profiling and received anticancer therapy: 85 were DDR+ (PS 1, 75%; med. melanoma training cohort (n=58) and the new cohort of patients, all treated with single agent
age, 59 yrs; males 56.5%; med prior therapies 4; PDL1+ 44.9%; TMB-H 18%,) and 425 DDR- (PS 1, 88%; med. anti-PD1/PDL1 therapy (n=137). Clinical and genetic differences were assessed using Fisher’s
age, 60 yrs; men 47%, med. No. prior therapies 3; PDL1+ 46%; TMB-H 6.5%). Results are shown in the Table. In exact test for categorical variables: pre-treatment, concurrent radiation, and SNP genotype
patients with DDR mutations, IO was associated with a higher ORR compared with “other” treatments
across 165 loci. Kruskal-Wallis tests were used for age, toxicity timing, and severity of toxicity.
(p=0.044); and with longer PFS compared with anti-DDR therapies (p = 0.044); no difference was noted in OS by
treatment type. In the DDR-neg group, IO was associated with longer PFS (p = 0.017), and longer OS (p = Predictive genetic models were constructed using elastic net, random forest, and boosted tree
0.0004) compared with anti-DDR therapy; OS was longer in the IO group compared with “other” treatments (p = algorithms and evaluated using leave-one-out cross-validation (LOOCV) metrics. Outcomes
0.029) and in the “other” treatments group compared with anti-DDR agents (p=0.006). Conclusions: Our data included cycle-specific toxicity (early #5 cycles, late $15 cycles). SNPs were pre-filtered for
indicate the complexity of assessing outcomes in patients with various tumor types and without DDR mu- inclusion in genetic models using Fisher or Jonckheere-Terpstra p-values (,0.2) for relevance
tations. Further work is needed to develop predictive biomarkers for IO and anti-DDR agents. Clinical trial to outcomes. Results: The training and current cohort were different in their toxicity timing,
information: NCT02152254. Research Sponsor: Steven McKenzie’s Endowment for Dr Tsimberidou’s per- with earlier toxicity onset in the new cohort (median 5 cycles vs. 20 cycles in melanoma,
sonalized medicined program; Katherine Russell Dixie Distinguished Endowed Professorship for Dr Tsim-
beridou; Jamie’s Hope for Dr Tsimberidou’s personalized medicined program; NIH National Cancer Institute
Kruskal p = 0.00018). Genetic analysis identified 12 significantly different SNP genotypes
award number P30 CA016672 (to The University of Texas MD Anderson Cancer Center).; Tempus, Inc. for the (Fisher p , 0.05) between cohorts, including mir146A rs2910164 (Fisher p = 0.0005). This SNP
IMPACT 2 study; Foundation Medicine, Inc for the IMPACT 2 study. was associated with early toxicity overall and within data subsets. Refining our model to
account for cycle-specific toxicity events significantly enhanced performance. For late toxicity
Rx DDR- Anti-DDR Anti-DDR DDR- Anti-DDR Anti-DDR
type pos N Outcome vs IO IO vs Other vs Other neg N Outcome vs IO IO vs Other vs Other ($15 cycles), the refined model achieved a LOOCV AUC of 0.793 (genetics + clinical). A newly
Overall re- IO 34 6 (25%) OR* 0.78; OR* 20.73; OR* 16.09; 116 9 (11%) OR* 0.34; OR* 2; OR* 0.67; developed early toxicity model (#5 cycles) using genetics alone demonstrated robust pre-
sponse (%)
Anti- 18 1 (16.7%)
p=0.81 p=0.044 p=0.1
52 0 (0%)
p=0.46 p=0.13 p=0.79
dictive accuracy with an AUC of 0.753. Conclusions: These findings emphasize the impor-
DDR
Other 33 0 (0%) 257 12 (5.9%)
tance of defining clinical and genetic diversity in refining predictive models for anti-PD1/PDL1
PFS, med. IO 34 4.26 (2.3, HR 2.01 HR 0.91 HR 1.82 116 5.42(3.02,6.61) HR 2.02 HR 0.85 HR 1.71 outcomes. The development of an early toxicity model offers significant clinical utility. Next
(95% CI) 7.4) (1.02,3.96) (0.56,1.47) (0.95,3.49) (1.14,3.61) (0.68,1.06) (0.98,2.99)
(p=0.044) p=0.69 p=0.07 p= 0.017 p= 0.15 p=0.058 steps will be to use time to event (toxicity) versus cycle number. This study provides a
Anti- 18 2.58 (1.68, 52 1.64 (1.45, NA)
DDR NA) foundation for the application of personalized genetic tools to predict the safety of ICIs for
Other 33 4.54 (3.06,
6.54)
257 3.98 (3.45,
4.57)
currently treated patient cohorts. Research Sponsor: U.S. National Institutes of Health;
OS, med. IO 34 14.37 HR 1.68 HR 0.73 HR 1.23 116 12.46 (8.75, HR 2.98 HR 0.76 HR 2.27 R01CA238998.
(95% CI) (10.22, NA) (0.82,3.42) (0.43,1.23) (0.62,2.42) 20.78) (1.62,5.48) (0.6,0.97) (1.27,4.07)
p=0.16 p=0.24 p=0.56 p=0.0004 p=0.029 p=0.006
Anti- 18 8.98 (3.48, 52 4.31 (2.66, NA) New LOOCV performance metrics in expanded PD1 data (n=234).
DDR NA)
Other 33 10.82 (7, 257 9.4 (8.28, Outcome Covariates Sensitivity Specificity PPV NPV F1 AUC
16.83) 11.21)

*OR, Odds Ratio, p values, unadjusted for multiple comparisons.

Late Toxicity (>= 15 cycles) SNPs + Clin 0.692 0.894 0.450 0.959 0.545 0.793
Early Toxicity (<=5 cycles) SNPs Only 0.600 0.907 0.486 0.939 0.537 0.753

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 177s

2662 Poster Session 2663 Poster Session

Overall survival according to timing of immune checkpoint inhibitors ad- The impact of lymphocyte count dynamics on the predictive value of tumor
ministration in patients with advanced cancer: Results from a large single- mutational burden (TMB) for immune checkpoint inhibitors (ICI) outcomes
centre cohort analysis. First Author: Tommaso Bosetti, The Christie NHS Foundation in patients (pts) with cancer. First Author: Mustafa Jamal Saleh, Dana-Farber Cancer
Trust, Manchester, UK, United Kingdom Institute, Boston, MA
Background: Immune checkpoint inhibitors (ICIs) have revolutionised cancer treatment Background: TMB has become a reliable biomarker for ICI response in several cancer types,
but are only effective in a subset of patients. Evidence of a circadian dependence of the leading to the pan-cancer FDA approval of the PD-1 inhibitor pembrolizumab in tumors
immune system has led to retrospective studies which suggested that the time of day of with a TMB ³10 mut/Mb. Lymphocyte count dynamics (lymphocyte stability LS) have been
ICIs infusion influences treatment response, with better outcomes for early treatment reported to be associated with both increased immune-related adverse events and improved
times. Previous studies are limited by small sample size and methodological bias. We overall survival (OS) on ICI. We aimed to investigate the role of LS as a risk stratification tool,
performed a retrospective study in patients with advanced solid tumours treated with in combination with TMB, in patients with cancer treated with ICIs. Methods: We identified
ICIs at a large tertiary cancer centre. Methods: Patients who received regimens 1215 pts from the Dana-Farber Cancer Institute, who had received ICI between 2015 and
comprising ICIs for advanced/metastatic disease between January 2018 and December 2024. The change in relative blood lymphocyte counts was calculated between pre- (up to
2023 were grouped according to whether they received $50% (“late group”) or , 50% 30 days) and post-treatment (between 21 and 49 days). Lymphocyte counts were con-
(“early group”) of cycles after the median time of all treatments. The primary endpoint sidered stable (LS $80%) if the drop in lymphocyte count did not exceed 20% after ICI-
was overall survival (OS). Hazard ratios (HRs) for OS after multivariable adjustments for exposure. TMB was assessed from targeted panel sequencing and categorized into high and
low, based on a cutoff of 10 mut/Mb. Overall survival was assessed using a multivariable Cox
age, sex, Eastern Cooperative Oncology Group (ECOG) performance status, cancer type
regression model, adjusting for several baseline characteristics. Results: The most
and treatment were estimated using Cox models with and without time-dependent
prevalent cancer types were non-small cell lung cancer (n = 190), breast carcinoma (n =
variables that allowed for group (early vs. late) assignment to change after baseline.
160), glioma (n = 104), and melanoma (n = 100). Median TMB was 6.8 (IQR: 0–265.4). In
Results: 2631 patients with lung cancer (45%), melanoma (23%), renal carcinoma total, 846 (69.6%) patients had LS (drop , 20%), while 369 (30.4%) did not. Higher TMB
(18%), head and neck (9%) and urothelial cancer (5%) were included. The median age ($10) and LS were both independently associated with better survival on ICI after adjusting
was 68.2 years and 1602 (61%) were men. The median follow up was 38 months. The for age, sex, tumor purity, line of therapy, ICI type and cancer type (HR 0.58, CI: 0.50–0.67;
median infusion time was 12:49h. Median OS was 13.1 (95% confidence interval [CI] p , 0.001) and 0.75 (CI: 0.63–0.90; p = 0.002) respectively. Overall, patients with LS and
11.8-14.4) vs. 21.4 (95% CI 19.8-24.5) months for late and early group. The late group high TMB demonstrated the longest median OS while the combination of
was associated with shorter OS compared to the early group on both the standard unstable lymphocyte counts and low TMB was associated with the worst survival (Table).
multivariate analysis (HR 1.48, 95% CI 1.33-1.63) and the time dependent Cox model (HR Conclusions: LS provides additional value, irrespective of TMB, in predicting response to
1.30, 95% CI 1.16-1.44). The association was significant for most regimens with ICIs ICI, suggesting distinct underlying immunological pathways. These findings emphasize the
alone (n = 1886) and for ICIs plus tyrosine kinase inhibitors (n = 163). There was no need for further research to validate LS and explore its integration into clinical decision-
difference for regimens comprising chemotherapy (n = 582). A sensitivity analysis based making in ICI-treated pts. Research Sponsor: None.
on exposure at 3 months showed no difference between the groups with the standard Results from multivariable Cox regression.
model (HR 1.09, 95% CI 0.98-1.20) and a higher risk of death for the late group with the
Groups Median OS (mo) N HR 95% CI P-value
time dependent model (HR 1.14, 95% CI 1.02-1.26). Conclusions: This study suggests a
benefit in OS with early administration of ICIs with a large sample size and a time- Stable lymphocytes (LS ‡80%) & 34.27 245 Ref. Ref. Ref.
High TMB (Cutoff = 10)
dependent Cox model which reduces the risk of immortal time bias. These findings could Stable (LS ‡80%) & Low TMB 18.99 601 1.3 1.04 - 1.63 0.019
have a major clinical impact after small changes in the way services are provided. Unstable (LS <80%) & High TMB 17.05 96 1.66 1.22 - 2.26 0.001
Exclusive morning administration for all doses in the first 3 months could be a feasible Unstable (LS <80%) & Low TMB 9.59 273 2.28 1.79 - 2.90 ,0.001
approach to minimise complexity of treatment slot allocation. Research Sponsor: None. Model is adjusted for age, sexe, tumor purity, lines of treatment, PD1/PDL1 therapy, and CTLA-4 therapy.
Abbreviations: TMB: Tumor mutational burden; LS: Lymphocyte Stability; OS: Overall Survival; mo: Months.

2664 Poster Session 2665 Poster Session

Association of intestinal exfoliome and Prevotellaceae with toxicity and A Bayesian population-based framework for detecting hyperprogressive
clinical outcome during immune-checkpoint blockade. First Author: Giacomo disease on cancer immunotherapies. First Author: Madison Stoddard, Fractal
Vitali, MetaGenoPolis, INRAE, Paris-Saclay University, Jouy-En-Josas, France Therapeutics, Lexington, MA
Background: Immune-related adverse events (irAEs) are autoimmune side effects Background: Hyperprogressive Disease (HPD), defined as an unexpected treatment-
related to ICI, varying in severity, onset and organ involvement that may be hard to induced rapid increase in tumor growth rate relative to the tumor burden pretreatment
differentiate from non-irAEs. Some reports have demonstrated that gut microbiota (GM) growth rate, has been reported in 9% of patients receiving immune checkpoint inhibitor
plays a role in modulating the risk of AEs. In this study, we combine gut mamalian and (ICI) therapies (Champiat et al, Clin Cancer Res (2017)). The definition of HPD used in
microbial metagenomics sequencing (MGS) to explore the influence of GM on treatment that work was an increase in the on-treatment “Tumor Growth Rate” (TGR) by a factor of
response and risk of AEs. Methods: NCT04567446 allowed fecal MGS at baseline and 2 or greater over the pre-treatment TGR as calculated from three consecutive CT scans:
longitudinally in patients (pts) with advanced non-small cell lung cancer (), renal cell pre-baseline, baseline and on-treatment. This method of directly calculating TGR from
carcinoma and bladder cancer treated with ICI alone (ICI cohort, n = 542pts) or in slopes between successive tumor measurements, however, does not consider uncer-
combination with chemotherapy (CT+ICI cohort, n = 122 pts) in France and Canada. Pts tainty in the CT-assessed sum of diameters (SoD) of target lesions (~8% per Zhao et al,
who experienced severe ($ grade 3) irAEs after ICI+/-CT were compared to those who Radiology (2009)), nor prior knowledge of the distribution of responses under ICI
did not, using microbial MGS parameters (Shannon diversity, TOPOSCORE, PCoA and treatment. We sought to develop and test a Bayesian approach to HPD detection and
LEfSe). Multivariate Cox regression models to analyze factors influencing overall compare its receiver operating characteristics (ROC) to the published TGR ratio
survival (OS) included microbiota composition and the host exfoliome (i.e., mammalian threshold of 2-fold as well as the TGR ratio considered as a continuous classifier of HPD.
eukaryotic DNA read counts within stools). Results: Pts with severe irAEs (10%) Methods: We represented the prior distribution of exponential TGR (eTGR) pre-
presented a less diverse microbiome, a lower TOPOSCORE and a distinct microbial treatment based on population modeling of historical data of untreated NSCLC tu-
community compared to those without severe irAEs, showing an overabundance of mor dynamics. We then calibrated the on-ICI-treatment prior distribution based on
several members of the Prevotellaceae family. Interestingly, CT+ICI pts who experienced reported rates of HPD and tumor regression. Next, we developed a Bayesian parameter
severe irAEs had the most dysbiotic microbiome, characterized by a lower alpha- estimation system to take three successive SoD assessments to calculate a patient’s
diversity and TOPOSCORE, dominated by oral taxa (Ligilactobacillus salivarius) and posterior probability of being in a state of HPD, attenuated tumor growth (ATG) or tumor
tolerogenic Hungatella spp. and Enterocloster spp. Among pts screened for exfoliation, regression (REG). We then simulated a cohort of 1000 virtual patients (VPs) with known
44% presented hstool mammalian DNA than healthy subjects, showed a GM enriched state and tested the ability of the Bayesian method, the TGR ratio method, and the TGR
with pathobionts, including the Enterocloster genus, and exhibited worse OS (HR 1.212, p ratio . 2 method to correctly classify each VP. We additionally developed a user-friendly
= 0.0135) in univariate and multivariate analyses (HR 1.18, p = 0.049). Although there web-based prototype tool ([Link] to solicit feedback
was no significant correlation between toxicity and exfoliation, either in CT+ICI or ICI from a potential future user community. Results: ROC analysis estimated the area
alone, pts enriched with Prevotellaceae family appeared to exhibit the highest levels of under ROC curve (AUROC) to be 0.94 for the Bayesian method, a significant improvement
exfoliation. Conclusions: Host-microbial interactions influence immunity and therefore in classification accuracy over the TGR ratio method (AUROC = 0.77) as well as better
ICI prognosis and toxicity. We found Prevotellaceae members as potential biomarkers maximal sensitivity (Se) and specificity (Sp) than the TGR ratio . 2 method (Se = 80% &
for ICI-related toxicity. The host exfoliome is an interesting parameter that may reflect Sp = 90% vs Se = 90% & Sp = 40%). Conclusions: We have developed a Bayesian
gut fitness, requiring further investigation. Clinical trial information: NCT04567446. population-based methodology and prototyped a web-based tool which under simulated
Research Sponsor: RHU IMMUNOLIFE, RHU LUMIERE. conditions consistently outperformed previous methods for detecting HPD. We believe
that this approach, trained on a larger patient-level dataset, can potentially support and
improve clinical management and decision-making for cancer patients taking immune
checkpoint inhibitor therapies. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
178s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

2666 Poster Session 2667 Poster Session

Safety outcomes of intravenous immunoglobulin (IVIG) in treatment of Vedolizumab or infliximab: Treatment option in immune checkpoint
steroid-refractory immune-checkpoint inhibitor pneumonitis. First Author: inhibitor–induced colitis. First Author: Shreya Shambhavi, RWJBH Rutgers
Mary Metkus, Department of Medicine, Johns Hopkins University, Baltimore, MD Health Community Medical Center, Toms River, NJ
Background: Pneumonitis is a potentially severe adverse event of immune checkpoint in- Background: ICI use is linked to severe gastrointestinal (GI) immune-related adverse
hibitors (ICIs). While the first line treatment for ICI pneumonitis is corticosteroids, there are events (irAEs), which affect morbidity and mortality and often require treatment pauses.
subsets of patients who either fail to respond, deemed steroid-refractory, or who cannot be Among these, immune-mediated colitis (IMC)—primarily associated with CTLA-4
tapered off steroids, deemed steroid-dependent. In these patients, there is no clear consensus therapy—occurs in 5.7% to 39.1% of patients receiving CTLA-4 inhibitors and 0.7%
on the best approach to treatment. IVIG has emerged as a potentially efficacious treatment for to 31.6% of those receiving PD-1/PD-L1 inhibitors; combination therapy can raise this
its immunomodulatory effects. It is also less immunosuppressive than other potential incidence to 40.4%. IMC symptoms range from mild diarrhea to severe colitis, typically
therapies, which is beneficial in a population which respiratory infection could lead to rapid requiring urgent intervention within six to eight weeks of immunotherapy to prevent
clinical decline. However, there are concerns about treatment safety, including volume
complications such as colonic perforation or sepsis. Corticosteroids are the usual first-
overload, thrombosis, and renal injury, which can be devastating in a patient population with
pre-existing respiratory compromise. In this study, we aimed to assess the safety of IVIG
line treatment, with TNF-alpha inhibitors (e.g., infliximab) considered when patients do
treatment in patients with steroid-refractory ICI pneumonitis. Methods: A retrospective review not improve after three to seven days. Vedolizumab, a gut-selective a4b7 integrin
was conducted of patients with steroid-refractory ICI pneumonitis treated with IVIG at Johns antagonist that targets gastrointestinal-homing T-lymphocytes, offers an alternative
Hopkins Hospital from 2018 to 2024. Patient characteristics, treatment features, disease approach. Both infliximab and vedolizumab—referred to as Selective Immunosup-
severity, clinical outcomes, and development of adverse events post-IVIG including renal pressive Therapies (SITs)—have shown promise, though their distinct mechanisms have
injury, volume overload, or thrombosis were evaluated. Results: A total of 31 patients were led to a lack of standardized protocols and reliance on provider discretion. This study
selected with steroid refractory pneumonitis treated with IVIG. Mean age at diagnosis was 68 compares infliximab, vedolizumab, and combined SITs (infliximab plus vedolizumab) in
(56-80). The most common primary tumor site was lung (n = 23) and majority of patients had managing IMC, focusing on remission rates, recurrence, and improved steroid tapering
stage IV malignancy (n = 19). 81% (25/31) had grade 3 or 4 pneumonitis at the time of IVIG success. Methods: A systematic search was conducted across the PubMed database.
therapy. In-hospital death occurred in 41% (13/31). Majority of in-hospital death was due to The Meta-Analysis was conducted using R version 4.4.1 to calculate odds ratios (ORs)
respiratory decompensation from pneumonitis and no deaths were directly related to com- and 95% confidence intervals (CIs). Results: A total of eight studies were included in the
plications of IVIG. Regarding adverse events, 9.68% (3/31) of patients developed a venous final analysis. In patients with immune checkpoint inhibitor–induced colitis, vedoli-
thromboembolism related to IVIG. 3.23% (1/31) developed acute kidney injury and 12.90% (4/ zumab was associated with higher rates of colitis recurrence (OR = 0.32, 95% CI =
31) developed hypervolemia in response to IVIG. Conclusions: IVIG was not associated with a 0.19–0.53) compared to infliximab. Patients receiving vedolizumab also had lower
high rate of toxicity when used in the treatment of steroid refractory ICI-pneumonitis. The overall corticosteroid usage (mean difference in days: -18.29, 95% CI = -21.88 to -14.71)
increased in-hospital mortality indicates the severity of illness in this patient population.
compared to infliximab recipients. There was no significant difference in remission rates
Despite this, IVIG-related adverse events were low overall. Although a small cohort, these
results suggest a favorable safety profile and support further study to evaluate efficacy of IVIG
between vedolizumab and infliximab monotherapy; however, higher remission was noted
as compared with other immunomodulatory agents. Research Sponsor: None. with combination therapy (vedolizumab plus infliximab) (OR = 0.40, 95% CI = 0.19–0.84)
compared to infliximab monotherapy. Conclusions: Vedolizumab was associated with a
IVIG treatment related adverse events. higher recurrence rate of colitis but resulted in significantly lower corticosteroid usage
VTE related to IVIG* (%; n) 9.68 (3) compared with infliximab. Although remission rates were similar for both mono-
Acute kidney injury (%; n) 3.23 (1)
Volume overload (%; n) 12.90 (4) therapies, combination therapy (vedolizumab plus infliximab) demonstrated higher
remission rates than infliximab alone. Research Sponsor: None.
IVIG: intravenous immunoglobulin, VTE: venous thromboembolism.
*2 of the patients died.

TPS2668 Poster Session TPS2669 Poster Session

A phase 1, first-in-human study of DS-2243, an HLA-A*02/NY-ESO– Phase 2 expansions of OR502, an antibody targeting leukocyte
directed bispecific T-cell engager, in patients with advanced solid immunoglobulin-like receptor B2 (LILRB2) 6 cemiplimab in patients with
tumors. First Author: Sandra P. D’Angelo, Memorial Sloan Kettering Cancer Center, advanced solid tumors. First Author: Mohamad Adham Salkeni, Virginia Cancer
New York, NY Specialists, Fairfax, VA
Background: NY-ESO-1 and LAGE-1 are homologous proteins commonly expressed in Background: LILRB2 is an inhibitory receptor expressed on myeloid cells, including tumor-
various malignancies but not in normal tissues other than the testis and placenta. Tumor associated macrophages, which binds to HLA-class I proteins and is associated with poor
types showing prevalent NY-ESO-1 and/or LAGE-1 expression include synovial sarcoma outcomes in multiple cancers. OR502 is a humanized immunoglobulin G1 antibody that
(SS), myxoid/round cell liposarcoma (MRCLS), non-small cell lung cancer (NSCLC), and blocks LILRB2 binding to HLA-class I proteins. Preclinically, OR502 has demonstrated best-
urothelial carcinoma (UC). Both NY-ESO-1 and LAGE-1 undergo intracellular proteolytic in-class reversal and prevention of myeloid cell-mediated immune suppression and res-
processing to generate the same highly immunogenic 9-mer NY-ESO peptide toration of T cell functions. Using OR502 to tackle immunosuppression and improve T cell-
(SLLMWITQC), which is presented on the cell surface in association with HLA-A*02 mediated responses in the tumor microenvironment (TME) is a rational for combination with
major histocompatibility complex molecules. DS-2243 is a bispecific antibody and T-cell checkpoint inhibitors. Methods: This is an ongoing, first-in-human, Phase 1-2 study of
engager with an effectorless Fc region. It is designed to target HLA-A*02/NY-ESO OR502 6 cemiplimab in patients with advanced solid tumors (NCT06090266). The primary
peptide complexes on tumor cells and specific molecules on T-cells, redirecting T- objectives are to evaluate the safety/tolerability and identify a dose for further clinical
cell–mediated cytotoxicity toward the tumor. Methods: DS2243-054(NCT06644755) development. Secondary objectives include assessment of pharmacokinetics (PK), im-
is a Phase 1, first-in-human, open-label, multicenter, 2-part, dose-escalation and -ex- munogenicity and anti-tumor activity. We are also assessing the effects of OR502 on the
TME and associations between response and pharmacodynamic (PD) markers. Dose es-
pansion trial of DS-2243. Patients must be $18 years of age and have HLA-A*02-
calation enrolled 39 patients at OR502 doses of 100–1600 mg, once every 3 weeks (Q3W) 6
positive advanced or metastatic SS, MRCLS, squamous or adenocarcinoma NSCLC, or
standard dose cemiplimab (350 mg), using a modified toxicity probability interval-2 design.
UC, and be unable to tolerate standard treatments, or have relapsed disease after or be As dose escalation completed, it became clear that to satisfy the FDA’s Project Optimus,
refractory to such treatment. Patients with NSCLC or UC in dose escalation and all adaptations were needed to provide dose-response proof and identify the minimal effective
patients in dose expansion must have NY-ESO protein expression confirmed in tumor dose before proceeding with development. The protocol’s adaptive elements, in conjunction
tissue by immunohistochemistry in a central laboratory. Further inclusion criteria in- with Safety Committee oversight, enabled modifications without amendment. Prior to dose-
clude the presence of $1 measurable lesion per Response Evaluation Criteria in Solid response optimization, we adapted the design in order to explore the efficacy signals from
Tumours, version 1.1 (RECIST 1.1) and Eastern Cooperative Oncology Group perfor- phase 1, specifically in patients with melanoma and NSCLC. Based on efficacy signals and
mance status of 0 or 1. The primary objective of dose escalation is to evaluate the safety excellent safety, PK and PD results, we selected OR502 800 mg Q3W for both expansion
and tolerability of DS-2243 and determine the maximum tolerated dose and/or rec- cohorts. Two new mini-expansion cohorts are now actively recruiting 10–20 patients each:
ommended dose for expansion (RDE). The dose-expansion part includes 4 cohorts monotherapy in patients with cutaneous melanoma and combination in patients with
defined by tumor type—SS/MRCLS, squamous NSCLC, adenocarcinoma NSCLC, and NSCLC. The sample size was chosen pragmatically, to exclude a response rate of ~ 10%,
UC—in which patients receive DS-2243 at the RDE. The primary objectives of dose with a target of ~ 35%. If , 2 responses are seen in the first 10 patients, the cohort will be
expansion are to evaluate safety and determine the objective response rate (ORR) discontinued. All patients must have a histological diagnosis of measurable disease that has
assessed by the investigator per RECIST 1.1. Safety endpoints include dose-limiting progressed with $ 2 lines of treatment, $ 12 weeks of prior PD-(L)1-based therapy, resolved
toxicities (dose escalation only) and treatment-emergent adverse events. Secondary prior toxicity with a 2–4 week washout, adequate organ function, ECOG # 2, and no
outcome measures include ORR (dose escalation only), time to response, duration of significant ascites, pleural effusion or CNS metastases, recent infections or autoimmune
response, progression-free survival (all assessed by the investigator per RECIST 1.1), disease requiring steroids or immunosuppressants. Cycles 1 and 3 include serial PK
and overall survival. The planned sample size is ~150 patients; enrollment is ongoing. sampling, while Cycles 2, 4 and beyond require only one visit on Day 1. Efficacy is assessed
Clinical trial information: NCT06644755. Research Sponsor: Daiichi Sankyo, Inc. Q6W for 1 year, then Q6 months. Safety follow-up at end of treatment is at 120 days. Clinical
trial information: NCT06090266. Research Sponsor: OncoResponse, Inc.; The Cancer
Prevention and Research Institute of Texas.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 179s

TPS2670 Poster Session TPS2671 Poster Session

A phase 1/2a, multicenter, first-in-human, open-label clinical trial evaluating ARC101-P1-101: A first-in-human phase 1 study of ARC101, a next gen-
MDX2001, a tetraspecific T cell engager-expander in patients with advanced eration T cell engager (TCE), in patients with advanced solid tumors. First
solid tumors. First Author: Ecaterina Elena Dumbrava, Department of Investigational Author: Prachi Bhave, Peter MacCallum Cancer Centre, The University of Melbourne,
Cancer Therapeutics, The University of Texas MD Anderson Cancer Center, Houston, TX Melbourne, VIC, Australia
Background: MDX2001 is a multispecific antibody recognizing CD3 and CD28 on T cells, Background: T-cell Engagers (TCEs) are emerging as a promising immuno-therapeutic
and c-MET and TROP2 on tumors. Anti-CD3 provides the primary signal for T cell modality in the treatment of solid tumors, demonstrating outstanding potency and a
activation; anti-CD28 delivers the secondary signal for enhanced T cell activation, manageable safety profile. Claudin 6 (CLDN6) is an oncofetal protein that has recently
survival, and proliferation. Combinatorial targeting of c-MET and TROP2 by MDX2001, emerged as a particularly attractive tumor-associated antigen (TAA) for TCE therapy
either on the same or different cancer cells, provides more effective engagement on because of its highly tumor-restricted pattern of expression. ARC101 is a bispecific
tumor cells, and may better address tumor heterogenicity and the development of antibody that targets CLDN6 on tumor cells with high specificity and selectivity, and CD3
resistance due to antigen downregulation. In vitro and in vivo studies with MDX2001 on T cells. In pre-clinical models, ARC101 demonstrated potent cytolytic activity at low
demonstrate potent antitumor activity with no CD28-superagonist activity and minimal concentrations against a panel of CLDN6-expressing tumor cells in vitro and an ovarian
T cell activation in the absence of tumor cells. Methods: This Phase 1/2a, multicenter, cancer xenograft in vivo. Methods: First-in-human, multicenter, phase 1 study ARC101-
first-in-human, open-label clinical trial explores intravenous MDX2001 in patients with P1-101 (NCT06672185) aims to determine the optimal dosing, safety, pharmacokinetics
advanced solid tumors (NCT06239194). The study design consists of Phase 1a dose (PK), pharmacodynamics (PD), and preliminary antitumor efficacy of ARC101 as mono-
escalation guided by a Bayesian Optimal Interval design with a target maximum tol- therapy in patients with locally advanced or metastatic CLDN6 expressing solid tumors.
erated dose toxicity rate of 30%, Phase 1b dose expansion, and Phase 2a indication The study will be conducted according to the Bayesian Optimal Interval (BOIN) design in
expansion. Patients with non-small cell lung, renal cell, prostate, breast cancer and 10 two parts: Part 1 (dose escalation) and Part 2 (dose expansion). Part 1 is designed to select
other selected tumors known to have significant levels of TROP2 or c-MET expression the Maximum Tolerated Dose (MTD), Recommended-Phase 2-Dose (RP2D) and dosing
are eligible for Phase 1a. In Phase 1b, patients will be randomized into 2 dose cohorts schedule of ARC101. Part 1 will start with an ‘Accelerated Titration Phase’, with cohorts of
using a Bayesian Optimal Phase 2 (BOP2) design. Once a recommended Phase 2 dose at least one, but no more than three patients and a fixed dose, intravenous regimen. Once a
(RP2D) is determined, Phase 2a will enroll patients in search of initial efficacy signals single event of clinically significant toxicity of Grade $2 occurs, the ‘Standard Titration
Phase’ will be initiated with cohorts of at least three patients per ARC101 target dose level.
using a BOP2 design. The primary objectives of this study are to characterize the safety,
Once immune-related toxicity is observed, the regimen may be changed to a ‘Fractionated
tolerability, and anti-tumor activity of MDX2001 in patients with advanced solid tumors.
Step-up Dosing’ IV regimen. The study design allows for backfill cohorts and intra-patient
Secondary endpoints include time to response, disease control rate, duration of re-
dose escalations. Part 2 will further explore the safety, PK/PD characteristics, and pre-
sponse, pharmacokinetics, immunogenicity and evaluation of the relationship between
liminary efficacy of ARC101 administered at the RP2D and schedule identified in Part 1 in
baseline tumor target protein expression and clinical benefit. Patients will have ra- patients with testicular and ovarian cancer. Key eligibility criteria include patients with any
diologic tumor assessments every 8 weeks and will continue to receive treatment until advanced or refractory solid tumor malignancy that expresses CLDN6 and is metastatic or
disease progression per RECIST v1.1 (as assessed by the investigator), unacceptable unresectable. Patients must be $18 years of age and have Eastern Cooperative Oncology
toxicity, withdrawal of consent, another protocol-defined discontinuation criterion is Group (ECOG) performance status 0 or 1. Patients must have received standard therapy for
met, or the sponsor terminates the study, whichever occurs first. The study will be advanced or metastatic disease, and disease must be measurable per Response Criteria in
conducted in United States, Europe, and Asia. Recruitment is ongoing. Clinical trial Solid Tumors (RECIST) v1.1 or evaluable. Mandatory requirement of a pre-study tumour
information: NCT06239194. Research Sponsor: None. sample for IHC analysis will facilitate the exploratory objective of biomarker analysis,
including correlating CLDN6 expression with treatment response. The study is actively
enrolling participants for the dose escalation phase. Contact clinicaltrials@thirdarcbio for
additional information. Clinical trial information: NCT06672185. Research Sponsor: None.

TPS2672 Poster Session TPS2673 Poster Session

EGL-121, a first-in-human phase 1/2 trial of EGL-001 in adult patients with SUPRAME: A phase 3 trial comparing IMA203, an engineered T-cell receptor
selected advanced and/or metastatic solid tumors. First Author: Thiziri expressing T cell therapy (TCR-T) vs investigator’s choice in patients with
Nait Achour, EGLE Therapeutics, Suresnes, France previously treated advanced cutaneous melanoma. First Author: Jason J. Luke,
Background: Regulatory T cells (Tregs) play a key role in the resistance to immune University of Pittsburgh, Pittsburgh, PA
checkpoint inhibitors therapy (ICI). Disarming Tregs could therefore restore/enhance Background: Frequent recurrence and limited long-term survival in unresected or
anti-tumor responses and increase the number of patients benefiting from these metastatic melanoma after relapse from 1L treatment with a checkpoint inhibitor (CPI)
treatments. EGL-001, a novel therapeutic agent, is designed to provide checkpoint highlight the critical need for new therapies that deliver deeper, more durable responses
inhibition by antagonizing the CTLA-4-CD80/86 interaction while selectively depleting (Knight Cancers 2023; Switzer JCO Oncol Pract 2022). ACTengine IMA203 is an au-
intratumoral Tregs by downregulating CD25 and inhibiting IL-2 signaling specifically tologous T cell receptor (TCR)-engineered T cell therapy (TCR-T) targeting PRAME, an
within these cells. This dual mechanism of action effectively unleashes potent anti- intracellular protein displayed as peptide antigen at high density on the surface of multiple
tumor immunity even in anti-PD-1 resistant models, independent of FcgR activity. In solid tumors, including melanoma. IMA203 TCR-T demonstrated a favorable tolerability
murine models, EGL-001 shows preferential distribution and persistence in the tumor profile and durable objective responses in heavily-pretreated patients with different tumor
until Treg get depleted/inactivated. Our data demonstrated complete anti-tumor activity types. In melanoma, IMA203 showed 54% confirmed ORR (14/26), 12.1 months mDOR and
of EGL-001 as a single agent across various tumor models and it overcomes resistance 6 months mPFS. mOS was not reached at a mFU of 8.6 months (Wermke et al., SMR, Oct
to anti-PD-1 treatment in many tumor models, highlighting its broad therapeutic po- 10, 2024). Based on these observations, a registration-enabling randomized phase 3 trial,
tential. Additionally, EGL-001 effectively depletes Tregs and exhibits activity in ex-vivo SUPRAME, was initiated to evaluate IMA203 in 2L patients with advanced cutaneous
human tumor samples, where other ICI showed no significant effect. In NHPs, EGL-001 melanoma after treatment with a CPI. Methods: SUPRAME (NCT06743126) is a phase 3,
was well tolerated across all tested doses, with rapid peripheral clearance preventing multicenter, open-label, randomized, actively controlled, parallel-group trial that will
lymphoid tissue hyperplasia in the spleen and lymph nodes. Methods: A Phase I/II evaluate the efficacy, safety and tolerability of IMA203 compared to investigator’s choice
clinical trial (NCT06622486) is currently underway in eight sites in France and Spain to of treatment in patients with previously treated, unresectable or metastatic cutaneous
evaluate EGL-001 as monotherapy and in combination with checkpoint inhibitors in melanoma (incl. acral melanoma). Eligible patients are $18yo, HLA-A*02:01-positive,
selected tumor types characterized by tumor Treg implication in induction of mech- with measurable disease (RECIST v1.1), ECOG PS of 0-1 and disease progression on or
anism of resistance to ICI. The selective targeting of tumor-infiltrating Tregs could after at least one PD-1 inhibitor. Patients with BRAF mutation should have been treated
effectively improve anti-tumor immune response and limit systemic immune-related with one prior line of BRAF-directed therapy (6 MEK inhibitor) prior to initial eligibility
toxicities. This first-in-human, multicenter, open label Phase 1/2 study evaluates the assessment. Patients with asymptomatic stable brain or leptomeningeal metastases will
safety, tolerability, and initial activity of EGL-001 in adult patients with selected ad- be assessed for eligibility. Patients with active brain metastases or with primary mucosal,
vanced and/or metastatic solid tumors. The study consists of a Part 1 (Phase 1) dose uveal melanoma and melanoma of unknown primary are excluded. The study will ran-
escalation of EGL-001 administered as a single agent (from 0.3 mg/kg to 12 mg/kg), and domize ~360 patients 1:1. Patients in the experimental arm will undergo leukapheresis to
in combination with pembrolizumab treatment, according to a BOIN design, followed generate the PRAME-specific TCR-T product, IMA203. Following lymphodepletion with
by a Part 2 (Phase 2) dose expansion of EGL-001 administered at the selected doses as cyclophosphamide (500 mg/m2 x 4 days) and fludarabine (30 mg/m2 x 4 days), 1-10x109
monotherapy and/or in combination therapy with anti-PD(L)1. Eligible patients are those IMA203 TCR-T cells will be administered, followed by low-dose IL-2 (1mio IU daily x5 days,
who have initially benefited (secondary resistance) from an ICI treatment as mono- twice daily x5 days). Patients in the control arm will receive approved investigator’s choice
therapy or in combination as SoC as defined by a CR, PR, or SD $3 months as best of standard treatment (nivolumab/relatlimab, nivolumab, ipilimumab, pembrolizumab,
response by RECIST Version 1.1. As of January 2025, the first 3 Cohorts of EGL-001 lifileucel (US), chemotherapy). The primary efficacy endpoint is BICR-assessed (RECIST
(0.03, 0.1, 0.3 mg/kg) have been completed. EGL-001 was well tolerated with no DLTs v1.1) PFS. Secondary endpoints include OS, ORR, safety and patient-reported outcomes
reported. Clinical trial information: NCT06622486. Research Sponsor: None. (EORTC QLQ-C30, EQ-5D-5L). The trial will enroll patients in the US and Europe. Clinical
trial information: NCT06743126. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
180s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

TPS2674 Poster Session TPS2675 Poster Session

Autologous tumor-infiltrating lymphocytes (HS-IT101) with low-dose lym- A phase 1/2 study of KSQ-004EX: Autologous tumor
phodepletion and IL-2 infusion for the treatment of advanced solid tumors: A infiltrating lymphocytes, engineered to inactivate genes encoding SOCS1
phase I clinical trial. First Author: Ning Li, Department of Clinical Trial Center, National and Regnase-1, in patients with select advanced solid tumors. First Author:
Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Rodabe Navroze Amaria, The University of Texas MD Anderson Cancer Center, Melanoma
Academy of Medical Sciences and Peking Union Medical College, Beijing, China Medical Oncology, Houston, TX
Background: Adoptive cell therapy with tumor-infiltrating lymphocytes (TIL-ACT) has Background: The effectiveness and durability of TIL therapy may be limited by the
demonstrated great therapeutic potential in numerous solid tumors and has become an immunosuppressive tumor microenvironment and baseline functionality of transferred
effective treatment for melanoma. However, high-dose lymphodepletion chemotherapy T cells. Through KSQ Therapeutics’ CRISPR2 platform, a novel method for screening
and IL-2 infusion during the treatment could cause serious safety risks, even death. The optimal combinatorial targets for enhancing T cell anti-tumor efficacy in vivo, SOCS1
purpose of this study is to develop a TIL cell therapy product (HS-IT101) that requires and Regnase-1 were identified as the most potent gene editing combination. KSQ-
low-dose lymphodepletion and IL-2 infusion to reduce safety risks and to improve 004EX, an engineered TIL product with CRISPR/Cas9 mediated dual-inactivation of
clinical accessibility. Currently, the Phase I clinical trial (NCT06342336) for the SOCS1 and Regnase-1, is anticipated to in enhance T cell tumor infiltration, persistence,
treatment of advanced solid tumors with HS-IT101 has been initiated. Methods: HS- and efficacy. This first-in-human clinical study (NCT06598371) evaluates KSQ-004EX in
IT101 is an autologous non-genetically modified TIL-ACT product independently de- patients with melanoma, non-small cell lung cancer (NSCLC), head and neck squamous
veloped by Sino-cell Biomed. The tumor tissue of culture require is $0.05g, and the cell carcinoma (HNSCC), colorectal carcinoma (CRC), pancreatic cancer, and cervical
manufacture time needed is 14 days. This study is a single-arm, multi-center, open-label cancer. Methods: The phase 1/2, single-arm, open-label study will assess the safety,
Phase I clinical trial of HS-IT101 for advanced solid tumors. The plan is to enroll 20 - 44 tolerability, and efficacy of KSQ-004EX in patients with select advanced solid tumors.
patients to explore the safety and preliminary efficacy under low-dose lymphodepletion Patients with melanoma, NSCLC, HNSCC, CRC, pancreatic, and cervical cancer who have
and IL-2 infusion. Before HS-IT101 infusion, subjects will receive lymphodepletion progressed following treatment with 1 to 3 lines of prior standard therapy including
chemotherapy consisting of cyclophosphamide (Cy) and fludarabine (Flu) for 3 - 4 days standard directed therapy (as applicable), are eligible. KSQ-004EX is manufactured from
(Cy: 900/2250mg/m2 & Flu: 90/120mg/m2). After HS-IT101 infusion, 1/2MIU/m2 of IL-2 the patient’s tumor, which is collected through surgical resection or core needle biopsy.
will be subcutaneously injected once a day for a maximum of 3 doses. The primary All patients must have at least 1 measurable lesion following resection. Patients receive
endpoint is the occurrence of adverse events (AE) and serious adverse events (SAE) after lymphodepleting chemotherapy with cyclophosphamide and fludarabine prior to KSQ-
HS-IT101 infusion. The secondary endpoints include the objective response rate (ORR), 004EX infusion. Patients in the initial dose escalation cohorts do not receive dosing with
disease control rate (DCR), time to response (TTR), duration of response (DOR), IL-2; IL-2 dosing may be included in subsequent cohorts. Approximately 6 patients will
progression-free survival (PFS), overall survival (OS) in efficacy evaluation, and changes be enrolled in Phase 1 dose escalation, in escalating dose levels. The primary objective
in relevant pharmacokinetic (PK) indicators. The exploratory endpoint is the change in of Phase 1 is to evaluate the safety and tolerability of KSQ-004EX. In Phase 2, patients
pharmacodynamic (PD) indicators. Clinical trial information: CTR20234065. Research will be enrolled in indication-specific cohorts. The primary objective of Phase 2 is to
Sponsor: Qingdao Sino-Cell Biomedicine Co., Ltd. assess the anti-tumor activity of KSQ-004EX in patients with advanced solid tumors by
ORR per RECIST v1.1. This is currently a single-institution study that is actively enrolling/
recruiting patients. Clinical trial information: NCT06598371. Research Sponsor: KSQ
Therapeutics, Inc.

TPS2676 Poster Session TPS2677 Poster Session

A phase 1, first-in-human study of IB-T101, an OUTLAST CAR-T product for Logic-gated, allogeneic Tmod chimeric antigen receptor T-cell (CAR T)
the treatment of CD70-positive clear cell renal carcinoma. First Author: therapy targeting epidermal growth factor receptor (EGFR) in advanced
Matthias Schroff, Inceptor Bio, Morrisville, NC solid tumors with human leukocyte antigen (HLA) loss of heterozygosity
Background: Relapsed or treatment-resistant clear cell renal cell carcinoma (ccRCC) (LOH): DENALI-1 trial. First Author: Kedar Kirtane, Moffitt Cancer Center, Tampa, FL
poses a significant, unmet medical challenge, as patients contend with scarce thera- Background: Despite the success in hematologic malignancies, CAR T therapies face
peutic alternatives and unfavorable clinical prognoses. CD70 is expressed in the majority significant challenges in solid tumors due to the lack of tumor-specific targets that
of ccRCC and presents an attractive target for chimeric antigen receptor T cell (CAR-T) distinguish cancer from normal cells. EGFR plays a critical role in oncogenesis across
therapy. IB-T101, an autologous CAR-T expanded under OUTLAST conditioning, targets several cancers and is often upregulated (TCGA 2022). While monoclonal antibodies
CD70 for the treatment of ccRCC. OUTLAST conditioning has been demonstrated to targeting EGFR have demonstrated efficacy, these approaches are often limited by on-
result in CAR-T cells that exhibit an early memory T cell phenotype, are resistant to target, off-tumor toxicities, such as skin rash, which constrains dose escalation and
suppressive signals from the tumor microenvironment, and exhibit increased persis- efficacy (Macdonald, et al. J Am Acad Dermatol. 2015). A2B395 is an allogeneic, logic-
tence. The effects of OUTLAST conditioning are expected to lead to superior clinical gated, EGFR-targeted Tmod CAR T therapy designed to address these limitations and
outcomes for IB-T101 CAR-T cells in the ccRCC solid tumor setting. Methods: Here we provide a convenient and consistent off-the-shelf option. This therapy incorporates 2
report an in-progress phase 1, first-in-human, open label, investigator-initiated clinical CARs: an activator targeting EGFR, and a blocker targeting HLA-A*02. The activator
trial aimed at evaluating the safety and efficacy of IB-T101 in ccRCC. Patients eligible for recognizes EGFR on both tumor and normal cells, while the blocker inhibits CAR T activity
inclusion had previously relapsed following VEGF targeting therapies alone or in against normal cells with preserved HLA expression and decreases the risk for graft-
combination with an immune checkpoint inhibitor. Autologous patient T cells are versus-host disease (Hamburger, et al. Mol Immunol. 2020). To address potential host-vs-
transduced with a lentiviral vector encoding a CD70-targeting CAR and are CRISPR Cas9 graft response, an shRNA expression module targeting B2M is included in the Tmod
gene edited to knock out endogenous CD70, followed by expansion under OUTLAST construct, which significantly reduces major histocompatibility complex class I levels and
conditioning. Escalating doses of IB-T101 CAR-T cells (150 – 500 x 106) will be infused subsequent host immune response (DiAndreth, et al. Clin Immunol. 2022). Importantly,
following lymphodepletion. Primary endpoints of the study will assess the safety and the Tmod system is modular and adaptable to multiple targets. Initial data on autologous
tolerability of IB-T101. Additional objectives of the study are to assess the anti-tumor Tmod CAR T therapy suggest reduced off-tumor toxicity and encouraging clinical efficacy
activity and the pharmacokinetics of IB-T101. Correlative assessments will include pre- (Grierson, et al. SITC 2024. Abstract 588). A2B395 represents a novel approach for EGFR-
treatment biopsies to assess the level of CD70 expression in the tumor. Research expressing solid tumors with HLA-A*02 LOH. Methods: DENALI-1 (NCT06682793) is a
Sponsor: None. phase 1/2, open-label, nonrandomized study evaluating the safety and efficacy of A2B395
in adults. Patients are enrolled through BASECAMP-1 (NCT04981119), a master pre-
screening study that identifies patients with HLA LOH at any time in the course of their
disease via next-generation sequencing (Tempus AI, Inc.). Key inclusion criteria include
histologically confirmed recurrent unresectable, locally advanced, or metastatic cancers
associated with EGFR expression, including colorectal, non-small cell lung, squamous cell
head and neck, triple negative breast, and renal cell cancers. Patients must have
received $1 line of prior therapy, such as a checkpoint inhibitor, molecular targeted
therapy, or chemotherapy. The primary objective of phase 1 is to evaluate safety, tol-
erability, and the recommended phase 2 dose (RP2D) using a Bayesian optimal interval
design for dose escalation. The dose-expansion phase will confirm RP2D and collect
biomarker data. Phase 2 will assess overall response rate per RECIST v1.1. Clinical trial
information: NCT06682793. Research Sponsor: A2 Biotherapeutics, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 181s

TPS2678 Poster Session TPS2679 Poster Session

+ +
A phase I study of AFNT-211, autologous CD4 and CD8 T cells engineered Safety and efficacy of HLA-G–targeted CAR T cells (IVS-3001) in patients
to express a high avidity HLA-A*11:01-restricted, KRAS G12V-specific with advanced HLA-G–positive solid tumors: Clinical trial in progress. First
transgenic TCR; CD8a/b coreceptor; and FAS-41BB switch receptor in Author: Samer Ali Srour, The University of Texas MD Anderson Cancer Center, Houston,
patients with advanced or metastatic solid tumors. First Author: Soumit K. TX
Basu, Affini-T Therapeutics, Inc., Watertown, MA Background: Immunotherapies have transformed cancer treatment, yet only a small
Background: Activating mutations in KRAS (including KRAS G12V) are well-described proportion of patients experiences durable responses. IVS-3001 is an innovative au-
oncogenic drivers in solid tumors, conferring poor prognosis to patients due to a lack of tologous chimeric antigen receptor (CAR) T-cell therapy specifically targeting Human
effective therapies for cancers with such KRAS driver mutations. T cell receptor (TCR)- Leukocyte Antigen (HLA-G). HLA-G is an immune-modulatory checkpoint molecule
T cell therapies targeting mutant KRAS have demonstrated proof of concept in the clinic, expressed on various solid tumors, positioning it as an ideal a tumor-specific targeted
but duration of response remains a challenge.1,2 AFNT-211 represents a novel strategy antigen. Our third-generation CAR construct features enhanced T cell activation and
to address the immunosuppressive tumor microenvironment and improve response rate persistence against HLA-G. By harnessing IVS-3001 to target HLA-G and revitalize
as well as duration of response in solid tumors. Methods: This ongoing Phase 1, first-in- immune cells, we aim to overcome the suppressive tumor microenvironment and
human, multicenter, open-label study of AFNT-211evaluates safety/tolerability, as well improve antitumor activity, potentially leading to better outcomes for patients with
as its clinical (antitumor) activity with the goal to identify an optimal biological dose advanced solid tumors who otherwise have no standard options known to confer clinical
(OBD) and recommended Phase 2 dose (RP2D) in patients with HLA 11:01 who suffer benefit. Methods: Study NCT05672459 is a First-in-Human, phase 1/2a, safety and
from cancers driven by the KRAS G12V mutation. The initial dose escalation part of the efficacy study of IVS-3001 in subjects with previously treated advanced HLA-G-positive
study follows Bayesian optimal interval Phase 1/2 (BOIN12), which quantifies the solid tumors. Phase 1 (n#24 patients) is a Bayesian Optimal Interval Design (BOIN) with
desirability of a dose in terms of toxicity-efficacy tradeoff and adaptively allocates primary objective to determine the safety, tolerability and the recommended phase 2
patients to the dose with the highest estimated desirability. After determination of OBD dose. The primary objective for phase 2 (n#90 patients) is to evaluate the anti-tumor
and RP2D based on the totality of the risk/benefit assessment and the BOIN12, the study activity of IVS-3001. The secondary objectives of the study are to evaluate i) phar-
is planned to proceed to the dose expansion phase which will consist of cohorts enrolling macokinetic profile of IVS-3001 (persistence, expansion); ii) the clinical activity of IVS-
patients with tumors with high KRAS G12V prevalence (pancreatic cancer, colorectal 3001 in selected HLA-G+ solid tumor types; iii) assess the long-term safety of IVS-3001.
cancer, non small cell lung cancer) as well as a tumor agnostic arm (any other solid Exploratory endpoints include functionality of CAR-T cells, immune biomarker changes,
tumor with KRAS G12V). This study has started enrolling patients $ 18 years old and relationships with clinical response. Key inclusion criteria: adults with advanced
positive for HLA-A*11:01-positivewithadvanced/metastatic solid tumors harboring a solid tumors expressing HLA-G; ECOG , 2; adequate organ function. Key exclusion
KRAS G12V mutation who have proven intolerant of or refractory to at least one prior criteria: uncontrolled brain metastasis; prior exposure to HLA-G targeted therapy.
standard of care systemic therapy. Patients undergo leukapheresis to collect T cells for Subjects undergo lymphodepletion with fludarabine and cyclophosphamide on days -5
the manufacturing of AFNT-211, and receive lymphodepleting chemotherapy prior in- to -3, followed by CAR-T cell infusion on day 0 and a 28-day monitoring period for dose
fusion of their autologous AFNT-211 product. Following this, patients proceed into a 28- limiting toxicity. Response assessment per RECIST criteria. Study is currently accruing
day dose-limiting toxicity observation period (during dose escalation) followed by a post- at Dose level 3. Active recruitment and enrollment are ongoing at The University of Texas
treatment follow-up period for 24 months/until disease progression. The study is open MD Anderson Cancer Center, Houston, Texas. Clinical trial information: NCT05672459.
for recruitment in the United States (NCT06105021). References: 1. Cook J, Melloni G, Research Sponsor: Invectys; National Cancer Institute.
Gulhan D, et al. The origins and genetic interactions of KRAS mutations are allele- and
tissue-specific. Nat Commun 2021;12:1808. 2. Hofmann MH, Gerlach D, Misale S, et al.
Expanding the reach of precision oncology by drugging all KRAS mutants. Cancer
Discov. 2022;12:924–937. Clinical trial information: NCT06105021. Research Sponsor:
Affini-T Therapeutics, Inc.

TPS2680 Poster Session TPS2681 Poster Session

QUILT 3.076 phase 1 study of memory-like cytokine-enriched natural killer A phase I, multicenter, open-label study of UB-VV111 in combination with
(M-CENK) cells plus N-803 in locally advanced or metastatic solid tumors. rapamycin in relapsed/refractory CD19+ B-cell malignancies. First Author:
First Author: Chaitali Singh Nangia, Chan Soon-Shiong Institute for Medicine, El Jacob Randolph Garcia, Umoja Biopharma, Seattle, WA
Segundo, CA Background: Autologous, ex vivo-manufactured chimeric antigen receptor (CAR) T cells
Background: Lymphopenia and low levels of natural killer (NK) cells may contribute to poor directed against CD19 have demonstrated clinical activity. These products have gained
prognosis and response to therapy in cancer patients, conditions that may be addressed by approvals in the relapsed/ refractory (R/R) setting in multiple B-cell malignancies (BCMs),
infusion of memory-like cytokine-enriched NK (M-CENK) cells stimulated ex vivo by IL-12, IL- including large B-cell lymphoma (LBCL) and chronic lymphocytic leukemia/small lym-
18, and the IL-15 agonist N-803 (ANKTIVA). M-CENK cells express elevated IFN-g and phocytic lymphoma (CLL/SLL). However, challenges in product availability due to limited
granzyme B compared to healthy donor NK cells, and display toxicity against multiple tumor manufacturing capacity, the need for apheresis and lymphodepletion, failure to prior
cell lines including SCLC lines [Fousek 2023 JITC 11 ab358]. The phase 1 study QUILT-3.076 ex vivo CAR T therapy, and the level of patient fitness needed to wait for and receive ex vivo
(NCT04898543) assesses the safety and preliminary efficacy of M-CENK cells plus N-803 in autologous CAR T therapy all pose significant challenges to the field, presenting significant
participants with locally advanced or metastatic solid tumors. Methods: In this first-in- unmet clinical need. UB-VV111 is a third-generation, self-inactivating, replication-
human study, cohort 1 (up to n = 40) includes participants with newly diagnosed solid incompetent lentiviral vector (LVV) investigational drug product comprising an enve-
tumors who have not received prior 1st line treatment; cohort 2 (up to n = 21) includes lope with cocal virus fusion glycoprotein (cocal) and surface engineered with a membrane-
participants with relapsed/refractory solid tumors who progressed after $ 2 prior therapies. bound multidomain fusion (MDF) protein. The MDF protein contains CD58, CD80, and anti-
Both cohorts undergo apheresis (part A), but only cohort 2 undergoes treatment with M- CD3 single-chain variable fragment (scFv) components that provide both T-cell tropism
CENK cells and N-803 (part B). During M-CENK cell generation, cohort 2B participants and activation signals thought to be critical for effective CAR T-cell generation. UB-VV111
receive oncologist-recommended therapy. Cohort 1 participants may subsequently enroll in addresses the limitations of currently available autologous CD19-directed CAR T therapies
cohort 2B if they have progressive disease (PD) after $ 2 prior therapies or within 12 months to deliver a product that would generate CD19-directed cells in the patient. UB-VV111 is to
of receiving neoadjuvant/adjuvant chemotherapy. In part B, M-CENK cells are administered be administered by either intranodal (IN) or intravenous (IV) route of administration (ROA).
weekly up to 10 times and N-803 SC for up to 5 doses every 2 weeks prior to every other dose Administration of UB-VV111 by either the IN or IV ROA is expected to transduce T cells to
of M-CENK cells. Key inclusion criteria are age $ 18 years, ECOG performance status of 0 to generate CAR T cells designed to bind to CD19 antigen to mediate cell killing and express
2, and histologically confirmed locally advanced or metastatic solid tumor, with at least 1 the rapamycin-activated cytokine receptor (RACR) system which, in the presence of
measurable lesion and/or non-measurable disease in accordance with RECIST v1.1. There rapamycin, is designed to enhance specific enrichment and expansion of transduced cells.
are no exclusion criteria for part A (apheresis). Key exclusion criteria for part B are life Methods: Study UB-VV111-01 (INVICTA, [NCT06528031CO]) is a first-in-human, global,
expectancy , 16 weeks, involuntary weight loss of . 10%, serious uncontrolled con- multicenter, dose-finding study of UB-VV111 administered IN or IV +/- rapamycin in CAR-
comitant disease, systemic autoimmune disease requiring medical treatment, and/or naı̈ve and CAR-exposed subjects with R/R LBCL and CLL/SLL. Dose escalation will proceed
currently receiving or received antibiotics since enrollment. The primary objective is independently for each ROA using a Bayesian optimal interval (BOIN) design. Confirmation
safety as assessed and recorded by TEAEs, SAEs, and clinically significant changes in of CD19 expression will be required for all subjects with prior CD19-directed therapy. Major
laboratory tests and vital signs. Toxicities are graded using CTCAE v5.0 or a specified eligibility criteria include adults with R/R LBCL/CLL/SLL following at least 2 lines of prior
grading system for CRS. Secondary measures evaluate the quantity and quality of the therapy who have standard organ function, measurable disease according to Lugano 2014
investigational M-CENK cells (number of MNCs for manufacturing M-CENK cells, number of (LBCL) or iwCLL 2018 (CLL/SLL), ECOG 0 or 1, and no prior allogeneic transplant. Primary
cryopreserved M-CENK aliquots, % NK cells, and number, phenotype, and function of M- objectives include determining the safety profile, maximum tolerated/administered dose,
CENK cells). Preliminary efficacy objectives in cohort 2B are objective response rate and recommended Phase 2 dose of UB-VV111 +/- rapamycin. Secondary/exploratory
(RECIST v1.1 and iRECIST criteria) and progression-free and overall survival evaluated using objectives include measuring preliminary antitumor activity (magnitude and durability), as
Kaplan-Meier methods. As of January 27, 2025, 15 participants have been enrolled in cohort well as translational correlates of safety/efficacy. Clinical trial information: NCT06528301.
1, 21 participants in cohort 2 have undergone apheresis, and 10 participants have been Research Sponsor: Umoja Biopharma.
treated with study therapies. Clinical trial information: NCT04898543. Research Sponsor:
ImmunityBio, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
182s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

TPS2682 Poster Session TPS2683 Poster Session

Phase 1 clinical trial of autologous T-cells genetically engineered with a INVOKE: A phase 1 study of OKN4395, a first-in-class EP2/EP4/DP1 triple
chimeric receptor to target the follicle-stimulating hormone receptor (FSHR) prostanoid receptor antagonist, in patients with advanced solid tumors. First
in recurrent ovarian cancer (OVCA). First Author: Robert Michael Wenham, Author: Neal Shiv Chawla, Sarcoma Oncology Center, Santa Monica, CA
Department of Gynecologic Oncology, H. Lee Moffitt Cancer Center and Research In- Background: Immunotherapy is an established cancer therapy, although mechanisms
stitute, Tampa, FL of non-response & resistance are emerging, leaving few options post-relapse. The
Background: FSHR is a tissue specific antigen expressed in . 55% of high-grade immunosuppressive pathway of prostaglandin E2 (PGE2), part of the cyclooxygenase
epithelial OVCAs with negligible FSHR expression in non-ovarian tissues. OVCA xe- (COX) pathway, is upregulated in certain cancers and has been implicated in tumor
nografts treated with FSHCER T (FSH-Chimeric Endocrine Receptor + T-Cell (CER T)) evasion of CD8 T, NK, and dendritic immune cells, allowing tumor growth and metastasis
cells demonstrated cytotoxic activity against patient-derived FSHR+ ovarian carcino- (Jin et al., 2023). COX2 inhibitors, aspirin and nonsteroidal anti-inflammatories (NSAIDS)
mas. We hypothesize targeting FSHR in women with FSHR+ OVCA will result in improved have shown some survival benefit in patients with colon, lung, prostate, and endometrial
response rates due to engraftment, expansion, and survival of these adoptively cancer (Cao et al., 2016; Lim et al., 2012; Huang et al., 2014; Takiuchi et al., 2018),
transferred FSHCER T-cells and will have acceptable toxicity. Methods: The primary however results are inconsistent, likely due to toxicity limiting complete blockade of the
objective of this phase 1 dose-escalation study (NCT05316129) in high-grade epithelial pathway, highlighting the need for more potent but selective COX pathway inhibitors.
OVCA using T-cells genetically modified to express CER targeting FSHR is to assess the OKN4395 is a first-in-class, highly selective, equipotent inhibitor of EP2, EP4 and DP1,
safety of the intraperitoneal (IP) and intravenous (IV) infusions of FSHCER T-cells. downstream receptors for COX-derived PGE2, and PGD2, respectively. DP1 has described
Secondary objectives include antitumor efficacy, persistence of transferred FSHR roles in immunosuppression and inhibition of apoptosis, supporting the therapeutic
T cells, expansion of endogenous tumor-targeted cells, and comparison of IP and IV rationale (Luo et al., 2024; Peinhaupt et al., 2017). OKN4395 is hypothesized to modulate
administration routes. Patients unable to be treated in the IP arm may be treated in the the tumor microenvironment to allow an effective immune response as monotherapy,
IV arm in the lowest unfilled cohort for that arm. Cohorts of 3 to 6 patients will be infused and to potentiate the effect of immunotherapies such as checkpoint inhibitors, both of
with escalating doses of FSHCER T-cells to establish the maximum tolerated dose (MTD) which are evaluated in INVOKE. Methods: INVOKE (OKN-4395-121; NCT06789172) is a
with 6 planned dose levels from 1 x105 to 1 x 107 cells/kg with the 5th level receiving Ph1a/1b, first-in-human study of OKN4395 (oral, BID) as monotherapy (mono) or in
lymphodepleting chemotherapy. Following MTD determination, an expansion phase will combination with pembrolizumab 200mg IV 3-weekly (combo), in patients with ad-
be initiated. Nine patients have been enrolled in the first three dose-level cohorts. Eight vanced solid tumors that have evidence of COX-associated immunosuppression. Ph1a
have cleared the DLT period and one patient is currently being treated. One patient is a Bayesian dose escalation in mono, followed by combo dose confirmation, primarily
received a second dose of 3 x 105 cells/kg after 20 months apparent stable disease. assessing safety, establishing the optimal dose for Ph1b. Using multimodal artificial
Cohorts 1 and 2 correlates are being processed. NCT05316129. Moffitt Scientific Review intelligence (AI) drug-matching algorithms, Ph1b tumor types were selected, and re-
#21113. Advarra Institutional Review Board #00000971. Clinical trial information: sponse will be assessed (cohorts of n = 20 each): select sarcomas (mono), pancreatic
05316129. Research Sponsor: Anixa BioSciences Inc., San Jose CA, USA. carcinoma (mono), non-small cell lung cancer (combo), colorectal carcinoma (combo),
head and neck squamous cell carcinoma (combo). Key inclusion criteria include COX-
active (Ph1a) or above-listed (Ph1b) tumors, performance status 0-1, biopsy-amenable
lesions, and adequate organ function. Active CNS metastases, upper GI bleed risk
factors, untreated H. pylori infection, and concomitant NSAIDs/COX inhibitors/
prostaglandins are exclusionary. Ph1b mono cohorts will include exploratory ana-
lyses including evaluation of the effect of food & gastric pH on OKN4395 pharmaco-
kinetics. Trial data, paired pre- and on-treatment biopsies, and exploratory biomarkers
will be used to enhance development using advanced agentic AI systems, including a
synthetic digital twin control arm. Ph1a of the study is currently recruiting in the US, UK,
and Australia. Clinical trial information: 06789172. Research Sponsor: Epkin.

TPS2684 Poster Session TPS2685 Poster Session

An open-label, phase Ib dose-expansion study to assess the efficacy of A phase 1, first-in-human study of CTIM-76, a claudin-6 (CLDN6)-directed
CD137/FAP agonist BI 765179 plus pembrolizumab as a first-line treatment bispecific antibody, in patients with recurrent ovarian cancer and other
in metastatic or incurable, recurrent programmed cell death ligand-1 (PD- advanced solid tumors. First Author: Roisin Eilish O’Cearbhaill, Memorial Sloan
L1)-positive head and neck squamous cell carcinoma (HNSCC). First Author: Kettering Cancer Center and Weill Cornell Medical College, New York, NY
Rachna T. Shroff, University of Arizona Cancer Center, Tucson, AZ Background: CLDN6 is an oncofetal protein expressed at high levels in many solid
Background: HNSCC is the seventh most common cancer globally and is often as- tumors while expressed at very low levels in adult normal tissues. The high target
sociated with poor quality of life and a dismal prognosis. Median overall survival for antigen density and slow rate of internalization makes it an attractive target in cancer
advanced HNSCC with first-line standard-of-care pembrolizumab 6 chemotherapy is therapeutics. CTIM-76, a CLDN6 x CD3 T cell engager bispecific antibody is engineered to
approximately 13 months, highlighting the need for new therapies. Fibroblast activation bind with high selectivity to CLDN6 and redirect the immune system’s T cells to rec-
protein (FAP)-positive fibroblasts are frequently present in the tumor stroma of HNSCC ognize and kill CLDN6-expressing cancer cells. CTIM-76 effectively inhibited tumor
tumors, representing a potential therapeutic target. BI 765179 is a bispecific antibody growth, inducing complete responses in ovarian cancer xenograft models. The first in
that simultaneously binds to FAP and CD137 expressed on T-cells, leading to local human study of CTIM-76 in patients with advanced ovarian, endometrial, and testicular
activation of tumor-specific CD137-positive T-cells. The Phase Ia part of the present cancers (NCT06515613) is described here. Methods: Part 1 dose escalation exploring 9
study (NCT04958239) determined safety and doses for dose escalated BI 765179, both ascending dose levels with a 3+3 design. The first two dose levels are single patient
as monotherapy and in combination with an anti-programmed cell death protein 1 (PD-1) cohorts (22.5 mg starting dose), CTIM-76 delivered as weekly iv infusions, with step
antibody in patients with advanced solid tumors. Here we present the design of the dosing and steroid premedication to minimize cytokine release syndrome. Approxi-
Phase Ib dose-expansion part, which aims to assess the preliminary efficacy of two mately 40 patients with platinum resistant ovarian cancer, or endometrial or testicular
doses of BI 765179 in combination with pembrolizumab in patients with metastatic or cancers relapsed after standard of care will be enrolled. Tumors from patients with
incurable, recurrent HNSCC whose tumors express PD-L1. Methods: In the Phase Ib ovarian or endometrial cancer require prospective CLDN6 + confirmation by IHC (10% $
dose-expansion part, approximately 60 patients with a histologically or cytologically 1+), testicular cancer patients will not require prospective screening due to the known
confirmed diagnosis of metastatic or incurable, recurrent HNSCC will be enrolled. Key uniformly high prevalence of CLDN6. The primary objective is to evaluate safety and
inclusion criteria are: no prior systemic therapy administered in the metastatic or in- tolerability (incidence and severity of adverse events per NCI CTCAE v5.0) and establish
curable recurrent setting; primary tumor locations of oropharynx, oral cavity, hypo- the recommended dose for expansion. Secondary objectives include assessment of
pharynx, or larynx; at least one measurable lesion outside of the central nervous system antitumor activity (RECIST v1.1, iRECIST), pharmacokinetics, and pharmacodynamic
(modified RECIST v1.1); a PD-L1-positive tumor (combined positive score $1, local correlates of immune activation. Part 2 will evaluate two doses in approximately 30
assessment); and Eastern Cooperative Oncology Group performance status 0–1. Pa- patients with one tumor type, with efficacy and further safety as primary objectives. This
tients who have previously received CD137-targeted or anti-PD-1/PD-L1 agents are not multicenter study has currently five sites open for enrollment. The first patient was
eligible. Patients will be randomized 1:1 to receive either Dose 1 or Dose 2 of BI 765179 dosed in January 2025. Clinical trial information: NCT06515613. Research Sponsor:
intravenously in combination with pembrolizumab. The primary endpoint is objective None.
response (OR), defined as best overall response of confirmed complete or partial re-
sponse (RECIST v1.1). Secondary endpoints include occurrence of adverse events (AEs)
and serious AEs, OR (immune-related RECIST v1.1), duration of response, progression-
free survival, and overall survival. Copyright 2025 AACR. Reused with permission.
Clinical trial information: NCT04958239. Research Sponsor: Boehringer Ingelheim.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 183s

TPS2686 Poster Session TPS2687 Poster Session

A phase 1a/1b study to evaluate the safety, tolerability, pharmacokinetics, Phase 2 dose expansion of START-001: A phase 1/2 study of invikafusp alfa
and anti-tumor activity of IMGS-001 in patients with relapsed or refractory (STAR0602), a first-in-class, selective T cell receptor (TCR)-targeting, bi-
advanced solid tumors. First Author: David S. Hong, The University of Texas MD functional antibody-fusion molecule, as monotherapy in patients with
Anderson Cancer Center, Houston, TX antigen-rich tumors resistant to anti-PD(L)-1. First Author: Claire Frances
Background: IMGS-001 is a fully human, dual specific immunoglobulin G1 (IgG1) Friedman, Memorial Sloan Kettering Cancer Center, New York, NY
monoclonal antibody (mAb) that binds both PD-L1 and PD-L2, silencing the entire PD-1 Background: Many patients do not respond to anti-PD(L)-1-based therapies and most
inhibitory circuit, with an engineered fragment crystallizable (Fc) region designed to responders eventually develop resistance. Thus, the development of effective therapies
induce robust antibody-dependent cell-mediated cytotoxicity (ADCC) and phagocytosis for anti-PD(L)-1 resistance is a significant unmet medical need. Invikafusp, a selective,
(ADCP). IMGS-001 mediated killing of PD-L1+ and PD-L2+ tumor and stromal cells can dual T cell agonist targeting Vb6/Vb10 T cells, is being evaluated in START-001: a
reduce the level of multi-modal immune suppression throughout the tumor microen- multicenter Phase 1/2 monotherapy trial in patients with anti-PD(L)1-resistant, antigen-
vironment while catalyzing cross presentation of tumor antigens to the adaptive immune rich (TMB-H, MSI-H/dMMR, or virally associated) solid tumors. The completed Phase 1
system. IMGS-001 also blocks binding of the T cell co-inhibitory receptor PD-1 with its dose escalation of intravenous invikafusp, Q2W, per 3+3 design, identified a recom-
ligands, restoring activation and function to tumor-specific T cells. In addition, IMGS-001 mended Phase 2 dose (RP2D) of 0.08 mg/kg, and demonstrated clinically meaningful
blocks binding of PD-L1 to B7-1, increasing costimulation of tumor-specific T cells. A single-agent anti-tumor activity in patients with anti-PD(L)-1 resistant tumors, including
phase 1a/1b study has been opened to investigate IMGS-001 safety, anti-tumor activity, confirmed partial responses in TMB-H, microsatellite stable, colorectal cancer (CRC)
and pharmacokinetics (PK) in solid tumor patients (Protocol IMGS-001-011; patients with one durable response lasting ~12 months. It promoted potent and selective
NCT06014502). Methods: This multi-center, first-in-human study is enrolling sub- expansion of mainly CD8+ Vb6/ Vb10 T cells with a novel central memory T cell
jects with advanced solid tumors refractory to standard of care therapy. Phase 1a uses a phenotype, and led to ctDNA decrease and expansion of antigen-specific T cells. Based
Bayesian optimal interval (BOIN) dose-escalation design to investigate doses from 0.3- on these results, the US FDA granted Fast Track Designation for invikafusp in TMB-H
15 mg/kg (Q2W). Phase 1b is a two-part design in subjects with PD-L1+ expression $ 5% CRC. Methods: Study design: Using an optimal Simon’s 2 stage design, Phase 2 of
across 5 tumor types: triple negative breast, bladder, gastric/esophageal, colorectal, START-001 is a dose expansion at the RP2D, to further investigate the safety and anti-
ovarian. Part 1 will enroll up to 10 subjects per cohort. Cohorts meeting prespecified tumor activity of invikafusp in 9 cohorts of patients who have the following solid tumors:
efficacy criteria will proceed to Part 2 dose optimization randomly assigning 40 subjects 1) tissue-agnostic, TMB-H; 2) tissue-agnostic, dMMR/MSI-H; 3) CRC (both Ras wild-type
(1:1) between two doses. The primary objective of Phase 1a is to assess IMGS-001 and mutant) TMB-H and/or MSI-H/dMMR); 4) virally associated tumors such as Merkel
safety, and of Phase 1b is to define the pharmacologically optimal dose (POD). Both cell carcinoma, cervical, oropharyngeal, anal, penile, vaginal, and vulvar cancers, or EBV-
study phases will assess tolerability, PK, immunogenicity, and anti-tumor activity in- related solid tumors; 5) metastatic triple-negative breast cancer; 6) platinum-resistant
cluding objective response rate and progression free survival, as well as exploratory epithelial ovarian cancer; 7) metastatic castration-resistant prostate cancer; 8) primary
tissue and serum biomarker analyses. The study will enroll approximately 25 patients in stage IV or recurrent non-small cell lung cancer; and 9) immunogenic tumors (e.g., cSCC,
Phase 1a and up to 250 in Phase 1b. The first two cohorts (0.3 and 1 mg/kg) have melanoma and RCC). Major Eligibility criteria: # 3 lines of prior cancer therapies [anti-
completed without any dose limiting toxicities (DLTs), and cohort 3 (3 mg/kg) is enrolling PD(L)-1s allowed] for advanced or metastatic disease; intolerance to standard therapies
as of the submission date. Clinical trial information: NCT06014502. Research Sponsor: including anti-PD(L)-1s allowed; no liver metastases or adequately treated liver me-
ImmunoGenesis; Cancer Prevention and Research Institute of Texas (CPRIT); Cancer tastases either locally (e.g., by surgery, radiofrequency ablation, or chemoembolization)
Focus Fund. or systemically and stable for 3 months. Primary objective: to further evaluate anti-
tumor activity of invikafusp as monotherapy in each of the above-described 9 cohorts of
patients with anti-PD(L)-1-resistant, unresectable, locally advanced, or metastatic solid
tumors. Primary endpoint: overall response rate (ORR) per iRECIST. The enrollment to
the first three cohorts has begun. Clinical trial information: NCT05592626. Research
Sponsor: Marengo Therapeutics, Inc.

TPS2688 Poster Session TPS2689 Poster Session

A phase 1 first-in-human study of the novel anti-LLT1 antibody (ZM008) ELEPHAS-01, ELEPHAS-02 and ELEPHAS-04: Multi-institutional observa-
alone and in combination with anti-PD1 antibody in patients with advanced tional prospective clinical trials to assess the accuracy of an ex vivo live
solid tumors. First Author: Maloy Ghosh, Zumutor Biologics, Bangalore, India tumor fragment platform for predicting immunotherapy response. First
Background: ZM008 is a first-in-class, fully human, IgG1 monoclonal antibody targeting Author: Hinco J. Gierman, Elephas, Madison, WI
the LLT1 antigen. It disrupts the interaction of LLT1-CD161, an NK-mediated innate Background: Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment. However,
immunity checkpoint. LLT1 expression on tumor cells has been associated with poor existing FDA approved companion diagnostic biomarkers like PD-L1, dMMR/MSI-H and TMB have low
overall survival in multiple solid tumors. Ex vivo experiments with lung and bladder accuracy in predicting response. Ex vivo cytokine profiling of live tumor samples has shown promise
as an improved means of predicting response to PD-1 blockade (Voabil, et al. Nat Med. 2021), but this
cancer biopsies showed significant tumor reduction and immune cell infiltration were approach has been limited to tumor resections given the need for large amounts of tissue. Here we
observed with ZM008 monotherapy. Synergistic anti tumor effects were observed with present three clinical trials that leverage a novel approach using limited tissue from a single core
ZM008 in combination with pembrolizumab. An open-label, phase 1, first-in-human needle biopsy (CNB) (20 gauge or larger). A sequential ex vivo treatment strategy is used, eliminating
study evaluating the safety, tolerability, pharmacokinetics (PKs), preliminary anti-tumor the need for a separate control arm and addressing challenges with tumor heterogeneity, particularly
activity and the Recommended Phase 2 Dose (RP2D) of ZM008 alone and in combination in CNBs where tissue is limiting. Using a specialized instrument, CNBs are cut into live tumor
with pembrolizumab in advanced solid tumors is now ongoing at 3 US sites fragments (LTFs) which are viable in culture and retain the native tumor microenvironment, enabling
(NCT06451497). Methods: The study includes a dose-escalation Part 1 and a dose- cytokine profiling in response to ICI treatment ex vivo. Methods: ELEPHAS-01 (NCT05478538),
ELEPHAS-02 (NCT05520099) and ELEPHAS-04 (NCT06349642) are observational prospective clinical
expansion Part 2. In dose escalation (part 1), ZM008 monotherapy follows 3+3 standard
trials initiated to characterize the accuracy of this approach for predicting ICI response. Over 750
design starting with 0.15 mg/Kg and up to 18 mg/Kg IV Q3W. A staggered parallel arm patients that are being considered for standard of care (SOC) ICI therapy in the metastatic/relapse or
will explore ZM008 in combination with pembrolizumab (200mg Q3W) from dose level 6. neoadjuvant setting will be enrolled (Table). Fresh live CNBs are collected prior to treatment start and
Histologically confirmed advanced or metastatic non-small cell lung, head & neck, processed within 24 hrs enabling prediction of results within 72 hrs of receipt. LTFs are treated using a
pancreatic, biliary, prostate, colorectal, triple negative breast, urothelial, ovarian and strategy where control (IgG) and SOC ICI treatments are performed sequentially on the same tissue in a
diffuse large B cell malignancies with no standard alternative are included. Measurable single well. Changes in the cytokine secretion rates are then compared between ICI and control to
disease by RECIST v1.1, adequate haematological, hepatic and renal functions are characterize immunotherapy response. Additionally, tissue viability and tumor content measurements
required. In Part 2, two or more doses of ZM008 will be used to select RP2D and are used to assess tissue quality. Clinical response is measured using pathologic response in patients
receiving neoadjuvant ICI therapy, while RECIST v1.1 is used in all other patients. The primary objective
indications of interest. Major exclusion criteria include, patients with history of un- of these trials is to determine the platform’s ex vivo accuracy (e.g., sensitivity, specificity) for
controlled brain metastasis, autoimmune disease, pneumonitis, active infections, and predicting clinical response to ICIs and comparing it to the accuracy of PD-L1, dMMR/MSI-H and TMB.
significant cardiovascular diseases. The primary objective is to determine the maximum Clinical trial information: NCT06349642, NCT05478538, NCT05520099. Research Sponsor: None.
tolerated dose (MTD) and RP2D of ZM008. Secondary objectives include PKs, incidence
ELEPHAS-01 (Lung) ELEPHAS-02 (Hoosier) ELEPHAS-04 (Mayo)
and severity of treatment-emergent AEs as per common terminology criteria for adverse
events (CTCAE) v.5.0, immunogenicity, pharmacodynamic changes, and preliminary Setting Metastatic & recurrent Metastatic & recurrent Metastatic, recurrent & neoadjuvant
Tumor type Lung Bladder, kidney, colorectal, Metastatic/recurrent: lung,
anti-tumor activity. Exploratory biomarkers will evaluate pharmacodynamics changes, head and neck, lung, skin, esophageal, cervical, endometrial,
receptor occupancy, immune and cytokine profiling, ctDNA, and transcriptomics. Paired melanoma, endometrial colon, liver, kidney, bladder
Neoadjuvant: breast–TNBC, lung
pre- and on-treatment biopsies will be analysed using immunohistochemistry and the Enrollment as of 26 44 20
spatial distribution of immune and tumor cells in the tumor microenvironment. At the 1/27/2025
time of submission, enrollment of 9 subjects were completed in three dose cohorts with Est. total enrollment 216 216 324
Clinical endpoints RECIST v1.1 RECIST v1.1 RECIST v1.1 & Pathologic response
no reported DLTs. The study is ongoing and open for enrolment at NEXT Oncology (San at surgery
Antonio and Austin sites) and Dana-Farber Cancer Institute, Boston. Clinical trial in-
formation: NCT06451497. Research Sponsor: Zumutor Biologics Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
184s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

TPS2690 Poster Session TPS2691 Poster Session

Phase II basket study to evaluate the tissue-agnostic efficacy of anti-PD1 in Phase 2 trial of TU2218, TGFb-RI, and VEGF-R2 dual inhibitor in combina-
patients with advanced rare tumors: The ANTARES trial. First Author: Camila M. tion with pembrolizumab in patients with biliary tract cancer and head and
Venchiarutti Moniz, Instituto do Câncer do Estado de S~
ao Paulo (ICESP), Universidade de neck cancer. First Author: Do-Youn Oh, Department of Internal Medicine, Seoul
S~
ao Paulo and Instituto D’Or de Pesquisa e Ensino (IDOR), S~ ao Paulo, Brazil National University College of Medicine, Seoul, South Korea
Background: Rare tumors account for 25-30% of all malignancies; however, patients Background: TU2218 is a low molecular weight dual kinase inhibitor highly specific to
(pts) with these cancers are underrepresented in clinical trials. The limited evidence on TGFbR1 and VEGFR2 and has a potential to be an efficacious therapy against cancer growth.
sequential oncologic treatment strategies in this population leads to a poorer prognosis In vitro and in vivo nonclinical studies have shown that TU2218 reduced the growth and
compared to pts with more common malignancies. The predictive role of the tissue- migration/invasion of tumor cells and increased antitumor effects in combination with anti-
agnostic biomarker PD-L1 and the combined positive score (CPS) in determining the PD-1/anti PD-L1 antibodies. To investigate safety and tolerability of TU2218 Phase 1a trial was
efficacy of anti-PD1 therapy remains poorly understood in this population. conducted with 6 dose level escalation (30mg/day → 60mg/day → 105mg/day → 150mg/day
→ 195mg/day → 270mg/day) of TU2218 alone, and it was confirmed that TU2218 was safe
Methods: ANTARES TRIAL (NCT06638931) is a basket phase 2 single-arm multicentric
and tolerated in all dose levels. And to explore the synergistic effect of TU2218 in combination
study to evaluate the efficacy of anti-PD 1 in rare tumors. Key inclusion criteria are
with Pembrolizumab and to decide RP2D Phase 1b trial was conducted with 3 dose level
invasive neoplasia with incidence lower than 6/100.000 people-year, expressing PD-L1 escalation (105mg/day → 150mg/day → 195mg/day) of TU2218 in combination with
with a combined positive score (CPS) $10, ECOG 0-1, measurable disease by RECIST Pembrolizumab in patients with advanced solid tumors. The RP2D of TU2218 was established
v1.1, progression or intolerance to all available treatments for metastatic disease. as 195mg/day in combination with Pembrolizumab, the total 19 patients received the
Patients will receive nivolumab 480 mg intravenously every 4 weeks until disease treatment and most frequently observed TRAE was pruritus and proteinuria, and three Grade 3
progression or for a maximum duration of 12 months. The primary endpoint was the TRAEs (Pruritus, Rash Maculo-Popular, Malaise) were observed. The MTD was not identified
disease control rate (DCR) assessed by RECIST v1.1. Based on Simon’s two-stage design during dose escalation period. The ORR of overall dose levels demonstrated 19%, and DCR was
(DCR under alternative hypothesis . 25%; DCR under null hypothesis #5%), nine about 63%. In particular, 80% DCR was observed in TU2218 195mg/day in combination with
patients were accrued in the first stage. If $1 responses are observed, the trial will Pembrolizumab. The trial was expanded to the specific cancer types, Biliary Tract Cancer and
accrue an additional 16 pts. The study will be considered positive if 4 or more pts achieve Head and Neck Cancer using the established RP2D for Phase 2 trial. Methods: Locally ad-
DCR among 25 pts in the second stage. Considering a drop-out rate of 10%, a sample size vanced unresectable or metastatic biliary tract cancer (BTC) patient whose tumor has pro-
of 28 patients will be needed to attain 90% power and alpha 0.05. Secondary endpoints gressed on/after first line standard anticancer therapy and anti-PD-(L)1 agent-naı̈ve
include progression-free survival, overall survival, response duration, and response time. metastatic or with unresectable, recurrent head and neck squamous cell carcinoma (HNSCC)
Blood samples for circulating tumor DNA, microvessels, and seric immune checkpoint patient whose tumor express PD - L1 (CPS $1) as determined by an FDA-approved test or
biomarkers will be collected at screening, at 8, 20, 32, 44 weeks, and at the final visit. recurrent or metastatic HNSCC with disease progression on or after platinum-containing
chemotherapy are eligible for this non-randomized, open-label multicenter trial. All patients are
Enrollment started in Brazil on June/24 at Instituto do Câncer do Estado de S~ ao Paulo
administered with TU2218 195mg/day (97.5mg BID) on a 2 weeks-on/1 week-off in combi-
(ICESP) and Instituto D’Or de Pesquisa e Ensino (IDOR); 8 sites are planned to open later
nation with Pembrolizumab 200mg IV Q 3 weeks and will be evaluated by investigator-
in 2025. Clinical trial information: NCT06638931. Research Sponsor: FINEP - Finan- assessed objective response rate (ORR) defined as the proportion of patients with a best
ciadora de Estudos e Projetos, Brazil; Reference 1676/22 protocol FADDE222-E1AE- overall response of complete response (CR) or partial response (PR) according to RECIST
45D9-A318- 0C8477BEA1D9. version 1.1. If 2 or less patients out of 22 evaluable BTC patients are observed with CR/PR and
3 or less patients out of 22 evaluable HNSCC patients are observed with CR/PR, this suggests
futility and the cohort may be stopped. Up to 40 BTC patients and up to 36 HNSCC patients are
planned to be enrolled and a dropout rate of up to 10% is expected. As of this abstract
submission date, 14 BTC patients and 8 HNSCC patients have been enrolled. Clinical trial
information: NCT05784688. Research Sponsor: TiumBio., Co., Ltd.; Merck Sharp & Dohme
LLC, a subsidiary of Merck & Co., Inc.

TPS2692 Poster Session TPS2693 Poster Session

5
A multi-center, single-arm, phase II study of pemigatinib combined with IMMUNORARE : A national platform of 5 academic phase II trials coordi-
immune checkpoint inhibitor in FGFR1/2/3 alteration advanced solid tumor. nated by Lyon University Hospital to assess the safety and the efficacy of
First Author: Tao Qin, Phase I Clinical Trial Centre, Sun Yat-sen Memorial Hospital, Sun the immunotherapy with domvanalimab + zimberelimab combination in
Yat-sen University, Guangzhou, China patients with advanced rare cancers—The Anaplastic Thyroid Carcinomas
Background: FGFR mutations are a significant genetic factor contributing to the onset Cohort. First Author: Hélène Lasolle, Lyon 1 University, Lyon, France
and progression of various cancers. FGFRs are aberrantly activated, including single- Background: In patients with rare cancers, there is an unmet medical need for in-
nucleotide variants, gene fusions, and copy number amplifications in human cancer. vestigating innovative therapeutics beyond standard first-line treatment. Indeed, these
FGFR mutation alterations are most commonly observed in urothelial carcinoma, breast diseases are rarely assessed in clinical trials. Anaplastic thyroid carcinomas (ATC)
cancer, endometrial cancer, squamous cell carcinoma of the lung, etc. The preliminary represent 2-3 % of thyroid carcinomas, but are responsible for 15-40% of thyroid cancer
efficacy of FGFR inhibitors in solid tumors has been established ORR ranged from 20% to mortality. Most cases ( . 90%) are diagnosed with advanced unresectable disease. In
30% [1,2,3]. However, the efficacy of FGFR inhibitors as monotherapy in treating FGFR such patients carrying the BRAFV600 mutation (20-30%), the standard 1st-line treat-
mutations of solid tumors has not yet met the clinical needs. Evidence from preclinical ment relies on dabrafenib & trametinib. In patients without BRAF mutation, the 1st line
research suggested that a combination of FGFR inhibition and PD-1 suppression ex- treatment is chemoradiation. There is no validated 2nd line treatment, but immuno-
panded the T-cell clones and caused immunological changes in the tumor microen- therapy combinations seem promising. In DUTHY trial (Durvalumab + tremelimumab),
vironment to enhance anti-tumor immunity and survival [4]. Based on the synergistic the 6month-OS was 65.6% in ATC. Moreover, TIGIT expression increased during ICI
interplay between the FGFR signaling pathway and immune mechanisms, this study treatment, suggesting potential synergistic effects by simultaneous blockade of TIGIT
aims to evaluate the safety and efficacy of combining the FGFR inhibitor pemigatinib and PD-1. Methods: IMMUNORARE5 (NCT06790706) is a platform of 5 single arm phase
plus PD-1 inhibitor to treat solid tumors harboring FGFR mutations. Methods: 1. This II trials testing the efficacy and safety of DOMVANALIMAB (anti-TIGIT) and ZIMBER-
study is a single-arm, multicenter, prospective Phase II clinical trial. Gene testing ELIMAB (anti PD-1) in 5 independent cohorts of rare cancers. The trial, sponsored by
confirms FGFR1/2/3 variants, including but not limited to mutations, fusions/ Lyon University Hospital, is conducted in 15 French centers, led in partnership with the
rearrangements in solid tumors. 2. Patients have not previously used specific small corresponding French national reference centers. The ATC cohort, led in collaboration
molecule multi-target inhibitors of the FGFR pathway, as assessed by investigators, and with the French network ENDOCAN-TUTHYREF ([Link] will enroll
have been treated with immune checkpoint inhibitors. 3. Patients receive pemigatinib 24 patients with either non-mutated BRAF tumours with persistent disease at the first
(13.5 mg QD, orally, 2 weeks on 1 week off, 21 days per-cycle), with immune checkpoint evaluation after chemoradiation or disease progression/relapse after the end of che-
inhibitor therapy (strictly follow instructions ). Treatment should continue until disease moradiation, or with mutated B-RAF tumors in progression after a standard B-RAF
progression or unacceptable toxicity occurs of intolerable toxicities. 4. At least one inhibitor. Patients will receive intra-venous DOMVANALIMAB and ZIMBERELIMAB, every
measurable lesion per RECIST v1.1 criteria. 5. The safety of the study will be assessed three weeks, until disease progression. The primary endpoint is the survival rate at
using the NCI-CTCAE v5.0 criteria. The primary outcome measures: objective response 6 months. The secondary objectives are overall response rate and duration of the
rate (ORR). Secondary outcome measures: disease control rate (DCR), progression-free response, progression-free survival and tolerability. The trial is designed with a two-
survival (PFS); overall survival (OS); safety and quality of life. Clinical trial information: stage Simon design, with early termination for futility (5% one-sided alpha level, 80%
NCT06551896. Research Sponsor: None. power. The treatment would be considered interesting if the survival rate at 6 months is
statistically higher than 25%; 50% is expected. Translational research projects will be
developed aiming at deciphering cellular and molecular mechanisms involved in re-
sponse to treatment. Moreover, data from the prospective database of the ENDOCAN-
TUTHYREF network will be investigated to build a synthetic historical arm representative
of the efficacy of the standard treatments in a similar population of patients. Clinical trial
information: NCT06790706. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 185s

TPS2694 Poster Session TPS2695 Poster Session

Trial in progress: A first-in-human (FIH) phase I study of PTX-912 in patients Phase 1/2 study of tiragolumab and atezolizumab in patients with relapsed
with locally advanced or metastatic solid tumors. First Author: Yan Xing, City of or refractory SMARCB1- or SMARCA4-deficient tumors: PEPN2121. First
Hope Comprehensive Cancer Center, Duarte, CA Author: Mary Frances Wedekind, Pediatric Oncology Branch, National Cancer Institute,
Background: High-dose IL-2 (HD IL-2) received FDA approval for metastatic melanoma National Institutes of Health, Bethesda, MD
(mM) and metastatic renal cell carcinoma (mRCC), but its use is limited by severe Background: The SMARCB1/A4 gene products are core subunits of the SWItch/Sucrose Non
systemic toxicities. While PD-1 blockade has improved overall survival in 20–30% of Fermentable (SWI/SNF) chromatin remodeling complex. Tumors with defects in SWI/SNF are
cancer patients, resistance remains a significant challenge. Notably, HD IL-2 has shown histologically distinct aggressive cancers occurring in children and young adults. SMARCB1/
durable anti-tumor effects in mM and mRCC patients who have progressed on anti-PD-1 A4 deficient tumors, particularly malignant rhabdoid tumor (MRT), atypical teratoid rhabdoid
therapy. Moreover, combining IL-2 with pembrolizumab in mRCC demonstrated a du- tumor (ATRT), poorly differentiated chordoma (PDC), epithelioid sarcoma (ES), and renal
rable response rate of 70%, compared to objective response rates (ORR) of 20% and 33% medullary carcinoma (RMC), have immune cell infiltrates and programmed death ligand 1
with IL-2 and pembrolizumab monotherapy, respectively (Chatzkel et al., Clin Genitourin (PD-L1) expression. Responses to immune checkpoint inhibition (CI) have been observed in
Cancer(2022)). These findings suggest that combining IL-2 receptor (IL-2R) activation SMARCB1/A4 deficient tumors; however, responses are not durable. T cell immunoreceptor
with Ig and ITIM domains (TIGIT) is a novel inhibitory receptor expressed on multiple immune
with PD-1 blockade may be a promising strategy to overcome PD-1 resistance and
cells. TIGIT inhibits T and NK cells by binding to its ligand poliovirus receptor (PVR) and
enhance clinical outcomes. PTX-912 is a novel, first-in-class bifunctional PD-1-proIL-2v
Nectin2 on both tumor and antigen-presenting cells. Utilizing RNAseq data, SMARCB1/A4
fusion protein designed to synergize PD-1 blockade with PD-1-cis-directed IL-2R deficient tumors demonstrate high expression of PVR and Nectin2. Tiragolumab is an an-
agonism specifically within the tumor microenvironment (TME), reducing systemic tibody to the TIGIT receptor. The combination of tiragolumab and atezolizumab has shown
toxicities typically associated with high dose IL-2 therapy. Methods: This first-in-human promising activity in early phase studies, and phase 3 studies are ongoing in multiple adult
(FIH), multi-center Phase I study (NCT06190886) evaluates the safety, tolerability, and indications. Thus, there is rationale that the addition of tiragolumab to CI may also enhance
preliminary efficacy of PTX-912 in patients with locally advanced or metastatic solid response rates in patients with SMARCB1/A4 deficient tumors. Methods: This is a phase 1/2
tumors who have had disease progression on all available standard of care and/or trial of tiragolumab monotherapy (300 mg if # 15 kg; 420 mg if . 15 kg to # 40 kg; 600 mg
refused available standard of care therapies that would confer clinical benefit. Eligible if . 40 kg or $ 18 years) and in combination with atezolizumab (15 mg/kg [max 1200 mg])
patients must have measurable disease per RECIST v1.1 and may have received any if , 18 yrs or 1200 mg if $ 18 years) administered IV on Day 1 of 21-day cycles in patients .
number of prior therapies. Key exclusions include immunodeficiency, unresolved 12 months of age with SMARCB1/A4 deficient tumors. Part A evaluating the safety of
toxicities . Grade 1 per NCI CTCAE from prior therapy, active autoimmune disease, tiragolumab monotherapy in patients , 18 years based on cycle 1 dose limiting toxicities is
primary CNS or leptomeningeal involvement, history of transplant, recent major surgery, complete. Part B estimates the antitumor activity of tiragolumab in combination with
and significant cardiac or pulmonary dysfunction. The study includes dose escalation atezolizumab in 6 histology-specific cohorts (RMC, MRT, ATRT, PDC, ES, and other SMARCB1/
(Part 1a) and dose expansion (Part 1b) cohorts. In Part 1a, seven dose levels (DL1–7) will A4 deficient tumors) and is now open to all eligible age groups. Each cohort is conducted
be tested, with DL1–3 following an accelerated titration design and DL4–7 using a using a 6+4 Simon’s two stage design. Enrollment for each cohort is as follows: Part A 6/6,
standard 3+3 design. The primary objectives are to determine the maximum tolerated Part B RMC 1/6, MRT 1/6, ATRT 4/6, PDC 2/6, ES 3/6, other 6/6. Radiographic imaging central
dose (MTD), optimal biological dose (OBD), and/or the recommended Phase II dose response assessment for the first stage of the “other” cohort is ongoing. Cycle 1 toxicities of
(RP2D) of PTX-912, assessed via dose-limiting toxicities (DLTs). Patients with mela- the combination therapy are monitored in Part B patients , 12 yrs using a Bayesian Optimal
noma, renal cell carcinoma (RCC), non-small cell lung cancer (NSCLC), or other pop- INterval (BOIN) design with a target toxicity of 17%. Secondary objectives are to characterize
ulations identified based on Part 1a data will be enrolled in Part 1b. In Part 1a, patients the pharmacokinetics/anti-drug antibody development and to estimate progression free
survival, overall survival, and duration of response. Enrollment is open at all Pediatric Early
will receive intravenous infusions of PTX-912 every two weeks (Q2W), followed by
Phase Clinical Trial Network sites. Data cutoff: Jan 10, 2025. Clinical trial information:
subsequent cycles with a 28-day DLT observation period. Study enrollment began in
NCT05286801. Research Sponsor: National Cancer Institute; UM1CA22882; Cookie for Kids
June 2024 in the United States at 3 centers. Cohorts 1 to 4 (6 patients) have been Foundation; Genentech, A Member of the Roche Group,; NIH, NCI Intramural research
completed without DLT. Enrollment to cohort 5 is currently ongoing. Clinical trial in- program.
formation: NCT06190886. Research Sponsor: Proviva Therapeutics.

TPS2696 Poster Session TPS2697 Poster Session

A phase 1b study of combined treatment with dupilumab (anti-IL-4Ra) and A phase 1 trial of APX-343A, NOX inhibitor targeting CAF-mediated immu-
cemiplimab (anti-PD-1) in patients with early-stage, resectable NSCLC. First nosuppression, as monotherapy or in combination with pembrolizumab in
Author: Fionnuala Crowley, Division of Hematology & Medical Oncology, The Tisch patients with advanced solid tumors. First Author: Hyesung Shin, Aptabio
Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY Therapeutics, Yongin-Si, South Korea
Background: For resectable stage II/III non-small cell lung cancer (NSCLC), neoadjuvant Background: Cancer-associated fibroblasts (CAFs), a key component of tumor stroma,
chemoimmunotherapy has become standard of care. Patients with Stage I disease (as per promote tumor growth and resistance to anticancer therapy. They contribute to immune
AJCC 8) were excluded from chemoimmunotherapy studies given prior data demonstrating suppression within the tumor microenvironment (TME), with evidence linking CAFs to
no survival benefit from perioperative chemotherapy. However, even patients with Stage immune checkpoint inhibitor (ICI) resistance and T-cell exclusion. Nicotinamide adenine
1A (, 2cm) tumors have a 30% chance of recurrence (Altorki et al, NEJM 2023). Recent dinucleotide phosphate (NADPH) oxidase (NOX), which is clinically upregulated by CAF
research has revealed that tumor-infiltrating myeloid cells express an IL-4 responsive in many human cancers, has been reported to be a critical effector of myofibroblast
transcriptional signature, and IL-4 signaling within monocyte-derived macrophages plays transformation during fibrosis. Inhibiting NADPH oxidases, NOX2 and NOX4, restored
an essential role in NSCLC progression and tumor microenvironment remodeling. cluster of differentiation 8 + T-cell proliferation by reducing reactive oxygen species
Dupilumab, a monoclonal antibody targeting IL-4 receptor alpha (IL-4Ra), is currently (ROS) generation in CAF-induced myeloidderived suppressor cells (MDSCs). A pivotal
approved for treating asthma and allergic rhinitis, and preclinical studies have demon- role of CAFs in regulating monocyte recruitment and differentiation demonstrated that
strated that blocking IL-4 signaling can significantly reduce lung tumor burden by acti- CC-chemokine receptor 2 inhibition and ROS scavenging abrogate the CAF-MDSC axis,
vating dendritic cells and effector T cells to generate a robust immune response against illuminating a potential therapeutic path to reversing the CAF-mediated immunosup-
tumor antigens. These findings are supported by early clinical evidence from a phase 1/2 pressive microenvironment. APX-343A, a selective NOX1, NOX2, and NOX4 inhibitor, has
trial showing that dupilumab can work synergistically with PD-(L)1 inhibition to induce been shown to ameliorate the fibrotic and immunosuppressive properties of CAFs. In
sustained tumor responses in some patients with metastatic NSCLC who had previously CAF-rich tumor mouse models that do not respond to ICIs, APX-343A demonstrated
progressed on immunotherapy. Whether similar synergy would be seen in the pre-operative significant anticancer efficacy by modulating both fibrosis and immunosuppression via
setting in patients with Stage 1 tumors, or patients not suitable for chemoimmunotherapy, NOX inhibition. Methods: This is a Phase 1, open-label, dose-escalation study designed
is not known, though an immunotherapy-alone approach may enable much more brief pre- to assess the safety, tolerability, PK, and preliminary efficacy of APX-343A as mon-
operative treatment given that T cell changes peak at one week in the metastatic setting,
otherapy (Part A) and in combination with pembrolizumab (Part B) in patients with
and prior studies show PD-1 blockade alone can cause robust responses in some patients
advanced solid tumors. The trial aims to determine the MTD and/or RP2D. Part A is a
within only a few weeks. Methods: This Phase 1b/2a single-arm trial will enroll patients
dose-escalation study of APX-343A monotherapy, starting at a dose of 100 mg BID
with early-stage (. T1b), resectable NSCLC. Patients will receive one dose each of
(Cohort 1) and escalating up to 600 mg BID (Cohort 6) until dose-limiting toxicity (DLT) is
dupilumab (600mg SC) and cemiplimab (350mg IV) on day 1, followed by surgical resection
within 15-21 days, with delays beyond 8 weeks considered a delay of surgery. The trial
identified. APX-343A will be administered on a continuous daily dosing schedule in 21-
consists of a 3+3 safety run-in (Phase 1b, up to 6 patients) followed by a Simon’s two-stage day cycles. Part B is a dose-escalation study of APX-343A in combination with
expansion (Phase 2a, up to 24 total patients). The primary endpoints are safety/feasibility pembrolizumab (200mg IV, Q3W). Using the BOIN design, the dose level of APX-343A will
(Phase 1b) and major pathological response rate, defined as #10% viable tumor at re- escalate from 200 up to 600 mg BID without exceeding the MTD. The BOIN design will
section (Phase 2a). Secondary endpoints include time to surgery, pathological complete guide dose escalation based on safety, with decisions made by the Safety Review
response rate, event-free survival, and overall survival. Comprehensive correlative studies Committee (SRC). Dose finding will be conducted independently for Parts A and B. APX-
will characterize the immune response through serial blood sampling (days 1, 4, 8, 15, 343A selective NOX inhibitor has the potential to become an effective treatment option
surgery, and 30 days post-op), matched proteomic and transcriptomic tumor tissue in combination with ICIs for patients with CAF-rich solid tumors that are unresponsive to
analysis (pre-treatment and operative samples), and stool microbiome profiling to identify current immunotherapies. By inhibiting CAF activity in the TME and resensitizing tumors
potential biomarkers of response. Clinical trial information: NCT06088771. Research to cancer immunotherapy, APX-343A offers a promising therapeutic approach. Phase 1
Sponsor: Cancer Research Institute. results are anticipated in Q1 2026. Research Sponsor: Aptabio Therapeutics Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
186s DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY

TPS2698 Poster Session TPS2699 Poster Session

A phase Ib study of a pooled synthetic long peptide mutant KRAS vaccine A phase I study of a pooled synthetic long peptide mutant KRAS vaccine in
combined with balstilimab/botensilimab in metastatic pancreatic cancer patients with pancreatic cystic neoplasms at risk for developing pancreatic
and metastatic MMR-proficient colorectal cancer in the maintenance cancer. First Author: Kai-li Liang, Department of Oncology, Johns Hopkins Sidney
setting. First Author: Kai-li Liang, Department of Oncology, Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, Baltimore, MD
Kimmel Comprehensive Cancer Center, Baltimore, MD Background: Mutant KRAS (mKRAS) is an oncogenic driver expressed in . 90% of
Background: Expressed in . 90% of all patients with pancreatic ductal adenocarcinoma patients with pancreatic ductal adenocarcinoma (PDAC) and the majority of pancreatic
(PDAC) and ~40% in mismatch repair-proficient colorectal cancer (MMRp CRC), the mutated precursors, including . 90% of intraductal papillary mucinous neoplasms (IPMNs) and
oncoprotein KRAS (mKRAS), is an attractive neoantigen vaccine target. Efforts to sensitize pancreatic intraepithelial neoplasia (PanIN) (Kanda et al., 2012). If left untreated, ap-
these immunologically ‘cold’ tumors to immune checkpoint inhibitors have just started to proximately 40-60% of high-risk IPMNs will have malignant transformation (Fonseca et al.,
yield encouraging clinical data with novel agents. In an ongoing pilot study in patients with 2018). mKRAS vaccines have recently demonstrated encouraging results in generating
resected PDAC and metastatic MMRp CRC (NCT04117087), we demonstrated that a pooled mKRAS-specific T cell responses that correlate with clinical benefit in patients with
synthetic long peptide (SLP) mKRAS vaccine in combination with ipilimumab and nivolumab resected PDAC. We previously reported that a mKRAS-targeted Listeria-based vaccine
was safe and well tolerated. This combination induced robust de novo mKRAS-specific given with Treg-depleting agents results in slowing of PanIN progression to PDAC in a
T cells in peripheral blood associated with improved disease-free survival (Haldar et al., murine model (Keenan et al., 2014). Based on these data, we have initiated a clinical trial
2023). Recently, the Fc-enhanced anti-CLTA-4 antibody, botensilimab (bot), in combination testing this vaccine in individuals at high-risk of developing pancreatic cancer. In our first
with balstilimab (bal; anti-PD-1) has been shown to demonstrate clinical activity in met- Cohort [A], we have tested this vaccine in individuals at high-risk due to a known germline
astatic relapsed/refractory MMRp CRC (Bullock et al., 2024). Based on these encouraging mutation or familial predisposition (n = 20). Our current study [Cohort B] aims to determine
data, our study combines mKRAS vaccine with dual checkpoint blockade to assess safety the safety and immunogenicity of a pooled synthetic long peptide (SLP) mKRAS vaccine
and early clinical efficacy in patients with metastatic PDAC and metastatic MMRp CRC in the with poly-ICLC adjuvant in patients with pancreatic cystic neoplasm at risk for developing
maintenance setting. Methods: This is a first-in-human, single-arm, open-label phase Ib PDAC and who are scheduled to undergo surgical resection. Methods: This is a single-
trial evaluating mKRAS vaccine with bal/bot in patients with metastatic PDAC (Cohort A, n = arm, open-label phase I trial evaluating mKRAS vaccine in patients with pancreatic cystic
21) and metastatic MMRp CRC (Cohort B, n = 21). The vaccine consists of SLPs corre- neoplasms at risk for developing PDAC and scheduled to undergo surgical resection (n =
sponding to six common mKRAS alleles: G12D, G12V, G12R, G12C, G12A, G13D admixed 10). The vaccine consists of SLPs corresponding to six common mKRAS mutations: G12D,
with poly-ICLC adjuvant. In the priming phase (Cycle 1) the mKRAS vaccine is given on days G12V, G12R, G12C, G12A, G13D admixed with poly-ICLC adjuvant. A two-dose series of the
1, 8, 15 and 22 along with bal/bot on day 1 and bal on day 15. In the boost phase, (Cycle 2 and mKRAS vaccine is administered at weeks 1 and 2 followed by pancreatic surgery at week 4.
beyond), patients receive bal every 2 weeks and boost vaccines starting on Cycle 4 and every Peripheral blood will be collected pre-vaccination (week 1) and post-vaccination (weeks 4
other cycle for a maximum of 2 years. Eligible patients must have metastatic PDAC or MMRp and 8). Following completion of the treatment phase, patients have the option to continue
CRC and measurable disease per RECIST 1.1 amenable to biopsies at baseline and week 9. annual follow-up visits until study closure. Eligible patients must have clinical, radio-
Patients must have one of the six KRAS mutations contained in the vaccine. Patients must graphic, or histologic evidence of a pancreatic cystic neoplasm with features warranting
have received 4-6 months of 1st line standard chemotherapy without disease progression. surgical resection per the discretion of the treating hepatobiliary surgeon. Co-primary
The primary endpoints are safety and tolerability, 4-month progression free survival (Cohort endpoints include the safety profile per NCI CTCAE v5.0 and maximal percent change of
A), and objective response rate (Cohort B). Secondary endpoints include disease control mutant-KRAS-specific T cells measured by IFNg ELISPOT at weeks 4 and 8 post-
rate, objective response rate (Cohort A), and progression free survival (Cohort B). Correlative vaccination compared to pre-vaccination baseline. Correlative studies of resected
studies will examine T cell receptor (TCR) clonal expansion in peripheral blood and paired specimens will include characterization of the pre-malignant microenvironment and
tumor specimens pre- and post-vaccination by next generation TCR sequencing. Patient mKRAS-specific T cell trafficking post-vaccination. Methods of analyses include bulk RNA
accrual began in October 2024 with the safety run in completed. Enrollment is currently and T cell receptor (TCR) sequencing, spatial transcriptomics, and imaging mass
ongoing. Study drug support provided by Agenus. Trial information: NCT06411691. Clinical cytometry. Patient accrual began in December 2024 and is currently ongoing. Clinical trial
trial information: NCT06411691. Research Sponsor: U.S. Department of Defense; U.S. information: NCT05013216. Research Sponsor: Lustgarten Foundation for Pancreatic
National Institutes of Health. Cancer.

TPS2700 Poster Session TPS2701 Poster Session

A phase 2 basket trial of tarlatamab in patients with advanced DLL3- EXPAND-1, a phase I/II study with ANV600, a novel PD-1 targeted IL-2R-bg
expressing tumors: University of California Lung Cancer Consortium agonist, in monotherapy and in combination with pembrolizumab, in pa-
UCCC-01/UCLA L-10. First Author: Michael Oh, David Geffen School of Medicine tients with advanced solid tumors. First Author: Iphigenie Korakis, Inst University
at University of California, Los Angeles, Los Angeles, CA Du Cancer De Toulouse, Toulouse, France
Background: Delta-like ligand 3 (DLL3) is an inhibitory Notch ligand that is aberrantly Background: ANV600 is a novel PD-1 targeted, interleukin-2 receptor beta/gamma (IL-
expressed on the surface of tumor cells, in particular on those with neuroendocrine 2Rb/g) selective agonist. This bispecific agent comprises two functionally distinct
differentiation. Tarlatamab is a bispecific T cell engager that binds to DLL3 and CD3 to arms: a PD-1 targeting arm consisting of an anti-PD-1 antibody binding to an epitope
promote T cell killing of DLL3-expressing cells. Prior studies of tarlatamab have that does not overlap with pembrolizumab or other PD-1 checkpoint inhibitors and an IL-
demonstrated encouraging antitumor activity and manageable toxicity in patients with 2 receptor (IL-2R) agonistic arm, composed of an interleukin-2 (IL-2)/anti-IL-2 antibody
small cell lung cancer (SCLC; DeLLphi-301) and neuroendocrine prostate cancer (NEPC; fusion protein which selectively signals through IL-2Rb/g. ANV600 is expected to
DeLLpro-300). Meanwhile, DLL3 has been reported to be highly expressed in multiple promote anti-tumor activity by preferentially stimulating and expanding antigen-
tumor types, including in many neuroendocrine neoplasms (NENs) other than SCLC and experienced PD-1+ CD8+ T cells and be combinable with existing anti-PD-1 clinical
NEPC. The role of anti-DLL3 therapies in these cancers has not been established. therapies. ANV600 will be studied as single agent and in combination with pem-
Methods: This is a phase 2, multicenter, open-label, basket study designed to evaluate brolizumab for the treatment of advanced solid tumors. Methods: Study ANV600-001
the efficacy of tarlatamab in patients with DLL3-expressing cancers. Key inclusion (EXPAND-1) is a global, multicenter, open-label, first-in-human Phase I/II study to
criteria include presence of advanced stage disease with progression following $1 prior characterize the safety, tolerability, pharmacokinetics (PK), pharmacodynamics (PD),
line of therapy and positive tumor DLL3 expression by immunohistochemistry (Ventana immunogenicity and antitumor activity of ANV600 administered as a single agent or in
SP347 assay). Patients with de novo SCLC or NEPC are excluded, but all other tumor combination with pembrolizumab in patients with advanced solid tumors. The Phase I
types and NENs are eligible, including large cell neuroendocrine carcinoma and SCLC will determine the maximum tolerated dose (MTD) and/or recommended phase 2 dose
transformed from previously treated NSCLC. Tarlatamab will be administered at an initial (RP2D) of ANV600 administered intravenously every 2 weeks (Q2W) either as single
step-up dose (1 mg on D1 and 10 mg on D8 and D15 of cycle 1) followed by 10 mg every agent or in combination with pembrolizumab in previously treated advanced solid
2 weeks. Treatment will continue until unacceptable toxicity, progressive disease, or tumors. A Bayesian Optimal Interval (BOIN) design will guide the dose escalation to
withdrawal of consent. The study will follow a Simon’s two-stage design: in Stage 1, 10 determine the MTD and/or RP2D. Once the RP2D has been determined, ANV600 will be
patients with tumor DLL3 expression $25% will be enrolled, and the study will be further evaluated as monotherapy and in combination with pembrolizumab in the Phase
stopped if #1 patient achieves an objective response; otherwise, an additional 19 II part of the study for efficacy and safety in PD-1 experienced patients with advanced
patients with tumor DLL3 expression $1% will be enrolled for Stage 2. The primary melanoma, NSCLC and HNSCC. Additional cohorts may be selected based on emerging
endpoint is the objective response rate. Secondary endpoints include safety, progression data. Tumor response will be assessed using RECIST v1.1. Enrolment began in June
free survival, duration of response, and overall survival. Exploratory studies will evaluate 2024, with 10 patients enrolled in the monotherapy arm and 4 in the combination arm. Up
correlation of antitumor activity with tissue and blood-based biomarkers, such as DLL3 to 240 participants will be enrolled in 7 countries: Belgium, France, Germany, the
expression on tumor and liquid biopsies. This study is currently enrolling patients Netherlands, Spain, Switzerland and the USA. Research Sponsor: ANAVEON AG. Clinical
through the University of California Lung Cancer Consortium (UCLCC). Clinical trial trial information: NCT06470763. Research Sponsor: None.
information: NCT06788938. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—IMMUNOTHERAPY 187s

TPS2702 Poster Session TPS2703 Poster Session

Fecal microbiota transplantation combined with sintilimab and SOX as first- A first-in-human phase 1 clinical trial of INI-4001, a novel TLR7/8 agonist, in
line treatment for advanced gastric cancer (FMT-JSNO-01): A prospective, patients with advanced solid tumors. First Author: Shannon Marilee Miller,
multicenter, double-blind, randomized placebo-controlled phase II trial. First Inimmune Corp., Missoula, MT
Author: Wenyu Zhu, Department of Oncology, The Second People’s Hospital of Background: Inimmune has developed INI-4001, a novel TLR7/8 agonist as an im-
Changzhou, the Third Affiliated Hospital of Nanjing Medical University, Changzhou, China munotherapy treatment for cancer. Pre-clinically, the lead formulation of INI-4001 was
Background: Chemotherapy in combination with immunotherapy has emerged as the able to eliminate Lewis Lung Carcinoma (LLC) flank tumors in mice after just two
first-line(1L) standard of care for gastric cancer (GC) patients(pts); nonetheless, the treatments. Moreover, INI-4001 slowed the growth of MC38 and B16F10 tumors and
overall prognosis remains suboptimal. Fecal microbiota transplantation (FMT) holds synergized when combined with anti-PD-1 therapy, leading to an increased cure rate in
promise in modulating the patient’s gut microbiota and immune milieu, thereby aug- both MC38 and B16F10 flank tumors in mice when both drugs were used compared to
menting the efficacy of tumor immunotherapy and enhancing long-term survival either treatment alone. In July of 2024, we dosed our first patient in a Phase 1 clinical
outcomes. We propose to integrate FMT into the regimen of chemotherapy plus im- trial in patients with advanced solid tumors. Methods: INI-4001 will be evaluated in a
munotherapy, aiming to assess its efficacy and safety in pts with advanced GC Phase Ia/Ib, open-label, dose-escalation, and dose expansion study. This study will be
(NCT06405113 ). Methods: FMT-JSNO-01 is a prospective, multicenter, randomized, conducted in two parts: Phase Ia (dose escalation) and Phase Ib (dose expansion).
double-blind, placebo-controlled phase II trial designed to enroll pts with previously Phase Ia will initially seek to establish the MTD or OBD of INI-4001 administered as
untreated, unresectable advanced gastric or gastroesophageal junction adenocarci- monotherapy. Using a BOIN design, we have planned six ascending 1-3-subject cohorts
noma (GAC/GEJAC) that is human epidermal growth factor receptor 2 (HER2) negative. with weekly dosing on continuous 21-day cycles. Imaging shall occur after each 3 cycles,
The physical status score of Eastern Tumor Collaboration Group (ECOG) was [Link] and combination therapy with a checkpoint inhibitor is allowable under certain con-
study will be conducted in more than 15 multidisciplinary treatment centers for GC in ditions after 3 cycles of monotherapy. Combination with checkpoint inhibitor is allowed
[Link] eligible pts were randomly assigned to arm A and arm B. Using a network if the subject has progressed or achieved stable disease according to iRECIST criteria
random system, subjects are randomly assigned in a 1:1 ratio to the experimental group and has a tumor type for which a checkpoint inhibitor is approved. Following identi-
and control group, and competitive random enrollment is conducted at each [Link] fication of the MTD or OBD, Phase 1b allows any dose level at or below the MTD to be
in arm A received fecal microbiota capsule transplantation combined with sintilimab expanded with up to 20 additional subjects to further explore the safety, PK, PD, and
immunotherapy plus S-1 and oxaliplatin (SOX) chemotherapy, while pts in arm B re- preliminary efficacy of INI-4001 alone or as combination therapy. Currently in Phase Ia,
ceived placebo combined with sintilimab plus SOX . If there is no progression of the Cohorts 1, 2, and 3 have been completed without DLT. Enrollment to Cohort 4 will begin in
disease after 4-6 cycles of 1L treatment, both arms of pts will enter the 1L maintenance February 2025. INI-4001 may continue as monotherapy or combination as long as the
treatment stage: S-1 plus sintilimab, until disease progression, intolerance, or death subject receives benefit. Following cessation of INI-4001, patients will be requested to
[Link] primary endpoint of the study is the 2-year overall survival rate (2-year OS participate in long-term follow-up to assess overall survival. Clinical trial information:
rate), with secondary endpoints including median progression-free survival (mPFS), NCT06302426. Research Sponsor: None.
objective response rate (ORR), incidence of adverse events (AEs), diversity of fecal
microbiota, and quality of life (QoL). Additionally, exploratory endpoints will encompass
efficacy prediction markers in the gut microbiota and proteomics. This study began
recruiting pts in June 2024 and is currently ongoing. Clinical trial information:
NCT06405113. Research Sponsor: the 2022 Clinical Research project of Changzhou
Medical Center, Nanjing Medical University; CMCC202201; 2022 Changzhou 8th Batch of
Science and Technology Project (Applied Basic Research); CJ20220086; 2023 Clinical
Research Project of Changzhou Medical Center,Nanjing Medical University;
CMCC202307; 2023 Changzhou Health Commission Science and Technology Project;
QN202320.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
188s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3000 Oral Abstract Session 3001 Oral Abstract Session

DB-1310, a HER3-targeted ADC, in pts with advanced solid tumors: Pre- Phase I study of iza-bren (BL-B01D1), an EGFR x HER3 bispecific antibody-
liminary results from the phase 1/2a trial. First Author: Aaron Lisberg, Division of drug conjugate (ADC), in patients with locally advanced or metastatic non-
Hematology/Oncology, University of California, Los Angeles, Santa Monica, CA small cell lung cancer (NSCLC) with driver genomic alterations (GA) outside
Background: DB-1310 is a novel ADC comprised of a humanized anti-HER3 IgG1 monoclonal of classic EGFR mutations. First Author: Yunpeng Yang, Department of Medical
antibody, cleavable peptide linker, and DNA topoisomerase I inhibitor. Here, we report the pre- Oncology, Sun Yat-Sen University Cancer Center, Guangzhou, China
liminary results of the FIH trial. Methods: This global, multi-center, open-label Ph 1/2a trial in- Background: iza-bren is a first-in-class ADC comprised of an EGFR x HER3 bispecific antibody conjugated
cludes dose escalation and expansion. Pts with advanced solid tumors who had failed standard to a novel topo-I inhibitor payload (Ed-04) via a stable tetrapeptide-based cleavable linker. Safety/efficacy
therapy were enrolled. In Ph1, DB-1310 was planned to be administered at doses from 1.5 mg/kg to data from the phase Ib study are presented, focusing on NSCLC patients (pts) with driver mutations outside
6.5 mg/kg, Q3W, iv, using a 3+3 design, with additional pts enrolled to determine the RP2D. Ph 2a of classic TKI-sensitizing EGFR mutations. Methods: Phase Ib part of this study included the expansion
will include approximately 30-40 pts per cohort to optimize the RP2D and assess efficacy. cohorts, each defined by a pre-specified GA, including EGFR exon 20 insertions, non-classical EGFR
Results: As of Jan 17, 2025, 123 pts were enrolled and treated with DB-1310 monotherapy in Ph1 mutations, mutations in HER2, ALK, ROS1, BRAF (V600E and others), KRAS (G12C and others), SMARCA4,
(ECOG PS 1, 80.5%; White, 39.0%, Asian, 52.8%; NSCLC, 65.0%, EGFRm NSCLC, 37.4%; brain MET (Exon 14), RET, and NTRK. Pts with these GA who progressed on standard targeted therapies (if
metastasis, 17.1%), median prior lines of systemic therapy was 3 (range, 1-11). Of the 42 efficacy- available) and no more than one prior line of chemotherapy were enrolled. iza-bren was given at 2.5 mg/kg
evaluable pts with EGFRm NSCLC, 92.9% had previously received 3rd generation EGFR TKI, 92.9% D1D8 Q3W. Results: As of Dec 5, 2024, a total of 73 NSCLC pts with listed GA were enrolled. Five pts were
still on treatment, but were excluded from the analysis due to insufficient follow-up for the first post-
had received platinum-based chemotherapy. The unconfirmed ORR was 25.5% (95% CI, 17.63,
baseline scan (see table below). Among 7 pts with EGFR exon 20 insertions, 85.7% (6 out of 7) achieved
34.65) across all tumor types and 35.7% (95% CI, 21.55, 51.97) in EGFRm NSCLC. Median PFS was cPR. Among 8 pts with KRAS G12C mutations, 3 cPR and 1 PR pending confirmation were observed.
5.4 months overall and 7.0 months for EGFRm NSCLC. 38 (30.9%) pts experienced $ G3 TRAEs, Efficacy for subgroups will be presented. The most frequent hematologic TRAEs (all grades) were anemia
while 7 (5.7%) had drug-related SAEs. TRAEs led to dose reduction in 14 (11.4%) pts and dis- (87.7%), leukopenia (74.0%), thrombocytopenia (74.0%), and neutropenia (72.6%); the most frequent non-
continuation in 5 (4.1%) pts. No TRAE leading to death was reported. Most common TRAE ( . 20%, hematologic TRAEs were asthenia (42.5%), nausea (41.1%), stomatitis (37.0%), diarrhea (32.9%), and
any grade/$G3) were nausea (36.6%/0.8%), anemia (35.8%/4.1%), neutrophil count decreased alopecia (31.5%). Grade 3 and above TRAEs which were predominantly hematologic in nature, were able to
(34.1%/17.9%), platelet count decreased (31.7%/9.8%), white blood cell count decreased (29.3%/ be effectively managed with standard supportive measure including dose reductions, as demonstrated by
8.9%), decreased appetite (23.6%/0.8%), and vomiting (21.1%/0%). Interstitial lung disease oc- the TRAE leading to discontinuation rate of 2.7%. Only 1 case of G2 ILD was observed. Notably, no iza-bren
curred in 7 pts (5.7%, 6 G1 and 1 G2). PK exposure was increased through dose escalation, with low related death was reported. No new safety signals were observed. Conclusions: In NSCLC pts with these
systemic payload exposure and no accumulation of DB-1310 upon repeated administration. GAs, iza-bren showed promising activity with a manageable safety profile, supporting further evaluation of
Conclusions: DB-1310 showed a manageable safety profile and encouraging antitumor activity in iza-bren in these populations. Clinical trial information: NCT05194982. Research Sponsor: None.
pts with heavily pretreated advanced solid tumors, particularly EGFRm NSCLC. Clinical trial in- EGFR mut
formation: NCT05785741. Research Sponsor: None. exon20ins/non- ALK/ROS1/ KRAS/BRAF/
Total classical HER2 mut RET fusion MET mut SMARCA4
Tumor response by dose (efficacy-evaluable). (N = 68) (N=12) (N=13) (N=19) (N=22) (N=2)
Dose
(mg/kg) 1.5 3 4.5 5.0 5.5 6 Total Prior lines of therapy, 1 (1-5) 1 (1-2) 1 (1-3) 3 (1-5) 1 (1-2) 1 (1-1)
median (range)
All tumors, 3 10 25 53 16 3 110 BOR, n
n PR 31 9 8 5 9 0
uORR, n (%) 0 (0) 1(10.0) 8 (32.0) 13 (24.5) (13.76, 6 (37.5) (15.20, 0 (0) 28 (25.5) (17.63, cPR 24 8 7 3 6 0
(95% CI) (0.00, (0.25, 44.50) (14.95, 53.50) 38.28) 64.57) (0.00, 70.76) 34.65) PR pending confirmation 6 1 1 2 2 0
70.76) SD 25 3 5 10 7 0
DCR, n (%) 3 (100.0) 8 (80.0) 23 (92.0) 42 (79.2) (65.89, 11 (68.8) (41.34, 2 (66.7) (9.43, 89 (80.9) (72.31, PD 9 0 0 3 5 1
(95% CI) (29.24, (44.39, 97.48) (73.97, 99.02) 89.16) 88.98) 99.16) 87.78%) NE [1]
3 0 0 1 1 1
100.00) ORR, % 45.6 75.0 61.5 26.3 40.9 0
EGFRm 0 7 9 16 8 2 42 cORR, % 35.3 66.7 53.8 15.8 27.3 0
NSCLC, n DCR, % 82.4 100.0 100.0 78.9 72.7 0
uORR, n (%) - 1 (14.3) 4 (44.4) (13.70, 5 (31.3) (11.02, 5 (62.5) (24.49, 0 (0) 15 (35.7) (21.55, mDOR (mo) (95% CI) 7.0 (5.6, NR) NR (5.6, NR) 5.7 (4.2, NR) 4.5 (2.7, NR) NR (NR, NR) /
(95% CI) (0.36, 57.87) 78.80) 58.66) 91.48) (0.00, 84.19) 51.97)
mPFS (mo) (95% CI) 6.7 (4.1, 11.2) NR (6.9, NR) 8.4 (2.1, NR) 2.8 (1.3, 4.1) 6.7 (1.5, NR) 1.4 (1.3, NR)
DCR, n (%) - 6 (85.7) (42.13, 9 (100.0) (66.37, 14 (87.5) (61.65, 7 (87.5) (47.35, 2 (100.0) 38 (90.5) (77.38,
(95% CI) 99.64) 100.00) 98.45) 99.68) (15.81, 97.34) Note:
100.00)
[1]Including pts w/o post-baseline scan.

3002 Oral Abstract Session 3003 Oral Abstract Session

Phase I study of iza-bren (BL-B01D1), an EGFR x HER3 bispecific antibody- Safety and efficacy of TQB2102, a novel bispecific anti-HER2 antibody–
drug conjugate (ADC), in patients with locally advanced or metastatic small drug conjugate, in patients with advanced solid tumors: Preliminary data
cell lung cancer (SCLC). First Author: Yan Huang, Department of Medical Oncology, from the first-in-human phase 1 trial. First Author: Rui-Hua Xu, Department of
Sun Yat-Sen University Cancer Center, Guangzhou, China Medical Oncology, University Cancer Center State Key Laboratory of Oncology in South
Background: iza-bren is a first-in-class ADC comprised of an EGFR x HER3 bispecific China, Collaborative Innovation Center of Cancer Medicine, Guangzhou, China
antibody conjugated to a novel topo-I inhibitor payload (Ed-04) via a stable tetrapeptide- Background: TQB2102 is an antibody-drug conjugate (ADC) comprised of a recombinant, hu-
based cleavable linker. Unlike existing ADCs targeting general tumor antigens, iza-bren manized anti-human epidermal growth factor receptor 2 (HER2) bispecific antibody conjugated
uniquely targets the EGFR and HER3 pathways, which are implicated in the aggressive to a topoisomerase I inhibitor via an enzyme-cleavable linker. The bispecific antibody component
biology of SCLC. It has shown promising clinical activity and a manageable safety profile can target both extra-cellular domains II (pertuzumab binding site) and IV (trastuzumab binding
site) of HER2. We conducted a multicenter, dose escalation and expansion first-in human (FIH)
in patients (pts) with advanced or metastatic solid tumors. Results for safety/efficacy phase 1 study of TQB2102 in advanced solid tumors. Methods: In the dose escalation phase,
from this phase I study in SCLC pts are presented. Methods: Pts with locally advanced eligible patients (pts) with advanced solid tumors whose disease had progressed after standard
or metastatic SCLC who had progressed on prior systemic therapies were enrolled and systemic treatments, were enrolled in a 3+3 dose escalation study of TQB2102(1.5, 3, 4.5, 6, 7.5 or 9
treated at 2.0, 2.5 mg/kg D1D8 Q3W, or 4.5, 5.0 mg/kg D1 Q3W. Tumor scans were done mg/kg) IV, every 3wks (Q3W). In the dose expansion phase, pts with HER2 positive cancers and
every 6 weeks. Efficacy was evaluated in the overall cohort and specific subgroups, HER2 low (HER2 1+ or HER2 2+ and FISH negative) metastatic breast cancer (MBC) received the
with a particular focus on pts with limited prior treatment exposure. Results: As of Dec selected recommended phase 2 dose (RP2D). The primary objectives were to evaluate the safety
5, 2024, a total of 58 SCLC pts were enrolled. All pts who received at least one dose of iza- and tolerability, dose limiting toxicities (DLTs) and maximum tolerated dose (MTD) of TQB2102.
bren are included in the analysis. The median follow-up was 16.4 mo, ORR was 55.2%, Results: As of October 1, 2024, 181pts (41 pts in dose escalation phase and 140 pts in dose
confirmed ORR was 44.8%, median PFS was 4.0 mo, and median OS was 12.0 mo. expansion phase) were enrolled from 12 centers. Most common tumor types included MBC (N = 80),
Colorectal cancer (N = 37) and Gastric cancer (N = 23).Twenty-five (31%) MBC received prior anti-
Among the 52 pts at 2.5 mg/kg, 20 pts received only 1 prior line of PD(L)-1 and PBC
HER2 ADCs, including 21 pts received T-DM1, 8 pts received [Link] median duration of
combination treatment. In this subgroup, ORR was 80.0%, confirmed ORR was 75.0%, follow-up was 8.15 months. TQB2102 was well-tolerated with no DLTs occurred and MTD was not
median DOR was 5.6 mo, median PFS was 6.9 mo, and median OS was 15.1 mo. The reached. The most common (occurring in $5%) grade $3 AEs were neutrophil count decrease
most frequent hematologic TRAEs (all grades) were anemia (84.5%), leukopenia (74.1%), (21.7%), WBC count decreased (10.6%), anemia (8.9%), platelet count decreased (6.1%), diarrhea
thrombocytopenia (72.4%), and neutropenia (70.7%); the most frequent non- (5.0%). Only one patient had grade 2 interstitial lung disease (ILD) until the cutoff date. 6 or 7.5mg/
hematologic TRAEs were asthenia (41.4%), hypoalbuminemia (39.7%), stomatitis kg was selected for dose expansion. Objective response rate (ORR) per RECIST v1.1 was 41.2% (68
(34.5%), nausea (31.0%), and vomiting (31.0%). Grade 3 and above TRAEs which were partial responses [PR]) in 165 responses evaluable pts who had $1 response assessment.
predominantly hematologic in nature, were able to be effectively managed with standard Surprisingly, 7 pts reached PR in 10 HER2+ MBC pts with brain metastases, one of whom the brain
supportive measure including dose reductions, as demonstrated by the TRAE leading to metastatic lesions reached complete response after 4 cycles of treatment. This trial is ongoing
discontinuation rate of 12.1%. Two infection-related deaths (1 respiratory failure, 1 now. Conclusions: TQB2102 is well tolerated with promising anti-tumor activity in pts with HER2-
expressing cancer. These early signs of activity support a phase 3 trial in patients with HER2-low
gastrointestinal infection) associated with iza-bren were reported. No ILD was observed.
MBC that has been initiated (NCT06561607). Clinical trial information: NCT05735496. Research
No new safety signals were identified. Conclusions: In SCLC pts, iza-bren has dem- Sponsor: Chia Tai Tianqing Pharmaceutical Group Co., Ltd.
onstrated an encouraging efficacy with a manageable safety profile. Notably, the high
confirmed response rate of 75% in pts with limited prior treatment underscores its 6mg/kg and above ORR(%) DCR(%) 6-months PFS rate, (%)
potential as a novel therapeutic option for SCLC, a disease with limited therapeutic HER2 positive MBC (N=39) 51.3 84.7 87.0
advancements over decades. The phase III study of iza-bren in SCLC pts who received 1 HER2 low MBC (N=33) 51.5 87.9 63.0
HER2 3+ colorectal cancer (N=23) 34.8 87.0 88.4
prior line of PD(L)-1 and PBC combination treatment is ongoing (NCT06500026). Clinical HER2 positive gastric cancer (N=10) 70.0 90.0 90.0
trial information: NCT05194982. Research Sponsor: None. HER2 positive Other (N=5) 60.0 100.0 NE

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 189s

3004 Oral Abstract Session 3005 Oral Abstract Session

212
Efficacy and safety of the DLL3/CD3 T-cell engager obrixtamig in patients [ Pb]VMT-a-NET therapy in somatostatin receptor 2 (SSTR2) expressing
with extrapulmonary neuroendocrine carcinomas with high or low DLL3 neuroendocrine tumors (NETs): Dose-limiting toxicity (DLT) observation
expression: Results from an ongoing phase I trial. First Author: Jaume Capdevila, participants after 1 year follow-up and preliminary report for expansion
Department of Medical Oncology, Vall d’Hebron University Hospital & Vall d’Hebron participants. First Author: Thorvardur Ragnar Halfdanarson, Mayo Clinic Compre-
Institute of Oncology, Barcelona, Spain hensive Cancer Center, Rochester, MN
Background: Delta-like ligand 3 (DLL3) is highly expressed in neuroendocrine carcinomas (NEC). Background: [212Pb]VMT-a-NET, a next generation 212Pb-based, SSTR2-targeted alpha-
Obrixtamig (BI 764532) is a DLL3/CD3 IgG-like T-cell engager that targets DLL3-positive (DLL3+) particle radiopharmaceutical therapy (RPT), was designed to achieve superior bio-
tumors. NCT04429087 is an ongoing, Phase (Ph) I dose-escalation trial of obrixtamig in patients (pts) distribution via optimized tumor uptake and retention and rapid renal clearance. Results
with DLL3+ pulmonary and extrapulmonary NEC (epNEC), who had failed to respond to standard
reported here are from the Phase 1/2a first-in-human study [NCT05636618].
treatment (Tx). This analysis examined the efficacy and safety of obrixtamig in pts with epNEC with
high vs low DLL3 expression. Methods: Obrixtamig was given IV in 4 dose-escalation regimens (R): Methods: Safety, pharmacokinetics, dosimetry and efficacy using RECIST v1.1 were
RA (fixed dose q3w); RB1 (fixed dose qw); RB2 (step-up dose, then qw) and RB3 (step-up dose, then investigated in the treatment of participants with [212Pb]VMT-a-NET who had SSTR2-
qw for 3 weeks, then q3w), until disease progression or unacceptable toxicity. Efficacy was assessed expressing, well-differentiated adult NETs of any grade. Participants received $1 prior
through objective response rate (ORR) and disease control rate (DCR) using RECIST v1.1. Results are therapy. No prior peptide receptor radionuclide therapy was allowed. The trial has
reported for pts who received obrixtamig RB2 or RB3, categorized as having high vs low DLL3, using a multiple dose cohorts including 92.5 MBq (2.5 mCi, cohort 1) and 185 MBq (5 mCi, cohort
threshold of $50% of tumor cells stained with an investigational antibody for DLL3 (SP347, Roche 2) of administered activity based on a Bayesian modified toxicity probability interval
Diagnostics). Results: As of June 21, 2024, 60 pts with epNEC were included (gastro- (mTPI-2) design. Up to 8 participants per cohort were treated with 4 doses of [212Pb]
enteropancreatic [GEP]: 45.0%, genitourinary [GU]: 30.0%, other/unknown primary site: 25.0%); 30 VMT-a-NET for DLT observation. Cohort 2 enrollment was expanded to further define the
each DLL3-high and DLL3-low. Mean age: 63.9 years in DLL3-high; 59.1 in DLL3-low pts. Baseline
characteristics were well-balanced across DLL3 groups. All pts had received prior systemic therapy;
safety and efficacy profile at this dose level. Results: Nine (9) gastroenteropancreatic
30.0% of DLL3-high and 50.0% of DLL3-low pts had received . 2 lines of prior Tx. Efficacy data are NET participants were enrolled into cohorts 1 and 2 for DLT observation. The ninth of
shown in the Table. After obrixtamig Tx, pts with high DLL3 expression had greater ORR, DCR, and these participants was enrolled more than 1 year prior to presentation of these data.
duration of response (DoR) than DLL3-low pts. Responses were seen most frequently amongst pts Among these participants at the time of abstract submission, no DLTs were observed,
with DLL3-high GEP (50.0%) or GU (60.0%) epNECs. Seven DLL3-high pts are still receiving Tx. Most and there were no grade 4, 5 or serious adverse events (SAEs). Specifically, no renal
treatment-related AEs (TRAEs) were mild to moderate for both groups (Table). insufficiency or dysphagia were observed. Hematologic AEs were low grade and few in
Conclusions: Analyses from this ongoing Ph I study show greater obrixtamig efficacy in patients number. No treatment discontinuations due to AE occurred. Three (3) of the 7 cohort 2
with epNEC with high DLL3 expression compared with low DLL3 expression, with a manageable participants enrolled for DLT observation achieved investigator-assessed partial re-
safety profile that is comparable across both groups. The ORR of 40.0% and median DoR of
sponses (PRs). Two PRs were unconfirmed at the time of abstract submission. Durable
7.9 months in heavily pretreated epNEC tumors with DLL3 high expression are encouraging, and
support further development of obrixtamig for this subgroup. Clinical trial information: progression-free survival (PFS) was consistently observed. More than 15 additional
NCT04429087. Research Sponsor: Boehringer Ingelheim. participants were enrolled in the cohort 2 expansion. Preliminary data for these patients
will be reported at the congress. Conclusions: [212Pb]VMT-a-NET is a well-tolerated,
Efficacy/safety parameter DLL3-high (n=30) DLL3-low (n=30) next generation RPT showing signs of clinical activity at early dose-levels in this phase
ORR, % (95% CI) 40.0 (24.6–57.7) 3.3 (0.6–16.7) 1/2a study. Based on these clinical data, further dose-escalation and development of
DCR, % (95% CI) 66.7 (48.8–80.8) 26.7 (14.2–44.4) this promising therapy are warranted. Clinical trial information: NCT05636618. Research
Median DoR (95% CI), months 7.9 (6.2–NC) 2.8 (NC–NC)
TRAEs, all G/G ‡3, (%) 100.0/23.3 90.0/20.0 Sponsor: Perspective Therapeutics, Inc.
Cytokine release syndrome, all G/G ‡3, (%) 70.0/3.3 60.0/3.3
Neurotoxicity, including immune effector 16.7/6.7 10.0/3.3
cell-associated neurotoxicity syndrome*, all G/G ‡3, (%)
*Evaluated with a customised MedDRA query.
CI, confidence interval; NC, not calculable.

3006 Oral Abstract Session 3007 Oral Abstract Session

Comprehensive genomic profiling of matched ctDNA and tissue from pa- Ultra-sensitive pan-cancer molecular residual disease assessment using
tients with less common cancers enrolled in but not eligible for a treatment whole-genome sequencing-based personalized ctDNA panel: Initial results
arm of the NCI-MATCH trial. First Author: Biswajit Das, Frederick National Labo- from the MONSTAR-SCREEN-3 project. First Author: Tadayoshi Hashimoto,
ratory for Cancer Research, Frederick, MD Department of Gastroenterology and Gastrointestinal Oncology, National Cancer Center
Background: During NCI-MATCH (NCT02465060) clinical trial screening, 5961 advanced Hospital East, Kashiwa, Japan
cancer patients underwent next-generation sequencing to assess eligibility. About 60% of Background: While circulating tumor DNA (ctDNA) demonstrates promise as a molecular
these patients had less common tumors (i.e., cancers other than colon, rectal, breast, non- residual disease (MRD) biomarker, its clinical implementation has been primarily limited to
small cell lung, or prostate). Most patients lacked a study eligible mutation of interest (MOI) tumors with favorable ctDNA shedding characteristics. We are evaluating an ultra-sensitive
and thus didn’t receive a trial therapy. Analysis of plasma samples from these patients may whole-genome sequencing (WGS)-based MRD assay in the MONSTAR-SCREEN-3 study to
illuminate circulating tumor DNA (ctDNA) profiles, potentially guiding ctDNA testing for establish a comprehensive pan-cancer MRD platform inclusive of traditionally low-shedding
clinically relevant mutations in less common cancer types. Here we report the molecular tumors. Methods: MONSTAR-SCREEN-3, a prospective multicenter study targeting 1,100
profiles of ctDNA and matched tumor from a subset of the NCI-MATCH screened patients. patients with solid tumors undergoing curative-intent treatment in the definitive cohort,
Methods: Comprehensive genomic profiling of ctDNA (from blood collected at enrollment) utilizes personalized panels constructed via Precise MRD (Myriad Genetics). These panels
was performed using the TSO500 ctDNA v2 assay (523-genes) and sequenced on the incorporate up to 1,000 tumor-specific alterations identified through WGS of matched tumor
Illumina NovaSeq 6000. Matched tumor was sequenced with the Oncomine Comprehensive tissue, including both short variants and insertion-deletions. Serial plasma samples were
Assay v2, a 143-gene panel. Positive percent agreement (PPA) between mutations of collected at baseline, post-neoadjuvant chemotherapy (when applicable), 1-month post-
interest (MOI) identified in plasma ctDNA and tissue-based screening was calculated with surgery, quarterly in year 1, and biannually thereafter up to 2 years. The assay perfor-
tumor tissue as referent (PPAref_tumor). Results: We tested 2253 patients from the less mance was evaluated across multiple cancer types for ctDNA detection and recurrence
common tumor cohort. 2194 samples were evaluable with 98.6% pass and 1.4% failure monitoring. Results: As of December 2024, 114 patients across 15 cancer types were enrolled,
rates. A subset of five tumor histologies with larger representation (n . 35) in sample size including colorectal (n = 33), gastric (n = 22), head and neck (n = 13), renal cell (n = 10),
esophageal (n = 8), and pancreatic (n = 7) cancers. Treatment strategies included upfront
were further analyzed: cholangiocarcinoma (CCA, n = 90), small cell lung cancer (SCLC, n =
surgery (n = 76) and neoadjuvant chemotherapy (n = 38). The median follow-up time was
59), adenocarcinoma of the esophagus (EAC, n = 37), adenocarcinoma of the pancreas
2.4 months (range, 0.5–7.7). WGS analysis identified a median of 6,089 panel-eligible al-
(PDAC, n = 232), and salivary gland cancer (SGC, n = 47). Overall, PPAref tumor was 83.4%
terations per patient (range: 214-14,112), with high variants counts observed in a deficient
(range: 76.5%-97.9%) in these five histologies. In patients with concordance , 75%, median mismatch-repair colorectal cancer, enabling comprehensive personalized panel design.
tumor fraction (as determined by maximum somatic allele frequency) was much lower Customized panel creation was successful in 69/71 patients (97.2%) across 8 cancer types,
(0.37%) than for specimens with concordance . = 75% (6.49%). The most frequently with two pancreatic cancer cases deferred to surgical specimens due to insufficient variants in
mutated genes identified in CCA were TP53, KRAS, and IDH1; in SCLC were TP53 and RB1 FNA samples. The assay demonstrated 100% baseline sensitivity (41/41), detecting tumor
loss; in EAC were TP53, KRAS, and ERRB2 amplification; in PDAC were TP53 and KRAS; and fractions ranging from , 0.001% to 45.2% across all cancer types, including traditionally low-
in SGC was TP53. Additionally, there were several clinically relevant mutations detected shedding tumors. Post-operative 1-month MRD assessment revealed 35.7% positivity (10/28),
only in ctDNA such as IDH1 for CCA; and BRAF, TP53, and PIK3CA in several histologies. with tumor fractions ranging from , 0.001% to 0.27%. Two MRD-positive patients developed
Microsatellite instability, as measured only in ctDNA, was most prevalent in SCLC, followed radiological recurrence with lead times of 2.5 and 3 months before conventional imaging
by CCA. Conclusions: Concordance of rare tumors in the NCI-MATCH trial is 83.4% in the detection. Conclusions: These interim results demonstrate successful pan-cancer imple-
representative histologies analyzed, which is similar to concordance of clinically relevant mentation of WGS-based personalized ctDNA detection, achieving universal baseline sensi-
MOIs in common cancer studies. Liquid biopsy may be a viable screening option for tivity and ultra-sensitive MRD detection across tumor types, including those traditionally
matching targeted therapies in clinical trials, especially when a tumor biopsy is not practical challenging to assess. Updated molecular and clinical outcome data will be presented. Clinical
or evaluable. The detection of some mutations in ctDNA only may suggest the presence of trial information: UMIN000053975. Research Sponsor: SCRUM-Japan Funds (http:// www.
tumor heterogeneity in multiple lesions in patients with less common cancers. Research scrum -japan. ncc. go. jp/ index. html); Myriad Genetics.
Sponsor: National Cancer Institute, National Institutes of Health; Illumina Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
190s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3008 Oral Abstract Session 3009 Rapid Oral Abstract Session

Circulating tumor DNA (ctDNA) in patients with stage 2/3 HR+HER2- Initial phase 1 dose escalation data for emiltatug ledadotin (Emi-Le), a novel
negative breast cancer (BC) treated with neoadjuvant endocrine therapy B7-H4-directed dolasynthen antibody-drug conjugate. First Author: Erika P.
(NET) in the I-SPY2 endocrine optimization pilot (EOP) trial. First Author: Silver Hamilton, Breast Cancer Research Program, Sarah Cannon Research Institute, Nashville,
Alkhafaji, University of California, San Francisco, San Francisco, CA TN
Background: Numerous studies have demonstrated the prognostic value of ctDNA analysis after neoadjuvant Background: B7-H4 is a transmembrane protein over-expressed in breast (BC), ovarian
chemotherapy for early-stage high-risk breast cancer. Few studies have characterized ctDNA in pts receiving (OC), endometrial (EC), and adenoid cystic carcinoma type 1 (ACC-1) cancers, with limited
NET for early-stage hormone receptor positive (HR+)/HER2- BC that is predicted to benefit less from che-
motherapy. Methods: Cell-free DNA (cfDNA) was isolated from 432 plasma samples from 108 pts enrolled in the expression in healthy tissues. Emi-Le (XMT-1660) is a B7-H4-directed Dolasynthen ADC
I-SPY2 EOP trial. Pts had Stage 2/3 HR+/HER2-, MammaPrint low or high risk 1 BC. Pts were randomized to one designed with a proprietary auristatin F-HPA microtubule inhibitor payload with controlled
of 7 neoadjuvant-based treatment arms including arms containing AI, Z-endoxifen, Lasofoxifene, vepdegestrant bystander effect. Methods: The Phase 1 trial is investigating Emi-Le monotherapy in adult
(ARV-471), and Abemaciclib. Pts were treated for 6 months prior to surgery. Blood was collected at baseline (T0), patients (pts) with advanced/metastatic TNBC, HR+/HER2- BC, OC, EC and ACC-1. In dose
3 weeks (T1), 12 weeks (T2), and 6 months (T3). A personalized ctDNA test (Signatera) was designed to detect up escalation, eligible pts received Emi-Le at doses of 7.2-115 mg/m2 per cycle, with all
to 16 patient-specific mutations (from whole-exome sequencing of pretreatment tumor) in cfDNA by ultra-deep
sequencing. The chi-square test was used to assess associations between categorical variables, and the collected data informing the recommended doses for the expansion (EXP) portion of the
Wilcoxon rank-sum test was used to evaluate differences in medians. Results: ctDNA information was available trial. Tumors were evaluated retrospectively for B7-H4 expression by IHC, with the pre-
for 101 patients at baseline (T0) (Table 1). At T0, 36 (35.6%) patients were ctDNA-positive. 23/36 (63.9%) liminary high cutoff set at TPS$70. Results: As of December 13, 2024, 130 pts were
became ctDNA-negative and 13/36 (36.1%) remained ctDNA-positive. At T0, 65/101 patients (64.4%) were dosed. Across all tumor types, median age of pts was 55; median 4.5 prior lines of therapy
ctDNA-negative. Of these, 57 (87.7%) remained negative, while 8 (12.3%) became ctDNA-positive at T1 before
(range 0-15). B7-H4 status was evaluated for 103 pts, with 44% determined to be B7-H4
reverting to ctDNA-negative. A higher percentage of ctDNA-positive patients at T0 were cN+ compared to ctDNA-
negative pts (p = 0.036, 64% vs. 40%). Additionally, ctDNA-positivity at T0 was strongly associated with higher TPS high. Overall, Emi-Le was generally well tolerated. The most common TRAEs were
Ki67 (p = 0.03) and larger functional tumor volume by MRI at baseline (p = 0.03). T3/T4 and high-grade tumors at transient AST increase (38%, G3 14%), proteinuria (31%, G3 9%), nausea (29%, G3 1%) and
baseline were also associated with having ctDNA positivity at baseline, though this was not statistically fatigue (28%, G3 0%). The only G3 TRAEs in $5% of pts were AST increase and proteinuria.
significant (p = 0.34 and 0.058, respectively). Conclusions: In this study of pts with Stage 2/3 HR+ HER2- BC No G4 or 5 TRAEs were reported. No observed dose-limiting treatment-related neutropenia,
with largely MammaPrint low risk signatures, over one-third of pts had detectable ctDNA at baseline. Detectable
neuropathy, ocular toxicity, interstitial lung disease or thrombocytopenia. TRAEs leading to
ctDNA at baseline was associated with cN+ disease, larger FTV, and high baseline Ki67. The majority of pts with
positive ctDNA at baseline cleared the ctDNA on NET. Clinical trial information: NCT01042379. Research discontinuation were observed in 2.3% of pts. Clinical activity was correlated with both
Sponsor: NIH/NCI ctDNA/MR; Breast Cancer Research Foundation; Give Breast Cancer the Boot; Quantum Leap dose and B7-H4 expression. For pts treated with doses ranging from 38.1–67.4 mg/m2 per
Healthcare Collaborative. cycle (intermediate dose range), the confirmed ORR in evaluable pts with high B7-H4
Clinicopathologic characteristics of EOP patients. expression was 23% (6/26), including a 23% (3/13) confirmed ORR in evaluable pts with
All Patients TNBC, with all 13 pts having previously received at least one topoisomerase-1 inhibitor
(N=101) (topo-1) ADC. At doses $76.2 mg/m2 per cycle (high dose range), the confirmed ORR in
Age at screening 55 (27-80) years* evaluable pts with high B7-H4 expression was 22% (2/9), with 78% (7/9) having $30%
Clinical N Stage
Node+ (cN+) 49 (48.5%) reduction in target lesions. Of the 8 pts with confirmed responses at doses $38.1 mg/m2,
Node- (cN-) 52 (51.5%) 5 had reduction in target lesions . 60%, including 1 CR. All 4 pts with high B7-H4 ex-
SET status
High 84 (83.2%) pression treated at the initial EXP dose of 67.4mg/m2 Q4W had tumor reductions and were
Low 14 (13.9%) on treatment with durations of $16 weeks as of data cutoff. Conclusions: Based on the
Missing 3 (2.97%)
MammaPrint (MP) risk** initial reported data, Emi-Le appears to have encouraging clinical activity and tolerability
High risk 1 (H1) 15 (14.9%) in a heavily pretreated population. Further clinical development is ongoing in the EXP
Low risk 86 (85.1%)
portion of the trial at a dose of 67.4 mg/m2 Q4W in pts with advanced/metastatic TNBC
*Mean(min-max). who have received 1-4 prior lines of systemic therapy, including at least one topo-1 ADC.
**H1: MP score between 0 and -0.57; low: MP score between 0 and 0.355.
Dose exploration is ongoing to identify a potential second higher EXP dose. Clinical trial
information: NCT05377996. Research Sponsor: Mersana Therapeutics.

3010 Rapid Oral Abstract Session 3011 Rapid Oral Abstract Session
Sigvotatug vedotin (SV), an investigational integrin beta-6 (IB6)–directed BMS-986504 in patients (pts) with advanced solid tumors with homozygous
antibody‒drug conjugate (ADC), and pembrolizumab combination therapy: MTAP deletion (MTAP-del): Clinical update and first report of pharmacoki-
Initial results from an ongoing phase 1 study (SGNB6A-001). First Author: netics (PK) and pharmacodynamic (PD) analyses from CA240-0007. First
Kartik Sehgal, Dana-Farber Cancer Institute, Boston, MA Author: Kathryn C. Arbour, Memorial Sloan Kettering Cancer Center, New York, NY
Background: IB6, a tumor-associated membrane protein, is overexpressed in many solid tumors, including Background: BMS-986504 selectively binds to the PRMT5-MTA complex, which represents a synthetic lethal
non-small cell lung cancer (NSCLC) and head and neck squamous cell carcinoma (HNSCC). SV, an IB6- target in MTAP-del cancer cells, while sparing MTAP–wild-type cells. In the first-in-human phase 1/2 CA240-
directed ADC, demonstrated encouraging antitumor activity and manageable safety as a monotherapy in 0007 study in advanced, unresectable or metastatic solid tumors with homozygous MTAP-del, BMS-986504
patients (pts) with advanced NSCLC in SGNB6A-001, an ongoing phase 1 study (Peters, ASCO 2024). Due to was found to be well tolerated and demonstrated antitumor activity in multiple tumors. Here, we report clinical
immunogenic cell death induction and innate immune system activation, SV activity may be enhanced when results and the first PK and PD analyses of BMS-986504 from the dose escalation and expansion phases of
combined with pembrolizumab (P; SV+P). We report initial results of SV+P in pts with advanced solid tumors. CA240-0007. Methods: Pts with measurable/evaluable disease and no available treatment (Tx) with curative
Methods: SGNB6A-001 (NCT04389632) is an open-label, multicenter, dose-escalation and dose-expansion intent were enrolled; 7 doses were evaluated (50 to 800 mg) in dose escalation. Objective response rate (ORR),
phase 1 study evaluating the safety, pharmacokinetics (PK), and antitumor activity of SV. Part C is evaluating disease control rate (DCR), duration of response (DOR), time to response (TTR), safety, PK, PD, including
safety of SV+P in pts with advanced solid tumors; part D is currently enrolling to evaluate SV+P in treatment- plasma SDMA were assessed. Results: As of 2 Dec 24, 152 heavily pretreated pts were enrolled across all
naive pts with locally advanced, unresectable, or metastatic NSCLC and HNSCC. Pts receive SV 1.8 mg/kg by doses: NSCLC (n = 34), PDAC (n = 41), cholangiocarcinoma (n = 12), and mesothelioma (n = 12) were the most
adjusted ideal body weight IV Q2W and P 400 mg IV Q6W. Primary endpoint is safety; secondary endpoints common tumor types. With a median f/u of 9.0 mo (95% CI 7.6–9.9), continued durable antitumor activity and
include efficacy and PK. Results reported here are from parts C and D. Results: As of Nov 26, 2024, 31 pts deepening tumor regression were seen across tumor types and doses (ORR = 23%, DCR = 70%, median DOR =
received $1 dose of SV+P in parts C and D (19 NSCLC, 11 HNSCC, and 1 esophageal); median (95% CI) follow- 10.5 mo, TTR = 4.6 mo). No new safety signals were identified. Most Tx-related adverse events (TRAEs) were
up was 2.9 (1.6-5.0) months, and 26 pts remain on treatment. Median (range) age was 65 (34-80) years, 61% grade (Gr) 1 or 2; 13% had Gr $ 3 TRAEs (Gr 3 = 12%, Gr 4 = , 1%, Gr 5 = 0). Doses from 50 to 600 mg QD were
were male, and 52% had ECOG PS 0. Of pts with NSCLC, 12 (63%) had non-squamous tumors and 11 (58%) assessed for PK/PD. The AUC(0-24) after multiple doses was approximately dose proportional at 200 to 600 mg,
had tumors with PD-L1 TPS $1. All pts with HNSCC had tumors with PD-L1 CPS $1. Any-grade (Gr) and and the terminal t1/2 after a single dose was approximately 24 h (table). There were dose-dependent reductions
Gr $3 treatment-emergent adverse events (TEAEs) occurred in 87% and 35% of pts, respectively. Most in predicted plasma SDMA, with the 400 and 600 mg doses approaching the plateau. Conclusions: With longer
common TEAEs are shown in the Table. Any-Gr and Gr $3 immune-mediated TEAEs occurred in 61% and f/u, BMS-986504 continued to show increasingly durable antitumor activity. BMS-986504 demonstrated a
10% of pts, respectively. Pneumonitis/interstitial lung disease occurred in 3 pts (9.7%), with no Gr $3 events. favorable PK/PD profile, supporting QD dosing at 400 and 600 mg. These results support further investigation
Renal TEAEs led to discontinuation of both SV and P in 2 pts (6%); 3 other pts discontinued treatment (1 of BMS-986504 at 400 and 600 mg QD as a potential first-in-class synthetic lethal Tx option in pts with
progressive disease, 2 consent withdrawal). There were no treatment-related deaths. In 7 efficacy-evaluable advanced solid tumors with MTAP-del. Clinical trial information: NCT05245500. Research Sponsor: Bristol
pts with TPS$1 NSCLC, 1 confirmed (c) complete response (CR), 1 c partial response (PR), and 2 PRs Myers Squibb.
pending confirmation were observed (ORR 57%; cORR 29%). In 8 efficacy-evaluable pts with 1L HNSCC, 2 cCR PK and PD.
and 1 cPR were observed (cORR 37.5%). Conclusions: SV+P demonstrated manageable safety and en- 50 mg QD 100 mg QD 200 mg QD 400 mg QD 600 mg QD
couraging preliminary efficacy. These data support the ongoing phase 3 Be6A-Lung-02 study
(NCT06758401) comparing SV+P vs P as first-line treatment for pts with PD-L1 high (TPS$50) advanced PK after multiple doses
NSCLC. Clinical trial information: NCT04389632. Research Sponsor: Seagen, which was acquired by Pfizer in tamax (min–max), h 2.0 (2.0–4.0) 2.0 (1.0–2.0) 2.0 (0.5–6.0) 2.0 (0.5–4.0) 2.0 (1.0–6.0)
n=3 n=3 n=8 n = 10 n = 10
December 2023. Cmax,b ng/mL 107 377 685 1240 2110
n=3 n=3 n=8 n = 10 n = 10
All Treated Pts (n=31) AUC(0-24),b h∙ng/mL 948 3060 7150 11,800 26,700
Any Gr (>25%) Gr ‡3 n=3 n=3 n=7 n = 10 n=9
TEAEs n (%) n (%) AUC(0-24),b,c h∙ng/mL/mg 19.0 30.6 35.8 29.5 44.6
n=3 n=3 n=7 n = 10 n=9
Fatigue 13 (42) 1 (3) Terminal t1/2 after single dose,d h 21.7 80.7 21.5 21.7 24.1
Decreased 13 (42) 3 (10) n=1 n=1 n = 14 n = 14 n=1
appetite PD
Nausea 12 (39) 0 Translational exposure targets 0.08 0.21 0.5 1.1 1.7
Alopecia 11 (35) 0
(Cavgss), X
Asthenia 9 (29) 2 (6)
Decreased weight 8 (26) 0 Predicted plasma SDMA 30.0 38.8 48.3 55.0 57.4
Dysgeusia 8 (26) 0 reduction,a % (90% PI) (23.2–37.8) (27.1–47.5) (42.0–53.0) (50.7–57.6) (54.3–59.0)
a
Median. bGeometric mean. cDose-normalized. dArithmetic mean.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 191s

3012 Rapid Oral Abstract Session 3013 Rapid Oral Abstract Session
Preliminary results from a first-in-human, phase I/II study of VLS-1488, an A first-in-human phase I/II study of GFH375, a highly selective and potent
oral KIF18A inhibitor, in patients with advanced solid tumors. First Author: oral KRAS G12D inhibitor in patients with KRAS G12D mutant advanced solid
Ecaterina Elena Dumbrava, The University of Texas MD Anderson Cancer Center, tumors. First Author: Xinghao Ai, Department of Medical Oncology, Shanghai Chest
Houston, TX Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China
Background: VLS-1488 is an oral small molecule inhibitor of KIF18A, a mitotic kinesin Background: Kirsten rat sarcoma (KRAS) G12D is one of the most prevalent RAS
protein important for successful division of cancer cells with chromosomal instability mutations in human cancers and suggests poor survival. GFH375 is an orally bio-
(CIN) but not required for mitosis in normal cells. Preclinical studies of VLS-1488 showed available, highly selective and potent KRAS G12D inhibitor targeting both “ON” (GTP-
dose-dependent inhibition of tumor growth in CIN models. Methods: VLS-1488-2201 bound) and “OFF” (GDP-bound) states. Here we report the preliminary results of GFH375
is a phase I/II study of patients (pts) with advanced solid tumors consisting of two parts, in patients (pts) with advanced KRAS G12D mutant solid tumors. Methods: This is a
Dose Escalation and Dose Expansion. During Dose Escalation, a Bayesian Optimal Phase I/II study (NCT06500676) evaluating the safety, tolerability, pharmacokinetics
Interval design was utilized to enroll pts to dose escalation cohorts with additional pts and efficacy of GFH375 in pts with advanced solid tumors harboring KRAS G12D
enrolled to backfill cohorts at dose levels (DLs) that did not meet de-escalation/ mutation. Pts with locally advanced or metastatic solid tumor failed to prior standard
elimination rules. Primary objective was to assess safety/tolerability of VLS-1488 at therapies are eligible for enrollment. Accelerated titration, plus Bayesian Optimal Interval
various DLs to determine the Maximum Tolerated Dose (MTD). Secondary objectives (BOIN) and back filling design are employed in the phase I part with safety and tol-
included evaluating preliminary efficacy and pharmacokinetics (PK). Eligible pts had erability as the primary objective, and pharmacokinetics and anti-tumor activity as the
exhausted standard of care treatments and had measurable disease per RECIST v1.1. secondary objectives. Results: As of 03Jan2025, thirty-two pts were treated, including
Pts received VLS-1488 once daily, orally for 28-day cycles until disease progression, 11 pancreatic ductal adenocarcinoma (PDAC), 11 non-small cell lung cancer (NSCLC), 5
unacceptable toxicity or other stopping criteria. Results: 52 pts (ITT) were enrolled colorectal cancer (CRC) and 5 others (median age: 59.5 yrs; 62.5% female). No dose-
across 5 DLs including 50mg (n = 4), 100mg (n = 12), 200mg (n = 14), 400mg (n = 12) and limiting toxicities (DLTs) were observed at the tested dose levels of 100 mg, 200 mg,
800mg (n = 10). Tumor types were high grade serous ovarian (HGSOC; n = 20), colorectal 400 mg, 600 mg, 750 mg, 900 mg once daily (QD) and 300 mg twice daily (BID). Eight pts
(n = 14), triple negative breast (n = 7), squamous lung (n = 3), endometrial (n = 3), ovarian (25%) experienced at least one G3/G4 treatment related adverse event (TRAE) and no G5
carcinosarcoma (n = 2), esophageal (n = 2) and bladder (n = 1). The median number of TRAEs. Five pts (15.6%) experienced at least one serious adverse event. Eight pts (25%)
prior lines was 4 (range 1-8). As of data cutoff, 52 pts (100%) received .1 dose of VLS- had treatment interruptions, and 2 (6.3%) discontinued treatment due to treatment
1488. No dose-limiting toxicities (as assessed during the first 28 days) were observed emergent adverse events (TEAEs). No dose reduction occurred. The most common
and MTD was not reached. Treatment-related AEs (TRAEs) occurred in 22 pts (42%), with TRAEs were gastrointestinal events including diarrhea (71.9%), vomiting (71.9%) and
fatigue (17.3%; G1 13.5%, G2 3.8%), aspartate aminotransferase increased (13.5%; G1 nausea (62.5%); all were grade 1 or 2. Anti-tumor activities were observed starting from
7.7%, G2 1.9%, G3 3.8%) and rash (11.5%; G1 3.8%, G2 1.9%, G3 5.8%) observed in .10% 100 mg QD. Among 22 pts who had at least one post-treatment tumor assessment,
of pts. 6 pts (12%) experienced G3 TRAEs and no . G3 TRAEs were observed. Drug objective response rate (ORR) was 27.3% (6/22), and disease control rate (DCR) was
exposures exceeded preclinically defined efficacious thresholds and were approximately 86.4% (19/22). Nine out of 13 pts with stable disease (SD) had tumor shrinkage. Among
dose proportional at analyzed DLs. 41 pts (79%) were evaluable for response per RECIST the 7 pts with PDAC, all exhibited tumor shrinkage with 3 partial response (PR) and 4 SD.
v1.1. In the 16 HGSOC pts evaluable for response (where the median number of prior Among the 9 pts with NSCLC, 3 achieved PR, 5 SD, and 1 progression disease (PD).
lines was 4.5; range 2-8), 3 partial responses (PRs; including 2 pts with sustained GFH375 demonstrated good oral bioavailability with a Tmax of 2~4 h and a terminal half-
PR .24 weeks) and 6 with stable disease (SD; including 4 pts with tumor reductions) life of 18.5-21.6 h. Conclusions: According to the preliminary data from ongoing FIH
were observed across multiple DLs, with 5 pts continuing with study treatment. study, GFH375 monotherapy has demonstrated good tolerability and promising anti-
Conclusions: VLS-1488 was found to be safe and tolerable, with encouraging anti-tumor tumor activities in pts with advanced solid tumor supporting further clinical develop-
activity observed in heavily treated HGSOC pts. VLS-1488 will be evaluated further in the ment. Clinical trial information: NCT06500676. Research Sponsor: GenFleet Thera-
Dose Expansion phase of the study. Clinical trial information: NCT05902988. Research peutics (Shanghai) Inc.
Sponsor: Volastra Therapeutics, Inc.

3014 Rapid Oral Abstract Session 3015 Rapid Oral Abstract Session
Efficacy and safety of selumetinib in adults with neurofibromatosis type 1 Phase 2 evaluation of the nilotinib-paclitaxel combination in patients with
(NF1) and symptomatic, inoperable plexiform neurofibroma (PN): Primary rare solid tumors: Rapid analysis and response evaluation of combination
analysis of KOMET (NCT04924608), a phase 3, international, randomized, anti-neoplastic agents in rare tumors trial 1 (RARE CANCER 1). First Author:
placebo-controlled study. First Author: Alice P. Chen, Developmental Therapeutics Sarah Shin, Developmental Therapeutics Clinic/Early Clinical Trials Development Pro-
Clinic, Division of Cancer Treatment and Diagnosis, National Cancer Institute, National gram, Division of Cancer Treatment and Diagnosis, National Cancer Institute, Bethesda,
Institutes of Health, Bethesda, MD MD
Background: No globally approved therapies exist for adults with NF1 and symptomatic, in- Background: Rare tumors constitute a heterogeneous group of cancers with limited treatment
operable [Link] is evaluating the efficacy and safety of selumetinib (SELU; ARRY-142886, options and poor outcomes. To address the need for novel therapeutic options in this
AZD6244) in adults. Methods: KOMET is an ongoingPhase 3, randomized, double-blind, challenging population, patient-derived xenograft models of rare cancers were developed to
placebo-controlled trial. Adults ($18 yrs) with NF1 and symptomatic, inoperable PN were screen combinations of anticancer agents. Based on these preclinical data, the NCI Devel-
randomized 1:1 to 28-day cycles of oral SELU 25 mg/m2 BID or placebo (PBO) with crossover to opmental Therapeutics Clinic designed a series of phase 2 clinical trials to assess promising
SELU at progression or the end of Cycle (C) 12. Among others, baseline (BL) PAINS-pNF target novel combination therapies in patients (pts) with rare tumors. RARE1—the first trial in this
PN chronic pain intensity score (, 3 or $3) was a stratification factor; 70% of patients (pts) series—evaluates the nilotinib-paclitaxel combination, for which a preceding phase 1 study
were required to have a score $3. Primary analyses were conducted after the last pt completed (NCT02379416) recently identified the recommended phase 2 dose (RP2D) and promising
C16 (data cutoff: Aug 5, 2024). The primary endpoint was objective response rate (ORR; clinical activity, including confirmed partial responses (PR) in 3 pts (2 adult granulosa cell
confirmed partial/complete response) per ICR REiNS by the end of C16. Key secondary end- ovarian tumors [AGCOT], 1 endometrial cancer) and 1 unconfirmed PR (anal cancer). Given this
points were change from BL to C12 in PAINS-pNF chronic pain score in pts with a BL score $3 clinical experience and preclinical activity in rare tumor models, RARE1 (NCT04449549) aims
and PlexiQoL total score in all randomized pts (SELU vs PBO). A planned sample of 73 pts per to evaluate the response and mechanism of action of the nilotinib and paclitaxel combination
arm with a 2-sided 5% alpha Fisher’s exact test had . 99% power to detect the difference in rare, refractory solid tumors. Methods: Pts with rare tumors meeting the RARECARE
between a SELU ORR of 20% and PBO ORR of 0%. Key secondary endpoints were analyzed with a definition were treated at the RP2D: nilotinib 300 mg orally twice a day and paclitaxel 80 mg/m2
mixed model for repeated measures. Results: Of 145 randomized pts (SELU: 71; PBO: 74), IV on days 1, 8, and 15 in 28-day cycles. Response was assessed by RECIST v1.1. Tissue
51.7% were male; median age was 29 yrs (range 18–60). SELU led to a rapid onset of response biopsies and research blood were collected at multiple timepoints for pharmacodynamic and
(median 3.7 mos), with an ORR of 19.7% (95% CI 11.2, 30.9) by C16 vs 5.4% (95% CI 1.5, 13.3) genomic analyses. Results: This study enrolled 31 pts of diverse rare cancers as of the data
with PBO (p = 0.011). At C12, pts with a BL chronic pain score $3 had a greater reduction in pain cut-off. Of the 30 evaluable pts, 2 (7%) had confirmed PRs: 1 Ewing sarcoma and 1 ovarian
score with SELU (LS mean 22.0; 95% CI 22.6, 21.4) vs PBO (LS mean 21.3; 95% clear cell cancer, completing 12 and 11 cycles (C) on study, respectively. Stable disease (SD)
CI 21.8, 20.7); and clinically meaningful improvement (meaningful score difference 22 points) was the best response in 15 pts (50%), of which 5 pts (17%) had prolonged SD: 1 AGCOT (23+
vs BL, but this was not statistically significant vs PBO (p = 0.070). Reduction in chronic pain C), 1 testicular embryonal rhabdomyosarcoma (22 C), 1 non-uterine leiomyosarcoma (21 C), 1
intensity was observed with SELU vs PBO in the full analysis set (all pts regardless of BL chronic salivary gland (12 C), and 1 ampullary adenocarcinoma (6 C). Overall, the median number of
pain intensity, nominal p = 0.024). Change from BL to C12 in PlexiQoL total score between cycles completed is 2 (range 0 - 23) and the median progression free survival is 3.8 months. No
treatment arms was not statistically significant (LS mean difference 20.1; 95% CI 21.2, 1.1). unexpected treatment-related adverse events have occurred. No grade 3-4 peripheral neu-
Adverse events (AEs) in the randomized period were consistent with the known safety profile of ropathy has been observed. Conclusions: Preliminary clinical outcomes for the nilotinib-
SELU. The most common AEs ($10% of pts) were dermatitis acneiform (59%), increased blood paclitaxel combination in patients with rare tumors showed encouraging signals of activity. To
creatine phosphokinase (45%), and diarrhea (42%) with SELU, and COVID-19 (20%), nausea facilitate further evaluation of response and the underlying mechanism of action for this
(16%), and fatigue (14%) with PBO. Fourteen pts on SELU and 1 pt on PBO reported CTCAE combination, enrollment now focuses on the 4 tumor types that have previously demonstrated
Grade $3 treatment-related AEs; 9 SELU and 5 PBO pts discontinued due to AEs. response: Ewing sarcoma, ovarian clear cell carcinoma, AGCOT, and anal cancers. Phar-
Conclusions: In the first international, randomized, placebo-controlled trial in adults with NF1- macodynamic and genomic analyses are also ongoing. This project has been funded in whole
PN, SELU achieved a significant ORR vs PBO (C16), meeting the primary endpoint, and a or in part with federal funds from the NCI, NIH, under contract HHSN261201500003I. Clinical
clinically meaningful reduction in PN-associated chronic pain (C12). Clinical trial information: trial information: NCT04449549. Research Sponsor: U.S. National Institutes of Health.
NCT04924608. Research Sponsor: Alexion, AstraZeneca Rare Disease and Merck Sharp &
Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
192s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3016 Rapid Oral Abstract Session 3017 Rapid Oral Abstract Session
Clinical utility of circulating tumor RNA (ctRNA) in a combined circulating Enhancing prognostic precision in bladder cancer: AI-driven tumor micro-
tumor DNA (ctDNA) and ctRNA next-generation sequencing (NGS) pan- environment analysis from H&E images. First Author: Evelyn Ramberger,
cancer liquid biopsy assay. First Author: Jonathan Poh, Lucence Diagnostics Pte Aignostics GmbH, Berlin, Germany
Ltd, Singapore, Singapore Background: Bladder cancer (BC) represents a significant healthcare burden. Despite
Background: Liquid biopsies are increasingly used in the real world for cancer therapy advancements in diagnostics and treatment, the survival rate remains low, underscoring
selection, disease monitoring, and early detection. Combining RNA with DNA sequencing the need for improved prognostic tools. The current UICC staging system often lacks
for the detection of FDA-actionable gene fusions is already recommended as standard of precision in patient stratification. Moreover, there is a paucity of scalable methods that
care in guidelines as fusions are inherently challenging to detect by DNA-only methods. explore and quantify tumor microenvironment (TME) features and their influence on
However, the clinical utility of profiling plasma ctRNA in addition to ctDNA has not yet been patient outcome. Here, we present an artificial intelligence (AI) framework that operates
quantified in large studies. Here we report results from the first large real-world study on routine hematoxylin & eosin-stained (H&E) slides to enable systematic TME char-
establishing the clinical utility of combining ctRNA and ctDNA in a single liquid biopsy acterization and improve prognostic accuracy. Methods: In a bicentric cohort of over 700
assay. Methods: A total of 1,007 consecutive plasma samples from 979 cancer patients resected BC patients, we developed and validated a deep learning approach for TME
across the USA and Asia underwent real-world liquid biopsy testing with a combined ctDNA analysis. The model was trained using multiplex immunofluorescence-validated anno-
and ctRNA assay (LiquidHALLMARK) in two CAP-accredited CLIA-certified laboratories tations but operates solely on H&E-stained images, maximizing clinical applicability. Key
from Jun 2021 to Dec 2024. The combined amplicon-based assay profiles genomic al- features included tissue compartment segmentation, cell classification, and spatially
terations in 80 genes on ctDNA and up to 37 genes on ctRNA. Only gene fusions (including resolved cell patterns. We evaluated the model’s performance and integrated TME
MET exon 14 skipping) were included in this analysis. The limit of detection for gene features with clinicopathological variables to improve prognostic stratification beyond
fusions of the ctDNA and ctRNA panel were validated as 0.5% and 10 copies. UICC staging. Results: The model demonstrated robust tissue compartment segmen-
Results: Plasma ctRNA was successfully analyzed in 99.6% (1003/1007) of samples tation (F1-score = 0.91) and accurately identified key immune cell populations in tissue
across 30 cancer types. The top 5 cancer types (lung, prostate, breast, colorectal, and regions. When integrating the spatially resolved cellular features with clinicopathological
pancreas) comprised 84.2% of the clinical volume. Gene fusions were detected across 11 variables, we observed significant improvements in prognostic capabilities for overall
genes (ALK, RET, ROS1, MET, NRG1, NTRK1, FGFR2, FGFR3, ESR1, ERG, ETV4) in 7.8% (78/ survival. The integrated approach demonstrated a 22% relative improvement over the
1003) of cases, primarily in prostate and lung, but also in breast, bile duct, thyroid, liver, and conventional UICC staging system alone (C-index increased from 0.59 to 0.61, p , 0.01),
bladder cancers. A total of 80 fusions were detected, of which 25 were detected by both measured against the random baseline C-index of 0.5. Our integrated model displayed a
ctDNA and ctRNA, while ctDNA and ctRNA each exclusively detected 27 and 28 fusions. hazard ratio of 1.859 (95% CI: 1.530-2.259, p = 4.390e-10), markedly stronger than
Among the 28 ctRNA-only fusions, 5 were not covered by the ctDNA panel (1 ATP1B1- traditional risk stratification which showed a hazard ratio of 1.477 for high versus low risk
NRG1, 1 ESR1-CCDC170, 1 ESR1-AKAP12, and 2 SLC45A3-ERG fusions). Eleven (11/28) groups (95% CI: 1.219-1.791, p = 7.117e-05). These findings demonstrate that AI-driven
ctRNA-only fusions were actionable; all 11 were found in lung cancer. Two of these co- analysis of the tumor microenvironment provides valuable prognostic information beyond
occurred with another lung driver mutation, highlighting potential resistance mechanisms current clinical staging methods, suggesting promising opportunities for enhancing
to targeted therapy. Of the remaining 9, 3 were treated with fusion-matched targeted patient risk stratification. Conclusions: We show that a combination of AI with UICC
therapy. Two patients had real-world response rates available; both exhibited partial staging shows improved patient stratification compared to stratifying by UICC alone. This
response to treatment. Overall, inclusion of ctRNA analysis increased the diagnostic yield study demonstrates the feasibility of automated TME characterization from routine H&E
of all fusions by 53.8% and actionable fusions by 36.7%. Conclusions: This is the first large slides in BC and suggests that incorporating TME features into prognostic models en-
study showing that adding ctRNA to ctDNA liquid biopsy increases total actionable di- hances accuracy and could support personalized patient management in individualized
agnostic yield by 36.7%, highlights potential resistance mechanisms, and can broaden oncology. While further validation in larger, multicentric-datasets is required, our ap-
panel coverage to include gene fusions not amenable to detection by conventional DNA- proach shows potential for facilitating systematic biomarker development and improving
based methods. These findings support recommendations for combined DNA/RNA testing clinical decision-making in BC care. Research Sponsor: BMBF.
of fusions in both tissue and liquid samples. Research Sponsor: None.

3018 Poster Session 3019 Poster Session

Preliminary results from a first-in-human phase 1 dose escalation trial of BC3195, a novel ADC targeting cadherin-3 (CDH3): Updated results of a
ADRX-0706, a next generation Nectin-4 ADC, in subjects with advanced first-in-human phase I study in patients with advanced solid malignancies.
solid tumors. First Author: Alexandra Drakaki, Jonsson Comprehensive Cancer Center, First Author: Haiyan Tu, Guangdong Lung Cancer Institute, Guangdong Provincial
David Geffen School of Medicine at UCLA, Los Angeles, CA People’s Hospital (Guangdong Academy of Medical Sciences), Southern Medical Uni-
Background: ADRX-0706 is a Nectin-4 targeting ADC designed to provide an increased versity, Guang Zhou, Guang Dong, China
therapeutic window through stable conjugation of a novel microtubule inhibitor payload Background: Cadherin-3 (CDH3), a calcium-dependent cell-cell adhesion glycoprotein, is overexpressed
(AP052) to an IgG1 monoclonal antibody at a drug-to-antibody ratio of 8. Preliminary on lung, breast, head and neck and other malignancies, and associated with cancer invasiveness and poor
safety, anti-tumor activity, pharmacokinetic (PK) and Nectin-4 expression results are prognosis. BC3195 is known as the only antibody drug conjugate (ADC) in clinical stage, targeting CDH3
with cleavable linker and payload of monomethyl auristatin E (MMAE). Methods: A phase I, open-label,
presented from the dose escalation part of the ongoing Phase 1 trial (NCT06036121). first in human study whose objectives were to evaluate the safety, tolerability, pharmacokinetics (PK), and
Methods: Eligible subjects with select advanced solid tumors (urothelial [UC], cervical preliminary antitumor activity of BC3195 is being performed in patients (pts) with advanced solid
[CC], breast [BC], head and neck [HNSCC], ovarian [OC], non-small cell lung [NSCLC], and malignancies. BC3195 is administered as 1-hr IV infusion every 3 weeks (Q3W) or every week (QW). An
pancreatic [PC]) were enrolled in cohorts of escalating dose levels (1-16 mg/kg, Q3W, IV) evaluation of seven dose levels (DLs) is planned: 0.3, 0.6, 1.2, 1.8, 2.1, 2.4 mg/kg Q3W and 1.2 mg/kg QW
using a BOIN design with backfill. Nectin-4 expression was evaluated retrospectively. with a BOIN design guiding dose escalation. Results: As of the data cut-off-date (Dec 26th, 2024), 56 pts
Primary and secondary endpoints included dose-limiting toxicities [DLTs], adverse events have been enrolled. The number of pts in each DL is shown in the table. Twenty-five (44.6%) pts had
[AEs], laboratory value changes, PK, immunogenicity, and response per RECIST v1.1. received $3 prior lines of treatment. Stomatitis (71.4%), rash (60.7%) and anemia (53.6%) were the main
adverse events (AEs). Stomatitis and rash typically occurred in the first cycle and were manageable.
Results: As of the 13Dec24 data cutoff, 53 subjects with a median age of 59 years and
Twenty-one pts (37.5%) experienced Grade$3 treatment related adverse events (TRAEs). Among the 50
median of 4 (1-14) prior therapies were enrolled. One DLT (G3 stomatitis) occurred at the pts who were evaluable for tumor response, 5 pts in 2.4 mg/kg Q3W had partial response (PR). Of the 20
highest dose of 16 mg/kg. The most common treatment related AEs (TRAE $15%) were NSCLC pts treated in 2.4 mg/kg, 4 pts had confirmed PR (cPR), and 14 pts had stable disease (SD) as their
arthralgia (32%), fatigue (21%), rash (19%), anemia (17%), and nausea (15%). The majority best response; the objective response rate (ORR) was 50% (4/8) in previously-treated EGFR-mutant
of TRAEs were G1-2 in severity and manageable, including only 3 (5.7%) subjects with NSCLC pts, and mPFS was 168 days (Table). PK results demonstrated that exposure of the ADC, total
peripheral neuropathy and 2 (3.8%) with liver enzyme increase. The most common $G3 antibody (TA) and MMAE increased in a non-linear manner at dose up to 2.4 mg/kg. Median Tmax values
TRAE was neutropenia (11%). ADC exposure increased in a dose-proportional manner with for ADC and TA were 1 h, and median Tmax for free MMAE was 25-169 h. In addition, elimination t1/2 values
averaged 54 h, 78 h and 63 h for the ADC, TA, and MMAE at 2.4 mg/kg, respectively. Conclusions: BC3195
minimal deconjugation and the ADC half-life was 15 days. There were 5 subjects who
has a manageable safety profile and favorable PK characteristics and demonstrated impressive pre-
achieved objective response across different tumor types (UC, NSCLC, CC) and 9 with liminary antitumor activity in heavily-pretreated pts with NSCLC, of which most had EGFR-mutations
stable disease per RECIST among 30 response-evaluable subjects treated at doses $8 mg/ (ORR=50%). Dose optimization and expansion are ongoing. Clinical trial information: NCT05957471.
kg (ORR 16.7%, DCR 46.7%), including 2 triple negative BC (TNBC) subjects with 27% and Research Sponsor: Biocity Biopharmaceutics Co. Ltd.
29% decrease in tumor size who remain on treatment. ADRX-0706 demonstrated Nectin-4 Safety and efficacy data of study BC3195-101 (safety analysis set).
expression-dependent anti-tumor activity with all responses observed in tumors with H- 0.3 mg/kg 0.6 mg/kg 1.2 mg/kg 1.8 mg/kg 2.1 mg/kg
score $100, including a confirmed complete response (CR) in a CC subject (H-score 250). Efficacy Q3W Q3W Q3W Q3W Q3W 2.4 mg/kg All*
Two responses were observed after prior progression on other Nectin-4 targeting MMAE and Safety (N = 3) (N = 3) (N = 3) (N = 9) (N = 6) Q3W (N =31) (N=56)
drugs and three responses remain ongoing with subjects on treatment for 9+ to 23+ weeks. All tumor types EGFR-mut
Based on these data, 10 mg/kg Q3W was selected as the Phase 1b dose. (N = 31) NSCLC
(N = 8)
Conclusions: ADRX-0706 demonstrated a preliminary safety profile differentiated from ORR, n (%) 0 0 0 0 0 5 (16.1) 4 (50.0) 5 (8.9)
MMAE-conjugates and with manageable toxicities. The antibody-like PK profile together DCR, n (%) 1 (33.3) 2 (66.7) 1 (33.3) 3 (33.3) 2 (33.3) 22 (71.0) 7 (87.5) 32 (57.1)
with minimal deconjugation supports Q3W dosing. Encouraging anti-tumor activity was mPFS, days 40 82 40 40 39 130 168 91
Grade‡3 TRAE 0 0 0 2 (22.2) 1 (16.7) 17 (54.8) 3 (37.5) 21 (37.5)
observed in multiple heavily pretreated tumors with moderate-high Nectin-4 expression. n (%)
Enrollment in Phase 1b cohorts of UC, CC, and TNBC is ongoing. Clinical trial information:
*The subject in 1.2 mg/kg QW dose level is not presented in a dedicated column in this table.
NCT06036121. Research Sponsor: Adcentrx Therapeutics.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 193s

3020 Poster Session 3021 Poster Session

A phase 2 basket trial of ado-trastuzumab emtansine for patients with HER2 The safety, tolerability, and efficacy of BRY812 in patients with advanced
amplified cancers. First Author: Jessica Ross, Memorial Sloan Kettering Cancer solid tumors: Preliminary results from the phase I clinical study. First Author:
Center, New York, NY Herui Yao, Sun Yat-sen Memorial Hospital, Sun Yat-Sen University, Guangzhou, China
Background: The HER2 gene is commonly amplified (amp) across a variety of tumor types. Ado- Background: Antibody-conjugated drugs (ADCs) have demonstrated outstanding
trastuzumab emtansine (TDM1) is a potent antibody-drug conjugate targeting HER2 that is ap- clinical efficacy in treating a wide range of solid tumors as well as hematological tumors
proved in HER2+ breast cancer. The efficacy of TDM1 in other HER2-amp solid tumors is unknown. currently. An ongoing multicenter, open-label, phase I clinical study assessed the safety,
Methods: We conducted a single-arm, phase 2 basket trial of TDM1 in which patients (pts) were tolerability and preliminary efficacy of BRY812, the ADC targeting transmembrane
enrolled in one of 5 HER2-amp cohorts: non-small cell lung cancer (NSCLC), colorectal cancer,
endometrial cancer, salivary gland cancer, or other solid tumor. HER2 amp was identified through
protein LIV-1 with MMAE as cytotoxic payload, in advanced solid tumors. Here we report
next generation sequencing by MSK-IMPACT, defined as two-fold change, or by in-situ hybrid- the interim analysis results. Methods: In Phase Ia (dose escalation), the eligible patients
ization (ISH) with HER2/CEP17 ratio $2.0 in a CLIA-certified laboratory. In tumors sequenced by with advanced solid tumors were enrolled in each of 7 dose groups (0.25, 0.5, 1.0, 2.0,
MSK-IMPACT, precise level of the ERBB2 amplification (i.e. integer copy number) was assessed 2.8, 3.6 and 4.4 mg/kg) for evaluation to determine the MTD of BRY812. An "accelerated
and correlated with clinical response. All pts received TDM1 3.6mg/kg IV every 21 days. The titration" method (0.25, 0.5 mg/kg) as well as a modified toxicity probability interval-2
primary endpoint was overall response rate (ORR). For each cohort, a Simon two-stage optimal method (subsequent doses) was applied. Subsequently dose expansion was conducted
design was used. In the first stage, 7 pts were accrued in each cohort; if 0/7 responses, the cohort for dose levels of 2.0, 2.8 and 3.6 mg/kg that demonstrated tolerability and relative
was closed. Otherwise, up to 11 additional pts were accrued. Cohorts 1, 3, and 4 were expanded by efficacy. The patients received treatment every 3 weeks until intolerable toxicity or
up to 5 pts (max 23 pts) due to durable responses seen early on. Response and progression of disease progression. The primary endpoint was to evaluate DLT and MTD to determine
disease was evaluated using RECIST version 1.1. Modified PERCIST was allowed if pts did not have
RP2D, other endpoints included ORR. Results: Overall, as data cut-off date (Dec 13,
RECIST measurable disease. Toxicity was graded as per CTCAE v4.1. Circulating tumor DNA was
collected pre-, post-, and on-treatment for all pts when feasible. Results: 88 pts were accrued
2024), 36 patients (including 30 patients with breast cancer) with advanced solid tumors
between 2016 and 2023. The ORR by cohort is listed in Table 1. The most common toxicities were were enrolled, including 20 patients in dose escalation phase and 16 patients in dose
decreased platelet count and elevated ALT. There was one incident of grade 5 pneumonitis in a expansion respectively. Treatment is still ongoing for 19 patients. No DLT was observed
patient in cohort 1. Median time on treatment was 2.5 months (range 0.03 – 53.7 months); in pts up to 3.6 mg/kg. 4.4 mg/kg was not tolerated due to DLTs (2 of 4 patients experienced
with salivary gland tumors, median time on treatment was 17.6 months (0.3 – 53.7). DLT events). The most common grade $3 TEAE was neutropenia, and grade 4 neu-
Conclusions: TDM1 demonstrated efficacy in multiple HER2-amp tumor types. The highest ORR tropenia was observed in 3.6 mg/kg (2/12) and 4.4 mg/kg (4/4). 4 patients discontinued
was seen in salivary gland tumors, with a median time on treatment of about 1.5 years. More work treatment due to AE including 2 injury corneal (each of 3.6 & 4.4 mg/kg), 1 peripheral
is needed to understand the enhanced efficacy of TDM1 in these tumors. Clinical trial information: neuropathy (2.8 mg/kg) and 1 hepatic enzyme increase (3.6 mg/kg). Among 34 patients
NCT02675829. Research Sponsor: Genentech. in efficacy analysis, 8 (23.5%) patients and 7 (20.6%) patients had PR and SD, re-
Response rate by cohort and disease site. spectively, and 17 (50%) patients had PD, leading to ORR of 23.5% (95% CI: 10.7, 41.1)
Cohort RECIST-only ORR Combined* ORR and DCR of 44.1% (95% CI: 27.2, 62.1). The patients with higher LIV-1 expression showed
1: HER2-amp lung 4/18 (22.2%) 4/19 (21.1%) better efficacy, as among 14 patients with PS2+ enriched, ORR was 43% (6/14 in breast
2: HER2-amp colorectal 0/7 (0.0%) 0/7 (0.0%) cancer). In addition, ADC and Total antibody clearance were similar. BRY812 showed
3: HER2-amp endometrial 5/23 (21.7%) 5/23 (21.7%) dose-dependent decrease of clearance in the dose range of 0.25 to 2.0 mg/kg, while
4: HER2-amp salivary 8/10 (80.0%) 14/16 (87.5%)
5: Other HER2-amp solid tumors 2/23 (8.7%) 2/23 (8.7%) approximately linear clearance in the dose range of 2.0 to 4.4 mg/kg, and the half-life
Biliary 1/8 (12.5%) 1/8 (12.5%) were ~7 days. No accumulation was observed after multiple dosing. ADA positive rate
Bladder & urinary tract 0/5 (0.0%) 0/5 (0.0%) was 11.43% (4/35). Conclusions: BRY812 demonstrated favorable safety and tolera-
Cervical 0/2 (0.0%) 0/2 (0.0%)
Ovarian 1/7 (14.3%) 1/7 (14.3%) bility profile, with promising clinical efficacy in patients with advanced solid tumors.
Pancreatic 0/1 (0.0%) 0/1 (0.0%) Further dose optimization and clinical efficacy will be explored in Phase Ib. Clinical trial
TOTAL 19/81 (23.5%) 25/88 (28.4%) information: NCT06038058. Research Sponsor: BioRay Pharmaceutical (Hangzhou) Co.,
ORR=overall response rate. *Combined: RECIST when available, PERCIST if non-RECIST-evaluable. Ltd.

3022 Poster Session 3023 Poster Session

A pooled analysis of JSKN003, a biparatopic anti-HER2 antibody conjugate Initial results from a first-in-human phase 1 study of LY4170156, an ADC
(ADC), in patients with advanced HER2-overexpressing (IHC 3+) gastroin- targeting folate receptor alpha (FRa), in advanced ovarian cancer and other
testinal tumors. First Author: Dan Liu, Beijing Cancer Hospital, Beijing, China solid tumors. First Author: Isabelle Laure Ray-Coquard, Centre Léon Bérard, and
Background: JSKN003 is a biparatopic HER2-targeting ADC conjugated with a topo- GINECO, Lyon, France
isomerase I inhibitor (TOP1i) payload via a dibenzocylooctyne tetrapeptide linker. The Background: Folate receptor alpha (FRa) is overexpressed in several solid tumors.
efficacy and safety of JSKN003 in several solid tumors have been highlighted in previous LY4170156 is an Fc-silent, FRa specific humanized IgG1 ADC linked to exatecan, a topo-I
reports. Methods: JSKN003-101 and JSKN003-102 are dose escalation and expansion inhibitor, via a novel cleavable polysarcosine linker at a homogenous DAR of 8. LY4170156
studies involving Australian and Chinese patients (pts) with metastatic solid tumors. demonstrated in vivo preclinical efficacy in tumor models, across all FRa expression levels.
This pooled analysis of two studies was performed to assess the efficacy and safety in Methods: This is a multicenter, open-label, first-in-human phase 1a/b study of LY4170156 in
advanced HER2-overexpressing (IHC 3+) gastric or gastroesophageal cancer (GC/GEJC) patients (pts) with advanced FRa-expressing ovarian, endometrial, cervical, and other solid
and colorectal cancer (CRC) pts. Results: As of data cutoff (18 Dec 2024), 40 patients tumors. Pts with prior ADCs targeting FRa with payloads other than topo-I (including
with HER2-overexpressing (IHC 3+ by local lab) gastrointestinal tumor (23 in GC/GEJC mirvetuximab soravtansine-gynx [mirv]) were allowed. Dose escalation followed the mTPI-2
method. LY4170156 was administered Q3W IV (dose range of 2-6 mg/kg); dose limiting
and 17 in CRC) were enrolled across 7 dose levels: 2.1 mg/kg (n = 1), 4.2 mg/kg (n = 1),
toxicity (DLT) evaluation period was 21 days. Dose escalation included a randomized dose
5.2 mg/kg (n = 1), 6.3 mg/kg (n = 33), 7.3 mg/kg (n = 1), 8.4 mg/kg (n = 2), 10.5 mg/kg (n
optimization cohort in PROC. Key endpoints were safety, PK, and antitumor activity per
= 1). The median follow-up time of two studies was 7.16 months. Most pts were heavily RECIST v1.1. Efficacy evaluable pts were those who had a post baseline response assessment
pretreated (37.5% had $3 lines of prior treatment; 45.0% received irinotecan; 67.5% or discontinued treatment prior to the response assessment. Results: As of 27 Nov 2024, 45
received anti-HER2 therapy; 42.5% received IO therapy). Four of the 17 CRC pts were pts were treated with LY4170156. Median age was 63 yrs (range, 24-85), 100% had ECOG PS
RAF/RAS mutations (n = 2 RAS-mut, n = 2 RAF-mut). Thirty-nine patients had at least 0-1, and 32 (71%) had high-grade serous ovarian cancer (HGSOC). Among the HGSOC pts (32),
one tumor assessment after baseline. The overall response rate (ORR) per RECIST v1.1 in median lines of prior therapy was 5 (range, 1-10), 19% had received prior mirv, and 44% had
HER2-overexpressing gastrointestinal tumor was 66.7% and the disease control rate FRa expression , 75% by local or central testing. PK of LY4170156, total antibody, and
(DCR) was 94.7%. Among 22 GC/GEJC pts, the ORR was 68.2% and DCR was 95.5%. The exatecan were linear and dose-proportional within the tested dose range. Unconjugated
median progression-free survival (PFS) was 9.59 months (95% CI: 2.96, NE) with 66.3% payload release from ADC at Cmax was , 4% at 4 mg/kg; median half-life of LY4170156 was
(95% CI: 29.4, 87.1) PFS rate at 6 months. Among 17 CRC pts, the ORR was 64.7% (66.7% 5.7-7.0 days and exatecan was 7.2-8.6 days. Main toxicities were myelosuppression and GI-
in RAF-wild pts, n = 15) and DCR was 94.1%. The mPFS was 13.77m (95% CI: 7.1, NE) related, as expected from an exatecan payload. Across all doses, the most common
with 94.1% (95% CI: 65, 99.2) PFS rate at 6 months. The median overall survival (OS) was treatment-emergent adverse events (TEAEs; $15%) were nausea (58%, 2% gr 3), fatigue
not yet mature. Notably, one BRAF-mut patient achieved PR at first tumor assessment (44%, 0% gr 3-4), anemia (33%, 24% gr 3), vomiting (27%, no gr 3-4), diarrhea (22%, 4% gr 3),
after baseline, two RAS-mut pts achieved PR and duration was over 48 weeks. The most and neutropenia (20%, 11% gr 3-4). Febrile neutropenia (FN) was observed in 3 pts (7%). To
common treatment-related adverse events (TRAEs) included nausea, diarrhea, neu- date, no pulmonary or ocular toxicity were noted. Two DLTs were observed (1 gr 3 FN [6 mg/
tropenia, decreased appetite, vomiting, rash, anemia and fatigue. Grade 3/4 neutropenia kg]; 1 gr 3 anemia [2 mg/kg]); no MTD has been established to date. Among 13 efficacy
was observed in 2 (5.0%) pts, Grade 3/4 anemia was observed in 1 (3.0%) pts. No TEAEs evaluable HGSOC pts (6 FRa $75%; 6 , 75%; 1 pending data), 9 showed reduction in target
led to death or treatment discontinuation. Interstitial lung disease (ILD) occurred in 3 lesions; preliminary ORR was 38% (n = 5) with 1 CR, 4 PR, and 4 SD across all dose levels.
(7.5%; n = 2 G1; n = 1 G2) pts. Conclusions: JSKN003 demonstrated promising efficacy Combined ORR for 4 and 6 mg/kg dose levels was 55%. All responses were unconfirmed and
in heavily pretreated pts with advanced HER2-overexpressing gastrointestinal tumors, ongoing at the time of data cutoff. Three of 5 responders had FRa expression , 75% and
two $75% including 1 who was mirv refractory. Conclusions: LY4170156 was well-tolerated
with a manageable and predictable safety profile. Clinical trial information:
with encouraging clinical activity among HGSOC pts, including those with FRa expression ,
NCT05494918, NCT05744427. Research Sponsor: Jiangsu Alphamab Bio-
75% and those with prior mirv. Randomized dose optimization is ongoing and updated data
pharmaceuticals Co., Ltd. will be presented. Clinical trial information: NCT06400472. Research Sponsor: Eli Lilly and
Company.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
194s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3024 Poster Session 3025 Poster Session

BAT8008, a TROP-2 antibody-drug conjugate (ADC), in patients with ad- A phase I clinical study to evaluate the safety, tolerability, and pharmaco-
vanced solid tumor: Results from a phase 1 study. First Author: Jianli Zhao, Sun kinetic characteristics of HLX43 (anti-PD-L1 ADC) in patients with
Yat-sen Memorial Hospital, Sun Yat-Sen University, Guangzhou, China advanced/metastatic solid tumors. First Author: Jie Wang, Internal Medicine,
Background: BAT8008 is a monoclonal ADC that delivers exatecan to cells expressing Department of Medical Oncology, National Cancer Center/National Clinical Research
TROP-2. TROP-2 is a cell surface glycoprotein that can be expressed in certain normal Center for Cancer/Cancer Hospital, Beijing, China
tissue but is frequently overexpressed in multiple carcinomas including cervical cancer Background: Antigens highly expressed in tumors such as HER2 and TROP2 have been
(CC) and esophageal cancer (EC). Here we report the safety and efficacy data of BAT8008. widely investigated as antibody-drug conjugate (ADC) targets, leading to their approval for
Methods: BAT8008 was administered by intravenous infusion at doses of 0.8-2.7mg/kg on various cancers due to their promising efficacies. However, limited targets for approved
days 1 of each14-day cycle(the first cycle is 21-day). The study included dose escalation, ADCs and resistance to the cytotoxic agents render the imperative need for new ADCs. This
dose expansion and cohort expansion which included CC, EC and other solid tumors that study aimed to evaluate the safety, tolerability, and preliminary efficacy of HLX43, a novel
progressed after . l systemic treatments (Tx). Primary objectives were assessment of anti-PD-L1 ADC in patients with advanced/metastatic solid tumors. Methods: This phase 1
safety and preliminary efficacy. Results: As of Jan 15, 2025, 170 patients (pts) were study consisted of 2 parts. Parts 1 and 2 were dose escalation and dose expansion phases,
enrolled with doses ranging from 0.8 to 2.7mg/kg. 2 out of 6 pts in 2.7mg/kg group had respectively, to explore different doses of HLX43. In Part 1, patients with histologically or
dose limiting toxicity (1 with G3 increased lipase, 1 with G4 thrombocytopenia and G4 cytologically confirmed advanced/metastatic malignant solid tumors refractory to or not
febrile neutropenia). The maximum tolerated dose and the RP2D was selected as 2.4mg/ amenable to standard therapies received intravenous HLX43 at 0.5 mg/kg, 1 mg/kg, 2 mg/
kg. 147 pts were enrolled at dose of 2.4mg/kg. The most common TRAEs of 2.4mg/kg dose kg, 3 mg/kg, or 4 mg/kg, Q3W. In Part 2, patients with advanced/metastatic non-small cell
group ($20%, all grade/$5%, $G3) were anemia (78.8%,13.7%), white blood cell count lung cancer (NSCLC) refractory to standard treatment received HLX43 at 2 mg/kg, 2.5 mg/
decreased (62.3%, 18.5%), nausea (59.6%,0%), stomatitis(59.6%,19.2%), neutrophil count kg, or 3 mg/kg, Q3W. The primary endpoints for Part 1 were the proportion of subjects
decreased (52.7%,19.2%), platelet count decreased (38.4%,8.2%), vomiting experiencing dose-limiting toxicity (DLT) in each dose group within three weeks after the
(38.4%,0.7%), lymphocyte count decreased (35.6%,5.5%), fatigue (34.9%, 0%), body weight first drug administration and the maximum tolerable dose (MTD) while that for Part 2 were
loss (29.5%, 0.6%), constipation (26.7%,0%), anorexia (23.3%,2.7%). 22 CC pts and 13 EC the recommended Phase 2 dose and IRRC-assessed objective response rate (ORR).
pts were enrolled at 2.4mg/kg and evaluable for tumor assessment. The obiective response Results: As of 27 June 2023, 18 patients with non-small cell lung cancer (n = 12, 66.7%),
rate (ORR) was 36.4% and 23.1%, respectively. Median prior lines of Tx were 2(range, 1-5). head and neck squamous carcinoma, cervical squamous carcinoma, thymic squamous cell
50% CC and 93% EC pts progressed after platinum-based chemotherapy and immune carcinoma, nasopharyngeal cancer, uterine carcinosarcoma, or small cell lung cancer (n = 1,
checkpoint inhibitors, respectively. 23% EC had previously used topoisomerase I inhibitors. 5.6% for each) were enrolled in Part 1 and received HLX43 at 0.5 mg/kg (n = 3), 1 mg/kg (n =
Objective responses were also observed in pts with other solid tumor types. 3), 2 mg/kg (n = 3), 3 mg/kg (n = 3), or 4 mg/kg (n = 6). All the patients experienced
Conclusions: The data indicated encouraging efficacy of BAT8008 in advanced CC and EC. treatment-emergent adverse events (TEAEs) that were mostly grades 1-2. One patient in the
The safety profile showed adequate tolerability. Clinical trial information: NCT05620017. 4 mg/kg dose group experienced DLTs of febrile neutropenia and decreased white blood cell
Research Sponsor: Bio-Thera Solutions, Ltd. count. Investigator-assessed ORR was 31.3% (95% CI 11.0-58.7). In Part 2, only data from 21
patients enrolled to receive HLX43 at 2 mg/kg is available and presented here. Among these
Tumor Type CC EC patients, 15 (71.4%) had squamous NSCLC and 6 (28.6%) had nonsquamous NSCLC.
n 22 13 Investigator-assessed ORR and disease control rate were 38.1% (95% CI 18.1-61.6) and
CR 1 1 81.0% (95% CI 58.1-94.6); no complete response was achieved, and 8 patients (6 sqNSCLC
PR 7 2 and 2 nsqNSCLC) had partial response. All the patients experienced TEAEs, most of which
ORR,% 36.4 23.1 were grades 1-2; grade $3 TEAEs occurred in 7 (33.3%) patients. Conclusions: HLX43 was
cORR,% 31.8 15.4
well tolerated with no new safety signals across different dose and exhibited encouraging
DCR,% 77.3 100
PFS, months 6.8 5.3 preliminary efficacy in patients with advanced solid tumors, including those with NSCLC,
(95% CI) (3.4-10.2) (3.1-7.4) who had failed standard therapies, which warrants further investigation. Clinical trial in-
formation: NCT06115642. Research Sponsor: Shanghai Henlius Biotech, Inc.

3026 Poster Session 3027 Poster Session

First in human phase I study of TQB2103, a Claudin18.2 (CLDN18.2) Association of genomic alterations in circulating tumor DNA (ctDNA) with
targeted antibody-drug conjugate (ADC), in patients with advanced solid clinical response to telisotuzumab vedotin (Teliso-V) in 2L+ EGFR wildtype
tumors. First Author: Xiangdong Cheng, Zhejiang Cancer Hospital, Hangzhou, Zhejiang (EGFRwt) non-squamous non-small cell lung cancer (NSCLC) patients (pts)
Province, China with c-Met overexpression (OE). First Author: David Ross Camidge, University of
Background: Claudin18.2 is a promising target for CLDN18.2-expressing cancers such Colorado Cancer Center, Aurora, CO
as gastric and pancreatic cancers. TQB2103 is a novel ADC comprised of a humanized Background: Teliso-V is an antibody-drug conjugate comprising the c-Met–targeting anti-
anti-CLDN18.2 IgG1 monoclonal antibody, a cleavable linker and topoisomerase I in- body telisotuzumab linked to the microtubule inhibitor monomethyl auristatin E. In the
hibitor, with a drug-to-antibody-ratio (DAR) of 8. Methods: This is a multicenter, first-in- LUMINOSITY trial (NCT03539536), Teliso-V monotherapy demonstrated efficacy in EGFRwt
human study of TQB2103 in patients (pts) with previously treated advanced solid tu- pts with c-Met OE ($25% tumor cells at 3+ intensity by IHC) (Camidge et al. JCO 2024;42:
mors. This study comprised of a dose escalation part in patients regardless of Clau- 3000-11). We used ctDNA molecular profiling to investigate baseline (BSL) and longitudinal
din18.2 expression level and a dose expansion part in patients specified by CLDN18.2 changes in pts’ tumor mutational spectrum, and to identify potential mechanisms of tumor
positive expression. The primary objectives were to assess safety and tolerability and response and drug resistance to Teliso-V. Methods: Pts received 1.9 mg/kg Teliso-V in-
determine the recommended phase 2 dose. Secondary objectives were to assess the travenously Q2W. In total, 83 pts with ctDNA data and evaluable tumor assessments in Stage
2 were included in the analysis. Plasma ctDNA was collected at multiple timepoints and
pharmacokinetics and preliminary anti-tumor activity. Results: As of December 16
analyzed for genomic alterations using the PGDx elio Complete NGS assay (521 genes).
2024, 59 pts were enrolled to receive TQB2103 intravenously every 3 weeks at 7 dose
Variants with allele frequency (VAF) , 0.3% or from putative clonal hematopoiesis of in-
level (range from 0.5 to 6.0mg/kg). One patient experienced dose-limiting toxicity (DLT) determinate potential genes were removed. High ctDNA levels were defined as having a mean
of grade 3 transaminase elevation at 0.5mg/kg which might also be related to the (m)VAF $median (2.05%), and low ctDNA levels as mVAF , median. Molecular response
comorbidity of choledocholithiasis. Fifty-six (94.9%) patients experienced at least one (MR) was defined as having $50% reduction in the mVAF vs BSL levels without gene
treatment-related adverse event (TRAE). The most frequent TRAEs were nausea (72.9%), amplification. Mutational profiles and their association with RECIST-defined tumor response
vomiting (64.4%), appetite decreased (57.6%), hypoalbuminemia (49.2%), anemia and/or drug resistance were assessed. Results: Overall, pts with high BSL ctDNA levels had
(49.2%), white blood cell decreased (44.1%), and asthenia (44.1%). The most frequent an ORR (28.6%, 12/42) similar to all EGFRwt pts with c-Met OE (28.6%, 46/161) and were not
grade $3 TRAEs were anemia (11.9%) and neutrophil count decreased (10.2%). Most of statistically different vs pts with low BSL levels. However, pts with low BSL ctDNA had longer
AEs were grade 1 or grade 2 and manageable. Among the 30 response evaluable pts with median OS (16.3 vs 8.5 mo) and mPFS (8.1 vs 5.4 mo) vs those with high BSL levels. Although
CLDN18.2 expression, the ORR and DCR were 20% and 76.7%, respectively. Shrinkage of the total pts with genomic alterations (GA) in this analysis was limited, KRAS GA were one of
the target lesions occurred in 17(56.7%) patients. In patients with CLDN18.2 moderate to the most common mutations detected at BSL (24%, 20/83 pts). ORR to Teliso-V among pts
high expression, the ORR was 42.9% of gastric cancer at 5mg/kg. Surprisingly, all of 3 with the actionable GA (AGA) of KRAS G12C was 100% (5/5). Conversely, among pts with non-
response-evaluable biliary tract cancer had shrinkage of the target lesions at the first AGA KRAS G12V/D/A and Q61H/L, the ORR was 23% (3/13). Additional AGAs were found in
assessment, and 1/3 achieved partial response. Conclusions: TQB2103 demonstrated BSL ctDNA, including 1 BRAF V600E, 3 MET ex14del, 1 EGFR G719C, and 2 RET1-KIF5B
encouraging anti-tumor activity in CLDN18.2 positive solid tumors, with a favorable translocations; none had response to Teliso-V. Pts with a MR at week 6 had higher ORR (35%
safety profile. The findings support further development of TQB2103 monotherapy or in vs 23%), longer median OS (15.5 vs 12.2 mo), and median PFS (8.5 vs 5.7 mo) vs those who did
combination with other anti-cancer therapies. Clinical trial information: NCT05867563. not. One pt with stable disease had several new AGAs detected in circulation at week 6,
Research Sponsor: Chia Tai Tianqing Pharmaceutical Group Co., Ltd. including activating EGFR ex20ins. At week 24, clinical progression was accompanied by gene
amplifications of ERBB2, FGF4, FGFR4, and FGFR3. Other types of pharmacodynamic changes
in ctDNA that could predict clinical response or drug resistance will be presented.
Conclusions: ctDNA is a promising biomarker in predicting Teliso-V activity. Confirmatory
research is planned in larger pt cohorts and/or with tissue-based NGS analyses. Clinical trial
information: NCT03539536. Research Sponsor: AbbVie, Inc.; n/a.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 195s

3028 Poster Session 3029 Poster Session

First-in-human phase 1 dose escalation trial of OMTX705, a novel anti- A first-in-human clinical study of 9MW2921, a novel TROP-2 antibody-drug
fibroblast activation protein (FAP) antibody drug conjugate (ADC), in mono- conjugate (ADC), in patients with advanced solid tumors. First Author: Shuiping
therapy and in combination with pembrolizumab in patients with solid Gao, Fudan University Shanghai Cancer Center, Shanghai, China
tumors. First Author: Javier Torres-Jiménez, Hospital Universitario 12 de Octubre, Background: TROP-2 (trophoblast cell surface antigen 2) is commonly overexpressed in
Madrid, Spain multiple solid tumors and associated with poor prognosis. 9MW2921 is a novel TROP-2
Background: Cancer Associated Fibroblasts (CAFs) are key components of tumor micro- ADC developed with a site-specific linker to conjugate the class of novel camptothecin-
environment and have immunosuppressive functions. FAP is expressed in a restricted fashion based payload Mtoxin, with a drug-to-antibody-ratio (DAR) of 4. Here we report the
on CAFs. OMTX705 is a first-in-class ADC targeting FAP with a novel tubulysin payload. safety and efficacy data of 9MW2921 in patients with advanced solid tumors in a phase 1
OMTX705 demonstrated a good safety profile in relevant toxicology models and high linker study. Methods: 9MW2921 was administered by intravenous infusion at doses of 1.0-
stability in plasma. We report the dose escalation phase 1 trial of OMTX705 in monotherapy 6.0 mg/kg once every 3 weeks. Primary objectives were assessment of dose-limiting
and in combination with pembrolizumab (PEM). Methods: Patients (pts) with advanced toxicity, safety and the recommended phase 2 dose/maximum dose. Results: As of 12
carcinomas or sarcomas received 1-18 mg/kg of OMTX705 monotherapy and 2-10 mg/kg with November, 2024, thirty-nine patients (pts) were enrolled and treated at dose levels of 1.0
standard PEM. Escalation used a classical 3+3 design with backfilling. OMTX705 schedule is (N = 1), 2.0 (N = 3), 2.5 (N = 12), 3.0 (N = 20) and 4.5 (N = 3) mg/kg. The average age of all
Day 1, 8 every 21 days. The primary endpoint is safety and key secondary are efficacy,
patients was 55.6 (range: 37-72) years with 20.5% male and 79.5% female. The median
pharmacokinetics and biomarkers. Biopsy and blood samples were collected for biomarker
prior therapy lines were 2 (range: 1~11); 48.7% pts treated after immunotherapy. Three
analysis. Results: A total of 78 pts have been dosed: 31 pts in monotherapy in 9 dose cohorts
and 47 in combination in 7 cohorts of 3 pts each plus 2 backfilling cohorts at 4 and 7.5 mg/kg
patients at 4.5mg/kg experienced at least one dose limiting toxicity (DLT), and this dose
in pancreatic adenocarcinoma (PDAC) and microsatellite stable (MSS) colorectal cancer level was considered intolerable. No other pts was observed DLTs at the 1.0~3.0 mg/kg
(CRC). Median age was 60, 43% male and ECOG PS 0 in 44%. Main histologies were PDAC 33% groups. The most common $grade 3 ($5% pts)TRAEs were stomatitis, anemia, white
and MSS CRC 21% with median of 2 (1 to 5) and 3 (1 to 5) prior lines of therapy, respectively. blood cell (WBC) count decreased, neutropenia, lymphocyte count decreased, rash,
Median treatment exposure was 44 days (range 8 to 113) in monotherapy and 92 days (1 to vomiting and platelet count decreased. There were no TRAEs leading to death. 38 pts
422) in combination. OMTX705 relative dose intensity was ~100% in all dose levels. No DLT were evaluable for efficacy with at least one post-baseline tumor assessment, 12 pts
has been observed. The most frequent related TEAEs were asthenia 35%, AST increased 14%, achieved partial response and 16 pts maintained stable disease. The ORR of 3.0 mg/kg
diarrhea 8%, anemia 8%, and nausea 8%. Grade 3 related TEAEs (pts): anemia (2), immune- was 42.1% (8/19) and DCR was 84.2% (16/19). The ORR, DCR of 3.0 mg/kg in patients
mediated hepatitis (2), GGT increased (1), neutropenia (1) and asthenia (1). In monotherapy, diagnosed with endometrial cancer (4 pts), HR+/HER2- breast cancer (4 pts), HER2-
best response was SD in 26%. In combination, PR was achieved in 4% (1 MSS CRC and 1 gastric cancer (4 pts) and Non-squamous non-small cell lung cancer (4 pts) were 75%,
PDAC; DOR 11+ and 8 months, respectively), SD in 33%, PD in 51%, and NE in 13%. In 13 pts 100%; 50%, 75%; 50%, 100%; 25%, 100%, respectively. Conclusions: The data indicated
there was target lesion reduction: median -17% (-46 to -1%). Median PFS was 1.4 months (0 to that 9MW2921 has acceptable tolerability and promising anti-tumor activity in patients
14+). In combination, 20% PDAC (4/20) and 21% CRC (3/14) pts showed PFS . 4 months. 2/3 with advanced EC, HR+/HER2- BC, HER2- GC and nsq-NSCLC. Clinical trial information:
NSCLC with previous checkpoint inhibitor treatment, 2 PDAC, and 2 MSS CRC showed PFS . NCT05990452. Research Sponsor: Mabwell (Shanghai) Bioscience Co., Ltd.
7 months. High FAP expression (H-score . 30) was observed in 86% PDAC, 64% CRC and 50%
other carcinomas. CD8+ and CD56+ immune-cell infiltration in tumor biopsies and down-
regulation of immunosuppressive cytokines in plasma samples were observed in PDAC and
CRC best responders. OMTX705 tubulysin payload was detected in both CAFs and tumor
epithelial apoptotic regions. Conclusions: OMTX705 is a novel anti-FAP ADC with excellent
safety profile. The combination with PEM showed disease control in some heavily pretreated
PDAC, MSS CRC and NSCLC. Changes in immune infiltrates and cytokines suggest that
OMTX705 may revert CAF-mediated immunosuppression. Clinical trial information:
NCT05547321. Research Sponsor: Oncomatryx Biopharma.

3030 Poster Session 3032 Poster Session

Efficacy and safety results of a multi-center phase I study of utidelone RC48-ADC combined with radiotherapy and immunotherapy as salvage
capsule, a novel oral microtubule inhibitor, in advanced solid tumor patients. therapy for advanced solid tumors with HER2 expression: A multicenter,
First Author: Judy S. Wang, Florida Cancer Specialists/Sarah Cannon Research Institute, phase II trial. First Author: Meiling Xu, Center for Cancer Diagnosis and Treatment, The
Sarasota, FL Second Affiliated Hospital of Soochow University, Suzhou, China
Background: Utidelone is a novel microtubule inhibitor, whose injectable formulation Background: Antibody-drug conjugates (ADC) have demonstrated efficacy in treating
(UTD1) has been approved for advanced breast cancer in China since 2021. Attempts to tumors with HER2 over expression. However, the clinical benefits of ADCs are limited in
develop oral microtubule inhibitor have not made significant progress; no oral micro- tumors with lower HER2 expression. The combination of ADCs, radiotherapy, and im-
tubule inhibitors have been approved in the United States to date. Utidelone is insus- munotherapy has shown promising feasibility with HER2-expressing tumors across
ceptible to P-glycoprotein-mediated efflux, thereby optimizing it for oral administration on various cancer types. This approach can enhance spatial and physicochemical synergistic
an intermittent schedule. Utidelone capsule (UTD2) can significantly improve medication effects, resulting in a more diverse and increased release of tumor antigens. Subsequently,
compliance and the convenience of clinical application. This is the first-in-human study of PD-1 inhibitors activate effector T cells, generating a robust immune response that targets
UTD2, and the trial has been completed with final results presented here. and eliminates tumor cells. A single-arm, multicenter phase II trial was initiated to evaluate
Methods: Eligible patients were aged $18, with an ECOG PS of 0-1, life the clinical efficacy of RC48-ADC combined with radiotherapy and immunotherapy, in
expectancy $12weeks, pathologically confirmed advanced solid tumor refractory to prior HER2-expressing advanced solid tumors. The findings from this trial may establish a new
standard therapies. Patients were treated with UTD2 monotherapy. The starting dose was salvage treatment strategy for tumors with low HER2 expression. Methods: This study
5-day 25 mg/m2/d for 2 patients, with planned escalation to 5-day 50, 75, 100 mg/m2/d enrolled patients with advanced, HER2-expressing (IHC 1+, 2+, or 3+) solid tumors that had
and 7-day 70 mg/m2/d for 2, 6, 3 and 2 patients, repectively in a 21-day cycle. The primary progressed following standard therapies or due to intolerance. Participants received RC48
objective was to determine DLT and the MTD. Secondary objectives included efficacy, PK (disitamab vedotin, 2.0 mg/kg on day 1), followed by radiotherapy every other day (2-3
profile and RP2D. Results: 18 advanced solid tumor patients were enrolled (3 didn’t fractions of 5-8 Gy), GM-CSF 200 mg on days 3-7), sequential IL-2 (2 million IU on days 8-
complete DLT observation) with median age of 60.8 years (range 29.0-81.0), 9 females 12), and a PD-1 inhibitor administered within one week after completing radiotherapy. This
and 9 males. All patients had received prior treatment in advanced settings with maximal regimen was repeated every three weeks. The primary endpoint was the objective response
9 lines. Two DLTs of Grade 3 and Grade 4 diarrhea occurred, one at 5-day 100 mg/m2/d rate (ORR). Results: As of the cutoff date (December 31, 2024), 52 patients were enrolled,
and one at 7-day 70 mg/m2/d. Considering the total dose per cycle for both cohorts were including 10 with gynecological cancers, 10 with pancreatic cancer, and 32 with various
similar, the MTD was determined to be 5-day 75 mg/m2/d via SMC. 11 patients were other tumor types (including breast, gastric and colorectal cancers). All participants had
evaluated for efficacy with an outcome of 1 CR (ovarian cancer), 1 PR (ovarian cancer), 7 evaluable data. According to RECIST 1.1, the overall ORR was 36.5%, with two patients
SD (testicular Sertoli cell tumor, NSCLC*2, pancreatic adenocarcinoma*2, appendiceal achieving a complete response (CR) that lasted nearly two years, maintaining minimal
adenocarcinoma and soft tissue sarcoma), with the longest DoT of 12 cycles. The ORR residual disease negative status. The ORRs for patients with HER2 expression of 1+, 2+,
was 18.2% and the CBR was 81.8%. PK results showed that the characteristics of and 3+ were 29.0%, 43.4%, and 60.0%, respectively. The median progression-free survival
utidelone were consistent with a two-compartment model. Compared to single-dose (PFS) for all patients was 5.9 months (95% CI: 4.1–9.7 months). The median overall
administration, there was no accumulation of utidelone in plasma upon multi-dose. The survival (OS) for all patients was 14.3 months (95% CI: 8.6–15.7 months).Treatment-
most frequent TEAEs were Grade 1/2, including diarrhea, fatigue, nausea, peripheral related adverse events were predominantly mild (grade 2 or lower), including fatigue, hair
sensory neuropathy, vomiting, and decreased appetite ($20% incidence rate), which loss, nausea, fever, and rash. Only three patients (5.8%) experienced grade 3 adverse
recovered with supportive treatments. The $Grade 3 TRAE included diarrhea (27.8%) and events. Conclusions: The results indicate promising efficacy and manageable safety,
fatigue (5.6%). Conclusions: This completed study demonstrates encouraging anti- with a favorable short-term tumor response rate. This suggests that the combination of
tumor activity with manageable safety of UTD2 in patients with heavily pre-treated RC48-ADC, radiotherapy, and immunotherapy could serve as an effective salvage therapy
advanced solid tumors. The results support continuing development of UTD2 for the option for patients with HER2-expressing advanced solid tumors. The combination therapy
upcoming phase II/III studies for gastric and ovarian cancers. Clinical trial information: appears to enhance the synergistic effects of radiotherapy and immunotherapy. Clinical
NCT05681000. Research Sponsor: Biostar Pharma, Inc. trial information: NCT0511550. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
196s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3033 Poster Session 3034 Poster Session

First-in-human study of BG-C9074, a B7-H4-targeting ADC in patients with Efficacy and safety of XNW27011, a Claudin 18.2 targeting antibody drug
advanced solid tumors: Preliminary results of the dose-escalation phase. conjugate with topoisomerase 1 inhibitor payload, in patients with Claudin
First Author: Cesar Augusto Perez, Sarah Cannon Research Institute at Florida Cancer 18.2 positive gastric/gastroesophageal junction cancer: Results from on-
Specialists, Orlando, FL going phase I/II study. First Author: Jinming Yu, Affiliated Hospital of Shandong First
Background: B7-H4 is a transmembrane glycoprotein in the B7 superfamily with limited Medical University, Jinan, China
expression in normal tissue but is upregulated in solid tumors including chol- Background: CLDN18.2 is a clinically validated target for cancer treatment. XNW27011 is a
angiocarcinoma, breast, ovarian, and endometrial cancers. BG-C9074 is an investiga- CLDN18.2 targeted ADC conjugated with a novel topoisomerase 1inhibitor (topo1i). Dose-
tional topoisomerase I inhibitor antibody-drug conjugate. This abstract presents the escalation study of XNW27011 demonstrated favorable safety, pharmacokinetics, and prom-
initial results of monotherapy dose escalation from the ongoing phase 1 study. ising preliminary efficacy in advanced solid tumors. Here we report the results of XNW27011 in
Methods: BG-C9074-101 (NCT06233942) is a first-in-human, multicenter study CLDN18.2+ GC/GEJC pts from ongoing expansion cohorts. Methods: Pts with CLDN18.2+
(TC$5%, IHC $ 2+), advanced/metastatic solid tumors progressed on standard therapy and an
designed to evaluate the safety, tolerability, pharmacokinetics, and preliminary anti-
ECOG PS of 0-2 are eligible to be enrolled in dose expansion cohorts. Pts received XNW27011 iv
tumor activity (per RECIST v1.1) of BG-C9074 as monotherapy and in combination with infusion Q3W at doses of 2.4, 3.0 and 3.6 mg/kg. 1st endpoint is ORR, 2nd endpoints include
tislelizumab in patients with advanced solid tumors. Patients with histologically or safety, other efficacy parameters, PK, ADA, and correlation between CLDN18.2 expression and
cytologically confirmed locally advanced, unresectable, or metastatic solid tumors, efficacy. Results: As of Dec 28th, 2024, a total of 116 pts with CLDN18.2+ solid tumors including
irrespective of B7-H4 expression, received BG-C9074 intravenously every 3 weeks in 84 GC/GEJC pts were enrolled in expansion cohorts at doses of 2.4 mg/kg, 3.0 mg/kg and 3.6
sequentially escalating dose cohorts ranging from 1 to 7 mg/kg. Results: As of January mg/kg, with median age of 59 years, median lines of prior treatment 2, 80.2% received
22, 2025, 55 patients with advanced tumors (n = 25, ovarian cancer; n = 16, breast checkpoint inhibitors and 18.6% topo1i-containing therapies. The most common any grade
cancer; n = 10, cholangiocarcinoma; n = 4, other tumor types) received BG-C9074 TEAE ($ 20%) in all patients were nausea, vomiting, anemia, appetite ↓, WBC ↓, neutrophil ↓,
monotherapy. Three patients experienced dose-limiting toxicities including fatigue (6 asthenia, hypoalbuminemia, platelet ↓, body weight ↓, and hypokalemia. The most common $
mg/kg), and febrile neutropenia and thrombocytopenia (7 mg/kg). Treatment-emergent G3 TEAEs ($5%) were neutrophil ↓, WBC ↓, anemia, lymphocyte ↓, and asthenia. TEAEs leading
adverse events (TEAEs) were reported in 48 patients (87.3%) with grade $3 TEAEs to dose interruption at 2.4, 3.0 and 3.6 mg/kg were 15.2%, 26% and 60%, dose reduction 10.9%,
occurring in 27.3% of patients. The most common TEAEs were nausea (45.5%), fatigue 26%, and 65%, and dose discontinuation 10.9%, 8%, and 0%. 1 pt at 3.0 mg/kg experienced TEAE
(38.2%), and neutropenia (32.7%), with neutropenia being the most frequent grade $3 leading to death (pneumonia). Safety profile was consistent with that of dose escalation part. In
the 84 GC/GEJC pts enrolled in dose expansion, 75 pts were evaluable with at least one post
TEAE (16.4%). Among 39 efficacy-evaluable patients, eight (20.5%) partial responses (n
baseline scan. The BOR and DCR across dose groups were 46.7% and 88.0%, respectively.
= 4, confirmed; n = 4, unconfirmed) were observed. Conclusions: BG-C9074 showed a
Efficacy in each dose group was summarized in the table below. The median follow up was 4.3M,
manageable safety/tolerability profile in patients with B7-H4 advanced solid tumors. 4.0M and 7.0M for 2.4, 3.0, and 3.6 mg/kg. Preliminary anti-tumor activity was also observed in
Preliminary clinical responses were observed at multiple dose levels across various pts who had prior CPI and topo 1i containing treatments, as well as in other CLDN18.2+ solid
tumor types without selection for B7H4 expression. Dose-escalation and dose-level tumor pts. Conclusions: In the expansion cohorts,XNW27011 demonstrated promising anti-
expansion are ongoing and updated clinical data will be presented at the conference. tumor activity and favorable safety profile in GC/GEJC pts with wide expression level of
Clinical trial information: NCT06233942. Research Sponsor: BeOne Medicines Ltd. [Link] results support further development of XNW27011 in CLDN18.2+ GC/GEJC.
Clinical trial information: CTR20231735. Research Sponsor: Evopoint Biosciences, Co. Ltd.
2.4 mg/kg 3.0 mg/kg 3.6 mg/kg
N=27 N=30 N=18
BOR, n (%) 7 (25.9%) 16 (53.3%) 12 (66.7%)
PR (confirmed) 4 (14.8%) 9 (30%) 6 (33.3%)
cPR Pending 2 (7.4%) 5 (16.6%) 3 (16.7%)
DCR, n (%) 23 (85.2%) 27 (90.0%) 16 (88.9%)

3035 Poster Session 3036 Poster Session

Results from a phase 1/2 study of 7MW3711: A novel B7-H3 antibody-drug Results from a phase 1/2 study of 7MW3711: A novel B7-H3 antibody-drug
conjugate (ADC) incorporating a topoisomerase I inhibitor in patients with conjugate (ADC) incorporating a topoisomerase I inhibitor in patients with
advanced solid tumors. First Author: Zhiye Zhang, Department of Medical Oncology, lung cancer. First Author: Ziming Li, Shanghai Chest Hospital, Shanghai Lung Cancer
The First Affiliated Hospital of Henan University of Science and Technology, Luoyang, Clinical Medical Center, Shanghai, China
China Background: 7MW3711 is a B7-H3 targeting ADC comprised of a recombinant hu-
Background: 7MW3711 is a B7-H3 targeting ADC comprised of a recombinant hu- manized monoclonal anti-human B7-H3 antibody conjugated to the topoisomerase I
manized monoclonal anti-human B7-H3 antibody conjugated to the topoisomerase I inhibitor via a protease cleavable linker. B7-H3 is upregulated in several malignant
inhibitor via a protease cleavable linker. B7-H3 is upregulated in several malignant cancers, such as lung, ovarian, breast, prostate and esophageal cancer, which plays an
cancers, such as lung, ovarian, breast, prostate and esophageal cancer, which plays an important role in multiple processes such as tumor occurrence, development, and
important role in multiple processes such as tumor occurrence, development, and immune escape. Here we present the safety and efficacy data of 7MW3711 from a first-
immune escape. Here we present the safety and efficacy data of 7MW3711 from a first- in-human phase 1/2 study. Methods: The study enrolled patients (pts) with advanced
in-human phase 1/2 study. Methods: The study enrolled patients (pts) with advanced solid tumor across three segments: dose-escalation (D-esc), dose-expansion (D-exp)
solid tumor across three segments: dose-escalation (D-esc), dose-expansion (D-exp) and cohort-expansion. In the D-esc and D-exp phase, 7MW3711 was administered
and cohort-expansion. In the D-esc and D-exp phase, 7MW3711 was administered intravenously at doses of 1.5, 3.0, 4.5, 5.0, 6.0 mg/kg every three weeks (Q3W).
intravenously at doses of 1.5, 3.0, 4.5, 6.0 mg/kg every three weeks (Q3W); 4.0 mg/kg Results: As of the data cutoff on Jan 8, 2025, 37 pts with lung cancer were enrolled and
every two weeks (Q2W). Results: As of the data cutoff on Jan 2, 2025, 43 pts were received at least one dose of 7MW3711 (D-esc, n = 25; D-exp, n = 12), which included 16
enrolled and received at least one dose of 7MW3711 (D-esc, n = 15; D-exp, n = 28). At pts with small cell lung cancer (SCLC) and 21 pts with non-small cell lung cancer
baseline, the median of prior lines of therapy for all pts was 2 (range, 1-9). No dose- (NSCLC). At baseline, the median of prior lines of therapy for all pts was one (range, 1-5).
limiting toxicities (DLTs) were observed in the D-esc phase. The maximum tolerated Five pts experienced dose-limiting toxicities (2 pts at 5.0 mg/kg; 3 pts at 6.0 mg/kg),
dose (MTD) has not yet been reached. The most common Grade $3 TRAEs ($5% of pts) including decreased platelet count, decreased neutrophil count, myelosuppression and
were decreased white blood cell count, anemia, decreased neutrophil count, decreased appetite. The maximum tolerated dose (MTD) has not yet been determined.
decreased lymphocyte count, decreased platelet count, diarrhea, and hypokalemia. The most common Grade $3 TRAEs ($5% of pts) were decreased neutrophil count,
Among 33 pts treated with 7MW3711 at 4.0 mg/kg or above and reaching tumor as- decreased white blood cell count, anemia, decreased lymphocyte count, decreased
sessment, 8 partial responses (PRs) were observed. The objective response rate (ORR) platelet count, hyponatremia, hypokalemia, and myelosuppression. Among 25 pts
and disease control rate (DCR) were 24.2% and 84.8%, respectively. 15 pts diagnosed treated with 7MW3711 at 4.5 mg/kg or above and reaching tumor assessment, 9 partial
with esophageal cancer (EC), ovarian cancer (OC) and prostate cancer (PC) were enrolled responses (PRs) were observed. The overall objective response rate (ORR) and disease
at 4.5 mg/kg or above and were evaluable for tumor assessment. All EC pts had control rate (DCR) were 36.0% and 96.0%, respectively. 8 pts diagnosed with SCLC were
previously progressed after receiving platinum-based chemotherapy and immune enrolled at 4.5 mg/kg and were evaluable for tumor assessment. All 8 SCLC pts had
checkpoint inhibitors. All OC pts were platinum-resistant. All PC pts had previously previously progressed after receiving platinum-based chemotherapy and immune
progressed after receiving docetaxel and endocrine therapy. The ORR of EC, OC and PC checkpoint inhibitors. The ORR and DCR of them were 62.5% and 100.0%, respectively.
was 33.3%, 60.0% and 50.0%, respectively. The DCR of EC, OC and PC was 100.0%. Among pts with B7-H3 H-score . 5, the ORR and DCR of lung squamous cell carcinoma
Objective responses were also observed in pts with other solid tumor types, such as lung (Sq-NSCLC) at 4.5 mg/kg or above were 37.5% (3/8) and 87.5% (7/8), respectively.
adenocarcinoma and breast cancer. Conclusions: The data indicated encouraging Conclusions: The data indicated encouraging efficacy of 7MW3711 in SCLC and Sq-
efficacy of 7MW3711 in advanced EC, OC and PC. The safety profile showed adequate NSCLC. The safety profile showed adequate tolerability. The dose optimization and
tolerability. The dose optimization and expansion study is continuing to establish the expansion study is continuing to establish the RP2D for 7MW3711. Clinical trial in-
RP2D for 7MW3711. Clinical trial information: NCT06008366. Research Sponsor: formation: NCT06008379. Research Sponsor: Mabwell (Shanghai) Bioscience Co., Ltd.
Mabwell (Shanghai) Bioscience Co., Ltd.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 197s

3037 Poster Session 3038 Poster Session

First-in-human trial of SYS6010 combined with SYH2051 in patients with Precemtabart tocentecan (M9140), an anti-CEACAM5 ADC with exatecan
advanced gastrointestinal tumors. First Author: Rongbo Lin, Gastrointestinal payload, in patients with metastatic colorectal cancer (mCRC): Results from
Medical Oncology, Fujian Cancer Hospital, Fuzhou, China the dose optimization of the phase 1 PROCEADE CRC-01 study. First Author:
Background: SYS6010 is an antibody-drug conjugate (ADC) composed of an EGFR- Scott Kopetz, The University of Texas MD Anderson Cancer Center, Houston, TX
specific antibody, a cleavable linker, and JS-1, a topoisomerase I inhibitor, as its cy- Background: CEACAM5 is overexpressed in ~90% of CRCs, with limited expression on
totoxic payload. It targets EGFR, a transmembrane receptor tyrosine kinase overex- healthy cells. Precemtabart tocentecan (M9140), the first anti-CEACAM5 ADC with an
pressed in malignancies such as lung, breast, gastric, and colorectal cancers. SYS6010 exatecan payload (topoisomerase 1 inhibitor), showed a predictable, manageable safety
induces DNA damage leading to apoptosis, while resistance may occur through ATM- profile and promising early clinical activity in the dose escalation of the Phase 1 PROCEADE-
mediated DNA repair. SYH2051, an ATM inhibitor, disrupts DNA repair, enhancing CRC-01 study (NCT05464030) in heavily pretreated patients with mCRC. Methods: This
SYS6010-induced apoptosis. This combination is hypothesized to exert synergistic anti- global Phase 1 study in 3L adult patients with locally advanced/mCRC (ECOG PS #1; previous
tumor effects. Methods: This first-in-human clinical trial employed a single-center, irinotecan therapy) evaluates clinical activity, safety, and tolerability of precemtabart
open-label, non-randomized design to evaluate the safety, tolerability, and preliminary tocentecan. Here, we report on dose optimization of precemtabart tocentecan tested at 2.8
mg/kg Q3W (Arm A1) or 2.4 mg/kg Q3W (A2; 1:1 randomization) to select the recommended
efficacy of SYS6010 combined with SYH2051. Patients with advanced gastrointestinal
phase 2 dose (RP2D). Results: As of Jan 2025, 60 patients (recruited Apr–Oct 2024) had
tumors expressing EGFR who had progressed on at least one prior line of standard
been treated (A1, n = 29; A2, n = 31). Median age was 60.0 years, and 51.7% were male. In A1,
therapy were enrolled. SYS6010 was administered intravenously at a dose of 3.2 mg/kg 18 (62.1%) patients remained on treatment and 16 (51.6%) in A2. Treatment-emergent AEs
on day 1 of each 14-day cycle, while SYH2051 was given orally at doses of 40 mg or (TEAEs) were reported in all patients; grade $3 in 38 (63.3%) patients (A1: n = 19 [65.5%]; A2:
80 mg, once daily, for five consecutive days within the same cycle. Safety was assessed n = 19 [61.3%]); anemia and neutropenia (any grade; grade $3) were most common. Serious
using CTCAE v5.0, and efficacy was evaluated according to RECIST v1.1 criteria. TEAEs were reported in 18 (30.0%) patients (A1: n = 8 [27.6%]; A2: n = 10 [32.3%]). Grade $3
Results: As of December 31, 2024, 25 patients were enrolled, including 18 with co- hematologic AEs were reported in 32 (53.3%) patients: anemia (A1, n = 9; A2, n = 10),
lorectal cancer and 7 with gastric cancer. Twelve patients had received $3 prior lines of neutropenia (A1, n = 14; A2, n = 12), thrombocytopenia (n = 6 both), leukopenia (A1, n = 7; A2, n
therapy. Among 6 evaluable gastric cancer patients, 3 achieved partial response (PR) = 6), lymphopenia (A1, n = 1; A2, n = 2), febrile neutropenia (n = 3 both), and pancytopenia (A1,
and 3 stable disease (SD), resulting in an objective response rate (ORR) of 50% and a n = 0; A2, n = 1). Treatment was discontinued in 26 (43.3%) patients (A1: progressive disease
disease control rate (DCR) of 100%. The median progression-free survival (PFS) was (PD), n = 9, patient withdrawal, n = 1, other, n = 1; A2: PD, n = 14, death, n = 1). No treatment-
approximately 5.8 months (data not mature), and 3 patients remained on treatment. related deaths were reported. Overall, PK profiles were consistent with previous data, with
Among 18 colorectal cancer patients (9 with KRAS mutations and 9 wild-type), pre- overlap attributed to high between-subject variability. Partial responses were reported in 7
liminary analysis showed a median PFS of approximately 4.2 months in wild-type KRAS (24.1%; n = 4 [13.8%] confirmed) patients in A1 and 3 (9.7%; n = 1 [3.2%] confirmed) in A2 (all
patients (data not mature). Common treatment-related adverse events (TRAEs) included responders remain on treatment), stable disease in 15 (51.7%) and 21 (67.7%), and PD in 5
hematologic toxicity, gastrointestinal symptoms, and fatigue. Frequently observed (17.2%) and 6 (19.4%) patients, respectively. DCR at 12 weeks was 72.4% in A1 and 67.7% in
TRAEs were fatigue (60%), decreased appetite (56%), leukopenia (56%), anemia (48%), A2. Conclusions: These preliminary results corroborate the encouraging efficacy and safety
neutropenia (48%), nausea (48%), thrombocytopenia (36%), and hypoalbuminemia data from the dose escalation part of the PROCEADE CRC-01 study, with no new relevant
(32%). Grade $3 TRAEs occurred in 12 patients (48%), including neutropenia (7 pa- safety findings. ORR was higher at 2.8 mg/kg, with similar tolerability at both doses. The ORR
tients), anemia (6 patients), thrombocytopenia (3 patients), vomiting (3 patients), of 24.1% (13.8% confirmed) at 2.8 mg/kg compares favorably with current monotherapy SoCs
leukopenia (2 patients), interstitial lung disease (1 patient), infection (1 patient), and (ORRs 1-2%) and recent phase 3 data with trifluridine–tipiracil + bevacizumab (ORR 6.1%) in
3L+ mCRC. These results suggest 2.8 mg/kg as the RP2D for further development in CRC, and
elevated bilirubin (1 patient). No treatment-related deaths were reported.
other solid tumors (NCT06710132). More mature data, including PFS, will be presented at the
Conclusions: SYS6010 combined with SYH2051 was well tolerated and demonstrated
congress. Clinical trial information: NCT05464030. Research Sponsor: the healthcare
preliminary antitumor activity in advanced gastrointestinal tumors, particularly in gastric business of Merck KGaA, Darmstadt, Germany (CrossRef Funder ID: 10.13039/100009945).
cancer. Further evaluation is ongoing. Research Sponsor: CSPC Pharmaceutical Group
Limited.

3039 Poster Session 3040 Poster Session

Rinatabart sesutecan (Rina-S) for patients with advanced endometrial can- EVEREST-2: Initial data of the logic-gated Tmod chimeric antigen receptor
cer: First disclosure from dose expansion cohort B2 of the GTC1184-01 T-cell (CAR T) therapy A2B694 for patients with solid tumors associated with
study. First Author: Ira Seth Winer, Wayne State University, Barbara Ann Karmanos mesothelin (MSLN) expression and with human leukocyte antigen (HLA)
Cancer Center, Detroit, MI loss of heterozygosity (LOH). First Author: Salman Rafi Punekar, Perlmutter Cancer
Background: Rina-S is an investigational antibody-drug conjugate targeting folate receptor Center, NYU Langone Health, New York, NY
alpha with a novel hydrophilic protease-cleavable linker and a topoisomerase I inhibitor, Background: EVEREST-2 is a first-in-human, phase 1/2 trial to assess the safety and
exatecan payload. Patients (pts) with advanced endometrial cancer (EC) who progress after efficacy of A2B694, an autologous, logic-gated Tmod CAR T therapy targeted to MSLN,
programmed death-ligand 1 [PD-(L)1] inhibitor plus chemotherapy have very poor prognoses which is normally expressed in the mesothelium and can be upregulated in many solid
and limited, ineffective treatment options (objective response rate [ORR] , 16% and median tumors. A2B694 is designed to overcome challenges of on-target, off-tumor toxicity that
progression-free survival , 5 months with single-agent chemotherapy); thus, there is urgent have limited other MSLN-targeted approaches by combining a CAR-activating receptor
need for novel therapies. In the dose escalation cohort, single-agent Rina-S showed pre- targeting MSLN with a blocker CAR that recognizes HLA-A*02, to distinguish between
liminary anti-tumor activity in pts with heavily pretreated EC. Here we first report results for normal and tumor cells (Tokatlian, et al. J Immunother Cancer. 2022). Methods: Adults
single-agent Rina-S in pts with heavily pretreated EC from dose expansion cohort B2 of the with recurrent unresectable, locally advanced, or metastatic cancers with MSLN ex-
phase 1/2 GCT1184-01 study (NCT05579366). Methods: Pts with metastatic or unre- pression who have progressed after standard-of-care therapy are eligible for EVEREST-2.
sectable EC who received prior platinum-based chemotherapy and a PD-(L)1 inhibitor Enrollment to EVEREST-2 and collection of T-cells occurs through the ongoing pre-
received either Rina-S 100 mg/m2 or 120 mg/m2 every 3 weeks after initial enrollment with screening study BASECAMP-1 (NCT04981119). When clinically appropriate, A2B694 is
120 mg/m2 only. The primary endpoint was safety and tolerability of Rina-S. Secondary manufactured from cryopreserved T cells, and patients undergo lymphodepletion before
endpoints included ORR and disease control rate (DCR). Results: As of data cutoff No- A2B694 infusion. The dose-escalation phase is evaluating the safety and tolerability of
vember 22, 2024, 64 pts with heavily pretreated EC (median 3 prior lines [range 1-8]) received A2B694 to identify the recommended phase 2 dose (RP2D). Dose escalation was started at
Rina-S 100 mg/m2 (n = 22) or 120 mg/m2 (n = 42) for a median treatment duration of 1x108 Tmod positive cells (dose level [DL] 1) and will increase up to 14x108 in combination
15.9 weeks. Most pts had ECOG PS 1 (64.1%), approximately half (46.9%) were aged $70 with low-dose IL-2 (DL 5). The dose-expansion phase will confirm RP2D and collect
years, and 48.4% had received prior radiotherapy. Pts primarily had endometrioid carcinoma biomarker data to further characterize A2B694. Results: As of January 15, 2025, 5
(45.3%) followed by serous carcinoma (26.6%). The most common (. 25%) treatment- participants (median age: 60 years; range, 50-84) have enrolled on EVEREST-2 and re-
emergent adverse events (TEAEs) were similar across doses and were primarily cytopenias ceived A2B694 at DLs 1-2; participants had ovarian cancer (n = 3), pancreatic cancer (n =
and grade 1-2 gastrointestinal events (nausea, vomiting, decreased appetite). Grade 3-4 1), and non-small cell lung cancer (n = 1) and had received a median of 4 prior lines of
cytopenia included neutropenia (48.4%), anemia (35.9%) and thrombocytopenia (21.9%). therapy (range, 1-7). Lymphodepleting chemotherapy was well tolerated with no significant
TEAEs led to Rina-S dose reductions in 15.6% of pts and discontinuation of Rina-S in 3.1% of cytopenias observed. The most common adverse events were lymphopenia (7 [14.6%]) and
pts; 37.5% of pts had serious TEAEs. There was 1 related (assessed by investigator) grade 5 decreased appetite (6 [12.5%]). There were no dose-limiting toxicities, cytokine release
TEAE at 120 mg/m2 confounded by comorbidities; no fatal TEAEs occurred at 100 mg/m2. syndrome, nor related neurotoxicity. One participant was admitted to the hospital for
No signals of ocular toxicity, neuropathy, or interstitial lung disease were observed. In decreased appetite. No long-term toxicities have been noted up to 7.5 months post-
efficacy-evaluable pts (median follow-up: 18.7 weeks), the unconfirmed ORR was 50%, infusion. Of the participants who have received A2B694, 5 were efficacy evaluable at DLs 1-
including 2 complete responses, with Rina-S 100 mg/m2 (n = 22) and 45.5% with 120 mg/m2 2. A2B694 was detected post-infusion in the peripheral blood in all patients. Additionally,
(n = 33). DCR was 100% and 81.8% with 100 mg/m2 and 120 mg/m2, respectively. Responses A2B694 was detected in an abdominal tumor biopsy from a patient with pancreatic cancer
were ongoing for 9 of 11 (81.8%) and 12 of 15 (80.0%) responders with 100 mg/m2 and 42 days post-infusion. Conclusions: The logic-gated approach was successful at reducing
120 mg/m2, respectively. Conclusions: Rina-S showed encouraging anti-tumor activity in toxicity seen with prior MSLN-targeted CAR T therapies, and A2B694 showed successful
pts with heavily pretreated EC and had a manageable safety profile consistent with previous CAR T expansion and tumor infiltration. The maximum tolerated dose has not been
reports. Further evaluation of single-agent Rina-S in pts with advanced EC is ongoing. reached, and results from the dose-escalation phase continue to determine the RP2D.
Clinical trial information: NCT05579366. Research Sponsor: Genmab A/S. Clinical trial information: NCT06051695. Research Sponsor: A2 Biotherapeutics, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
198s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3041 Poster Session 3042 Poster Session

ZL-1310, a DLL3 ADC, in patients with extensive stage small cell lung cancer: First-in-human (FIH) phase 1 study of CUSP06, a cadherin-6 (CDH6)-
Ph1 trial update. First Author: Manish R. Patel, Florida Cancer Specialists/Sarah directed antibody-drug conjugate (ADC), in patients with platinum-
Cannon Research Institute, Sarasota, FL refractory/resistant ovarian cancer and other advanced solid tumors. First
Background: Treatment options are limited for patients (pts) with extensive-stage small Author: Manish R. Patel, Sarah Cannon Research Institute/Florida Cancer Specialists,
cell lung cancer (ES-SCLC) that progress after platinum-based chemotherapy (chemo). Sarasota, FL
ZL-1310, a DLL3-targeted antibody drug conjugate (ADC) with a topoisomerase 1 in- Background: CDH6 is a transmembrane glycoprotein involved in cancer metastasis
hibitor payload and cleavable linker, demonstrated promising preliminary results in pts expressed in various tumors including ovarian cancer (OC), renal cell carcinoma (RCC),
with relapsed/refractory (r/r) ES-SCLC (Spira et al, ENA 2024). Here, we report updated cholangiocarcinoma (CCA), and uterine cancer. CUSP06 is an ADC composed of a hu-
data with additional pts and follow-up (NCT06179069). Methods: This is a two-part manized IgG1 mAb against CDH6 conjugated with a cleavable linker to exatecan, a
Phase I study of ZL-1310 administered intravenously every 3 weeks to pts with r/r SCLC topoisomerase I inhibitor. In preclinical studies, CUSP06 showed CDH6-dependent cell
who have progressed after at least one platinum-based chemo regimen. Part 1A is a growth inhibition in OC cell lines and tumor regression in CDH6-expressing OC, RCC, and
monotherapy dose escalation; Part 2 is a randomized dose optimization/expansion. other tumor models including those with low CDH6 expression supporting its use across
Study endpoints include safety parameters, objective response rate (ORR) per RECIST various indications and CDH6 expression levels. We report here the initial results from a
v1.1, duration of response (DOR), disease control rate (DCR) and pharmacokinetics (PK). FIH study of CUSP06. Methods: CUSP06-1001 is a Phase 1a/1b, open-label, multi-center
Exploratory tumor biomarkers, including DLL3 expression (expressed as H-score), are dose escalation and expansion study to evaluate safety, tolerability, pharmacokinetics,
examined. Results: As of 28 Jan 2025, 28 pts were enrolled in the dose escalation Part pharmacodynamics, recommended Phase 2 dose, and preliminary efficacy of CUSP06 in
1A and received ZL-1310 at dose levels ranging from 0.8 mg/kg to 2.8 mg/kg. The patients (pts) with platinum-refractory/resistant ovarian cancer (PRROC), advanced RCC
median time on study is 5.1 months (range 2.4-10.1+). Median age was 66 years (range and other advanced CDH6-positive solid tumors. Prescreening for CDH6 expression was
36-79); 43% were female; 75% had an ECOG performance status of 1; 93% progressed required for those pts with solid tumors other than OC or RCC. CUSP06 was administered IV
after prior anti-PD-L1 therapy; 39% had prior lung irradiation, and 36% had baseline brain every 21 days. Phase 1a followed a standard 3+3 dose escalation design and included dose
metastases. Any-grade treatment-related adverse events (TRAEs) occurred in 89% of pts enrichment cohorts at doses that had demonstrated safety. Phase 1b consists of dose
(Grade$3 TRAEs, 39%). One pt (2.4 mg/kg) had dose limiting toxicities of neutropenia expansion cohorts for pts with OC, RCC, and other CDH6-positive solid tumors to assess the
and thrombocytopenia; 5 pts underwent drug reduction and 5 had drug discontinued due safety, tolerability and efficacy at the RDE. Results: As of 03JAN25, 26 pts were dosed
with data available for 22 pts in Phase 1a (18 OC, 2 RCC, and 2 CCA) at doses from 1.6 mg/
to TRAE. Grade$3 TRAEs occurring in more than 1 patient include anemia (6 pts),
kg to 5.6 mg/kg. The median age was 60.5 yrs and the median prior therapies was 3. Of the
neutropenia (5), thrombocytopenia (3), WBC decreased (2), and interstitial lung disease
18 pts with OC, all pts received prior platinum and taxane, 67% received bevacizumab, and
(2). Objective responses were observed in 19 of 28 pts (68%), including one pt pending
22% received mirvetuximab (MIRV). All patients with RCC received an immune checkpoint
response confirmation, and a DCR of 93%. Responses were observed across all dose
inhibitor and a TKI. Related TEAEs occurred in 20 pts (91%). The most common related
levels and all levels of DLL3 expression (H-score range: 0-260), including one pt with TEAEs ( . 20%) were anemia (50%), neutropenia (46%), thrombocytopenia (46%), fatigue
prior tarlatamab treatment. Pts with baseline brain metastases had an 80% response (46%), nausea (36%), diarrhea (23%), and vomiting (23%). The most common related
rate and 100% DCR. Fourteen of 19 (74%) responders remain on study. PK data from 25 Grade $3 TEAEs were neutropenia, thrombocytopenia, and anemia. AEs led to discon-
pts showed dose-proportional increase of systemic ADC and payload exposure, with tinuation in 3 (14%) pts. Five of 14 GCIG-evaluable OC pts (36%) had a CA-125 response.
relatively low exposure of the payload and no significant accumulation. Among the 20 RECIST-evaluable pts, 5 partial responses (4 confirmed, & 1 unconfirmed)
Conclusions: ZL-1310 demonstrated a tolerable safety profile and promising antitumor including MIRV-pretreated pts, and 11 stable disease were observed. All PRs were in pts
activity in r/r ES-SCLC, including pts with brain metastases, pt with prior tarlatamab, and with platinum-resistant high grade serous OC, with an ORR of 36% (5/14). 18 pts were
in the setting of low DLL3 expression. Updated data, including patients in the ran- ongoing at the cutoff date. Conclusions: The preliminary data from the Phase 1a dose
domized Part 2 dose optimization, will be presented. Clinical trial information: escalation portion of this study showed acceptable tolerability and encouraging efficacy in
NCT06179069. Research Sponsor: None. pts with OC, which support further evaluation of CUSP06 in the Phase 1b expansion
cohorts. Clinical trial information: NCT06234423. Research Sponsor: None.

3043 Poster Session 3044 Poster Session

Phase 1 dose-escalation trial of talazoparib in combination with belinostat Impact of stereotactic ablative radiotherapy (SABR) on detection of ctDNA in
in select advanced solid tumors. First Author: Monika L. Burness, University of patients with early-stage lung cancer: Interim findings from the prospective
Michigan Rogel Cancer Center, Ann Arbor, MI SABR-DETECT trial. First Author: Saurav Verma, Verspeeten Family Cancer Centre,
Background: Inhibitors of histone deacetylase (HDACi) may synergize with poly (ADP- Schulich School of Medicine and Dentistry, Western University, London, ON, Canada
ribose) polymerase inhibitors (PARPi). This Phase 1 dose escalation trial tested the Background: Stereotactic ablative radiotherapy (SABR) is the preferred curative treatment for in-
combination of the PARPi talazoparib and the HDACi belinostat. Methods: This open- operable patients with stage I/IIA non–small-cell lung cancer (NSCLC). In cases where the tumor is
label study was conducted with a combined dose escalation of talazoparib (0.75 mg-1 inaccessible or biopsy carries a high risk of complications, SABR is offered even in the absence of a
tissue diagnosis, based on a high likelihood of malignancy as calculated by validated predictive
mg) and belinostat (500-1000 mg/m2) in subjects with advanced breast, ovarian, models. In these situations, a blood based liquid biopsy detecting circulating tumor DNA (ctDNA) can
prostate and pancreatic cancers. Primary objectives were to identify the safety, tol- serve as an aide to confirm malignancy and allow molecular testing. However, low ctDNA yield in early
erability, and recommended phase 2 dose (RP2D) of the combination. A TITE-CRM model stage NSCLC presents a challenge for diagnosis. This study hypothesizes that ctDNA detection rates
was used for dose level assignment and identification of RP2D. Results: A total of 25 will improve by combining assessment of pre- and post-SABR plasma samples. Methods: This is a
evaluable subjects were enrolled. Tumor types included breast cancer (10 subjects), multi-institutional study including two cohorts: 1) patients with suspected stage I/IIA NSCLC, with a
ovarian cancer (5), prostate cancer (5), and pancreatic cancer (5). Treatment-related pretreatment likelihood of malignancy of $60% on Herder or Brock models, and 2) patients with
adverse events (AEs) included nausea (n = 8, 32%), fatigue (n = 8, 32%), thromboembolic biopsy-proven NSCLC. SABR was delivered according to standard guidelines. Plasma was collected for
ctDNA analysis before and 24-72 hours following the first fraction of SABR. SHIELDING ULTRA MRD
events (n = 6, 24%), vomiting (n = 5, 25%), and anemia (n = 4, 16%). Treatment-related
panel of hotspot regions in 2365 cancer-related genes with ultra-high sensitivity was used for ctDNA
serious adverse events (SAEs) encompassed thromboembolic events (n = 4) and anemia analysis (mutation + fragment profile + CNV). In this pre-planned interim analysis, we report on the
(n = 1). Dose limiting toxicities (DLTs) occurred in 3 subjects including decreased white secondary objective: to assess the impact of SABR on detection rates of ctDNA. Results: Paired
blood cell count, fatigue, anemia, and failure to thrive. Seven subjects experienced stable plasma samples (pre- and post-SABR) were tested for 69 patients. After quality control analysis, 66
disease (SD), for a clinical benefit rate (CBR) of 28% (7/25); of those with SD, 6 were paired samples were analyzed and included in this interim analysis. The median age was 76 years
assigned to the highest dose (dose level 4) and 1 subject was assigned to dose level 3. (range, 56-89) and 36 (54%) were male. The median concentration of circulating free DNA (ng/mL) did
Duration of enrollment ranged from 18-291 days. Conclusions: In subjects with select not increase from pre- (5.5, inter quartile range (IQR): 3.3-8.1) to post-SABR (5.7, IQR: 4.1-7.6)
advanced solid tumors, talazoparib and belinostat combination therapy exhibits a fa- (P=0.82). The ctDNA detection rate in pre-SABR samples was 22.7% versus 27.3% in post-SABR
samples (Table). Interestingly, in 10 patients (15.2%), ctDNA became detectable in post-SABR samples
vorable safety profile and manageable toxicity. Nausea and fatigue were the most and in 7 patients (10.6%) the ctDNA was no longer detectable in the post-SABR samples. The ctDNA
common adverse events. Further studies are warranted to determine the efficacy of this remained undetectable in 41 patients (62.1%). 37.9% of patients had detectable ctDNA either before or
combination. Clinical trial information: NCT04703920. Research Sponsor: Pfizer; after SABR. Conclusions: The diagnostic yield of ctDNA for confirming malignancy in early stage
Acrotech. NSCLC is improved by testing both the pre- and the post-SABR samples, collected within 24-72 hours
after the first fraction of SABR. This approach may improve the diagnostic rates of liquid biopsies for
patients with presumed NSCLC undergoing SABR, warranting further investigation of ctDNA detection
before and shortly after treatment. Clinical trial information: NCT05921474. Research Sponsor:
Verspeeten Family Cancer Centre Medical Oncology Research Fund (MORF); Lawson Internal Research
Fund; Lung Cancer Canada Geoffrey Ogram Memorial Research Grant; ‘Crush it with Bev’ fundraiser.
ctDNA detection rates (N=66).
Pre-SABR Post-SABR n (%)
detected detected 8 (12.1%)
not detected detected 10 (15.2%)
detected not detected 7 (10.6%)
not detected not detected 41 (62.1%)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 199s

3045 Poster Session 3046 Poster Session

Prognostic significance of preoperational circulating tumor DNA detection Improved detection of circulating tumor DNA in patients with leiomyosar-
in early-stage NSCLC using a tissue-free blood test. First Author: Naixin Liang, coma with fragment size restriction. First Author: Nensi M. Ruzgar, Dana-Farber/
Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Boston Children’s Cancer and Blood Disorders Center, Boston, MA
Peking Union Medical College, Beijing, China Background: Prior work has shown that detection of circulating tumor DNA (ctDNA) at
Background: There is a growing need for risk evaluation and treatment monitoring in time of diagnosis of leiomyosarcoma (LMS) is associated with lower likelihood of objective
cancer care. However, current methods, mainly imaging, can be burdensome for patients response and patients with detectable ctDNA after two cycles of chemotherapy have worse
over time and prone to variability among readers. Recent research has highlighted the survival. However, because ctDNA exists at much lower concentrations in plasma
potential of tumor-informed circulating tumor DNA (ctDNA) testing for identifying compared to cell-free DNA of non-tumor origin, the sensitivity of these prognostic
postoperative minimal residual disease (MRD) due to its high sensitivity by tracking in- measures to tumor signals remains unclear. With increasing evidence of ctDNA fragments
dividualized mutations. Nevertheless, its use in early-stage patients prior to surgery is being shorter than the background non-tumor cell-free DNA, we sought to test whether
constrained by limited tissue availability and extended turnaround times. MUSETALK- restricting our analysis to smaller fragment sizes would improve detection of ctDNA in
Lung01 (multiomics sequencing technique application kick-start) is a prospective, lon- LMS. Methods: Plasma was serially collected from patients with LMS undergoing che-
gitudinal, observational study designed to evaluate the clinical utility of a tumor-naı̈ve motherapy. Cell-free DNA extracted from these samples was profiled by ultra-low-passage
ctDNA assay in patients with early-stage non-small cell lung cancer (NSCLC). whole-genome sequencing (ULPWGS). Copy number alterations were identified and used
Methods: Pretreatment plasma samples were prospectively collected from participants to detect ctDNA using the ichorCNA algorithm before and after restricting the dataset to
with stage I-IIIA NSCLC. Cell-free DNA was extracted and analyzed using a blood assay that fragments of 90-150bp (short). We compared detectability of ctDNA via ichorCNA between
interrogates both epigenetic and genetic information. The detection status and the es- analyses using all sequencing data and those using data restricted to short fragments.
timated fraction of ctDNA were reported by a machine learning classifier and an inde- Results: From 28 patients, 126 plasma samples were profiled. The median fragment
pendent statistical model, respectively. The calling threshold corresponding to a 99% length of cell-free DNA was 240bp (IQR 142-349) for patients with LMS, compared to 307bp
clinical specificity was verified in a subgroup from the THUNDER study (NCT04820868). (IQR 171-465, p , 0.001) in samples collected from healthy controls. Short fragments
Longitudinal data, including vital status, cancer status, and treatment, were collected for made up 19.48% of LMS libraries at diagnosis when ctDNA levels were highest, 15.27% of
up to 5 years. The study was approved by the institutional review board and all participants all LMS samples, and 12.96% of libraries from healthy controls. While ctDNA was de-
were required to provide informed consent. Results: A total of 289 participants from the tectable by ULPWGS in 17% of all samples, detection increased to 40% when analyzed
MUSETALK-Lung01 study were analyzed. Of these, 49% (141/289) reached the 5-year using only short cell-free DNA fragments (p , 0.0001). The proportion of diagnostic
follow-up, with a median follow-up duration of 59 months. To assess the prognostic value samples with detectable ctDNA was nominally higher when analyzed by short fragments
of preoperative ctDNA levels, relapse-free survival (RFS) and overall survival (OS) were (39% vs. 64% with 90-150bp size restriction, p = 0.1078, n = 28) and was significantly
evaluated across different stages and pathological subtypes separately. In stage I LUAD higher in samples collected after two cycles of chemotherapy (5% vs. 40% with 90-150bp
patients (N = 179), ctDNA-positive patients (N = 20) had significantly inferior RFS size restriction, p = 0.0197, n = 20). Increases in ctDNA detectability with fragment size
compared to ctDNA-negative patients (2-year RFS: 70% [95% CI: 46%–88%] vs. 94% [95% restriction were also observed in each of localized and metastatic subgroups (metastatic:
CI: 90%–97%]; log-rank p , 0.001). In contrast, no association was found between 16% without restriction, 38% with restriction, n = 97, p = 0.0012; localized: 17% vs. 45%, n =
preoperative ctDNA detection and RFS in stage II-IIIA LUAD or non-LUAD NSCLC, irre- 39, p = 0.04). Conclusions: Our results demonstrate that detection of ctDNA is improved
spective of the clinical stage. Specifically, among the 179 stage I LUAD patients, 11 by analyzing short fragments of cell-free DNA in samples collected from patients with LMS.
relapsed within 2 years, and 6 of these had positive ctDNA test results. This rate was These findings represent a potential to increase the sensitivity of an affordable, low-
significantly higher than in patients who relapsed between 2 and 5 years (1/12) or never coverage liquid biopsy assay and may enhance the prognostic value of ctDNA detection in
relapsed (13/156; x2 test, p , 0.001). Conclusions: These findings indicate that pre- these patients. Further study of the association between ctDNA detection and outcome is
surgical ctDNA can serve as a prognostic indicator in early-stage NSCLC. Tumor-naive needed to fully validate the impact of fragment size restricted analysis of cell-free DNA
ctDNA testing may enhance the standard workflow by identifying high-risk patients who samples in patients with LMS and is currently ongoing in a prospective study. Research
could benefit from innovative treatments. Clinical trial information: NCT04820868. Sponsor: National Cancer Institute/U.S. National Institutes of Health.
Research Sponsor: None.

3047 Poster Session 3048 Poster Session

Accurate differentiation of malignant and benign gastric lesions using cell- Development of a methylation-based, tissue-free test for the detection of
free DNA biomarkers. First Author: Hengzhen Li, Harbin Medical University Cancer molecular residual disease by circulating tumor DNA. First Author:
Hospital, Harbin, China John Paul Y.C. Shen, Department of Gastrointestinal Medical Oncology, The University of
Background: Early detection of gastric cancer is challenging due to the invasive nature of Texas MD Anderson Cancer Center, Houston, TX
current diagnostic methods and the difficulty in distinguishing cancer from benign gastric Background: Clinical validation studies support tumor-informed molecular residual
conditions. Cell-free DNA (cfDNA) features have emerged as promising biomarkers for non- disease (MRD) as a prognostic biomarker for disease recurrence across multiple solid
invasive detection. This study aims to develop and evaluate a machine learning model tumor types. However, these tests are not always feasible due to the occasional lack of
utilizing cfDNA features for early gastric cancer detection. Methods: We developed an tumor tissue for next-generation sequencing. Here, we discuss the design of a test for
ensemble machine learning model incorporating four cfDNA features: repeat elements, tissue-free (tf)MRD detection and its application to a cohort of patients with colorectal
fragment-based methylation, focal copy number variation, and fragment size pattern. The cancer (CRC). Methods: A targeted panel composed of differentially methylated regions
model was trained using cfDNA data from 150 gastric cancer patients and 153 individuals was developed. A machine-learning model was trained on differential methylation
with stomach-related conditions. The ensemble model was validated using a cohort of 149 patterns in order to classify plasma samples as MRD-positive or MRD-negative. Per-
cancer patients, 149 individuals with high-risk benign lesions, and 50 low-risk benign formance of the trained classifier was assessed in an independent cohort of 246 patients
lesions. Risk is stratified according to the Correa’s Cascade. Results: The ensemble enrolled in the Bespoke CRC trial (NCT04264702). These patients had MRD results
machine learning model developed using four cfDNA features achieved an AUROC of 0.913 available using a tumor-informed circulating tumor DNA (ctDNA) assay (SignateraTM),
in the training cohort and 0.912 in the testing cohort for distinguishing gastric cancer of whom 163 were persistently MRD-negative without clinical progression, and 83 had
patients from individuals with stomach-related complications, which outperformed in- MRD-positive results. Tissue-free MRD results were compared to the tumor-informed
dividual cfDNA features. A decision threshold of 0.418, established via cross-validation, results by calculating the percent positive agreement (PPA) and negative percent
was set to ensure at least 95% sensitivity in the training cohort. This threshold enabled agreement (NPA). Clinical outcomes (recurrence-free survival [RFS]) were evaluated
accurate binary classification in the validation cohort, with model scores correlating with based on tfMRD results in all patients and stratified based on whether the patient
cancer stage and tumor differentiation, supporting its potential for clinical risk stratifi- received adjuvant chemotherapy (ACT). Results: In this clinical cohort from Bespoke
cation. Model scores effectively differentiated cancer from high-risk individuals, with
CRC (72% non-Hispanic White, 54% male, mean age 61.4612.3 years), 71 (28%) patients
significantly lower scores in non-cancer groups compared to precancerous or Stage I–III
had stage II CRC, and 146 (59%) had stage III CRC. Overall, PPA was 86% (95% CI: 77-
cancer cases (Non-cancer group median score = 0.35, precancer group median score =
93%) and NPA was 98% (95% CI: 95-100%). Patients with tfMRD-positive status showed
0.48, cancer group median score = 0.61). Additionally, the model assigned significantly
inferior RFS compared to tfMRD-negative patients (p , 0.001). Significant benefit from
higher cancer prediction scores to gastric cancer cases compared to high-grade intra-
epithelial neoplasia (p = 0.003). In the validation dataset, sensitivities were 92.9% (95% CI:
ACT was observed among tfMRD-positive (p , 0.001) but not among tfMRD-negative
85.3%–96.7%) for Stage I, 96.3% (95% CI: 81.7%–99.3%) for Stage II, and 100% (95% CI: patients (p = 0.19). For patients who did not receive ACT in this cohort, we observed
83.2%–100%) for Stage III. Sensitivities for well-, moderate, and poorly differentiated 100% PPV and 100% specificity. Conclusions: This is the first study of its kind
tumors were 91.7%, 92.2%, and 100%, respectively. Of note, specificity for the detection of demonstrating a high concordance between a tfMRD test and a clinically validated
cancer was 66% in the training cohort and 71% in the validation cohort. Conclusions: Our tumor-informed ctDNA assay. Similar to recently reported data using a tumor-informed
findings demonstrate the potential of cfDNA-based machine learning models as a non- ctDNA assay, patients with tfMRD-positive results appeared to derive benefit from ACT
invasive and accurate diagnostic tool for early gastric cancer detection. By reducing treatment. These findings demonstrate that in cases where tissue is not available or of
reliance on invasive procedures, this approach could enhance clinical workflow efficiency inadequate quality, a methylation-based tissue-free assay may serve as a potential
and improve patient outcomes. Further validation in larger, independent cohorts is needed alternative for MRD detection. Research Sponsor: None.
to support clinical implementation. Research Sponsor: Heilongjiang Provincial Key R&D
Program Projects; “Open competition mechanism” of Heilongjiang Province; “Climbing
program” of Harbin Medical University Cancer Hospital.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
200s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3049 Poster Session 3050 Poster Session

Prevalence of androgen receptor ligand-binding domain mutations (AR- A high-performance blood-based DNA methylation test for early detection of
LBDm) in circulating tumor DNA (ctDNA) versus tissue biopsies in partic- gastrointestinal cancers. First Author: Xiaosheng He, The Sixth Affiliated Hospital,
ipants (pts) with metastatic castration-resistant prostate cancer (mCRPC) Sun Yat-sen University, Guangzhou, China
and other tumor types. First Author: Emmanuel S. Antonarakis, University of Background: Early detection of gastrointestinal cancers (GICs), including esophageal
Minnesota, Masonic Cancer Center, Minneapolis, MN cancer (EC), gastric cancer (GC), and colorectal cancer (CRC), remains suboptimal in
Background: AR-LBDm is a common mechanism of resistance to AR-directed therapies in patients with China due to low screening adherence and limited access to endoscopic procedures.
mCRPC. However, data on the utility of ctDNA-based versus tissue-based AR-LBDm detection in mCRPC Multi-cancer early detection (MCED) tests present a convenient alternative, yet the
and the prevalence of AR-LBDm in nonprostate cancers are lacking. We evaluated AR-LBDm in ctDNA and
tumor tissue samples from pts with advanced solid tumors enrolled in various Merck & Co., Inc. trials.
efficacy in detecting early-stage GICs has been inadequate. Accurate tumor localization,
Methods: Samples came from pts with mCRPC and 25 other tumor types (eg, bladder, colorectal, ovarian, particularly differentiating between upper and lower GICs, is crucial for determining
and pancreatic cancers). AR-LBDm data were available from ctDNA via 2 fixed panel-based ctDNA NGS subsequent diagnostic procedures. In this study, we report the performance of an MCED
assays or from tumor tissue samples (mainly archival) via a fixed-panel NGS assay. Data obtained from the assay utilizing targeted DNA methylation sequencing to detect GICs. Methods: This
3 assays were nonoverlapping. AR-LBDm prevalence was evaluated in each tumor type and by BRCAm or multicenter, case-control study prospectively enrolled a cohort of 667 GIC patients (203
HRRm status in pts with mCRPC. Results: The analysis included 3026 samples assessed by ctDNA assay 1
EC, 263 GC, 201 CRC; stages: I 20.8%, II 26.7%, III 35.7%, IV 16.8%) and 667 non-cancer
(mCRPC, n = 1785; other, n = 1241), 2232 samples assessed by ctDNA assay 2 (mCRPC, n = 378; other, n =
1854), and 8181 samples from tissue biopsies (mCRPC, n = 833; other, n = 7348). AR-LBDm was detected in participants. Plasma cell-free DNA was sequenced using a panel targeting tumor-
21.0% of pts with mCRPC by ctDNA assay 1 and 19.8% of pts with mCRPC by ctDNA assay 2. Across all specific hyper- and hypo-methylation markers. A total of 5120 GIC-specific methylation
other tumor types, only 1 pt (with hepatocellular carcinoma) had an AR-LBDm in ctDNA. The prevalence of features were captured. The GIC model was trained using a gradient-boosted tree model,
selected AR-LBDm in tissue biopsies was 5.4% (45/833) in pts with mCRPC, 0.6% (1/159) in pts with salivary and nested cross-validation was implemented to determine the optimal parameters and
cancer, and 0% in all other tumor types. AR-LBDm prevalence in pts with mCRPC increased with later-line evaluate the model’s performance of cancer detection. To predict the tissue of origin
treatments (Tx) both by tissue and ctDNA analyses and was similar regardless of BRCAm or HRRm status
(table). The prevalence of the AR-LBD T878A mutation (by ctDNA) in pts with mCRPC was higher after prior
(TOO), the top 256 features were first selected based on pairwise mutual information for
Tx with abiraterone than with enzalutamide (20.9% [43/206] vs 1.7% [3/173]). AR-LBDm was observed in each cancer type. An XGBoost classifier combined with Synthetic Minority Over-
tissues biopsied from the mCRPC setting (prevalence, 12.8% [12/94]) and was associated with higher AR Sampling Technique (SMOTE) was trained to determine the TOO. Results: The GIC
transcriptional activity than in AR-LBD–negative tissues. Conclusions: Similar prevalence of AR-LBDm model exhibited robust performance, achieving an area under the curve (AUC) of 0.959
was observed in ctDNA of pts with mCRPC by 2 different ctDNA assays; AR-LBDm prevalence in archival (95% CI: 0.949-0.970), with an overall sensitivity of 86.4% (83.5%-88.8%) at a specificity
tissue samples was lower than by ctDNA analysis, likely due to the Tx-emergent nature of AR-LBDm. AR-
of 96.0% (94.2%-97.2%). Notably, for stage I-III GICs, which accounted for 83.2% of
LBDm prevalence was similar regardless of BRCAm or HRRm status in mCRPC, although prevalence of
certain AR-LBDm is impacted by the specific prior AR-directed Tx. These data support ctDNA-based (rather cases (a proportion consistent with that seen in prospective observational cohort
than tissue-based) AR-LBDm testing in pts with mCRPC. Research Sponsor: Merck Sharp & Dohme LLC, a studies of MCED, such as the SYMPLIFY study), the sensitivity reached 84.1% (80.9%-
subsidiary of Merck & Co., Inc., Rahway, NJ, USA. 87.0%). The sensitivities for EC, GC, and CRC were 87.2%, 82.9%, and 90.0% respectively.
For stage I CRC, the sensitivity of the GIC model reached 79.1% (64.8%-88.6%),
AR-LBDm prevalence, % (n/N) Frontline Tx Later-line Tx
comparable to that of multitarget stool DNA tests, and outperformed fecal immuno-
ctDNA assay 1 16.6 (122/734) 23.8 (252/1059)
ctDNA assay 2 Not available 19.8 (75/378)
chemical tests (FIT). Regarding tumor localization, the accuracy of TOO across all
Tissue biopsy 0.6 (3/535) 4.1 (147/3625) positive cases was 89.8% (87.0%-92.0%). For stage I-III GICs, the model maintained a
AR-LBDm +ve* AR-LBDm -ve* high accuracy of 88.7% (85.5%-91.2%) in predicting TOO. Moreover, the model dem-
BRCA
Mut 26.5 (9/34) 73.5 (25/34) onstrated exceptional accuracy in distinguishing between upper and lower GICs, with an
WT 27.5 (95/345) 72.5 (250/345) accuracy of 95.5% (93.5%-96.9%). Conclusions: This study demonstrated the high
HRR
Mut 28.8 (30/104) 71.2 (74/104) performance of the MCED test for early detection of GICs in a large-scale, prospective
WT 26.9 (74/275) 73.1 (201/275) cohort enriched with early-stage cancers. These findings highlight the potential of the
*Per ctDNA assay 1. GIC model to enhance early detection and precise localization of GICs, thus improving
the efficiency of subsequent diagnostic procedures. Research Sponsor: None.

3051 Poster Session 3052 Poster Session

A novel magnetic bead-based cfDNA extraction method for advanced ctDNA Liquid biopsy-informed precision oncology clinical trial to evaluate the
marker discovery and methylation profiling. First Author: Zhuoran Jiang, utility of ctDNA genomic profiling in patients with advanced or metastatic
Shanghai Xiaohe Medical Laboratory Co., Ltd., Shanghai, China solid tumors. First Author: Amna Jamali, Sidney Kimmel Comprehensive Cancer
Background: Cell-free DNA (cfDNA) extraction and circulating tumor DNA (ctDNA) enrichment Center, Johns Hopkins School of Medicine, Baltimore, MD
are critical for liquid biopsy-based cancer diagnostics. However, the QIAamp Circulating Nucleic Background: Genomic profiling through liquid biopsies (LB) has enabled precision oncology
Acid Kit (QIA), despite its widespread use, has limitations, including a labor-intensive manual decision making, however a key challenge lies in critically interpreting LB data to optimize
workflow and suboptimal performance in ctDNA marker enrichment and contaminant removal, patient care. Methods: We report results from the first planned interim analysis of an
potentially affecting downstream methylation analyses. To overcome these limitations, we observational biomarker trial, designed to evaluate the clinical utility of serial LB in patients
developed a novel, automatable magnetic bead-based cfDNA extraction method to improve with advanced/metastatic solid tumors (NCT05585684). Primary endpoints were to de-
ctDNA enrichment and biomarker discovery. Methods: The optimized extraction protocol in- termine feasibility, prevalence of actionable alterations in LB and the fraction of patients with
corporates refinements in pre-treatment, lysis, and binding steps to enhance cfDNA purity and enacted genotype-matched therapies. Secondary endpoints included progression-free (PFS)
ctDNA enrichment. Plasma samples from 8 lung adenocarcinoma (LUAD) patients, 10 healthy
and overall survival (OS), time to subsequent therapy and concordance between LB and
donors, and five simulated plasma samples spiked with varying proportions of fragmented
tumor next generation sequencing (NGS). Exploratory endpoints included correlation of
genomic DNA from H838 (cancer) and NA12878 (healthy) cells were processed using both our
ctDNA dynamics with survival. Serial LBs were obtained at baseline, 1-3 weeks on therapy
optimized assay and the QIA kit. Results: The optimized method achieved cfDNA yields
comparable to the QIA kit in LUAD and healthy samples but exhibited significantly higher ex- and at progression using a CAP/CLIA validated NGS panel (Labcorp, MD). Patient-matched
traction yields in simulated samples (39.91 ng vs. 31.17 ng, p , 0.05). Digital PCR demonstrated white blood cell (WBC) NGS was utilized to identify clonal hematopoiesis (CH)-derived
superior enrichment of 136 bp and 400 bp cfDNA fragments, while library preparation showed a 1- variants. Actionability of genomic alterations was assessed by an ensemble multi-resource
fold and 1.46-fold increase in pre-library yield and a 16.55% and 6.45% increase in mapped ratios programmatic approach; results were reviewed at the Johns Hopkins Molecular Tumor Board
for LUAD and healthy donors, respectively. These results demonstrate that our method achieves (JH MTB). Results: Between March 2023 and July 2024, 51 patients with NSCLC, SCLC and
superior cfDNA enrichment efficiency compared to QIA, making it better suited for NGS-based esophageal cancer were enrolled, with 45 evaluable baseline and 12 progression LBs
workflows. Library complexity rose from 24.07% to 31.24% in LUAD samples and from 29.71% to reviewed at JH MTB. Median turnaround time from baseline and progression LB to MTB
35.04% in healthy donors, with coverage depth of target regions improving by 56.13% and 22.88%, recommendation was 14 and 13 days respectively. Patient-matched analyses of baseline
respectively. This enabled the detection of more CpG sites at equivalent sequencing depths. WBC samples revealed 30.1% (n = 22) CH-derived alterations. The frequency of actionable
Simulated sample analysis confirmed that our optimized method better preserved methylation variants was 28.8% (n = 21) at baseline, 23.3% (n = 7) on therapy and 21.7% (n = 5) at
accuracy, achieving higher consistency between extracted and unextracted DNA. Furthermore, progression. Of the 45 patients reviewed at baseline, 33 received a recommendation for
the optimized method identified significantly more cancer-specific haplotypes across 1,517 LUAD genotype-matched therapies; 48.5% (n = 16) based on tumor molecular profiling, 15.2% (n =
markers, improving ctDNA detection sensitivity, particularly in low tumor burden samples. 5) based on LB alone and 36.3% (n = 12) based on LB and tissue NGS. Thirteen patients were
Conclusions: Our automatable magnetic bead-based cfDNA extraction method outperforms QIA treated according to MTB recommendations. Patients who were treated with genotype-
in library quality, complexity, and methylation accuracy while enabling enhanced ctDNA en- matched MTB recommended therapies had longer OS and PFS compared to those who
richment and biomarker detection. This approach provides a robust and scalable solution for received alternate therapies (not reached-NR vs. 14.8 months, log-rank p = 0.028 and NR vs
advancing liquid biopsy-based cancer diagnostics. Research Sponsor: None. 6.2 months, log-rank p = 0.21 respectively). Among the 12 patients reviewed at progression, 5
Handling Beta bias with Unmethylated Methylated received an MTB recommendation for genotype-tailored therapies based on LB alone (n = 3)
time per run Cost unextracted marker counts marker counts or in combination with tissue NGS (n = 2). Early on-therapy ctDNA clearance was associated
Technique Protocol Throughput (min) ($) DNA median median with longer PFS and OS (log rank p = 0.02 and p = 0.06). Conclusions: Our findings highlight
QIAamp Vacuum- Manual 24 180~240 25 0.01319 382 526 the value of a multidisciplinary MTB when supported by comprehensive liquid biopsy
(QIA) column
Our Magnetic Automatic 24 30 1 0.00365 590 734
molecular information to inform therapy selection and improve patient outcomes. Clinical
assay bead trial information: NCT05585684. Research Sponsor: LabCorp; National Cancer Institute;
1U01CA274631-01A1; Oncology Center of Excellence, Food and Drug Administration (FDA);
U01FD0005042.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 201s

3053 Poster Session 3054 Poster Session

Real-time clinical validation of a blood cell-free, mRNA-based GeneVerify Molecular profiling of body fluid cfDNA: Advancing diagnostics and thera-
test for screening and early diagnosis of prostate cancer. First Author: Sudhir K. peutic decisions. First Author: Aditya V. Shreenivas, City of Hope National Medical
Rawal, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India Center, Duarte, CA
Background: Prostate cancer screening methods, including prostate-specific antigen Background: Effusions in cancer patients pose several critical challenges for clinicians. In known
(PSA) testing, radiological imaging, and pathological assessments, often fail to achieve cancer patients, an effusion may signal recurrence, whereas in newly diagnosed, seemingly localized
reliable early detection. Despite advancements, gene-based diagnostic tests tailored for cases, it indicates a more advanced stage. In many patients the effusion may be secondary to
complications of treatment or comorbidities, rather than malignant. Diagnosing malignant involve-
early detection of prostate cancer remain underdeveloped. To bridge this gap, we ment of body fluids remains a challenge due to the limitations of conventional cytology. This study
evaluated the clinical utility and diagnostic accuracy of the plasma cell-free mRNA- explores the potential of molecular profiling of body fluids to identify actionable molecular alterations
based GeneVerify test. This cutting-edge diagnostic approach aims to deliver faster and and its role in diagnosing malignant effusions. Methods: We analyzed cfDNA from body fluids—ascitic
more precise results while avoiding the risks and complications associated with surgical fluid (N=26), cerebrospinal fluid (N=7), pleural fluid (N=11), and pericardial fluid (N=1), collected from
biopsies. Methods: Blood samples were prospectively collected from patients sus- 45 patients with solid tumors, including lung (N=12), breast (N=9), ovarian (N=9), pancreas (N=4),
pected of prostate cancer who presented with urinary symptoms, persistently elevated gastrointestinal cancers (N=4), cervix (N=2), and one each of CNS, endometrial cancer, HCC, lip-
PSA levels (.4 ng/mL), and PIRADS 4-5 lesions. All patients were recommended for a osarcoma, and melanoma. In a subset, results from fluid samples were compared with tissue and
plasma samples to assess concordance across different sample types. Results: Pathogenic alter-
transperineal biopsy. Plasma cell-free RNA was isolated and analyzed for 25 prostate
ations were identified in 89% (40/45) of fluid samples. The most frequently mutated genes were TP53
cancer-specific genes using the GeneVerify real-time PCR kit (Hayward, USA). The log2 (53%), EGFR (20%), KRAS (18%), PIK3CA (9%), CTNNB1 (7%), FGFR3 (7%), GNAS (7%), MYC (7%), and
fold change in gene expression for each gene was calculated, and a genetic risk score ESR1 (4%). Simultaneous analysis of body fluid and tissue samples (n=11) revealed that 7 patients
(GeneVerify Dx) was assigned to each patient. The transperineal biopsy results served as (64%) had at least one concordant pathogenic alteration. Similarly, analysis of body fluid and plasma
the reference standard. The diagnostic performance of the test in distinguishing be- samples (n=16) showed that 8 patients (50%) had at least one concordant pathogenic alteration. Body
tween benign and cancerous cases was assessed and ROC (receiver operating char- fluid analysis identified acquired resistance alterations, such as EGFR T790M and ALK C1156Y, which
acteristic) curve was plotted. Results: We tested a total of 45 subjects, comprising 35 influenced therapy decisions. Among the alterations detected exclusively in fluid samples were ERBB2
suspected prostate cancer cases and 10 healthy controls. The median age of the cases amplification and ESR1 D538G mutation in two breast cancer patients. In evaluating molecular
profiling against cytology for detecting malignant effusions, 17 of 25 samples were positive by both
was 66 years (range: 45–85), while the controls had a median age of 41.5 years (range: methods, while 4 of 5 cytology-negative samples were ctDNA-positive. Notably, 3 of 4 ctDNA-negative
39–47). The genetic risk score effectively differentiated prostate cancer patients from cases were cytology-positive. These results emphasize the potential role of molecular profiling for
benign cases. The Area Under the Curve (AUC) was 0.83 6 0.07 (P = 0.002; 95% CI: diagnosis when cytology is inconclusive. Conclusions: This study highlights the importance of body
0.70–0.96), demonstrating the test’s strong discriminatory ability. Using a risk score cut- fluid ctDNA profiling (ascites, pleural, pericardial, CSF) in identifying actionable mutations, including
off value of 10, the test achieved a sensitivity of 72% and a specificity of 90%. Ad- unique druggable alterations not found tissue or liquid biopsies. The ability to detect ctDNA in
ditionally, patients with higher Gleason grades and those experiencing chronic in- cytology-negative samples underscores the potential of body fluid ctDNA as a valuable complement to
flammation exhibited elevated gene expression levels and higher risk scores compared fluid cytology for diagnosing malignant involvement. Research Sponsor: None.
to benign subjects. Conclusions: In summary, this prospective study is the first to ESCAT classification of pathogenic variants identified in body fluids from 45 patients.
validate a blood cell-free RNA-based GeneVerify test for real-time screening and early Tier Level Incidence (%) Number of unique patients
detection of prostate cancer. The test exhibits high precision in identifying early-stage IA 28.9% 13
prostate cancer, with strong concordance to biopsy results. Research Sponsor: None. IIA 0% 0
IIIA 35.6% 16
IIIB 4.4% 2
IVA 55.6% 25
IVB 2.2% 1
X 22.2% 10

3055 Poster Session 3056 Poster Session

Circulating tumor DNA and late recurrence in high-risk, hormone receptor- A plasma proteomics-based model for predicting response to neoadjuvant
positive, HER2-negative breast cancer: An updated analysis of the CHiRP chemotherapy in ovarian cancer. First Author: Coren Lahav, Oncohost Ltd,
study. First Author: Tae-Kyung Robyn Yoo, Dana-Farber Cancer Institute, Boston, MA Binyamina, Israel
Background: Risk of recurrence for patients (pts) with HR+/HER2- breast cancer persists Background: Neoadjuvant chemotherapy (NACT) is a standard treatment option for ad-
for decades. Most distant recurrences occur in the ‘late’ adjuvant setting, . 5 years (yrs) vanced high-grade serious ovarian cancer (HGSOC). Following interval debulking surgery,
from diagnosis. In CHiRP (ASCO 2022), we showed that minimal residual disease (MRD) pathologists assess tumor response to initial chemotherapy using a standardized che-
was detectable in the late adjuvant setting: ctDNA was detected in 8/83 (9.6%) pts in the motherapy response score (CRS) corresponding with patient survival. We sought to de-
cohort and 6/8 (75%) pts with positive ctDNA (+ctDNA) had developed distant recurrence termine the feasibility of developing a proteomic-based biomarker to predict response to
when initially reported (median follow-up 2 years from first plasma sample collected on NACT based on pre-treatment plasma samples. Methods: Pre-treatment samples were
study). Here, we report updated clinical outcomes and investigate the meaning of a ctDNA collected from 71 HGSOC patients receiving platinum-taxane combination NACT. Deep
test result during surveillance with longer follow-up. Methods: In CHiRP, pts with stage II- plasma proteomic profiling was performed using SomaLogic’s 7K aptamer-based tech-
III HR+/HER2- breast cancer at high risk of recurrence diagnosed . 5 yrs prior with no nology. Based on the proteomic profiles, a computational model was developed to predict
evidence of recurrence were prospectively identified. All pts provided informed consent CRS, focusing on differentiating between poor CRS (CRS1) versus partial or near-complete
for prospective plasma collection every 6-12 months at routine follow-up visits for CRS (CRS2/3). The model counted the number of response-associated proteins per patient.
batched, retrospective ctDNA testing using RaDaR, a tumor-informed whole exome Patient scores were derived by resampling into training and test sets, averaging test set
sequencing-based assay. Pts were followed at the discretion of the clinical provider results. Patients were stratified into groups (i.e., ’Chemo-responsive’ or ’Chemo-resistant’)
without any routine surveillance imaging, as per guideline-concordant care. See CHiRP based on median score. Bioinformatic analysis of the HGSOC-specific proteomic biomarkers
ASCO 2022 presentation for additional methods. Results: Of 83 pts in the analytic cohort, was performed to gain insight into the potential mechanisms driving NACT therapeutic
57 (68.7%) pts had stage III disease, and most (n = 75, 90.4%) underwent (neo)adjuvant benefit and resistance. Results: Our proteomics-based predictive model differentiated
chemotherapy. All pts received endocrine therapy. In this update, median follow-up from between patients with CRS1 versus CRS2/3 (ROC AUC = 0.67, p = 0.008). CRS association
with disease-free survival (DFS) and overall survival (OS) in this cohort was consistent with a
first sample collection was 4.4 yrs (interquartile range 4.0, 4.9). 214 plasma samples were
previous meta-analysis, though not reaching statistical significance (CRS3 versus CRS1/2,
collected prior to any known recurrences and included in this analysis. 8/83 (9.6%) pts had
HR = 0.62, 95% CI: 0.27-1.42, p = 0.25 for DFS; HR = 0.67, 95% CI: 0.21-2.09, p = 0.48 for OS).
+ctDNA at any timepoint including 4/83 (4.8%) with +ctDNA on first study plasma sample.
Hazard ratios for patient classification as ‘Chemo-responsive’ versus ‘Chemo-resistant’
In pts initially ctDNA-negative (-ctDNA; n = 4), median time from first sample collection to
trended in the same direction (HR = 0.70, 95% CI: 0.37-1.30, p = 0.25 for DFS; HR = 0.65, 95%
MRD detection was 1.29 yrs (range, 0.72 – 3.05). During follow-up, 8 (9.6%) pts developed
CI: 0.25-1.72, p =0.39 for OS). The predictive model incorporated 62 proteins. Of these, 27
distant recurrence and 1 (1.2%) pt had a local recurrence. With additional follow-up were elevated in the plasma of patients with poor CRS compared to those with partial or
included in this update, all 8/8(100%) pts with +ctDNA developed distant recurrence with a near-complete CRS. These proteins were notably enriched in pathways related to cell death
median lead time of 1.39 years (range 0.01 – 4.24). Among -ctDNA plasma samples resistance. Moreover, several of these elevated proteins have previously been linked to
with . 2 yrs of follow-up (n = 185), the negative predictive value (NPV) of a -ctDNA test for HGSOC, particularly in the context of chemotherapy resistance. Conversely, patients with
lack of clinical recurrence for . 2 yrs post-test was 98.4% (3/185). The NPV indicating partial or near-complete CRS showed significant enrichment of proteins associated with
freedom from recurrence . 1 and . 3 yrs was 100% (0/196) and 96.6% (5/147), re- genome instability and mutation. Conclusions: This study demonstrates the feasibility of
spectively. Conclusions: In pts with high-risk HR+/HER2- breast cancer in the late CRS prediction using plasma proteomics at baseline, potentially complementing existing
adjuvant setting, all pts with +ctDNA developed distant metastasis. A -ctDNA test was imaging approaches. While current treatment options limit immediate clinical utility, these
strongly associated with lack of recurrence over a 3 yr follow-up period. Future studies are findings provide novel insights into biological determinants of chemotherapy response and
needed to determine if ctDNA-guided intervention can impact clinical outcomes for early- may become predictive biomarkers for novel treatment protocols. Research Sponsor:
stage breast cancer and to determine the optimal role of MRD surveillance during follow- OncoHost; Israel Science Foundation (ISF); 2972/21; Israel Cancer Research Fund (ICRF);
up. Research Sponsor: None. 21-302-MI.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
202s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3057 Poster Session 3058 Poster Session

NeoCircle: Investigating circulating tumor DNA dynamics as a predictor of Monitoring populations of tumor-macrophage fusion cells in blood prog-
survival in primary breast cancer. First Author: Anthony M George, Lund University, nosticates PFS and OS in pan-metastatic cancers over 2 years. First Author:
Lund, Sweden Sonia Muthuraj, Rutgers University, New Brunswick, NJ
Background: Persistent circulating tumor DNA (ctDNA) detection during neoadjuvant Background: Tumor associated macrophages are known to fuse with cancer cells in the
treatment (NAT) of early-breast cancer (EBC) indicates high-risk disease. Following blood through a dysfunctional CD47 phagocytic immune pathway resulting in the
surgical resection, ctDNA-positivity indicates molecular residual disease (MRD) and formation of tumor- macrophage fusion cells (TMFCs) which are observed as het-
heralds occult metastatic disease relapse. To incorporate ctDNA into EBC management, erokaryon (incomplete fusion), synkaryon (full fusion) or hetero-to-synkaryon transition
scalable and widely accessible diagnostic methods are necessary. Here we apply an (partial fusion). Previous studies in lung and breast cancer demonstrated that subtypes
ultrasensitive, personalized tumor-informed approach to ctDNA analysis leveraging of TMFCs in blood may correlate with highly aggressive disease unlikely to respond to
structural variant (SV) detection using a novel multiplex digital PCR (dPCR) technology. certain systemic therapies (i.e. chemotherapy). We initiated a prospective study to
Methods: 116 patients with stage I-III EBC (31.0% TNBC, 43.1% HR+/HER2- and 24.1% evaluate the blood of n = 100 metastatic pan-cancer patients (pts) receiving systemic
HER2+) and eligible for NAT were recruited through the prospective SCAN-B study therapy for the presence of full, partial, and incomplete TMFCs to compare their
(NCT02306096, substudy NeoCircle) between December 2014 and March 2019 and have progression-free survival (PFS) and overall survival (OS). Methods: We conducted a
been analyzed for ctDNA. Whole genome sequencing was performed on tumor material prospective pilot study of n = 100 pathologically confirmed metastatic cancer pts with
and personalized multiplex dPCR assays tracking up to 16 SVs were used for ctDNA breast (n = 23), prostate (n = 21), pancreas (n = 17), colon (n = 19), or lung (20) with active
monitoring. Plasma samples were collected at baseline, during NAT, pre- and post- progressive disease, prior to the induction of new systemic therapies, i.e. chemotherapy
surgery and at 6-monthly intervals during follow up. Results: High baseline detection (n = 39), PD-L1 immunotherapy (n = 27), hormone therapy (n = 20), or targeted therapy (n
was observed across all stages and subtypes (90.5% overall), and ctDNA-positivity at = 23). TMFCs were isolated from 7.5ml peripheral blood using CellSieve microfiltration
end-of-NAT (end-NAT) was a significant predictor of eventual disease relapse and death and identified by their enlarged multinucleated structure (. 30 mm), which was cat-
(relapse-free interval, RFI, hazard ratio, HR, 3.7, 95% CI 1.4-9.7; overall survival HR 7.7, egorized into 3 distinct subtypes: full fusion marked by a single multinucleated nuclei,
95% CI 2.2-26.6). A significant association was observed between end-NAT ctDNA partial fusion marked by 2 contacting nuclei, incomplete fusion marked by 2 distinct
clearance and pathological complete response (pCR), whereas non-pCR by itself was non-contacting nuclei. TMFC subtypes were compared to pts’ PFS and OS by cox
not a significant predictor of relapse or death in this cohort. At one or more post- proportional univariate and multivariate analysis over 24 months. Results: We identified
operative timepoints, MRD+ was detected in 10 patients who experienced distant re- TMFCs in 78% of all pts (n = 78/100), averaging 10 per pt. 37% of pts were found to have
currence, with lead times up to 4 years (median 13.9 months, range 1.8-47.7 months). more than one TMFC subtype in their sample, with 70 pts having full fusion, 23 partial
Similarly, ctDNA was detected in 3 of 4 patients with local recurrences and 1 of 2 fusion, and 28 incomplete fusion. At 24 months, pts with incomplete fusion TMFCs had
patients with CNS-only recurrences. For 2 patients without presentation of clinical significantly worse PFS (HR, 2.9; 95% CI, 1.5 to 5.7; P = 0.0023) and OS (HR, 2.5; 95% CI,
recurrence to date, ctDNA was detected post-operatively, with subsequent clearance 1.2 to 5.3; P = 0.0202). Interestingly, pts with incomplete fusion and treated with
during follow-up. Post-operative MRD associated with poor RFI (HR 45.5, 95% CI 13.0- systemic targeted therapy (n = 7) were found to have significantly improved PFS (HR,
159.8) and OS (HR 15.3, 95% CI 4.5-52.9). Conclusions: In this analysis of 116 patients 4.8; 95% CI, 1.8 to 13.0; p = 0.0052), but not OS (HR, 3.0; 95% CI, 1.0-9.2; p = 0.0999)
from a prospective study in patients with EBC receiving NAT, we monitored ctDNA using versus other therapy types. There was no significant PFS differences in pts without
an ultrasensitive tumor-informed dPCR assay tracking patient-specific SVs. ctDNA incomplete fusion TMFCs being treated with targeted therapies. Conclusions: In a pan
detection post-NAT and prior to surgery was associated with high-risk of disease relapse metastatic cancer setting, we found that incomplete fusion in circulating TMFCs as-
and death, outperforming pCR. Moreover, post-operative ctDNA detection was also sociates with poorer outcomes at 24 months. Further, it appears that pts with in-
significantly associated with disease relapse and death, with long lead-times over complete TMFCs may have had better outcomes when treated with targeted therapies
standard-of-care clinical assessments. These findings further validate the feasibility of compared to other therapy types. These preliminary findings suggest the need for larger
SVs as an MRD analyte and support the clinical use of this approach in EBC. Research scale prospective studies to further evaluate relationships between TMFCs and ther-
Sponsor: None. apeutic responses in specific disease populations. Research Sponsor: Creatv MicroTech.

3059 Poster Session 3060 Poster Session

Whole-genome bisulfite sequencing of cell-free DNA to investigate molec- Beyond tumor-shed markers: AI driven tumor-educated polymorphonuclear
ular contributors to racial survival differences in advanced-stage triple- granulocytes monitoring for multi-cancer early detection. First Author: Rajan
negative breast cancer. First Author: Chun Wang, Thomas Jefferson University - Datar, Datar Cancer Genetics, Nashik, India
Department of Medical Oncology, Philadelphia, PA Background: Tumor educated Polymorphonuclear Granulocytes (tPMNG) are a distinct
Background: Black women are twice as likely to be diagnosed with triple-negative phenotypic set of Neutrophils (N2) with corrupted programmed death pathways
breast cancer (TNBC), a highly aggressive and difficult-to-treat subtype with poor resulting in apoptosis resistance. The presence of tPMNGs in the peripheral blood
survival outcomes. While well-defined factors such as socioeconomic status have been indicates up-regulation of pro-tumoral factors and resistance to apoptotic signals. Our
recognized, the impact of genomic and molecular factors on the racial disparity in TNBC platform leverages this unique anti-apoptotic characteristic of tPMNGs through a
survival remains understudied. Liquid biopsy, a non-invasive and real-time method for proprietary culture process and AI-based digital imaging analysis to detect cancer in its
studying the molecular landscape of tumors, has yet to be explored in the context of early stages. This approach represents a paradigm shift from detecting rare tumor-shed
racial survival disparities in TNBC. Methods: Ten Black TNBC patients were matched analytes such as ctDNA and CTCs to monitoring relatively abundant tPMNGs, whose
with ten White TNBC patients based on age, family history, tumor stage, grade, and numbers typically exceed conventional analytes by 3-4 orders of magnitude. We studied
inflammatory breast cancer (IBC) status. Baseline blood samples were collected prior to detection of tPMNGs in a case-control study to evaluate their suitability for a multi-
the initiation of a new therapy. Cell-free DNA (cfDNA) was extracted from plasma, cancer early detection test (MCED). Methods: We collected 10 ml of peripheral blood in
bisulfite-converted, and analyzed through whole-genome bisulfite sequencing (WGBS). EDTA tubes from 892 asymptomatic healthy volunteers above 18 years of age [463, 52%
After quality control, ichorCNA was used to identify copy number alterations (CNAs), and male; 429, 48% females with mean age of 48 (20 to 89) years], 24 individuals diagnosed
MethylKit was applied for methylome analysis. Associations between CNAs, differen- with non-malignant conditions including prostatitis, polycystic ovarian disease and
tially methylated regions (DMRs), and progression-free survival (PFS) were evaluated acute pancreatitis, and 90 individuals recently diagnosed with surgically resectable early
and compared between Black and White patients. Results: The Black-White pairs were stage cancers (Stage 1/ 2) comprising Head and Neck (N=32, 36%), Breast (N=20, 22%),
well-matched across key clinical variables, including age (P = 0.99), family history (P = Colorectal (N=14, 16%), others (N=24, 27%). Nucleated cells were isolated from the
1.00), tumor stage (P = 1.00), grade (P = 1.00), and IBC status (P = 0.37). Black TNBC samples after RBC lysis and centrifugation. These cells were seeded in six well culture
patients had poorer PFS compared to their White counterparts. Significant CNAs were plates and subjected to controlled apoptotic stress under serum-free, hypoxic conditions
identified on chromosome 1 (chr1:20400000-23900000_p36.12 and chr1:7200000- with specific growth factor supplementation for 5 days. Surviving cells were set on
9200000_p36.23), regions associated with known breast cancer prognosis genes imaging slides and stained with H&E. 60X images were obtained and analyzed using a
such as EPHB2 and E2F2. Thirteen DMRs were found to be significantly associated with convolutional neural network (CNN) based AI algorithm to detect tPMNGs per ml.
the racial differences in PFS, with key genes such as CDH13 and TMEM132C identified in Results: The mPMNG detection method demonstrated 84% sensitivity (95% CI: 84.44%)
these regions. Notably, the methylation status of the CDH13 promoter has previously across multiple cancer types. The platform demonstrated 97% specificity (95% CI:
been associated with breast cancer risk. The predictive power for PFS was significantly 97.40%) among healthy asymptomatic cohort. In samples from individuals from indi-
enhanced in a model combining these 13 DMRs with race (Concordance Index [C-index] = viduals with non-malignant conditions, specificity was 96% (95% CI: 95.83%).
0.96), compared to a model using race alone (C-index = 0.75). Conclusions: In this pilot Conclusions: This first-in-class immune cell-based MCED approach offers several
study utilizing WGBS of cfDNA, we identified significant CNAs and DMRs associated with advantages over tumor-shed analyte detection: 1. Leverages amplified host response
the racial disparity in survival outcomes among advanced-stage TNBC patients. These rather than rare tumor products. 2. Provides robust detection across cancer types and
findings provide insight into the genomic and molecular contributors to this disparity stages. 3. Utilizes existing laboratory infrastructure and a scalable protocol. 4. Dem-
and highlight the potential of liquid biopsy for future studies. Larger studies are needed onstrates potential for screening, diagnosis, and monitoring applications. The high
to validate these results and further investigate the underlying mechanisms. Research sensitivity and specificity, combined with practical advantages, suggest potential for
Sponsor: National Cancer Institute/U.S. National Institutes of Health; R01CA207468; clinical implementation in cancer screening and monitoring programs either using
Sidney Kimmel Comprehensive Cancer Center. tPMNG alone or in conjunction with CTCs / cfDNA evaluation. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 203s

3061 Poster Session 3062 Poster Session

Next-generation U-Net Encoder: Decoder for accurate, automated CTC de- Clinical outcomes of a prospective multicenter study evaluating a combined
tection from images of peripheral blood nucleated cells stained with EPCAM circulating tumor DNA (ctDNA) and RNA (ctRNA) liquid biopsy assay in
and DAPI. First Author: Tim Crook, Cromwell Hospital, London, United Kingdom metastatic non-small cell lung cancer (NSCLC). First Author: Richa Dawar,
Background: Direct circulating tumor cell (CTC) detection is a promising biomarker for University of Miami Sylvester Comprehensive Cancer Center, Miami, FL
early cancer detection and monitoring. Traditional fluorescence microscopy and AI- Background: Genomic profiling of metastatic NSCLC to inform targeted therapy selection
driven methods have limitations such as subjectivity and labor-intensiveness. We is endorsed by numerous guidelines. While tissue biopsy is the mainstay of molecular
developed a deep-learning pipeline using a U-Net–type encoder–decoder architecture profiling, liquid biopsy offers a practical real-world approach to non-invasively identify
for precise pixel-level CTC discrimination in peripheral blood nucleated cells (PBNCs). guideline-recommended biomarkers. LIQUIK was a prospective, multicenter, observational
This method preserves morphological and fluorescence details, overcoming convolu- cohort study to evaluate the performance of a combined ctDNA and ctRNA liquid biopsy
tional neural network (CNN) limitations by maintaining fine features through skip assay, LiquidHALLMARK (LHM ctDNA and ctRNA) in comparison to the ctDNA-only liquid
connections for better discrimination. We present specificity and sensitivity data from a biopsy Guardant360 (G360 ctDNA) and tissue next-generation sequencing (NGS) for
case-control study. Methods: We collected 5 ml of peripheral blood in EDTA tubes from biomarker detection in metastatic NSCLC. Diagnostic performance of the primary cohort
1383 asymptomatic healthy volunteers (744, 54% male; 639, 46% females with mean age has been previously presented. Here, we report clinical outcomes after 1-year follow-up of
of 49 [(20 to 93) yrs], 38 individuals diagnosed with non-malignant conditions including the cohort. Methods: LIQUIK (NCT04703153) enrolled 151 non-squamous NSCLC patients
prostatitis, PCOD and acute pancreatitis, and 143 individuals recently diagnosed with across the USA and Singapore from Apr 2021 to Dec 2022. Enrolled patients were
surgically resectable early stage cancers (Stage 1/ 2) - Head and Neck (N=50, 35%), genotyped using tissue NGS, LHM ctDNA and ctRNA, and G360 ctDNA for 9 biomarkers
Breast (N=31, 22%), Colorectal (N=17, 12%), Pancreas (N=8, 6%), Prostate (N=8, 6%), (EGFR, ALK, RET, ROS1, BRAF, KRAS, MET, ERBB2, NTRK1/2/3). Patients were treated
Lung (N=5, 3%), Ovary (N=5, 3%) others (N=19, 13%). Nucleated cells were isolated from according to their physician’s choice of first-line therapy following biomarker testing.
the samples after RBC lysis and centrifugation and stained with EPCAM and DAPI and Tumor assessments were performed at baseline and within 6 months (mo) of treatment
set on imaging slides. 60X images were obtained and processed by AI utilizing U- initiation. Overall response rate (ORR), progression-free survival (PFS), and the clinical
Net–Based Encoder–Decoder Architecture and context discrimination to detect CTCs. utility of ctRNA were investigated. Results: Among the 151 patients, 129 were subse-
The customized U-Net pipeline encodes spatial information through successive con- quently treated in the first-line setting (49.6% on targeted therapy, 41.1% on chemo-
therapy, and 30.2% on immunotherapy), with 27 on combination therapy. Of the 64 patients
volutional and pooling layers, generating a highly compressed representation of cells in
on targeted therapy, 47 had matched biomarker findings from tissue NGS, 47 from LHM
the bottleneck. By employing transposed convolutions in the decoder stage—and in-
ctDNA and ctRNA, and 43 from G360 ctDNA. ORRs of patients on targeted therapy and
corporating skip connections from the encoder layers—the AI model reconstructs a
chemo/immunotherapy were 40.4% and 16.1% respectively. Among patients treated with
pixel-wise segmentation mask to identify potential CTCs with cell diameter .10 mi-
targeted therapy, ORR was similar between patients with biomarker-matched findings from
crons. This approach aims to surpass existing methods that rely on bounding-box– tissue NGS (45.2%), LHM ctDNA and ctRNA (40.5%), and G360 ctDNA (36.8%). PFS of
based detection by offering enhanced sensitivity and specificity through end-to-end patients on targeted therapy (median 23.6 mo) was significantly longer than those not on
learned feature extraction. Ground truth annotations were established via expert targeted therapy (median 3.8 mo; HR = 0.26; p , 0.001). Median PFS was similar between
cytopathology review, and training procedures involved cross-validation to ensure patients with biomarker-matched findings from tissue NGS (23.6 mo), LHM ctDNA and
generalizable performance. Results: Analysis of total 1564 samples showed that our U- ctRNA (18.6 mo), and G360 ctRNA (20.1 mo). Overall, incorporation of ctRNA into LHM
Net–based model achieved a sensitivity of 89% (95% CI: 88.81) and specificity of 97% identified 2 additional biomarker-positive patients. Both ctRNA-exclusive biomarkers were
(95% CI: 97.98) for detecting CTCs. Performance remained consistent across solid confirmed by tissue NGS, and both patients were treated with biomarker-matched targeted
tumors, highlighting the flexibility and adaptability of the architecture in various therapy. While one patient was lost to follow-up, the second patient had a partial response
fluorescence staining conditions. Conclusions: Our U-Net pipeline uses pixel-level to treatment. Conclusions: Treatment outcomes based on liquid and tissue biopsies are
segmentation and skip connections to enhance CTC detection accuracy. Integrating comparable. The inclusion of ctRNA in liquid biopsy increases its diagnostic yield of
fluorescence and morphology, it can streamline cancer screening and disease moni- actionable biomarkers. Clinical trial information: NCT04703153. Research Sponsor: None.
toring. Research Sponsor: None.

3063 Poster Session 3064 Poster Session

Microbial metabolic pathways to abrogate immunotherapy toxicity and Tumor-educated platelets as a source of potential biomarkers for colorectal
promote anti-tumor response in metastatic renal cell cancer. First Author: cancer. First Author: Rodnei Macambira, Oncológica do Brasil Cancer Center, Belém,
Shahla Bari, Duke Cancer Institute Center for Prostate and Urologic Cancers, Duke Pará, Brazil
University School of Medicine, Durham, NC Background: Colorectal cancer (CRC) is the third most common cancer and the second
Background: Metastatic RCC has a poor prognosis. Despite improvement in treatment leading cause of cancer-related death worldwide. Current diagnostic methods rely on
outcomes with ICB and targeted therapy, many patients fail to respond to first line therapy invasive procedures and serum markers with limited sensitivity, highlighting the need for
and immune mediated adverse events(irAE) remains a major challenge, often leading to novel, minimally invasive biomarkers. Tumor-educated platelets (TEPs) have emerged
treatment discontinuation. Therefore, mitigating irAE without compromising antitumor as a promising source, as malignant cells reprogram platelets through molecular al-
immunity is a critical unmet need. Tryptophan microbial metabolic pathway is known to terations. This study aimed to identify differentially expressed genes in TEPs from
play a major role in immune homeostasis through its action on Aryl hydrocarbon receptor patients with CRC that could serve as potential diagnostic biomarkers. Methods: We
(AhR) balancing immune suppresser with immune effector responses. We hypothesize that downloaded gene expression data from platelets from the Gene Expression Omnibus
microbial metabolism of tryptophan to indole metabolites may play a role in ICB resistance (GEO; GSE183635), tissue gene expression data from TCGA-COAD and TCGA-READ, and
as well in irAE, identification of which may help us predict patients most likely to respond, vesicle data from Vesiclepedia. The data were preprocessed to remove low-quality reads,
without life threatening toxicity. Methods: We prospectively collected paired stool and and high-quality reads were aligned to the human reference genome GRCh38.p13. We
blood samples of treatment naı̈ve metastatic RCC patients, treated with ICB +/- Tyrosine performed differential expression analysis through DESeq2 (|log2FoldChange| . 1,
kinase inhibitors (TKI) at treatment initiation and at time of first response assessment adjusted p-value , 0.05). Gene ontology (GO) enrichment analysis (p-value , 0.05) was
(12+/-3 weeks). We evaluated stool metagenomics and untargeted stool and plasma conducted, and shared genes between TEPs, tumor tissues, and vesicles were identified.
metabolomics among responders (R) and non-responders (NR). We focused on ROC curves (AUC . 0.75) assessed diagnostic potential. Results: A total of 3,211
kynurenine/tryptophan and indoles/tryptophan ratio to evaluate differential host and differentially expressed genes were identified in TEPs, with 55 upregulated and 3,156
microbial metabolism of tryptophan. A responder was classified as progression free downregulated. Six genes (TLN1, IGF2, IFITM3, DKK1, MYL9, TNNC2) were consistently
survival (PFS) greater than 6 months while patients with grade 3 or higher irAE was upregulated in TEPs, tumor tissues, and observed in microvesicles. These genes
classified as serious IrAE. Results: Among 120 patients accrued, 49 were treated with exhibited AUC values ranging from 0.76 to 0.83, indicating high sensitivity for dis-
combination ICB, while 71 patients were treated with ICB + TKI. Median follow up was
tinguishing CRC patients from healthy controls. Conclusions: Our study identified six
27 months. 28 patients (23%) had a Grade 3 or higher irAE. 3 patients died from com-
DEGs in TEPs with elevated expression in CRC tissue, highlighting their potential as
plications attributable to irAE. The median duration to development of any irAE was
minimally invasive biomarkers for CRC diagnosis. These findings pave the way for
3.5 months. Using negative binomial regression model evaluating baseline relative
developing more precise diagnostic tools to improve early detection and patient out-
abundance of microbial tryptophan metabolites that were associated both with response
as well as serious irAE, we noted significant higher abundance of Indole acetic acid (IAA),
comes. Research Sponsor: Conselho Nacional de Desenvolvimento Cientı́fico e
indole acetonitrile(ACN), indole acetyl phenylalanine (IAAP)and IAA/kynurenine (Kyn) and Tecnológico.
lower abundance of tryptophol, indole 3 pyruvic (IPA), (coefficient of 6.4, 1.8, 7.15, 4.6, Potential biomarker genes in TEPs for CRC, ranked by Log2FC.
0.04, 0.46, with adj p value , 0.05 ) with serious irAE as well as ICB resistance (Coefficient- Gene Log2FoldChange P-adjusted AUC
5.42, 1.83, 6.15, 4.05, 0.03, 0.4, p , 0.05). Conclusions: This is one of the first studies DKK1 2.011554 1,5031023 0.78
evaluating microbial metabolic pathways that may play a role in predicting patients who IGF2 1.5006689 3,5631029 0.81
are more likely to respond with lower likelihood of serious irAE in RCC, thus helping to IFITM3 1.102034 3,0131029 0.81
TLN1 1.098185 1,2431025 0.83
identify strategies to decouple tumor immunity from autoimmunity to improve ICB MYL9 1.060944 1,91310215 0.77
outcomes. Further the results can be extrapolated to many other solid tumor treated with TNNC2 1.023068 7,2231028 0.76
immunotherapy, as tryptophan metabolism plays a immune homeostatic role across
AUC: Area Under the Curve.
cancers. Research Sponsor: Conquer Cancer foundation-ASCO YIA.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
204s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3065 Poster Session 3066 Poster Session

Molecular landscape and therapeutic vulnerability of RRAS- and RRAS2- Larotrectinib resistance in TRK fusion cancers: Analysis of a tumor-
mutant solid tumors. First Author: Alexander James Pfeil, Memorial Sloan Kettering agnostic, global clinical trial dataset. First Author: Alexander E. Drilon,
Cancer Center, New York, NY Memorial Sloan Kettering Cancer Center and Weill Cornell Medical College, New York, NY
Background: The RRAS subfamily of small GTPases shares considerable sequence ho- Background: Larotrectinib (laro) is the first-in-class, highly selective, TRK inhibitor ap-
mology with the canonical RAS oncoproteins KRAS, NRAS, and HRAS. Mutations in RRAS proved for tumor- and age-agnostic use in TRK fusion cancers. This is the seminal report of
and RRAS2 that are homologous to KRAS hotspot mutations promote transformation in primary and secondary laro resistance based on an analysis of the regulatory dataset that
vitro. Most diagnostic sequencing panels do not profile the RRAS subfamily, and therefore, supported drug approval across multiple countries. Methods: Genomic data from patients
the prevalence and clinical relevance of these mutations and their treatment have not been (pts) with non-primary CNS TRK fusion cancer enrolled in a global, prospective, multicenter
fully established. Methods: Based on a clinical targeted DNA sequencing assay (MSK- database of three laro clinical trials including adult and pediatric pts were analyzed. Tumor
IMPACT), an institutional cohort of 51,040 solid tumor cases prospectively sequenced DNA (Illumina TruSight Oncology [TSO] Comprehensive, TSO 500, or FoundationOne CDx)
between 2016-2024 was analyzed to identify RRAS/RRAS2 mutations. Lung cancers were or circulating tumor DNA (Guardant360 or GuardantOMNI) NGS was performed pre-laro
excluded and analyzed in a separate study. Hotspot mutations in RRAS/RRAS2 were (baseline; BL) and post-laro initiation. On-target NTRK (solvent front [SF], gatekeeper [GK],
determined based on homology to KRAS hotspot mutations and/or previous literature xDFG) mutations and COSMIC-classified tier 1/2 off-target alterations were identified.
demonstrating potential oncogenicity (hotspot RRAS: G38, G39 and Q87; hotspot RRAS2: Primary laro resistance analysis set included pts with no meaningful clinical benefit (PD/
G23, G24, A70T, Q72 and in-frame insertions in G23/G24). The sensitivity of cells harboring SD , 4 months). Secondary (acquired) laro resistance analysis set included pts who
RRAS and RRAS2-mutations to the clinically active pan-RAS inhibitor RMC6236 was developed resistance after meaningful clinical benefit (CR/PR/SD $4 months). Data
examined in vitro and in vivo. Western blotting was utilized to determine changes in protein cutoff: July 20, 2024. Results: Of 304 adult and pediatric pts enrolled, 216 had BL genomic
expression and activation. Results: Among the 51,040 cases analyzed, hotspot RRAS and data. Primary laro resistance was observed in 24 pts. Only 1 pt had an on-target mutation
RRAS2 mutations were detected in 6 (0.01%) and 270 (0.5%) patients, respectively. (NTRK3 G623R), likely attributable to prior crizotinib; 9 pts (38%) had off-target alterations
Hotspot RRAS mutations were seen in various cancer types, and most had other mitogenic involving AKT, BRAF, FGFR1, GNAS, KRAS, NRAS, and PIK3CA. Secondary laro resistance
drivers (4/6). Amongst tumors with hotspot RRAS2 mutations, the most common cancer was observed in 55 pts with valid post-BL ctDNA (the most common of these TRK fusion
types included endometrial (n = 172), ovarian (n = 27), and germ cell tumors (n = 26); these cancers were infantile fibrosarcoma [22%], other soft tissue sarcoma [18%], thyroid [11%],
variants were seen in 5%, 0.9%, and 5% of these respective patient populations. Most lung and salivary gland [9% each]); acquired alterations were identified in 16 of these pts.
endometrial, ovarian, and germ cell tumors lacked other K/H/NRAS mutations (83%, 89%, On-target resistance alone was observed in 5 of 16 pts (31%) and were mainly SF or GK
85%, respectively). Other tumor types with hotspot RRAS2 mutations included esoph- single or double mutation-mediated (NTRK1 F589L, NTRK1 G595R, NTRK3 G623R [n = 2],
agogastric cancers (n = 9), cholangiocarcinomas (n = 5), and breast cancers (n = 5, 2 of NTRK3 G623R/G696A). One xDFG mutation was identified. Off-target resistance alone was
which were triple negative). Treatment with RMC-6236 reduced ERK and P90 RSK observed in 7 of 16 pts (44%) and included hotspot KRAS G12D/A/S/V or G13D, PIK3CA
phosphorylation in CAL-51 (human triple-negative breast cancer line) and A2780 cells E545K or E542A, BRAF V600E, and GNAS R844H/C mutations. Complex, combined on-
(human ovarian carcinoma cell line), both harboring the RRAS2Q72L mutation. Treatment of target and off-target resistance was observed in 4 of 16 pts (25%): on-target SF or GK
mice bearing CAL-51 xenograft tumors with RMC-6236 (50 mg/kg, once daily) significantly alterations (NTRK1 G595R, NTRK1 F589L/G595R, NTRK3 G623R, NTRK3 G623R/F617L)
reduced tumor growth. Growth of cells harboring RRAS and RRAS2 mutations was also co-occurred with KRAS G12D or G12D/G13D, and NRAS G12D or Q61H. An analysis of
blocked by MEK1/2 and ERK1/2 inhibitors. Conclusions: Hotspot RRAS2 mutations are resistance profiles by cancer type and age will be presented. Conclusions: In this analysis,
rare but recurrently found in endometrial, ovarian, and germ cell tumors. These mutations on-target resistance to laro, including potential double NTRK resistance mutations, was
are predominantly mutually exclusive with other canonical RAS mutations, although co- commonly observed. Off-target, largely MAPK or PI3K/AKT pathway reactivating resis-
mutations with RAS do occur in a subset. RRAS2Q72L-mutant cancer cells are sensitive to tance, also occurred. In select cases, complex and likely polyclonal resistance including
inhibition of the MAPK pathway including pan-RAS inhibition both in vitro and in vivo. both on-target and off-target alterations were identified. These observations impact novel
These preliminary findings may inform future therapeutic strategies for patients with therapy development for TRK fusion cancers. Clinical trial information: NCT02637687,
RRAS2-mutated solid tumors. Research Sponsor: Memorial Sloan Kettering Cancer Center NCT02576431, NCT02122913. Research Sponsor: Bayer HealthCare Pharmaceuticals, Inc.
Department of Pathology and Lab Medicine.

3067 Poster Session 3068 Poster Session

Non-invasive PD-L1 prediction in NSCLC patients using 3D self-supervised Development and implementation of molecular oncology test e-consult at
deep learning and radiomics. First Author: Xavier Rafael-Palou, Quantitative Im- the VA North Texas Health Care System (VANTHCS). First Author: Lucas Jiaxue
aging Biomarkers (Quibim), Valencia, Spain Wang, UT Southwestern Medical Center, Dallas, TX
Background: Non-small cell lung cancer (NSCLC) is the most common subtype of lung Background: The VA National Precision Oncology Program (NPOP), launched in 2016 as
cancer for ~85% of all cases. Despite therapeutic advances, prognosis remains poor, part of the White House’s Cancer Moonshot Initiative, aimed to revolutionize cancer care
especially in advanced stages. Immunotherapy has revolutionized NSCLC treatment, with through precision medicine for Veterans. Oncologists at the VANTXHCS utilized mo-
immune checkpoint inhibitors (ICIs) targeting the programmed death-ligand 1 (PD-L1) lecular testing to identify potentially actionable mutations to tailor treatment strategies
pathway. While PD-L1 expression is typically measured via immunohistochemistry (IHC), more accurately. This study aims to evaluate the impact of the identification and impact
predictive modeling using CT images could offer a non-invasive alternative to enhance of potentially actionable mutations with treatment decisions and overall survival.
patient stratification and treatment planning in hard-to-biopsy cases and tumor follow-up Methods: We conducted a retrospective chart review of Veterans with molecular on-
of clonal resistance. We propose a solution for non-invasive prediction of PD-L1 ex- cology testing (MOT) at VANTXHCS from August 2019 to May 2024 to assess cancer
pression leveraging radiomics and AI. Methods: This multicentric retrospective study type, prevalence of potentially actionable mutations, treatment decisions, targeted
included NSCLC patients from five real-world data sources who underwent CT and biopsy. therapy utilization, and overall survival (OS). Potential actionable mutations were de-
CT scans were quality-checked and annotated by imaging experts supervised by radi- termined based on AMP/ASCO/CAP Variant Categorization mapped to the OncoKB FDA-
ologists (. 5 years’ experience) to delineate primary tumors. PD-L1 levels were obtained Recognized Human Genetic Variant Database at the time of review. Results: 570
via IHC. The cohort was randomly split into training and test sets (80/20%), with 5-fold Veterans, almost exclusively male (N = 523, 92%) with solid tumors had MOT during the
cross validation for model building and fine-tuning. Three methods—radiomics, deep study period. Lung (N = 211, 37%), GI (N = 105, 21%), prostate (N = 49, 9%), pan-
learning, and deep radiomics—were proposed for binary PD-L1 prediction (cut-off . 1%) creatobiliary (N = 46, 8%), head and neck cancers (N = 35, 6%) and GU (N = 30, 5%) were
based on 3D lesion-centered patches. The radiomics pipeline included 3D feature ex- most common with a median overall survival (OS) of 15.93 months. Potential actionable
traction, standardization, dimensionality reduction and classifier selection. The deep mutations were present in 107 (18.7%) tumors, however only 41 (38%) of the mutations
learning approach used a self-supervised 3D network, pretrained on 2420 lung lesion were associated with an FDA approved targeted therapy at the time of testing. Six
patches from 751 CTs by minimizing dissimilarity between augmented pairs and then fine- Veterans (5.6%) received targeted therapy. Reasons for not receiving targeted therapy
tuned on the training set to predict PD-L1 expression. The deep radiomics method fused included secondary mutations (i.e., KRAS), ECOG performance status, hospice or
both methods via weighted averaging of predicted probabilities. Results: A total of 324 community care and death. Veterans with potentially actionable mutations had sig-
patients (41% women, 63 6 10 years) with varying PD-L1 expression (63% with levels . nificantly worse. Conclusions: In the VANTXHCS Veteran population males with lung
1%), were included. The deep radiomics approach achieved the highest performance, with cancer represents the main tumor type with molecular oncology testing. We observe
AUCs of 75.9% 6 5.3% and 70.1%, and F1-scores of 78.1% 6 1.2% and 76.7% on the actionable mutations in nearly 20% of patients tested, with only a minor subset receiving
validation and test sets, respectively, with a per-instance processing time of 1.2 6 1.3 targeted therapy. Our data showed that patients with a potentially actionable mutation
seconds. Compared to the radiomics method, it improved AUC by 2.6% and 1.2%, and F1- have worse overall survival than those that do not. However, treatment with a targeted
score by 2.7% and 3% on the validation and test sets, respectively. Compared to the deep therapy can significantly improve survival. In conclusion, these findings underscore the
learning approach, it showed AUC gains of 0.3% and 2%, and F1-score gains of 11.3% and need for further research to identify barriers to increase appropriate use of targeted
18.4% on the validation and test sets, respectively. Conclusions: This study demon- therapies in a timely fashion to improve patient outcomes and advance precision
strates the effectiveness of a novel 3D image-based approach combining radiomics and oncology practices. Research Sponsor: None.
3D self-supervised learning to predict PD-L1 expression in a heterogeneous NSCLC cohort
using real-world data. The model executes in seconds and could be regulatory cleared and
deployed in clinical practice as a medical device performing non-invasive PD-L1 ex-
pression identification from CT scans. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 205s

3069 Poster Session 3070 Poster Session

Phase 2 multicenter clinical trial to evaluate the safety and efficacy of Analysing the impact of size of NGS panel in defining first line therapeutic
abenacianine for injection (VGT-309), a tumor-targeted, intraoperative mo- strategies in NSCLC. First Author: Kshitij Joshi, MOC Cancer Care & Research Centre,
lecular imaging agent, for patients undergoing surgery for cancer in the lung. Mumbai, India
First Author: Sunil Singhal, University of Pennsylvania Perelman School of Medicine, Background: NCCN recommends the analysis of 8 genes (EGFR, ALK, ROS1, BRAF,
Philadelphia, PA KRAS, MET, RET, ERBB2, and NTRK1/2/3) for NSCLC patients to identify efficacious
Background: Molecular targeted agents have revolutionized cancer diagnosis and target therapies. Concerned with the rising incidence of sub-optimal response to first-
treatment. Intraoperative molecular imaging (IMI) is a novel technique that entails line therapy and rather early progression of disease we performed retrospective analysis
performing real-time, in vivo optical imaging during surgery to ensure cancer clearance. in a subset of patients treated at our hospital in order to streamline molecular evaluation
A multicenter clinical trial was conducted using abenacianine, a cathepsin targeted near- strategies. Methods: In this study, we retrospectively evaluated the impact of NGS
infrared (NIR) IMI agent to visualize pulmonary nodules during surgery. The primary panel sizes in therapy-naı̈ve NSCLC patients. 242 therapy naı̈ve patients evaluated for
objective was to assess the efficacy of abenacianine in localizing pulmonary lesions, molecular genetic profiling were stratified into three groups based on gene panel size: a)
evaluating surgical margin, and identifying unsuspected disease. Methods: Participants Small panel (,20 genes): Focused on NCCN-recommended genes, b) Medium panel
scheduled to undergo surgery for known or suspected cancer in the lung received (50–100 genes): Included organ agnostic genes, c) Comprehensive panel (.100 genes):
abenacianine (0.32 mg/kg intravenous) preoperatively 12 to 36 hours before surgery. Included genomic signatures like TMB, MSI & HRD scores. Results: Of 242 therapy-
During a standard-of-care surgery, the lung underwent IMI and additional disease not naı̈ve NSCLC patients, 60% (145/242) were evaluated using a small panel of which 13%
identified by conventional methods were identified and analyzed. Efficacy was measured (19/145) had no detectable genetic alterations while 37% (54/145) had 1st line targetable
by frequency of clinically significant events (CSEs) defined as localization of lesions not mutations, 31% (45/145) exhibited both targetable and resistance causing mutations,
found by standard surgical techniques, identification of additional cancers, identification and 19% (28/145) showed only resistance causing mutations. In the 50–100 genes
of positive surgical margins confirmed by histology, and detection of cancerous lymph Panel, comprising 29% (70/242) of patients, 10% (7/70) had no genetic alterations, while
nodes. Results: 89 participants were included in the study. The mean age was 67 years 26% (18/70) had 1st line targetable mutations, 30% (21/70) demonstrated both tar-
and 57 (64%) were female. Of 89 participants administered abenacianine who underwent getable and resistance mutations, and 34% (24/70) harboured only resistance causing
standard of care surgical resection for known or suspected cancer in the lung, 40 (45%) mutations. Finally, in the comprehensive NGS group (.100 genes), which accounted for
had at least one CSE. Abenacianine with NIR imaging identified lesions that were not 11% (25/242) of cases, only 4% (1/25) lacked detectable genetic alterations; while, 12%
found by standard surgical methods in 34 (38%) participants, synchronous and occult (3/25) had 1st line targetable mutations, 32% (8/25) exhibited both targetable and
cancers that were not found by pre-operative imaging in 2 (2%), margins within 10 mm of resistance causing mutations, and 52% (13/25) showed only resistance causing mu-
the closest staple line in 8 (9%), and lymph nodes determined to be cancerous in 1 (1%) tations. Conclusions: a. Increase in gene panel size results in reduction of true neg-
participant. Tumors visualized by IMI with abenacianine included non-small cell lung atives. Hence smaller panels may not necessarily capture resistance causing mutations.
cancers (adenocarcinoma, squamous cell carcinoma, neuroendocrine tumor) and b. As the gene panel size increases, the detection of actionable driver mutations (e.g.,
cancers that metastasized to the lung (breast, colorectal, prostate, thymoma, renal cell, EGFR, ALK) remains consistent; however, there is a notable shift in the mutation profile,
sarcoma). Abenacianine was safe and well tolerated in this study; there were no drug- with a decrease in cases harbouring only targetable mutations and an increase in those
related serious adverse events. Conclusions: This Phase 2 multicenter study dem- exhibiting both actionable and resistance causing mutations. Hence opting for com-
onstrated that using IMI with abenacianine during standard-of-care lung cancer surgery prehensive NGS profiling at baseline may increase diagnostic costs marginally, but will
markedly improved clinical outcomes. Abenacianine localized tumors intraoperatively, have significant impact in designing more effective 1st line therapeutic strategies.
identified synchronous and occult lesions, helped assess negative margin status, and Research Sponsor: None.
identified cancerous lymph nodes enabling a more complete oncologic resection.
Clinical trial information: NCT06145048. Research Sponsor: Vergent Bioscience, Inc.;
Vergent Bioscience Australia Pty Ltd.

3071 Poster Session 3072 Poster Session

68 18
Evaluation of Ga-FAPI PET/CT and F-FDG PET/CT for the primary stag- Clinical utility of comprehensive transcriptome testing in advanced solid
ing of non-small cell lung cancer (NSCLC). First Author: Daniel Udayan C, P K Das tumors. First Author: Fei Su, University of Texas MD Anderson Cancer Center, Houston,
Institute of Medical Science, Palakkad, India TX
Background: Fibroblast activation protein inhibitor (FAPI) radiolabeled with Gallium-68 Background: Despite breakthroughs in genomically-matched therapies, many patients lack
or Fluorine-18 have emerged as promising tracers for targeting cancer-associated fi- actionable genomic alterations. Consequently, innovative new strategies to identify targets
broblasts (CAFs) within the tumor microenvironment. Previous studies in lung cancer for effective therapies are imperative. To assess the potential of comprehensive whole
demonstrated that FAPI PET/CT is more sensitive in detecting metastases in the brain, transcriptomic sequencing (WTS) in clinical decision-making, we conducted a prospective
lymph nodes, pleura, and bone. This study aimed to evaluate the clinical utility of FAPI trial to perform WTS in patients with metastatic or advanced solid tumors who had prior DNA-
PET/CT compared to FDG PET/CT for the primary staging of newly diagnosed NSCLC based panel testing ($100 genes) with no reported AMP/ASCO/CAP Tier 1 actionable ge-
patients. Methods: This prospective study was done at Amrita Institute of Medical nomic alterations; along with whole exome sequencing (WES) to compare RNA expression to
Sciences, Kochi between August 2022 to December 2023 among patients with newly copy number alterations. We also developed an informed actionability classification scheme
diagnosed NSCLC who consented to undergo both scanning i.e. FDG as well as FAPI PET to determine actionability of transcriptomics findings. Methods: Between August 2022 and
CT before initiating any treatment. This study was approved by institutional review board. December 2024, 100 patients at MD Anderson Cancer Center with advanced cancers were
Kolmogorov Smirnov one sample test was used to check the normality of data. To test the enrolled and underwent comprehensive profiling. Alterations in RNA expression of 147 genes
were considered as potentially actionable if the gene’s protein product is the direct target of a
statistical significance of the difference in the average values of tumor volume, stan-
clinically available therapy (including cell surface targets of antibody-drug conjugates
dardized uptake value (SUV), target to background ratio (TBR) and background uptake
(ADCs)) and/or CNAs of the gene are predictive of response or resistance to a clinically
between FAPI and FDG, paired sample t test was used for normality and Wilcoxon signed
available therapy. A clinical trial was considered a match if RNA expression was an enrollment
rank test was used for non-normality. Results: In this study, 42 patients with newly criterion, or the gene alteration can be directly or indirectly targeted with a therapy utilized
diagnosed NSCLC (32 with adenocarcinoma (AC) and 10 with squamous cell carcinoma within the trial. Results: Actionable RNA expressions (AREs) were detected in all patients
(SCC)) were included. Comparison of the results between FAPI and FDG PET in pa- (100%). In total, 2,216 AREs were detected from 17,800 selected-reported RNA expressions; a
rameters like TBR, SUVmax and background value in metastatic lesions, FAPI performed median of 22 ARE changes were reported per patient [interquartile range (IQR), 17.0-26.0].
better. In case of lymph nodes FAPI vs FDG, mean TBR was (5.06 6 4.19 v/s 3.02 6 2.89) The majority (86.0%) of AREs were RNA overexpressions, defined as the distribution of tpm
p value=0.002 and in SUVmax value was (9.07 6 5.1 v/s 6.59 6 4.2) p value=0.01. Similar values of the gene expression . 83% of the pan-cancer reference cohort. A median of 13 RNA
benefits were seen in TBR and SUVmax with FAPI on metastatic site like pleura and bone. expression-matched trials were identified per patient [IQR, 10.0-16.0]. The most frequent drug
In the primary lung lesion, mean tumour volume (MTV) was larger with FAPI but there was class of actionable RNA overexpressions were ADCs, with a median of 10 ADC targets per
no difference in SUVmax, TBR and back ground uptake value. If we compare AC vs SCC patient. Out of 11 distinct genes of which at least 1 amplification is reported, a concordance
cases, there was no advantage for SCC with FAPI in metastatic lesions as well as primary rate of 61.5% and a Concordance Correlation Coefficient (CCC) value of 0.70 (95% CI, 0.56-
lesions. Driver mutated AC cases performed extremely well with FAPI. In brain, mean SUV 0.84) between the actionable gene amplifications (n = 13) and the actionable RNA over-
max with FAPI was 4.52 6 0.89 compared to 8.45 6 3.31 with FDG. But the brain lesions expressions detected in the same patient sample from the tumor profile, indicates a sub-
identified with FAPI were higher than FDG due to higher TBR. In liver, SUV max was similar stantial concordance by transcriptional profiling with copy number gain. Conclusions: WTS
between FAPI and FDG but higher TBR was seen with FAPI. Conclusions: 68Ga-FAPI PET/ identified actionable RNA expressions in all patients– including for novel ADC targets. These
CT performed better than 18F-FDG PET/CT in the primary staging of NSCLC. FAPI PET may results underscore the utility of comprehensive transcriptional profiling to identify additional
be considered instead of FDG PET in staging of NSCLC. Patient compliance was also actionable targets beyond DNA-based comprehensive profiling. Clinical trial information:
better because fasting and glycemic control was not required prior to FAPI. This is the MDACC 2021-1049. Research Sponsor: Strategic Alliance: BostonGene2MD Anderson
only study where histology (AC and SCC) has been directly compared with both FAPI/PET Cancer Center Feasibility and Clinical Utility of Combined Genomics/ Transcriptomics with
vs FDG/PET – it shows FAPI/PET performs better with AC, and this South East Asian study Systems Biology for Personalized Cancer Therapy; MD Anderson Cancer Center Support
showed FAPI/PET performs better with driver mutated NSCLC. Research Sponsor: None. grant; Center for Clinical and Translational Science.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
206s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3073 Poster Session 3074 Poster Session

Application of an epigenomic-based classifier to identify cancer signal of HRD status prediction in patients with advanced breast, prostate, ovarian
origin on liquid biopsy in cancer of unknown primary cases. First Author: Elmira and pancreatic cancers in a liquid biopsy assay. First Author: Pegah Safabakhsh,
Forouzmand, Guardant Health, San Diego, CA Guardant Health, Palo Alto, CA
Background: Cancer of unknown primary (CUP) lacking resolution to a cancer type Background: Homologous recombination and repair (HRR) deficiency (HRD) is charac-
(Cancer Signal Origin; CSO) leads to suboptimal outcomes. While several approaches terized by genomic instability associated with mutations in BRCA1/2 or other HRR genes.
currently exist for identifying a cancer type for CUP samples, they rely on clinical HRD can also be detected by copy number variant (CNV) features, indels, and SNVs (HRD
approaches that often require tissue biopsies for immunohistochemistry (IHC), and even signature). Patients with canonical BRCA-associated cancers harboring an HRD signature
still they often fail to resolve a CSO (reported to occur in 50-80% clinical cases). This with or without HRR mutations derive clinical benefit from PARPi therapy. Here, we
requirement for tissue, and typically lengthy diagnostic journey, create a large unmet present a method of predicting HRD status using Guardant Infinity in patients with these
need for CSO identification in CUP individuals. Here, we present feasibility data from a advanced cancers. Methods: We developed an ensemble logistic regression model to
high accuracy Liquid Biopsy method for CSO identification in CUP, bypassing the need predict HRD status, inferred from genome-wide somatic SNV and CNV signatures indicative
for a tissue biopsy and quickly returning a CSO to individuals with CUP. Methods: We of deficiency in HRR genes, including ploidy-adjusted large-scale state transitions (LST),
developed a CSO prediction algorithm on Guardant360 utilizing DNA methylation sig- whole-genome tumor loss of heterozygosity (LOH) and telomeric allelic imbalance (TAI). The
natures across thousands of cancer-specific differentially methylated regions for 14 model was trained on clinical samples processed on Guardant Infinity, a next-generation
cancer types. We applied the CSO classifier to 1,128 CUP samples in which circulating platform evaluating both genomics and epigenomics, to assess the sensitivity and accuracy
of detecting biallelic loss-of-function in BRCA1/2. The aggregated model was tested on an
tumor DNA was detected. Accuracy was assessed by comparing CSO predictions to
independent pan-tumor clinical cohort and pre-treatment samples from a subset of patients
suspected diagnoses based on clinicopathologic and molecular findings. Results: The
enrolled in TRITON2, a phase 2 single arm study evaluating rucaparib in metastatic cas-
CSO prediction algorithm was evaluated on 1,128 CUP samples; lung (285/1128, 25.3%) tration resistant prostate cancer (mCRPC) patients with HRR mutations. HRD status as-
and bile duct (166/1128, 14.7%) were the most common predicted CSOs, aligning with sociation with radiographic progression free survival (rPFS) was evaluated with Cox-
reported prevalence in the literature. Of the 1,128 samples, 12 had a suspected clinical proportional hazards model. Results: Our model demonstrated high sensitivity in pa-
diagnosis. These 12 spanned 8 tumor types. The top CSO prediction aligned with tients with BRCA1/2 biallelic loss and high specificity in HRR-wildtype patients, with an AUC
suspected diagnosis in 91.6% (11/12) of cases. The CSO algorithm also provides of 0.95 in a pan-tumor cohort with tumor fraction (TF) .10%. In an independent cohort of
confidence scores to quantify the confidence of CSO prediction. Out of the total 12 breast prostate ovarian and pancreatic samples, HRD detection ranged from 79-100% in
cases, 7 CSOs had high confidence, 3 had moderate confidence, and 2 had low con- samples with BRCA1/2 biallelic loss and .10% TF. In breast (n = 703) and prostate (n = 655)
fidence. 7/7 high, 2/3 moderate, and 2/2 low confidence predictions were correct. The cancers, HRD was detected in 14.9% and 13.5% of samples with . 10% TF (3.6% and 3.7% in
single incorrect moderate sample was diagnosed as CRC but predicted to be lung. all TF), with 5.5% and 6.2% attributed to samples not harboring deleterious mutations in HRR
Conclusions: These findings show the feasibility of using plasma-based epigenomic genes, respectively, potentially reflecting non-genomic drivers of HRD. In a pilot cohort (n =
profiling to assign CSO with acceptable accuracy. While interventional clinical studies 15) from TRITON2, HRD was detected in 100% of patients enrolled with either BRCA1/2, or
are necessary to demonstrate clinical utility, this has significant potential for guiding PALB2 mutations (n = 10), where rucaparib demonstrated meaningful activity as measured
treatment decisions and improving outcomes in CUP patients without ready access to by independent radiology review objective response rate. HRD was not detected in patients
tissue. Research Sponsor: Guardant Health, Inc. with mutations in CDK12, FANCA, or NBN (n= 5). HRD detected status was associated with
prolonged rPFS (HR = 0.07, p = 0.03). Conclusions: Guardant Infinity can predict HRD status
in patients with advanced canonical BRCA-associated cancers, with preliminary results
indicating potential for predicting PARPi benefit in mCRPC. Further studies are warranted to
determine PARPi response for breast, prostate, ovarian, and pancreatic cancers with de-
tected HRD status. Research Sponsor: None.

3075 Poster Session 3076 Poster Session

Efficacy and safety of distinct regimens for individuals with advanced EGFR- Preliminary results of a first-in-human phase 1b (aCCeleR8-001) study of S-
mutated non-small-cell lung cancer who progressed on EGFR tyrosine- 531011, a humanized anti-CCR8 monoclonal antibody, in patients with
kinase inhibitors: A systematic review and network meta-analysis. First advanced solid tumors. First Author: Toshihiko Doi, National Cancer Center Hos-
Author: Zhang Wengang, Shanghai Pulmonary Hospital, Shanghai, China pital East, Chiba, Japan
Background: Targeted therapy with EGFR tyrosine-kinase inhibitors (TKIs) is the pre- Background: C-C motif chemokine receptor 8 (CCR8) is selectively upregulated in tumor-
ferred first-line treatment for EGFR-mutated advanced non-small cell lung cancer infiltrating regulatory T cells (TI-Tregs) in multiple cancers, inhibiting anti-tumor activity of
(NSCLC), but acquired resistance inevitably occurs in almost all responding individuals. the host immune system. S-531011, a humanized IgG1 monoclonal antibody, is anticipated
We aimed to comprehensively review the literature to investigate the efficacy and safety to deplete CCR8-positive TI-Tregs, restoring anti-tumor immunity without inducing auto-
of distinct regimens in the subsequent-line setting, thereby identifying the optimal immunity. Methods: An aCCeleR8-001 study is a Phase 1b/2, multicenter, open-label study
regimen for these TKI-resistant NSCLC patients. Methods: The PubMed, Embase, of S-531011 which consists of Phase 1b Dose Escalation part (Parts A-1 and A-2) and Phase
Cochrane Library databases, and abstracts of ASCO, ESMO, and WCLC were searched from 2 Dose Expansion part. The safety/tolerability, pharmacokinetic (PK), pharmacodynamic, and
database inception to 3 November 2024, to identify eligible randomized controlled trials anti-tumor activity of S-531011 as monotherapy and in combination with pembrolizumab
(RCTs) that assessed distinct regimens for individuals with advanced EGFR-mutated (Merck & Co., Inc.) were evaluated in patients with various types of locally advanced or
NSCLC who progressed on TKIs. The outcomes of progression-free survival (PFS), overall metastatic solid tumors. S-531011 monotherapy was administered at 8, 24, 80, 240, 800, or
survival (OS), objective response rate (ORR), disease control rate (DCR), and grade 3 or 1600 mg/kg intravenously every 3 weeks (Q3W) in Part A-1, whereas patients in Part A-2
higher adverse events ($3AEs) were compared and ranked in overall patients and various received S-531011 at 80, 240, 800, or 1600 mg/kg in combination with pembrolizumab 200
subgroups among 8 regimens by network meta-analysis and the surface under the mg/kg Q3W. The data were analyzed when all patients in Dose Escalation cohorts completed
cumulative ranking curve, respectively. The protocol is registered with PROSPERO, the dose-limiting toxicity (DLT) observation period. Results: As of the data cutoff date (30
CRD42024601619. Results: 14 RCTs, involving 3177 participants and 8 treatment reg- Sep 2024), 40 and 35 patients were enrolled in Parts A-1 and A-2, respectively. No DLTs were
imens (chemotherapy plus ivonescimab (PD-1/VEGF inhibitor) [CT+IVO]; CT+amivanta- reported at any dose level and the maximum administered dose of S-531011 was 1600 mg in
mab+lazertinib [CT+AMI+LAZ], CT+immunotherapy+bevacizumab [CT+IO+BEV], CT+AMI, both parts. One patient reported an infusion-related reaction in Part A. Immune-related
adverse events (irAEs) were reported in two patients (5.0%; Grade 1/2 only) in Part A-1,
CT+BEV, CT+IO, CT, and IO), were included. In overall patients, the most pronounced PFS
whereas 15 irAEs were reported in 10 patients (28.6%; including four Grade 3 irAEs in four
benefit was observed with the CT+IVO, followed by CT+AMI+LAZ, CT+IO+BEV, and
patients) in Part A-2. PK of S-531011 was approximately dose proportional with a terminal
CT+AMI, ranked second, third, and fourth, respectively. In terms of OS, the regimen of
elimination half-life of 10 to 12 days regardless of dose level. CCR8 receptors in PBMCs were
CT+AMI ranked the best, followed by CT+IVO. However, the comparisons of OS among
occupied at doses of 80 mg or higher. PK/CCR8 receptor occupancy modeling analysis
different regimens did not reach statistical significance, possibly due to immature data.
indicated that . 90% of receptors in tumor tissues were occupied in the range of 80 to
The results for ORR and DCR were similar to those for OS, with CT+AMI topping the 800 mg. Multiplex immunohistochemistry analysis demonstrated proof of mechanism as
rankings, followed by CT+AMI+LAZ. In terms of safety, the incidence of $3AEs was evidenced by CCR8-positive Treg depletion in tumor tissue at doses of 24 mg or higher.
highest in CT+AMI+LAZ, followed by CT+AMI. In subgroup analysis, CT+IVO demonstrates Among 62 evaluable patients dosed at 80 to 1600 mg in Part A, four patients (6.5%) had
stable PFS benefits across clinicopathological characteristics, ranking first in most confirmed partial response, three of whom had colorectal cancer (CRC). Twenty patients
subgroups. Due to the unavailability of OS subgroup data in most RCTs, many regimens (32.3%) had disease control for $ 6 weeks. Response rate was not correlated with dose (80
were missing in the OS subgroup analysis. Conclusions: Integrating the results of dif- to 1600 mg). Following a comprehensive data review, tentative recommended Phase 2 doses
ferent clinical outcomes and subgroup analyses, we conclude that CT+IVO is the optimal were determined to be 80 to 800 mg in both parts. Conclusions: S-531011 was well tolerated
treatment option with an acceptable safety profile for patients with advanced EGFR- up to 1600 mg as monotherapy and in combination with pembrolizumab. A higher response
mutated NSCLC who have progressed on TKIs. CT+AMI+LAZ and CT+AMI are alternative rate in patients with CRC warrants further exploration of this tumor type in Phase 2 Dose
subsequent line options as well, with superior efficacy compared to immunotherapy- Expansion part. Phase 2 CRC cohorts are currently ongoing. Clinical trial information:
based or chemotherapy regimens, yet elevated toxicity profiles requiring vigilant NCT05101070. Research Sponsor: Shionogi & Co., Ltd.
management. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 207s

3077 Poster Session 3078 Poster Session

First report of ROR2 directed therapy with a conditionally active antibody Glutaminase isoform expression in cancer: Implications for metabolic ad-
drug conjugate in advanced melanoma. First Author: Jacob Keeling, Sarah Cannon aptation and therapy. First Author: Mohammed Osama Ahamd Bader, University of
Research Institute at HealthONE, Denver, CO Khartoum Faculty of Medicine, Khartoum, Sudan
Background: The receptor tyrosine-kinase like orphan receptors (ROR) are mediators of non- Background: Glutamine is a critical amino acid involved in various metabolic pathways,
canonical WNT signaling and tissue patterning. ROR2 upregulation is observed in malignancy and particularly in cancers where its importation is significantly elevated via multiple
has been implicated in metastasis. Development of ROR2 directed conditionally active antibodies transporters. Glutaminase, the enzyme catalyzing the deamination of glutamine to
with enhanced affinity in the tumor microenvironment is a promising treatment strategy. Here we glutamate, has two isoforms: kidney-glutaminase 1 (GLS1) and liver-glutaminase 2
report our institution’s experience treating 5 patients with advanced cutaneous or uveal melanoma
treated with the anti-ROR2 antibody drug conjugate ozuriftamab vedotin (MMAE) on the first in
(GLS2). This study investigates the expression of glutaminase isoforms in cancers
human phase 1 trial. (NCT03504488). Methods: Adults with advanced solid tumors naı̈ve to vinca originating from tissues with high glutaminase activity—namely, clear renal cell car-
binding site therapies who had failed all standard of care therapy were eligible. The charts of all 5 cinoma (KIRC), chromophobe renal carcinoma (KICH), papillary renal carcinoma (KIRP),
patients with advanced melanoma treated at our institution were reviewed. Patients were treated hepatocellular carcinoma (LIHC), and glioblastoma (GBM)—to understand the fate of the
with ozuriftamab vedotin at concentrations of 1.8 mg/kg or 3.0 mg/kg IV Q2W. Safety and efficacy imported glutamine. Methods: The Cancer Genome Atlas (TCGA), Tumor Immune
data were collected. Patient samples were evaluated for pharmacokinetics and clinical correlates. Estimation Resource ([TIMER] 2.0), Gene Expression Profiling Interactive Analysis
Results: 4 out of the 5 patients achieved an objective response in target lesions, per RECIST v1.1 ([GEPIA] 2.0), and the University of Alabama at Birmingham Cancer Data Analysis
(Table 1). Two of these patients have maintained disease control, including one patient in complete ([UALCAN]) Portal were used to investigate GLS1 and GLS2 expression. TIMER 2.0
remission (CR) .5 years and the other responding .1 year after starting treatment. Adverse analyzed 533 KIRC (72 normal), 66 KICH (25 normal), 290 KIRP (32 normal), 153 GBM (5
events requiring dose reduction or interruption included neuropathy and neutropenia, both of which
normal), and 371 LIHC (50 normal) samples. GEPIA 2.0 analyzed 523 KIRC (100 normal),
recovered with reduced dosing and/or colony stimulating factor. No patients discontinued therapy
for adverse drug reactions. Pharmacokinetics showed predictable plasma concentrations of both
66 KICH (53 normal), 286 KIRP (60 normal), 163 GBM (207 normal), and 369 LIHC (160
drug and free MMAE. Anti-drug antibodies were not identified. Biopsies were assessed by IHC for normal). UALCAN analyzed 533 KIRC (72 normal), 66 KICH (25 normal), 290 KIRP (32
ROR2. The biopsy belonging to the patient who achieved CR was strongly positive for ROR2. All normal), 153 GBM (5 normal), and 371 LIHC (50 normal) samples. Additionally, datasets
other biopsies showed low/negative ROR2 staining of malignant cells. Conclusions: Ozuriftamab from NCBI GEO were used, including GSE15641 (23 normal, 32 KIRC, 6 KICH, and 12
vedotin showed early promising antitumor activity in this first report describing ROR2 directed KIRP), GSE7696 (4 normal, 40 GBM), and GSE41804 (20 normal, 20 LIHC). These
treatment in refractory advanced cutaneous and uveal melanoma. Clinical trial information: platforms detected GLS1 expression in KIRC, KICH, KIRP, and GBM and GLS2 in LIHC by
NCT03504488. Research Sponsor: None. comparing tumor and normal samples. Results: GLS1 was significantly downregulated
Efficacy and safety of ozuriftamab vedotin. in KIRC, KICH, KIRP, and GBM across TIMER 2.0, GEPIA 2.0, and UALCAN (P , 0.05).
Case 1 Case 2 Case 3 Case 4 Case 5 GLS2 was also significantly downregulated in LIHC (P , 0.05). NCBI GEO datasets
Melanoma Subtype Uveal Uveal Cutaneous Cutaneous Cutaneous
(GSE15641 for kidney cancers, GSE7696 for GBM, and GSE41804 for LIHC) supported
Initial Dose (mg/kg) 1.8 1.8 3.0 1.8 1.8 these results, showing consistent GLS1 downregulation in KIRC, KICH, KIRP, and GBM,
Final Dose (mg/kg) 1.5 1.5 1.8 1.8 1.8 and GLS2 downregulation in LIHC (adjusted P , 0.05; |Log2FC| . 1). Conclusions: The
Best Target Lesion 10% increase 31.9% de- 89% decrease 38.2% de- 43.7% de- tissue-specific downregulation of glutaminase isoforms—GLS1 in kidney and brain
Response crease (PR) (CR, w/ lymph crease crease
nodes , 5 mm) (PR) (PR) cancers, GLS2 in liver cancer—highlights an adaptive mechanism in cancer cells to limit
Time to Progressive 35 +475* +2079* 127 84 glutamine deamination and preserve imported glutamine for other metabolic needs.
Disease (days) Enhancing the deamination process could deprive cancer cells of essential precursors
Major Adverse G3 Neutropenia G2 G2 Neuropathy, G4 None None
Events Neuropathy Neutropenia for nucleotide synthesis, disrupting their growth and survival. These isoforms represent
Other Adverse G2 Transaminitis, G2 None G1 Salivary Inflam- G2 G2 potential diagnostic markers and therapeutic targets. Research Sponsor: None.
Events Myalgia, G2 Arthralgia mation, G1 Alopecia Neuropathy Neuropathy
*Patient has not developed progressive disease since starting treatment.

3079 Poster Session 3080 Poster Session

A phase 1 study of PARP inhibitor (niraparib) plus HSP90 inhibitor (pimi- Development and validation of an AI-enabled prediction of prostate cancer
tespib) in solid tumors: Dose-expansion results from the NiraPim (PCa) using urine-based liquid biopsy. First Author: Marvin S. Hausman, Genetics
(EPOC2102) study. First Author: Yasuyuki Kawamoto, Cancer Center, Hokkaido Institute of America, Delray Beach, FL
University Hospital, Sapporo, Japan Background: Prostate cancer (PCa) remains a major cause of malignancy-related
Background: Heat shock protein 90 (HSP90) inhibitors have shown potential in desta- mortality among men. Current diagnostic techniques, including PSA testing, lack ac-
bilizing homologous recombination repair (HRR) proteins, thereby inducing homologous curate early detection capabilities, while global barriers include limited access to
recombination deficiency and enhancing PARP inhibitor efficacy. The NiraPim (EPOC2102) specialized facilities and cultural sensitivities around transrectal biopsy and digital
study is a phase 1 study to evaluate this combination therapy in humans, investigating the rectal examination. This study evaluates a non-invasive, urine-based liquid biopsy assay
safety and efficacy of combining niraparib, a PARP inhibitor, with pimitespib, a novel for diagnosing PCa through disease-specific biochemical profiles using an artificial
HSP90 inhibitor, in patients with advanced solid tumors. Following establishing the intelligence pipeline. Methods: We collected urine from men scheduled for prostate
recommended dose (RD) in the dose-escalation part, we present primary analysis results biopsy (biopsy-positive PCa n=197) and healthy controls (n=84). Samples were pro-
from the dose-expansion part. Methods: In the dose-expansion part, patients received cessed using NUTEC slides, underwent heat cycling, and were converted to digital
pimitespib 80 mg (5-day on/2-day off) combined with niraparib 200 mg daily. Cohort A images for AI analysis. Using 5x2 cross-validation with a random forest classifier, we
included patients with BRCA-associated cancers (breast, ovarian, prostate, and pancreatic) evaluated cancer detection performance and analyzed cohorts with specific Gleason
harboring BRCA pathogenic variants and immediately after progression to prior PARP scores (Gle): Gle 6 (n=70), Gle 7 (3+4) (n=55), Gle 7 (4+3) (n=34), and Gle 8,9,10 (n=38).
inhibitors. Cohort B included patients with breast/pancreatic cancer without gBRCA, Results: Our classifier demonstrated strong overall performance in distinguishing
prostate cancer without tBRCA, and other solid tumors (excluding ovarian cancer) not cancer versus non-cancer subjects (F1=0.843) with notably high recall (R=0.967).
previously treated with PARP inhibitors. Results: As of August 2024, 30 patients were Importantly, performance remained robust across Gleason score cohorts (F1=0.799-
enrolled: 14 in cohort A and 16 in cohort B. Cohort A included breast (n=6), ovarian (n=5), 0.838), maintaining high recall (R.0.89) while preserving clinically relevant precision.
prostate (n=2), and pancreatic (n=1) cancers. Cohort B included breast (n=3), prostate The classifier showed particular strength in detecting intermediate- (Gle 7 (3+4):
(n=4), pancreatic (n=4), and other tumors (n=5). The median follow-up period was F1=0.838) and low- (Gle 6: F1=0.822) Gleason grade cancers. Conclusions: AI-enabled
6.0 months. The median treatment cycle was 2 (range 1–18). Treatment-related adverse prediction of PCa using urine-based liquid biopsy demonstrates accurate, rapid, and
events $Grade 3 occurred in 33.3%. Common adverse events ($20.0%) included nausea
accessible early cancer detection, with consistent performance across disease grades.
(73.3%), diarrhea (40.0%), anorexia (23.3%), vomiting (20.0%), fatigue (20.0%), and de-
This non-invasive approach addresses both clinical and cultural barriers to prostate
creased platelet count (20.0%). No treatment-related deaths occurred during the study
cancer diagnostics. Research Sponsor: None.
period. The objective response rate was 10.0% (95% CI: 2.1, 26.5), with disease control rate
of 36.7% (19.9, 56.1) and 3-month PFS of 27.7%. In cohort A, one patient with hormone TASK F1 P R AUC ACC
receptor-positive breast cancer achieved partial response post-olaparib progression. In
Cancer v. Controls 0.843 0.748 0.967 0.768 0.748
cohort B, two patients (leiomyosarcoma and urothelial carcinoma) with BRCA pathogenic Gle6 v. Controls 0.822 0.770 0.893 0.776 0.746
variants achieved partial response, and one prostate cancer patient with CDK12 patho- Gle347 v. Controls 0.838 0.757 0.940 0.752 0.754
genic variant maintained stable disease $3 months. Conclusions: The dose-expansion Gle437 v. Controls 0.800 0.715 0.913 0.695 0.691
part demonstrated a manageable safety profile and potential efficacy at the recommended Gle8910 v. Controls 0.799 0.722 0.900 0.695 0.698
dose of niraparib plus pimitespib. Clinical benefit was observed in both BRCA-associated
cancers resistant to PARP inhibitors and PARP inhibitor-naive non-BRCA associated AI-enabled prediction of PCa using urine-based liquid biopsy demonstrates accurate, rapid, and
accessible early cancer detection, with consistent performance across disease grades. This
cancers, supporting further investigation in biomarker-selected populations. Clinical trial non-invasive approach addresses both clinical and cultural barriers to prostate cancer
information: jRCT2031220179. Research Sponsor: Takeda Pharmaceutical Co., Ltd.; Taiho screening.
Pharmaceutical Co., Ltd.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
208s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3081 Poster Session 3082 Poster Session

Genomic landscape of 5’methylthioadenosine phosphorylase (MTAP) de- Preliminary efficacy results from an ongoing phase I/II trial of CTS2190, a
leted (MTAP loss) non-squamous carcinoma of unknown primary site PRMT1 inhibitor, in patients with advanced/metastatic solid tumors. First
(nsCUP). First Author: Parth J. Sampat, SUNY Upstate Medical University, Syr- Author: Jianan Jin, Oncology Department/Phase I Clinical Center, Zhejiang Cancer
acuse, NY Hospital, Hangzhou, Zhejiang, China
Background: MTAP, a key enzyme in the polyamine pathway breaks down 5’Deoxy-5- Background: Epigenetic gene regulation, including arginine methylation holds significant
Methylthioadenosine (MTA) into methionine and adenine. MTAP loss reduces adenine and accumu- promise in immunomodulation and long survival outcomes. It represents a potential clinical
lates MTA, which inhibits protein arginine methyltransferase 5 (PRMT5). This suggests MTAP loss approach to address the highly unmet needs of patients (pts) with advanced solid tumors who
cancers may respond to PRMT5 inhibition. Methionine adenosyl transferase 2a (MAT-2A) is a primary failed PD-(L)1 immune checkpoint inhibitors (ICIs) or standard of cares（SoCs) therapies.
producer of donor S-adenosylmethionine (SAM) and the depletion of MAT-2A has antiproliferative effect
CTS2190, the orally available, first-in-class small molecule, specifically inhibits arginine
in cancers with MTAP loss. Based on the synthetic lethality concept, MTAP loss is being used as a
biomarker for accrual in multiple trials with PRMT5 and MAT-2A inhibitors. We queried the genomic methyltransferase 1 (PRMT1) with significant reduction of intra-tumor asymmetric dime-
landscape of MTAP loss in patients with nsCUP. Methods: DNA extracted from formalin-fixed paraffin- thylarginine (ADMA) level, DNA damage response (DDR), androgen receptor (AR) level, and
embedded (FFPE) tissue of 7,440 nsCUP cases from 2020 to 2024 underwent hybrid capture-based oncogenic proliferation through epigenetic modulation in various solid tumors. Here we
comprehensive genomic profiling (CGP) to assess all classes of genomic alterations (GA). All cases present clinical data from an ongoing Phase I/II study of CTS2190 (NCT06224387).
underwent central pathology review to confirm that at the time of sequencing, a primary site for the cases Methods: Eligible pts in the dose-escalation stage received 60~300 mg of CTS2190 orally,
was not established. Microsatellite instability (MSI) status and tumor mutational burden (TMB) were while pts in the dose-expansion stage were treated with 180 or 240 mg until disease pro-
derived from the CGP data. Programmed death-ligand 1 (PD-L1) was determined by immunohisto- gression or intolerable toxicity. Efficacy, safety, PK, PD and biomarker profiles were evaluated.
chemistry (IHC) using the DAKO 22C3 system. Results: 853 (11.5%) of nsCUP cases had either complete Results: As of January 24, 2025, 38 pts had received CTS2190 treatment, 32 of them were
or partial MTAP loss with 0.7% 1 exon, 1.2% 2 exons, 2.9% 3 exons, 5.1% 4 exons, 0.5% 5 exons, 2.5% 6 response-evaluable. In the PD-(L)1 primarily resistant group, the objective response rate (ORR)
exons, 32.8% 7 exons and 54.3% 8 exons lost. The median age of the MTAP loss patients was higher (68 vs
and disease control rate (DCR) were 18.2% (2/11) and 72.7% (8/11), respectively. In addition, in
65; p,.0001) and the gender distributions were similar (52% to 54% female; not significant (NS)). Cyclin-
dependent kinase inhibitor 2A (CDKN2A) loss co-occurred in 99.8% in patients with MTAP loss. MSI-high
PD-(L)1 primarily resistant non-small cell lung cancer (NSCLC) subgroup, the ORR and DCR
status was uncommon in both MTAP loss vs MTAP wildtype (0.4% vs 0.7%; NS). The MTAP wildtype group were 28.6% (2/7) and 71.4% (5/7), respectively, with significantly prolonged median
had higher tumor mutational burden (TMB) . 10 mutations/mb (15.7% vs 11.1%; p=.0004) and TMB .20 progression-free survival (PFS) (summarized in the table below). Among 2 response-evaluable
mutations/Mb (5.4% vs 3.5%; p=.017) rates. MTAP loss nsCUPs had higher frequencies of KRAS GA and pts with metastatic castration-resistant prostate cancer (mCRPC), one achieved partial re-
KRAS G12C, whereas MTAP wildtype cases had greater frequencies of ERBB2, PTEN, MET and EGFR GA sponse (PR) while the other exhibited stable disease (SD) with tumor shrinkage. Most
(Table). GA in BRCA1/2 and FGFR2 were similar in both groups. GA in ALK, RET, ROS1, RET and TRK were treatment-related adverse events (TRAEs) were grade 1/2 and manageable. The only TRAE $
extremely uncommon in both groups (all less than 1%). Conclusions: At 11.5%, nsCUP features a grade 3 with an incident rate . 15% was platelet count decreased (31.6%). No TRAEs led to
relatively high frequency of MTAP loss, with the vast majority involving either all (8 of 8) or nearly all (7 of treatment discontinuation or death. CTS2190 exposure increased proportionally with esca-
8) exons. MTAP loss patients are slightly older and have reduced TMB levels which may impact their lating doses, and a PK-PD-efficacy model demonstrated a relationship between CTS2190
responsiveness to immunotherapy-based combination regimens with PRMT5/MAT-2A inhibitors. Clinical exposure, efficacy and PD marker changes. The correlation between clinical efficacy and intra-
trials for the development of targeted therapies to use PRMT5 inhibition and MAT-2A in nsCUP are
tumor PRMT1 expression, as detected by immunohistochemistry (IHC), is under investigation.
warranted. Research Sponsor: None.
Conclusions: CTS2190 demonstrated a favorable safety profile and promising efficacy in
nsCUP MTAP Loss (N=853) nsCUP MTAP wildtype (N=6,587) P value heavily pretreated pts with advanced solid tumors, particularly in immunologically cold mCRPC
KRAS all/G12C 45.9%/7.6% 31.2%/4.2% ,.0001/,.0001 and PD-(L)1 primarily resistant NSCLC. These results position CTS2190 as a promising
ERBB2 all/amp only 6.7%/4.3% 10.9%/8.1% ,.0001/,.0001 therapeutic option to fulfil unmet medical needs following ICIs therapies. Clinical trial in-
PIK3CA 6.2% 7.1% NS formation: NCT06224387. Research Sponsor: CytosinLab Therapeutics Co., Ltd.
BRAF 6.0% 5.0% NS
FGFR2 4.3% 4.0% NS PFS of pts with PD-(L)1 primary resistance.
PTEN 4.1% 6.1% .02 Patients, n ‡ 3 prior lines of therapy, n (%) Event Median PFS (weeks)
BRCA1/2 1.8%/2.3% 2.1%/2.4% NS/NS
MET 2.1% 4.6% .0004 All comers 11 7 (63.6%) 7 12.7 (95% CI: 8.0~35.3)
EGFR 2.6% 4.1% .031 NSCLC 7 4 (57.1%) 5 24.9 (95% CI: 8.0~35.3)

3083 Poster Session 3084 Poster Session

The first-in-human phase 1/2 study of TSN1611, a highly selective KRAS Artificial intelligence (AI)-powered evaluation of protein drug-targetability
G12D inhibitor, in patients with advanced solid tumors. First Author: Siqing Fu, through subcellular-level expression profiling from immunohistochemistry
The University of Texas MD Anderson Cancer Center, Houston, TX (IHC) images. First Author: Sukjun Kim, Lunit Inc., Seoul, South Korea
Background: TSN1611 is a novel small molecule KRAS G12D inhibitor targeting both active Background: As a standardized methodology for quantifying the targetability of proteins in drug
(GTP-bound) and inactive (GDP-bound) forms of KRAS G12D protein. TSN1611 showed high development has yet to be established, we developed an AI-powered analyzer capable of scalably
potency and selectivity against KRAS G12D mutant tumor cells in vitro and effectively measuring cellular and subcellular-level expression to assess 74 membrane-specific targets in de-
inhibited tumor growth in several pancreatic ductal carcinoma (PDAC), colorectal cancer velopment. Methods: A total of 160K cancer and normal IHC images from Human Protein Atlas (HPA)
were analyzed, including 47,591 on 74 target genes. The AI model trained on pathologist-annotated
(CRC) and non-small cell lung cancer models (NSCLC) in vivo. Methods: A phase 1/2 study
histology images, took the IHC images as input to predict cell types and subcellular compartments
of TSN1611 was developed to enroll patients (pts) with advanced solid tumors harboring (nucleus, cytoplasm, and membrane) along with intensity scores. Target genes were evaluated by 1)
KRAS G12D mutation. The study comprised a phase 1a dose escalation part following a Tumor cell specificity (TCS): normalized ratio of positive tumor cells to the total positives, 2) Inverse
BOIN design with accelerated titration to determine the maximum tolerated dose (MTD), normal score (INS): inverse ratio of positive normal cells to the total normal cells, 3) Membrane
recommended phase 2 dose and pharmacokinetics (PK), followed by a phase 1b part for intensity score (MIS) and 4) Membrane specificity (MBS): ratio of MIS to the intensity scores from 3
dose optimization to compare different recommended doses and a phase 2 part to evaluate subcellular compartments. Finally, the targetability score (T score) was calculated as ZTCSx2 + ZINSx2 +
the efficacy of TSN1611 across various tumor types. Alternative dose levels or regimens ZMISx0.5 + ZMBSx0.5. Also, Tumor infiltrating lymphocytes (TIL) were compared between Tumor
could be explored based on the emerging data. Pts received oral TSN1611 twice daily (BID), Proportion Score (TPS)$1 and TPS,1 groups in each target. Results: The IHC analyzer assessed
until disease progression, unacceptable toxicity, or patient withdrawal. Here we report the 528M cells including 147M cancer cells. In 34 cancer types, the average T score for the 74 targets was
preliminary data from phase 1a part. Results: As of Jan 05, 2025, 18 pts received TSN1611 0.62, which was higher than -0.07 observed for the other 699 targets that have never been explored as
from 50 to 600 mg BID (9 with CRC, 5 PDAC, 2 NSCLC, and 1 each with ampullary cancer and drugs. The average T score of the top 10 targets in pan-cancer was 4.27, which was significantly higher
than the average (0.0). Among the top 10 targets in pan-cancer (Table), MUC16 was ranked high in non-
gallbladder cancer, all with pre-identified KRAS G12D mutation). Median age was 61 years
squamous lung, ovary, uterine, cervical cancers; and CEACAM5 and TACSTD2 were ranked high in 7
(range 36-81). The median prior lines of systemic therapy were 3 (range 1-6). No dose- and 10 cancer types, respectively. Most targets showed an association with lower TILs and higher TPS,
limiting toxicity was reported and MTD was not reached. The most common ($ 10%) whereas CEACAM5 demonstrated significantly higher TILs (x1.39) in the TPS$1 group in bladder
treatment related adverse events (TRAEs) were grade 1 or 2 vomiting (44.4%), nausea and cancer. Conclusions: We developed a pipeline leveraging AI-powered and big-data-driven approaches
diarrhea (38.9% each), fatigue (16.7%), ALT increased, blood CPK increased, hyperkalemia to assess the cancer and membrane-specific expression of target proteins in IHC images. The current
and hyperuricemia (11.1% each). No treatment related grade 3 or higher AE or SAE was pipeline reproduces the targetability of developed targets as well as novel targets with a potential
reported. Four (30.8%) out of the 13 evaluable pts demonstrated stable disease per RECIST synergy with immuno-oncology agents. Research Sponsor: None.
v1.1. Tumor reductions were observed in 3 pts (CRC, PDAC, and NSCLC, n = 1 each) at 200 or Top 10 targets and their association with TILs.
400 mg BID, with treatment ongoing. Serial assessment of plasma ctDNA revealed declines TIL fold
in KRAS G12D variation allele frequency at 200 mg BID and above, echoing that the exposure Targets T score Top 5 ranked cancer types change
at 200 mg BID reached that of ED90 in the CRC GP2D model. TSN1611 was rapidly absorbed
MUC16 5.82 LUAD, OV, UCEC, CESC 0.37
with Tmax around 2 hours and half-life around 15 hours. The PK profile indicated a general SEZ6 5.14 CESC, PAAD, Skin, Brain, UCEC 0.33
dose proportionality in exposure across the evaluated dose ranges and low to moderate CLDN4 4.67 PAAD, BLCA, CRAD, PRAD, STAD, UCEC, THCA 0.21
accumulation after multiple BID dosing. Dose escalation is ongoing, and more data will be DLK1 4.49 LN, HCC, LUAD, UCEC, RCC, Brain 0.60
available at the conference presentation. Conclusions: TSN1611 was well tolerated, TM4SF4 4.27 BLCA, LUSC, BRCA, HNSC, PRAD, CESC, THCA, PAAD 0.26
CLDN1 4.11 STAD, HNSC 0.22
demonstrating acceptable PK characteristics as predicted, with preliminary tumor shrinkage CLDN3 3.77 RCC, UCEC 0.08
observed in pts with refractory KRAS G12D mutant tumors. Phase 1b/2 studies are planned CEACAM5 3.76 STAD, CRAD, LUSC, BRCA, HNSC, LUAD, CESC 0.40
to evaluate TSN1611 both as a monotherapy and in combination with standard of care and/ 1.39 (BLCA)
NECTIN4 3.42 HNSC, BLCA, THCA 0.39
or novel agents treating cancer. Clinical trial information: NCT06385925. Research Sponsor: TACSTD2 3.27 BLCA, LUSC, BRCA, HNSC, PRAD, CESC, THCA, PAAD 0.31
Tyligand Pharmaceuticals (Suzhou) Limited.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 209s

3085 Poster Session 3086 Poster Session

Seizure-related homolog 6 (SEZ6) expression and ctDNA methylation pro- EphA2 siRNA in DOPC nanoliposomes (EPHARNA): A phase I clinical trial in
files in patients with high-grade neuroendocrine carcinomas (NECs)/ patients with solid tumors. First Author: Ravali Annam Reddy, The University of
neuroendocrine tumors (NETs) from a phase 1 study of ABBV-706 in ad- Texas MD Anderson Cancer Center, Houston, TX
vanced solid tumors. First Author: Song Wang, AbbVie, Inc., North Chicago, IL Background: EphA2 overexpression is common in human cancers and has an important role in promoting
Background: SEZ6 is a transmembrane protein with overexpression in small cell lung tumor growth and metastasis. Further, EphA2 has kinase-dependent and independent functions, making it
ideal for RNAi-based targeting. EphA2 siRNA incorporated in DOPC nanoliposomes (EPHARNA) was effective
cancer (SCLC) and other neuroendocrine neoplasms (NENs) and minimal expression in in reducing EphA2 protein levels and reducing tumor growth in preclinical studies. This is a first-in-human
normal tissues, making it a promising therapeutic target for these NENs that have a phase I clinical trial of EPHARNA in patients with solid tumors. Methods: Adult patients with advanced solid
significant unmet need for treatments. ABBV-706 is a novel SEZ6-targeting antibody-drug tumors received escalating doses of intravenous EPHARNA twice weekly in 3-week cycles. A total of 8 dose
conjugate with a potent topoisomerase 1 inhibitor payload and is being evaluated in a levels were explored under a BOIN design (Table 1). Study objectives included evaluation of safety, tolerability,
phase 1 study (NCT05599984) in patients (pts) with advanced solid tumors. Preliminary maximal tolerated dose, and efficacy. Adverse events were assessed per NCI CTCAE Version 4.03 and efficacy
per RECIST v1.1. Patients were evaluable for response if they completed at least 2 cycles. Clinical benefit was
data from ABBV-706 monotherapy dose escalation demonstrated a manageable safety defined as objective response or stable disease for 4 or more cycles. Results: A total of 48 patients were
profile and promising efficacy in pts with SCLC and NECs/NETs (JCO 2024;42[suppl 16]: treated. Most common diagnoses were colorectal (29.2%) and ovarian cancer (12.5%). 20.8% of treated
abs 3001). Herein, we describe SEZ6 expression at the protein and mRNA levels in tumor patients were Black and 8.3% were Hispanic. Median age was 60.3 years (range 24.5-78.8). Median number of
tissues of pts with NENs outside of SCLC, as well as detection of high-grade NEN cancer prior therapies was 4 (range 0-12). Among treated patients, 36 (75%) experienced an AE. Most common AEs
signal of origin (CSO) among these pts by investigating ctDNA methylation prior to ABBV- ($20%) were fever (33.3%), infusion-related reaction (25%), and chills (20.8%). 5 (10.4%) treated patients
experienced Grade 3 AEs that were dose-limiting toxicities including chills (2.1%), dyspnea (2.1%), hyper-
706 treatment. Methods: This phase 1, open-label study enrolled pts ($18 yr) with
tension (2.1%), infusion-related reaction (2.1%), and nausea/vomiting (2.1%). No Grade 4 AEs were noted. Of
relapsed/refractory high-grade NECs/NETs (well-differentiated grade 3 NETs and poorly the 25 patients evaluable for response, disease control rate was 44% (95% CI: 24.5-63.5%) with 11 patients
differentiated NECs), atypical lung carcinoid, and medullary thyroid cancer (MTC) in dose- demonstrating stable disease for at least 2 cycles. No patients demonstrated partial or complete response.
escalation and -expansion cohorts. Pts received ABBV-706 monotherapy IV at 1.3–3.5 mg/ Clinical benefit was observed in 4 (16%, 95% CI: 1.6-30.4%) patients who demonstrated stable disease for at
kg Q3W. FFPE tumor tissues of these pts, when available, were subjected to a proprietary least 4 cycles. One patient received 16 cycles with stable disease before withdrawing consent and dis-
IHC assay for SEZ6 and RNAseq analysis. ctDNA samples collected prior to ABBV-706 continuing the trial due to desire for a treatment break. The study was closed to enrollment prior to con-
firmation of the MTD due to unavailability of the drug. Conclusions: EPHARNA demonstrated an acceptable
treatment were subjected to the Cancer Research Solution (RUO; GRAIL, Inc.). ctDNA safety profile with manageable adverse events in patients with advanced solid tumors. Further investigation of
abundance and CSO were assessed by examining cancer-specific methylation patterns of this novel therapeutic approach is warranted to fully elucidate efficacy and optimal dosing strategy. Clinical
ctDNA. Results: As of Aug 27, 2024, in the NEC/NET cohort of 64 pts, median age was trial information: NCT01591356. Research Sponsor: University of Texas MD Anderson Cancer Center; National
63 yr (range 33–86) and the median number of prior therapies was 3 (range 1–8). High Cancer Institute; P30CA016672; Gateway for Cancer Research; G-18-300; U.S. National Institutes of Health;
prevalence of moderate to strong SEZ6 expression (SEZ6 cytomembrane IHC H- 5P50CA098258; T32 Institutional Training Grant; T32CA101642.
score $100) was observed across NEC/NET histologies: 78% of extrapulmonary small Dose levels, patients treated, and DLTs.
cell NEC of diverse anatomic sites (n = 9); 80% of neuroendocrine prostate carcinoma (n = Dose EphA2 siRNA-DOPC Dose Number of Patients Number of Patients Experi-
5); 43% of large cell NEC of diverse anatomic sites (n = 14); 50% of MTC (n = 4); 40% of Level (mg/m2) Treated encing DLTs DLTs
gastroenteropancreatic NENs (n = 10); 50% of atypical lung carcinoid (n = 6). SEZ6 mRNA 1 450 14 2 G3 chills
levels were highly correlative with SEZ6 IHC scores. ctDNA positivity rate was 95% from G3 nausea
baseline plasma samples of the NEC/NET cohort (n = 60); 67% of ctDNA-positive samples G3 vomiting
2 675 6 1 G3 hypertension
(n = 57) were predicted to have high-grade NEN as their primary CSO. Conclusions: Robust 3 1012.5 5 0 N/A
SEZ6 expression was observed with some heterogeneity across histologies of the NEC/NET 4 1518.75 5 0 N/A
monotherapy cohort. High ctDNA detection rate at baseline indicates the feasibility of 5 2278.13 7 0 N/A
monitoring molecular response longitudinally without an invasive procedure and identi- 6 3417.2 2 2 G3 infusion-related
reaction
fying predictive biomarker(s) for ABBV-706. Clinical trial information: NCT05599984. G3 dyspnea
Research Sponsor: AbbVie, Inc.; n/a. 7 3600 3 0 N/A
8 7200 6 0 N/A

3087 Poster Session 3088 Poster Session

A phase 2 study of olaparib in IDH1 and IDH2 mutant advanced chondro- The effect of HIFU treatment on liver metastasis of colorectal cancer in mice
sarcomas and other solid tumors. First Author: Philippos Apolinario Costa, Yale and its impact on immunity. First Author: Shasha Wang, Department of Medical
Cancer Center, New Haven, CT Oncology, Affiliated Hospital of Qingdao University, Qingdao, China
Background: Pre-clinical data have shown that mutations in isocitrate dehydrogenase Background: The liver represents the predominant site for metastasis in colorectal
(IDH) 1 and 2 can lead to a “BRCAness” phenotype by impairing homologous recom- cancer, with over 85% of cases exhibiting the microsatellite stable (MSS) phenotype,
bination (HR) repair. Olaparib, a PARP inhibitor effective in BRCA-mutated cancers such which typically shows limited response to immunotherapy. High-Intensity Focused
as ovarian, prostate, pancreas and breast cancer, may also be effective in IDH1/2 mutant Ultrasound (HIFU) not only facilitates direct destruction of tumor tissues but also has the
solid tumors. IDH1/2 mutations are frequently present in gliomas and chol- potential to remodel the tumor immune microenvironment, thereby enhancing systemic
angiocarcinomas but also in other solid tumors, such as chondrosarcomas, and mel- anti-tumor immunity. This study aimed to explore the effects of HIFU on liver metastasis
anomas. This study evaluated the efficacy of olaparib in treating advanced IDH1/2 in a murine model and its subsequent impact on immune modulation. Methods: A
mutated solid tumors other than cholangiocarcinoma and gliomas. Methods: NCI 10129 BALB/c mice model of colorectal cancer liver metastasis was established and validated.
was a 3-arm, open-label Phase II clinical trial performed in the NCIExperimental The animals were treated with HIFU, followed by transcriptomic profiling and immu-
Therapeutics Clinical Trials Network (ETCTN) evaluating olaparib 300 mg twice daily for nohistochemical staining to assess immune-related markers (CD8, F4/80, FOXP3, PD-L1,
IDH mutated solid tumors refractory to standard treatment. Patients with solid tumors, IL-6) in the liver metastasis tissues. Results: Transcriptomic sequencing performed on
excluding cholangiocarcinoma and glioblastoma, were enrolled in cohort 3 of the Phase liver metastatic tissues collected seven days after HIFU treatment revealed a notable
II trial. The primary endpoint was the overall response rate (ORR), and the secondary upregulation of CXCL14 expression, which was corroborated by protein immunoblotting.
endpoints were progression-free survival (PFS) and overall survival (OS). Results: From Immunohistochemical analysis demonstrated an increased infiltration of cytotoxic
March 2019 until January 2024, a total of 26 patients with IDH1/IDH2-mutant tumors T cells (CD8+), a reduction in macrophage populations (F4/80), and a significant de-
were enrolled in the study across 10 sites. Of these, 14 (53%) had chondrosarcomas, with crease in T regulatory cells (FOXP3 expression). Additionally, both PD-L1 and IL-6 levels
5 (35%) being dedifferentiated. Following, the most prevalent histologies were gas- were substantially reduced in the treated tissues. Conclusions: Seven days post-HIFU
trointestinal adenocarcinomas (4, 15%), other sarcomas (3, 12%) and other tumors (5, treatment, significant immune modulation was observed in liver metastatic tumor
19%). Most tumors had IDH1 mutations (n = 20, 77%), with R132C (n = 12, 60%) being the tissues, including enhanced infiltration of cytotoxic T cells, a reduction in immune-
most common substitution. The median age was 62 years (range 43-78), and 18 (69%) suppressive cell populations such as macrophages and Tregs, and the attenuation of
participants were male. Patients had received a median of 1.5 prior line of therapy (0-9). inflammatory cytokines. These findings suggest that HIFU not only enhances the anti-
After a mean follow-up time of 8.6 months (0.8-62.6), no objective responses were seen, tumor immune response but also facilitates the transition of liver metastases from an
leading to the closure of enrollment. The median PFS was 2 months (95% CI 1.8-2.2), and immune "cold" to a "hot" tumor microenvironment, potentially improving the efficacy of
the median OS was 7.5 months (95% CI 1.3-13). Only two patients had a clinical benefit, subsequent immunotherapeutic strategies. Research Sponsor: None.
defined as PFS . 6 months. Olaparib was tolerable, with most adverse events scored as
grades 1-2. Conclusions: Olaparib did not demonstrate activity in IDH-mutant chon-
drosarcomas and other solid tumors. This study underscores the remarkably poor
outcome associated with IDH mutant tumors, emphasizing the urgent need for addi-
tional therapeutic options. Further evaluation of the correlative data, including as-
sessment of HR proficiency, is required to elucidate why pre-clinical evidence
suggesting potential efficacy did not translate into clinical benefit in IDH mutant solid
tumors. Clinical trial information: NCT03212274. Research Sponsor: National Cancer
Institute; NCI-2017-01182.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
210s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3089 Poster Session 3090 Poster Session

A phase 2 study of the olaparib and AZD6738, an ATM/ATR inhibitor, in Effect of extrachromosomal DNA (ecDNA) on MYCN amplified neuroblas-
isocitrate dehydrogenase (IDH) mutant solid tumors. First Author: toma and patient outcomes. First Author: Mihika Sonalkar, UCSD, La Jolla, CA
Philippos Apolinario Costa, Yale Cancer Center, New Haven, CT Background: Recurrent cytogenetic abnormalities represent candidate therapeutic
Background: Pre-clinical data has shown that mutations in isocitrate dehydrogenase targets for children with neuroblastoma (NB). MYCN oncogene amplification is asso-
(IDH) 1 and 2 can lead to impaired homologous recombination repair. IDH1/2 mutations ciated with significantly worse survival rates for children with NB and remains one of the
are frequently present in gliomas and cholangiocarcinomas but also in other solid primary predictors of patient prognosis. MYCN amplifications in NB can be found both
tumors, such as chondrosarcomas. AZD6738 is an ATR inhibitor, and Olaparib is a PARP within the linear genome and on circular extrachromosomal DNA (ecDNA), and ther-
inhibitor. Preclinical evidence showed a synergistic effect of this combination in models apeutic targeting of the mechanisms underlying MYCN amplification represents a novel
with DNA damage repair effects. This study aims to evaluate the efficacy of Olaparib and and promising strategy in NB. However, the molecular features and clinical and bio-
AZD6738 in treating advanced IDH1/2 mutated solid. Methods: NCI 10222 is an open- logical significance of these amplifications in NB tumors are not sufficiently understood.
label Phase II clinical trial performed in the NCI National Clinical Trials Network Methods: Whole genome and RNA sequencing data were analyzed for NB cell lines and
evaluating olaparib 300 mg twice daily with AZD6738 160 mg daily for IDH mutated solid NCI TARGET NB samples using AmpliconSuite software for ecDNA identification and
tumors refractory to standard treatment. The primary endpoint was the overall response characterization. GISTIC was used for identification of recurrently amplified regions.
rate (ORR), and the secondary endpoints were progression-free survival (PFS) and Gene expression levels were determined using StringTie, and gene clustering heatmaps
overall survival (OS). Results: From January 2020 until March 2023, a total of 24 were generated using FeatureCounts software. For differential gene expression ana-
patients with IDH1/IDH2 mutant tumors were enrolled in the study across 8 sites. Of lyses, samples were divided into ecDNA+ and ecDNA-, and genes contained on ecDNA
these, 14 (58%) had cholangiocarcinoma, 4 (17%) had chondrosarcomas, and 6 (25%) were compared to the same regions on linear DNA across samples using DESeq2.
had other tumors. Most tumors had IDH1 mutations (n = 16, 70%). The median age was Associations between ecDNA quantity, content, and patient survival were performed
59 years (range 29-83), and 15 (63%) participants were male. Patients had received a using multivariate Cox regression survival analysis. Associations of gene expression
median of 3 prior lines of therapy (0-6). After a mean follow-up time of 3 months (0.2- with patient survival were performed using the R2 Platform. The efficacy of targeting
ongoing), no objective responses were seen, leading to the closure of enrollment. The ecDNA-associated gene products was assessed using live cell imaging and cell viability
median PFS was 2 months (95% CI 2-4), and the median OS was 7 months (95% CI 3-NE). assays. Results: WGS analysis confirmed 7/20 NB patient tumors from the TARGET
Only three patients had a clinical benefit, defined as PFS . 6 months, with one patient database to be ecDNA amplified with 1-5 independent ecDNA elements and MYCN gene
diagnosed with G1 chondrosarcoma still on treatment with stable disease. Combination expression correlated with the ecDNA copy number. ecDNAs in MYCN-amplified
of Olaparib with AZD6738 resulted in G3 AE in 9 (38%) patients, leading to 4 (17%) neuroblastoma cell lines contained distinct gene combinations and possessed unique
discontinuations. Conclusions: Olaparib with AZD6738 did not demonstrate activity in structures. MYCN overexpression in NB cells has been shown to be associated with
IDH mutant solid tumors. However, the stability seen in the patient with low-grade replication stress (RS), and tumor cells containing ecDNA are hyper-reliant on the DNA
tumors could suggest that the effect is restricted to lower-grade tumors, still dependent damage response (DDR) kinase CHK1 to manage heightened replication stress. Ex-
on IDH mutations. Further evaluation of the correlative data is required to elucidate why pression of the CHK1 gene was associated with neuroblastoma patient outcomes and
pre-clinical evidence suggesting potential efficacy did not translate into clinical benefit neuroblastoma was most significantly associated with CHK1 RNA dependency. We
in IDH mutant solid tumors. Clinical trial information: NCT03878095. Research Sponsor: further validated CHK1i as a promising therapeutic strategy in MYCN amplified NB, as
None. CHK1 inhibition with the novel inhibitor BBI-2779 was most effective against ecDNA+,
MYCN-amplified neuroblastoma cell lines. Conclusions: Our results emphasize the
critical role of ecDNA in NB. We identify a synthetic lethality axis shaped by ecDNA
MYCN amplification and CHK1 dependence. We further demonstrate the feasibility of
targeting this vulnerability through CHK1 inhibition, thus offering new avenues for
treatment in MYCN amplified tumors. Research Sponsor: Curebound Foundation;
Boundless Bio, Inc.

3091 Poster Session 3092 Poster Session

AI-driven design of novel PARP inhibitors. First Author: Juan Velasco, Yale A first-in-class KIF11 degrader antibody conjugate (DAC) as a potential
University, New Haven, CT therapy targeting a broad spectrum of cancers. First Author: Yan Feng, Accutar
Background: Inhibitors of the Poly (ADP-ribose) polymerase (PARP) family play a role in Biotechnology, Cranbury, NJ
treating HER2-negative locally advanced or metastatic breast cancer with germline Background: Kinesin family member 11 (KIF11) plays a critical role in mitotic spindle
BRCA1/2 (gBRCA) mutations, as well as in the maintenance treatment of gBRCA- formation and centrosome separation during cell division, making it an attractive anti-
associated metastatic pancreatic ductal adenocarcinoma. However, the design of de cancer target. Despite its promise, the clinical success of KIF11 inhibitors has been
novo small molecules targeting proteins like PARP remains time consuming and re- hindered by a narrow therapeutic window, largely due to on-target toxicities, such as
source intensive. It is hypothesized that generative models trained on molecular graph myelosuppression. Antibody-drug conjugates present a promising strategy to overcome
encodings could accelerate the design of novel PARP inhibitors. Objective:This study this limitation by enhancing the therapeutic window of KIF11-targeting therapies.
aims to develop a generative model capable of designing novel, orally bioavailable PARP However, creating a KIF11-targeting inhibitor payload with sub-nanomolar potency
inhibitors. Methods: A large language model was pre-trained on 1 million chemical remains a significant challenge. Using Accutar’s chimeric degrader platform, we de-
structures sourced from the ChEMBL database. Each structure was represented as a veloped dKIF976, a first-in-class KIF11 degrader with sub-nanomolar potency. This
Simplified Molecular Input Line Entry System (SMILES) string, which was tokenized into degrader was further used as payload to create DACs demonstrating potent cell growth
discrete atomic and functional group-level tokens. The model leverages an Average- inhibition coupled with KIF11 degradation, thereby providing a robust foundation for
Stochastic Gradient Descent Weight-Dropped Long Short-Term Memory (AWD-LSTM) further in vivo evaluation of novel KIF11-targeting therapy. Methods: dKIF976, a CRBN-
architecture. Transfer learning was applied to adapt the pre-trained model to specific based KIF11 degrader, was designed via Accutar’s chimeric degrader platform. Western
target chemical structures, enabling domain-specific fine-tuning for the de novo design blot analysis was used to evaluate KIF11 degradation and mitotic arrest, as indicated by
of PARP inhibitors. Results: The model demonstrated robust performance in generating Histone H3 phosphorylation, in cancer cell lines. Cell growth inhibition was assessed
chemically valid, unique, novel, and diverse PARP inhibitors. It achieved a validity rate, using ATP-based assays. Mechanism and selectivity of dKIF976 were confirmed through
uniqueness rate, and novelty rate of 100%, along with a diversity score of 81.53%. specific cellular assays and proteomic analyses. Cell surface antigen-dependent activity
Furthermore, the generated molecules exhibited favorable physicochemical properties, of dKIF976 DACs were evaluated in cell lines with different levels of antigen expression.
including a molecular weight of 417.52 Da, a logarithm of the partition coefficient (LogP) Results: dKIF976 demonstrated rapid, dose-dependent KIF11 degradation and signifi-
of 2.58, a topological polar surface area (TPSA) of 93.21 Angstrom squared, an average cant upregulation of p-Histone H3 across all tested cell lines, achieving sub-nanomolar
of 4.05 rotatable bonds, 1.84 hydrogen bond donors, and 5.14 hydrogen bond acceptors. potency. In side-by-side comparisons, dKIF976 displayed significantly greater cell growth
Conclusions: A generative model was developed to design novel, orally bioavailable inhibition than multiple published KIF11 inhibitors. KIF11 degradation and the resulting
PARP inhibitors. It provides an efficient and automated tool for de novo small molecule cell growth inhibition induced by dKIF976 were confirmed to be dependent on the E3
design with tailored molecular and pharmacological properties, potentially accelerating ligase CRBN and the proteasome. Proteomic analysis via mass spectrometry validated the
the development of PARP inhibitors. Research Sponsor: None. selective degradation of KIF11. When conjugated to antibodies, dKIF976 DACs exhibited
antigen-dependent KIF11 degradation and cell growth inhibition, with enhanced potency
observed in cell lines with high target expression. Conclusions: dKIF976 achieves
specific and potent KIF11 degradation, inducing mitotic arrest and robust cancer cell
growth inhibition. Its superior efficacy and unique mechanism of action establish it as a
highly promising payload for antibody conjugates. dKIF976 DACs demonstrated strong
antigen-dependent KIF11 degradation and cell growth inhibition. This innovation high-
lights the potential of using chimeric degraders as payloads for antibody conjugates,
offering a promising strategy to enhance the therapeutic window of KIF11-targeted
therapies and pave the way for their future success. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 211s

3093 Poster Session 3094 Poster Session

Comparative analysis of fibroblast activation protein Inhibitor (FAPI) O4 Role of Lu177 FAPI-09 therapy in combination with chemotherapy or im-
PET/CT versus flurodeoxyglucose (FDG) PET/CT in staging gastrointestinal munotherapy for chemo-resistant progressive cancers: Early clinical
tumours. First Author: S.P. Somashekhar, Aster International Institute of Oncology, experience. First Author: Prathap H. J, Aster International Institute of Oncology,
Bangalore, India Aster Hospitals Whitefield, Bangalore, India
Background: FAPI 04 PET/CT has emerged as a promising alternative to FDG PET/CT, Background: Chemo resistant progressive tumours represent a major challenge in
offering potential advantages in sensitivity, specificity, and patient comfort due to its oncology. These tumours often adapt and develop resistance mechanisms that limit the
unique targeting of fibroblast activation protein (FAP), which is up regulated in the efficacy of standard treatments. Lu177 FAPI-09 therapy, which targets fibroblast ac-
stromal cells of many cancers. This study aims to evaluate the diagnostic performance tivation protein (FAP) expressed on cancer-associated fibroblasts (CAFs) within the
of FAPI 04 PET/CT compared to FDG PET/CT for staging gastrointestinal (GI) tumours. tumor microenvironment, has emerged as a promising treatment option. By binding to
Methods: 55 patients with suspected or confirmed GI malignancies underwent both FAP, Lu177 FAPI-09 selectively delivers radiation to the tumor stroma, potentially
FDG PET/CT and FAPI 04 PET/CT scans within a 7-day window. A total of 115 lesions altering the tumor environment and making it more susceptible to subsequent therapies.
were identified across the cohort, including colorectal, gastric, pancreatic, and oeso- This study aimed to evaluate the safety, feasibility, and effectiveness of combining
phageal cancers. Tumor localization, lesion size, and uptake characteristics were Lu177 FAPI-09 therapy with chemotherapy or immunotherapy in patients with advanced
compared between the two imaging modalities. Sensitivity, specificity, and diagnostic chemoresistant cancers. Methods: Eighteen patients with advanced progressive and
accuracy were calculated using histopathology as the gold standard. Results: Of the 55 chemo resistant malignancies were included in this study. All underwent Ga-68 FAPI-09
patients, 31 were male and 24 female, with a median age of 58 years (range: 42–78). PET/CT scans prior to treatment to confirm FAP expression in their tumors, ensuring that
FAPI 04 PET/CT demonstrated superior sensitivity (92%) compared to FDG PET/CT Lu177 FAPI-09 therapy would be beneficial. The patient group consisted of GI(7), ovarian
(84%) for the detection of GI tumours, particularly in pancreatic and colorectal cancers, (11). Prior to Lu177 FAPI-09 therapy, all patients had demonstrated resistance to one or
where stromal fibrosis is prevalent and metabolic activity may be low. Specificity of FAPI more prior lines of chemotherapy or immunotherapy. Following the administration of
04 PET/CT was also higher (95%) compared to FDG PET/CT (88%), reflecting the lower Lu177 FAPI-09 therapy, patients received chemotherapy or immunotherapy 5 days later,
background activity in non-tumor tissues. Lesion detection rates were significantly tailored to their specific tumor type. Disease responses were assessed two months after
improved with FAPI 04 PET/CT, with 112 out of 115 lesions identified, while FDG PET/CT the combined treatment. Results: After receiving Lu177 FAPI-09 therapy, 80% of
detected 98 lesions (p = 0.02). The accuracy of FAPI 04 PET/CT was 94%, whereas FDG patients showed either stable disease or a positive response to the subsequent che-
PET/CT had an accuracy of 86%. Notably, FAPI 04 PET/CT showed a higher diagnostic motherapy or immunotherapy. The safety profile of Lu177 FAPI-09 therapy was
yield in detecting metastases in liver, peritoneum, and lymph nodes compared to FDG favourable, with no major grade 3 or 4 adverse events reported. The most common side
PET/CT. The average scan time for FAPI 04 PET/CT was 30 minutes, significantly shorter effects were mild reductions in blood counts, including neutropenia or anaemia. Overall,
than FDG PET/CT (1 hour 15 minutes). Conclusions: FAPI 04 PET/CT outperforms FDG Lu177 FAPI-09 therapy in combination with chemotherapy or immunotherapy was well-
PET/CT in terms of sensitivity, specificity, and diagnostic accuracy for staging gas- tolerated by the majority of patients. Conclusions: Lu177 FAPI-09 therapy, when used in
trointestinal tumours, particularly in cases with abundant stromal involvement. The combination with chemotherapy or immunotherapy, shows promise in enhancing
reduced background activity and higher lesion detection rate enhance the utility of FAPI treatment responses in patients with chemo resistant progressive cancers with minimal
04 PET/CT in clinical practice. Additionally, the walk-in basis for FAPI 04 PET/CT with side effects and no major toxicities. The use of Ga-68 FAPI-09 PET/CT scans to identify
shorter fasting and scan time offers significant improvements in patient comfort and patients with tumors expressing FAP ensures appropriate patient selection, optimizing
convenience. Given these advantages, FAPI 04 PET/CT represents a promising alter- the likelihood of a positive outcome. These findings suggest that Lu177 FAPI-09 therapy
native to FDG PET/CT for staging GI cancers, with potential implications for improved could be an effective approach to overcome resistance and improve treatment out-
treatment planning and monitoring. Research Sponsor: None. comes in patients with advanced malignancies. Research Sponsor: None.

3095 Poster Session 3096 Poster Session

Dosing tolerability and adverse events (AEs) in dihydropyrimidine dehydro- Monocentric pilot trial of trametinib in severe extracranial arteriovenous
genase (DPYD) variant carriers receiving genotyping-guided fluoropyrimi- malformations. First Author: Emmanuel Seront, Institut Roi Albert II, Department of
dine (FP) dosing. First Author: Grace Nguyen, Atrium Health Levine Cancer Institute, Medical Oncology, Saint Luc University Hospital, Brussels, Belgium
Charlotte, NC Background: The Mitogen Activated Protein Kinase (MAPK) pathway is crucial for cell
Background: We previously showed that DPYD genotype-guided FP (5-FU, capecitabine) dosing growth, proliferation, and survival. Overactivation of MAPK is observed in many cancers
reduces severe AEs and hospitalizations in variant carriers. However, optimal dose reduction and leading to evaluation of targeted therapies such as MEK inhibitors. Vascular malfor-
tolerability for individual DPYD variants are not well understood. This study aims to evaluate dosing mations, including arteriovenous malformation (AVM) share many oncogenic mutations
tolerability and AEs among DPYD variant carriers. Methods: This is a retrospective cohort study of
with cancer. For example, AVM present KRAS, RASA1, MAP2K1 mutation that result in
patients (pts) receiving FP-based chemotherapy at a multisite cancer center who underwent routine
in-house DPYD genotyping covering 5 variants (Table). Test results and dose recommendations
excessive activity of the RAS-RAF-MEK cascade. This trial aimed to assess trametinib
were provided to oncologists per Clinical Pharmacogenetics Implementation Consortium guidelines safety and efficacy in adult patients with stage III AVM refractory to conventional
(i.e., 50% dose reduction in DPYD heterozygous carriers and slow titrations in subsequent cycles therapy, causing deformities, pain, bleeding, or ulceration. This is the first trial to
based on AEs). Clinicodemographics were collected via chart review, and AEs were graded using evaluate targeted therapies in AVM. Methods: We conducted a prospective Phase II trial
Common Terminology Criteria for Adverse Events criteria version 5.0. Data was collected for on ten adult AVM patients. Trametinib was administered orally for 12 months, with initial
3 months of FP treatment unless discontinued early. Results: From March 2020-October 2024, dosage escalation based on patient tolerance. Clinical and radiological outcomes were
1,645 pts were genotyped with 85 (5.2%) identified as heterozygous DPYD variant carriers. This assessed at baseline, during treatment, and at follow-up. Primary outcomes included
analysis included 49 carriers who were tested pretreatment and started FP chemotherapy (median safety and clinical efficacy (pain reduction, ulceration healing, thrill and deformation
age 65, 35% male, 67% White, 27% Black, 71% gastrointestinal cancers, 53% 5-FU, 47% capeci-
improvement). Secondary outcomes included radiological responses assessed via MRI,
tabine). All pts started on dose-reduced FP at cycle 1 (Table). Of 49 carriers, 18 (37%) had at least
one dose escalation, most occurring in cycles 2 or 3. Of these, 5 were eventually escalated to full
Doppler ultrasound, and angiography. Results: Of the ten patients (6 female, 4 male),
dose, 5 had AEs preventing further escalation, 5 had subsequent dose reduction due to AEs, 2 eight had facial AVM, one auricular, and one foot AVM. All experienced deformities, with
completed therapy while escalating, and 1 had one dose escalation with no documented AE. Dose seven reporting severe pain, five ulceration, and two bleeding. Trametinib was initiated at
escalation was not performed in 31 (63%) pts due to the following reasons: any-grade AE (n = 27, of 2mg daily for three patients but, due to skin toxicities, subsequent patients started at
which 5 had further dose reductions due to AEs), poor performance status (n = 1), early dis- lower doses, with only three reaching the target dose. Trametinib led to clinical im-
continuation (n = 2, 1 disease progression, 1 declined further therapy), treatment completion (n = 1). provement in 80% of patients. Pain alleviation occurred in all symptomatic patients (VAS
Conclusions: This is the largest retrospective cohort study evaluating dosing tolerability and 5–7 to 0–5), deformation improved in 55%, and ulceration healed in 20%. Radiological
toxicity in DPYD carriers using real world data. There is interindividual variability in tolerability within assessment showed a reduction in vessel size in one patient and nidus disappearance in
each DPYD variant, particularly with decreased function variants. Findings will help inform pro-
two. Acneiform rash was the most frequent toxicity (100%), including two cases of grade
spective studies and clinical guidelines on variant-specific dosing strategies. Research Sponsor:
None.
3, requiring early drug discontinuation. Severe mucosal bleeding led to premature
cessation in two patients with mucosal AVMs. Correlations with genomic-alteration will
Median (range) Median (range) be presented at congress. Conclusions: Trametinib demonstrated clinical benefit in
first dose final dose Grade AE related AE related refractory AVMs, supporting MAPK inhibition as a therapeutic approach. Skin and
Variant N, % intensity, % intensity, % 3+ AE hospitalization discontinuation
mucosal toxicities necessitate dose adjustment, dermatological co-management, and
All 49 50 (40-81) 50 (27-100) 13 (27%) 9 (18%) 11 (22%) cautious use in mucosal AVMs. Further studies are warranted to optimize therapeutic
c.1236G>Aa 27 (55%) 50 (40-75) 54 (33-100) 6 3 6
c.557A>Ga 12 (25%) 50 (40-81) 60 (40-100) 2 1 3 regimens and assess long-term outcomes. Clinical trial information: 2019-003573-26.
c.2846A>Ta 6 (12%) 50 (47-54) 50 (47-54) 2 3 2 Research Sponsor: None.
c.1905+1G>Ab 3 (6%) 51 (45-54) 38 (27-45) 2 1 0
b
c.1679T>G 1 (2%) 50 45 1 1 0
a
Decreased function variant.
b
No function variant.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
212s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3097 Poster Session 3098 Poster Session

Cost analysis of pre-treatment dihydropyrimidine dehydrogenase (DPYD) Predicting immunotherapy response in advanced solid tumors using quan-
genotyping to reduce hospitalizations at a cancer center in the United States titative imaging features from CD8 PET/CT exams. First Author: Michael A.
(U.S.). First Author: Sarah Morris, Atrium Health Levine Cancer Institute, Charlotte, NC Postow, Memorial Sloan Kettering Cancer Center, New York, NY; Weill Cornell Medical
Background: Patients with certain DPYD variants are at increased risk of fluoropyrimidine (FP) College, New York, NY
related adverse events (AEs) and mortality at standard doses. We previously showed pre-treatment Background: CD8-PET/CT imaging with 89Zr crefmirlimab berdoxam (ImaginAb, Inc),
DPYD testing and genotype-guided FP dosing reduced severe AEs and hospitalizations in variant which targets CD8-expressing T-lymphocytes, is being explored as an imaging tool to
carriers (PMID 38935897), but testing cost remains a barrier to widespread adoption in the U.S. predict responses and monitor immune checkpoint inhibitors (ICI) in patients with ad-
Herein, we performed a cost analysis of pre-treatment DPYD genotyping. Methods: Variant carrier
vanced solid malignancies. Here we explore how quantitative imaging features from CD8
rates, hospitalization rates, and AEs were derived retrospectively from our institutional cohort
(n=442) of patients with no observed variant, dose reduced variant carriers, and standard dose
PET/CT may predict ICI responses. Methods: We studied quantitative imaging features
variant carriers (identified reactively) receiving FP primarily for gastrointestinal cancers. All (PET parameters and radiomics) in 45 patients from the ImaginAb IAB-CD8-201 phase II
patients were genotyped and followed for three months for FP-related AEs and hospitalizations. trial (NCT03802123). Tumoral lesions, peritumoral ring (ring shaped margin extending
Hospitalization cost was the weighted average cost of treating the most expensive AE experienced 0.5 cm inwards and outwards from the segmented tumor surface), healthy tissue, benign
by the hospitalized patient. Input parameters (Table 1) were modeled using a decision tree to and pathological lymph nodes were segmented from baseline and first on-treatment (4-
compare the cost of pre-treatment testing (no variant and dose reduced variant carriers) to no pre- 6 weeks after standard of care treatment including ICI blockade). Imaging features from
treatment testing (no variant and standard dose variant carriers) from a health-system perspective CD8-PET/CT scans were extracted. Predictive models for best overall response (BOR)
with a three-month time horizon. The model accounted for hospitalization and genotype test costs according to RECIST 1.1 were developed. Models’ performance was evaluated by the
only. Results: Pre-treatment testing resulted in a cost savings of $36.98 per patient compared to ability to distinguish responders (complete or partial response, n = 13) from non-
no pre-treatment testing (average per patient cost = $1,655.81 and $1,692.79, respectively). Cost responders (stable or progressive disease, n = 32). A survival random forest analysis
savings increase to $124.39 per patient if half of those tested have insurance that reimburses the
was also conducted to estimate time to BOR. Results: Significantly greater delta values
test cost. Additional savings are expected if costs for outpatient management of AEs and use of
uridine triacetate in the inpatient setting are included in the model. Conclusions: Pre-treatment
were identified in the tumor and peritumoral ring compared to healthy tissues, suggesting
DPYD genotyping led to cost savings by reducing AE related hospitalizations among variant the tumor and peritumoral ring may reveal early treatment-induced changes important for
carriers. Cancer centers should adopt pre-treatment DPYD genotyping to reduce severe AEs, predicting response. Eighteen predictive models were developed, with models using
hospitalizations, and costs. Research Sponsor: None. imaging features from the peritumoral ring showing comparable performance to those
using features from lesions and pathological lymph nodes. The simplest BOR predictive
Model inputs.
model, which yielded the highest performance, used delta values extracted from the
Parameter Value Source
peritumoral ring (AUCs = 0.895, sensitivity = 0.900, specificity = 0.615). The inclusion of
No variant population prevalence 94% Institutional cohort clinical variables (including age, sex, body mass index, cancer type, received treatments,
Variant carrier population prevalence 6% Institutional cohort
No variant hospitalization rate 11% Institutional cohort number of lines of received treatment, white blood cell count) did not significantly
Dose reduced variant carrier hospitalization rate 25% Institutional cohort enhance model accuracy, emphasizing the robustness of the imaging data alone. A final
Standard dose variant carrier hospitalization rate 64% Institutional cohort model integrating key imaging features from multiple regions successfully predicted time
No variant hospitalization cost $12930.67 HCUP NC Inpatient to BOR with a C-index, a generalizable AUC that considers censored data, of 0.86.
Database 2021
Dose reduced variant carrier hospitalization cost $8858 HCUP NC Inpatient Conclusions: This study highlights the potential of quantitative imaging analysis of CD8-
Database 2021 PET/CT scans as a tool for predicting responses to cancer immunotherapy. Results
Standard dose variant carrier hospitalization cost $8928.29 HCUP NC Inpatient suggest the effectiveness of delta imaging features from the peritumoral ring as a
Database 2021
Genotype test cost $174.81 CMS clinical laboratory potential indicator of patient’s ability to respond to ICI treatment, simplifying the analysis
fee schedule 2023 without sacrificing accuracy. Further validation in trials with more homogeneous pop-
HCUP; Healthcare Cost and Utilization Project, NC, North Carolina; CMS, Centers for Medicare and Medicaid
ulations and treatment regimens (eg. NCT05013099) is warranted, with the potential to
Services. advance personalized cancer care. Research Sponsor: ImagingAb.

3099 Poster Session 3100 Poster Session

212
Interim safety and efficacy data of [ Pb]VMT01 in MC1R expressing SPYK04, a novel RAF-MEK molecular glue: Dose escalation (DE) in first-in-
melanoma. First Author: Zachary Scott Morris, University of Wisconsin School of human study for MAPK pathway-altered solid tumors. First Author: Sarina A.
Medicine and Public Health, Madison, WI Piha-Paul, Department of Investigational Cancer Therapeutics, The University of Texas
Background: Immune checkpoint inhibitors (ICI) are effective in melanoma, but many MD Anderson Cancer Center, Houston, TX
patients experience progression on or after approved ICI +/- MAPK inhibitor therapy. Background: SPYK04 is a novel MEK1/2 inhibitor designed to enhance RAF-MEK binding and
Melanocortin-1 receptor (MC1R) is a novel target for radiopharmaceutical therapy (RPT) potentially inhibit feedback activation of MEK1/2. This approach was developed to address the
and is highly expressed on melanoma tumor cells. VMT01 is an MC1R-targeted RPT that limited efficacy of conventional MEK inhibitors in RAS-mutated cancers, which is thought to be
can be radiolabeled with either 203Pb (patient selection and dosimetry assessments) or due to feedback activation of the MAPK pathway. Methods: This first-in human study of SPYK04
212
Pb (alpha particle therapy). Here, we present data on the first in-human evaluation of is conducted in the US and Japan (NCT04511845). Patients with locally advanced or metastatic
[203Pb/212Pb]VMT01 in patients with metastatic melanoma. FDA granted Fast Track solid tumors harboring MAPK pathway alterations were eligible in its DE part. The primary
Designation to the product on the bases of preclinical experiments combining [212Pb] endpoints included assessment of pharmacokinetics, adverse events (AEs) and dose-limiting
VMT01 with immunotherapy. Methods: This is a first-in-human dose-finding study to toxicities (DLTs); the secondary endpoint was objective response rate (ORR). SPYK04 was
determine the safety, pharmacokinetics, dosimetry and preliminary efficacy of [212Pb] administered orally once daily in continuous 28-day cycles. Results: A total of 23 patients were
enrolled in DE. SPYK04 was evaluated at doses ranging from 0.1 to 1.3 mg/day. An accelerated
VMT01 in subjects with MC1R-positive metastatic melanoma who progressed on at least 1
titration design was employed for doses of 0.1, 0.2, and 0.4 mg/day, followed by a transition to a
approved first-line therapy (NCT05655312). Phase 1 of the trial includes escalating dose
3+3 design starting from 0.8 mg/day. Over the dose range of 0.1 to 1.3 mg/day, systemic
cohorts. The first two cohorts incorporate dosimetry evaluations (reported separately) with exposure demonstrated a dose-dependent increase. Grade 3 or higher treatment-related adverse
the imaging surrogate [203Pb]VMT01 prior to receiving up to 3 treatment cycles of [212Pb] events (TRAEs) were observed in 30.4% of patients. TRAEs observed in .= 20% of patients were:
VMT01 therapy (injected activity of 111 MBq (3 mCi) or 185 MBq (5 mCi) for Cohort 1 and 2, dermatitis acneiform and blood creatinine phosphokinase increased (60.9% each); AST increased
respectively). Participants are evaluated for any DLT for the first 6 weeks after cycle one. (30.4%); nausea and stomatitis (26.1%, each). DLTs were observed in one patient each at dose
Efficacy is assessed by RECIST 1.1 criteria by the investigator. Following the start of the levels of 0.8, 1.0 and 1.3 mg/day. All DLTs were due to receiving ,75% of the planned dose. A
study the anticipated combination arm with nivolumab was opened as an amendment. decision was made to not escalate beyond 1.3 mg/day. The ORR was 8.7% (n = 2 of 23), with both
Results: Cohort 1 (DCO 04Sep24) was completed with 3 enrolled participants who received partial responses (PRs) occurring amongst the 5 enrolled ovarian cancer patients. The disease
3 treatment cycles without any DLTs or SAEs. Cohort 1 participants showed prolonged control rate (DCR; PR+SD) was 52.2% (n = 12 of 23). Conclusions: SPYK04 was tolerated at
stabilization of disease from start of treatment (mean: 11.1 months); one participant doses up to 1.3 mg/day in patients with advanced solid tumors. PRs were observed in two
developed a confirmed objective response (PR) after completion of all three [212Pb]VMT01 patients with ovarian cancer. Further evaluation of safety and efficacy will occur in the expansion
administrations and is still on trial after 13.1 months from start of treatment. Cohort 2 has part of this study. Clinical trial information: NCT04511845. Research Sponsor: Chugai Phar-
completed with 7 enrolled participants. No DLTs or related SAEs have been observed. All maceutical Co., Ltd. (Contact Person: Satoe Kawakami, email: kawakamiste@chugai-
participants in this Cohort progressed after either the first cycle (3 participants) or the [Link]).
second cycle (4 participants).Based on these preliminary results showing anti-tumor effect SPYK04 dose, mg/day 0.1 0.2 0.4 0.8 1 1.3
at the lower dose, additional cohorts for both monotherapy and in combination with
Patients, n 1 1 1 8 6 6
nivolumab were introduced at a de-escalated dose of 55.5 MBq (1.5 mCi). Both cohorts are Safety, n (%)
now open for enrollment. Updated safety and efficacy data will be analyzed and presented TRAE 0 1 (100.0) 1 (100.0) 8 (100.0) 6 (100.0) 6 (100.0)
at ASCO. Conclusions: At 111 MBq and 185 MBq activity levels, [212Pb]VMT01 was safe and Grade 3 or higher TRAE 0 0 0 1 (12.5) 4 (66.7) 2 (33.3)
well-tolerated. An objective response and prolonged stabilization of disease were observed TRAE leading to discontinuation of 0 0 0 0 1 (16.7) 0
study treatment
at the 111 MBq activity level while no effect was seen at 185 MBq. The study will continue to Confirmed best response, n (%)
explore potentially immunostimulating lower doses of administered activity of [212Pb] Responders 0 0 0 1 (12.5) 1 (16.7) 0
VMT01 either as a monotherapy or in combination with nivolumab. Clinical trial information: Partial Response 0 0 0 1 (12.5) 1 (16.7) 0
NCT05655312. Research Sponsor: Perspective Therapeutics. Stable Disease 0 0 1 (100.0) 2 (25.0) 3 (50.0) 4 (66.7)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 213s

3101 Poster Session 3102 Poster Session

Zanzalintinib (zanza) + nivolumab (nivo) 6 relatlimab (rela) in patients (pts) Phase II dose optimization update with EZH2/EZH1 inhibitor tulmimetostat
with advanced solid tumors: Results from two dose-escalation cohorts of in patients with ARID1A-mutated ovarian clear cell carcinoma or endometrial
the phase 1b STELLAR 002 study. First Author: Benjamin Garmezy, Sarah Cannon carcinoma. First Author: Linda R. Duska, University of Virginia Health System,
Research Institute, Nashville, TN Charlottesville, VA
Background: Zanza (XL092) is a novel, multi-targeted tyrosine kinase inhibitor (TKI) that Background: EZH2 inhibition antitumor activity occurs through various mechanistic pathways in multiple
inhibits VEGFR, MET, and TAM kinases. In tumor models, zanza showed antitumor activity tumor types, including via synthetic lethality in advanced ARID1A-mutated ovarian clear cell carcinoma
alone and in combination with an anti-PD-1 immune checkpoint inhibitor (ICI). The addition (OCCC) and endometrial carcinoma (EC). Oral, next-generation, dual EZH2/EZH1 inhibitor tulmimetostat is
in Phase II evaluation in multiple disease cohorts (NCT04104776; Oaknin et al. ASCO 2024, ESMO 2024). We
of a VEGFR-TKI to ICI combinations of anti-PD-1 + anti-LAG-3 may enhance clinical activity. report updated efficacy and safety data from the ARID1A-mutated OCCC/EC cohorts, including dose
STELLAR-002 (NCT05176483) is a phase 1b, open-label, dose escalation and expansion optimization and expansion arms. Methods: Phase II Stage 1 evaluated tulmimetostat 350 mg once daily
study evaluating the safety and efficacy of zanza as monotherapy and in combination with (QD). Stage 2 dose-optimization design randomizes further patients with OCCC (M2) or EC (M3) to 200 mg or
ICIs in pts with advanced solid tumors. Data from dose escalation cohorts treated with zanza 300 mg tulmimetostat QD in Stage 2a, with an efficacy gateway for each arm to open Stage 2b. Primary
+ nivo and zanza + nivo/rela are presented. Methods: Adults with advanced/metastatic endpoint is objective response rate (complete response [CR] + partial response [PR]), and secondary
solid tumors were enrolled. Starting doses were zanza 100 mg po qd + nivo 360 mg IV q3w objectives include safety. Results: As of October 15, 2024, enrollment into the M2/M3 200 mg, 300 mg, and
(zanza + nivo cohort), and zanza 60 mg po qd + nivo/rela 480/480 mg IV q4w (zanza + nivo/ 350 mg arms included 20/10, 21/21 and 14/11 patients, respectively. A total of 56.4% M2 and 61.9% M3
patients received $3 prior lines of therapy. Most responses were seen in the M2 200 mg arm and in the M3
rela cohort). Pts were enrolled in a rolling six design. Sparse pharmacokinetic (PK) samples 350 mg arm (n=4 each; Table). The safety profile across arms was consistent with the EZH1/2 drug class. In
were collected. The primary endpoint was safety. Exploratory endpoints included M2/M3 cohorts, treatment-emergent adverse events (TEAEs) leading to dose modifications were reported
investigator-assessed ORR per RECIST 1.1 and PK. Results: Among 19 pts in the zanza in 55.0%/60.0%, 71.4%/85.7%, and 92.9%/90.9% of patients at 200 mg, 300 mg, and 350 mg, respectively.
100 mg + nivo cohort, the most common tumor types were prostate cancer (26%) and TEAEs leading to treatment discontinuation were reported in 5.0%/20.0%, 4.8%/4.8%, and 14.3%/9.1%,
colorectal cancer (26%). Median number of prior therapies was 6 (range: 2–16). No dose- respectively. Serious TEAEs considered at least possibly related (TRAEs) to tulmimetostat treatment were
limiting toxicities (DLTs) were observed in the first 11 DLT-evaluable pts; zanza 100 mg is reported in 5.0%/0%, 9.5%/14.3%, and 21.4%/27.3%, respectively. Grade $3 TRAEs were mainly hema-
the recommended dose (RD). The most common treatment-emergent adverse events tologic (Table); no TRAEs leading to death were reported. Conclusions: Tulmimetostat showed an im-
proved and acceptable safety profile in OCCC and EC at 200 mg and 300 mg doses (versus 350 mg) with
(TEAEs) were fatigue (68%) and diarrhea (58%). The most common grade (G) 3/4 TEAEs were promising antitumor activity, supporting further clinical investigation. Clinical trial information:
fatigue (26%) and hypertension (16%). Palmar-plantar erythrodysesthesia (PPE) occurred in NCT04104776. Research Sponsor: MorphoSys GmbH.
11% of pts (all G1/2). No responses were observed; the disease control rate (DCR:
Best confirmed responses and most common grade ‡3 related TEAEs.
CR+PR+SD) was 42%. In the zanza + nivo/rela cohort, 24 pts received zanza 60 mg + nivo/
Cohort M2: OCCC M3: EC
rela and 25 pts received zanza 100 mg + nivo/rela. One DLT was observed in the first 6 DLT-
evaluable pts at zanza 60 mg (G3 ALT increase) and none in the first 5 DLT-evaluable pts at Dose, mg 200 300 350 200 300 350
Efficacy evaluable*, N 20 20 14 10 15 11
zanza 100 mg. Zanza 100 mg is the RD with nivo/rela. In zanza 100 mg-treated pts, the most Best confirmed response†, n CR 0 0 0 0 1 0
common tumor type was renal cell carcinoma (RCC; 56%). Median number of prior therapies PR 4 2 1 0 1 4
was 4 (range: 0–15). The most common TEAEs were diarrhea (68%) and fatigue (56%). The Stable disease 10 10 7 8 5 2
Progressive disease 5 7 6 1 7 4
most common G3/4 TEAE was fatigue (20%). PPE occurred in 28% (4% G3/4). ORR was 28% No post-baseline response assessment 1 1 0 1 1 1
and DCR was 80%; these rates were 36% and 86% in pts with RCC (n = 14). Plasma zanza Safety evaluable, N 20 21 14 10 21 11
concentrations 2 hrs after the first dose (Cmax, mean 6 SD) were 595 6 353 and 838 6 689 Grade ‡3 related TEAEs‡, n (%) Thrombocytopenia 1 (5) 3 (14) 4 (29) 0 4 (19) 2 (18)
Anemia 3 (15) 0 7 (50) 0 5 (24) 1 (9)
ng/mL for the 60-mg and 100-mg dose levels, respectively. Conclusions: The tolerability of Neutropenia 0 0 2 (14) 0 0 4 (36)
zanza + nivo and zanza + nivo/rela was manageable and consistent with each monotherapy Diarrhea 0 4 (19) 0 0 0 2 (18)
agent. Preliminary safety, PK, and response data support selection of the 100-mg zanza Data cut off: October 15, 2024.
dose in combination with nivo or nivo/rela for further investigation. Expansion cohorts are *Patients who received $1 dose, had $1 post-baseline response assessment, or discontinued treatment prior to first
ongoing in various tumor types. Clinical trial information: NCT05176483. Research Sponsor: post-baseline assessment for any reason.
†
RECIST 1.1.
Exelixis, Inc. ‡
.10% in any M2/M3 arm.

3103 Poster Session 3104 Poster Session

OMX-0407: A novel spectrum-selective small molecule kinase inhibitor in AI molecular search engine paired with RNA sequencing analysis to develop
advanced/metastatic solid tumors. First Author: Valentina Boni, NEXT Madrid potent and selective VEGFR-3 inhibitors. First Author: Maxim Sorokin, OmicsWay
Universitary Hospital Quironsalud Madrid, Pozuelo De Alarcon, Spain Corp., San Mateo, CA
Background: OMX-0407 is an orally available spectrum-selective kinase inhibitor that Background: Tumor angiogenesis (TA) is driven by several VEGF factors and corre-
targets key oncology-relevant tyrosine kinases and salt-inducible kinases and is being sponding receptors (VEGFRs). Targeting TA inhibits tumor growth, however there are no
developed as a first-in-class treatment for solid tumor indications. Preclinical inves- selective small molecule TA inhibitors on the market yet. Here we report development
tigations indicate a dual mode of action by sensitizing tumor cells to immune cell of a selective VEGFR-3 inhibitor with potential anti-tumor activity in multiple cancers.
induced apoptotic cell death as well as direct inhibition of tumor growth promoting Methods: The Bioptic virtual screening pipeline employs two models. The first is a
kinases. Methods: This is a phase Ia/Ib dose escalation and expansion study of OMX- SMILES-based LLM fine-tuned on binding affinity data of contrastive molecular pairs.
0407 (NCT05826600). Eligible patients for the phase Ia dose escalation part had ad- The screening was performed on the ultra-large 40-billion-compound virtual library
vanced solid tumors and exhausted available therapies. Results: As of the 30th of Enamine REAL. For selectivity, the top-ranked molecules were re-scored using a
October 2024, 24 patients have been treated at dose levels 10 through 140 mg p. o. BID secondary model, a GNN specifically designed for this task and trained to differentiate
and the phase Ia part has been completed. Solid tumor histologies included melanoma, activity across similar kinases. The Oncobox algorithm was used to select cancer types
non-small cell lung cancer, sarcoma and colorectal cancer. OMX-0407 was generally well with the highest sensitivity to VEGFR1-3 inhibitors based on VEGF(R)s expression and
tolerated, adverse reactions were mainly gastrointestinal. Two dose limiting toxicities TA pathways activation. It was applied to RNA-seq profiles from TCGA (11428 profiles
were observed: One case of facial swelling secondary to drug allergy at the 90 mg BID from 33 primary sites) and the internal relevant RWD cohort (1056 profiles from 89
dose, and one case of fatigue at the 140 mg BID dose. One durable complete response cancer types). Results: Compounds with IC50 , 10 mM in Eurofins VEGFRs Kin-
in a patient with cutaneous angiosarcoma secondary to radiotherapy, resistant to two aseProfiler were considered active. Among 110 tested compounds, 1 was active against
prior lines of chemotherapy treatment, was observed at 30 mg BID which is ongoing at VEGFR-1, 1 - against VEGFR-2, and 4 - against VEGFR-3. One compound was active
16 months (at the time of data cutoff). Pharmacodynamic analyses demonstrated against all VEGFRs, and 3 - against a single VEGFR. One VEGFR-3 active showed . 45-
phosphorylation inhibition of target kinases. A recommended phase II dose of 100 mg fold selectivity against both VEGFR1 & 2, while no compound was specific to VEGFR-1 or
BID was identified. Conclusions: OMX-0407 has been well tolerated at pharmacolog- VEGFR-2. All 3 selective VEGFR-3 inhibitors showed minimal activities ( , 50% at
ically and therapeutically active dose levels. One very durable response has been 10 mM) against the 12 off-target kinases including B-Raf, c-Raf, c-Kit, FGFR1, FGFR2,
observed in an angiosarcoma patient at a low dose level. The phase Ib expansion part at FGFR3, FGFR4, Flt3, Met, PDGFR-a, PDGFR-b, Ret. In order to select cancer types for
the recommended phase II dose is currently recruiting patients with angiosarcoma and further pre-clinical validation, Oncobox algorithm was used to simulate predicted ef-
clear cell renal cell carcinoma. Clinical trial information: NCT05826600. Research ficiency of a VEGFR-3 inhibitor in multiple cancer types from TCGA and internal RWD
Sponsor: iOmx Therapeutics AG. cohort. The highest response rate is expected for papillary thyroid cancer, followed by
clear-cell renal, pancreatic, ovarian cancers, and sarcomas. A selective VEGFR-3 in-
hibitor may be beneficial when compared to already developed pan-VEGFR inhibitors due
to lower toxicity. Finally, identification of potential responders to selective anti-VEGFR-3
therapy via RNA-seq analysis may enable patient enrichment in further clinical trials and
development of a companion diagnostic for the drug. Conclusions: Our results suggest
that Bioptic’s molecular search engine significantly enhances identification of potent
and selective inhibitors for a specific target, and, paired with the Oncobox algorithm, may
facilitate development of novel anti-cancer drugs. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
214s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3105 Poster Session 3106 Poster Session

Efficacy and safety of alectinib in pediatric and adult patients with ALK Phase 1 study of TT-00973-MS, a highly selective and potent AXL inhibitor,
altered advanced solid tumors: Results from the TACKLE phase II trial, a in patients with advanced solid tumors. First Author: Bo Yang, The First Affiliated
MASTER KEY substudy (NCCH1712/MK003). First Author: Hitomi Sumiyoshi Hospital of Bengbu Medical University, Bengbu, China
Okuma, Department of International Clinical Development, National Cancer Center Background: AXL is a member of the TAM family activated by the high-affinity ligand
Hospital, Tokyo, Japan Gas6. The Gas6/AXL signaling pathway plays a critical role in drug resistance, tumor
Background: Alectinib is an orally administered tyrosine kinase inhibitor that targets an- proliferation, metastasis, invasion, epithelial-mesenchymal transition and immune
aplastic lymphoma kinase (ALK) and is approved in ALK-positive non-small cell lung cancers regulation, implicating AXL as an important target in cancer treatment. TT-00973-MS
and anaplastic large cell lyphomas in Japan. Beyond these, there is a rare population carrying is a highly selective and potent AXL inhibitor which exhibited significant anti-tumor
the same ALK alteration regardless of cancer type, often referred to as “ALKomas”. Treatment activities in both SK-OV-3 and H1299 derived CDX model with AXL over-expression. Here
options are limited for these ALK altered solid tumors. Methods: This open-label phase II is first time to present the first-in-human study of TT-00973-MS. Methods: Dose es-
study evaluated alectinib for ALK altered locally advanced or metastatic cancer. Patients calation is performed using“3+3”design. Adverse events (AE) are evaluated per CTCAE
received alectinib as either a capsule or a suspension. Primary endpoint was central-assessed v5.0 criteria. Tumor responses are evaluated per RECIST 1.1. Pts receive TT-00973-MS
confirmed objective response rate (ORR) according to RECIST v1.1. A Bayesian approach was once daily continuously for 28-day cycles. The primary endpoint is to evaluate dose
used to evaluate eligible patients in the main cohort, which were patients able to ingest
limiting toxicity (DLT) and identify the maximum tolerated dose (MTD) and/or rec-
capsules. Expected ORR was set at 40% and a threshold at 10%. Secondary endpoints
ommended phase II dose (RP2D). Results: As of the data cut-off date on December 24,
included safety, disease control rate (DCR), progression-free survival (PFS), and overall
survival (OS). Results: 26 patients, aged from 8 months to 78 years, across 11 tumor types
2024, 18 pts have received TT-00973-MS treatment in 2 mg (n = 1), 5 mg (n = 3), 10 mg (n
received treatment. The most common tumor type was soft tissue sarcoma (STS) (n = 11) = 3), 17 mg (n = 7), 25 mg (n = 4) at QD dose levels. Median age was 56.5 (37~69), 8
followed by embryonal neoplasm (n = 5). Alterations included fusion/rearrangements (n = 19), (44.4%) were males, all had ECOG PS #1. 66.6% of pts had Stage III/IV disease. 50%
mutations (n = 5), and amplifications (n = 2). The median follow-up was 15.0 months. In the had $3 prior lines of systemic therapies. 66.7% had prior immunotherapies. One DLT
evaluable patients of the main cohort (n = 16), the central-assessed ORR was 43.8% (95% CI, was observed in a subject at 17 mg QD dose level (Grade 3 peripheral motor nerve
19.8 to 70.1), meeting the decision criteria of the Bayesian design for the primary endpoint. disorder). The MTD was not reached. Treatment-related AEs (TRAEs) were reported in all
For the entire evaluable patients (n = 24), the central-assessed ORR was 54.2% (32.8 to 74.4), pts, grade 3 in 6 (33.3%), and no grade 4 or 5. The most common TRAE ($30%) included
the DCR was 70.8% (48.9 to 87.4); the median PFS was 24.9 months (3.9 to not estimable); increased blood lactate dehydrogenase (83.3%), increased aspartate aminotransferase
and the median OS was 38.8 months (13.9 to not estimable). In patients with ALK fusions/ (72.2%), increased alanine aminotransferase (66.7%), hypercholesterolaemia (55.6%),
rearrangements (n = 17), the ORR was 76.5 % (50.1 to 93.2), DCR 82.4% (56.6 to 96.2) and the hypoalbuminaemia (38.9%), hypertriglyceridemia (33.3%) and proteinuria (33.3%).
median PFS and OS were both not reached. All patients with inflammatory myofibroblastic Fourteen pts were efficacy evaluable. Two confirmed partial remission (PR) were
tumor showed a response (ORR 100%, n = 8/8). In pediatric patients aged 15 and under (n = achieved in pts with renal pelvis cancer (n = 1) and ovarian cancer (n = 1). TT-00973-MS
11), the ORR was 63.6% (30.8 to 89.1) with a DCR of 100.0% (71.5 to 100.0). Although no is slowly eliminated from body, with a half-life of about 55 h. After multiple dosing (QD),
responses were observed in the patients with ALK mutations or amplifications, a clinically steady state was reached within 15 days, and the mean accumulation factor of AUC0-24h
meaningful stable disease was observed in 3 of 5 ALK mutated patients. Grade $3 drug- was approximately 6. Preliminary PK analysis showed a linear increase on exposure.
related adverse events were observed in 15.4% (n = 4) among all treated patients (n = 26), with Plasma levels of soluble AXL (sAXL) increased to approximately 1.7 times (range:
no drug related deaths. Conclusions: Our study showed that patients with ALK alterations 0.9~2.8) the baseline levels at C1D28. Conclusions: The preliminary findings from this
treated with alectinib achieved sustained clinical benefit, meaningful survival outcomes, and phase I study demonstratedthat the AXL inhibitor TT-00973-MS monotherapy exhibits a
safety consistent with previous data. Greatest benefit was observed for the ALK fusion well-tolerable safety profile, promising pharmacodynamic activity, with early signs of
population including pediatric patients. These data support the potential role of alectinib as a
efficacy in pts with heavily pre-treated advanced solid tumors. Further studies are
tumor-agnostic therapy for both pediatric and adult patients with ALK altered solid tumors.
warranted to comprehensively evaluate the efficacy and safety of TT-00973-MS in large
Clinical trial information: jRCT2091220364. Research Sponsor: None.
patient populations and specific tumor types. Clinical trial information: NCT05673538.
Research Sponsor: None.

3107 Poster Session 3108 Poster Session

A phase 1 dose escalation study of LNP7457 (PRMT5 inhibitor) in patients An open-label phase 1 dose-escalation and dose-expansion trial to evaluate
with advanced or metastatic solid tumors. First Author: Shashank Deoghare, the safety, tolerability, and efficacy of TQ-B3234 in adults with neurofibro-
Lupin Limited, Pune, India matosis type 1 (NF1). First Author: Jun Liu, Neurofibromatosis Type 1 Center and
Background: Protein arginine transferase 5 (PRMT5) overexpression plays an important Laboratory for Neurofibromatosis Type 1 Research, Shanghai Ninth People’s Hospital,
role in pathogenesis of several cancers. LNP7457 is an investigational, orally active, highly Shanghai Jiao Tong University School of Medicine, Shanghai, China
potent, S-adenosylmethionine (SAM) competitive PRMT5 inhibitor with wider therapeutic Background: NF1 is an autosomal-dominant genetic disease that can manifest as neuro-
window compared to other investigational PRMT5 inhibitors. Here, we report results of dose fibromas, including cutaneous neurofibromas (cNFs) affecting almost all NF1 patients (pts),
escalation (DE) and food-effect (FE) sub-studies of LNP7457 from a Phase 1 study in and plexiform neurofibromas (PNs, up to 50%), which are benign nerve sheath tumors.
patients with advanced or metastatic solid tumors. Methods: This is first-in-human, However, PNs can cause substantial pain, disfigurement and impairment in pts and are at risk
multicenter, open-label study in patients with advanced or metastatic solid tumors with of transforming into malignant peripheral nerve sheath tumors (MPNSTs). Currently, there is
failed prior standard therapies or for whom no standard therapy exists. DE followed initial no medical cure for adult pts with NF1-related PN (NF1-PN), and fewer options for adult pts
modified acceleration followed by 3+3 design. LNP7457 was administered orally in 21-day with cNFs. TQ-B3234 is a highly selective MEK1/2 inhibitor. A phase 1 dose-escalation and
cycles [first 16 days (on-period) and last 5 days (off-period)] at escalating doses (1 to 4mg dose-expansion trial (NCT05107037) evaluated the safety and efficacy of TQ-B3234 in adult
QD). Primary objective was to determine maximum tolerated dose (MTD) assessed by dose- pts with inoperable NF1-PN, MPNSTs or cNFs. Methods: Eligible adult pts with inoperable
limiting toxicities (DLT) in cycle 1; secondary objectives included safety, pharmacokinetics, NF1-PN, MPNSTs or cNFs were included in the dose-escalation phase using a 3+3 design,
pharmacodynamics and preliminary antitumor activity measured by RECIST v1.1. FE study while only NF1-PN pts proceeded to the dose-expansion phase. TQ-B3234 was administered
was conducted as a single-dose cross over study at MTD to assess impact of food on PK of as a capsule, with doses ranging from 5 mg to 100 mg, once daily in 28-day cycles during the
LNP7457. Results: 17 patients (9 male & 8 females) in DE received LNP745 at doses of 1mg dose-escalation phase. A dose of 50mg was chosen for the dose-expansion phase. The
(n = 1), 2mg (n = 8), 4mg (n = 2), 1.5mg (n = 6); median age 51.0 (range 23 - 76) years; primary endpoints were safety and objective response rate (ORR). ORR was assessed by
Diagnosis included head and neck cancer (n = 9), cervix cancer (n = 2), breast cancer (n = 1), investigators using REiNS for NF1- PN pts, RECIST 1.1 for MPNSTs pts, and paper frames for
ovarian cancer (n = 1), pancreatic cancer (n = 1), gall bladder cancer (n = 1), prostate cancer the cNF population. Results: As of September 30, 2024, 40 adult pts (31 pts of PNs, 7 pts of
(n = 1) and uterine cancer (n = 1). Sixteen patients discontinued treatment; progressive cNFs, and 2 MPNSTs) were enrolled. The distribution of doses was as follows: 4 pts at 5 mg, 1
disease (n = 8); consent withdrawal (n = 4); AE (N = 2); death (n = 1) and lost to follow up (n = at 10 mg, 3 at 15 mg, 24 at 50 mg, 5 at 70 mg, and 3 at 100 mg. One patient, a cNF case,
1). One patient is ongoing at data cut-off of 31-Oct-2024. A total of 16 (94.1%) patients had experienced dose-limiting toxicity (DLT), in the form of G3 diarrhea (16.7%) at the 100 mg
at least 1 treatment emergent adverse event (TEAE). The majority of TEAEs were unrelated dose during the dose-escalation phase. After a median follow-up of 12 months, treatment-
to LNP7457 (76.4%). Anemia (47.1%) and thrombocytopenia (23.5%) were the most emergent adverse events (TEAEs) were observed in 39 pts (97.5%), with the majority were
common TEAEs. A total of 6 (35.3%) patients had serious TEAEs, thrombocytopenia (11.8%) grade 1 or 2. Grade 3 TEAEs were reported in 17.5% of pts; the principal reasons for dose
being the most common serious TEAE. One patient in 4mg cohort had DLT of thrombo- reductions (15.0%) TEAEs included rash acneiform (5.0%), diarrhea (2.5%), and edema (2.5%).
cytopenia during cycle 1. None of the patients in 2mg cohort had DLT in cycle 1, hence 2 mg No death occurred during the study. Among the 30 pts with NF1-PN who had at least one
was determined as MTD. After single and multiple dosing in DE study, peak plasma tumor assessment, 29 pts (96.7%) experienced tumor size reduction, and 11 (36.7%) achieved
concentrations of LNP7457 were observed between 2-4 hrs. A single-dose cross over FE partial response (PR) based on REiNS criteria. The largest reduction in tumor size was 37.7%.
study (n = 6) at 2mg dose showed no impact of food on PK of LNP7457. Target engagement For the 6 cNF pts who had at least one tumor assessment using paper frames, all 6 pts (100%)
based on reduction in plasma SDMA levels was observed at all dose levels in DE study. experienced reduced tumor size, with the largest reduction being 85.2%. Conclusions: TQ-
Preliminary efficacy in the form of stable disease was observed in 8 (66.7%) patients across B3234 demonstrated manageable safety and a significant ORR in adult NF1- PN pts. These
8 different tumor types. Conclusions: LNP7457 was well tolerated with desirable safety, results support the potential of TQ-B3234 to become a new treatment option for NF1- PN pts.
PK/PD and preliminary efficacy profile. Clinical trial information: CTRI/2023/07/054753. Additionally, TQ-B3234 showed deep and durable volume reduction in adult cNF pts, as
Research Sponsor: Lupin Limited, India. assessed by paper frames. This method could serve as a valuable tool in clinical research for
achieving accurate quantitative phenotype for NF1. Clinical trial information: NCT05107037.
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 215s

3109 Poster Session 3110 Poster Session

A phase 1, multicenter, open-label study of HSK42360, a brain-penetrant Anti-tumor activity of BH-30643, a novel macrocyclic kinase inhibitor, in
BRAF inhibitor, in patients with BRAF V600-mutated solid tumors. First EGFR-mutant lung cancer models. First Author: Wei Deng, BlossomHill Thera-
Author: Jian Li, Beijing Cancer Hospital, Beijing, China peutics, Inc., San Diego, CA
Background: Limitations of approved BRAF V600E inhibitors include toxicity from Background: Outcomes on tyrosine kinase inhibitor (TKI) treatment in EGFR-mutant
paradoxical activation of RAF dimerization as well as limited brain penetration. In non-small cell lung cancer (NSCLC) fall short of the durable benefit observed with next-
contrast to approved agents, investigational pan-RAF inhibitors both inhibit mutant RAF generation targeted therapies in ALK and ROS1-driven NSCLC. Novel targeted therapies
proteins and wild-type (wt) RAF proteins, leading to a narrow therapeutic index. are needed to address treatment resistance and offer prolonged patient benefit with
HSK42360 is a next-generation, small-molecule BRAF paradox breaker with high brain reduced toxicity. We recently described (AACR 2025) the design and discovery of BH-
penetration. It displays significantly less paradoxical activation than approved BRAF 30643, a first-in-class macrocyclic reversible TKI targeting the active conformation of
inhibitors and spares wtBRAF-containing RAF dimers. Treatment with HSK42360 results mutant EGFR and offering potent, mutant-selective EGFR inhibition across classical and
in excellent and durable anti-tumor effect in BRAF Class I and II mutant CDX or PDX non-classical EGFR mutations. Here we study diverse preclinical models to assess the
models. Here we report the interim results from a Phase 1 study of HSK42360 in patients breadth of activity from this novel approach. Methods: Anti-tumor activity of BH-30643
(pts) with BRAF V600 mutations (NCT06536400). Methods: This multicenter, open- was evaluated in cell-derived xenograft (CDX) or patient-derived xenograft (PDX) tumor
label, two-part study enrolled adult pts with advanced BRAF V600-mutated solid tumors, models carrying classical or atypical EGFR mutations. CNS activity of BH-30643 was
including those with recurrent or metastatic solid tumors or primary CNS tumors. investigated in an intracranial xenograft model. BH-30643 was administered twice daily
Previous BRAF6MEK inhibitor treatment is permitted. In the dose-escalation (Part 1), via oral gavage; osimertinib when used as a comparator was dosed daily. Studies were
HSK42360 (200-3600 mg/day) monotherapy was given orally. Escalation followed a “3+3 done with n $ 5. The activity of BH-30643 against diverse EGFR exon 20 insertions
design” with dose-limiting toxicities assessed during Cycle 1. Part 2 was cohort ex- (ex20ins) was evaluated in 34 engineered Ba/F3 cell lines in cell proliferation assays
pansion. Primary objectives were maximum tolerated dose and recommended phase 2 in vitro. Results: In the PC-9 (exon 19 del) CDX model, BH-30643 led to deep tumor
dose of HSK42360. Secondary objectives included safety, tolerability, pharmacokinetics, regressions, similar to what was observed with osimertinib at the 25 mg/kg dose level.
pharmacodynamics and preliminary efficacy. Results: As of January 15, 2025, 17 pts Similarly deep responses with BH-30643 were observed in double mutant CDX models
(47.1% male; median age 57.0 years) have been treated with HSK42360 monotherapy including those derived from H1975 cells (cis L858R / T790M) and Ba/F3 cells engi-
across five dose levels (200-3600 mg/day). Of these, 64.7% pts experienced adverse neered with exon 19 del / T790M. In a triple-mutant PDX model (cis exon 19 del / T790M /
events (TEAEs), most frequently increased ALT (23.5%) and increased AST (23.5%). Most C797S) and a Ba/F3 triple-mutant CDX model (cis L858R / T790M / C797S), deep
(89.7%) of TEAEs were grade 1. Two pts had drug-related grade 3 AEs (increased responses were observed with BH-30643 while osimertinib demonstrated no anti-tumor
creatinine and increased ALT) and one had drug-related serious AEs (SAEs) (increased effect. BH-30643 activity was also evident in the HCC827-luc (exon 19 del) intracranial
creatinine). There were no DLT, grade 4 TEAEs, treatment-related discontinuations, or xenograft model with 90% tumor reduction. In two Ba/F3 CDX models carrying atypical
treatment-related deaths. Among 11 efficacy evaluable pts, the ORR was 18.2%. Two (1 mutations (cis G719A / S768I and cis G719A /L861Q), BH-30643 maintained strong anti-
CRC and 1 ganglioglioma) had a partial response (PR) and three had stable disease (SD) tumor activity. Finally, we explored the activities of BH-30643 against 34 different EGFR
with shrinkage (per RECIST or RANO). This trial is ongoing. Conclusions: HSK42360 ex20ins in engineered Ba/F3 cell lines and BH-30643 showed anti-cell proliferation
monotherapy was well tolerated without unexpected safety issues. Preliminary efficacy activity with a median IC50 value of 6.06 nM. Conclusions: These preclinical studies
data demonstrate favorable activity of HSK42360 in pts with BRAF V600-mutated solid demonstrate broad activity of BH-30643 against classical and atypical EGFR activating
tumors, including primary CNS tumors. Clinical trial information: NCT06536400. mutations, EGFR ex20ins, as well as acquired resistance EGFR mutations. Such an
Research Sponsor: Haisco Pharmaceutical Group Co., Ltd. "OMNI-EGFR" inhibitor may be able to overcome some of the limitations of earlier
agents. Supported by favorable ADME and preclinical safety profiles, BH-30643 is now
being assessed in a first-in-human study in locally advanced or metastatic NSCLC
harboring EGFR and/or HER2 mutations (NCT06706076, SOLARA). Research Sponsor:
BlossomHill Therapeutics, Inc.

3111 Poster Session 3112 Poster Session

Phase II trial of trametinib in patients with advanced solid tumors harboring Updated efficacy and safety of zurlectrectinib in adult patients (pts) with
genomic alterations in the MAPK pathway: Results from the BELIEVE trial locally advanced or metastatic NTRK fusion–positive (NTRK+) solid tumors.
(NCCH1901). First Author: Hideyuki Hayashi, Center for Cancer Genomics, Keio First Author: Dan-yun Ruan, Department of Clinical Research, Sun Yat-Sen University
University Hospital, Shinjuku-Ku, Japan Cancer Center, State Key Laboratory of Oncology in South China, Guangzhou, China
Background: The MAPK pathway is one of the most mutated oncogenic pathways in solid Background: NTRK gene fusion is one of the most defined driving factors of carcino-
tumors. However, effective treatments targeting this pathway have not been well- genesis, which occurs in various adult tumor types. Zurlectrectinib, a highly selective
established. The BELIEVE trial aimed to evaluate the efficacy of trametinib, a selective next-generation TRK tyrosine kinase inhibitor, previously demonstrated encouraging
MEK inhibitor, in patients with solid tumors harboring genomic alterations in the MAPK efficacy and manageable toxicity in pts with NTRK+ tumors in a phase I/II clinical trial
pathway. Methods: The BELIEVE trial is a multi-cohort, tumor agnostic phase II trial. El- (NCT04685226). The pivotal phase II clinical trial (NCT05745623) is currently ongoing.
igibility criteria included patients with solid tumors for which no standard treatment was Here we present the integrated results of adult pts from both trials. Methods: Pts with
available or those who had shown resistance or intolerance to standard therapies. In the locally advanced or metastatic solid tumors, who failed from clinical standard of care or
trametinib arm, participants received 2 mg/day of trametinib continuously until disease for whom there was currently no effective therapy, were enrolled in this study. The primary
progression or intolerable toxicity occurred. The primary endpoint was the objective re- endpoint was confirmed objective response rate (ORR) per independent review committee
sponse rate (ORR) within 16 weeks, and secondary endpoints were overall survival (OS), (IRC). Tumor responses were assessed by IRC and investigators per RECSIT1.1 and RANO
progression-free survival (PFS), disease control rate (DCR), and safety. The clinical hy- (BM) criteria. Treatment-emergent adverse events were evaluated and graded according
pothesis was that patients would respond to the genotype-matched drugs. Bayesian to CTCAE v5.0. Results: As of 23 Nov 2024, a total of 229 adult pts were enrolled in the
analysis was performed using a prior distribution with an expected response rate of 30% two trials. Forty-nine TRK inhibitor naı̈ve adult pts were evaluable for efficacy repre-
[Beta (0.6, 1.4)]. Results: Between October 2019 and October 2023, 60 patients with senting 12 different solid tumor types. Among the efficacy population, the distribution of
measurable disease and 9 without measurable disease were enrolled. The top three primary NTRK1, NTRK2 and NTRK3 fusions was 53.1%, 2.0% and 44.9% respectively. The median
tumor sites were the central nervous system (n=20), pancreas (n=7), and ovary (n=6). The age was 51.0 years (range: 18-77). Pts had received a median of two prior lines of
targeted genes for trametinib included non-BRAF V600 (n=26), NF1 (n=18), MAP2K1 (n=10), systemic therapies, with ECOG performance status between 0-1. Median follow-up was
NRAS (n=6), KRAS (n=5), and RAF1 (n=4). Among the full analysis set of 49 patients with 11.7 months. The confirmed ORR by IRC was 83.7% (95% CI: 70.3, 92.7), 5 pts (10.2%) with
measurable disease, the confirmed ORR was 12.2% (95% CI, 4.6% to 24.8%), and the
complete response. Median duration of response (DOR) and median progression-free
expected value of posterior distribution [Beta (6.6, 44.4)] was 12.9%. Partial responses were
survival (PFS) by IRC were not reached. The DOR rate and PFS rate by IRC at 12 months
observed in patients with genomic alterations in non BRAF V600 (n=3), KRAS (n=2), and NF1
was 92.0% and 90.5%, respectively. Two of the three pts (66.7%) who had brain me-
(n=1). However, the confirmed ORR of 12.2% fell below the prespecified threshold of 20%,
tastasis at the baseline achieved intracerebral ORR, which is consistent with the good
therefore the primary endpoint was not achieved. The median OS was 11.9 months (95% CI,
8.5 to 22.9 months), and median PFS was 3.9 months (95% CI, 2.4 to 44 months). The DCR brain penetration and strong intracranial activity of zurlectrectinib. In the safety pop-
was 46.9% (95% CI, 32.5% to 61.7%). Among 59 patients in the safety analysis, severe ulation of adults (N = 229), treatment-related adverse events (TRAEs) were predominantly
adverse events (Grade $3) were observed in 55.9% of patients. The most frequent adverse grade 1 or 2. The most common TRAEs ($20%) were anemia (28.4%), increased alanine
effects were acneiform dermatitis (22%), blood creatine phosphokinase increased (20%), transferase (27.9%) and increased aspartate transferase (25.8%). Grade $3 TRAEs
and stomatitis (15%). Conclusions: The BELIEVE trial demonstrated limited efficacy of ($2%) were weight gain (3.5%) and dizziness (2.2%). No Serious TRAEs occurred in $2%
trametinib in patients with advanced solid tumors harboring genomic alterations in the pts. TRAEs led to dose interruption, reduction and discontinuation in 9.2%, 3.9% and 0.4%
MAPK pathway. While the confirmed ORR did not meet the primary endpoint, the outcomes of safety population, respectively. Conclusions: In line with previously reported results,
for OS, PFS, and DCR were consistent with the clinical hypothesis, suggesting potential zurlectrectinib continued to demonstrate a deep and durable responses in adult pts with
benefit in a subset of patients. Clinical trial information: jRCTs031190104. Research NTRK+ advanced solid tumors with or without brain metastasis. Zurlectrectinib was also
Sponsor: Japan Agency for Medical Research and Development; Health and Labour Sci- well-tolerated and showed favorable safety profile in adult pts with various tumor types.
ences Research Grant. Clinical trial information: NCT05745623. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
216s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3113 Poster Session 3114 Poster Session

Discovery of potent degraders of pan-KRAS based on a novel KRAS binder. Development and validation of a biology-based novel therapeutic agent
First Author: Wei Sun, Global Drug R&D Center, Huadong Medicine Company Limited, targeting the LIN28/let-7 pathway in cancer. First Author: Patrick Sipila,
Hangzhou, China University of Calgary, Calgary, AB, Canada
Background: As the most frequently mutated oncogene, KRAS alteration occurring in Background: Currently, brain tumors that are diagnosed in infants and young children
approximately 25% of all malignancies. Despite extensive efforts, targeting KRAS has carry an exceptionally high risk for treatment resistance and toxicities. The LIN28 family
proven to be challenging due to its structure and complex function. Recently, direct of RNA-binding proteins regulate stem cell biology and pluripotency. In addition to
KRASG12C inhibitors, such as Sotorasib and Adagrasib have occurred first successes. embryonic development, they have also been implicated in oncogenesis through in-
However, therapeutic approaches targeting other variants beyond G12C are under teraction with the tumor suppressor micro-RNA (miRNA), let-7. In cancer, LIN28 ex-
significant unmet needs. Recent developments in protein degradation technologies, pression leads to let-7 loss-of-function, oncogenic activation, and tumorigenesis.
such as Proteolysis-Targeting Chimeras (PROTACs), bring new hope for targeting pan Importantly, LIN28 expression has been associated with stemness and subsequent
KRAS. Based on our novel warheads, potent pan-KRAS degraders were designed and tumor aggressiveness and poor survival in early childhood brain tumors. Thus, LIN28
synthesized. Methods: The ability of compounds to degrade KRAS protein was eval- may offer an effective therapeutic strategy to prevent relapse by specifically targeting
uated using western blotting. The anti-tumor efficacy was assessed in vitro through cancer stem cells. In this study, we describe a novel therapeutic inhibitor of LIN28 in
different mutant cell lines. As a proof of concept study in vivo, an experiment in cancer. Methods: Using an in silico approach, we designed and synthesized a panel of
subcutaneous xenograft mouse model was performed. Pharmacokinetic studies were novel compounds predicted to bind and inhibit the critical molecular interaction between
conducted in mice, with serial blood samples analyzed by (LC2MS)/MS. Results: A LIN28 and let-7. Cytotoxicity was evaluated by alamar blue viability assay in a panel of
novel series of warheads exhibiting exceptional enzymatic and cellular activity against LIN28-positive cancer cell lines derived from atypical teratoid rhabdoid tumor (ATRT),
pan-KRAS has been successfully obtained. Based on these warheads, more than 100 embryonal tumor with multilayered rosettes (ETMR), and germ cell tumor. LIN28-
degraders were meticulously designed and synthesized incorporating a diverse array of negative cells were used as control. LIN28 protein expression and let-7 miRNA levels
linkers and E3 ligands. Through a comprehensive evaluation, two series of compounds were determined by immunoblot and reverse transcription quantitative polymerase
have demonstrated a remarkable KRAS degradation property and downstream inhibition chain reaction (RT-qPCR), respectively. Self-renewal capacity was analyzed by sphere
at concentrations below 10 nM in SW620G12V and GP2DG12D cells. Cell proliferation formation assay. Mice carrying xenografts were treated to investigate LIN28 inhibition in
assay demonstrated that the IC50 values of these degraders are ranging from 0.01 to 30 vivo. Results: Preliminary screening of small molecule inhibitors identified a lead
nM in the MIA PaCa-2G12C, GP2DG12D SW620G12V and LOVOG13D cell lines, without af- compound, designated THNB-3, that induces cell death at micromolar concentrations in
fecting the viability of KRAS-independent cell lines (selectivity . 500-fold). These LIN28-positive cell lines established from various tumors, without affecting LIN28-
compounds also possess favorable PK properties in mice (clearance , 10 mL/min/kg; negative controls. Treatment with THNB-3 increased the level of let-7 tumor suppressor,
IV, 2 mpk, AUC . 5000 ng$hr/mL) and good safety profile (hERG IC50 . 30 mM). confirming effective inhibition of LIN28. In addition to cytotoxicity, THNB-3 significantly
Moreover, these compounds showed strong antitumor activity in xenograft mouse inhibited sphere formation in brain tumor cells, reducing self-renewal and multipotency
model in vivo. Conclusions: The innovative linker elongation and branching, coupled of cancer stem cells. Lastly, the anticancer activity of THNB-3 was validated in vivo
with modifications of KRAS binder portion significantly contributed to potent pan-KRAS against LIN28-positive xenografts and the drug also demonstrated systemic tolerability
degraders, which demonstrate excellent pharmacokinetics and exhibit remarkable ef- in mice. Conclusions: Our studies provide the first evidence for an effective, targeted
ficacy both in vitro and in vivo. The IND-enabling studies are being conducted and the therapeutic agent against the LIN28/let-7 pathway for the treatment of cancer in the
regulatory IND filing will be completed in 2025. Research Sponsor: None. future. THNB-3 selectively induces cytotoxicity in LIN28-positive cancers by restoring
let-7 miRNA, confirming effective target modulation. Further, LIN28 inhibition by THNB-3
may reduce self-renewal and multipotency of cancer stem cells. Together, our preclinical
data supports further development of THNB-3 for the treatment of high-risk LIN28-
positive tumors. Research Sponsor: Kids Cancer Care Foundation of Alberta.

3115 Poster Session 3116 Poster Session

A first-in-human, phase 1a/b, dose-escalation/expansion study of BG- Efficacy and safety of pralsetinib in RET fusion-positive solid tumors: Final
68501, a selective CDK2 inhibitor, as monotherapy or in combination with data from the ARROW trial. First Author: Vivek Subbiah, University of Texas MD
fulvestrant for patients with HR+/HER2- breast cancer and other advanced Anderson Cancer Center and Sarah Cannon Research Institute, Nashville, TN
solid tumors: First disclosure of clinical data. First Author: Rohit Joshi, Cancer Background: Pralsetinib is an oral tyrosine kinase inhibitor that selectively and potently targets
Research SA, Adelaide, Australia oncogenic RET fusion and mutation proteins. RET fusions or mutations are present in various tumor
Background: CDK2 inhibition could represent a novel treatment (tx) option for patients (pts) types. We report the final results from the phase 2 portion of ARROW, a phase 1/2, open-label, multi-
cohort, dose-expansion study evaluating the efficacy and safety of pralsetinib (NCT03037385) in
with resistance to CDK4/6 inhibitors (CDK4/6i) and/or increased cyclin E1 activity. BG-
patients with RET fusion-positive solid tumors other than non-small-cell lung cancer (NSCLC) and
68501 is a highly potent CDK2 inhibitor with high CDK2 selectivity (~100x) vs other CDK thyroid cancer. Methods: Eligible pts were $18 years of age with a pathologically documented,
family members. We present dose-escalation data of BG-68501 as monotherapy or in definitively diagnosed advanced solid tumor with an oncogenic RET fusion or mutation, had previously
combination with fulvestrant in pts with HR+/HER2- metastatic breast cancer (BC) and received standard of care appropriate for their tumor type, and were not eligible for any other study
advanced solid tumors (NCT06257264). Methods: This is the dose-escalation phase of a groups. Overall response rate (ORR) and safety were primary endpoints of the study. Key secondary
first-in-human, phase 1a/b, open-label, multicenter study to evaluate the safety/tolerability, endpoints included duration of response (DOR), progression-free survival (PFS), and overall survival
pharmacokinetic (PK), and pharmacodynamic (PD) profiles, and preliminary antitumor (OS). The final database lock was May 20, 2024. Results: Twenty-nine patients were enrolled with 11
activity of BG-68501 in pts with advanced, nonresectable, or metastatic solid tumors, different solid tumor histologies. Twenty-six (90%) received prior systemic therapy. Median age was
including HR+/HER22 BC. During dose escalation, sequential cohorts received increasing 58 years (range 25-75); 59% were female. Twenty-eight patients were included in the efficacy analysis.
doses of BG-68501 as monotherapy or in combination with fulvestrant. Eligible pts are $18 ORR (by RECIST) was 46.4% (13/28); 10.7% (3/28) achieved complete response (pancreatic cancer,
yrs, with histologically or cytologically confirmed advanced or metastatic solid tumors n=2; cancer of unknown primary, n=1) and 35.7% (10/28) achieved partial response. Median PFS was
7 months (95% CI: 3.9, 12.8). Median DOR was 11.1 months (95% CI: 5.5, 25.1). Median OS was
associated with CDK2 dependency who have received $1 line of tx for advanced or
10.3 months (95% CI: 6.8, 25.2). Twenty-five (86%) patients experienced treatment-related adverse
metastatic disease and prior endocrine therapy and a CDK4/6i in either the adjuvant or events (TRAEs); 19/29 (66%) reported TRAEs $grade 3. The most common TRAEs included increased
advanced or metastatic setting for HR+/HER22 BC, or prior standard of care for all other aspartate aminotransferase (11/29; 38%), increased alanine aminotransferase (10/29; 35%), and
advanced solid tumors. Results: As of Jan 22, 2025, 41 pts (median age 63 yrs) have been anemia (9/29; 31%). Four (13.8%) patients experienced hypertension, and 1 (3.4%) patient had $grade
enrolled. Eleven pts had BC (all received prior CDK4/6i), 12 had ovarian cancer (OC), 7 had 3 hypertension. No new safety risks were identified; AEs remained manageable with supportive care
endometrial cancer, and the remaining 11 pts had other tumor types. To date, 6 dose levels and/or dose modifications. Conclusions: In the phase 2 portion of this trial, responses were observed
(DLs) of BG-68501 monotherapy and 1 DL in combination with fulvestrant have been in many tumor types (Table). Pralsetinib demonstrated robust and durable anti-tumor activity with an
assessed. The median duration of exposure is 1.5 months. Treatment-emergent adverse ORR of 46.4%. These data validate RET fusions as a tissue-agnostic target with sensitivity to RET
events occurred in 39 pts (95.1%; grade $3, 26.8%), with the most common being nausea inhibition and activity beyond NSCLC and thyroid cancer, further supporting the promising potential of
(56.1%; grade $3, 0%), vomiting (48.8%; grade $3, 0%), and fatigue (24.4%; grade $3, 0%); pralsetinib to address the unmet medical need in these patients. Clinical trial information:
no DLTs have been observed. BG-68501 demonstrated a linear PK profile with clinical NCT03037385. Research Sponsor: Blueprint Medicines; Genentech/Roche; Rigel Pharmaceuticals,
Inc.
characteristics consistent with preclinical predictions; signs of TK1 reductions have been
observed across DLs tested, including in heavily pretreated pts. Of the 24 efficacy-evaluable Overall response rate by tumor type.
pts, 1 extensively pretreated HR+/HER22 BC pt experienced PR and 10 pts showed SD. ORR
Cancer Type (patient n) n (%)
Dose escalation is ongoing for both monotherapy as well as in combination with fulvestrant.
Conclusions: BG-68501 demonstrates a favorable safety/tolerability profile, with no DLTs Pancreatic (5) 5 (100)
Cancer of unknown primary (1) 1 (100)
observed to date during dose escalation. Extensively pretreated patients achieving PR and Neuroendocrine (3) 2 (67)
SD with monotherapy, coupled with signs of PD responses and a favorable safety profile, Sarcoma (3) 2 (67)
support continued assessment of BG-68501; updated clinical data will be presented at the Head and neck (2) 1 (50)
Small cell lung (2) 1 (50)
time of the conference. Clinical trial information: NCT06257264. Research Sponsor: Bei- Hepatobiliary (4) 1 (25)
Gene, Ltd. Colorectal (5), gastric (1), ovarian (1), thymic (1) 0

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 217s

3117 Poster Session 3118 Poster Session

Exploring the efficacy and mechanism of action of combined pan-Raf and Atropisomeric pyrrolopyrimidine inhibitor as a targeted approach for RET
MEK inhibition in halting the growth of non-V600 BRAF mutated tumors. tyrosine kinase in neuroblastoma. First Author: Ananya Bharathwaj, UCSD, La Jolla,
First Author: Islam E. Elkholi, Lady Davis Institute, Sir Mortimer B. Davis Jewish General CA
Hospital, Montreal, QC, Canada Background: Increased RET expression is associated with poor prognosis in children
Background: Class 2 & 3 non-V600 BRAF mutations mediate RAF dimerization to hyper- with solid tumors such as neuroblastoma (NB), prompting an interest in RET inhibition. A
activate the MAPK signaling pathway. Encorafenib (Enco; BRAF monomer inhibitor) + number of kinase inhibitors currently in use for cancer patients have RET inhibitory
Binimetinib (Bini; MEK inhibitor) elicit responses in ,15% of patients with non-V600E BRAF activity, but these inhibitors also display activity against other kinases, resulting in
mutations (NCT03839342). We hypothesized that Belvarafenib (Belva), a novel pan-RAF unwanted side effects and limiting their safety and efficacy. However, developing more
dimer inhibitor, is more potent than Enco in inhibiting the growth of Class 2 & 3 non-V600 specific RET inhibitors remains a drug design challenge due to high levels of con-
BRAF mutated tumors. Methods: We performed in vitro colonogenic assays to compare the servation between kinase binding pockets. Using novel chiral chemistry leveraging
growth inhibitory effect of 5 independent doses of individual inhibitors (Bini, Belva, or Enco) in atropisomerism to convert a promiscuous, rapidly interconverting pyrrolopyrimidine
parallel to 25 different combinations of either Belva+Bini or Enco+Bini in 6 non-V600 BRAF compound into an atropisomerically stable analog, we have developed a new atropi-
mutated melanoma (WM3629, HMV-II), colorectal (CRC) (NCI-H508, HT55), and lung (NCI-
somerically stable, highly selective and specific RET inhibitor, getretinib, with similar
H1666, NCI-H2087) cancer cells. Low nanomolar doses (10-1000 nM) were used to inves-
potency and improved selectivity to that of other next generation RET inhibitors but with
tigate the synergistic potential of either combination using the SynergyFinder tool. Belva+Bini
(15 mg/kg each) and Enco+Bini (75 mg/kg + 15 mg/kg) combinations were assessed in 4 non-
half the molecular weight and significantly improved ligand efficiencies towards RET.
V600 BRAF (3 Class 3, 1 Class 2) metastatic CRC patient-derived xenograft (PDX) models. The Methods: Associations of gene expression with patient survival and prognostic features
inhibitory effect of Belva and Enco on MAPK activity in the outlined 6 cell lines was assessed were performed on available neuroblastoma tumor databases using the R2 Genomics
by immunoblotting. Transcriptomic (RNA-Seq) analysis was performed on PDXs. Analysis and Visualization Platform. The efficacy of RET inhibition was assessed
Results: Belva+Bini was 2-6-fold more effective than Enco+Bini in inhibiting the growth of against a panel of NB cell lines using live cell imaging and cell viability assays,
the 6 cell lines. Belva+Bini achieved overall higher synergy scores in the 6 cell lines and was comparing results with the active and selective RET kinase inhibitor, (R)-getretinib to
synergistic in 5/6 cell lines (synergy score . 10) vs. Enco+Bini that was synergistic in 1/6 cell results with the inactive atropisomer, (S)-getretinib. Mechanisms of cell death and
lines. In the 6 cell lines, Belva inhibited MAPK activity more robustly than Enco (assessed by impacts on RET signaling in cells treated with (R)- and (S)-getretinib were evaluated by
pERK levels). In vivo, Belva+Bini was significantly more effective than vehicle or Enco+Bini in Western blots. Results: (R)-getretinib reduced NB cell confluence in a dose-dependent
halting the growth of 3 out of 4 PDXs. Both Belva+Bini and Enco+Bini significantly inhibited manner, while (S)-getretinib had no significant effect on cell confluence over time. R-
MAPK activity vs. vehicle (as assessed by the transcriptional MAPK Pathway Activity Score). getretinib treatment of NB cells resulted in reduced phosphorylation of RET in a dose-
However, there was no statistically significant difference between both combinations. Gene dependent manner, while treatment with (S)-getretinib resulted in paradoxical increase
Set Enrichment Analysis revealed that Belva+Bini significantly downregulated genes me- in RET phosphorylation. Conclusions: We present (R)-getretinib as an atropisomerically
diating the interconnected mTORC1 pathway activity and cholesterol metabolism dynamics in stable and potent inhibitor of RET and have shown its efficacy in in vitro models of NB.
the 3 PDXs where Belva+Bini had anti-growth effect. Specifically, among the top down- The high selectivity of (R)-getretinib towards RET has the potential to minimize un-
regulated genes by Belva+Bini was PCSK9, a druggable key regulator of cholesterol wanted side effects caused by off-target kinase binding, thereby increasing its potential
metabolism. Conclusions: These results from 10 preclinical models, tested so far, put for clinical utility. Research Sponsor: None.
forward combined Pan-Raf and MEK inhibition as a potential effective treatment choice for
patients with non-V600 BRAF mutated tumors to be investigated in clinical trials. In parallel,
they unravel novel insights into the mechanism of action of this therapeutic approach and in
return the druggable vulnerabilities of the non-V600 BRAF mutated tumors, a notion we are
further investigating in the outlined models. Research Sponsor: Canadian Cancer Society;
707457; Conquer Cancer, the ASCO Foundation; Canadian Cancer Society; 708442.

3119 Poster Session 3120 Poster Session

Combined RAF- and MEK-inhibition in solid cancers with kinase-impaired Safety and efficacy of a small-molecule c-Myc degrader WBC100 in solid
BRAF mutations (SORATRAM phase I trial). First Author: Anna Lena Illert, Center tumors: A first-in-human, phase I trial. First Author: Qi Zhang, Department of
for Personalized Medicine (ZPM) Technical University of Munich (TUM) and Department Hepatobiliary and Pancreatic Surgery, The First Affiliated Hospital, Zhejiang University
of Medicine III, TUM University Hospital, Technical University of Munich (TUM) and School of Medicine, Hangzhou, China
German Cancer Consortium (DKTK), Munich Partner Site, Munich, Germany Background: c-Myc amplification or overexpression is involved in the development and
Background: BRAF is a frequently mutated gene in cancer, with most mutations (mut) at the progression of many human cancers, and is often associated with poor outcomes. It is
activating hotspot V600 codon. Recently, kinase-inactive class III BRAF mut emerged as on- an extraordinarily desirable target, but is also considered undruggable. WBC100 is an
cogenic driver and potential therapeutic target, as they lead to paradoxical cross-activation of oral active molecule glue that selectively degrades c-Myc protein. Methods: This is a
RAF1- and RAS-dependent downstream signaling. Here we report the Phase I toxicity results of first-in-human, dose-escalation study conducted in China. Patients with solid tumors
combinatory inhibition of RAF kinases by sorafenib (S), a multi-kinase inhibitor (e.g. RAF1, BRAF,
that have progressed or relapsed after standard systemic therapy were enrolled.
c-KIT, FLT-3) and MEK/ERK signaling by trametinib (T) in patients (pts) with inactivating BRAF
mut. Methods: SORATRAM is a prospective, molecularly stratified, multicenter phase I trial.
WBC100 was orally administered every other day (QOD) according to 3+3 design. The
Primary objective is to determine the maximal tolerated dose (MTD) of T combined with S and the primary endpoints were safety, dose-limiting toxicity (DLT) and maximal tolerated dose
recommended phase II dose (RP2D). Adult pts with metastatic malignancies, confirmed or known (MTD). Results: As of Dec 9, 2024, 28 patients were enrolled in seven dose levels (DLs)
impaired kinase BRAF mut (according to in vitro testing), ECOG # 2 and no available therapy from 0.5 to 3.5 mg. The median age was 59 (range, 45-71) years, comprising 15 males
options were eligible. S was given in the approved dose (800 mg) from day (d)1 cycle (c)1, and 13 females. Three (11%) patients had an ECOG PS score of 0, while the remaining 25
combined with T on c1d8 for max 12c or until progression or unacceptable toxicity. Dose levels (89%) had a score of 1. The median number of prior systemic therapy lines was 3 (range,
(DL) are defined by T dose (0.5mg DL1, 1.0mg DL2 and 1.5mg DL3) and escalated in a con- 1 to 6). One DLT of prolonged QT interval was observed in DL7, and MTD has not been
ventional 3 + 3 design. MTD is defined as highest dose at which 0/3 pts or , 2/6 pts experience a reached. Six patients (21%) experienced grade 3 or higher treatment-related adverse
dose limiting toxicity (DLT) during c1. DLT is defined as toxicity related to S+T combination, events, including five (17.9%) neutropenia and two (7.1%) leukopenia and one (3.6%)
unrelated to disease progression, intercurrent illness or concomitant medications, that requires prolonged QT interval. Increased aspartate aminotransferase, thrombocytopenia, pro-
dose reduction or drug withdrawal. Results: Since 2020, 236 cases from 9 sites were classified
teinuria, increased alanine aminotransferase, fatigue, nausea, anemia, and hypo-
for mutational SORATRAM eligibility with 42% being kinase impaired (e.g. D594G, N581I, G466E),
21% known intermediate/high activity (excluding V600E/K) (e.g. L597V, K601E) and 36% with
albuminemia were the most commonly reported grade 1 or 2 adverse events. Nineteen
novel/unclear/unknown kinase activity (e.g. G469I, W531S). Eligible pts with inactivating BRAF patients were evaluable for efficacy, one (5.3%) showed partial regression (PR), and six
mut proceeded to SORATRAM screening. 15 pts received dose finding treatment: 3 in DL1 and DL2 (31.6%) showed stable disease (SD), including two patients with hepatocellular car-
and 9 in DL3. Median age was 57 years (34-75) with 12f/3m pts. Included entities were colorectal cinoma, one with duodenal adenocarcinoma, and three with pancreatic cancer. Notably,
cancer (60%), duodenal carcinoma/carcinoma of papilla vateri (20%), lung adenoid cystic car- we enrolled eight patients with pancreatic cancer at DL6 and DL7, and six of them were
cinoma (6.7%), bone sarcoma (6.7%) and ovarian cancer (6.7%). 3/3 pts in DL1 and DL2 and 6/9 evaluable for efficacy, with one (16.7%) PR and two (33.3%) SD. Conclusions: WBC100
pts at DL3 fulfilled the minimum safety evaluation requirements ($ 80% of S+T doses in c1; 28ds showed a tolerable safety profile and preliminary anti-tumor activity in advanced solid
observation). No DLT was observed in DL1 and DL2. 1/6 pts in DL3 developed a DLT (reduction of tumors especially in PDAC. Dose escalation is ongoing and expected to proceed to dose
left ventricular ejection fraction (LVEF)). 5/15 pts (33.3%) experienced grade 3 adverse events expansion soon. To our knowledge, this is the first study of a small-molecule c-Myc
(AEs) during c1: Hypertension (13.3%), gastrointestinal bleeding (6.7%; rated as SAE), anemia degrader for further clinical development in cancer. Clinical trial information:
(6.7%), LVEF reduction (6.7%), diarrhea (6.7%) and fatigue (6.7%). 79 AEs grade 1/2 were reported NCT05100251. Research Sponsor: Weben Pharma.
in c1. Conclusions: Combination of sorafenib/ trametinib is feasible and can be safely ad-
ministered to pts. MTD was determined as DL3 (800mg S + 1.5mg T), RP2D as DL2 (800mg
S + 1mg T). Dose expansion part of SORATRAM is open for enrollment. Clinical trial information:
EU – CT No. 2024-512887-77-00. Research Sponsor: German Cancer Consortium (DKTK); German
Cancer Consortium (DKTK) Freiburg site; Sorafenib was kindly supplied by Bayer.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
218s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3121 Poster Session 3122 Poster Session

In vitro efficacy of CDK9 inhibitor tambiciclib (SLS009) in ASXL1 mutated Safety and efficacy of EIK1003, a selective PARP1 inhibitor, as monotherapy
colorectal cancer cell lines. First Author: Dragan Cicic, Sellas Life Sciences Group, in participants with advanced solid tumors. First Author: Guru P. Sonpavde,
New York, NY AdventHealth Cancer Institute, Orlando, FL
Background: ASXL1 (Additional sex combs-like 1) gene encodes ASXL1 protein thought to Background: PARP inhibitors (PARPi) selectively kill tumor cells with genetic mutations
disrupt chromatin, enhancing transcription of certain genes while repressing the tran- in critical DNA repair genes (eg, BRCA1/2). While approved nonselective PARPi may
scription of others. ASXL1 mutations occur in ~20% of acute myeloid leukemia (AML) provide antitumor activity, they are associated with hematologic toxicities. Drugs
patients and ~50% of ASXL1 AML mutations are frameshift or nonsense mutations inhibiting PARP1 but not PARP2 may improve the risk-benefit profile by retaining
generating a truncated ASXL1 form with oncogenic gain of function in AML. Recently antitumor activity while avoiding PARP2-related toxicities. EIK1003 (IMP1734) is a
ASXL1 mutated AML patients were treated with a CDK9 inhibitor tambiciclib (SLS009) with potent PARP1-selective inhibitor that may widen the therapeutic index in susceptible
promising results. ASXL1 mutations were also reported in 55% of Colorectal Carcinoma tumors. Methods: EIK1003-001 (IMP1734-101) is an ongoing global, multi-center,
with High Microsatelite Instability (CRC MSI-H) cell lines, but it is not known whether those Phase 1/2 study evaluating the safety and efficacy of EIK1003 (once daily oral) as
mutations are similar to ASXL1 mutations observed in AML and whether CDK9 inhibitors monotherapy or in combination with anticancer agents in participants (pts) with ad-
have enhanced cytotoxic effect in CRC MSI-H cells with those mutations. vanced solid tumors (NCT#06253130). Pts must be $ 18 yrs with deleterious or
Methods: Twelve CRC MSI-H cell lines were treated with SLS009 at various concentra- suspected deleterious mutations in select homologous recombination repair genes. This
tions. Staurosporin was used as positive control. Cytotoxicity analysis was performed by abstract reports the interim safety and efficacy from Part 1 monotherapy (dose es-
CellTiter-Glo 2.0 assay. Data analyses were performed using GraphPad Prism 9. NGS was calation). Results: At the data cut (10 Jan 2025), 32 pts were treated in the first 4
used to determine mutations in studied cell lines. The experiment was designed to completed dose levels (DLs; n = 3 to 15 per DL, including backfill) of monotherapy dose
compare ASXL1 mutations in CRC MSI-H cell lines to those observed in AML and determine escalation. There were no dose-limiting toxicities and a maximum tolerated dose has not
efficacy of SLS009 cytotoxicity in CRC MSI-H cells with and without ASXL1 mutations. been reached. The majority (29/32) of pts were female. Pts had a median age of 60 years
Highly effective concentrations in this experiment were considered those with IC50 values (31 to 76), and cancers represented included ovarian (n = 15), HER2-negative breast (n =
below 100 nM. Results: Among the 12 tested cell lines, 8 (67%) had non-synonymous 10), pancreatic (n = 3), fallopian tube (n = 2), and prostate cancer (n = 1). Pts received a
ASXL1 mutations of any kind, similar to the literature reported 55% ASXL1 mutations rate
median of 3 (range 1 to 11) prior lines of therapy for metastatic disease with 50%
in CRC MSI-H. Among cell lines with ASXL1 mutations, 4 (50%) had high impact frameshift
receiving prior PARPi. To date, EIK1003 demonstrated a tolerable safety profile. All pts
mutations, similar to estimated rate of high impact frameshift mutations in AML (~50%).
experienced at least one treatment-emergent adverse event (TEAE), and 13/32 expe-
Among cell lines with high impact frameshift mutations, all had mutations in protein
rienced at least one $ Grade 3 TEAE. 27/32 pts experienced a treatment-related AE
position regions 581-582 and 642-643. Three out four had in addition high impact
frameshift mutations in the region of protein position 637-638. Protein positions of these
(TRAE), 6 of which experienced a $ Grade 3 TRAE. Hematologic toxicities (Grade 3)
frameshift ASXL1 mutations were similar to those observed in AML (591 – 592 and 635 – included neutropenia (3/32) and anemia (1/32). 7/32 experienced a serious adverse
646). Among the cell lines with any ASXL1 mutation, 4/8 (50%) had IC50 values for SLS009 event (including one related case of Grade 3 vomiting). There were no Grade 4 AEs or
below 100 nM (highly efficacious) vs 0/4 (0%) among cell lines without ASXL1 mutations. deaths due to AE, and no trends in AEs by DL were observed. EIK1003 PK was linear,
Among cell lines with ASXL1 frameshift mutations, SLS009 was highly efficacious in 3/4 with a half-life . 24 hours. Of the 13 ovarian cancer patients with post-treatment scan
(75%) cell lines vs 1/8 (12.5%) in cell lines without ASXL1 frameshift mutations. High assessments, 3 experienced partial response (PR; one each at DL2, DL3, and DL4) and 3
efficacy was observed in all cell lines (3/3, 100%) with frameshift mutations in the protein experienced stable disease (SD; one at DL1 and two at DL3) by RECIST v1.1. For these 3
position region 637-638. Presence of high impact TP53 mutations did not appear to PRs, all had a CA125 response. Of the 6 breast cancer patients with post-treatment scan
significantly affect SLS009 efficacy. Conclusions: Results indicate that ASXL1 mutations assessments, there was 1 PR (DL3) and 1 SD (DL1) via RECIST v1.1. Conclusions: To
may be oncogenic drivers in some solid tumors, like CRC MSI-H, similar to those in AML and date, EIK1001 has demonstrated tolerable safety and encouraging preliminary efficacy.
that efficacy of CDK9 inhibition with SLS009 may be similar in some solid tumors to the Dose escalation is ongoing. Updated safety and efficacy data will be reported at the time
efficacy observed in AML. Research Sponsor: None. of presentation. Clinical trial information: 06253130. Research Sponsor: None.

3123 Poster Session 3124 Poster Session

A first-in-human phase I/Ib study of ATG-037 monotherapy and combina- Phase 1 study of zavondemstat (TACH101), a first-in-class KDM4 inhibitor,
tion therapy with pembrolizumab in patients with advanced solid tumors: in patients with advanced solid tumors: Results on safety, pharmacokinet-
STAMINA-01. First Author: Janine Margaret Lombard, Calvary Mater Hospital ics, and anti-tumor activity. First Author: Apostolia Maria Tsimberidou, University of
Newcastle, Newcastle, NSW, Australia Texas MD Anderson Cancer Center, Houston, TX
Background: ATG-037 is a highly potent oral small molecule inhibitor of CD73. Background: Zavondemstat is an epigenetic targeting inhibitor of KDM4 histone demeth-
STAMINA-01 is an open-label, first-in-human, phase 1/1b study (NCT05205109) ylase. Dysregulation of KDM4 enzymes (isoforms A-D) has been implicated in various cancers
designed to evaluate the safety, pharmacokinetics, and optimal dosing of ATG-037 as where they drive oncogenesis and resistance pathways by regulating gene transcription.
monotherapy and in combination with pembrolizumab in patients with refractory/ Preclinical studies of zavondemstat demonstrated robust anti-proliferative effects and
relapsed solid tumors. Methods: The study successfully completed enrollment of significant inhibition of tumor growth across numerous xenograft and PDX models. Targeting
dose escalation of ATG-037 with optional addition of pembrolizumab following two KDM4 offers the potential to reprogram the epigenetic dysfunction of cancer cells and inhibit
cycles of monotherapy in May 2024. The primary objectives were to evaluate the safety drivers of tumor dedifferentiation and proliferation leading to apoptotic cell death. This is the
and define the optimal biological dose of ATG-037 as monotherapy and combination first clinical evaluation of a pan-isoform KDM4 inhibitor. Methods: TACH101-CS-0001
(NCT05076552) was an open-label Phase 1 study assessing zavondemstat’s safety, toler-
treatment. As of 20 January 2025, 43 patients were enrolled across the following doses -
ability, pharmacokinetics (PK), and recommended Phase 2 dose (RP2D) in patients (pts) with
20mg BID (n=3), 60mg BID (n=6), 120mg BID (n=10), 240mg BID (n=6), 400mg BID (n=12)
advanced solid tumors. Pts received zavondemstat orally on a weekly schedule in 28-day
and 600mg BID (n=6). The trial is currently recruiting the second part of the study for cycles. Dose escalation followed a Bayesian optimal interval (BOIN) design and explored both
dose optimization of upfront combination therapy at two dose levels (120mg BID and intermittent and continuous dosing. Inclusion criteria included heavily pre-treated advanced/
400mg BID). Results: Efficacy: As of the data cut-off (20 Jan 2025), 43 patients were metastatic solid tumors that progressed or were non-responsive to available therapies and for
enrolled on study and received monotherapy. While on ATG-037 monotherapy, 21 which no standard therapy exists. Pts must have measurable disease according to RECIST
patients had a best response of stable disease (SD) with a disease control rate (DCR) of (v1.1) and ECOG score of 0 to 1. Exclusion criteria included severe hematologic, hepatic, or
49%. Twenty-eight patients with a history of acquired checkpoint inhibitor resistance renal insufficiency. Primary endpoints included safety/tolerability, MTD and RP2D. Secondary
received combination therapy; 7 of which (5 melanoma and 2 NSCLC patients) endpoints included PK and radiographic response per RECIST v1.1. Results: Thirty patients
achieved a confirmed partial response (PR) with an overall response rate (ORR) of 25% were enrolled across 6 dose cohorts. MTD was not reached; RP2D was not determined. The
(95% CI: [51.33, 86.78]). Additionally, 15 patients had a best response of SD with a DCR of most common treatment-related adverse events (TRAEs) were diarrhea (12%), fatigue (7%),
79% (95% CI: [8.30, 40.95]). Of the 11 enrolled melanoma patients who received decreased appetite (7%), nausea (7%), and hyponatremia (7%). All TRAEs were Grade 1 or 2
combination, 5 achieved a PR for an ORR of 45% and 6 achieved SD for an DCR of 100%. (no TRAEs $ Grade 3 were reported). No treatment-related serious adverse events (SAEs) or
Of the 9 enrolled NSCLC patients who received combination, 2 achieved a PR for an ORR dose limiting toxicities (DLTs) were reported. In 23 response-evaluable patients, 10 patients
of 22% and 4 achieved SD for an DCR of 67%. Safety: While on monotherapy, 24/43 (56%) (44%) achieved stable disease (SD) across dosing cohorts. Two patients (9%) had SD $
patients reported treatment-related adverse events (TRAEs). While on combination 6 months, including a patient with castration-resistant prostate cancer (CRPC) and a patient
therapy, 17/28 (61%) patients reported TRAEs. The majority of TRAEs were grades 1-2. with leiomyosarcoma. Another patient with leiomyosarcoma is continuing to receive
The only dose limiting toxicity was a grade 3 rash which occurred at the monotherapy zavondemstat under compassionate use, demonstrating SD for a total of 6 months at time of
400mg BID dose. Only one serious TRAE (grade 3 immune mediated hepatitis) was abstract submission. Zavondemstat demonstrated a dose-proportional exposure profile
reported at the data cut-off. Conclusions: In relapsed/refractory solid tumor patients, with a short half-life of about 1.5 hours. There was no to minimal drug accumulation ob-
ATG-037 appears to be well tolerated as monotherapy and in combination with pem- served. Conclusions: Zavondemstat was very well tolerated and showed encouraging
preliminary signals of clinical benefit in very heavily pre-treated metastatic cancer patients.
brolizumab. The preliminary efficacy data is encouraging and suggests that the
Continued evaluation of zavondemstat is warranted. Clinical trial information: NCT05076552.
combination regimen may provide a new therapeutic option for CPI resistant NSCLC and
Research Sponsor: Tachyon Therapeutics; California Institute for Regenerative Medicine
melanoma patients. Clinical trial information: NCT05205109. Research Sponsor: None. (CIRM).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 219s

3125 Poster Session 3126 Poster Session

Phase I dose-escalation study of the safety and pharmacokinetics of PAS- Pertuzumab plus trastuzumab (P+T) in patients (pts) with bladder (BC) and
004, a macrocyclic MEK inhibitor, for the treatment of patients with MAPK ovarian cancer (OC) with ERBB2/3 alterations (alt): Results from the Tar-
pathway–driven advanced solid tumors. First Author: Tiago Reis Marques, geted Agent and Profiling Utilization Registry (TAPUR) study. First Author:
Pasithea Therapeutics Corp, Miami, FL John K. Chan, Sutter Cancer Research Consortium, San Francisco, CA
Background: PAS-004 is a small molecule allosteric inhibitor of MEK 1/2 and the first Background: TAPUR is a phase II basket study evaluating antitumor activity of commercially available
macrocyclic structure MEK inhibitor in clinical development. Macrocycles are large cyclic targeted agents in pts with advanced cancers with genomic alt. Results of two cohorts of pts with BC or OC
with ERBB2/3 alt treated with P+T are reported. Methods: Eligible pts had measurable disease, ECOG
molecules that can bring increased potency, metabolic stability, and oral bioavailability.
performance status (PS) 0-2, adequate organ function, and no standard treatment (tx) options. Genomic
PAS-004 was developed to reduce metabolic liabilities and overcome the limited exposure testing was performed in CLIA-certified, CAP-accredited site selected labs. Recommended dosing was P at
and stability of known MEK inhibitors. We report initial results of an ongoing Phase I dose an initial dose of 840 mg intravenously (IV), then 420 mg IV every 3 weeks (wks) and T at an initial dose of 8
escalation, multicenter study of PAS-004 in monotherapy in patients with advanced re- mg/kg IV, then 6 mg/kg IV every 3 wks until disease progression. Primary endpoint was disease control (DC)
fractory solid tumors. Methods: The Phase 1 clinical trial is a multi-center, open-label, per investigator defined as complete (CR) or partial (PR) response per RECIST v.1.1, or stable disease (SD)
dose escalation 3+3 study design to evaluate the safety, tolerability, pharmacokinetic (PK), of at least 16 wks duration (SD16+). CR was based on radiographic assessment. For both cohorts, Simon 2-
pharmacodynamic (PD), and preliminary efficacy of PAS-004 in patients with MAPK stage design was based on a null DC rate of 15% vs. 35% (power = 0.85; a = 0.10). If $2 of 10 pts in stage I
had DC, 18 more pts were enrolled; otherwise, the cohort was closed. If $7 of 28 pts had DC, the null DC rate
pathway driven advanced solid tumors with a documented RAS, NF1 or RAF mutation, or was rejected. Secondary endpoints were objective response (OR), progression-free survival (PFS), overall
patients who have failed BRAF/MEK inhibition (NCT06299839).Eligible pts, $18 years with survival (OS), duration of response and SD, and safety. Results: 28 pts with ERBB2/3 alt were enrolled in
MAPK-driven advanced solid tumors, are being enrolled in dose escalation at 8 different each cohort. The table shows demographics and outcomes. For the BC cohort, 2 CRs (ERBB2 amplification
dose cohorts in monotherapy. Results: As of 3 Jan 2025, a total of 9 patients have been [amp, n=1] and ERBB2 amp and ERBB3 mutation [mut, n=1]), 5 PRs (ERBB2 amp [n=3], ERBB2 amp and mut
enrolled in dose escalation in 3 dose cohorts (2mg, 4mg, and 8mg). 55.6% of the patients [n=1], and ERBB2 mut [n=1]) and 3 SD16+ (ERBB2 mut [n=3]) were observed for DC rate of 37% (90% CI, 24 to
were female with a median age of 60 years. Most common tumor types included colorectal 100) and OR rate of 25% (95% CI, 11 to 45). The null DC rate was rejected (p=0.005). For the OC cohort, 2 PRs
(ERBB2 amp [n=1] and ERBB2 mut [n=1]) and 3 SD16+ (ERBB2 amp [n=2] and ERBB2 amp and mut [n=1])
(n = 5, 55.56%), pancreatic (n = 2, 22.22%), gastroesophageal (n = 1, 11.11%), and of
were observed for DC rate of 25% (90% CI, 10 to 100) and OR rate of 7% (95% CI, 1 to 24). The null DC rate
unknown type (n = 1, 11.11%). Treatment Related Adverse Events (TRAE) were reported in was not rejected (p=0.29). Across both cohorts, 4 pts had 6 tx-related serious adverse events (SAE)
44.44% of patients. TRAEs were all low grade (n = 7, 100% were Grade 1-2). No Grade 3, 4 or including: infusion-related reaction, confusion, diarrhea, and fever, and 2 pts had 1 grade 3 tx-related
5 TRAEs were reported. The most common TRAEs were gastrointestinal disorders (n = 4, adverse event (AE) each including: GGT increase and lymphopenia. No pts had grade 5 SAEs.
57.14%), dehydration (14.29%), arthralgia (14.29%) and urinary incontinence (14.29%). No Conclusions: P+T met prespecified criteria to declare clinical activity in pts with BC with ERBB2 alt, but not
rash was observed in any dose cohort. No dose limiting toxicities were detected, and the in pts with OC. Additional study is warranted to confirm the efficacy of P+T in pts with BC with ERBB2 alt.
MTD has not been reached. Preliminary PAS-004 PK analysis suggests linear PK with Clinical trial information: NCT02693535. Research Sponsor: Genentech; Astrazeneca, Bayer, Boehringer
Ingelheim, Bristol Myers Squibb, Eli Lilly and Company, Merck, Pfizer, Seagen (now a wholly owned
estimated t1/2 of 70h, Cmax/Cmin ratio of 1.4 at steady state, achieving potentially subsidiary of Pfizer Inc.), Taiho Oncology.
sufficient exposures for target engagement at the highest dose tested. In the efficacy
evaluable population (n = 6), early response evaluation reveals stable disease (SD) by Demographics and efficacy outcomes.
RECIST 1.1 was observed in 2 patients, with progression free survival of up to 159 days and BC (N=28) OC (N=28)
overall survival of up to 253 days. Conclusions: To date, PAS-004 is shown to be a safe and ECOG PS, N (%) 0 6 (21) 11 (39)
well-tolerated novel MEK inhibitor, with dose-dependent PK profile and preliminary clinical 1 17 (61) 16 (57)
activity in monotherapy in patients with heavily pre-treated refractory solid tumors. PAS- 2 5 (18) 1 (4)
Prior systemic regimens, N (%) 1-2 6 (21) 13 (46)
004 has the potential to achieve prolonged target inhibition and once-daily dosing (QD) due $3 22 (79) 15 (54)
to its long half-life and low Cmax to Cmin ratio. These findings provide a compelling DC (OR plus SD16+) rate, % (90% CI), p-value 37 (24, 100), p=0.005 25 (10, 100), p=0.29
rationale to continue to test PAS-004 into clinical trials for the treatment of MAPK-driven OR rate, % (95% CI) 25 (11, 45) 7 (1, 24)
Median PFS, wks (95% CI) 13 (7, 22) 8 (8, 16)
opportunities. Clinical trial information: NCT06299839. Research Sponsor: Pasithea Median OS, wks (95% CI) 32 (17, 54) 44 (26, 89)
Therapeutics.

3127 Poster Session 3128 Poster Session

Digital spatial profiling of advanced solid tumors and lymphomas from a Long non-coding RNA (lncRNA) SNHG11 as a prognostic and predictive
phase 1 trial of copanlisib and nivolumab. First Author: Sayak Ghatak, Frederick biomarker in metastatic colorectal cancer (mCRC): Insights from CALGB
National Laboratory for Cancer Research, Frederick, MD (Alliance)/SWOG 80405. First Author: Michela Bartolini, Division of Medical On-
Background: Digital spatial profiling (DSP) is an innovative technique that facilitates cology, Norris Comprehensive Cancer Center, Keck School of Medicine, University of
spatially resolved proteotranscriptomic analysis within tissue sections, providing essential Southern California, Los Angeles, CA
insights into the tumor microenvironment (TME). As part of the exploratory objectives in Background: LncRNAs have emerged as key regulators of cancer progression and therapeutic
the Phase 1B trial (NCT03502733) evaluating adult patients(pts) with solid tumors and responses. In CRC, several lncRNAs have been implicated in modulating tumor growth, metastasis
lymphomas copanisib (C), nivolumab (N) + ipilumumab (I), we employed the NanoString- and treatment resistance by interacting with critical tumorigenic pathways. Here, we investigate
Bruker GeoMx DSP platform to evaluate spatial heterogeneity in differentially regulated the potential prognostic and predictive value of lncRNA expression in patients (pts) with mCRC
biomarkers. Methods: The study analyzed samples from pts in the trial’s doublet enrolled in CALGB/SWOG 80405 (NCT00265850) trial. Methods: We analyzed 433 mCRC pts
treatment arm, which included C + N. Pts receive C on days 1 and 15 or days 1, 8, and 15 of treated with either bevacizumab (bev, n = 226) or cetuximab (cet, n = 207) plus first-line
each cycle and nivolumab on day 1 or days 1 and 15 of each cycle. Core biopsies collected chemotherapy. Tumor RNA expression (Illumina HiSeq 2500) of 13 candidate lncRNAs (SNHG11,
HOTAIR, FGF14-AS2, H19, YWHAE, NEAT1, MIR100HG, UCA1, LINC00973, SLCO4A1-AS1,
from 8 pts at cycle 1 day 1 pre-dose (C1D1, baseline), cycle 1 day 8 post-dose (C1D8, C only)
POU5F1P4, MALAT1, HCG18) was explored. Median overall survival (mOS) and progression-free
and cycle 2 day 15 post-dose (C2D15, C+N) were selected for GeoMx analysis. Tissue
survival (mPFS) in months (mo) were compared between pts grouped by tertiles (low [L] vs
sections (5-mm) from archival FFPE blocks were prepared on glass slides and hybridized medium [M] vs high [H]) of gene expression. Likelihood ratio tests, hazard ratios and 95%
with photocleavable tag-conjugated antibodies (targeting 85 proteins) and oligonucleotide confidence intervals were computed from multivariable Cox proportional hazards models,
probes (whole transcriptome- WTA) for protein and WTA analyses respectively. Tissue adjusting for age, sex, ECOG PS, tumor location, number of metastatic sites, KRAS, Consensus
imaging was performed via high-resolution fluorescent microscopy using morphology Molecular Subtypes (CMS), and treatment. Results: Overall, SNHG11 was strongly associated
markers (cytokeratin AE1/AE3, CD45, CD3, CD20, Syto-13 nuclear marker). At least three with OS and PFS after adjusting for multiple tests (Benjamini-Hochberg False Discovery Rate ,
rectangular (660 x 784 mm) regions of interest (ROI) per timepoint were analyzed using 0.05). High SNHG11 expression (H group, n = 144) was associated with improved mPFS (H: 14.3
NanoString nCounter for proteomics and NGS for WTA. Data quality control and analysis vs M: 11.2 vs L: 8.3 mo; p = 0.038) and mOS (H: 39.6 vs M: 31.1 vs L: 20.5 mo; p = 0.033) in the
were conducted using the GeoMx DSP Control Center (V-3.0) with a significance threshold combined treatment analysis. Among cet-treated pts, SNHG11-H showed a numerically longer
of a = 0.05. Results: In two lymphoma pts with stable disease (SD) or partial response mPFS (H: 14.2 vs M: 11.1 vs L: 7.6 mo; p = 0.19) and significantly longer mOS (H: 41.1 vs M: 32.4
(PR), PI3K downstream signaling showed downregulation at C1D8 due to C-mediated PI3K- vs L: 14.3 mo; p = 0.012). In contrast, no statistically significant OS or PFS differences were
AKT signaling inhibition, possibly through PIK3IP1 overexpression. This signaling returned observed in bev-treated pts. SNHG11-H tumors had a significant OS benefit from cet compared to
to baseline by C2D15, likely due to C’s elimination half-life. In a follicular lymphoma case bev (mOS 41.1 vs 36.5 mo, respectively; p = 0.016), with a nominally significant treatment
(#18, PR), FOXP3 expression decreased at both C1D8 and C2D15, while CD4 and CD8 levels interaction observed for OS (p = 0.030). No significant differences were observed in the L or M
remained constant. Immune marker expression (PD-L1, PD-1, CTLA-4, CD80) progressively expression groups. Additional analyses showed that SNHG11 expression was high in the CMS2
(canonical) subtype and substantially lower in CMS1 (immune). Conclusions: LncRNA SNHG11
declined. However, in diffuse large B-cell lymphoma (#14, SD), no changes in T-cell
plays a significant role in CRC progression and metastasis via tumorigenic pathways, including c-
markers were observed. Solid tumor cases (#12, #24, SD) showed PI3K-AKT signaling Myc and HIF-1a. Moreover, elevated circulating SNHG11 levels show promise as a non-invasive
downregulation at C2D15, along with CD3+/CD8+ T-cell infiltration into the tumor. FOXP3 biomarker for early CRC detection. In CALGB/SWOG 80405, high SNHG11 expression correlated
levels slightly decreased in tumor and immune compartments. Progressive disease cases with improved PFS and OS, particularly in cet-treated pts, supporting its role as a prognostic and
showed no change in T-cell markers. Conclusions: GeoMx DSP demonstrated its capa- predictive biomarker. Its strong association with CMS2 aligns with its reported involvement in c-
bility to investigate phospho-signaling and immune profiles in tumor and stromal com- Myc-driven pathways. Further validation is needed to confirm the clinical utility of this biomarker
partments of small biopsies, highlighting its potential to enhance the understanding of and elucidate underlying mechanisms. Research Sponsor: National Cancer Institute;
TMEs in clinical studies. Further applications may provide critical insights for clinical P30CA014089, U10CA180821, U10CA180882, U10CA180888; Genentech; https://
cancer trials. Clinical trial information: NCT03502733. Research Sponsor: None. [Link].

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
220s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3129 Poster Session 3130 Poster Session

Using single-cell sequencing to identify endothelial expression of immune Assessment of homologous recombination deficiency and BRCA status in
checkpoint ligands in advanced hepatocellular carcinoma, pre- and post- ovarian cancer: Analytical performance and relevance of a decentralized
atezolizumab plus bevacizumab in the phase II INTEGRATE study. First NGS assay for comprehensive genomic profiling. First Author: Mohit Gupta,
Author: Florence T.H. Wu, Clinician Investigator Program, University of British Columbia, Thermo Fisher Scientific, South San Francisco, CA
Vancouver, BC & Princess Margaret-UHN, Toronto, ON, Canada Background: Homologous recombination deficiency (HRD) is a complex biomarker with
Background: Atezolizumab (atezo; anti-PD-L1) plus bevacizumab (bev; anti-VEGF-A) became a predictive value in ovarian cancer. Understanding both the causes of HRD, such as
standard treatment for advanced hepatocellular carcinoma (HCC) after demonstrating an overall pathogenic alterations in homologous recombination repair (HRR) genes, and its
survival advantage over sorafenib (inhibitor of VEGFR2 & other kinases) in the phase III clinical trial, consequences, like genomic instability (GI), is crucial for exploring various therapeutic
IMbrave150. However, the mechanisms of primary and acquired resistance to atezo-bev are poorly
understood. VEGFR2+ endothelial cells (ECs) are potential cellular targets of bev and may play a key
strategies, including the potential use of poly (ADP-ribose) polymerase inhibitors
immunomodulatory role in response to atezo-bev. In this study, we utilized single-cell sequencing to (PARPi). This study evaluates the analytical performance and clinical research relevance
identify potential mediators of resistance within EC subsets. Methods: Eight patients with unre- of the Oncomine™ Comprehensive Assay Plus (OCA Plus), a distributable next-generation
sectable HCC were enrolled on the INTEGRATE study, treated with atezo-bev, and underwent intensive sequencing (NGS) research use assay that offers in a single workflow comprehensive
biospecimen collection (NCT04563338). Serial tumor biopsies were collected and viably cryopreserved genomic profiling, including HRD evaluation. Methods: The OCA Plus panel was used for
including pre-treatment (n=6), 21-28 days after first dose (n=6), and at disease progression (n=2). comprehensive genomic profiling of a series of 299 ovarian cancer research samples
Single-cell analysis via cellular indexing of transcriptomes and epitopes (CITEseq) has been performed from the PAOLA-1 trial, part of the ARCAGY biorepository. Research samples were
and data from four patients have been analysed to date. Aggregating their nine biopsies, 8,569
analyzed to assess agreement with orthogonal method, specifically for BRCA1 and
hepatocytes (ALB+ FABP1+ FGB+), 29,072 immune cells (CD45+), and 8,028 ECs (CD31+ vWF+ KDR+)
were annotated. The differential expression of VEGFR2, PD-L1, and other immune checkpoint ligands BRCA2 mutational status, GI status and overall HRD status which combined BRCA1/2
by tumor vs. immune vs. endothelial cellswere interrogated (Table). Results: VEGFR2 (receptor for mutational status and GI status. GI status was determined using Genomic Instability
VEGF-A) is predominantly expressed by ECs, at high prevalence & intensity. PD-L1 and PD-L2 (ligands Metric (GIM), a quantitative method that characterizes unbalanced copy number
of PD-1) are expressed by ECs at low prevalence & intensity. Galectin3 (LAG3 ligand) is widely changes. Progression-free survival (PFS) was retrospectively studied to determine
expressed by hepatocytes, immune cells and ECs; while L-SECtin (LAG3 ligand) is predominantly future clinical relevance. Results: The success rate for DNA sequencing was 100%,
expressed by ECs but at low prevalence & intensity. ECs had the highest prevalence of galectin9 (TIM3 starting from a minimal sample input of 20ng of genomic DNA isolated from FFPE tissue
ligand) expression. Nectin2 (TIGIT ligand) is expressed by both hepatocytes and ECs at high prev- blocks. The OCA Plus panel provided a detailed genomic profile in a single workflow,
alence & intensity. Conclusions: Liver ECs express a broad array of immune checkpoint ligands, which
are more frequent than previously anticipated. These EC subsets may potentially drive resistance by
achieving high success rates across all biomarkers tested, including single nucleotide
contributing to exhaustion of T cell subsets entering the tumor microenvironment. Complete CITEseq, variants/indels and HRD (100%). Overall percent agreement (OPA) for HRD status with
TCR sequencing, and correlative studies from the full cohort are underway. Clinical trial information: orthogonal method was 87%. OPA for BRCA1/2 variants was 98%, while OPA for GI
NCT04563338. Research Sponsor: This research was a collaborative effort made possible through status was 80%. PFS analysis demonstrated a significantly better hazard ratio (HR: 0.51,
support from F. Hoffmann-La Rohce for the imCORE Network. p , .005) for the cases positive for OCA Plus HRD solution compared to the cases
Hepatocytes Immune cells ECs
negative for the OCA Plus HRD solution (HR: 0.84, p = 0.43). Conclusions: The OCA Plus
solution enables robust and reliable comprehensive genomic profiling with high OPA for
VEGFR2 (KDR) 0.06% (,0.01) 0.05% (,0.01) 66% (0.8) BRCA1/2 and HRD status compared to commonly used orthogonal method. Albeit
PD-L1 (CD274) 0.2% (,0.01) 4% (0.03) 2% (0.01)
PD-L2 (PDCD1LG2) 0.02% (,0.01) 2% (0.02) 3% (0.02) additional studies are due, overall, the reported data suggests its future clinical utility in
L-SECtin (CLEC4G) 0% 0.1% (,0.01) 3% (0.05) predicting treatment outcomes in ovarian cancer. Research Sponsor: None.
Galectin-3 (LGALS3) 68% (0.8) 39% (0.5) 43% (0.5)
Galectin-9 (LGALS9) 5% (0.04) 28% (0.3) 34% (0.3)
PVR (CD155) 14% (0.1) 1% (,0.01) 18% (0.1)
Nectin2 (CD112) 60% (0.5) 5% (0.04) 44% (0.4)
% = proportion of cells with positive expression. ( ) = normalized mean expression.

3131 Poster Session 3132 Poster Session

Cost-effectiveness of NTRK testing strategies for detecting NTRK fusions in Impact of sample characteristics on RNA-based next-generation sequenc-
solid tumors in China. First Author: Jian Wang, Fudan University Shanghai Cancer ing (NGS) for fusion gene detection in non-small cell lung cancer (NSCLC).
Center; Shanghai Medical College, Fudan University, Shanghai, China First Author: Jun Liu Jr., First People’s Hospital, the Second Affiliated Hospital of South
Background: Neurotrophic tyrosine receptor kinase (NTRK) gene fusions are oncogenic China University of Technology, Guangzhou, China
drivers in many solid tumors. With the inclusion of the targeted drug Entrectinib in Background: RNA-based next-generation sequencing (NGS) has been widely employed for
China’s national reimbursement drug list, the demand for NTRK testing has also in- detecting fusion genes in NSCLC, due to its superior sensitivity and simplified design
creased. This study evaluates the cost-effectiveness of a two-step testing strategy compared to DNA-based NGS. However, the impact of sample quality on fusion variant
(initial pan-TRK IHC testing followed by next-generation sequencing (NGS) confirmation detection using RNA-based NGS remains unclear. Methods: The study analyzed 5,386 and
for positive results) compared to directly conducting NGS testing, with treatment of 5,538 NSCLC samples using DNA- or RNA-based NGS to detect common fusion genes (ALK,
positive cases using Entrectinib. Methods: A decision tree model was established RET, ROS1, NTRK, NRG1, MET exon 14 skipping, and FGFR). NGS libraries were constructed
from a health system perspective, based on clinical practices in China. The study in- using capture-based or amplicon-based methods for DNA and RNA samples, respectively,
cluded 17 cancer types as reported in the latest clinical trial of Entrectinib (n = 194). focusing on pathogenic mutations. Results: RNA-based NGS detected 2.44% more fusions
Diagnostic performance data were sourced from literature and validated by pathologists than DNA-based NGS [9.50% (526/5538) vs. 7.06% (380/5386)], with notable advantages for
and clinicians. Clinical efficacy, treatment phase costs, and utility for progression-free NTRK (0.13% vs. 0.02%), NRG1 (0.25% vs. 0.06%), MET exon 14 skipping (2.15% vs. 1.36%),
survival (PFS) were obtained from open databases and literature. Turnaround time and and FGFR fusions (0.40% vs. 0.02%). Tumor cell content analysis showed no significant
costs for testing was gathered from expert interviews. The time horizon was set to impact on fusion detection rates within the 20%-90% range for either method. However,
higher tumor cell content ($80%) significantly increased RNA-based NGS detection rates
include the duration of NTRK testing and the period of PFS associated with the
compared to DNA-based NGS, nearly doubling the total detection rate (17.3% vs. 8.88%),
medication. A one-way sensitivity analysis was conducted to assess the model’s ro-
primarily due to increased ALK fusion detection (8.97% vs. 5.02%). The type of sampling
bustness. Results: In China, the NGS testing alone produced 0.55507 life years (LY) and (surgical, biopsy, or others) did not significantly affect overall fusion detection rates for
0.42646 quality-adjusted life years (QALY) at a total cost of $5932.57，whereas the either method (p . 0.05). However, gene-specific analyses showed significantly higher
pan-TRK IHC + NGS testing strategy yielded 0.55530 LY and 0.42665 QALY at a total cost detection rates for ROS1, MET, and RET using RNA-based NGS in biopsy samples compared
of $4176.67. The pan-TRK IHC+NGS testing strategy was dominant, offering higher to DNA-based methods (ROS1: 11.83% vs. 1.18%, MET exon 14 skipping: 2.87% vs. 1.62%,
QALY at lower costs than NGS testing alone. Additionally, the average wait time for pan- RET: 1.24% vs. 0.79%). Conversely, RNA-based detection of ALK and NRG1 fusions was
TRK IHC + NGS testing was reduced by 10 days. The robustness of the base case results higher in surgical samples (ALK: 4.00% vs. 3.25%, NRG1: 0.34% vs. 0.08%) compared to
was confirmed through sensitivity analysis. Conclusions: Initial pan-TRK IHC testing, DNA-based methods. Regarding sample types, pleural/peritoneal effusions showed higher
followed by NGS confirmation for positive results, is the optimal strategy for NTRK detection rates than FFPE samples, though not statistically significant. RNA-based NGS
fusion detection in patients with locally advanced or metastatic solid tumors in China, consistently showed superior detection rates for ALK and MET exon 14 skipping in all
providing superior cost-effectiveness compared to NGS testing alone. Research sample types compared to DNA-based methods, with the most substantial increase for MET
Sponsor: None. exon 14 skipping in pleural/peritoneal effusions (2.14% vs. 0.98%). Conversely, RNA-based
Cost-effectiveness analysis results. NGS for NRG1 and ROS1 fusions showed a greater relative increase in detection rate in 10%
Total Effectiveness Effectiveness Incremental Incremental neutral formalin-fixed tissue/FFPE sections/unstained slides compared to pleural/
Testing Cost(USD$) (LYs) (QALY) cost ($) QALYs ICER (Cost/QALY) peritoneal effusions. Conclusions: Sample characteristics did not significantly impact
NGS $5932.57 0.55507 0.42646 pan-TRK IHC +NGS
the overall detection rate of RNA-based fusion assays. However, detection rates for specific
pan-TRK $4176.67 0.55530 0.42665 -$1755.90 0.000187 was dominant fusions like ALK, NRG1, and MET exon 14 skipping varied with sample type, sampling
IHC+NGS method, and tumor cell content. Optimizing testing strategies and sample handling is crucial
ICER: Incremental cost-effectiveness ratio.
to improving diagnostic accuracy in NSCLC. Research Sponsor: Medical Scientific Research
Foundation of Guangdong Province, China; A2022519.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 221s

3133 Poster Session 3134 Poster Session

Effect of irradiation on the killing effect of NK cells in colon cancer through Combined prognostic value of post-surgery circulating tumor DNA and
MYB/TIM3 axis. First Author: Xiuli Guo, Zhongnan Hospital of Wuhan University, tumor-stroma ratio in patients with stage III colon cancer treated with
Wuhan, HuBei, China adjuvant chemotherapy. First Author: Ingrid Franken, University Medical Center
Background: Natural killer (NK) cells play a crucial role in tumor progression and anti- Utrecht, Department of Medical Oncology, Utrecht, Netherlands
tumor immunity. However, they often exhibit an exhausted phenotype within the tumor Background: Patients with stage III colon cancer (CC) are routinely treated with resection
microenvironment (TME), limiting their full cytotoxic potential. T-cell immunoglobulin followed by adjuvant chemotherapy (ACT). About half of patients are cured by surgery and
and mucin-domain containing-3 (TIM-3) has emerged as a novel immune checkpoint hence overtreated with ACT, yet another ~30% experience recurrence and are currently
that is highly expressed on NK cells and suppresses their cytotoxic function. TIM-3 is undertreated. Only ~20% of patients are cured by ACT and we are unable to identify these
closely associated with immune evasion and anti-tumor immune tolerance. This study patients. Prognostic value of circulating tumor DNA (ctDNA) and the tumor-stroma ratio
aims to investigate the effects and mechanisms by which radiation modulates NK cell (TSR) has been shown in separate studies. This study aimed to integrate these biomarkers
function, providing a foundation for developing strategies that specifically target TIM-3 with pTNM substage to better predict outcome in stage III CC patients treated with ACT.
on NK cells. Methods: First, mRNA high-throughput sequencing, RT-qPCR, and Western Methods: Patients with stage III CC who received radical resection followed by ACT were
blot experiments were used to analyze changes in the expression of related genes in selected from the Prospective Dutch ColoRectal Cancer cohort (PLCRC) substudy PRO-
NK92 cells after radiotherapy. The LDH release assay was employed to evaluate the VENC3 (Rubio-Alarcon AACR 2024). Blood was collected between surgery and ACT, to
effect of radiation on the viability of NK92 cells. ELISA was conducted to detect changes determine ctDNA status using Labcorp Plasma Detect. Based on a diagnostic H&E slide
in the release levels of tumor necrosis factor TNF-a and other factors after radiotherapy. from the CC resection, the TSR was determined by a trained observer according to the
Dual-luciferase reporter assays and chromatin immunoprecipitation (ChIP) experiments United study (Polack ESMO open 2024). A stroma content of #50% was considered low
confirmed that the transcription factor MYB mediates radiation-induced regulation of and .50% high. The primary outcome was recurrence risk (RR), calculated from date of
NK cell activation by targeting and binding to the TIM-3 promoter region. A non-contact resection. Results: In the overall cohort (N = 207), the 3-year RR was 23.4% [17.3-29.1]. In
co-culture system was established, and flow cytometry demonstrated that radiation total, 88 patients (43%) were stroma-high and had a higher recurrence risk (3-year RR
combined with MYB overexpression enhanced the cytotoxicity of NK cells against tumor 33.1% [22.5-42.3]) than the 119 stroma-low patients (3-year RR 16.0% [8.9-22.5]; HR 2.7
cells. A colon cancer mouse model was constructed to evaluate the anti-tumor effect of [1.6-4.6]). CtDNA was detectable after surgery in 28 patients (13.5%; HR 5.8 [3.3-10]), of
whom 11 (39%) were stroma-high. T4/N2 stage was observed in 82 patients (HR 2.9 [1.7-
combining anti-TIM-3 antibodies with radiotherapy. Results: We found that radiation
5.0]), of whom 46 (56%) were stroma-high. TSR (HR 2.6 [1.5-4.6]) had added prognostic
can activate NK92 cells in vitro and enhance TIM-3 expression,promoting the secretion
value to ctDNA (HR 7.6 [4.3-13]) and pTNM substage (HR 2.9 [1.7-5.0]) in a multivariable
of granzyme B, perforin, TNF-a, IFN-g, and other cytokines and chemokines that
cox model (LRT p,0.001). Patients with no detectable ctDNA and stroma-low T1-3N1 CC
modulate the TME and enhance anti-tumor immune responses. Moreover, the tran-
were at low recurrence risk (N = 71; 3-year RR 2.9% [0-6.8]). In comparison, patients with no
scription factor MYB inhibits TIM-3 expression by directly binding to the TIM-3 promoter detectable ctDNA and a tumor that was either stroma-high or T4/N2 were considered
region, mediating the effects of radiation on the TME through NK cell activation. In vivo, intermediate risk (N = 68; 3-year RR 17.2% [7.4-26.0]; HR 5.4 [1.5-19]). Patients with
the combination of radiotherapy and anti-TIM-3 antibodies effectively controlled the detectable ctDNA and/or stroma-high T4/N2 CC had a high risk (N = 68; 3-year RR 50.2%
growth of subcutaneously transplanted colon cancer tumors in C57BL/6 mice. However, [36.7-60.8]; HR 19 [5.9-62]). Conclusions: The tumor-stroma ratio has added value to post-
this combined treatment effect was significantly diminished after NK cells were depleted surgery ctDNA and pTNM substage in predicting outcome in stage III CC patients treated
by the anti-NK1.1 antibody. Conclusions: This study elucidates a novel mechanism by with ACT. The recurrence risk in the third of patients with no detectable ctDNA and stroma-
which radiation activates NK cells in the tumor microenvironment through the MYB/TIM- low T1-3N1 CC was only 3%. It is of interest to investigate whether this low risk would
3 pathway. It provides new insights for enhancing the efficacy of radiotherapy and persist in a cohort treated with surgery only, to suggest whether these patients could be
offers a theoretical basis for the potential clinical application of these cells in future spared ACT in the future. The third of patients with detectable ctDNA, and/or stroma-high
research. Research Sponsor: None. T4/N2 CC, had a 50% recurrence risk despite ACT, highlighting the need for alternative
adjuvant treatment options for these patients. Research Sponsor: None.

3135 Poster Session 3136 Poster Session

Regorafenib response prediction in metastatic colorectal cancer by a novel Relationship between FOLR1 expression and pan-cancer subgroup of tu-
genomic and transcriptomic model. First Author: Andreas Seeber, Department of mors with specific transcriptomic profile. First Author: Andrew Ip, John Theurer
Hematology and Oncology, Comprehensive Cancer Center Innsbruck, Medical University Cancer Center, Hackensack University Medical Center, Hackensack, NJ
of Innsbruck, Innsbruck, Austria Background: Mirvetuximab soravtansine is an antibody–drug conjugate currently approved
Background: The multi-kinase inhibitor regorafenib (Rego) is approved for the treat- for the treatment of advanced platinum-resistant ovarian cancer. The efficacy of this therapy
ment of refractory metastatic colorectal cancer (CRC). However, its efficacy is limited, is correlated with high expression of folate receptor alpha (FRa), encoded by the FOLR1 gene.
and its use is frequently associated with substantial toxicities. Identifying biomarkers Treatment requires $ 75% of tumor cells to be stained positively for FRa by immunohis-
predicting Rego-response could improve therapeutic outcomes and reduce unnecessary tochemistry (IHC). Here we defined the FOLR1 RNA level that distinguishes ovarian cancers
treatment-related adverse effects in non-responders. Methods: A predictive model for with very high FRa ($ 75% by IHC), then compared the transcriptomic profile of these cases
Rego-response was developed based on transcriptomic and genetic data from 41 CRC (FOLR1-H) with the transcriptomic profile of cases with very low FRa expression (FOLR1-L).
cell lines. Cell lines were classified into Rego-sensitive versus -resistant groups based on We further explored the presence similar FOLR1-H signature in various types of cancers.
Methods: RNA was extracted from 1450 solid tumor FFPE samples and sequenced using a
drug sensitivity data from the CTRP2 database. Several machine-learning algorithms
targeted RNA panel of 1600 genes. The RNA expression levels of various genes were
were evaluated, with the Generalized Linear Model via Elastic Net (GLMNET) achieving
quantified and expressed as transcript per million (TPM). IHC for FRa protein was performed
the highest predictive performance. Model accuracy was assessed using leave-one-out on ovarian cancers (N = 49) using VENTANA FOLR1 RxDx assay. Results: Based on
cross-validation. Further validation was performed using transcriptomic (WTS) data comparing IHC with RNA expression of FOLR1, FOLR1-H samples were defined with RNA $
from 24,384 real-world CRC patients assessed by Caris Life Sciences, which included 300 TPM while FOLR1 mRNA , 100 was correlated with very low FRa by IHC and classified as
720 patients treated with Rego. Results: The predictive model identified key cell line FOLR1-L. Of the 312 ovarian cancers, 21% were classified as FOLR1-H and showed sig-
features associated with Rego-response, including gene expression signatures (e.g., nificantly (Log10FDR , -2) higher expression in 39 genes as compared with FOLR1-L. The
ZNF441, CCDC82, ZFP69) and specific mutations (e.g., RALGAPA1, MORC1). Tran- Log10FDR was , -10 in 19 genes. The top highly expressed genes in FOLR1-H cases were
scriptome profiling showed that Rego responders exhibited enrichment in cell-cycle TROP2, NECTIN4, ROR1, ROR2, ACVRL1, and NTHL1. In breast cases (N = 199) FOLR1-H was
regulation and DNA-repair mechanisms, while non-responders showed a stroma-rich detected in 14.6% of cases and the most highly expressed genes were ACVRL1, NECTIN4,
microenvironment with significant endothelial and fibroblast infiltration. External val- ROR1, and ACVRL1 (Log10FDR , -5). Of the 932 cases of lung cancer, 21.5% classified as
idation using WTS data from real-world Rego-treated CRC patients revealed that pre- FOLR1-H and had significantly (Log10FDR , -2) high expression of 137 genes, but similar to
dicted responders had a prolonged time-on-treatment (p = 0.02, HR = 0.79) and median ovarian cancer TROP2, NECTIN4, ROR1, ROR2, ACVRL1, and NTHL1 were top expressed
overall survival (p = 0.01, HR = 0.76) compared to predicted non-responders. This genes. Of the 174 pancreatic cancers 9.8% were FOLR1-H and top expressed genes were
association was specific to Rego-response, as there was no survival difference between NECTIN4, ROR1, and ACVRL1. In sarcoma (N = 166), 8.4% had FOLR1-H and only three genes
predicted responders and non-responders among patients not treated with Rego (p = (NECTIN4, NTHL1 and SLC47A1) were significantly high. Of the 327 colorectal cancers, 8%
0.72, HR = 1.0). Conclusions: This novel predictive model successfully identified and met the criteria for FOLR1-H and 16 genes were significantly higher in FOLR1-H including
validated molecular features associated with Rego-response in CRC. The transcriptomic NECTIN4, NTHL1, ACVRL1, ROR1/2. In 64 esophageal cancers 10.9% were FOLR1-H, but only
and genetic signature holds significant potential for improving personalized treatment 3 genes (GALNT12, ACVRL1 and NECTIN4) were significantly higher. Conclusions: This data
strategies by identifying patients most likely to benefit from Rego and prevents un- suggests that cancers with significantly high expression of FOLR1 mRNA are a special
necessary Rego-associated toxicities in non-responders. Research Sponsor: None. subtype of tumors characterized by the expression of embryonic cell surface markers (FOLR1,
TROP2, NECTIN4, ROR1/2). The pan-cancer marked overexpression of these genes suggests
that cancers with FOLR1-H represent a subtype of cancers with similar biology. This subtype
may benefit from combination therapy targeting more than one of these markers (e.g. anti-
FOLR1 with anti-TROP2, or anti-NECTIN4) and clinical trial with such combination may be
justified. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
222s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3137 Poster Session 3138 Poster Session

Multi-omics cohort-based prediction model for early relapse of hepatocel- Comprehensive clinicogenomic profiling of signet ring cell carcinoma
lular carcinoma post-surgery. First Author: Penghong Song, The First Affiliated across multiple organ sites. First Author: Lawrence Wen Wu, Columbia Univer-
Hospital, School of Medicine, Zhejiang University, Hangzhou, China sity Irving Medical Center, New York, NY
Background: Postoperative early relapse (PER) of hepatocellular carcinoma (HCC) Background: Signet ring cell carcinoma (SRCC) is a rare, aggressive histological subtype of ade-
presents significant challenges in clinical management. Identifying reliable predictive nocarcinoma that is associated with earlier age of onset and poor prognosis. It most commonly arises
markers and therapeutic strategies for PER is crucial for improving patient outcomes. from the stomach but can originate elsewhere. Few studies have compared molecular alterations in
SRCC across various primary sites. Utilizing the American Association for Cancer Research (AACR)
Methods: We constructed a predictive model for PER using multi-omics data from 177 Project Genomics Evidence Neoplasia Information Exchange (GENIE) v17.0, we performed a com-
HCC patients with follow-up information. Transcriptomic and proteomic profiling of HCC prehensive analysis of clinicogenomic variables in SRCC across different primary sites. Methods: The
tissues was performed, followed by differential expression analysis and Weighted Gene AACR GENIE v17.0 database was used to select tumor samples classified as SRCC by the Oncotree
Co-expression Network Analysis (WGCNA) to identify molecular markers associated with Code. We excluded primary SRCC sites with , 10 tumor samples. Samples were analyzed for
PER. Univariate, LASSO, and multivariate Cox regression analyses were employed to clinicogenomic characteristics including gender, race, ethnicity, age at sequencing, and oncogenic
refine the marker set, resulting in a three-gene signature. The model’s accuracy was molecular alterations by OncoKB classification (somatic mutations, structural variants, copy number
validated using a proteomic cohort and The Cancer Genome Atlas (TCGA) database. alterations). We classified “early onset” SRCC as age of sequencing , 50. Chi-square testing was used
to compare categorical variables, and Benjamini-Hochberg procedure was used to control the false
Functional enrichment, drug sensitivity, and immune infiltration analyses were con-
discovery rate (statistical significance for q , 0.05). Results: From 355 patients with SRCC, 358
ducted to explore the biological characteristics and therapeutic implications of high-PER tumor samples were analyzed, with the following distribution among primary sites: stomach (n = 168),
risk patients. Patient-derived organoid (PDO) models were used for further validation. colon/rectum (n = 125), appendix (n = 38), and bladder (n = 27). There were high rates of early onset
Results: We identified 31 molecular markers associated with PER, which were narrowed SRCC in the stomach (29.2%), colon/rectum (47.2%), and appendix (36.8%). Female gender was
down to a robust three-gene signature (MIK67, GPD1, and MBL2) with an area under the numerically higher in stomach (56.9%) and appendix (55.3%) SRCC cases compared to colon/rectum
curve (AUC) of 0.868 for predicting early relapse. Functional enrichment analysis (47.2%) and bladder (33.3%) SRCC cases. The most prevalent altered genes included TP53 (45.0%),
revealed that high-PER risk patients exhibited enhanced DNA damage repair and cell CDH1 (19.4%), ARID1A (14.8%), KRAS (12.9%), and SMAD4 (12.3%). There was differential enrichment
cycle pathways. Drug sensitivity analysis suggested potential benefits from gemcitabine of molecular alterations across various sites in TP53, CDH1, KRAS, SMAD4, TERT, APC, and BRAF (q ,
0.05). Conclusions: To our knowledge, this study represents the largest molecular analysis of SRCC
and paclitaxel, which were validated using PDO models. Immune infiltration analysis across multiple organ sites, revealing high rates of early onset SRCC and distinctive molecular al-
showed reduced NK cell and M2 macrophage infiltration in high-PER risk patients, teration patterns. These findings underscore the further need to investigate functional implications
confirmed by single-cell sequencing and immunohistochemical validation. and potential therapeutic targets for site-specific molecular alterations in SRCC. Research Sponsor:
Conclusions: This study provides a novel multi-omics-based predictive model for early U.S. National Institutes of Health; 5T32CA203703-09.
recurrence in HCC, highlighting potential therapeutic options for high-risk patients. The Highlighted clinicogenomic features.
findings underscore the importance of DNA damage repair and cell cycle pathways in Colon/
PER and suggest targeted therapies that could improve clinical outcomes for high-PER Clinical Variable/ Alteration Total Stomach Rectum Appendix Bladder q-value
risk patients. Research Sponsor: None. Age of Sequencing <50 (%) 34.4% 29.2% 47.2% 36.8% 3.7% ,0.001
Female Gender (%) 51.5% 56.9% 47.2% 55.3% 33.3% 0.24
TP53 45.0% 40.5% 48% 31.6% 77.8% 0.011
CDH1* 19.4% 26.2% 3.4% 7.9% 63% ,0.001
KRAS 12.9% 8.3% 17.6% 26.3% 0% 0.0055
SMAD4* 12.3% 3.6% 20.3% 21.0% 11.1% ,0.001
TERT* 8.7% 1.5% 7.1% 0% 66.7% ,0.001
APC 5.7% 1.2% 15.3% 0% 0% ,0.001
BRAF 3.9% 1.2% 8.8% 2.6% 0% 0.017
*Not all samples profiled for specific alteration; % reflects percentage of samples with alterations of those
profiled.

3139 Poster Session 3140 Poster Session

Comparative analysis of T-cell subsets and vessel features in matched Characterisation of ductal carcinoma in situ (DCIS) using mass spectrom-
primary colorectal tumors and corresponding resected liver metastases. etry imaging towards near realtime margin assessment. First Author: Hemali
First Author: Pia J. Osterlund, Tampere University Hospital and Tampere Univeristy, Chauhan, Imperial College, London, London, United Kingdom
Tampere, Finland Background: Imprecision in breast-conserving surgery leads to high national average high rates
Background: Outcome after liver resection for colorectal cancer metastases (CRLM) is partly of reoperative intervention. In line with updated margin guidelines, accurate differentiation
determined by factors such as the number, size, and vitality of the metastases, as well as the T- between non-invasive and invasive breast cancer is essential. This study aimed to assess
and N-stage of the primary tumor. The tumor microenvironment—particularly immune cell whether mass spectrometry can distinguish between normal breast tissue, benign, non-invasive
infiltration and vascular features—also influences outcome. Data on how these factors compare and invasive disease towards the development of an intraoperative margin assessment tool.
between paired primary colorectal tumors and matched CRLM are limited. Methods: We used Methods: Breast tissue samples were collected from patients undergoing mastectomy.
TMAs from matched primary tumors and CRLM samples of 50 patients, of which 15 were Samples were flash-frozen, sectioned, and analysed using a Xevo G2-XS QTof mass spec-
untreated and 35 had received neoadjuvant therapy (cytotoxic 6 VEGF-/EGFR-targeted agents) trometer (Waters Corp.). Selected sections were ionised using a pulsed optical parametric
before liver resection. Each tumor consisted of 1–3 tissue cores (1 mm) from both the tumor oscillator laser (OpoletteTM 2731/3034, OPOTEK) which operated at 2940 nm wavelength and
center and the invasive margin. Multiplex immunofluorescence was performed to assess T-cell 20 Hz repetition rate. The laser focused on the tissue through a 20 mm focal distance convex
(e.g., CD3, CD4, CD8, PD1, FOXP3, TIM3, Ki67), and vessel (claudin-5, aSMA, PDGFRb) markers. lens generating aerosol which was aspirated into the spectrometer. The data was combined
Nonparametric Wilcoxon signed rank and Spearman correlations were used for cell density using spatial distribution and chemical information from characteristic ions to generate 2D
comparisons. Cox regression for continuous variables was used for disease-free survival (DFS) chemical images and labelled using consecutive H&E-stained sections annotated by a Con-
associations. Results: We observed significantly lower densities of CD3+CD8+Ki67+, sultant Histopathologist for ground truth cross-validation. Results: Over 1 million mass spectra
CD3+CD4+Ki67+, and CD3+CD4+FOXP3+ cells in CRLM than in paired primary (all p , .006) in were collected from imaging 52 breast tissue sections. This includes 720 mass spectra from 31
both untreated and pretreated cohorts. The aSMA+PDGFRb– vessel subset was more prevalent DCIS breast tissue sections, compared to 6 spectra from 2 DCIS breast tissue samples in
in CRLM compared with the primary tumor in the untreated cohort (p , .001). In the untreated previous work. A pixel size of 50 mm and scan rate was 250 mm/s was utilised. An ex-vivo
cohort, larger vessel size in CRLM (but not in the primary tumor) showed a positive correlation classification model was built using n=6,796 and achieved .99% sensitivity for tumour de-
with CD3+CD8+PD1+, CD3+CD4+PD1+, and CD3+CD4+FOXP3+ densities (Spearman r = .54–.60, tection (DCIS and IBC) and 100% specificity for identifying normal tissue. Principal Component
p = .02–.04). In the pretreated cohort, higher tumor vitality and/or CDX2+ expression in CRLM Analysis demonstrated accurate separation of IBC, DCIS, benign breast disease, and normal
(indicative of poor treatment response) were each negatively correlated with cytotoxic breast tissue. Six possible metabolites were identified following Recursive Feature Elimination
(CD3+CD8+) and helper T-cell (CD3+CD4+) subsets (r = -.42 to -.63, p , .01). DFS after (RFE) was used to identify the most significant features which differentiate the tissue types,
metastasectomy was associated with vessel and T-cell features. Regarding vessel metrics in the these were annotated using the Lipid Maps database ([Link] (Table 1).
small untreated cohort, aSMA–PDGFRb– vessel subset in primary showed a negative trend (p = Cancerous tissue showed higher levels of structural lipids (600-900 Da), while normal/benign
.06) as did smaller vessel size in CRLM (p = .06). CD3–CD4+TIM3+ in CRLM was negatively breast tissue had higher levels of small metabolites (50-300 Da) and fatty acids (200-400 Da).
associated with DFS in pretreated (p = .04), with a trend also in untreated (p = .10). Conclusions: Mass spectrometry imaging enables accurate differentiation of IBC, DCIS, benign
Conclusions: Densities of certain T-cell subsets are significantly lower in matched CRLM than in breast disease, and normal breast tissue. Research Sponsor: NIHR Imperial Biomedical
primary tumors indicating immune desert phenotype. Vessel subset profiling suggests dif- Research Centre.
ferences between primary tumors and CRLM, possibly relevant for treatment response. Poor Biological features identified from the most significant RFE selected features.
pretreatment effect, i.e., high vitality and CDX2+ density in CRLM, was negatively correlated with
m/z value Annotation Delta Theoretical m/z Ion Class
several T-cell subsets, a correlation not seen in untreated. The poor prognosis association of
CD3–CD4+TIM3+ cells in CRLM merits further investigation. Research Sponsor: Finska 255.2324 Palmitic acid 0.0006 255.2330 M-H Fatty acid
297.2751 FA 19:0 0.0037 297.2799 M-H Fatty acid
Läkaresällskapet; Sigrid Juselius Stiftelse; Medicinska understödsföreningen Liv & Hälsa; The 307.2019 FA 16:0 0.0026 307.2046 M+Cl Fatty acid
Finnish Cancer Foundation; The Competitive State Research Financing of the Expert Re- 766.5392 PE 38:4 0 766.5392 M-H PE
sponsibility Area of Tampere and Helsinki; Tampere University Hospital Fund; Mary and Georg C. 843.5053 PI 35:4 0.0025 843.5029 M-H PI
Ehrnrooth Foundation; Radiumhemmets fonder; Cancerfonden; Suomen onkologiayhdistys. 891.7444 TG 52:3 0.0003 891.7447 M+Cl TG

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 223s

3141 Poster Session 3142 Poster Session

A minimal comprehensive somatic panel to aid clinical decision making in a Clinical performance of Signatera Genome assay in a cohort of patients (pts)
low cost setting. First Author: Urvashi Bahadur, Strand Life Sciences, Bengaluru, India with solid tumors. First Author: Mridula Annette George, Rutgers Cancer Institute of
Background: Large next-generation sequencing (NGS) panels (. 300 genes) offer New Jersey, New Brunswick, NJ
multiple potential therapeutic options for patients with metastatic cancer. However a Background: Circulating tumor DNA (ctDNA) has emerged as a powerful, minimally invasive
large portion of the population in developing countries is unable to avail the benefits of biomarker of treatment response and pt prognosis. Signatera, a tumor-informed, mPCR-NGS
such testing due to limited availability of many drugs or suitable clinical trials, coupled ctDNA assay, offers high sensitivity and specificity for detecting molecular residual disease
with the high cost of these tests. There is an urgent need to offer a compact, affordable (MRD). Signatera (exome) uses its proprietary approach to select a highly curated set of tumor
and robust testing solution which can offer expanded but feasible therapeutic options. variants, followed by deep sequencing of plasma libraries at .100,000x per variant. Signatera
Hence we decided to develop a custom mid-sized panel to fulfil these unmet re- Genome uses the same proven technology and may provide an advantage over exome in
quirements. Methods: A targeted solid tumor (DNA + RNA) panel comprising 74 genes certain cases. However, beyond analytical improvements, it remains unclear if Signatera
Genome provides superior performance and utility compared to the clinically validated exome-
(SA74) was designed to cover all genes with a Tier 1 drug recommendation for therapy,
based version in the clinical setting. In this study, we assessed the clinical performance of the
evaluating single nucleotide variants (SNV), indels, copy number variants (CNV) and Signatera Genome assay in a cohort of pts with solid tumors. Methods: We performed a
gene fusions (GF). We also included certain Tier2 genes for prognosis or added clinical retrospective analysis of clinically annotated residual pt samples from commercial ctDNA
impact. GFs were evaluated by RNA. Inferior sample quality often results in poor quality testing (Signatera, exome-based, 16-plex mPCR-NGS assay). Adjuvant treatment decisions
data, so we added a DNA component for tiling select intronic regions to identify GF which and ctDNA-cadence of testing were at the provider’s discretion. Signatera Genome assays
to be used when RNA could not be analyzed. The analytical sensitivity was . 99% for were designed, consisting of 64 high-quality variants, from the respective pts’ matched tumor
SNV/Indels with a 5% limit of detection, . 99% for CNV and GF. The clinical sensitivity and normal whole genome sequencing data. These assays were used to detect ctDNA in the
was 100% for SNV, 95% for indels, 84.2% for CNV, 100% for RNA GF and 70% for DNA GF. associated pts’ plasma utilizing a sample calling strategy that combined the target confi-
Results: 239 formalin-fixed paraffin embedded tumor samples were evaluated using dences and sample-level noise into a final confidence score. ctDNA concentration was
SA74 and the data was scored for actionability across tumor types. This was compared measured in mean tumor molecules per mL of plasma (MTM/mL). Longitudinal plasma
with data from 706 samples across multiple cancers analyzed using the Illumina TSO500 samples represented postoperative time points until recurrence/end of follow-up. The cor-
panel. The average number of actionable alterations was 1.1 in SA74 and 1.6 in TSO500. relation between any time postsurgical ctDNA positivity and recurrence-free survival (RFS) was
The overall actionability (cases with at least one Tier1 or Tier2 actionable variant) of assessed using Cox regression analysis. Results: The Signatera Genome assay achieved a
SA74 was 63.4%; while that of TSO500 was 78.3%. Of this, the overlap with SA74 was high analytical sample-level specificity of 99.8% (healthy subjects). Clinical performance was
73.7%. The actionability in the remainder was due to Tier2 genes with lesser evidence assessed in a real-world cohort of .300 pts with several cancer types, including breast cancer,
altering the same pathway as an approved drug target or targeting investigational drugs. non-small cell lung cancer (NSCLC), melanoma, and renal cell carcinoma (RCC). Among pts
with relapse, the Signatera Genome assay detected ctDNA ahead of clinical recurrence as
The actionability for each cancer type was calculated and found to be: colon (34.1%; n =
confirmed by imaging. Pts with postsurgical ctDNA-positivity demonstrated significantly
44), non-small cell lung (NSCLC) (78.4%; n = 37), breast (91.7%; n = 24), carcinoma of inferior RFS compared to ctDNA-negative pts. This trend was consistent across all cancer
unknown primary (CUP) (45%; n = 20), uterine (53.3%; n = 15), gallbladder (75%; n = 12) types investigated. Multivariate analysis adjusted for tumor type and stage revealed ctDNA-
and sarcoma (40%; n = 10), among others. Smaller panels often do not include CNV and positivity to be the most significant prognostic factor associated with RFS. Performance
GF. Addition of these variant types increased actionability across cancers. The lack of metrics by cancer type will be presented. Conclusions: Here we report the largest Signatera
CNV and GF would have decreased total actionability from 62.9% to 54.5%. NSCLC and Genome ctDNA study to date across multiple solid tumor histologies. The data indicate robust
CUP were most impacted with a difference in actionability of 16.2% and 15% re- performance and concordance with Signatera Exome. Prospective clinical trials are underway
spectively. GF increased actionability mainly in NSCLC, while CNV contributed to in- evaluating clinical utility. Research Sponsor: None.
creases across all cancers. Conclusions: SA74 demonstrated high actionability across
cancers. It therefore presents a practical alternative to large panel testing by optimizing
actionability and affordability, useful in a cost-sensitive setting. Research Sponsor:
None.

3143 Poster Session 3144 Poster Session

Landscape of genomic alterations in genes implicated in the regulation of Analytical validation of EPISEEK, an epigenomic blood-based assay for
hypoxia inducible factor (HIF) signaling: A pooled analysis of two pan- multicancer detection. First Author: Thi Hanh Pham, Precision Epigenomics Inc,
cancer cohorts. First Author: Wassim Daoud Khatoun, Dana-Farber Cancer Institute, Tucson, AZ
Boston, MA Background: Early cancer detection significantly improves treatment outcomes and
Background: HIF2 alpha inhibitors (HIF2i) have been recently approved for VHL-related tumors, survival rates. However, cancer screening faces key challenges: (1) current tests detect
especially clear cell renal cell carcinoma (ccRCC). Alterations in genes that regulate HIF signaling may only 14% of new cases and cover a limited range of cancers, (2) each cancer requires its
modulate response to HIF2i and may help identify tumor types that could benefit from new therapeutic own costly and complex screening process, (3) limited patient awareness of suitable
approaches targeting HIF activity. Here, we describe the landscape of these alterations across screenings, and (4) poor adoption among marginalized and underinsured groups. EPISEEK
different cancer types. Methods: This study included patients (pts) from the TCGA PanCancer Atlas
was developed for multicancer detection using minimal cell-free DNA from plasma. Here,
and GENIE Cohort v17.0, across all cancer types. All pts underwent Next Generation Sequencing or
OncoPanel analysis of their tumors. Alterations in genes related to the HIF pathway (VHL, EPAS1, we present validation study results assessing its robustness and accuracy across 20+
EGLN1, and EGLN2) and citric acid cycle (SDHA, SDHB, SDHC, SDHD, SDHAF1, FH, IDH1, IDH2, and cancer types and stages. Methods: EPISEEK is a cfDNA-based methylation assay opti-
MDH2) were screened, and only driver mutations, identified by OncoKB, CIViC Variants, My Cancer mized for 20 ng of cfDNA input from plasma. Plasma cfDNA underwent bisulfite conversion
Genome, and the available literature, were included. Pts with a mutation in at least one gene were followed by methylation-specific quantitative PCR targeting 10 cancer biomarkers and 3
included. Mutation frequencies across different cancer types were analyzed for each cohort, then internal control markers. 251 plasma samples from four cancer stages across 25 cancer
pooled together. Results: We identified a total of 10,953 and 167,073 pts on TCGA and GENIE, types, and 57 samples from individuals over 40 with no known cancer history, were used to
respectively. Mutation frequencies for different cancer types are shown in the Table. Among the establish reference ranges and assess assay specificity and sensitivity. Additional con-
mutated cases, IDH1 was the most altered gene in glioma (90.20% [95% Confidence Interval (CI): trived and clinical samples were used to determine the assay’s analytical LOD, repro-
90.19; 90.21]), hepatobiliary cancers (75.48% [75.40; 75.56]) and melanoma (55.75% [55.57; 55.92]),
ducibility and stability. Results: 57 non-cancer samples were used to train the classifier,
and the second most common in leukemia (39.86% [39.68; 40.04]), while IDH2 was the most altered in
leukemia (57.52% [57.34; 57.70]), and the second most common in glioma (5.25% [5.24; 5.25]), and achieving 99% specificity at a 95% confidence level. Accuracy testing included 251 cancer
hepatobiliary cancers (19.10% [19.04; 19.17]). SDHA was the second most altered gene in melanoma samples representing . 20 primary cancer sites, including lung, colon, cervix, esophagus,
(12.95% [12.87; 13.03]). In RCC, VHL was the most mutated gene (92.79% [92.78; 92.79]), followed by head and neck, kidney, liver, breast, bladder, skin, testis, thyroid, ovary, pancreas, prostate,
FH (2.41% [2.40; 2.41]). EPAS1 was mutated in 64.19% of cases with driver mutations in miscellaneous stomach, brain, bone marrow, and others. The cancer samples spanned all stages: I (29%),
neuroepithelial tumors (MNET) ([57.07; 71.31]), but not in RCC. Among the altered cases in pheo- II (13%), III (29%), IV (22%), and cases with missing stage data (6.7%). Sensitivity increased
chromocytoma, SDHB was the most commonly altered gene (39.71% [35.31; 44.11], followed by VHL with advancing stage, with observed sensitivity rate of 52%. By stage, observed sensitivity
(18.41% [15.92; 20.91]). SDHB was also the second most common gene to be mutated in MNET was stage I: 42%, stage II: 46%, stage III: 57%, and stage IV: 64%. Due to the limited and
(23.12% [19.37; 26.86]). Conclusions: In this analysis, mutations in HIF-regulating genes were de- non-representative sample distribution for a typical multicancer screening population,
tected in multiple cancers, and were not limited to those studied in the context of HIF inhibitors. SEER data were utilized to estimate EPISEEK’s real-world performance by adjusting tumor
Further research is required to elucidate whether these gene alterations sensitize tumors to HIF
inhibition. Research Sponsor: None.
incidence and stage when estimating positive predictive value and negative predictive
value. At 99% specificity and based on adjusted performance, EPISEEK achieved a positive
Pooled mutation frequencies for all considered genes in different cancer types. predictive value (PPV) of 40% and a negative predictive value (NPV) of 99%. Seven markers
Cancer Type N % Pooled Mutation Frequency of any gene 95% CI were detectable with , 0.1 ng DNA, while the remaining three markers had an LOD95 of
RCC 3662 39.76 39.74; 39.77 0.1–0.37 ng. Comparing Ct values of each cancer target both intra and inter runs, EPISEEK
Glioma 12657 23.67 23.67; 23.67 demonstrated high reproducibility with standard deviation of 0.383 (high positive), 0.232
Leukemia 1857 15.49 15.47; 15.50 (low positive) and 1.063 (negative samples). Conclusions: EPISEEK is a sensitive, specific,
Hepatobiliary Cancer 4234 11.18 11.18; 11.19
MNET 253 8.61 8.54; 8.68 accurate, and reproducible multicancer detection test. Compared to comprehensive ge-
Pheochromocytoma 198 5.63 5.56; 5.69 nomic profiling techniques, it offers an affordable testing option for broader populations
Melanoma 6589 4.34 4.34; 4.34 with a fast turnaround time. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
224s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3145 Poster Session 3146 Poster Session

Small nucleolar RNAs (snoRNAs) expression and effects on patient (pt) Impact of MGMT methylation on overall survival in solid tumors: A system-
outcomes in metastatic colorectal cancer (mCRC): Data from CALGB atic review and meta-analysis. First Author: Abdulla Alzibdeh, King Hussein Cancer
(Alliance)/SWOG 80405. First Author: Francesca Battaglin, Division of Medical Center, Amman, Jordan
Oncology, Norris Comprehensive Cancer Center, Keck School of Medicine, University of Background: The protein O6-alkylguanine-DNA-alkyltransferase (AGT), encoded by the
Southern California, Los Angeles, CA MGMT gene, plays a crucial role in DNA repair by singularly removing alkyl lesions from
Background: SnoRNAs are non-coding RNAs that primarily guide the chemical modification of the O6 position of guanine, maintaining genomic stability. Loss of MGMT expression,
ribosomal RNA. Emerging evidence suggests snoRNAs play critical roles in cancer, including often due to promoter methylation, is linked to enhanced sensitivity to chemotherapy.
CRC, by regulating cell proliferation, apoptosis and tumor progression. Aberrant snoRNA ex- While MGMT methylation has been observed in various cancers, its impact on overall
pression has been linked to CRC development and poor prognosis, offering potential as di- survival (OS) in solid tumors remains uncertain. Methods: According to PRISMA
agnostic biomarkers and therapeutic targets. We investigated whether the tumor expression guidelines, we selected studies from PubMed that examined the impact of MGMT
levels of 3 types of snoRNAs (SCARNA, SNORA, SNORD) affect treatment response in pts methylation on OS in adult patients with solid tumors. Data were extracted where MGMT
enrolled in CALGB/SWOG 80405 (NCT00265850). Methods: 433 mCRC pts treated with bev- methylation status was clearly defined, and OS was reported through hazard ratios (HR)
acizumab (bev, n = 226) or cetuximab (cet, n = 207) in combination with first-line chemotherapy
from either uni- or multivariable analyses. We employed R version 4.4.2 and the ‘meta’
were analyzed. RNA was isolated from FFPE tumor samples and sequenced on the HiSeq 2500
(Illumina). 422 snoRNAs were evaluated (23 SCARNA, 140 SNORA, 259 SNORD). Overall survival
package for our meta-analysis, using both fixed-effects (Mantel-Haenszel method) and
(OS) and progression-free survival (PFS) were compared across tertiles of gene expression (high random-effects (DerSimonian and Laird’s method) models based on the I2 statistic for
[H], medium [M], low [L]) by multivariable Cox proportional hazards models adjusted for age, sex, heterogeneity. Subgroup analyses were conducted by cancer type, and publication bias
ECOG performance status, tumor side, number of metastatic sites, KRAS, CMS subtypes, and was assessed through funnel plot inspection and Egger’s regression. Statistical sig-
treatment. Interaction tests for the predictive effect (bev vs cet) were performed. P-values were nificance was set at p , 0.05. Results: The meta-analysis included 23 studies, with a
corrected for multiple testing using the Benjamini-Hochberg approach (q , 0.05). Results: Only total of 3,410 participants across all studies. The studies included an array of cancers, the
SCARNA21 achieved statistical significance for OS after false discovery rate (FDR) adjustment. most common being colorectal (n = 7), then head and neck (n = 6), and lesser-represented
Tumors with H levels of SCARNA21 had shorter survival compared to tumors with M or L groups like pancreatic neuroendocrine (n = 2) and others. The pooled analysis using a
expression (median OS 24.4 vs 32.4 vs 33.9 months, respectively; P = 0.0015, q = 0.033), random-effects model demonstrated that MGMT methylation status was not significantly
independent of treatment. No snoRNAs achieved FDR significance for PFS. However, several related with OS (HR of 1.1967; 95% CI: 0.9004 to 1.5904; p = 0.2040). Subgroup analysis
snoRNAs showed significant treatment interactions with biologic agents. SCARNA6-H and revealed that the impact of MGMT methylation on survival varied significantly across
SCARNA5-H tumors had longer PFS and OS when treated with cet, but shorter PFS and OS when different types of cancer. No significant association was yielded between MGMT
treated with bev, compared to the M and L expression groups (PFS interaction q = 0.0067 and methylation and OS for colorectal cancer (HR of 0.9496; 95% CI: 0.6252 to 1.4422), head
0.045, respectively; OS interaction q = 0.022 and 0.034, respectively). SCARNA7 also showed and neck cancer (HR of 1.1520; 95% CI: 0.8223 to 1.6137), NSCLC (HR of 1.0479; 95% CI:
significant treatment interaction for OS, favoring cet in the H expression group (q = 0.018); the 0.3343 to 3.2841) and pancreatic neuroendocrine cancer (HR of 1.5541; 95% CI: 0.6493 to
opposite was observed for SNORA63B, SNORA63D, SNORA35B, and SNORA36C (q = 0.027,
3.7195). Conversely, a significant association was yielded for less common cancers,
0.027, 0.027 and 0.047, respectively). No significant results were observed for any of the tested
SNORDs. Conclusions: SnoRNAs dysregulation affects key pathways such as cell cycle control
including melanoma, biliary and cervical cancers. The funnel plot and Egger’s test for
and immune evasion, making them promising players in CRC biology. Our study highlights the publication bias (t = -0.3999, p = 0.6933) suggested no significant asymmetry, indicating
prognostic and predictive potential of specific snoRNAs in mCRC. Notably, high SCARNA21 minimal publication bias within this meta-analysis. Conclusions: Our findings indicate
expression was linked to shorter OS, while SCARNA5 and 6 showed predictive value for that MGMT methylation does not universally predict OS across all solid tumors. The
treatment response, indicating their potential for guiding treatment decisions. Further validation variability in survival impact across different cancer types suggests that the prognostic
is needed to confirm these findings, and mechanistic studies are warranted. Research Sponsor: significance of MGMT methylation may be context-dependent, emphasizing the need for
National Cancer Institute; Genentech; [Link] tumor-specific studies. Research Sponsor: None.

3147 Poster Session 3148 Poster Session

A phase II basket trial evaluating the efficacy of tasurgratinib (E7090) in Efficacy and safety of larotrectinib in patients with non-primary central
patients with advanced solid tumors with fibroblast growth factor receptor nervous system TRK fusion cancer: An updated analysis. First Author: Rui-Hua
(FGFR) gene alteration: FORTUNE study. First Author: Junichi Matsubara, Xu, Sun Yat-sen University Cancer Center, Guangzhou, China
Department of Clinical Oncology, Kyoto University Hospital, Kyoto, Japan Background: NTRK gene fusions are oncogenic drivers in various tumor types. Larotrectinib
Background: Tasurgratinib is an orally available selective inhibitor of FGFR1-3 tyrosine (laro) is the first-in-class, highly selective, central nervous system (CNS)-active TRK inhibitor
kinase and is approved in Japan for biliary tract cancer with FGFR2 fusions or rear- approved for tumor-agnostic use in patients (pts) with TRK fusion cancer based on a robust
rangements based on a global phase 2 study. We previously identified FGFR gene alterations and durable objective response rate in pts with various cancers. Here, we report updated
that are highly sensitive to tasurgratinib using a high-throughput functional evaluation long-term efficacy and safety data in adult and pediatric pts with non-primary CNS TRK
method (MANO method) (npj Precision Oncology [2021] 5:66). We conducted a single-arm, fusion cancer treated with laro. Methods: Pts with TRK fusion cancer enrolled in 3 laro
investigator-initiated multicenter phase 2 basket trial to evaluate the efficacy and safety of clinical trials (NCT02637687 [SCOUT], NCT02576431 [NAVIGATE], NCT02122913) were
tasurgratinib in patients (pts) with advanced solid tumors harboring FGFR gene alterations, included. Laro was administered at 100 mg twice daily (BID) and 100 mg/m2 BID in most
including alterations identified by MANO method. Methods: Pts with advanced solid tumors adult and pediatric pts, respectively. Responses were independent review committee (IRC)-
with FGFR gene alterations detected by next-generation sequencing assays received assessed per RECIST v1.1. Pts enrolled in SCOUT were permitted to stop laro in the absence
tasurgratinib 140 mg QD. Pts were allocated to each Group based on FGFR gene alteration of on-treatment progression (“wait-and-see”). The data cutoff was July 20, 2024.
(Group A: FGFR1-3 fusion, Group B: FGFR1-3 sensitive mutations to tasurgratinib determined Results: At data cutoff,304 pts were eligible for efficacy assessment by IRC; 25 pts had
by MANO methods, Group C: FGFR1-3 activating mutation not applicable to group B or known CNS metastases at baseline. Median age was 45 years (range 0–90). There were 28
FGFR1, 2 gene amplification, Group D: cholangiocarcinoma with FGFR2 fusion and previous different tumor types, including soft tissue sarcoma (24%), infantile fibrosarcoma (16%),
treatment with a FGFR inhibitor except for tasurgratinib). The primary endpoint for Groups A, lung (11%), and thyroid (10%). A total of 101 pts (33%) received no prior systemic therapies
B, and C was objective response rate (ORR) by independent central review (ICR). Group D in the metastatic/unresectable setting; 115 (38%) received 2 or more. NTRK gene fusions
was an exploratory cohort, and ICR was not performed. The secondary endpoints included were detected by next-generation sequencing (NGS) in 267 (88%) pts. The overall response
ORR by investigator assessment (IA), progression-free survival (PFS), overall survival, and rate was 65% (95% confidence interval [CI] 59–70): 66 (22%) complete responses (CR), 20
safety. The threshold and expected response rates were 5% and 30%, respectively. With the (7%) pathological CR, 112 (37%) partial responses, 56 (18%) stable disease, 32 (11%)
one-sided significance level of 5%, the target enrolments were 10 (62% power), 15 (87%), and progressive disease, and 18 (6%) not evaluable/undefined. Median time to response was
15 pts (87%) in Groups A, B, and C, respectively. Group D’s target number was 1 to 5 pts 1.8 months (mo; range 0.9–22.9). Median duration of response (DoR), progression-free
without a statistical hypothesis. Results: From June 2021 to December 2022, 46 pts were survival (PFS), and overall survival (OS) were 43 mo (95% CI 34–not estimable), 28 mo (95%
registered. The full analysis set includes 41 pts (10, 15, 15, and 1 in Groups A, B, C, and D, CI 22–38), and not reached, respectively, at median follow-ups of 45, 42, and 57 mo. The 4-
respectively). The most common primary sites were brain in 4 pts (40.0%) in Group A, biliary year rates for DoR, PFS, and OS were 48% (95% CI 40–57), 39% (95% CI 32–46), and 63%
tract in 4 pts (26.7%) in Group B, and esophagus/stomach in 4 pts (26.7%) in Group C. ORRs (95% CI 57–68), respectively. Median duration of treatment was 19 mo (range 0–100+).
by ICR in Group A, B and C were 20.0% (90% CI: 3.7-50.7, p = 0.0861), 20.0% (90% CI: 5.7- Fifty-five of 99 pediatric pts in SCOUT had participated in “wait-and-see”; the median
44.0, p = 0.0362), 6.7% (90% CI: 0.3-27.9, p = 0.5367), respectively. ORRs by IA in Group A, B, duration of the first “wait-and-see” period was 33 mo (range 1–72). At data cutoff, 83 pts
C, and D were 20.0% (95% CI: 2.5-55.6), 40.0% (95% CI: 16.3-67.7), 13.3% (95% CI: 1.7-40.5) (27%) remained on trial (either on treatment or in “wait-and-see”). Treatment-related ad-
and 0.0% (95% CI: 0.0-97.5), respectively. Median PFS by IA in Groups A, B, C, and D were 2.5 verse events (TRAEs) were mainly Grade 1/2 (n = 189; 62%). Grade 3/4 TRAEs occurred in 71
(95% CI: 1.4-5.7), 7.2 (95% CI: 1.7-8.2), 2.2 (95% CI: 1.9-3.7) and 5.7 months (95% CI: not (23%) pts. Five (2%) pts discontinued due to TRAEs. Conclusions: Laro continues to
evaluable), respectively. There was no new safety signal compared to previous reports. demonstrate rapid and durable responses, extended survival, clinical benefit, and a favorable
Conclusions: In Group B, the primary endpoint was met. Tasurgratinib demonstrated safety profile in pts with TRK fusion cancer. This data supports the wider adoption of NGS
clinical activity in pts with selected FGFR-mutated tumors. Further study is needed to panels that include NTRK gene fusions to identify pts who may benefit from treatment with
validate these findings. Clinical trial information: NCT04962867. Research Sponsor: Eisai; TRK inhibitors. Clinical trial information: NCT02637687, NCT02576431, NCT02122913.
Japan Agency for Medical Research and Development; 20lk1403036h0001. Research Sponsor: Bayer HealthCare Pharmaceuticals, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 225s

3149 Poster Session 3150 Poster Session

Early prediction of prognosis in advanced solid tumor patients using tumor Picking needles in a haystack: Exploring rare variants of a pan-cancer target
growth rates with g score in early phase clinical trials. First Author: Kana in the RET landscape from 229,453 adult cancer patients. First Author: Niamh
Kurokawa, Department of Advanced Medical Development, The Cancer Institute Hospital Coleman, Trinity St. James’s Cancer Institute, Dublin, Ireland
of Japanese Foundation for Cancer Research, Tokyo, Japan Background: Advances in precision oncology have led to the approval of tumor-
Background: The primary objective of early phase clinical trials is to evaluate the safety agnostic therapies, and RET, due to its role as a driver of oncogenesis across multi-
of investigational drugs, which requires participants to have sufficient expected survival ple tumor types, is increasingly recognized as a pan-cancer target. RET alterations,
durations. Tumor growth rate using g scores, calculated using radiographic mea- including mutations and fusions, are relatively rare events, however, potent and selective
surements and timing after treatment, is gaining attention as a potential tool for RET inhibitors such as selpercatinib and pralsetinib have demonstrated remarkable
treatment efficacy assessment. This study aims to assess the utility of g scores before efficacy and changed clinical practice in RET-driven NSCLC, thyroid cancer and other
and after treatment in predicting prognosis, and explore suitable trial candidates for cancers. Here, we present a comprehensive analysis of RET alterations in pan-cancer
accelerating drug development in early phase clinical trials. Methods: We retrospec- adult malignancies. Methods: 229,453 samples from 196,244 patients available from
tively reviewed patients who participated in early phase clinical trials after standard AACR Project GENIE v.17 database were analyzed for the prevalence of RET mutations,
treatment at the Department of Advanced Medical Development, The Cancer Institute fusions and copy number alterations in a range of cancer types. Results: A total of 7011
Hospital of Japanese Foundation for Cancer Research between January 2020 to De- separate RET alterations were identified in 6690 separate pts (3%), including 660 fusions
cember 2023. A mathematical exponential growth model was applied to estimate tumor (9.4%), 5553 missense mutations (79.2%), 373 splice site mutations (5.3%), 339
growth rates (g) based on radiographic tumor measurements and interval time: f(t) = exp truncating mutations (4.8%), 86 in-frame mutations (1.2%). Most frequent tumor types
(g$ t), with pre-g scores derived from measurements before the clinical trial and post-g included NSCLC, colorectal cancer, melanoma, thyroid cancer, endometrial cancer and
scores from measurements after trial initiation. Pre-g scores were calculated using trial glioma (23%, 12.2%, 9.5%, 6.6%, 6.4%, 5.3% identified RET alterations, respectively). RET
baseline computed tomography (CT)s and the most recent CTs before trial enrollment, fusions were observed in 0.3% of tumor samples, most identified in NSCLC, thyroid and
while post-g scores were calculated using baseline CTs and the first evaluated CTs after colorectal cancer (53%, 24% and 4% of identified RET fusions). Most fusions were
treatment. We defined dichotomized g score levels (high/low) using the time-dependent considered driver events using OncoKB database (632, 96%);frequent fusion gene
ROC curve procedure. We evaluated independent predictors for survival outcomes partners included KIF5B, CCDC6, NCOA4, and intragenic events (34%, 25%, 9.7%, 8% of
according to each g score and patient characteristics. Results: Of the 173 cases who 660 fusion samples). Of the 5553 missense mutations, most (89%) were considered
participated in early phase clinical trials after standard treatment, 162 cases with variants of uncertain significance; 605 (11%) were considered oncogenic or likely
evaluable CT scans before and after the clinical trial were included in this study. Median oncogenic. Oncogenic missense mutations occurred across codons, most frequently
time to pre-trial CT was 29 days (range, 5–202), and median time to first post-treatment involving codon 918 (n = 215, 36%; M918M/K/T/V), 648 (n = 41, 6.8%; V648I/A), 886 (n =
evaluation was 49.5 days (range, 16–87). Log-rank testing showed both high pre-. and 28, 4.6%; R886W/Q/L), 630 (n = 21, 3.5%; C630G/R/S/F/Y/W), 891 (n = 31, 5%, S891A/L/
post-. scores correlated to shorter overall survival (OS) compared to low-score groups W). Documented on-target drivers of multi-kinase RET inhibitor resistance gatekeeper
(HR 2.16; 95% CI 1.22-3.81; P = 0.0067, HR 2.68; 95%CI 1.84-3.90; P , 0.001). mutations (V804M/L), and selective RET inhibitor resistance mutations were noted in 61
Multivariate analysis showed both high pre-g and post-g scores were independent samples, including G810C/S substitutions, solvent-front mutations K809R/N, activation
predictors of shorter OS (HR 2.06, 95% CI 1.13-3.75; P = 0.019, HR 3.80, 95% CI 2.44- loop mutations Y806C/N (33%, 53%, 3%, 3% of identified samples); most were classified
5.90; P , 0.001). Conclusions: This study is the first to incorporate pre-. scores as an as oncogenic or likely oncogenic (85%). Conclusions: RET fusions are rare events
independent prognostic factor and may serve as a valuable reference for patient en- across cancers; however, most are characterized as oncogenic. RET missense muta-
rollment in early phase clinical trials under late-line settings. Additionally, post-. scores tions occur in 2.4% of malignancies, and while most RET missense variants are de-
were also identified as an independent prognostic factor across multiple cancer types in scribed as variants of uncertain significance, oncogenic RET variants are diverse,
early phase clinical trials. These results indicate their potential use as surrogate occurring across codons. We confirm multiple documented oncogenic drivers of on-
endpoints to, which may help facilitate drug development. Research Sponsor: None. target resistance, and their distinct and diverse mechanisms underline the urgent need
to develop next generation RET inhibitors. Research Sponsor: None.

3151 Poster Session 3152 Poster Session

LODESTAR: A single-arm phase II study of rucaparib in solid tumors with A multicenter, randomized controlled trial of intrapleural drug-loaded vesicle
pathogenic germline or somatic variants in homologous recombination perfusion combined with systemic therapy for malignant pleural effusion.
repair genes. First Author: Sriram Anbil, University of Pennsylvania Health System, First Author: Jiani Wang, Department of Medical Oncology, National Cancer Center/
Philadelphia, PA National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of
Background: To explore PARP inhibitor (PARPi) utility across solid tumors and identify Medical Sciences and Peking Union Medical College, Beijing, China
biomarkers that predict sensitivity. Methods: This single-arm phase II study assessed Background: This study aimed to evaluate the efficacy and safety of drug-loaded
rucaparib monotherapy in patients with solid tumors and pathogenic variants (PVs) in vesicle (DLV) intrapleural perfusion combined with systemic therapy in patients with
BRCA1, BRCA2, PALB2, RAD51C, RAD51D (Cohort A) or BARD1, BRIP1, FANCA, NBN, lung or breast cancer and malignant pleural effusion (MPE). Methods: This multicenter,
RAD51B (Cohort B). The primary endpoint was ORR in Cohort A. Secondary endpoints randomized, controlled, open-label clinical trial included patients with pathologically
included DCR, PFS, OS and safety. A scar-based HRD signature (HRDsig) and platinum confirmed lung or breast cancer and MPE requiring thoracentesis. In total, 96 patients
sensitivity status were explored post-hoc. Results: Fifty-one patients in Cohort A and 12 were randomised 1:1 to arm 1 receiving DLV intrapleural perfusion (50 mL daily for four
in Cohort B were evaluable for efficacy. ORR of cohort A was 18% (95% CI 10-30%). A consecutive days) plus systemic therapy (ST) or arm 2 receiving interleukin-2 (IL-2)
significantly higher ORR was observed with HRDsig+ tumors compared to HRDsig- intrapleural perfusion (50 mL every three days for three sessions) with [Link] primary
tumors (32%, 95% CI 15-54, vs. 0%, 95% CI 0-14%, p , 0.01). In the entire study endpoint was the objective response rate (ORR) of pleural effusion at 4 weeks post-
population: DCR of 65% (95% CI 53-76%), mPFS of 5.5 mo (95% CI 3.68-7.82), and mOS perfusion, while secondary endpoints included overall survival (OS) and treatment-
of 12.1 mo (95% CI 10.6 – inf). PFS and OS were significantly longer for platinum related [Link] difference in ORR between the two cohorts was analyzed using the
sensitive tumors (mPFS: 7.8 mo vs. 3.5 mo, p = 0.02; mOS: NR vs 5.45mo, p = 0.01). Chi-square test. Kaplan-Meier survival analysis was performed for OS comparison
Tumor histology was not independently predictive of outcome. Tumors with PVs in between the two cohorts. Results: A total of 91 patients were evaluated for efficacy (50
Cohort A genes were more likely to be HRDsig+ than tumors with PVs in Cohort B genes. in arm 1 and 41 in arm 2). The DLV+ST arm 1 showed a significantly higher ORR for
Analysis of a large commercial database showed that in non-canonical tumors with pleural effusion than the IL-2+ST arm 2 (74.0% vs. 53.7%, P = 0.043). In the survival
BRCA PVs, 30.2% were HRDsig+. Conclusions: Rucaparib has activity in HRDsig+ solid analysis of 83 evaluable patients, median OS was 15.0 months (95% CI: 9.2–26.9) in arm
tumors with PVs in HRR genes, regardless of histology. Platinum sensitivity correlated 1 and 6.9 months (95% CI: 5.3–15.8) in arm 2, without a statistically significant dif-
with improved outcomes. Clinical trial information: NCT04171700. Research Sponsor: ference (HR = 0.75; 95% CI: 0.46–1.24; P = 0.266). The 1-, 2-, and 3-year OS rates for arm
Clovis Pharmaceuticals. 1 were 83.0% (95% CI: 72.9–94.4%), 59.6% (95% CI: 47.1–75.4%), and 51.1% (95% CI:
38.6–67.6%), compared to arm 2’s 69.4% (95% CI: 55.9–86.2%), 41.7% (95% CI:
28.3–61.3%), and 33.3% (95% CI: 21.0–52.9%). Both arms had similar safety profiles,
with chemotherapy-induced toxicities, including leukopenia, gastrointestinal reactions,
and liver dysfunction, being the most common treatment-related adverse events.
Conclusions: Drug-loaded vesicle intrapleural perfusion combined with systemic
therapy is a safe and effective treatment option for malignant pleural effusion in patients
with lung or breast cancer. This approach represents a promising treatment strategy for
MPE and warrants further clinical investigation and consideration in clinical practice.
Clinical trial information: ChiCTR1800017104. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
226s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

3153 Poster Session 3154 Poster Session

The efficacy and safety of a selective PARP1 inhibitor ACE-86225106 in Effect of biopsy requirement on patient enrollment to phase I trials in cancer.
patients with advanced solid tumors: Preliminary results from a first-in- First Author: Marisa Palmeri, Rutgers Robert Wood Johnson Medical School, New
human phase 1/2 study. First Author: Jiongjie Chen, Acerand Therapeutics (Hong Brunswick, NJ
Kong) Limited, Shanghai, China Background: The need for safer and more effective drugs for patients with cancer is a
Background: ACE-86225106 is a highly selective PARP1 inhibitor, exhibiting high constant unmet need. Given their narrow therapeutic index, determination of dose and
potency in enzymatic and DNA-trapping assays of PARP1, while maintaining significant toxicity through phase I clinical trials is confined to patients with cancer. Recently, there
selectivity over PARP2. Pre-clinical studies with ACE-86225106 have demonstrated has been a trend towards a greater demand for fresh tumor biopsies (Bx) from patients to
strong anti-cancer activities in in vivo CDX models, with excellent tolerability. Here we better understand the pharmacodynamic and target effect of the drugs. We sought to
report the preliminary clinical data of ACE-86225106 from the ongoing first-in-human determine whether the requirement for Bx among this vulnerable population had any
study (NCT06380660). Methods: This is a multicenter, open-label, phase1/2 study of detrimental effect. Methods: The study population included patients who enrolled to
ACE-86225106 in adult patients with locally advanced (unresectable) or metastatic solid phase I trials between June 2022 (new EMR system) through January 2025 at a single NCI-
tumors. Phase 1 includes a typical “3+3” dose escalation and backfill module, followed designated comprehensive cancer center. All charts were reviewed and data collected
by a dose expansion module in phase 2. The primary objective is to assess safety, included sex, race, age, diagnosis, and dates of consent, treatment start, last dose of study
tolerability, PK/PD profile, and pre-liminary efficacy of ACE-86225106 as a monotherapy. drug, off treatment, last contact/death; Bx site/approach/complications. Images were
Results: As the data cut-off (23 Jan 2025), 10 patients received ACE-86225106 at a reviewed by an interventional radiologist to assess safety, with targets deemed appropriate
dose of 5mg, 10mg or 20mg QD, and 5 patients backfilled at a dose of 10mg QD. Median biopsied under image guidance. For lung lesions or those , 1 cm, 20g core needle was
number of prior therapy lines was 3 (range 2-12). Two patients (squamous lung cancer used, while 18g core needle was used for others. Outcomes were analyzed by Mantel–Cox
and breast cancer each) did not complete the DLT evaluation period due to disease test using Prism GraphPad v 10. Results: 146 patients [male (n = 63, 43.2%), age 62, 23-
progression and were replaced. No DLTs were reported as of data cut-off. Among total 82(median, range), NHW-81 (55.5%), NHB-23 (15.8%), Hispanic-24 (16.4%), and Asian-18
fifteen patients (10 patients) who received at least one dose of ACE-86225106, no Grade (12.3%)] consented to 25 clinical trials. Of these, 8 mandated paired tumor Bx, 15 were
3 or higher treatment-related adverse events (TRAEs) were reported. There were no mandatory or optional Bx depending on cohort, and 2 did not require Bx. The most common
diagnoses were colorectal (37, 25.3%), other GI (21, 14.4%), pancreas (18, 12.3%), lung (11,
treatment discontinuations or dose reductions due to TRAE. The compound exhibited a
7.5%), breast (6, 4.1%), prostate (3, 2.1%) and others (50, 34.2%). Bx samples were to be
relatively flat PK curve with mild accumulation after multiple dosing. The steady-state
collected prior to the first dose of study drug (pre-dose) and repeated after the first 1-2
Ctrough was approximately 5 fold, 24 fold and 36 fold above target effective concentration
cycles (on-study). Image guidance included ultrasound (50, 57.5%), CT scans (34, 39.1%),
at dose level of 5mg, 10mg, 20mg respectively. The PARylation inhibition was . 90%
and others (3, 3.4%). Overall, 62 patients (42.4%) provided 87 Bx samples; 25 paired Bx, 20
confirming target engagement. Of seven patients having post-treatment tumor as- only pre-dose Bx, and 17 only on-study Bx. Five patients (3.4%) did not undergo Bx because
sessment and being considered efficacy-evaluable, two patients (one fallopian tube it was deemed unsafe or high risk. The sites of Bx included liver (45, 51.7%), lung (9, 10.3%),
cancer patient with BRCA mutation and one prostate cancer patient with BRCA wild type) lymph node (8, 9.2%), peritoneum (4, 4.6%), and others (21, 24.1%). Two patients ex-
achieved PR per RECIST1.1. Conclusions: Preliminary data indicate that ACE-86225106 perienced pneumothorax and recovered without sequelae. The median (mean) duration
is well tolerated and shows promising efficacy in heavily pre-treated advanced solid from consent to start of study treatment was 20 (20) days among Bx patients vs. 14 (16)
tumors. Clinical trial information: NCT06380660. Research Sponsor: Acerand Thera- among non Bx patients (p = 0.003). The median (mean) duration of time on study was 77
peutics (Hong Kong) Limited. (91) days among Bx patients vs. 77 (125) among non Bx patients (p = 0.046).
Conclusions: Over 40% of patients entering phase I trials underwent study specific Bx. The
patients who underwent a Bx had a median delay of 6 days in receiving the first dose of
study medication. Further in-depth review of medical records will help identify variables
that may have led to shorter time on study for patients undergoing clinical trial related
biopsies. Research Sponsor: None.

3155 Poster Session 3156 Poster Session

Rapid analysis and response evaluation of combination anti-neoplastic Body composition modulations during cyclic fasting-mimicking diet in pa-
agents in rare tumors (RARE CANCER) trial: RARE 2 talazoparib and tients with advanced solid cancers. First Author: Caterina Sposetti, Dana-Farber
temozolomide. First Author: Jibran Ahmed, National Cancer Institute, Bethesda, MD Cancer Institute, Harvard Medical School, Boston, MA
Background: Preclinical data generated from NIH/NCI Patient-Derived Models Repository Background: Fasting-mimicking diets (FMD) can induce favorable immune-metabolic changes in
(PDMR) demonstrated significant synergistic activity of talazoparib (a PARP inhibitor) humans and preclinical data suggest potential antitumor activity. Cyclic FMD impact on muscle mass and
combined with temozolomide (an alkylating agent) in patient-derived xenograft models of adiposity in patients (pts) is unclear. Here we evaluate body composition changes in pts with advanced
solid cancers undergoing FMD in the context of a phase Ib trial (NCT03340935). Methods: NCT03340935
rare adult and pediatric cancers. This clinical trial aimed to evaluate the objective response study evaluated safety and biological effects of a cyclic FMD consisting of a 5-day, calorie-restricted,
(OR) rate of this combination in patients with advanced rare cancers in exploratory fashion. plant-based diet repeated every 21-28 days for a maximum of 8 cycles, conducted in oncologic patients
Correlatives include genomic and transcriptomic profiling of tumor tissue, circulating receiving concomitant therapies. Here we included pts with advanced solid cancers and Computed
tumor DNA (ctDNA), circulating tumor cells, and assessments of apoptosis and epithelial- Tomography (CT) exams at baseline (BL), FMD end and disease progression (PD). Body composition
mesenchymal transition in relation to treatment activity. Methods: This open-label, non- parameters, i.e. Skeletal Muscle Index (SMI) and Visceral (VAT), Subcutaneous (SAT), Intermuscular
randomized, phase 2 trial used a Simon two-stage design. Patients aged $18 years with (IMAT) Adipose Tissues, were assessed via axial CT scan at 3rd lumbar vertebra using SliceOmatic
advanced rare cancers received temozolomide (37.5 mg/m² orally, days 2–6) and tala- software. Pts with advanced triple negative breast cancer (TNBC) receiving chemotherapy (ChT) without
FMD, having CT scans at BL and PD, were selected as control cohort. Wilcoxon tests were used for
zoparib (750 mcg orally, daily) in 28-day cycles. Tumor response was assessed per RECIST comparisons. Results: In the FMD cohort (n=36), 61% had BC, 39% had TNBC, 78% received ChT; median
v1.1, and adverse events (AEs) assessed using CTCAE v5.0. In the first stage, if 0/14 age was 54 (IQR 51-65), median completed FMD cycles was 5 (IQR 3-8), median time from FMD end to PD
(across all histologies) responses are observed, the trial will be closed for futility. Oth- was 3.5 months (IQR 1.0-16.0); 8 pts met criteria for sarcopenia (SMI ,38.5 cm2/m2) at BL, 6 of whom
erwise, additional 16 patients were planned to be enrolled. There are no selection criteria had TNBC. In the control TNBC cohort (n=17), median age was 54 (IQR 42-68), 6 pts were sarcopenic at
based aside from rare tumor to allow for exploration of activity. Results: Fourteen patients BL. In the FMD cohort, between BL and FMD end there was a significant reduction of VAT, SAT and SMI,
were enrolled, all evaluable for response and toxicity. Median age was 57 years; 11 were but no change in IMAT; between BL and PD only SMI was significantly decreased (Table). At FMD end and
female, and all had ECOG 0–1. Tumor histologies included uterine sarcoma (N = 3), PD, 12 pts were sarcopenic in the FMD cohort, 7 having TNBC. In the control TNBC cohort (n=17), IMAT
was significantly increased between BL and PD, with no changes in other parameters (Table); 8 pts were
cholangiocarcinoma (N = 2), and one each of adrenocortical carcinoma, adenoid cystic
sarcopenic at PD. Conclusions: In advanced cancers pts, cyclic FMD reduces adiposity as well as muscle
carcinoma, clear cell salivary carcinoma, MPNST, angiosarcoma, carcinoma of unknown mass. Tumor/therapy-related factors contribute to sarcopenia in advanced cancer pts, thus future trials
primary, squamous urothelial carcinoma, small cell neuroendocrine carcinoma, and SDHB involving FMD intervention should detect pts at risk and include supportive measures to preserve muscle
deficient renal cell carcinoma. Best responses included stable disease (N = 6), progressive mass. Support: Italian Association for Cancer Research (AIRC): AIRC-Bonadonna fellowship (C Sposetti),
disease (N = 5), and clinical progression (N = 3). One patient with clear cell salivary cancer AIRC fellowship (F Ligorio), AIRC IG 2024 ID 30499 (PI: C Vernieri); Giuliani Foundation. Clinical trial
and another with cholangiocarcinoma remained on treatment for 8 and 6 cycles, re- information: NCT03340935. Research Sponsor: Italian Association for Cancer Research (AIRC) / Gianni
spectively. The median progression free survival is 3.81 months. The most common Bonadonna Foundation; Italian Association for Cancer Research (AIRC); Italian Association for Cancer
treatment related AEs (TRAEs) overall as well as $Grade 3 were hematologic including Research (AIRC); Giuliani Foundation.
thrombocytopenia ( 13; $Grade 3 = 10), anemia (total12; $Grade 3 = 10), lymphopenia Skeletal
(total 12; $Grade 3 = 5), neutropenia (total 11; $Grade 3 = 6), and leukopenia (total Visceral Subcutaneous Intermuscular Muscle
Adipose Adipose Tissue Adipose Tissue Index
10, $Grade 3 = 4). No Grade 5 TRAEs were reported. Although none of the patients Tissue change - change - % change - % change -
% median p value median p value median p value % median p value
discontinued treatment due to TRAEs, planned dose reductions were needed for 7 patients.
Conclusions: Despite promising preclinical activity, this tumor agnostic exploratory trial FMD end vs BL
FMD cohort, n=36 -12.4 0.002 -11.9 ,0.001 +0.9 0.55 -4.2 0.014
did not meet strict goal more design for single histology. Future efforts will focus on
correlative analyses, exploring histology-specific expansion cohorts informed by pre- PD vs BL
FMD cohort, n=36 -6.2 0.25 -5.5 0.083 +1.2 0.5 -5.1 ,0.001
clinical response data, and optimizing dosing schedules to reduce overlapping toxicities. TNBC subset, n=14 -3.2 0.24 -5.0 0.042 +0.2 0.9 -4.7 0.025
Clinical trial information: NCT05142241. Research Sponsor: U.S. National Institutes of Control TNBC -1.3 0.64 -11.0 0.16 +11.7 0.023 -0.9 0.24
cohort, n=17
Health.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 227s

TPS3157 Poster Session TPS3158 Poster Session

IDeate-PanTumor02: A phase 1b/2 study to evaluate the efficacy and safety REJOICE-PanTumor01: A phase 2 signal-seeking study of raludotatug
of ifinatamab deruxtecan (I-DXd) in patients (pts) with recurrent or meta- deruxtecan (R-DXd) in patients with advanced or metastatic gynecologic
static solid tumors. First Author: Takahiro Kogawa, Department of Advanced Medical or genitourinary tumors. First Author: Laurence Albiges, Gustave Roussy, Paris
Development, The Cancer Institute Hospital of JFCR, Tokyo, Japan Saclay University, Paris, France
Background: B7-H3 is highly expressed in many solid tumors but has limited expression in Background: Cadherin-6 (CDH6), a transmembrane protein involved in cell–cell adhesion
normal tissues; high B7-H3 expression is associated with shorter overall survival (OS) in and epithelial–mesenchymal transition, is overexpressed in many cancer types. R-DXd is an
several tumor types. I-DXd is a B7-H3–directed antibody–drug conjugate (anti2B7-H3 mAb anti-CDH6 antibody–drug conjugate composed of a humanized CDH6 antibody covalently
covalently linked to a topoisomerase I inhibitor cytotoxic payload [DXd] via an enzymatically linked to a potent topoisomerase I inhibitor payload (DXd) via a plasma-stable linker. In an
cleavable peptide-based linker). It showed promising efficacy in pts with advanced solid ongoing Phase 1 study (NCT04707248), a subgroup of patients with heavily pretreated
tumors in the Phase 1/2 IDeate-PanTumor01 study, with objective responses in 6 of the 7 ovarian cancer (OC) who received R-DXd 4.8–6.4 mg/kg, had an objective response rate
tumor types with $5 pts (small cell lung cancer [SCLC], esophageal squamous cell car- (ORR) of 48.6% (95% confidence interval [CI], 31.9–65.6); median duration of response
cinoma, metastatic castration-resistant prostate cancer, squamous non-small cell lung (DOR) was 11.2 months (95% CI, 3.1–not estimable), and progression-free survival (PFS)
cancer, head and neck squamous cell carcinoma [HNSCC], and endometrial cancer). I-DXd was 8.1 months (95% CI, 5.3–not estimable), irrespective of CDH6 expression level (data
also showed encouraging antitumor activity in 88 pretreated pts with extensive-stage SCLC cut-off: July 14, 2023). The safety profile of R-DXd was manageable. In total, 11.1% of
in the Phase 2 IDeate-Lung01 study, with greater efficacy at the 12-mg/kg than the 8-mg/kg patients discontinued R-DXd due to treatment-emergent adverse events. These promising
dose (objective response rates [ORRs] of 54.8% [95% CI, 38.7–70.2] and 26.1% [95% CI, data warranted further investigation of R-DXd in REJOICE-Ovarian01 (NCT06161025), a
14.3–41.1], respectively). I-DXd has demonstrated a manageable and tolerable safety profile Phase 2/3 study in patients with platinum-resistant high-grade serous OC (HGSOC), and in
across tumor types. We describe a study investigating the efficacy and safety of I-DXd in pts the REJOICE-PanTumor01 Phase 2 study, which is described here. Methods: REJOICE-
with advanced solid tumors with substantial unmet medical needs. Methods: IDeate- PanTumor01 (NCT06660654) is a global, open-label Phase 2 study in patients with locally
PanTumor02 (NCT06330064) is a global, multicenter, open-label, single-arm, parallel-cohort, advanced or metastatic gynecologic (endometrial cancer [EC], cervical cancer, or non-
Phase 1b/2 study in ~520 adults with recurrent or metastatic solid tumors (endometrial HGSOC) or genitourinary (urothelial cancer [UC] or clear cell renal cell carcinoma [ccRCC])
cancer; HNSCC; pancreatic ductal adenocarcinoma; colorectal cancer; hepatocellular tumors. Cohorts are tumor type–specific; patients in all cohorts must have relapsed or
carcinoma [HCC]; esophageal/gastroesophageal/gastric adenocarcinoma; urothelial car- progressive disease after receiving $1 prior line (and #3 prior lines in the EC, UC, and ccRCC
cinoma; ovarian cancer; cervical cancer; biliary tract cancer; HER2-low breast cancer [BC]; cohorts only) of standard treatment. Adult patients with ECOG performance status 0–1 are
HER2-negative BC; and cutaneous melanoma). Eligible pts will have received $1 systemic eligible; there is no selection for tumor CDH6 expression. Approximately 40 patients will be
therapy for the selected tumor type and have an ECOG PS of #1. The study will be divided enrolled into each cohort to receive R-DXd 5.6 mg/kg IV every 3 weeks until disease
into 2 parts: Stage 1 and Stage 2 (n»20 per stage per cohort). Each cohort starts with Stage progression per RECIST 1.1, unacceptable toxicity, death, or other reason per protocol. In
1 and may continue to Stage 2 if sufficient safety and efficacy data are observed. All cohorts each cohort, a nonbinding futility interim analysis will be conducted after 20 patients
except the HCC cohort will receive I-DXd 12 mg/kg every 3 weeks (Q3W). The HCC cohort complete a minimum of 12 weeks of follow-up, the results of which may determine whether
includes a safety run-in part to assess tolerability and the potential need for dose ad- the remaining (~20) patients will be treated. Primary endpoints are ORR for the gyneco-
justment; the planned starting dose is 8 mg/kg Q3W, which may be escalated. Primary logical and UC cohorts, disease control rate (DCR) for the ccRCC cohort (both investigator-
endpoints are ORR per investigator (all cohorts) and safety (HCC safety run-in only). assessed), and safety and tolerability for all cohorts. Secondary endpoints are ORR (ccRCC
Secondary endpoints are safety, duration of response, progression-free survival, OS, disease cohort only), DCR (except ccRCC cohort), PFS, DOR, time to response (all investigator-
control rate, pharmacokinetics, and immunogenicity. The Kaplan–Meier method will be used assessed per RECIST 1.1), pharmacokinetics, and immunogenicity. No formal hypothesis
to estimate time-to-event endpoints, the Brookmeyer and Crowley method for median event testing will be performed; ORR and DCR will be analyzed using a Clopper–Pearson method to
times, and the Clopper–Pearson exact method to summarize descriptively endpoints with determine 95% CI. PFS and DOR will be analyzed using the Kaplan–Meier method (2-sided
proportion. Enrollment is ongoing. Clinical trial information: NCT06330064. Research 95% CI). Study enrollment began in January 2025. Clinical trial information: NCT06660654.
Sponsor: Daiichi Sankyo, Inc., Merck, Inc. Research Sponsor: Daiichi Sankyo, Inc., Merck Inc.

TPS3159 Poster Session TPS3160 Poster Session

A phase 1, open-label, multi-center study of the safety, tolerability, and Design of a first-in-human multicenter open-label study of ZW171, a mes-
efficacy of IPH4502 as a single agent in advanced solid tumors. First Author: othelin x CD3 targeting bispecific T-cell engager, in participants with ad-
Shiraj Sen, NEXT Oncology, Dallas, TX vanced solid tumors: ZWI-ZW171-101. First Author: Melissa Lynne Johnson,
Background: Nectin-4 is a cell adhesion molecule frequently overexpressed across Sarah Cannon Research Institute, Nashville, TN
multiple solid tumor types, including urothelial carcinoma (UC), esophageal cancer, non- Background: Mesothelin (MSLN) is a membrane glycoprotein overexpressed in several
small cell lung cancer, and triple-negative breast cancer. It plays a significant role in solid tumors, making it a promising target for cancer treatments, including T cell engagers
carcinogenesis and cancer progression and is associated with poor survival in several (TCEs). ZW171 is a humanized trivalent bispecific TCE antibody that targets a threshold
tumor indications. Targeting Nectin-4 with enfortumab vedotin (EV), an antibody-drug level of MSLN expression with 2 binding sites and CD3e receptor on T cells with 1 binding
conjugate (ADC) with a monomethyl auristatin E (MMAE) payload, demonstrated clinical site. Preclinical studies of ZW171 demonstrated favorable pharmacology, pharmacoki-
benefit in UC, which exhibits the highest Nectin-4 expression among all solid tumor netics (PK), and toxicology, showing it preferentially kills MSLN-overexpressing cells,
types. EV is now approved for the treatment of UC. IPH4502 is a differentiated Nectin-4 activates T cells without significant toxicity, inhibits tumor growth, and is well tolerated in
ADC conjugated with exatecan, a topoisomerase-1 inhibitor payload with a drug-to- cynomolgus monkeys, suggesting its potential for treating MSLN-expressing tumors1
antibody ratio of 8 via a cleavable hydrophilic linker. IPH4502 has been developed to while sparing healthy tissues with low levels of expression. This first-in-human, phase 1,
address the unmet medical need of UC patients who have progressed on, or are ineligible ongoing study (ZWI-ZW171-101) evaluates safety, tolerability, PK, and anti-tumor activity
for EV, as well as to treat tumor types with lower Nectin-4 expression beyond UC. In of ZW171 in participants with advanced solid tumors. Methods: This 2-part study enrolls
preclinical models, internalization capability and bystander effect of IPH4502 enable an eligible adult participants with unresectable MSLN-expressing ovarian cancer (OC), non-
efficient antitumor activity in Nectin-4 expressing tumor models, independent of Nectin- small cell lung cancer (NSCLC), or other MSLN-expressing cancers, with measurable
4 expression level, as well as in models resistant to EV. Finally, IPH4502 shows an- disease per RECIST v1.1, ECOG PS score of 0 to 1, adequate organ function, and a minimum
titumor activity in patient-derived xenograft models from UC and other tumor types. life expectancy of 12 weeks. Participants with additional progressing malignancies, recent
Methods: This is a first-in-human, open-label, multicenter, single-arm Phase 1 study to transplants, clinically significant ongoing toxicity, uncontrolled renal, pancreatic or liver
assess the safety profile (DLTs and MTD), tolerability according to NCI-CTCAE v5.0, and disease, or active autoimmune diseases requiring high-dose corticosteroids or immu-
RP2D of IPH4502 in patients with advanced solid tumors. Secondary objectives aim to nosuppressive drugs are excluded. Part 1 evaluates the safety and tolerability of ZW171
and Part 2 evaluates the anti-tumor activity while continuing to evaluate safety and
characterize the pharmacokinetic profile and evaluate the immunogenicity and pre-
tolerability. Part 1 is dose escalation to identify maximum tolerated dose (using modified
liminary efficacy of IPH4502. The study is being conducted in participants aged $18
toxicity probability interval [mTPI-2] design, n=40) among participants with OC or NSCLC
years withhistologically confirmed, unresectable, locally advanced, or metastatic solid
receiving subcutaneous ZW171 monotherapy on days 1, 8, and 15 of 3-week (21-day)
tumors known to express Nectin-4, including, but not limited to non-small cell lung,
cycles. Approximately 6 dose levels will be explored based on safety and tolerability. Step-
triple-negative breast, ovarian, esophageal, gastric, and colorectal cancers, as well as up dosing will be used for cycle 1. Dose level 1, determined by QSP-based MABEL ap-
UC. Part 1 (Dose Escalation) will use a Bayesian Optimal Interval Design (BOIN) with proach2, is administered at 4.2 mg (day 1), 12.6 mg (day 8), and 38.0 mg (day 15). Part 2 is
backfilling of safety-cleared dose levels. This approach will guide dose escalation and dose expansion in participants with OC, NSCLC, and other MSLN-expressing cancers
help establish the MTD/MAD. Part 2 (Dose Optimization)will begin after identifying the (MSLN expression evaluated retrospectively). Primary objectives are to evaluate safety
MTD/MAD, to select the RP2D. It will enroll participants with selected tumor indications and tolerability of ZW171 and determine the maximum tolerated dose. Key secondary
(up to 2), for whom a clinical benefit was observed in Part 1. Participants will be objectives are to assess PK, anti-drug antibodies, and anti-tumor activity. This is a global
randomized at a 1:1 ratio to 2 dose levels, to determine the RP2D. A maximum of 105 study with sites in North America, Europe, and Asia; and actively enrolling participants into
participants will receive treatment with IPH4502 in France and the US. Clinical trial Part 1. References: 1. Afacan N, et al. Presented at AACR Annual Meeting 2023; abstract
information: NCT06781983. Research Sponsor: None. 2942. 2. Afacan N, et al. Presented at SITC Annual Meeting 2024; abstract 1062. Clinical
trial information: NCT06523803. Research Sponsor: Zymeworks BC Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
228s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

TPS3161 Poster Session TPS3162 Poster Session

A phase 1, first-in-human study of AMT-676, an anti-CDH17 antibody-drug Phase I multicenter, open-label, dose escalation study of T-1201, a small
conjugate, in patients with advanced gastrointestinal tumors. First Author: molecule drug conjugate, to assess safety, pharmacokinetics, and antitu-
Charlotte Rose Lemech, Scientia Clinical Research, Randwick, Australia mor activity in advanced solid tumors. First Author: Hui-Ching Wang, Division of
Background: Cadherin-17 (CDH17), also known as liver-intestine-cadherin, is a Hematology and Oncology, Department of Internal Medicine, Kaohsiung Medical Uni-
transmembrane protein that is highly expressed in a variety of gastrointestinal cancers, versity Hospital, Kaohsiung, Taiwan
including colorectal, gastric, esophageal adenocarcinoma, cholangiocarcinoma, pan- Background: Phosphatidylserine (PS) is a phospholipid critical for maintaining cell membrane
creatic ductal, and gastrointestinal neuroendocrine tumors. The overexpression of integrity and functionality. In rapidly proliferating cancer cells, PS translocates to the outer leaflet
CDH17 is associated with tumor metastasis and progression to advanced tumor stages. of the membrane, making it a promising biomarker and therapeutic target for cancer treatment. The
AMT-676 is a novel antibody-drug conjugate (ADC) that targets CDH17. It is comprised investigational drug T-1201 is a proprietary small molecule drug conjugate combining a bioactive
topoisomerase I inhibitor, SN-38, with Zn-DPA complexes, which exhibit high affinity for PS.
of a humanized IgG1 monoclonal antibody specific to CDH17, conjugated to the potent Preclinical studies have demonstrated T-1201’s in-vivo antitumor activity across multiple human
topoisomerase I inhibitor exatecan, with a drug-to-antibody ratio of 4, linked through a tumor xenograft models. This study represents the first clinical evaluation of T-1201 in humans.
proprietary T-moiety technology. Preclinical studies have demonstrated significant anti- Methods: The primary objectives of this phase I study are to evaluate the safety profile of T-1201,
tumor activity of AMT-676 across multiple gastrointestinal cancer models and great determine dose-limiting toxicities (DLTs), establish the maximum tolerated dose (MTD), and
tolerability in safety studies, highlighting its potential as a therapeutic agent for CDH17- identify the recommended phase II dose (RP2D). Secondary objectives include characterization of
expressing malignancies. Methods: This phase 1, open-label, multicenter study aims to pharmacokinetics (PK) and assessment of antitumor activity for T-1201. The study comprises
determine the Maximum Tolerated Dose (MTD) and the Recommended Phase 2 Dose three dose-escalation parts. In Part A, T-1201 is administered intravenously once every four weeks
(RP2D) of AMT-676, as well as to assess its safety, tolerability, anti-drug activity, (Q4W), starting at 18 mg/m2 during Cycle 1. From Cycle 2 onward, the dosing interval can be
pharmacokinetics, pharmacodynamics, immunogenicity and preliminary efficacy in adjusted to once every two weeks (Q2W) at the investigator’s discretion, subject to agreement with
the Sponsor. When switching to the Q2W schedule, the dose level is halved compared to the Q4W
patients with advanced solid tumors. Tumor types that express CDH17 including
dose. Each treatment cycle spans four weeks, with dose escalation proceeding via a single-patient
gastrointestinal cancers, treated with or with no standard therapeutic options are to be cohort design (100% dose increments) initially, transitioning to a modified 3+3 design (40% dose
enrolled. AMT-676 will be administered intravenously on a 21-day cycle. The dose increments) based on DLTs observed in Cycle 1. In Part B, T-1201 is administered intravenously
escalation will be guided by the Bayesian Optimal Interval (BOIN) design, incorporating Q2W in a 28-day treatment cycle, starting at 100 mg/m2, which represents half of the MTD
an accelerated titration approach to evaluate 6 cohorts: 1.6, 3.2, 4.8, 6.4, 8, and 10 mg/ identified in Part A. In Part C, each treatment cycle is reduced to 21 days, with the starting dose not
kg. Three backfilling cohorts at doses that have demonstrated safety will also be in- exceeding the highest dose level deemed safe by the Safety Review Committee (SRC) in Part B.
cluded, each enrolling up to 18 patients, to gather additional data on safety, phar- Eligible patients are $18 years of age, have an Eastern Cooperative Oncology Group (ECOG)
macokinetics, pharmacodynamics, and preliminary efficacy, thereby supporting the performance status of 0–1, and possess radiographically or clinically evaluable tumors. As of now,
selection of an optimized dose for expansion. Mandatory pre-study biopsy sample 27 patients have been enrolled in the Part A dose-escalation stage. This study is registered with
collection for retrospective immunohistochemistry (IHC) analysis will facilitate a [Link] (NCT04866641). Clinical trial information: NCT04866641. Research Sponsor:
Taivex Therapeutics corporation.
comprehensive exploratory biomarker plan, potentially correlating CDH17 levels with
treatment responses. The study is actively enrolling participants for the dose escalation Dose escalation/de-escalation rule for the BOIN design.
phase. Cohorts 1-4 have been completed DLT evaluation and enrollment of cohort 5 Number of subjects treated at the current dose* 3 4 5 6 7 8 9
began in December 2024. Clinical trial information: NCT06400485. Research Sponsor: Escalate if # of DLT £ 0 1 1 1 1 2 2
Stay at current dose if # of DLT = 1 NA NA 2 2 3 3
Multitude therapeutics Inc. De-escalate if # of DLT ‡ 2 2 2 3 3 4 4
Eliminate if # of DLT ‡ 3 3 4 4 5 5 6
*The enrollment may stop when one of the following criteria is met: The planned sample size has been reached; at least 9
subjects have been treated and evaluable for DLT at one dose level; or all doses explored appear to be overly toxic, and the
MTD cannot be determined.

TPS3163 Poster Session TPS3164 Poster Session

A phase 1 study to evaluate the safety and tolerability of the antibody–drug TUB-030, a novel ADC targeting 5T4: A phase I/IIa multi-center, first-in-
conjugate (ADC) MesoC2 (PF-08052666) in patients with advanced solid human clinical trial (5-STAR 1-01) in patients with advanced solid tumors.
tumors. First Author: Amita Patnaik, START San Antonio, San Antonio, TX First Author: Shiraj Sen, NEXT Oncology, Dallas, TX
Background: MesoC2 (PF-08052666) is an ADC that targets mesothelin (MSLN), a cell- Background: TUB-030 is a novel antibody-drug conjugate (ADC) targeting 5T4, an
surface glycoprotein overexpressed in solid tumors including mesothelioma, ovarian oncofetal antigen expressed in various solid tumors with limited expression in healthy
cancer, pancreatic cancer, non-small cell lung cancer (NSCLC), endometrial cancer (EC), tissues. TUB-030 leverages optimized biophysical properties, an effector-silenced an-
and colorectal cancer (CRC), but with limited expression in normal tissues. MesoC2 is tibody, and an exatecan payload to maximize the therapeutic index and minimize off-
constructed from a recombinant human IgG1 anti-MSLN monoclonal antibody conjugated target toxicities. Preclinical studies demonstrated potent anti-tumor activity, including
to a cleavable tripeptide linker that carries a topoisomerase 1 inhibitor (TOP1) payload. long-lasting tumor regression at doses as low as 1 mg/kg and durable responses even in
The average number of TOP1 molecules per antibody is 8. Following high-affinity binding tumors with low 5T4 expression. Methods: 5-STAR 1-01 is a multicenter, first-in-human
to MSLN on the cell surface, MesoC2 is internalized, the linker is cleaved, and the released dose escalation and dose optimization Phase I/IIa clinical trial designed to investigate
payload inhibits DNA religation during amplification, leading to cell cycle arrest and cell safety, tolerability, pharmacokinetics (PK), and efficacy of the anti-5T4 ADC TUB-030 in
death. MesoC2 has shown potent antitumor efficacy in in vitro assays and xenograft patients with advanced and metastatic solid tumors. Eligible patients have one of the
models and an acceptable safety profile in cynomolgus monkeys. The aim of this first-in- following tumor types: head and neck squamous cell carcinomas (HNSCC), non-small-
human study is to explore the safety, tolerability, and preliminary efficacy of MesoC2 in cell lung cancer (NSCLC), small cell lung cancer, pleural mesothelioma, triple-negative
patients with certain advanced solid tumors. Methods: In this phase 1, open-label study, breast cancer, HR+/HER2- breast cancer, esophageal cancer, gastric cancer, pancreatic
up to 365 patients with mesothelioma, platinum-resistant ovarian cancer (PROC), adenocarcinoma, colorectal cancer, bladder cancer, prostate cancer, cervical cancer,
pancreatic ductal adenocarcinoma (PDAC), NSCLC, EC, or CRC will receive intravenous osteosarcoma, or soft tissue sarcomas and must have exhausted available standard-of-
infusion of MesoC2 in dose escalation (n=45), dose and schedule optimization (n=40), and care therapies. Phase I is an open-label, single-arm dose escalation trial, with ad-
disease-specific dose expansion cohorts (n=280; includes a biology cohort to evaluate ministration every 21 days. Dose escalation follows an accelerated titration design (ATD)
exploratory biomarkers). Key inclusion criteria are histologically or cytologically con- transitioning to Bayesian optimal interval (BOIN) upon predefined toxicity thresholds.
firmed metastatic or locally advanced mesothelioma, PROC, PDAC, NSCLC, EC, or CRC Backfill cohorts are planned in NSCLC and HNSCC to further evaluate the safety and
who have relapsed or progressed following standard therapies; aged $18 years; ECOG efficacy profile at, or near, the maximum tolerated dose (MTD). Primary endpoints
performance status score of 0 or 1; and available archival tumor tissue (a fresh biopsy is include safety and tolerability of TUB-030 as monotherapy, determination of the MTD
required if unavailable). Key exclusion criteria include prior or current treatment with and the recommended phase II doses; secondary endpoints assess pharmacokinetics,
systemic anticancer therapy or focal radiotherapy within 4 weeks prior to first dose of immunogenicity, and preliminary clinical activity using RECIST v1.1 criteria. Exploratory
MesoC2, prior anti-MSLN therapies, and any unresolved toxicities from prior therapy endpoints include analysis of circulating tumor DNA. In phase IIa, dose-optimization will
greater than G1 at the time of starting study treatment, except alopecia. Primary evaluate two dose levels in select indications in order to identify the optimal dose for
endpoints include type, incidence, and severity of adverse events (AEs), frequency of dose further development. Enrollment of approximately 130 patients across the US and
modifications due to AEs, incidence of dose-limiting toxicities, cumulative safety, Canada is planned, with dose escalation currently underway. This study investigates
pharmacokinetics, pharmacodynamics, and preliminary antitumor activity. Key additional TUB-030, a novel 5T4 targeted ADC as a therapy for advanced/metastatic solid tumors.
endpoints include objective and best response rates per Response Evaluation Criteria in Clinical trial information: NCT06657222. Research Sponsor: None.
Solid Tumors Version 1.1 (RECIST v1.1), duration of response, progression-free survival,
overall survival, MSLN expression in blood and tissue, and changes in tumor-specific
biomarkers. Enrollment is ongoing; clinical trial information: NCT06466187. A genAI tool
(01/06/25; Pfizer; GPT-4o) developed the 1st draft; authors assume content responsi-
bility. Clinical trial information: NCT06466187. Research Sponsor: Pfizer Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 229s

TPS3165 Poster Session TPS3166 Poster Session

PROCEADE PanTumor: A phase 1b/2, multicenter study of precemtabart A dose escalation and cohort expansion phase I/IIa study of ACR246, an
tocentecan (M9140), an anti-CEACAM5 antibody-drug conjugate (ADC) with innovative 5T4- antibody drug conjugate (ADC), in patients (pts) with ad-
exatecan payload, in patients with advanced solid tumors. First Author: Zev A. vanced solid tumors. First Author: Panpan Zhang, Key laboratory of Carcinogenesis
Wainberg, University of California, Los Angeles, Medical Center, Los Angeles, CA and Translational Research (Ministry of Education/Beijing), Department of Early Drug
Background: CEACAM5 is a cell surface glycoprotein that is overexpressed in various Development Centre, Peking University Cancer Hospital & Institute, Beijing, China
carcinomas, notably in gastric cancer (GC), non-small cell lung cancer (NSCLC), pancreatic Background: The oncofetal antigen 5T4 is overexpressed in many solid tumors with limited
adenocarcinoma (PDAC), and colorectal cancer (CRC), but shows limited expression on expression in normal adult tissues. Overexpression of 5T4 is associated with poor prognosis.
healthy adult cells. Precemtabart tocentecan is an investigational anti-CEACAM5 ADC (drug- 5T4 on tumor cell surface is rapidly internalized when bound to antibody and is thus an ideal
to-antibody ratio: 8) that utilizes a unique linker–payload combination to selectively deliver target for the development of ADC drugs. ACR246 is the first next- generation 5T4-ADC
the topoisomerase 1 inhibitor, exatecan, to CEACAM5 overexpressing tumor cells. Preliminary consisting of a fully human monoclonal antibody that is site-specifically conjugated to a
clinical data from the dose-escalation part of the first-in-human study of precemtabart novel DNA topoisomerase I inhibitor D2102, via a stable and cleavable linker, with a drug-to-
tocentecan in patients with metastatic CRC (PROCEADE CRC-01) demonstrated a man- antibody ratio (DAR) of 8. ACR246 was carefully designed to improve the safety and efficacy
ageable and predictable safety profile and promising preliminary efficacy in 40 heavily in treating 5T4 positive solid tumors. In preclinical studies, ACR246 demonstrated robust
pretreated patients. The PROCEADE PanTumor study is a Phase 1b/2, multicenter, open-label anti-tumor activity, superior to a Dxd-5T4 ADC (as a reference) both in CDX and PDX models,
study that aims to investigate the clinical activity of precemtabart tocentecan, either as including but limited to NSCLC, gastric cancer, pancreatic cancer and Esophageal cancer,
monotherapy or in combination with other anticancer agents, in patients with advanced GC, and excellent tolerability, supporting further development for clinical use. Methods: This is
advanced NSCLC and advanced PDAC. Methods: The study was designed as a matrix study an ongoing, phase I/IIa, open-label, multicenter, dose escalation and cohort expansion study
with a master protocol (applicable to all substudies) and three substudy protocols (GC; of ACR246 to be injected intravenously to adult pts with advanced solid tumors. For phase I
NSCLC; PDAC). Based on the master protocol, patients aged $18 years, with an Eastern study, a Bayesian optimal interval design is adopted to assess dose levels of ACR246, 0.6,
Cooperative Oncology Group performance status #1, adequate baseline hematological, renal, 1.2, 2.4, 3.6 and 4.5 mg/kg, administered every 3 weeks on a 21-day cycle，and inter-
and hepatic function, $1 lesion that is measurable using RECIST v1.1, who have received $1 mediate dose levels of 3.0, 4.0 and 5.0 mg/kg may be evaluated based on emergent safety or
prior line of treatment are eligible. Patients must have an archival formalin-fixed paraffin- pharmacologic data. The primary objectives are to evaluate safety and tolerability and
embedded tumor tissue or a fresh biopsy. In the respective substudies, patients with ad- determine the maximum tolerated dose (MTD) and the recommended phase 2 dose (RP2D);
vanced or metastatic, HER2-negative GC or gastroesophageal junction adenocarcinoma; the second objectives include PK, immunogenicity and preliminary clinical efficacy. Dose
patients with advanced (Stage III; ineligible for resection/curative radiation) or metastatic limiting toxicity (DLT) will be assessed at each dose level. The DLT evaluation period will be
NSCLC; or patients with advanced or metastatic PDAC will be included. Patient selection will 21 days. Once the RP2D is determined, phase IIa study will be conducted to further evaluate
be based on CEACAM5 expression level (both high and low in GC, only high in NSCLC and the safety, tolerability, efficacy, PK and immunogenicity of ACR246 in 5T4-positive ad-
PDAC [CEACAM5high: $50% tumor cells with immunohistochemistry [IHC] $2+ staining; vanced solid tumor pts (esophageal cancer, NSCLC, ovarian cancer, prostate cancer and
CEACAM5low: ,50% tumor cells with IHC $2+ staining]), and in patients with NSCLC, EGFR other types of tumors) under RP2D. Approximately 77 pts $ 18 years of age with advanced
mutation status (EGFR-wt and EGFR mut+). The primary endpoint is objective response solid tumors that have histologically or cytologically been diagnosed recurrent or metastatic
(proportion of patients with confirmed complete/partial response [CR/PR] per RECIST v1.1, unresectable advanced disease and have failed or are intolerant of systemic standard
assessed by investigator). Secondary endpoints include adverse events, duration of response therapy or standard therapy is not available, and having adequate ECOG performance status
(RECIST v1.1), disease control (CR, PR, stable disease, or non-CR/non-progressive disease (0-1), hematologic function, and end organ function are planned to be enrolled, with 37 pts in
[PD] at Week 12), time to response, progression-free survival, and pharmacokinetic as- phase I study and approximately 40 pts in phase IIa study. 5T4 expression is not required for
sessments. The study is planned to be initiated at multiple sites globally, with an estimated enrollment for phase I, but will be assessed retrospectively. The toxicity will be assessed by
enrollment of 250 patients. Copyright © 2025 AACR. Originally presented at AACR 2025. Common Terminology Criteria for Adverse Events v5.0 and the tumor response will be
Reprinted with permission. Clinical trial information: NCT06710132. Research Sponsor: the
determined per RECIST v1.1. Dose levels of 0.6 mg/kg and 1.2mg/kg has completed en-
healthcare business of Merck KGaA, Darmstadt, Germany (CrossRef Funder ID: 10.13039/
rollment with no DLT. Clinical trial information: NCT06238401. Research Sponsor: Hangzhou
100009945).
Adcoris Biopharma Co., Ltd.

TPS3167 Poster Session TPS3168 Poster Session

A phase 1 dose escalation and dose expansion study for LNCB74, a B7-H4 The EQUAL study: Utilizing plasma EGFR cfDNA detection as an accessible
targeted antibody drug conjugate, as monotherapy in participants with screening tool for lung cancer in underserved patients ineligible for routine
advanced solid tumors. First Author: Michael M Song, NEXT Oncology, San Anto- screening. First Author: Narjust Florez, Lowe Center for Thoracic Oncology, Dana-
nio, TX Farber Cancer Institute, Harvard Medical School, Boston, MA
Background: B7-H4 is a transmembrane receptor of the B7-family of immunomodulatory Background: Lung cancer (LC) among non-tobacco-users is increasing in the United States,
proteins whose expression correlates with poor clinical outcomes for ovarian and breast with no routine screening available. Among those patients, EGFR mutations (EGFRm) are
cancers. High expression in multiple tumor types and limited expression in normal tissues common, with the highest prevalence seen in East Asian and Hispanic women. Delays in
makes B7-H4 an attractive target for antibody drug conjugate (ADC) therapeutics. LNCB74 diagnosis and treatment are exacerbated in these marginalized groups and women, neg-
is a B7-H4 targeted ADC in which a humanized IgG1k antibody is conjugated to the atively impacting their cancer outcomes. At Dana-Farber Cancer Institute’s (DFCI) Belfer
microtubule disrupting payload monomethyl auristatin E (MMAE) with a drug-to-antibody Center for Applied Science, we developed a novel droplet digital PCR ctDNA assay to detect
ratio of 4 (DAR4). LNCB74 is designed to maximize therapeutic index through three key EGFR del19 and L858R mutations, which comprise 85-90% of total EGFRm in LC. Here, we
elements. First, site specific ConjuAll conjugation results in a homogeneous DAR to drive report the methodology of EQUAL, a study assessing the feasibility of a diagnostic assay
uniform PK. Second, our proprietary glucuronidase-cleavable linker reduces both on- and among non-tobacco using, historically marginalized East Asian and Hispanic populations at
off-target toxicity. Third, the antibody Fc was “LALA”-mutated to reduce Fc mediated high risk for EGFRm-LC. Methods: To assess the feasibility of our ctDNA screening tool, the
uptake into Fc receptor expressing cells such as immune and endothelial cells. Compared EQUAL study is recruiting two cohorts of participants. Cohort 1 (n=500) includes 50–80-
to other B7-H4 targeted ADCs in clinical development, LNCB74 has demonstrated a year-olds who self-identify as East Asian or Hispanic from the general population, while
superior safety profile in nonhuman primate toxicity studies and potent anti-tumor activity Cohort 2 (n=500) includes 40–80-year-olds of the same backgrounds with an additional risk
in multiple cell line- and patient-derived xenograft in vivo models, making it a promising factor for LC, with a focus on direct family members of patients with EGFRm-LC. Re-
ADC therapy for B7-H4-expressing solid tumors. Methods: LNCB74-01 is a phase 1, open- cruitment is beginning with these family members of patients with EGFRm-LC at DFCI main
label, first-in-human study that will include dose escalation, safety, and biomarker backfills campus, Beth Israel Deaconess Medical Center, Massachusetts General Hospital, DFCI
(Part 1) and randomized dose expansion/optimization (Part 2). The objectives of the study Merrimack Valley, DFCI regional campuses, and will expand to primary care clinics and
will be to determine safety and tolerability, define the maximum tolerated dose and/or community events. Blood samples are collected in clinics or at home via mobile phlebotomy.
recommended phase 2 dose, characterize the pharmacokinetics (PK) and pharmacody- Positive results are verified in a government-certified CLIA laboratory; a complementary
namics (PD), and to assess the preliminary efficacy in participants with metastatic solid chest CT will be arranged for those with positive assay results, and patients will receive
tumors treated with LNCB74. The tumor types include ovarian, breast, endometrial, biliary navigation until resolution. Patients with a positive assay but negative chest CT will be
tract cancer, and squamous NSCLC. Key eligibility criteria include measurable disease followed for 12 months and will receive a second annual chest CT. Recognizing how cultural
based on RECIST v1.1 and the ability to provide tissue samples to test B7-H4 expression by beliefs and tobacco’s association with LC may hinder screening participation, EQUAL in-
CLIA-certified immunohistochemistry assay in a central laboratory. Participants will re- cludes an optional survey and focus groups to explore perceptions and barriers surrounding
ceive LNCB74 on Day 1 of each 21-day cycle. Dose escalation will follow a Bayesian LC screening with our tool for future optimization efforts. The study is available in 8
optimal interval (BOIN) design. Dose expansion will occur in up to two tumor types. In each languages including Spanish, Portuguese, Korean, Vietnamese, Japanese, Chinese (sim-
tumor specific dose expansion, participants will be randomized to two dose levels plified and traditional), and Creole. EQUAL is the first study to implement EGFRm-LC blood-
stratifying for prior lines of therapy (1-3 vs $4) and B7-H4 expression (intermediate vs based screening for historically marginalized populations who are not eligible for LC
high). The PK profile, immunogenicity, preliminary anti-tumor activity per RECIST v1.1, and screening, thereby allowing for LC identification that can be effectively treated with targeted
correlation of baseline B7-H4 expression to anti-tumor activity of LNCB74 will be evaluated therapy approved for stages IB-IV. This pilot study seeks to lay the groundwork for future
as secondary endpoints. Biomarkers will be assessed in peripheral blood and tumor tissue. sensitivity and specificity trials that will confirm the value of the assay and expand the scope
Enrollment is ongoing in the United States. Clinical trial information: NCT06774963. of current screening guidelines to reduce health disparities and delays in LC diagnosis.
Research Sponsor: NextCure Inc. Clinical trial information: NCT06716580. Research Sponsor: Dana-Farber Cancer Institute
Philanthropic Funds.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
230s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

TPS3169 Poster Session TPS3170 Poster Session

A phase 1/2 study of FOG-001, a first-in-class direct b-catenin: TCF inhib- SCRUM-Japan MONSTAR-SCREEN-3: Comprehensive tumor microenviron-
itor, in patients with colorectal cancer (CRC) and other locally advanced or ment analysis via multi-omics in a large-scale prospective study. First Author:
metastatic solid tumors. First Author: Kyriakos P. Papadopoulos, START-San Mitsuho Imai, Translational Research Supporting Office, National Cancer Center Hospital
Antonio, San Antonio, TX East, Kashiwa, Japan
Background: Activation of the Wnt/b-catenin pathway, often as truncal APC mutations, Background: SCRUM-Japan is a multi-institutional, industry-academia collaborative
in 80–90% of CRCs and other solid tumors, is known to be a key driver of cancer cancer genome screening project launched in 2015, consisting of LC-SCRUM-Asia for lung
progression and has been associated with immune exclusion and resistance to immu- cancer and SCRUM-MONSTAR for other malignancies. The project has successfully
notherapy. Development of agents targeting this pathway at the key b-catenin: T-cell implemented organ-agnostic liquid biopsy-based precision oncology and molecular re-
factor (TCF) node has eluded the pharmaceutical industry to date. FOG-001 is a Helicon sidual disease (MRD)-guided therapeutic development, resulting in multiple regulatory
peptide that competitively inhibits interaction between b-catenin and TCF transcription approvals of therapeutic agents and diagnostics. In 2024, MONSTAR-SCREEN-3 was
factors. Helicon peptides are hyperstabilized a-helices that can be tuned for picomolar launched to expand the scope of multi-omics analysis beyond advanced solid tumors to
binding affinities, robust cell penetration, broad tissue distribution, no immune recog- include resectable solid tumors and hematologic malignancies, aiming for a compre-
nition, and long in vivo half-lives. In studies in a wide range of patient-derived xenograft hensive understanding of tumor microenvironment (TME) dynamics. The project
(PDX) CRC and HCC models, FOG-001 inhibited tumor growth and promoted tumor re- integrates a multi-omics platform, including spatial transcriptomics, ctDNA analysis, and
gression as monotherapy. Combinations with immune checkpoint inhibitors or standard- proteomics, to advance personalized medicine and accelerate drug development.
of-care therapies, including bevacizumab and 5-FU, showed strong additivity/synergy in Methods: MONSTAR-SCREEN-3 (UMIN000053975) is a large-scale, multi-institutional
PDX CRC models. Methods: This first-in-human, phase 1/2, multicenter, open-label, dose- prospective study involving 55 centers across Japan, aiming to enroll 3,200 patients
escalation (part 1) and dose-expansion (part 2) study evaluates the safety/tolerability, across three cohorts: Cohort A: Advanced solid tumors undergoing systemic therapy
pharmacokinetics (PK), pharmacodynamics (PD), and anti-tumor effects of FOG-001 (n=1,700); Cohort B: Resectable solid tumors receiving perioperative treatment (n=1,100);
monotherapy and combined with other anti-cancer therapies in patients with micro- Cohort C: Hematologic malignancies (n=400). Our analysis platform combines spatial
satellite stable (MSS) CRC or advanced/metastatic solid tumors known to harbor a Wnt transcriptomics with circulating tumor DNA/RNA sequencing, bulk tissue whole exome/
pathway-activating mutation (WPAM). Eligible patients must have received at least one transcriptome sequencing, plasma proteomics, and microbiome analyses. For resectable
prior systemic anti-cancer therapy and either progressed on, not responded to, or be unfit cases, standardized longitudinal monitoring with whole genome sequencing-based MRD
for available therapies. In Part 1, FOG-001 is administered intravenously every week, at analysis is implemented, while disease-specific MRD approaches are applied to hema-
escalating dose levels evaluated sequentially in a standard 3+3 design as monotherapy in tologic malignancies. Following SCRUM-Japan’s quality assurance system, standardized
patients with MSS CRC or any solid tumor with documented WPAM. PD effects are monitoring collects regulatory-grade clinical data, including key indicators such as re-
evaluated in a separate cohort of approximately six patients with MSS CRC. Combination sponse rate, progression-free survival, and overall survival. MONSTAR-SCREEN-3 applies
cohorts will evaluate FOG-001 + FOLFOX/bevacizumab (1L MSS CRC), FOG-001 + standardized protocols for tissue preservation and data acquisition across all centers,
nivolumab (3L MSS CRC or anti-PD-1/PD-L1-resistant CRC and solid tumors), and FOG- ensuring high-quality data. The project leverages the VAPOR CONE supercomputing in-
001 + trifluridine/tipiracil + bevacizumab (3L MSS CRC). Part 2 dose expansion will frastructure for real-time data integration and AI-driven analysis to identify biomarkers,
evaluate FOG-001 monotherapy in patients with MSS CRC and other solid WPAM+ tumors. elucidate resistance mechanisms, and deepen the understanding of tumor-immune in-
Combination dose expansion will evaluate combinations initially studied in Part 1. Primary teractions. The study aims to establish a framework for next-generation precision on-
endpoints are safety/tolerability of FOG-001 alone or in combination. Secondary end- cology. The latest enrollment status and initial operational results will be reported at
points are PK, PD, recommended phase 2 dose and schedule, and preliminary anti-tumor the ASCO meeting. MONSTAR-SCREEN-3 is expected to contribute to new therapies,
activity (e.g., ctDNA changes, overall response rate, best objective response, duration of cross-cancer MRD assays, and the resolution of drug lags in hematologic malignancies,
response, and progression-free survival). 156 patients are planned to be enrolled in Part 1, driving advancements in personalized cancer treatment. Clinical trial information:
which is currently enrolling in the USA. Clinical trial information: NCT05919264. Research UMIN000053975. Research Sponsor: None.
Sponsor: Parabilis Medicines.

TPS3171 Poster Session TPS3172 Poster Session

A first-in-human multi-center phase 1/2 study of a selective FGFR2/3 Phase IB/II study to evaluate safety and preliminary efficacy of the WEE1
inhibitor, CGT4859, in patients with intrahepatic cholangiocarcinomas or inhibitor Debio 0123 in combination with sacituzumab govitecan (SG) in
other advanced solid tumors. First Author: Lipika Goyal, Stanford Cancer Center, triple-negative or hormone receptor–positive (HR+)/HER2-negative
Stanford School of Medicine, Stanford, CA (HER2–) advanced breast cancer (ABC): The WIN-B study. First Author:
Background: Genetic alterations in fibroblast growth factor receptors 2 and 3 (FGFR2/3) Timothy J. Robinson, University of Bristol, Bristol, CT, United Kingdom
occur in nearly all cancer types. FGFR2 fusions and rearrangements occur in up to 10-15% Background: SG is a Trop-2 directed antibody drug conjugate that has shown an overall
of intrahepatic cholangiocarcinomas (iCCA) and alterations in FGFR3 occur in 15-30% of survival benefit for patients (pts) with HER2- ABC in two phase III trials. Unfortunately,
urothelial cancers. The clinical benefit from currently approved FGFR inhibitors (FGFRi) is most pts become refractory to this treatment, highlighting a critical need for strategies
often curtailed by development of acquired resistance, which may arise through on-target to overcome resistance to SG and improve therapeutic outcomes. WEE1 is a cyclin-
mutations in the FGFR2/3 kinase domain. Additionally, off-tumor effects on FGFR1 by pan- dependent kinase 1 regulator, which delays the G2/M transition and maintains genomic
FGFRi can lead to hyperphosphatemia and consequently to dose reductions or dose holds. stability during the cell cycle. Debio 0123, a highly selective and brain penetrant WEE1
Thus, there is an unmet clinical need for a selective FGFR2/3 inhibitor that has clinical inhibitor, has demonstrated synergistic activity in breast cancer preclinical models with
efficacy against activating alterations and resistance mutations without causing FGFR1- SG. The aim of the WIN-B study is to evaluate the safety and preliminary efficacy of
mediated hyperphosphatemia. CGT4859 is an orally bioavailable, ATP-competitive, re- combining the WEE1 inhibitor Debio 0123 with SG in pts with previously treated HER2-
versible inhibitor of FGFR2/3, with potency against clinically relevant FGFR2/3 kinase ABC. Methods: WIN-B (NCT06612203) is an international, multicenter, open-label,
domain mutations. In addition, CGT4859 demonstrates .140 fold selectivity over FGFR1, single-arm phase Ib/II trial. In phase Ib, 12-24 pts will be assigned to different
and shows robust efficacy in target altered in vivo tumor models without increases in Debio 0123 dose cohorts (200, 300, 400, or 520 mg orally once daily on days 1-3 and 8-
serum phosphorus. Nonclinical pharmacokinetics (PK) and safety data support evaluating 10) plus standard doses of SG (10 mg/kg intravenously on days 1 and 8) given in 3-week
CGT4859 in a first-in-human, open-label, dose-escalation and signal-seeking Phase I/II cycles. In phase II, 52 pts will be divided into cohorts A (triple-negative breast cancer
study (NCT06777316). Safety, tolerability, PK, pharmacodynamics, and antitumor activity [TNBC], n = 26) and B (HR+/HER2– tumors, n = 26), and will be treated with the
of CGT4859 will be assessed in adults with histologically confirmed unresectable or recommended doses determined during phase Ib. Key inclusion criteria are: pts
metastatic iCCA or other solid tumors with FGFR2/3 alterations. Methods: CGT4859 will aged $18 with TNBC or HR+/HER2- tumors who have experienced disease progression
be administered orally continuously in 28-day cycles to patients (N=~50) at a starting dose
after 1 or 2 lines of systemic therapy for ABC, ECOG performance status of 0-1, with
of 1 mg QD, and dose escalation will not exceed 40 mg QD as determined using a Bayesian
evaluable (for phase Ib) or measurable (for phase II) disease as per RECIST v.1.1. Pts will
optimal interval design with backfill (BF-BOIN). This approach will be used to guide dose
receive study treatment until progression, death, unacceptable toxicity, or study dis-
escalation and establish the maximum tolerated dose (MTD) and recommended Phase 2
continuation. Primary objectives are: in phase Ib, to establish the recommended phase 2
Dose (RP2D). BF-BOIN enables backfilling of participants to doses that are cleared for
safety during the dose escalation, generating additional data on safety and tolerability
dose of the combination of Debio 0123 plus SG and, in phase 2, to assess the objective
below the MTD. Objective response rate (ORR) and disease control rate will be determined response rate (ORR) as per RECIST v.1.1. Key secondary endpoints are progression-free
based on investigator assessment using RECIST v1.1. Phase II will enroll up to 4 cohorts, survival and overall survival, safety and toxicity. In phase Ib, dose escalation will be
each enrolling ~15 patients. Proposed cohorts will include participants who have iCCA and performed using a Bayesian Logistic Regression Model with overdose control. In phase
are either FGFRi-naı̈ve or FGFRi-exposed. Two additional cohorts with other advanced solid 2, A’Hern one-stage design will be set at one-sided type I binomial exact test of 5% to
tumors harboring FGFR2/3 alterations may be included based on signals detected in dose attain 80% power. The primary analyses will estimate ORR (H0: ORR#29% for TNBC and
escalation. The primary efficacy endpoint for Phase II is ORR per RECIST v1.1. The ORR#19% for HR+/HER2- tumors vs H1: ORR$55% for TNBC and ORR$41% for HR+/
preclinical data support the study of CGT4859 in this patient population with solid tumors HER2- tumors). The phase 2 part of the study will be deemed positive if at least 12
harboring FGFR2 and/or FGFR3 genetic alterations. The phase I dose escalation study is (46.2%) and nine (34.6%) pts with TNBC and HR+/HER2- tumors, respectively, achieve an
currently enrolling at sites in the United States. Clinical trial information: NCT06777316. objective response. Clinical trial information: NCT06612203. Research Sponsor:
Research Sponsor: Cogent Biosciences. Debiopharm. Gilead will provide the supply of SG.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 231s

TPS3173 Poster Session TPS3174 Poster Session

Trial in progress: First-in-human study of PFL-721/STX-721 in participants A phase 1/2 dose escalation study of the oral DNA polymerase theta
with locally advanced or metastatic non-small cell lung cancer harboring inhibitor (POLQi) GSK4524101 6 niraparib in adults with advanced or
EGFR exon 20 insertion mutations. First Author: Anas Gazzah, Gustave Roussy, metastatic solid tumors. First Author: Vivek Samnotra, GSK, Waltham, MA
Villejuif, France Background: In homologous recombination-deficient (HRd) tumors, use of a PARP
Background: Mutations in exon 20 of the EGFR gene account for approximately 4% to inhibitor (PARPi) leads to generation of DNA breaks that cannot be effectively repaired,
10% of all EGFR mutations in Non-Small Cell Lung Cancer (NSCLC). Most of these thus selectively killing cancer cells via synthetic lethality. An alternative DNA repair
mutations are insertions (EGFR ex20ins) that reduce the binding of first, second, and mechanism, microhomology-mediated end joining, is mediated by DNA polymerase
third generation tyrosine kinase inhibitors (TKI) to the ATP-binding pocket of the EGFR. theta (encoded by POLQ). In preclinical studies, POLQi + PARPi demonstrated superior
Amivantamab, a bispecific anti-EGFR/c-MET-receptor antibody, is approved for the efficacy vs PARPi alone in preventing HRd tumor growth. To evaluate the clinical
treatment of NSCLC with EGFR ex20ins mutations. However, there is significant unmet potential of this combination, this first-in-human study investigates treatment with
need for new oral agents that lack the limitations of intravenous administration and GSK4524101, an investigational POLQi, and niraparib, a PARPi, in patients with solid
associated infusion-related toxicities and possess improved target engagement, mutant tumors. Methods: This open-label, phase 1/2, multicenter study opened in October 2023
selectivity, and tolerability. PFL-721/STX-721 is an orally bioavailable, irreversible small- and includes a phase 1a/b, dose-escalation portion (part 1; potential enrollment to
molecule inhibitor targeting a broad range of EGFR- and HER2-activating ex20ins n»75). Sites in the US and Canada are enrolling patients for part 1, which aims to assess
mutations. PFL-721/STX-721 is highly selective for EGFR ex20ins mutations compared the maximum tolerated dose, pharmacokinetics (PK), and safety of oral GSK4524101 6
to wild type EGFR and exhibits greater selectivity compared to other EGFR mutant oral niraparib. Eligibility criteria include age $18 years, Eastern Cooperative Oncology
inhibitors. In addition, PFL-721/STX-721 has demonstrated superior anti-proliferation Group performance status of 0–2, life expectancy $3 months, and diagnosis of ad-
and antitumor effects compared to other investigational anti-EGFR ex20ins agents in vanced or metastatic solid tumor with all standard-of-care treatment options exhausted.
relevant tumor models in vitro and in vivo. These observations suggest a more robust Exclusion criteria include unresolved chemotherapy-induced adverse events (AEs) or
clinical risk-to-benefit profile and support further clinical investigation of PFL-721/STX- symptomatic uncontrolled brain or leptomeningeal metastases, uncontrolled hyper-
721. Methods: PFL-721/STX-721-101 (NCT06043817) is an open-label, first-in-human tension, history of myelodysplastic syndrome or acute myeloid leukemia, or another
(FIH), Phase 1/2 study evaluating the safety, tolerability, pharmacokinetic (PK) exposure, malignancy that has progressed or required active treatment in the past 2 years.
and preliminary antitumor activity of PFL-721/STX-721 in participants with locally Outcome measures include dose-limiting toxicity (DLT) incidence during the DLT ob-
advanced or metastatic NSCLC harboring EGFR/HER2 ex20ins mutations. It consists of servation periods (up to 28 days; primary); treatment-emergent AEs (TEAEs) and serious
3 parts: Part 1 Dose Escalation, Part 2 Recommended Phase 2 Dose (RP2D) selection, AEs (SAEs); percentage of patients receiving all planned doses; and percentage of
and Part 3 Dose Expansion. In Part 1, participants with NSCLC harboring EGFR or HER2 patients requiring AE-related dose interruptions, reductions, and discontinuations in the
ex20ins mutations will be enrolled into sequential cohorts to receive ascending oral DLT observation period. Secondary endpoints include the PK of niraparib and the
doses of PFL-721/STX-721 administered daily in 28-day treatment cycles. The main goal metabolite of GSK4524101 and incidence and duration of TEAEs and SAEs beyond the
is to identify the maximum tolerated dose (MTD) and optimal biological dose (OBD) of DLT observation period. The study is currently recruiting, with 17 patients having re-
PFL-721/STX-721. In Part 2, participants with NSCLC harboring EGFR ex20ins mutations ceived doses across 9 sites in 2 countries as of January 10, 2025. Clinical trial in-
who have received 1 to 2 prior lines of treatment, including a platinum-containing formation: NCT06077877. Research Sponsor: GSK.
chemotherapy regimen and excluding EGFR targeted therapies with the exception of
amivantamab, will be randomized 1:1 to receive PFL-721/STX-721 at the MTD or OBD in
order to determine the optimal RP2D. Finally, Part 3 will further test the anticancer
efficacy of PFL-721/STX-721 is administered at the RP2D. PFL-721/STX-721-101 is
actively enrolling at 18 sites in 7 countries globally. Clinical trial information:
NCT06043817. Research Sponsor: Scorpion Therapeutics, Inc.

TPS3175 Poster Session TPS3176 Poster Session

A phase 1 study to evaluate the safety, pharmacokinetics, and efficacy of the First in human phase 1 dose escalation and expansion clinical trial to
first-in-class cyclin A/B RxL inhibitor CID-078, an orally bioavailable, cell- evaluate the safety, pharmacokinetics and antitumor activity of intravenous
permeable macrocycle. First Author: Nehal J. Lakhani, The START Center for Cancer AROG4-01 in patients with advanced solid tumors. First Author: Sonia Macia,
Research, Grand Rapids, MI Applied Research using OMIC Sciences, Barcelona, Spain
Background: The cyclin-dependent kinase (CDK)-RB-E2F axis forms the core tran- Background: AROG4-01 is a synthetic compound with a first-in-class mechanism of
scriptional machinery driving cell cycle progression. Alterations in RB1 or other key action, targeting complex secondary structural elements in mRNA, including G-
components occur in many cancers, resulting in heightened oncogenic E2F activity. E2F quadruplexes(G4s). These secondary nucleic acid structures, characterized by
activation relies on the interaction between the cyclin’s conserved hydrophobic patch (HP) Hoogsteen base pairing, play pivotal roles in gene regulation and are abundant in cancer
and the RxL motif found on E2F and other cyclin/CDK substrates. Disrupting this cyclin A/ cells due totheir high proliferation rates and dysregulated gene expression patterns. By
E2F RxL interaction leads to hyperactivation of E2F and synthetic lethality in E2F-driven binding to G4s present in untranslated regions, AROG4-01 modulates gene expression
tumors. CID-078 is a novel, orally bioavailable, passively cell-permeable, potent and atthe post-transcriptional level, reducing tumor growth and survival. Preclinical studies
selective macrocycle that binds to the HP of cyclins A and B, blocking the RxL motif- have demonstrated that AROG4-01 achieves significant antitumor activity,inhibiting
mediated binding of E2F1 to cyclin A2-CDK2 and Myt1 to cyclin B1-CDK1. Consequently, cancer cell proliferation, with a strong effect of the compound on inhibit colony for-
CID-078 induces cell cycle arrest at the G2/M phase, leading to apoptotic tumor cell death. mation, evidencing the capacity of AROG4-01 to prevent the long-termsurvival and
In preclinical studies including small cell lung cancer (SCLC) and triple negative breast proliferation of cancer cells. This activity has been validated in vivo across multiple solid
cancer (TNBC) tumor types, CDX and PDX models with high E2F target pathway scores and cancer models. Methods: This study (NCT06652529,EudraCT2024-517569-18) is an
high E2F1 expression demonstrated tumor regression following single-agent CID-078 open label, Phase 1 dose escalation trial with two expansion cohorts to investigate the
treatment. Pre-clinical species demonstrate a well-tolerated safety profile and 20% oral safety, tolerability, pharmacokinetics,pharmacodynamics and preliminary antitumor
bioavailability. Preclinical to clinical predictions maintain a 20% bioavailability. activity of AROG4-01. The study consists of two parts. Part A is a dose escalation that
Methods: This is a phase 1, first-in-human, open-label, multicenter, dose escalation and will include 8-20 patients withadvanced solid tumors, covering up to 6 dose levels with
dose expansion study to evaluate the safety, tolerability, pharmacokinetics (PK) phar- the primary objective of determining the safety and tolerability of AROG4-01 and de-
macodynamics (PD) and preliminary anti-tumor efficacy of CID-078 in patients (pts) with fining an appropriaterecommended phase 2 dose (RP2D) for further evaluation in part B.
locally advanced or metastatic solid tumor malignancies (NCT06577987). Pts previously The study will start with an accelerated-titration dose escalation scheme enrolling one
treated with standard of care therapy and for whom no available curative therapy exists evaluablepatient per cohort for the first 2 dose levels followed by a classic 3+3 design.
are eligible. CID-078 will be administered orally, twice-daily in repeating 21-day cycles and Part B is a dose expansion, with two cohorts of ten patients: one cohort of patients
treatment will continue until disease progression, death, unacceptable toxicity or with- withadvanced mesothelioma (cohort 1) and a second cohort of patients with other solid
drawal from study. Part I dose escalation will be guided by a Backfill-Bayesian Optimal tumors (cohort 2). Serum samples collected from patients enrolled in part A when-
Interval Design (BF-BOIN) based on the incidence of dose-limiting toxicities (DLTs) and all receiving the first IMP dose during the first treatment cycle will be used to assess the PK
available safety and PK data. Under the BF-BOIN design, additional pts may be enrolled to
of AROG4-01. Three sites in Spain are expected to participate. Clinical trial information:
expand previous cohorts to better characterize the safety, PK, PD and preliminary efficacy
NCT06652529. Research Sponsor: Applied Research using Omic Sciences.
activity to support a recommended dose for expansion. A pilot food effect cohort is
planned as well. In Part II dose expansion, pts will be enrolled to one or more cohorts
defined by histologic tumor type or molecular alteration at the recommended doses of
expansion. Based on preclinical data generated to date, the study plans to include patients
with SCLC, TNBC, and RB1-mutated tumors with additional tumor types expanded based
on observed efficacy. Dose escalation is ongoing with no DLT reported in the initial 3 dose
cohorts evaluated. Clinical trial information: NCT06577987. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
232s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

TPS3177 Poster Session TPS3178 Poster Session

IMMUNONET: A multicenter, open-label, proof-of-concept phase II trial Trial in progress: Phase 1 study of the selective protein degrader ASP4396 in
evaluating NP137 as add-on therapy in advanced/metastatic solid tumors patients with locally advanced or metastatic solid tumors with KRAS G12D
treated with standard immunotherapies. First Author: Jérôme Fayette, Centre mutations. First Author: Shiraj Sen, NEXT Oncology, Dallas, TX
Léon Bérard, Lyon, France Background: KRAS G12D is the most common KRAS mutation at codon 12 found in solid
Background: PD-1/PD-L1 blockade has transformed oncology by offering durable re- tumors and is difficult to target. There are no approved therapies directly targeting KRAS
sponses in various cancers. However, many patients develop resistance, highlighting the G12D. Targeted protein degradation is emerging as a promising therapeutic approach for
need for novel therapeutic strategies. Epithelial-to-Mesenchymal Transition (EMT) undruggable targets. ASP4396, a novel protein degrader, targets KRAS G12D-mutated
plays a pivotal role in immune checkpoint inhibitor efficacy, with epithelial tumors protein for degradation via the ubiquitin-proteasome system. This mode of action may
exhibiting greater immunoreactivity than mesenchymal ones. NP137, a first-in-class offer higher efficacy and safety compared with inhibitors by blocking both enzymatic
anti-Netrin-1 monoclonal antibody, has shown in phase I study the ability to inhibit EMT, and scaffolding functions of proteins and by higher target selectivity. This first-in-human
potentially overcoming resistance (Cassier et al., Nature, 2023). This phase I data study aims to evaluate the safety and efficacy of ASP4396 in patients with advanced
demonstrated NP137’s ability to shift tumors toward an epithelial phenotype, supporting solid tumors with KRAS G12D mutations (NCT06364696). Methods: This Phase 1, open-
its combination with immune checkpoint inhibitor to sensitize tumors and alleviate label, multicenter, dose-escalation and dose-expansion study of ASP4396 is enrolling
resistance. The goal of IMMUNONET study (NCT05605496) is to evaluate NP137’s ability adult patients with locally advanced (unresectable) or metastatic solid tumors with
to re-sensitize advanced solid tumors to anti-PD-1/PD-L1 therapy. Methods: This proof- documented KRAS G12D mutations who have $ 1 measurable lesion per Response
of-concept study assess NP137 (14 mg/kg, IV, Q3W) as add-on therapy to standard PD- Evaluation Criteria in Solid Tumors (RECIST) version (v)1.1, ECOG performance status of
1/PD-L1 inhibitors across three independent cohorts of patients with advanced/ 0 or 1, adequate organ function, and who did not respond or who are ineligible for
metastatic solid tumors of any histological types: Cohort 1 (Stable Disease [SD]): standard therapies. Tumor-specific dose expansion cohorts may be enrolled at the
Radiological SD after $12 weeks of anti-PD-1/PD-L1 therapy. Cohort 2 (Primary Re- maximum tolerated dose (MTD) and/or candidate recommended phase 2 dose (RP2D).
fractory): Radiological progressive disease (PD) and no response under anti-PD-1/PD-L1 Patients who received prior treatment targeting KRAS G12D will be excluded. Primary
therapy. Cohort 3 (Secondary Refractory): Radiological PD following initial response endpoints are safety and tolerability (assessed by dose-limiting toxicities [DLTs], ad-
under anti-PD-1/PD-L1 therapy. Treatment continues until progression, unacceptable verse events, laboratory and other standard tests), and RP2D and/or MTD of ASP4396.
toxicity, or consent withdrawal. The primary endpoint is clinical activity: objective Secondary endpoints are antitumor activity (objective response rate, duration of re-
response rate (ORR)-12W for cohort 1 and progression-free rate (PFR)-12W for cohorts 2 sponse, disease control rate, and progression-free survival per RECIST v1.1 by inves-
and 3. Secondary endpoints include ORR-12W (cohorts 2 and 3), Time to Objective tigator assessment; and overall survival), and pharmacokinetic/ pharmacodynamic
Response (ToR), Duration of Response (DoR) and safety for all cohorts. Evolution of assessments. In the dose escalation cohort, patients will receive increasing doses of
EMT, Netrin-1, and receptor expression will be analysed and correlated with clinical ASP4396 intravenously in a 21-day cycle. The target enrollment for each dose level is set
outcomes. An adaptive 2-stage design is being used for this study (Lin and Shih, at 1 DLT-evaluable patient for dose levels 1–3 and $ 3 DLT-evaluable patients for each
Biometrics 2004). The target levels of clinical activity are set at 20% (relevant) and 25% subsequent dose level. The study will consist of 3 periods: screening (up to 28 days),
(high). In stage 1, 18 patients will be enrolled at 1-sided alpha of 5%. Depending on the treatment (every 21-day cycle until treatment discontinuation criteria are met), and
observed success rate, additional 11 patients (if 1 or 2 successes) or 5 patients (if . 2 follow-up. Data will be summarized descriptively (mean, standard deviation, median) for
successes) could be recruited into stage 2. Null hypotheses will be rejected if $4 continuous endpoints, and by counts and percentages for categorical endpoints. Study
successes are observed in 29 [test p0 = 0.05 vs. 0.20, 80% power] or 23 patients [test p0 = enrollment is ongoing. Clinical trial information: NCT06364696. Research Sponsor:
0.05 vs. 0.25, 90% power], respectively. Current Status: Cohort 1 has been closed due to Astellas Pharma Inc.
non-feasibility. Prespecified goals for the first stage were met, stage 2 enrolment is
underway. Cohort 2 has enrolled 21 patients, and cohort 3 has enrolled 19 of 23 planned
evaluable patients. Clinical trial information: NCT05605496. Research Sponsor: Euro-
pean Innovation Council.

TPS3180 Poster Session TPS3181 Poster Session

Phase 1 first-in-human clinical trial of AG01, a recombinant monoclonal Trial in progress: First-in-human study of ATX-559, an oral inhibitor of
antibody to progranulin/glycoprotein 88 (PGRN/GP88), to determine the DHX9, in patients with advanced or metastatic solid tumors, and molecularly
safety, tolerability, pharmacokinetics, and preliminary anti-tumor response defined cancers. First Author: Meredith Pelster, Sarah Cannon Research Institute,
in subjects with advanced solid tumor malignancies. First Author: Nashville, TN
Katherine H. R. Tkaczuk, University of Maryland Greenebaum Comprehensive Cancer Background: DHX9 is a multifunctional RNA helicase that is involved in the mainte-
Center, Baltimore, MD nance of genomic stability by resolving DNA/RNA secondary structures that may lead to
Background: GP88/PGRN is the largest member of the granulin/epithelin family. We DNA replication stress and DNA damage. High expression of DHX9 is evident in multiple
demonstrated GP88’s role as an autocrine growth & survival factor in breast cancer (BC): in cancer types. ATX-559, an oral inhibitor of DHX9, has been shown preclinically to induce
ER+BC cells, GP88 stimulates proliferation & confers resistance to anti-estrogen therapy & robust anti-tumor activity of a variety of different solid tumors with genomic instability,
aromatase inhibitors;GP88 is expressed in 80% of invasive ductal carcinomas & is negative in including models with BReast CAncer gene 1 and/or 2 alterations or deficiency (BRCA
normal mammary tissue; GP88 tumor expression is a prognostic indicator of recurrence & deficient) and microsatellite instability-high (MSI-H) and/or deficient mismatch repair
death in BC pts; Elevated GP88 serum level in metastatic BC patients (pts) is associated with (dMMR). Methods: This is a first-in-human, Phase 1, open-label, single-arm, dose-
disease progression. PGRN/GP88 is overexpressed in several other solid tumors (non-small escalation and expansion study to evaluate the safety profile of ATX-559 and to de-
cell lung carcinoma, colorectal, bladder, ovarian, prostate & brain). In advanced NSCLC & termine the recommended phase 2 dose (RP2D). In dose-escalation, patients with
prostate pts, elevated serum PGRN/GP88 have been found. These results make GP88/PGRN
locally-advanced or metastatic solid tumors, and molecularly-defined cancers will be
an ideal therapeutic & diagnostic target in BC and other solid tumors. An anti-human PGRN/
enrolled for safety assessment, guided by a model-assisted dose escalation design
GP88 monoclonal antibody (AG01) inhibiting PGRN/GP88 action was developed & expressed
as recombinant antibody in CHO cells. Pharmacology, GMP manufacturing, formulation,
(Yuan, 2019) to identify an acceptable dose. To assess evidence of preliminary antitumor
stability studies & GLP toxicology studies in non-human primates were done. The IND activity in the expansion study, participants with (1) BRCA deficient, HER2-negative,
application was cleared by the FDA to proceed with the first-in-human (FIH) AG01 study in metastatic breast cancer, and (2) dMMR/MSI-H solid tumors will be enrolled using a
adult subjects with advanced solid tumors. Methods: This IRB approved FIH study, will be Simon 2-stage design (Simon, 1989). Primary endpoints include identification of the
conducted in 2 stages, dose escalation (1A) and dose expansion (1B). The 1A part is ongoing, RP2D dose that is deemed acceptable per the model-assisted dose escalation design
with the 1 + (3+3) design. In the 1A part the AG01 is administered intravenously (IV) over 90 and to evaluate safety and tolerability as noted by the frequency and severity of adverse
min. every 14 days +/- 1 day, 1 cycle = 28days, DLT assessments occur in the first 28 days of events (AEs). Secondary endpoints will evaluate pharmacokinetics (PK), pharmaco-
treatment. Five dose levels of AG01 & a -1 level are planned (level -1-0.5mg/kg, & 1mg/kg, dynamics (PD) peripherally and in a biopsy sub-study, and preliminary anti-tumor activity
2mg/kg, 4mg/kg, 6mg/kg, 8 mg/kg). In 1A part of the study, initially an accelerated titration per RECIST v1.1. Exploratory objectives will explore potential biomarkers in relationship
design (1pt/dose level) was utilized to guide dose progression & estimation of the maximum to ATX-559 exposure, as well as those that may correlate with treatment outcomes. A
tolerated and/or administered dose (MTD/MAD). Eligibility criteria for 1A part include pts with randomized cohort has also been included during dose expansion in recognition of
advanced relapsed/refractory solid tumor malignancies who failed 1 or more standard of care Project Optimus. The study is open and enrollment is ongoing. Clinical trial information:
(SOC) therapies or for whom no SOC treatment exists or is not tolerated, at least 1 RECIST1.1 NCT06625515. Research Sponsor: Accent Therapeutics.
measurable lesion, ECOG , = 2, Life expectancy . = 12wks, adequate organ & bone marrow
function, willing to sign informed consent & follow study procedures. Primary objective (1A) is
to determine the MTD and/or MAD of AG01. Secondary objectives: to determine the rec-
ommended phase 2 dose (RP2D), safety, tolerability, the PKs, immunogenicity & the pre-
liminary anti-tumor activity of AG01. Exploratory objectives:todetermine PGRN/GP88
expression in tumor tissue & PGRN/ GP88 blood levels (A&G’s IHC & ELISA test). This study is
registered at NCT05627960. The study is supported by NCI grants NCI R44 CA224718 &
CA162629. Clinical trial information: NCT05627960. Research Sponsor: National Cancer
Institute; National Cancer Institute.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY 233s

TPS3182 Poster Session TPS3183 Poster Session

A phase 1/2 study of JK06, a 5T4 antibody drug conjugate, in patients with A phase I/Ib study of olaparib and ASTX727 in BRCA 1/2- and HRD-mutated
unresectable locally advanced or metastatic cancer. First Author: Nuria Kotecki, tumors. First Author: Pamela N. Munster, UCSF Helen Diller Family Comprehensive
Université libre de Bruxelles (ULB), Hôpital Universitaire de Bruxelles (H.U.B), Institut Cancer Center, San Francisco, CA
Jules Bordet, Brussels, Belgium Background: Patients with germline or somatic HRR pathway mutations often develop
Background: 5T4, a Type I transmembrane glycoprotein, plays a pivotal role in neonatal resistance despite initial response. Overlapping toxicities hinder combination strategies
development, but its expression in normal adult tissues is limited. In contrast, 5T4 in breast, ovarian, prostate, and pancreatic cancers, creating a need for safer and more
emerges prominently in a broad spectrum of solid tumors, including but not limited to effective approaches. Preclinical studies have shown that DNMT inhibition enhances
NSCLC, breast, ovarian, endometrial, bladder, pancreatic, esophageal, gastric and co- PARP inhibitor efficacy by promoting PARP trapping on DNA. This phase I study aims to
lorectal cancers. Furthermore, the expression of 5T4 is confirmed to be associated with assess the safety and tolerability of olaparib and AST727 in HRR-mutated patients and
advanced disease and worse clinical outcomes in multiple solid tumors. These features establish the RP2D for a phase II trial, to be supported by the NCI ComboMatch program.
make 5T4 an attractive, but as of yet unexploited, target for cancer therapeutics. JK06 is Correlative studies include the creation of PDX and organoid models for ex vivo analysis
an antibody-drug-conjugate (ADC) targeting 5T4-expressing cancer cells. The antibody of therapy response. Methods: Further studies include cfDNA and tumor tissue as-
moiety of JK06 has a high-affinity tetravalent binding capacity, compensating for sessment to elucidate mechanisms of resistance (reversion mutations, epigenetic
generally low 5T4 expression levels. Further, the JK06 binding specificity is biparatopic, markers) and PD markers of HR pathway modulation. Rad51 foci will be measured to
targeting two non-overlapping epitopes on 5T4 antigens. In this way, JK06 cross-links determine DNA repair function and CHIP assays (clonal hematopoiesis of indeterminate
5T4 on the surface of cancer cells, which enhances internalization and increases in- potential) to study the differential rate of CHIP as an early event in the evolution of AML/
tracellular release of the cytotoxic payload. The cytotoxic payload of JK06 is the MDS. Trial Design: This is a single center phase I/Ib clinical trial evaluating the
clinically proven microtubule-disrupting agent, MMAE, that inhibits cell division by combination of olaparib and ASTX727 (an oral formulation of decitabine with
preventing the polymerization of tubulin, leading to cell cycle arrest and apoptosis. JK06 cedazuridine, a cytidine deaminase inhibitor that allows for oral administration). All
mediates cytotoxicity in vitro, in a 5T4 receptor density dependent manner, and anti- participant enrollment and study participation will be conducted at UCSF as single site
tumor activity has been demonstrated in several murine xenograft models. JK06 has trial with collaboration from other centers for correlative/exploratory objectives. The
been shown to bind to recombinant human and cynomolgus 5T4, supporting the phase I dose escalation portion will follow a standard 3+3 design for enrollment and will
translation of pre-clinical toxicology studies. Preclinical toxicology studies showed no include adults with advanced/metastatic solid tumor malignancies with germline or
toxicity with JK06 at dose levels up to 17 mg/kg single dose and 9 mg/kg repeat dose. somatic mutations in the HRR pathway (i.e., BRCA1/2, PALB2, ATM, and/or CHEK2
Toxicokinetic analysis and PK modeling suggest that a Q3W dosing regimen should mutations). Patients will be treated in 2 escalating cohorts with a 12 patient phase Ib
provide adequately sustained exposure in clinical studies. In summary, preclinical dose expansion in the same population. At least 6 of 12 expansion patients must have
studies support clinical development of JK06 for the treatment of multiple 5T4 germline HRD mutations. Key Eligibility: The participant must have histologically
expressing solid tumors. Methods: The Phase 1/2 study of JK06 will enroll patients with confirmed advanced solid tumors with a germline and/or somatic mutation in one or
advanced relapsed/refractory solid tumors. The study will employ a 3+3 escalation more of the following genes: BRCA1/2, PALB2, ATM, and/or CHEK2. Patients must have
design to explore the safety, PK and preliminary anti-tumor activity of JK06. Back-fill adequate organ function and recovered from prior treatment associated toxicities. Prior
enrollment at specific dose levels is permitted but mandates fresh tumor biopsy. treatment with PARP inhibitors is allowed if the participant has not required dose
Patients will receive treatment with JK06 intravenously once every three weeks until reductions or delays due to toxicity. Participants with treated brain metastases are
confirmed disease progression or intolerable toxicity. Tumor specific expansion cohorts eligible if follow-up brain imaging shows no evidence of progression for at least 4 weeks.
will be initiated once dose and schedule are established from dose escalation; fresh Individuals with a prior or concurrent malignancy are eligible, however participants
tumor biopsies will also be collected from patients enrolled in expansion cohorts. diagnosed with MDS or AML are excluded from the study. Trial Status: The study is
Response will be assessed every 9 weeks per RECIST v1.1. Clinical trial information: ongoing and 4 patients have been enrolled to date. Clinical trial information:
NCT06667960. Research Sponsor: None. NCT06177171. Research Sponsor: National Cancer Institute/U.S. National Institutes of
Health.

TPS3184 Poster Session TPS3185 Poster Session

225
RYZ101 ( Ac-DOTATATE) in patients with estrogen receptor-positive, panSOHO: Phase II trial of BAY 2927088 in patients with unresectable or
human epidermal growth factor receptor 2–negative, locally advanced metastatic solid tumors other than NSCLC with HER2-activating mutations.
and unresectable, or metastatic breast cancer progressing after prior ther- First Author: Vivek Subbiah, Sarah Cannon Research Institute, Nashville, TN
apy: The phase 1b/2 TRACY-1 study. First Author: Erica L. Mayer, Medical On- Background: Human epidermal growth factor receptor 2 (HER2) gene mutations occur
cology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA in approximately 3.5% of solid tumors, with a frequency varying from less than 1% to 9%,
Background: RYZ101 (actinium-225 [225Ac]-DOTATATE) is a radiolabeled somatostatin depending on the tumor type. BAY 2927088 is an oral, reversible tyrosine kinase inhibitor
analog (SSA) for the treatment of patients with solid tumors expressing somatostatin that potently inhibits HER2 and mutant epidermal growth factor receptor and has shown
receptor-type 2 (SSTR2). RYZ101 is composed of the alpha-emitting radioisotope 225Ac, clinical benefit based on preliminary evidence from the Phase I/II SOHO-01 trial in
the chemical chelator DOTA (tetraxetan), and SSA octreotate (TATE). RYZ101 binds with patients with HER2-mutant non-small cell lung cancer (NSCLC; PL04.03 presented at
high affinity to SSTR2 on the cell surface and is internalized, whereupon the alpha-particle IASLC 2024 World Conference on Lung Cancer), an indication for which the FDA has
emission of 225Ac results in lethal double-strand DNA breaks. Although SSTR-directed granted Breakthrough Designation. Here we introduce the panSOHO trial evaluating the
therapy is widely used in patients with well-differentiated gastroenteropancreatic neu- efficacy and safety of BAY 2927088 in patients with unresectable, locally advanced or
roendocrine tumors (GEP-NETs), its relevance in non-GEP-NET SSTR-expressing neo- metastatic solid tumors with HER2-activating mutations. Methods: panSOHO is a Phase
plasms is still emerging. Clinical positron emission tomography (PET) imaging studies II, open-label, multicenter, multinational, single-arm basket trial of BAY 2927088 in
have reported SSTR expression in estrogen receptor (ER)-positive breast cancer. Available patients with unresectable or metastatic solid tumors with HER2-activating mutations
data support investigating the efficacy of RYZ101 in patients with ER-positive, HER2- (NCT06760819), and will be conducted in the USA, Europe, and the Asia-Pacific region.
negative, locally advanced and unresectable or metastatic breast cancer. Eligibility criteria include patients aged $18 years with: documented histologically or
Methods: TRACY-1 (NCT06590857) is a global, multicenter, open-label, two-part (dose cytologically confirmed, locally advanced or metastatic solid tumor cancer (colorectal,
escalation and expansion) phase 1b/2 study. Key inclusion criteria are: age $18 years; biliary tract, bladder and urothelial tract, cervical, endometrial, or other solid tumor);
histologically confirmed, ER-positive, HER2-negative locally advanced and unresectable or documented activating HER2 mutation; $1 measurable lesion per RECIST v1.1; and
metastatic breast cancer not amenable to curative-intent treatment; endocrine-refractory previous standard therapy or no satisfactory alternative treatment options. Key ex-
disease; documented progression (per RECIST v1.1) after $2 and #4 prior lines of clusion criteria include primary diagnosis of NSCLC, treatment with a HER2 tyrosine
chemotherapy and/or ADC ($1 must be ADC if the patient is a candidate for ADCs and kinase inhibitor, untreated active brain metastases, and leptomeningeal disease. Overall,
treatment is available); $1 RECIST-measurable SSTR2PET-positive lesion and $80% of
111 eligible patients will receive BAY 2927088 p.o. 20 mg twice daily in 3-week cycles
RECIST-measurable lesions being SSTR2PET-positive on screening scan. Key exclusion
until disease progression, unacceptable toxicity, or study withdrawal. The primary
criteria are: prior radiopharmaceutical therapy; prior anticancer therapy or external beam
outcome is BAY 2927088 efficacy on objective response rate per RECIST v1.1 as
radiotherapy in past 4 weeks; anticancer hormonal treatments in past 2 weeks. Primary
assessed by blinded independent central review (BICR). Secondary outcomes include
objectives are to determine the recommended phase 2 dose (R2PD) of RYZ101 (dose
escalation; anticipated 6224 patients), and the efficacy of RYZ101 at the RP2D defined as
BAY 2927088 efficacy on time to response, duration of response, disease control rate,
ORR as determined by BICR (dose expansion; approximately 100 patients). During dose and progression-free survival per RECIST v1.1 by BICR, and overall survival, and BAY
escalation, patients will receive RYZ101 by IV infusion every 6 weeks for up to 6 cycles at a 2927088 safety and tolerability. Impact of BAY 2927088 on patient quality of life will be
starting dose of 6.5 MBq (dose level [DL] 1), with escalation to DL 2 (8.3 MBq) and DL 3 evaluated by EORTC QLQ-C30. Enrollment is open. Clinical trial information:
(10.2 MBq), or dose de-escalation to 4.6 MBq if DL 1 is not tolerated, based on dose-limiting NCT06760819. Research Sponsor: Bayer AG.
toxicity rates. In the expansion phase, patients will receive RYZ101 at the RP2D. Con-
comitant amino acid IV infusions (containing L-arginine and L-lysine) will be co-infused
with RYZ101 for renal protection. The study is ongoing and enrolling patients in the USA.
Clinical trial information: NCT06590857. Research Sponsor: RayzeBio.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
234s DEVELOPMENTAL THERAPEUTICS—MOLECULARLY TARGETED AGENTS AND TUMOR BIOLOGY

TPS3186 Poster Session TPS3187 Poster Session

Molecular residual disease (MRD) in solid tumors. First Author: Majd T. Ghanim, Beamion PANTUMOR-1: A phase II, multicenter, multicohort, open-label trial
Flatiron Health, New York, NY to evaluate the efficacy and safety of the oral HER2-selective tyrosine
Background: Pragmatically designed clinical studies facilitate rapid accrual of representative kinase inhibitor zongertinib for the treatment of HER2-mutated or
populations by aligning research with routine care and enabling study execution in community overexpressed/amplified solid tumors. First Author: Alison M. Schram,
practice settings. In addition, the implementation of technologies to streamline patient as- Memorial Sloan Kettering Cancer Center, New York, NY
certainment and data collection further reduce site burden and improve efficiency. Herein we Background: While it is well known that HER2 overexpression, amplification, and mutation drives
describe the initial cohort under a platform study designed with pragmatic elements initiated various tumors, there remains an unmet need for effective, oral, HER2-targeted therapies. Zon-
within a technology-enabled community oncology research network. This substudy gertinib is an irreversible tyrosine kinase inhibitor (TKI) that selectively inhibits HER2 while sparing
establishes a prospective observational registry that collects routinely documented clinical EGFR, thereby limiting associated toxicities. In the ongoing Phase Ia/Ib trial (NCT04886804),
data plus intentionally collected biomarker samples, including blood, for the purpose of zongertinib showed manageable safety and confirmed responses in patients (pts) with HER2-
isolating circulating tumor DNA at specified intervals to enable exploration of MRD in patients overexpressed/amplified and HER2-mutant tumors (Wilding et al, Cancer Discov. 2024). Based on
with early stage solid tumors. Methods: This is a prospective, multicenter, observational, these encouraging data, the Beamion PANTUMOR-1 basket trial (NCT06581432) is evaluating the
biospecimen collection study in participants (pts) diagnosed with early stage cancers in select efficacy and safety of zongertinib monotherapy in pts with HER2-mutant or HER2-overexpressed/
solid tumors who have planned curative-intent surgery. The scientific objective is to collect amplified solid tumors. Methods: In this global Phase II basket trial, ~200 pts with HER2-driven
tumor tissue, longitudinal blood samples, and associated clinical data to explore applications (HER2-mutant or HER2-overexpressed/amplified) tumors will be enrolled at ~60 sites in 13
of blood and/or tissue-based cancer biomarkers for cancer detection, prognosis, therapy countries. Pts will be enrolled to 10 cohorts: 8 cohorts of specific tumor types and 2 tumor-agnostic
selection, surveillance, and therapy response. Approximately 1350 pts will be enrolled across cohorts (see Table). The specific tumor type cohorts will initially recruit 10 pts, with potential for
expansion to up to 20 pts after an interim analysis. In the tumor-agnostic cohorts, 20 pts will be
~30 Flatiron Research Network community oncology sites. Participants are grouped by tumor
recruited directly without an interim analysis. Pts will receive 120 mg zongertinib until disease
site of origin and histology into 7 cohorts (Table). Patients provide informed consent and are
progression, unacceptable toxicity, or withdrawal. Patients must be $18 years old, have docu-
enrolled before starting neoadjuvant or adjuvant therapy. Study visits correspond with routine mented HER2-positive (HER2-overexpressed/amplified) status or a HER2 mutation (established by
care. Research tissue and blood samples are obtained upon enrollment and at study-specified local testing), $1 measurable lesion outside the central nervous system, an ECOG performance
intervals up to 5.5 years or until disease recurrence for analysis by Exact Sciences laboratories. score of 0 or 1, and have progressed following prior treatment or have no alternative treatment
Technology enablement includes near real-time, AI-assisted, centralized patient ascertainment options. Exclusion criteria include HER2-mutant non-small cell lung cancer (NSCLC) and previous/
and integrated electronic health record-to-electronic data capture system data transfer. Under concomitant malignancies. Primary endpoint is objective response, as assessed by central inde-
the parent protocol mechanism, the study was IRB approved 65 days from commencement of pendent review according to RECIST v1.1. Secondary endpoints include duration of response,
protocol writing. Target enrollments are based on the number needed to enroll to observe at progression-free survival, disease control, occurrence of treatment-emergent adverse events, and
least 30 events in 3 years. Clinical trial information: NCT06605404. Research Sponsor: None. health-related quality of life. Enrollment is ongoing. Clinical trial information: NCT06581432.
Study cohorts.
Research Sponsor: Boehringer Ingelheim.
Disease Target HER2 overexpression/ HER2 mutation
Tumor type stage enrollment amplification cohorts Tumor type cohorts Tumor type
Muscle invasive urothelial carcinoma II-III 200 Cohort 1 Urothelial cancer Cohort 7 Urothelial cancer
Esophageal I-III 150 Cohort 2 Biliary tract cancer Cohort 8 Breast cancer
Gastric & gastroesophageal junction I-III 150 Cohort 3 Uterine cancer Cohort 9 Gastroesophageal
Melanoma II-III 300 cancer
Non–small cell lung cancer I-III 200 Cohort 4 Cervical cancer Cohort 10 Other HER2-mutant
Exocrine pancreatic cancer I-III 150 solid tumors†
Other solid tumors (excluding central nervous system, colorectal, II-III 200 Cohort 5 Non-squamous NSCLC
breast, skin squamous and basal cell, gastrointestinal stromal Cohort 6 Other HER2 overexpressed/
tumors, thyroid, uveal melanoma, and low or intermediate grade amplified solid tumors*
neuroendocrine tumors)
*Except breast cancer, gastric, gastroesophageal junction, or esophageal adenocarcinoma.
†
Except NSCLC.

TPS3188 Poster Session

Perfume trial: Phase II trial of binimetinib in patients with BRAF fusion-
positive low-grade glioma or pancreatic cancer. First Author: Tomoyuki Satake,
Department of Hepatobiliary and Pancreatic Oncology, National Cancer Center Hospital
East, Kashiwa, Japan
Background: BRAF fusion was reported to be a rare mutation found in 0.3% of all solid
tumors, but a high percentage of BRAF fusion has been reported in pilocytic astrocytoma
(30-77%) and pancreatic acinar cell carcinoma (24-67%). Although treatment for BRAF
V600 mutations has been developed, treatment for BRAF fusion has not yet been
established. Recently, tovorafenib has been granted accelerated approval by the FDA for
pediatric low-grade glioma (LGG) with BRAF alteration (including BRAF fusion). Still, it is
not yet approved in Japan, and an unmet need exists. In BRAF fusion-positive solid
tumors, the constitutively activated BRAF kinase domain forms dimers that cause
activation of the MAPK pathway. MEK inhibitors have been reported to show anti-tumor
effects against BRAF fusion-positive cell lines. Phase I/II trials with selumetinib or
binimetinib have shown efficacy in patients with BRAF fusion-positive LGG.
Methods: Perfume trial (NCCH2101/MK011) is an open-label, parallel, 2-cohort, mul-
ticenter, phase II, investigator-initiated registration-directed clinical trial to evaluate the
efficacy and safety of binimetinib in patients with advanced or recurrent LGG or
pancreatic cancer (PC) harboring BRAFfusion/rearrangement. Sample sizes of 16 and 11
patients are needed for LGG and PC at a one-sided significant level of 5% to achieve 85%
and 70% power, respectively. Key eligibility criteria for LGG (grade 1 and grade 2 tumors
according to WHO classification) include ageS12 (body weight S40 kg in 12-17 year
old) and KPS/LPSS70, regardless of history of cancer drug therapy. Key eligibility
criteria for PC include ageS12 (body weight S40 kg in 12-17 year old); ECOG PS 0-1;
refractory or intolerant to at least one prior cancer drug therapy. Enrolled patients
receive binimetinib 45mgadministered orally twice daily. The primary endpoint is the
objective response rate (ORR) using RECIST 1.1 by independent central review. The
secondary endpoints include ORR by investigators’ assessment, ORR by RANO in LGG,
progression-free and overall survivals, disease control rate, duration of response, and
safety. This study implemented a decentralized clinical trial system for patients living in
remote areas to reduce their time and economic burden. Enrollment started in March
2023 and is ongoing at 6 facilities in Japan. As of Dec 2024, 6 patients with LGG and
3 patients with PC were enrolled. Clinical trial information: jRCT2031230007,
NCT06159478. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 235s

LBA3500 Oral Abstract Session 3501 Oral Abstract Session

First-line encorafenib + cetuximab + mFOLFOX6 in BRAF V600E-mutant Nivolumab (NIVO) plus ipilimumab (IPI) vs chemotherapy (chemo) or NIVO
metastatic colorectal cancer (BREAKWATER): Progression-free survival and monotherapy for microsatellite instability-high/mismatch repair-deficient
updated overall survival analyses. First Author: Elena Elez, Department of Medical (MSI-H/dMMR) metastatic colorectal cancer (mCRC): Expanded analyses
Oncology, Vall d’Hebron University Hospital (HUVH), Vall d’Hebron Institute of Oncology from CheckMate 8HW. First Author: Heinz-Josef Lenz, University of Southern
(VHIO), IOB-Quiron, UVic-UCC, Barcelona, Spain California Norris Comprehensive Cancer Center, Los Angeles, CA
Background: In the phase 3 CheckMate 8HW study (NCT04008030), both dual primary endpoints of
progression-free survival (PFS) for first-line (1L) NIVO + IPI vs chemo (HR 0.21; P , 0.0001) and NIVO + IPI
vs NIVO across all lines (HR 0.62; P = 0.0003) in patients (pts) with centrally confirmed MSI-H/dMMR mCRC
were met. We report expanded analyses of NIVO + IPI vs NIVO (all lines) and longer follow-up results for
NIVO + IPI vs chemo (1L). Methods: The study design was described previously. Pts with MSI-H/dMMR per
local testing were enrolled. After randomization, IHC and PCR based tests were used for central con-
firmation. PFS2 (time from randomization to progression after subsequent systemic therapy, start of
second subsequent systemic therapy, or death) was a key exploratory endpoint. Results: In all randomized
pts (all lines), 296 of 354 (84%) in the NIVO + IPI arm, 286 of 353 (81%) in the NIVO arm, and 113 of 132
(86%) in the chemo arm had centrally confirmed MSI-H/dMMR. In all randomized 1L pts, 171 of 202 (85%) in
the NIVO + IPI arm and 84 of 101 (83%) in the chemo arm had centrally confirmed MSI-H/dMMR. Median
follow-up was 47.0 mo (range 16.7–60.5). 1L NIVO + IPI continued to show PFS benefit vs chemo (Table).
Subsequent systemic therapy was received by 27 (16%) and 61 (73%) pts after 1L NIVO + IPI and chemo,
The full, final text of this abstract will be available at respectively; 10 (6%) and 21 (25%) received subsequent non-study immunotherapy. In the 1L chemo arm,
39 (46%) pts crossed over to NIVO + IPI on study. PFS2 continued to favor 1L NIVO + IPI vs chemo (Table).
[Link] on the day of presentation and in the Across all lines, NIVO + IPI demonstrated superior PFS vs NIVO (Table). Subsequent systemic therapy was
online supplement to the June 10, 2025, issue of the Journal received by 54 (18%) pts in the NIVO + IPI arm and 83 (29%) in the NIVO arm; 20 (7%) and 31 (11%) received
subsequent non-study immunotherapy. PFS2 favored NIVO + IPI vs NIVO across all lines of therapy (Table).
of Clinical Oncology. In all treated pts, grade 3/4 treatment-related adverse events occurred in 78 (22%) and 50 (14%) pts in the
NIVO + IPI and NIVO arms, respectively. Additional analyses will be presented. Conclusions: NIVO + IPI
demonstrated sustained clinical benefit vs chemo (1L) and NIVO (all lines) despite use of subsequent
therapy, as shown by improved PFS2 in pts with centrally confirmed MSI-H/dMMR mCRC. No new safety
signals were observed. These results support NIVO + IPI as a standard of care treatment for MSI-H/dMMR
mCRC. Clinical trial information: NCT04008030. Research Sponsor: Bristol Myers Squibb.
NIVO + IPI Chemo
Centrally confirmed MSI-H/dMMR (1L) (n = 171) (n = 84)
Median PFS (95% CI), mo 54.1 (54.1–NE) 5.9 (4.4–7.8)
HR (95% CI) 0.21 (0.14–0.31)
Median PFS2 (95% CI), mo NR (NE–NE) 30.3 (15.2–NE)
HR (95% CI) 0.28 (0.18–0.44)
NIVO + IPI NIVO
Centrally confirmed MSI-H/dMMR (all lines) (n = 296) (n = 286)
Median PFS (95% CI), mo NR (53.8–NE) 39.3 (22.1–NE)
HR (95% CI) 0.62 (0.48-0.81); P = 0.0003
Median PFS2 (95% CI), mo NR (NE–NE) NR (NE–NE)
HR (95% CI) 0.57 (0.42–0.78)

NE, not evaluable; NR, not reached.

LBA3502 Oral Abstract Session 3503 Oral Abstract Session

Anlotinib versus bevacizumab added to standard first-line chemotherapy ctDNA-guided adjuvant chemotherapy escalation in stage III colon cancer:
among patients with RAS/BRAF wild-type, unresectable metastatic colo- Primary analysis of the ctDNA-positive cohort from the randomized AGITG
rectal cancer: A multicenter, prospective, randomised, phase 3 clinical trial dynamic-III trial (intergroup study of AGITG and CCTG). First Author: Jeanne Tie,
(ANCHOR trial). First Author: Ke-Feng Ding, Department of Colorectal Surgery and Department of Medical Oncology, Peter MacCallum Cancer Centre and Personalised
Oncology, Key Laboratory of Cancer Prevention and Intervention, Ministry of Education, Oncology Division, Walter and Eliza Hall Institute of Medical Research, Melbourne,
The Second Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China Australia
Background: Despite adjuvant chemotherapy (ACT) a proportion of patients (pts) with stage III
colon cancer (CC) will recur. Most at risk are those with detectable ctDNA, whereas those with
undetectable ctDNA have a reduced recurrence risk. The DYNAMIC-III study explored the impact
of ACT de-escalation or escalation as informed by post-surgery ctDNA results. Here, we report the
primary analysis on the impact of treatment escalation in ctDNA-positive pts. Outcome data for
treatment de-escalation in ctDNA-negative pts is immature. Methods: DYNAMIC-III is a multi-
center, randomized, phase II/III trial. Eligible pts had resected stage III CC and were fit for ACT. Pts
were randomly assigned 1:1 to ctDNA-informed or standard of care (SOC) management. Clinicians
nominated the selected SOC ACT regimen prior to randomization. For ctDNA-informed
management, a ctDNA-positive result at 5-6 weeks after surgery with a tumor-informed assay
prompted an escalation ACT strategy (from single agent fluoropyrimidine [FP] to oxaliplatin-based
The full, final text of this abstract will be available at doublet, from 3 months doublet to 6 months doublet or FOLFOXIRI [clinician choice], or from
[Link] on the day of presentation and in the 6 months doublet to FOLFOXIRI). The primary efficacy endpoint for the ctDNA-positive cohort was
2-year RFS. The target sample size of 250 provided 80% power with 90% confidence to confirm
online supplement to the June 10, 2025, issue of the Journal superiority of ctDNA-informed treatment escalation compared to SOC with a HR of 0.746.
of Clinical Oncology. Results: Of 961 eligible pts randomized between Oct 2017 and Apr 2023, 259 (27%) were ctDNA-
positive. Of these, 113 (44%) had clinical low risk disease (non-N2 + non-T4). Median follow-up
was 42.2 months (range 0.78 – 63.0). 115 (89%) of 129 ctDNA-informed pts received ACT
escalation, with 65 (56%) receiving FOLFOXIRI. Of 130 SOC pts, 14 (11%) and 112 (86%) received
single agent FP and oxaliplatin doublet, respectively. 2-year RFS for ctDNA-informed treatment
escalation was 52% (90% CI: 44 - 59%) vs 61% (90% CI: 54 - 68%) for SOC (HR 1.11, 90% CI: 0.83 -
1.48; P = 0.6). The 3-year RFS for ctDNA-positive pts receiving FOLFOXIRI and FOLFOX/CAPOX
was similar (47% vs 51%, HR 1.09, 90% CI 0.78 to 1.53; P = 0.7). In a pre-specified correlative
analysis of all ctDNA positive pts, recurrence risk increased with ctDNA burden, with 3-year RFS of
78%, 63%, 36% and 22% for tumor-derived mutant molecules/mL quartiles , 0.06, 0.06 – 0.17,
0.18 – 1.31, and . 1.31, respectively (P , 0.01). Treatment-related hospitalisation was similar for
escalated and SOC pts (OR 1.21, P = 0.58). Analysis of post-ACT ctDNA is underway.
Conclusions: In this first randomised study of ctDNA-informed management in stage III CC, we
confirm the prognostic significance of detectable ctDNA, with the novel finding of recurrence risk
increasing markedly with ctDNA burden. Treatment escalation, including to FOLFOXIRI, did not
improve RFS. Future studies in ctDNA positive pts should explore other escalation strategies.
Clinical trial information: ACTRN12617001566325. Research Sponsor: Marcus Foundation;
NHMRC; U.S. National Institutes of Health.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
236s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3504 Oral Abstract Session 3505 Oral Abstract Session

Tissue-free circulating tumor DNA assay and patient outcome in a phase III Perioperative systemic therapy for resectable colorectal peritoneal metas-
trial of FOLFOX-based adjuvant chemotherapy (Alliance N0147). First Author: tases: A multicenter randomized phase 3 trial (CAIRO6). First Author: Koen
Frank A. Sinicrope, Mayo Clinic Rochester, Rochester, MN Rovers, Catharina Cancer Institute, Eindhoven, Netherlands
Background: Among patients with resected node-positive colon cancer, nearly 30% will Background: In patients with resectable colorectal peritoneal metastases who qualify
relapse despite standard adjuvant chemotherapy. Analysis of molecular residual disease for cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (CRS-HIPEC),
(MRD) using circulating tumor DNA (ctDNA) may enable risk stratification for tumor there is no prospective data comparing the efficacy of perioperative systemic therapy
recurrence and inform adjuvant treatment decisions. Methods: Postsurgical ctDNA was with CRS-HIPEC alone. Methods: In this multicenter phase 3 superiority trial, patients
analyzed in patients with stage III colon carcinoma who participated in a phase 3 trial of with resectable colorectal peritoneal metastases without extraperitoneal metastases
adjuvant FOLFOX alone or combined with cetuximab (n = 3084) [NCCTG N0147]. We who did not receive systemic therapy within six months prior to enrollment were
utilized a tissue-free epigenomic assay for ctDNA detection (Guardant Reveal) with randomly assigned (1:1) to receive perioperative CAPOX, FOLFOX, or FOLFIRI with
sampling prior to start of adjuvant therapy. Among ctDNA positives, epigenomic tumor neoadjuvant addition of bevacizumab (perioperative systemic therapy group) or CRS-
fraction (TF) was estimated and ctDNA genotyping was done with Guardant360 (panel of HIPEC alone (surgery alone group). The primary outcome was overall survival. Key
739 genes). Median follow-up was 6.1 years (yr). Study endpoints included time-to- secondary outcomes were progression-free survival and 90-day major postoperative
recurrence (TTR), disease-free survival (DFS) and overall survival (OS) analyzed by morbidity and mortality. The trial needed 179 patients in each arm to detect a superior 3-
Kaplan–Meier method. Multivariable Cox proportional hazards models were used to year overall survival of 65% in the perioperative systemic therapy group versus 50% in
assess prognostic utility of ctDNA status adjusting for confounders. Interaction between the surgery alone group (corresponding hazard ratio [HR] for death 0.62) with 80%
ctDNA and clinicopathological features were assessed. Results: Among 2260 patients power, 5% drop-out, and a two-sided log rank test of p,0.05. The primary overall survival
with evaluable ctDNA data, 461 (20.4%) were ctDNA [Link] were significantly analysis was done after 171 events (88% power). Results: Of 358 randomized patients,
associated with higher T, N stage, BRAFV600E, high grade, obstruction/perforation, and 351 were eligible for primary analysis: 173 in the perioperative systemic therapy group
worse performance status. Positive vs negative ctDNA was significantly associated with and 178 in the surgery alone group. At a median follow-up of 41 months, median and 3-
shorter TTR (hazard ratio [HR] 4.33, 95% confidence interval [CI] 3.65-5.13, P , 0.0001), year overall survival were 44 months and 54% in the perioperative systemic therapy
poorer DFS (HR 3.74, CI 3.18-4.39, P , 0.0001] and OS (HR 3.17, CI 2.63-3.83, P , group and 39 months and 53% in the surgery alone group, respectively (HR for death
0.0001), adjusting for covariates and tissue MMR, KRAS and BRAFV600E. ctDNA positive 0.85, 95% CI 0.62-1.15, p=0.28). Median and 3-year progression-free survival were
vs negative patients had 3y DFS of 36.4% (95%CI 32.2-41.2%) vs 82.5% (95% CI 80.0- 13.5 months and 20% in the perioperative systemic therapy group and 7.0 months and
84.4%), respectively. Adverse prognosis was consistent across subgroups (all P , 0.05), 5% in the surgery alone group, respectively (HR for progression or death 0.51, 95% CI
with stronger detrimental effects for positive ctDNA in N1 (vs. N2), T1/2 (vs T3 or 4), and 0.41-0.65). In the per-protocol population of 292 patients who underwent macroscopic
mismatch repair deficient tumors [interaction P = 0.0002 to 0.041). Among patients with complete CRS-HIPEC, median and 3-year overall survival were 54 months and 64% in the
positive ctDNA, TF in those who recurred/died within 3 yr was double of those who perioperative systemic therapy group (138 patients) and 45 months and 59% in the
remained recurrence-free (P = 0.0001). High vs. low ctDNA TF ( . vs # median) further surgery alone group (154 patients), respectively (HR for death 0.73, 95% CI 0.51-1.05).
stratified TTR (HR 1.48, CI 1.17-1.88, P = 0.0011), DFS (HR 1.52, CI 1.21-1.92, P = 0.0004) Ninety-day major postoperative morbidity rates were 36% in the perioperative systemic
and OS (HR 1.58, CI 1.21-2.07, P = 0.0009), adjusting for confounders. ctDNA positive therapy group and 26% in the surgery alone group, with a 90-day postoperative mortality
cases, . Analyses of ctDNA detection by site of recurrence, ctDNA clearance, and of 1% in both groups. Conclusions: Among patients with resectable colorectal peri-
genomic variant detection are ongoing. Conclusions: In the largest study evaluating toneal metastases, perioperative systemic therapy did not result in superior overall
tissue-free epigenomic-based MRD detection, we demonstrate a robust prognostic survival as compared to CRS-HIPEC alone. Clinical trial information: NCT02758951.
utility of postsurgical ctDNA. Tumor fraction provided further patient stratification and Research Sponsor: Dutch Cancer Society; F. Hoffmann-La Roche.
analysis is ongoing to identify a subgroup based on TF that may be unlikely to clear
ctDNA despite adjuvant chemotherapy. Research Sponsor: National Cancer Institute;
U10CA180882.

3506 Oral Abstract Session 3507 Oral Abstract Session

Long-term safety and efficacy of sotorasib plus panitumumab and FOLFIRI Efficacy and safety of olomorasib, a second-generation KRAS G12C inhib-
for previously treated KRAS G12C-mutated metastatic colorectal cancer itor, plus cetuximab in KRAS G12C-mutant advanced colorectal cancer. First
(mCRC): CodeBreaK 101 (phase 1b). First Author: John H. Strickler, Duke Author: Antoine Hollebecque, Institut Gustave Roussy, Villejuif, France
University Medical Center, Durham, NC Background: Olomorasib, a potent and selective second-generation KRAS G12C inhibitor (G12Ci), has
Background: In the phase 3 CodeBreaK 300 trial (NCT05198934), the combination of demonstrated promising efficacy and a favorable safety profile in KRAS G12C-mutant cancers. Based
sotorasib (KRASG12C inhibitor) and panitumumab (monoclonal anti-EGFR antibody) on emerging nonclinical and clinical data, combining a KRAS G12Ci with cetuximab offers a compelling
opportunity to improve outcomes in patients (pts) with KRAS G12C-mutant colorectal cancer (CRC).
improved clinical outcomes in patients with chemorefractory KRAS G12C-mutated Here we report updated results from a phase 1/2 study (NCT04956640) on the safety, tolerability and
mCRC. CodeBreaK 101 is a phase 1b trial where FOLFIRI was added to sotorasib optimal dose of olomorasib + cetuximab in pts with KRAS G12C-mutant CRC. Methods: Pts with
and panitumumab in previously treated patients with KRAS G12C-mutated mCRC. For advanced KRAS G12C-mutant CRC (tissue or plasma) previously treated with $1 prior oxaliplatin- or
the first time, we report mature overall survival (OS) and progression-free survival (PFS), irinotecan-containing regimen were eligible and enrolled into dose escalation/expansion or optimi-
as well as updated safety and response data. Methods: Patients with KRAS G12C- zation at 2 doses of olomorasib (100 and 150 mg, orally BID). Dose escalation of olomorasib +
mutated mCRC who received $1 prior systemic treatment but were KRASG12C inhibitor- cetuximab followed a mTPI-2 method. Key objectives were safety and to determine the optimal dose of
naı̈ve, were enrolled into the expansion cohort of the CodeBreak 101 subprotocol H olomorasib + cetuximab. Antitumor activity per RECIST v1.1 was studied in pts with $1 post-baseline
response assessment or who discontinued before a first response assessment. Results: As of 13
(NCT04185883) phase 1b trial. As defined from dose exploration cohort, patients re-
November 2024, 93 pts received olomorasib + cetuximab in dose escalation/expansion (n=49) or
ceived the recommended phase 2 dose (RP2D) of sotorasib (960 mg orally daily) plus optimization (n=44). Median age was 58 yrs (range, 35-82) and median number of prior therapies was 3
panitumumab (6 mg/kg intravenous every 2 weeks [Q2W]) and standard dose FOLFIRI (range, 1-8). All grade TRAEs in $20% of pts were dermatitis acneiform (58%), diarrhea (38%), dry skin
(intravenous Q2W). The primary endpoint was safety and secondary endpoints included (31%), paronychia (28%), hypomagnesemia (26%), and rash (26%). The majority of TRAEs were grade
confirmed response, OS, and PFS, assessed by investigator per Response Evaluation 1-2, with grade $3 observed in 24% of pts. The most common TRAEs grade $3 were diarrhea,
Criteria in Solid Tumors (RECIST) v1.1. Results: By November 2024, 40 patients were hypokalemia, and rash, each occurring in 2 pts. TRAEs led to olomorasib dose reduction in 2% of pts,
enrolled (female: 47.5%; median age: 56.0 years; median [range] prior lines of systemic olomorasib dose hold in 22% of pts, and cetuximab dose hold in 16% of pts. Of the 61 pts who
therapy: 2 [1-6]). The most common treatment-related adverse events (TRAEs) were discontinued treatment, 57 were due to PD. Two pts discontinued cetuximab due to TRAEs and
continued on olomorasib. The AE profile was similar between doses. Median time on combination
dermatitis acneiform and dry skin (n = 27 [67.5%] each), decreased neutrophil count (n = treatment was 6.5 mo (range, 0.8-24.1) and 32 pts remained on treatment. See Table 1 for efficacy
20 [50.0%]), and stomatitis (n = 17 [42.5%]). Grade $3 TRAEs occurred in 20 (50.0%) data. Biomarker analysis will be reported. Conclusions: Olomorasib + cetuximab demonstrated
patients with no new safety signals. Discontinuation of sotorasib, panitumumab, or similar antitumor activity and favorable safety at both dose levels in pts with KRAS G12C-mutant CRC,
FOLFIRI (5-FU, irinotecan, or leucovorin/levoleucovorin) due to AEs was observed in 1 with the optimal dose of olomorasib + cetuximab determined as 100 mg BID. These results further
(2.5%), 1 (2.5%), and 16 (40.0%) patients, respectively. A total of 7 patients are still support combining second-generation KRAS G12Ci with other anticancer therapies to improve
continuing the study, of whom 5 are off-treatment and under follow-up. Updated ob- outcomes in previously treated pts with KRAS G12C-mutant CRC. Clinical trial information:
jective response rate (95% CI) was 57.5% (40.9, 73.0) and disease control rate (95% CI) NCT04956640. Research Sponsor: Eli Lilly and Company.
was 92.5% (79.6, 98.4). Median time to response was 1.6 months and duration of Olomorasib
response was 6.6 months. After a median follow-up of 29.2 months, the median (95% CI) (100 mg BID) + Olomorasib
Cetuximab (150 mg BID) + Total
PFS was 8.2 (7.0, 10.8) months and median OS was 17.9 (12.9, 25.1) months. Endpoint N=64 CetuximabN=29 N=93
Conclusions: Sotorasib plus panitumumab and FOLFIRI showed promising long-term
ORR, % (n/N) 44% (28/64) 38% (11/29) 42% (39/93)
safety and efficacy in pretreated KRAS G12C-mutated mCRC. AEs were consistent with BOR, n (%)
the safety profile of the drugs administered. The ongoing phase 3 study, CodeBreaK 301 PR 28 (44) 11 (38) 39 (42)
(NCT06252649), aims to evaluate this combination against standard of care in first-line SD 31 (48) 16 (55) 47 (51)
PD 5 (8) 2 (7) 7 (8)
patients with KRAS G12C-mutated mCRC. Clinical trial information: NCT05198934. mDOR, mo (95% CI) 8.3 (5.6-12.7) 6.2 (2.8-NE) 7.6 (6.0-12.2)
Research Sponsor: Amgen Inc. mPFS, mo (95% CI) 7.5 (6.7-9.7) 6.6 (4.2-7.6) 7.5 (6.6-8.8)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 237s

3508 Oral Abstract Session LBA3509 Clinical Science Symposium

The KRAS G12C inhibitor MK-1084 for KRAS G12C–mutated advanced Association between empirical dietary inflammatory pattern (EDIP) and
colorectal cancer (CRC): Results from KANDLELIT-001. First Author: survival in patients with stage III colon cancer: Findings from CALGB/
Iwona A. Lugowska, Maria Sklodowska-Curie National Research Institute and Oncology SWOG 80702 (Alliance). First Author: Sara K. Char, Department of Medical On-
Centre, Warsaw, Poland cology, Dana-Farber Cancer Institute, Boston, MA
Background: Preliminary data from the phase 1 KANDLELIT-001 trial (NCT05067283)
showed a manageable safety profile and preliminary antitumor activity for MK-1084, a next-
generation, selective KRAS G12C-GDP covalent inhibitor, in participants (pts) with previously
treated, KRAS G12C-mutant solid tumors, including non–small-cell lung cancer and CRC.
Here, we report data for MK-1084 monotherapy, MK-1084 + cetuximab, and MK-1084 +
cetuximab + mFOLFOX6 in pts with advanced KRAS G12C-mutant CRC.
Methods: KANDLELIT-001 enrolled pts with confirmed KRAS G12C mutation, RECIST-
measurable disease, and ECOG PS 0-1. Pts with any advanced solid tumor and $1 prior
systemic therapy received MK-1084 monotherapy PO QD or BID (total daily dose, 25-800 mg)
in arms 1 and 3. Pts with advanced CRC and 1-2 prior systemic therapies received MK-1084
QD (total daily dose, 25-200 mg) plus cetuximab 500 mg/m2 IV Q2W in arm 5. Pts with
advanced CRC and 0-1 prior systemic therapies received MK-1084 QD (total daily dose, 25-
100 mg) plus cetuximab 500 mg/m2 Q2W and mFOLFOX6 in arm 6. The primary endpoints
were dose-limiting toxicities (DLTs), AEs, and AEs leading to discontinuation. Secondary The full, final text of this abstract will be available at
endpoints included ORR per RECIST v1.1 by investigator review. ORR was assessed in all pts [Link] on the day of presentation and in the
who received their first MK-1084 dose $5 wk before the data cutoff date of August 12, 2024,
for arms 1 and 3 and November 6, 2024, for arms 5 and 6. Results: In arms 1+3, 99 pts, online supplement to the June 10, 2025, issue of the Journal
including 53 (54%) with CRC, received MK-1084 alone. In arm 5, 34 pts, including 23 (68%) of Clinical Oncology.
who had $2 prior lines of therapy, received MK-1084 + cetuximab. In arm 6, 20 pts, including
10 (50%) who had no prior therapy, received MK-1084 + cetuximab + mFOLFOX6. Median
(range) study follow-up was 14.8 mo (0.2-30.8) in arms 1+3, 5.3 mo (2.6-11.5) in arm 5, and
1.9 mo (0.1-5.4) in arm 6. One pt in arm 6 experienced a DLT (grade 3 febrile neutropenia);
there were no DLTs in arms 1, 3, or 5. Treatment-related AEs occurred in 62% of pts in arms
1+3, 97% of pts in arm 5, and 90% of pts in arm 6, were grade $3 in 9%, 18%, and 25%,
respectively, and led to discontinuation of any drug in 1%, 3%, and 15%, respectively. There
were no treatment-related deaths. The two most common treatment-related AEs in each arm
were increased AST (17%) and nausea (17%) in arms 1+3, dermatitis acneiform (47%) and
rash (24%) in arm 5, and nausea (55%) and rash (50%) in arm 6. ORR (95% CI) was 36% (23-50)
in pts with CRC in arms 1+3 (n = 53), 50% (32-68) in arm 5 (n = 34), and 14% (2-43) in arm 6 (n =
14); all responses were partial responses. Conclusions: Preliminary data suggest that MK-
1084 monotherapy, MK-1084 + cetuximab, and MK-1084 + cetuximab + mFOLFOX6 have
manageable safety profiles and show evidence of antitumor activity in pts with advanced,
KRAS G12C-mutated CRC. Pts continue to be followed, and enrollment continues. Clinical trial
information: NCT05067283. Research Sponsor: Merck Sharp & Dohme LLC, a subsidiary of
Merck & Co., Inc., Rahway, NJ, USA.

LBA3510 Clinical Science Symposium LBA3511 Clinical Science Symposium

A randomized phase III trial of the impact of a structured exercise program Aspirin as secondary prevention for colorectal cancer liver metastases
on disease-free survival (DFS) in stage 3 or high-risk stage 2 colon cancer: (ASAC): A multicenter, randomized, double-blind, placebo-controlled, phase
Canadian Cancer Trials Group (CCTG) CO.21 (CHALLENGE). First Author: 3 trial. First Author: Sheraz Yaqub, Department of Hepato-Pancreato-Biliary Surgery,
Christopher M. Booth, Queen’s University, Kingston, ON, Canada Oslo University Hospital and University of Oslo, Institute of Clinical Medicine, Oslo,
Norway

The full, final text of this abstract will be available at The full, final text of this abstract will be available at
[Link] on the day of presentation and in the [Link] on the day of presentation and in the
online supplement to the June 10, 2025, issue of the Journal online supplement to the June 10, 2025, issue of the Journal
of Clinical Oncology. of Clinical Oncology.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
238s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3512 Rapid Oral Abstract Session 3513 Rapid Oral Abstract Session
Upfront modified FOLFOXIRI plus panitumumab (pan) versus FOLFOX/pan FIRE-4 (AIO KRK-0114): Randomized study evaluating the efficacy of
for unresectable RAS and BRAF wild-type (wt) metastatic colorectal cancer cetuximab re-challenge in patients with metastatic RAS wild-type colorectal
(mCRC) patients: Overall survival (OS) results from the phase III TRIPLETE cancer responding to first-line treatment with FOLFIRI plus cetuximab. First
study by GONO. First Author: Veronica Conca, Department of Translational Research Author: Lena Weiss, Department of Medicine III, University Hospital Munich LMU,
and New Technologies in Medicine and Surgery, University of Pisa & Unit of Medical Munich, Germany
Oncology 2, Azienda Ospedaliera Universitaria Pisana, Pisa, Italy Background: Several smaller studies performed in later lines of treatment have
Background: TRIPLETE (NCT03231722) is a phase III trial where unresectable RAS/BRAF suggested a potential benefit from anti-EGFR re-challenge on survival of RAS wild-type
wt mCRC patients (pts) were randomized 1:1 to first-line FOLFOX/pan (Arm A) or (RAS WT) metastatic colorectal cancer (mCRC). FIRE-4 is a randomized phase-III study
modified FOLFOXIRI/pan (Arm B). The study failed to demonstrate an improved overall that prospectively evaluates re-challenge with chemotherapy plus cetuximab as
response rate, primary endpoint of the study, in Arm B, and did not show any benefit from compared to physician’s choice. Methods: The FIRE-4 study was performed with two
the intensification of the chemotherapy also in terms of progression-free survival (PFS), steps of randomisation. Within the first randomisation, pts were either attributed to
early tumor shrinkage, depth of response, R0 resection rate at the price of increased induction therapy with FOLFIRI plus cetuximab continued until disease progression (PD)
gastrointestinal toxicity. Here we report OS results. Methods: Eligible pts were stratified or intolerable toxicity (arm A) or to a switch maintenance using 5-FU plus bevacizumab
according to ECOG PS (0-1 vs 2), primary tumor location (right vs left), and liver-only (arm B). After first PD, an anti-EGFR-free “window therapy” was recommended. After
metastases (yes vs no). OS was assessed from randomization to death from any cause. diagnosis of second PD, RAS WT pts (again selected by liquid- or tumor-biopsy), who had
Survival curves were calculated using the Kaplan–Meier method and compared with the responded to cetuximab-based induction therapy within FIRE-4 (entry 1) or outside of
log-rank test stratified by the same factors as per randomization. Hazard ratios (HR) the study (entry 2), could then proceed to 2nd randomization attributing pts either to re-
with 95% confidence interval (CI) were estimated using Cox regression models. challenge with cetuximab or to physician’s choice. Overall survival (OS) after 2nd
Results: 435 pts (A/B: 217/218) were enrolled. Main pts’ characteristics were median randomization was evaluated as primary endpoint. Results: From August 2015 to
age 59/59 years, ECOG PS 0 80%/84%, synchronous metastases 88%/86%, liver-only February 2021, 672 pts were randomized and 657 pts were assigned to treatment in 120
disease 38%/39%, left-sided primary tumour 88%/88%; deficient MMR tumours 1%/3%. German and 10 Austrian centers. Within the 2nd randomization, 87 pts (entry 1: N = 62;
At a median follow up of 60.2 months (mos), 292 (67%, arm A/B: 71%/63%) OS events entry 2: N = 25) were attributed either to physician’s choice (A2: N = 42) or (FOLF)IRI plus
were collected. Significantly longer OS was observed in Arm B with a median OS of 41.1 cetuximab (Arm B2: n = 45). Baseline characteristics were comparable between groups
vs 33.3 mos in Arm A (HR: 0.79; 95%CI: 0.63-0.99; p = 0.049). No molecular or clinical without significant differences regarding parameters such as age, sex, ECOG perfor-
groups of interest emerged from the subgroup analyses. While no significant difference mance status, or primary tumor sidedness. All pts were RAS WT at the time of ran-
in PFS was confirmed (median PFS Arm A/B: 12.4/12.7 mos, p = 0.606), longer post domization. No statistically significant difference between arm A2 and B2 was observed
progression survival (PPS) was reported in Arm B (HR: 0.73; 95%CI: 0.57-0.93; p = regarding OS (15.1 months vs. 17.6 months; HR 0.84; P = 0.48) or PFS (4.6 months vs.
0.012). The proportion of pts receiving subsequent lines of therapy was similar between 5.8 months; HR 0.91; P = 0.64). ORR was greater in the experimental arm (11.9% vs.
arms (2nd-line Arm A/B: 73%/71%, 3rd-line: 52%/50%, 4th-line: 33%/33%), and no dif- 28.9%; OR 0.33; P = 0.07), while disease control rate was nearly identical (59.5% vs.
ferences were evident in the exposure to anti-EGFRs (Arm A/B 35%/38%) and oxaliplatin 60.0%; OR 0.98; P . 0.99). Conclusions: FIRE-4 did not meet its primary endpoint.
(26%/33%) after PD, while higher percentages of pts in ARM A received anti-angiogenics While the control arm using physician’s choice exceeded expectations, re-challenge with
(59%/45%) and irinotecan (66%/56%). Similar percentages of pts received locoregional anti-EGFR therapy in RAS WT pts obtained comparable results in terms of OS. Clinical
treatments with radical intent after PD (Arm A/B 16%/15%). Conclusions: Upfront trial information: NCT02934529. Research Sponsor: MERCK Serono.
modified FOLFOXIRI/pan provides a statistically significant and clinically meaningful
survival advantage compared to standard FOLFOX/pan in pts with RAS/BRAF wt mCRC,
with a 7.8 mos difference in median values, though in the absence of any significant
difference in treatment activity and PFS. Clinical trial information: NCT03231722.
Research Sponsor: GONO Foundation; Amgen.

3514 Rapid Oral Abstract Session LBA3515 Rapid Oral Abstract Session
Longitudinal ctDNA monitoring and prediction of anti-EGFR rechallenge Panitumumab retreatment followed by regorafenib versus the reverse se-
outcomes in RAS/BRAF wild-type metastatic colorectal cancer (mCRC): quence in chemorefractory metastatic colorectal cancer patients with RAS
The REMARRY & PURSUIT trials. First Author: Yoshinori Kagawa, Department of and BRAF wild-type circulating tumor DNA (ctDNA): Results of the phase II
Gastroenterological Surgery, Osaka International Cancer Institute, Osaka, Japan randomized PARERE trial by GONO. First Author: Chiara Cremolini, Unit of Medical
Background: Anti-EGFR monoclonal antibody (mAb) rechallenge involves re-administering Oncology 2, Azienda Ospedaliera Universitaria Pisana and Department of Translational
EGFR blockade after a treatment-free interval to exploit clonal evolution. Recent evidence Research and New Technologies in Medicine and Surgery, University of Pisa, Pisa, Italy
suggests that circulating tumor DNA (ctDNA)–based selection may improve outcomes;
however, the predictive value of longitudinal ctDNA monitoring remains uncertain.
Methods: The REMARRY study evaluated plasma RAS (pRAS) dynamics in patients with RAS/
BRAF V600E wild-type metastatic colorectal cancer (mCRC), ECOG PS 0–1, who had pre-
viously responded to anti-EGFR therapy and experienced progression within two months of
the last dose. pRAS status was assessed at progression on prior anti-EGFR, before
rechallenge, at cycle 3, and at discontinuation using BEAMing digital PCR. Patients meeting
additional criteria—namely, pRAS-negative, refractory or intolerance to standard chemo-
therapies, and an anti-EGFR–free interval of at least 4 months—were enrolled in the phase II
PURSUIT trial, which administered panitumumab (6 mg/kg) plus irinotecan (150 mg/m²)
biweekly. The primary endpoint was confirmed objective response rate (ORR) per RECIST
v1.1. In parallel, participants underwent next-generation sequencing (NGS) of ctDNA in the
The full, final text of this abstract will be available at
GOZILA trial at corresponding time points to identify resistance alterations in genes including [Link] on the day of presentation and in the
RAS, BRAF, EGFR-ECD, MAP2K, ERBB2, and MET. Results: Between May 2019 and May 2021, online supplement to the June 10, 2025, issue of the Journal
183 patients were enrolled in REMARRY, and 50 pRAS-negative patients before rechallenge
were included in PURSUIT. The confirmed ORR was 14.0% (90% CI, 7.0–23.0%), with a median
of Clinical Oncology.
progression-free survival (PFS) of 3.6 months and a median overall survival (OS) of
12.0 months. Although all patients were pRAS-negative at baseline, 10.0% converted to pRAS-
positive by cycle 3 and 36.0% by discontinuation. Patients who were pRAS-positive im-
mediately after prior anti-EGFR had higher conversion rates at cycle 3 (42.9% vs. 6.3%, p =
0.010) and at discontinuation (85.7% vs. 32.3%, p , 0.001) compared with those initially
negative. The ORR was 23.8% in patients remaining pRAS-negative versus 0% in those
converting to pRAS-positive (p = 0.254). Moreover, NGS-detected resistance alterations after
prior anti-EGFR were associated with no responses (0/13) compared to a 27.8% response rate
(5/23) in patients without such alterations (p = 0.038), and correlated with shorter PFS (HR
3.297, p = 0.002) and OS (HR 4.569, p , 0.001). In 86% (7/8) of patients who were pRAS-
positive post–anti-EGFR, the identical pRAS codons re-emerged during rechallenge, con-
sistent with NGS findings. Conclusions: ctDNA status immediately after progression on prior
anti-EGFR therapy predicts subsequent response, thereby supporting clinical decision-
making and personalized therapy. Trial Registration: PURSUIT (jRCTs031190096), REMARRY
(UMIN000036424), GOZILA (UMIN000029315). Clinical trial information: jRCTs031190096.
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 239s

LBA3516 Rapid Oral Abstract Session 3517 Rapid Oral Abstract Session
JMT101 in combination with irinotecan and SG001 versus regorafenib in Phase 2 dose expansion study of DSP107, a first-in-class bi-specific 4-1BB
patients with metastatic colorectal adenocarcinoma (mCRC): Results of a T-cell engager, with and without atezolizumab in metastatic MSS colorectal
randomized, controlled, open-label, phase II study. First Author: Jianmin Xu, cancer patients. First Author: Anwaar Saeed, University of Pittsburgh Medical Center
Department of Colorectal Surgery, Zhongshan Hospital Fudan University, Shanghai, (UPMC) and UPMC Hillman Cancer Center, Pittsburgh, PA
China Background: DSP107 is a bi-specific fusion protein composed of sequences from the
extracellular domain of SIRPa and 4-1BBL. The SIRPa arm selectively targets CD47
overexpressed on tumor cells, while simultaneously anchoring trimeric 4-1BBL to the tumor,
to engage and co-stimulate 4-1BB on activated immune cells in the tumor microenvi-
ronment. This results in tumor-localized, conditional activation of innate and adaptive
immune responses. Phase 1 data demonstrated an excellent safety profile with no RBC
binding and no hematological, hepatic or other dose limiting toxicities (DLTs). Here we
describe safety and efficacy data from a Phase 2 microsatellite stable (MSS) colorectal
(CRC) expansion cohort in which patients were treated with DSP107 alone or with ate-
zolizumab (NCT04440735). Methods: Metastatic/unresectable MSS CRC patients who
progressed following 2 lines of therapy including standard chemotherapy 6 targeted
antibodies (n = 50), were randomized to receive weekly IV DSP107 infusions (10 mg/kg/
dose) alone or with atezolizumab (1200 mg) Q3W during 3-week treatment cycles. The
The full, final text of this abstract will be available at majority (76%) had liver metastases. Study objectives were safety, tolerability and pre-
[Link] on the day of presentation and in the liminary efficacy. Restaging imaging was performed every 2 months and evaluated by
RECIST v1.1 criteria. Results: DSP107 monotherapy and with atezolizumab was well
online supplement to the June 10, 2025, issue of the Journal tolerated with no DLTs. The most frequent TRAEs were infusion-related reactions (IRR; 38%
of Clinical Oncology. Grade 1 or 2, 4% Grade 3), fatigue (12% Grade 1 or 2, 4% Grade 3), Grade 1 or 2 nausea (14%)
and Grade 1 or 2 anemia (10%). IRRs were managed during subsequent infusions by re-
ducing the infusion rate and administering IV fluids. The median OS from the efficacy-
evaluable patients who received DSP107 monotherapy (n = 19) and combination therapy
with atezolizumab (n = 21) has not been reached but currently (Dec 2024 cutoff) stands at
7.6 and 14.6 months, respectively. Disease control was demonstrated in 26% (monotherapy)
and 62% (combination) of evaluable patients including a patient who achieved complete
response ( . 2.5 years) and a patient with a deep (86% target lesion reduction) and durable
( . 16 months) confirmed partial response and disappearance of pulmonary and hepatic
metastases. Immunofluorescence analysis of baseline tumor biopsies (n = 16) demon-
strated moderate to high levels of CD47 expression in 15/16 biopsies ( . 120 H-Score), and
very high levels of CD47 expression ( . 170 H-Score) in all 7 samples collected from liver
metastases. Conclusions: These data suggest that the combination of DSP107 with PD(L)1
blockade has anti-tumor activity and provides clinical benefit in third line metastatic MSS
CRC including in patients with liver metastases. Updated survival data will be presented at
the conference. A Phase 2 randomized controlled study is currently in planning to confirm
this preliminary efficacy signal. Clinical trial information: NCT04440735. Research Sponsor:
KAHR Medical LTD.

3518 Rapid Oral Abstract Session 3519 Rapid Oral Abstract Session
Circulating tumor DNA as an early response indicator in anal squamous cell Short-course radiotherapy followed by sintilimab and CAPOX as total neo-
carcinoma treated with chemoradiation. First Author: Aron Bercz, Memorial Sloan adjuvant treatment in locally advanced rectal cancer: A prospective, ran-
Kettering Cancer Center, New York, NY domized controlled trial (SPRING-01). First Author: Feng Tian, Shandong
Background: Definitive chemoradiation (CRT) is a highly effective, organ-preserving treatment for Provincial Hospital Affiliated to Shandong First Medical University, Jinan, NA, China
localized anal squamous cell carcinoma (ASCC). However, a subset of patients experience locoregional Background: Neoadjuvant short-course radiotherapy (SCRT) combined with chemotherapy as
failure, leading to unfavorable oncologic outcomes despite salvage surgery. Circulating tumor DNA total neoadjuvant therapy (TNT) increases the pathological complete response (pCR) rate for
(ctDNA) has emerged as a promising tool for monitoring treatment efficacy and predicting prognosis; locally advanced rectal cancer (LARC). The potential synergistic effects of combining ra-
however, there is a paucity of studies evaluating the baseline detectability and kinetics of tumor-
diotherapy and immunotherapy might benefit patients with LARC. This study aimed to
informed ctDNA assays in ASCC. Methods: Patients with ASCC (N=88) undergoing definitive CRT
provided prospective consent for longitudinal ctDNA monitoring using a personalized, tumor-informed
compare the efficacy and safety of SCRT followed by 6 cycles of CAPOX chemotherapy with or
ctDNA assay (Signatera, Natera, Inc.). ctDNA testing was assessed at three key time points: pre- without immunotherapy as TNT in LARC patients. Methods: In this randomized controlled trial,
treatment (any time before CRT to within 5 days after initiation), mid-treatment (from .5 days after patients with T3-4, N+, EMVI(+), MRF(+) or lateral lymph node(+) rectal adenocarcinoma were
initiation to ,7 days before completion), and post-treatment (.7 days before completion to 42 days randomly assigned to receive SCRT followed by 6 cycles of CAPOX chemotherapy with or
after CRT). Surveillance testing continued every three months. Changes in ctDNA levels were analyzed without sintilimab. Total mesorectal excision (TME) was performed 2-3 weeks after the
in relation to clinical outcomes. Locoregional failure (LRF) was assessed using competing risk re- completion of TNT. The primary study endpoint was the pCR rate. Results: In this randomized
gression, stratified by ctDNA status (positive vs. negative). ctDNA results were also correlated with controlled trial, patients with T3-4, N+, EMVI(+), MRF(+) or lateral lymph node(+) rectal ad-
progression-free survival (PFS). Results: Pre-treatment ctDNA was detected in 79% of patients, with enocarcinoma were randomly assigned to receive SCRT followed by 6 cycles of CAPOX
92% achieving ctDNA-negativity at post-treatment (Table 1). Over a median follow-up of 18 months (IQR chemotherapy with or without sintilimab. Total mesorectal excision (TME) was performed 2-
11–26), 7 patients experienced LRF, and 5 experienced distant failure. The cumulative LRF incidence 3 weeks after the completion of TNT. The primary study endpoint was the pCR rate.
was 0% among patients with ctDNA negativity by mid-treatment. Conversely, 26% of patients with Conclusions: In LARC patients, SCRT combined with sintilimab and CAPOX as a TNT sig-
ctDNA positivity at mid-treatment and 61% of patients with ctDNA positivity at post-treatment ex-
nificantly increases the pCR rate while maintaining manageable safety in patients with LARC.
perienced LRF, respectively. Estimated one-year PFS was 100% for patients who achieved ctDNA
negativity by mid-treatment. In contrast, patients who remained ctDNA-positive at mid-treatment and
SCRT followed by sintilimab and CAPOX can be recommended as a superior neoadjuvant
post-treatment had estimated one-year PFS rates of 81% and 44%, respectively, from the date of the treatment option for these patients. Clinical trial information: ChiCTR2100052288. Research
corresponding ctDNA test. Among patients who achieved ctDNA negativity but subsequently developed Sponsor: National Natural Science Foundation of China; Special Foundation for Taishan
molecular recurrence during the surveillance period (N=7), all developed disease recurrence. Molecular Scholars Program of Shandong Province; Key Research and Development Program of
recurrence predated clinical or radiographic evidence of recurrence in all instances. Conclusions: This Shandong Province; China Postdoctoral Science Foundation.
tumor-informed ctDNA assay demonstrates high baseline detectability and rapid clearance during CRT, The efficacy of SIN+CAPOX and surgical and pathological results.
with molecular clearance correlating with favorable outcomes in ASCC. Notably, ctDNA-based detection
Intention-to-treat (ITT) population SIN+CAPOX (N = 49) CAPOX (N = 49)
of molecular recurrence consistently precedes conventional clinical and radiographic indicators of
disease recurrence. Further validation in large, prospective cohorts is warranted. Research Sponsor: Pathological complete response (ypT0N0, ITT 29（59.2） 16（32.7）
National Institutes of Health/National Cancer Institute (NIH/NCI); R37 CA248289; National Institutes of population) — no. (%)
Health/National Cancer Institute (NIH/NCI); K08 CA255574; National Institutes of Health/National (% [95% CI]) (45.4, 72.9) (19.5, 45.8)
Cancer Institute (NIH/NCI); 5T32 CA 9501-34; National Institutes of Health/National Cancer Institute Complete response 30（61.2） 16（32.7）
Surgical population SIN+CAPOX (N = 45) CAPOX (N = 44)
(NIH/NCI) Memorial Sloan Kettering Cancer Center (MSK) Support Grant; P30 CA008748. Tumor regression grading (AJCC
8th edition) — no. (%)
0 29（64.4） 16（36.4）
Pre-Treatment Mid-Treatment Post-Treatment
1 7（15.6） 7（15.9）
Positive 61 (79%) 29 (47%) 6 (8%) 2 5（11.1） 13（29.5）
Negative 16 (21%) 33 (53%) 67 (92%) 3 4（8.9） 8（18.2）
Unknown 11 26 15
Abbreviations: IQR, interquartile range; N, regional nodal category; T, primary tumor category; yp, pathologic.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
240s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3520 Rapid Oral Abstract Session 3521 Poster Session

Proposed changes to the pathologic staging for colon cancer (CC): AJCC Experience of patients with HIV and squamous cell carcinoma of the anal
Colon Cancer Expert Panel (AJCCCCEP). First Author: Qian Shi, Department of canal (SCAC) treated with retifanlimab. First Author: Jean-Philippe Spano,
Quantitative Science Research, Mayo Clinic Rochester, Rochester, MN Department of Medical Oncology, Groupe Hospitalier Pitie-Salpetriere, Paris, France
Background: Recent analyses highlight nonhierarchical outcomes using the 8th Edition AJCC staging system for Background: People living with HIV (PLHIV) usually have more advanced cancer at di-
CC. For instance, the 5-year survival rate for stages I and IIIa patients (pts) closely align. Additionally, tumor agnosis and a higher cancer-related mortality, posing a significant burden on health care.
deposits (TDs) have been established as significant prognostic indicators. The AJCCCCEP commissioned this
study to develop an updated pathological staging system for CC focused specifically on pts without distant However, clinical studies often exclude cancer patients with HIV, thereby limiting access to
metastasis (M0), while retaining the existing stage IV classification. Methods: Individual patient data (IPD) from therapies for this patient population. PLHIV have a 25- to 35-fold higher chance of being
pts diagnosed with cc (2010- 2017) in the NCDB were divided into training (70%) and internal validation (30%) diagnosed with SCAC than individuals who are HIV negative. We therefore evaluated safety
datasets. External validation used IPD from clinical trials. The primary endpoint was overall survival (OS). Risk and efficacy of retifanlimab in PLHIV with SCAC. Methods: The study designs for
classification development for M0 pts incorporated ungrouped data on pathologic T categories, the number of POD1UM-202 (NCT03597295) and POD1UM-303/InterAACT-2 (NCT04472429) have pre-
involved regional lymph nodes (LN+), and TD counts. Recursive partitioning and regression tree analyses were
applied to construct hierarchical staging levels. Pre-specified criteria required survival probabilities to be viously been described. Both trials permitted PLHIV to enroll if CD4+ count was $200/mL
consecutive and show clear separations using Kaplan-Meier (KM) estimates with pairwise log-rank test P of , with an undetectable viral load per standard of care assay, and who did not experience any
0.005 for the training and , 0.05 for validation analyses. Results: Data from 281,997 pts (median age 67 years, HIV-related opportunistic infection for $4 weeks prior to study enrollment. Patients
50% male, 81% white, 55% T3, 19% T4, 44% N+, 26% M+, and 11% with $1 TD) were analyzed, with a median continued to receive antiretroviral therapy (ART/HAART) without interruption or dose
follow-up of 7.3 years. The updated staging system (Table) met pre-specified criteria, with all observed pairwise
reduction. HIV viral load and CD4+ cell count was assessed every 8 weeks during the
P , 0.0001 in the development and internal validation sets. KM OS curves displayed a hierarchical separation
across all sub-levels after the 1st year of diagnosis. Consistent results were seen in pts treated with adjuvant studies and could be reduced to every 6 months during safety and disease follow-up.
chemotherapy in 4 trials (all pair-wise P , 0.0001). Conclusions: The proposed pathological staging system for Results: Patient and disease characteristics were similar among PLHIV and the overall
M0 pts fulfills pre-specified criteria for hierarchical risk stratification, validated both internally and externally, study populations. Among the 20 patients with HIV enrolled in these SCAC trials, median
and provides an evidence-based update. Pending review process, the AJCCCCEP will recommend that these age was 58 years, 70% (n = 14) were male, and 80% (n = 16) were White. Forty-five percent
changes be made to the Version 9 staging protocol for colon cancer to improve prognostication for CC pts.
(n = 9) of patients received retifanlimab and 30% (n = 6) received retifanlimab with
Research Sponsor: None.
platinum-based chemotherapy, whereas the remaining 5 patients were assigned to placebo
Stage T, # of LN+, # of TD M % of pts 1y OS (CI), % 3y OS (CI), % 5y OS (CI), % plus chemotherapy. During these studies, no patient experienced a sustained drop in CD4+
I T1, 0, 0 0 5 96 (95-97) 91 (90-92) 84 (83-86) T-cell counts or increase in HIV viral load of clinical significance. No treatment-emergent
IIa T2, 0, 0 0 10 95 (94-95) 88 (88-89) 80 (79-81)
IIb T1, 0, 1+ 0 27 93 (92-93) 84 (84-85) 75 (75-76) opportunistic infections were recorded. Immune-related adverse events (irAEs) and
T1, 1+, 0 grade $3 irAEs were consistent with the non-HIV population. Objective response rates
T2, 0, 1+
T2, 1-4, 0 were 22% (2/9) with retifanlimab in second-line and 67% (4/6) with retifanlimab and
IIIa
T3, 0, 0
T1, 1+, 1+ 0 14 92 (91-92) 80 (80-81) 71 (70-72)
chemotherapy in first-line (previously untreated). Patient-reported outcomes showed no
T2, 1-4, 1+ negative impact and based on Quality-of-Life Questionnaire for Anal Cancer, good scores
T2, 5+, 0
T3, 0, 1+
for bowel function, sexual, and symptom domains were maintained. Retifanlimab phar-
T3, 1-4, 0 macokinetics was independent of HIV status and not impacted by the HAART required for
IIIb T2, 5+, 1+
T3, 1-4, 1+
0 13 86 (86-87) 69 (68-70) 58 (57-59)
ongoing HIV management. Conclusions: Among PLHIV and advanced SCAC who received
T3, 5+, 0 treatment, retifanlimab showed significant clinical activity with efficacy qualitatively
T4a, 0-4, 0
T4b, 0-2, 0 similar to patients without HIV and no excess toxicity or reduced HIV control. The analysis
IIIc T3, 5+, 1+ 0 5 78 (77-80) 53 (51-54) 40 (38-41) indicates that retifanlimab is generally safe for PLHIV and SCAC and also supports in-
T4a, 0-4, 1+
T4a, 5+, any clusion of HIV-positive patients in other immunotherapy trials. The favorable outcomes in
T4b, 0-2, 1+ PLHIV are encouraging because infection with HIV is among the most important risk
T4b, 3+, any
IVa Any 1a 19 59 (58-60) 28 (28-29) 17 (16-18) factors for SCAC. Clinical trial information: NCT03597295 and NCT04472429. Research
IVb Any 1b 7 43 (42-44) 14 (13-14) 6 (6-7)
Sponsor: Incyte Corporation.
CI: 95% confidence internal; Peritoneum involvement data were not available before 2018 in NCDB. Thus, IVa/b were based on 7th Edition.

3522 Poster Session 3523 Poster Session

Investigating the immunogenomic profile of anal HPV driven disease for Immunotherapy combined with hypofractionated radiotherapy and chemo-
novel therapeutic discovery. First Author: Micol Lupi, Imperial College London, therapy for locally recurrent rectal cancer (TORCH-R): A prospective, single-
London, United Kingdom arm, two-cohort, phase II trial. First Author: Ruiyan Wu, Department of Radiation
Background: Anal squamous cell carcinoma (ASCC) is driven by Human Papilloma Virus and Oncology, Fudan University Shanghai Cancer Center, Shanghai, China
arises from high-grade squamous intraepithelial lesions (HSIL) which, when treated, reduces Background: To assess whether the integration of PD-1 inhibitor with hypofractionated
the risk of cancer progression. However, current treatment options for anal HSIL are limited radiotherapy and chemotherapy therapy can lead to an improvement in objective re-
with considerable morbidity. Although a proportion of patients with localised ASCC can be sponses in patients with proficient mismatch repair or microsatellite stable (pMMR/
successfully treated with pelvic chemoradiotherapy, there are significant associated side MSS) locally recurrence rectal cancer (LRRC). Methods: We did a prospective, single-
effects. Given the stepwise evolution of disease, there is an opportunity for the identification arm, two-cohort, phase 2 trial in LRRC patients without or with oligometastases. Eligible
of novel therapeutic approaches for cancer prevention. Methods: This study investigates the patients with previously untreated (cohort A) or progressive disease after first line
mutational and immune landscape of matched fresh frozen anal cancer and HSIL samples as therapy (cohort B), were assigned received 25-40 Gy/5 Fx irradiation or 15–30 Gy/5 Fx
well as blood samples from 8 patients using multi-omic analysis (WES, RNAseq and FUME-
reirradiation for pelvic recurrence, followdd by 18 weeks of chemotherapy, toripalimab,
TCRseq). We compared candidate driver mutations and pathways (including known immune
and stereotactic ablative radiotherapy (SABR) for all metastatic lesions between
escape mechanisms), copy number alterations (CNAs), differential gene expression, predicted
neoantigen profiles (using NeoPredPipe) and TCR repertoire composition and diversity. In
chemoimmunotherapy cycles. The primary endpoint was confirmed local recurrence
addition, digital cell classification of H&E stained sections was used to characterise the objective response rate (ORR). The study is registered with [Link],
distribution of 4 immune cell types. Results: There was considerable copy number profile NCT05628038. Results: Between Oct 15, 2022, and Aug 31, 2024, We enrolled 67
overlap between anal HSIL and cancer with 79.2% CNA concordance. Samples clustered by patients: 41 in cohort A and 26 in cohort B. Median follow-up duration was 10.5 months
patient rather than pathology on gene expression and few differentially expressed genes were (IQR 7.3-15.5 months). The local recurrence ORR was achieved at 82.9% (34 of 41
identified. There was no difference in predicted neoantigen burden (p= 0.11) nor the pro- patients) in cohort A and 65.4% (17 of 26 patients) in cohort B. Six patients (14.7%)
portion of unique or common neoantigens (p = 0.64), illustrating shared immunogenicity underwent radical resections (R0) in cohort A and two patients (7.7%) in cohort B. The
between anal cancers with corresponding pre-cancerous lesions. However, we observed a CR rate was 34.1% (12 cCR patients + 2 pCR patients) in cohort A and 11.5% (2 cCR
shift in TCR repertoire composition between HSIL and cancer in all patients, with HSIL regions patients + 1 pCR patients) in cohort B. The most frequent grade 3-4 toxicities were
containing larger clusters of related TCR clonotypes (p = 0.02). Driver mutations in PIK3CA, neutropenia (10.8% in cohort A and 25.0% in cohort B) and diarrhea (16.2% in cohort A
KMT2C, PBRM1, KLF5, STK11 and CUL1 were shared between matched samples. Enriched GO and 20.8% in cohort B). Conclusions: The PD-1 inhibitor remarkably improved ORR in
terms, Kegg pathways and Reactome pathways shared by 2 or more samples included pMMR/MSS LRRC compared with historical benchmark with acceptable toxicity. Up-
ubiquitination, lipid metabolism and glycosylation. There was higher PD-L1 and CTLA4 front immunochemotherapy combined with hypofractionated radiotherapy was selected
expression in anal cancer compared with HSIL, suggestive of immune escape at the transition for future definitive study. Clinical trial information: NCT05628038. Research Sponsor:
to invasive cancer. Pathogenic mutations in the Endoplasmic reticulum aminopeptidase 1 None.
(ERAP1) gene, responsible for modulating the peptide repertoire presented by MHC class I
molecules, were found in one HSIL sample and two cancer samples. Conclusions: For the
first time, this study demonstrates compelling overlap in the immunogenomic profiles of
advanced anal HSIL and neighbouring invasive cancer. The shared neoantigen burden and
overall immunogenicity supports future vaccine development in the treatment of anal HSIL
and subsequent anal cancer prevention. Furthermore, the evidence for immune escape at the
transition to invasion could motivate the use of immunotherapy in this setting. Research
Sponsor: NIHR Biomedical Research Centre at the Royal Marsden Hospital and Institute of
Cancer Research; The Syncona Foundation; The Royal Marsden Cancer Charity.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 241s

3524 Poster Session 3525 Poster Session

Harnessing transcriptome signatures and CD103+CD8+ immune infiltration POD1UM-303/INTERAACT2 subgroup analyses and impact of delayed reti-
for prognosis and treatment outcomes in anal squamous cell carcinoma. fanlimab treatment on outcomes in patients with squamous cell carcinoma
First Author: Ilma Soledad Iseas, Medical Oncology Department, Paris-St Joseph of the anal canal (SCAC). First Author: Marwan Fakih, City of Hope National Medical
Hospital, Paris,, France Center, Duarte, CA
Background: Anal squamous cell carcinoma (ASCC) is a rare malignancy linked to high- Background: SCAC is a rare cancer with high unmet medical need and no FDA-approved
risk HPV, with rising incidence among younger adults. While immunotherapy advances treatment options. POD1UM-303 is the only phase 3 study of systemic therapy completed to
have improved outcomes in metastatic ASCC, treatment for localized disease has date in advanced SCAC. The study met its primary endpoint of progression-free survival (PFS;
remained unchanged for decades, with high recurrence rates. This study investigates 9.3 mo in the retifanlimab group vs 7.4 mo in the placebo group [HR, 0.63; 95% CI, 0.47, 0.84;
molecular biomarkers and immune mechanisms predictive of chemoradiotherapy P= 0.0006]) (Rao S, et al. Ann Oncol. 2024;35:S1217). Based on these results, retifanlimab
outcomes in non-metastatic ASCC. Methods: This retrospective study analyzed 94 combined with carboplatin-paclitaxel represents a new standard of care (SOC) for inoperable
stage I-III non-metastatic anal squamous cell carcinoma (ASCC) patients treated with locally recurrent/metastatic SCAC. Here, we present outcomes for predefined subgroups of
curative chemoradiotherapy (CRT) at Hôpital Paris Saint Joseph (2010–2017) in France. interest in POD1UM-303 and exploratory analyses in patients who received open-label
retifanlimab in the crossover phase of the study. Methods: The POD1UM-303 study de-
Treatment response (CR) was assessed at 24 weeks by RECIST v1.1. Molecular analyses
sign and methods were previously presented at ESMO 2024. PFS comparisons for predefined
included whole-exome and RNA sequencing on FFPE samples to evaluate somatic
subgroups, including PD-L1 expression, region of enrollment, presence of liver metastases,
mutations, tumor mutational burden (TMB), and gene expression profiles. Immuno- extent of disease, as well as HPV and HIV status, were performed. Exploratory analyses of
histochemistry assessed immune markers (CD8, CD103). Statistical analyses identified investigator-assessed response to retifanlimab, overall survival (OS), and safety during
predictors of CR, progression-free (DFS), and overall survival (OS). Results: Complete crossover treatment were also performed. Results: A total of 308 patients were enrolled (1:1)
response (CR) was achieved in 71% of cases, with no significant differences between to receive retifanlimab or placebo with chemotherapy; 69 (45%) from the placebo + che-
treatment regimens (p . 0.05). Mutational analysis identified 172 alterations in novel motherapy group received crossover treatment with retifanlimab monotherapy upon con-
(SLAMF7 and GOLGA6L9) and previously described cancer driver genes (KMT2C, firmed progression. A consistent PFS benefit in favor of retifanlimab + chemotherapy was
KMT2D, and PIK3CA), with higher mutational burdens showing a non-significant trend observed for all predefined subgroups, including tumors with PD-L1 expression , 1%,
toward CR. Transcriptomic profiling revealed 350 differentially expressed genes among patients with liver metastases, and regardless of HPV or HIV status. Median PFS in the
CR vs. NCR patients (p-value , 0.01; FC . 2). CR was associated with modulation of retifanlimab + chemotherapy group was higher in the PD-L1 $1% vs PD-L1 , 1% groups (9.3
immune-related pathways, including TNFa/NFkB signaling (p , 0.01). Immune infiltrate mo; HR, 0.64 vs 7.5 mo; HR, 0.53) but was not impacted by presence of liver metastases.
analysis showed enrichment of CD8+ central memory T cells (p = 0.008) and CD4+ During crossover, investigator-assessed overall response rate was qualitatively similar to that
resting memory B cells (p = 0.01) in CR cases, correlating with improved OS (p = 0.0026) seen in the POD1UM-202 study, which enrolled a similar platinum-refractory population.
and DFS (p = 0.0098). CD103+CD8+ tumor-infiltrating lymphocytes emerged as the Median OS for patients receiving crossover treatment with retifanlimab was 24.3 mo,
strongest predictor of survival (OS: p = 0.011; DFS: p = 0.003), underscoring their compared with 29.2 mo for patients who were assigned to retifanlimab + chemotherapy at
potential as prognostic biomarkers and therapeutic targets in ASCC. randomization. Safety during crossover was consistent with earlier observations and
Conclusions: These findings underscore the potential of integrating molecular and comparable with experience in POD1UM-202. Conclusions: The benefits of retifanlimab
immune markers into clinical practice to better predict treatment response and guide combined with carboplatin-paclitaxel extend to the broad population of SCAC, including those
personalized therapies for CRT efficacy for ASCC patients. Further validation in inde- with tumors not expressing PD-L1 and liver metastases. Response rate and safety profile of
pendent cohorts is necessary to confirm the clinical relevance of these biomarkers and retifanlimab monotherapy in the crossover period were consistent with the previous
POD1UM-202 experience; however, exploratory analysis of survival in crossover patients
their application in therapeutic decision-making. Research Sponsor: None.
suggests first-line retifanlimab with SOC chemotherapy is preferable to sequential treatment
after progression on chemotherapy. Clinical trial information: NCT04472429. Research
Sponsor: None.

3526 Poster Session 3527 Poster Session

Genomic characterization of anal canal squamous cell carcinoma (ASCC) Monitoring botensilimab- and balstilimab-induced T-cell dynamics in re-
and outcomes on matched targeted therapy. First Author: Maliha Nusrat, fractory mismatch repair proficient metastatic colorectal cancer. First Author:
Memorial Sloan Kettering Cancer Center, New York, NY Gertjan Rasschaert, Gastrointestinal Oncology Department, University Hospitals Leuven,
Background: Anal canal squamous cell carcinoma (ASCC) is uncommon but increasing in Leuven, Belgium
incidence. 5-year survival of patients (pts) with metastatic ASCC is only 36%; new therapies Background: Mismatch repair proficient (pMMR) metastatic colorectal cancer (mCRC)
are an unmet medical need. Genomic alterations (GA) in phosphoinositol-3-kinase (PI3K) responds poorly to immune checkpoint inhibition (ICI). A better understanding of local
signaling pathway have been reported in small datasets of ASCC. Data on clinical outcomes and systemic immune cell activities is critical for improving ICI treatment efficacy.
with therapies targeting these GA in pts with ASCC are lacking. Methods: Tumor genomic T cells elicit anti-tumor specificity through their T cell receptors (TCR) and dynamic
data of pts with ASCC at Memorial Sloan Kettering Cancer Center (MSK) were obtained changes in the TCR repertoire are associated with clinical outcomes. Circulating T cells
using a targeted next generation sequencing assay (MSK IMPACT) from cBioPortal da- in the blood provide an accessible liquid biomarker to quantify and track T cell activity at
tabase. GA were annotated for biological significance using the OncoKB database, and only systems level and longitudinally. Here, we present temporal T cell tracking as a correlate
GA with known oncogenic potential were included. GA were categorized as mutations (mut), of ICI efficacy in refractory pMMR mCRC patients treated with botensilimab (BOT; Fc-
amplifications (amp), deletions (del) and fusions (fus). Clinical annotations were abstracted enhanced anti-CTLA-4 antibody) with or without balstilimab (BAL; anti-PD-1 antibody).
from electronic health records and outcomes of pts who participated in clinical trials were Methods: 10 patients from the open-label, phase 2 study (NCT05608044) with BOT in
assessed. Data were summarized using descriptive statistics and survivals were estimated refractory pMMR CRC (without metastatic liver disease) were included. In this trial
using Kaplan-Meier method. Results: Of 92,711 pts in cBioPortal, 218 (0.2%) pts had ASCC
patients were randomized into BOT (75mg or 150mg Q6W, 4x) monotherapy or in
(male n = 65, 30%). Of these 218 pts, 179 (82%) had at least 1 oncogenic GA. Oncogenic GA
combination with BAL (240mg Q2W, for 2 years), versus standard of care (regorafenib or
were most frequently identified in PIK3CA 40% (87 pts; mut 67, amp 36, with overlap),
KMT2D 19% (mut 42), BCL6 17% (amp 37), PTEN 12% (mut 18, del 9), EP300 11% (mut 24),
trifluridine/tipiracil). TCR dynamics were longitudinally assessed (0,2,4,6,12 weeks) from
KMT2C 11% (mut 20, del 3), and FBXW7 10% (mut 20, del 1). Oncogenic GA were most circulating T cells, based on deep TCR sequencing (OS-TCR, Omniscope) and quantified
frequent in the PI3K-AKT-mTOR signaling pathway (121 pts, 55%). Amps were also seen in using functional clustering. Results: 2 mCRC patients out of 10 (20%) showed partial
FGF3, FGF4, FGF19 and CCND1 in 4% pts each. Thirteen pts with metastatic treatment response (PR) while 8 (80%) had progressive disease (PD), however at variable time-
refractory ASCC participated in early phase clinical trials; 3 pts enrolled in . 1 studies (total points. However circulating T cells showed significant expansion of both pre-existing
18 trial participations). GA-matched targeted therapy was administered to 8 pts: oncogene and novel clonotypes in all patients, detectable at conserved frequencies at sequential
inhibitors in 6 pts (targeting PIK3CA E545K in 3, PIK3CA Q546K in 1, HER2 I767M in 1, FGFR2 time points. The magnitude of induced T cell clonotypes varied across treatment cycles,
amp in 1), and drugs selected for tumor suppressor gene GA in 3 pts (PTCH1 loss in 1, TP53- with repeated boosting effects observed in responders. Scoring T cell activity based on
wild in 1, FBXW7 in 1). Six pts received immunotherapy and 3 pts were treated with antivirals quantitative and qualitative TCR repertoire metrics, allowed to rank patients by their
drugs targeting Human Papillomavirus. Out of 4 pts treated with PI3K signaling inhibitors, 1 response. Intriguingly, TCR repertoire dynamics strongly correlated with clinical out-
had partial response and 2 had stable disease, with median progression free survival of 3.6 comes, establishing its potential as a quantitative biomarker for monitoring treatment
(95% CI 0-8.6) months and median overall survival of 9.1 (95% CI 5.7-12.5) months. Two out efficacy. Conclusions: Deep T cell repertoire profiling detected dynamics of circulating
of four pts treated with PI3K pathway inhibitors were on treatment for . 6 months. No T cells with quantitative and qualitative difference related to ICI response. Immune cell
response was seen in pts treated with drugs targeting GA other than PIK3CA or with tracking from liquid biopsies is a powerful tool to quantify ICI efficacy in real time.
immunotherapy; and one of three pts treated with anti-viral agents had best response of Research Sponsor: None.
stable disease. Conclusions: This is the largest characterization of GA with known on-
cogenic potential in ASCC. The PI3K signaling pathway is altered in over half of ASCC, and
PI3K-AKT-MTOR inhibitors have the potential for further investigation in pts with activating
GA in PIK3CA gene. Research Sponsor: Robert A. Winn Career Development Award.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
242s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3528 Poster Session 3529 Poster Session

Predictive role of circulating tumor DNA in pMMR locally advanced rectal Risk factors associated with de novo metastatic colorectal cancer in early
cancer patients receiving neoadjuvant chemoradiotherapy combined with onset colorectal cancer. First Author: Emma Schatoff, Memorial Sloan Kettering
sintilimab. First Author: Xiao-bin Zheng, The Sixth Affiliated Hospital, Sun Yat-sen Cancer Center, New York, NY
University, Guangzhou, Guangdong, China Background: Early onset colorectal cancer (EO CRC) is rising worldwide. Epidemiologic
Background: Circulating tumor DNA (ctDNA) has emerged as a potential biomarker for studies have shown that diets high in sugar and/or processed foods and limited exercise
various solid tumors, including colorectal cancer (CRC). It offers the advantage of longitudinal are associated with EO CRC when compared to healthy controls. However, risk factors
and dynamic surveillance of the tumor-specific genetic characteristics, eliminating the need associated with de novo metastatic disease are less well established. Identifying such risk
for repeated invasive biopsies. However, the predictive role of ctDNA in patients with factors may provide insight into modifiable lifestyle variables. Methods: All eligible EO
proficient mismatch repair (pMMR) locally advanced rectal cancer (LARC) receiving neo- CRC patients were enrolled in MSK’s Center for Young Onset Colorectal and Gastroin-
adjuvant chemoradiotherapy (CRT) combined with immunotherapy remains to be explored. testinal Cancer and completed a risk factor questionnaire (MSK-approved IRB #20-315).
Methods: In this prospective single-arm, phase II trial, pMMR LARC patients (cT3-4N0M0 and We analyzed questionnaire responses and compared risk factors between patients with
cT1-4N1-2M0) with an intermediate or high immunoscore (ISB) were enrolled (NCT05450029). de novo metastatic disease vs localized disease using Wilcoxon rank sum test or Fisher’s
Treatment-naı̈ve patients received radiotherapy (50 Gy/25 f) and 6 cycles of mFOLFOX6 plus exact test. Clinical outcomes, including overall survival (OS) from diagnosis and pro-
5 cycles of sintilimab, followed by total mesorectal excision (TME) 6-8 weeks post- gression free survival (PFS) from 1st-line chemotherapy, were estimated using Kaplan-
radiotherapy. Baseline tumor tissue DNA and serial ctDNA dynamic changes were evalu- Meier methods. The Cox regression model was used to assess association between risk
ated using next-generation sequencing. Baseline (T0) maximal somatic variant allelic fre- factors and survival outcomes. Tumors from a subset of patients (n = 206) were se-
quency (maxVAF) as well as its changes at the first (T1, two cycles after therapy) and the
quenced using MSK-IMPACT (MSK-approved IRB #12-245) and underwent genomic
second clinical evaluation (T2, four cycles after therapy) were assessed. Results: Tumor
analyses. Results: 303 patients completed the questionnaire (median age at diagnosis
somatic mutations and aligned ctDNA analyses were conducted in 43 patients. The most
frequently mutated genes in tumor tissue samples were APC (67%, n = 29), TP53 (65%, n = 28),
42; 51% Female; 88% left-sided tumors). 112 had de novo stage IV disease and 191 had
KRAS (47%, n = 20), and FBXW7 (28%, n = 12), which were also observed in plasma. Pathway stage I-III disease. Patients with de novo metastatic disease were younger, 40.9 [95%CI:
analysis indicated that mutations in SWI_SNF were more likely to be detected in patients 36.8 - 44.8] vs. 43.0 [95%CI: 38.4 - 46.4] (P-value 0.037). Analysis of dietary factors
achieving pathological complete response (pCR) (P = 0.02 for tumor tissue, P = 0.08 for showed no association with fruit, vegetable, fish, poultry, red meat, processed meat, or
plasma), suggesting a potential sensitization to sintilimab combined with CRT. For the dairy intake. However, high sugar diets were significantly associated with de novo
dynamic ctDNA analysis, 37 patients were assessed using a 950-gene panel relevant to metastatic disease, with 30 (45%) vs. 37 (29%) (P-value, 0.004) patients reporting daily
cancer. A significant decline in maxVAF from T0 to T1 was observed in the pCR group. An consumption of high sugar foods. Daily consumption of high calorie foods was also
optimal cut-off of 0.11 for the maxVAF ratio (T1/T0) was identified to discriminate complete frequently reported in patients with metastatic disease (P-value, 0.057). 3-year OS in the
responders from other patients (AUC = 0.768; sensitivity 83.3%; specificity 72.0%; P , 0.001; metastatic population was 72% [95%CI: 62% - 84%] vs. 99% [95%CI: 98% - 100%]. Within
95% confidence interval [CI] 0.597-939). Patients with a low maxVAF ratio ( , 0.11) were the metastatic group, no association was observed between daily high sugar consumption
more likely to achieve pCR following CRT plus sintilimab therapy (OR = 12.86; 95% CI: 2.23- and non-daily consumption in terms of PFS or OS. 3-year OS in the daily high sugar group
74.08; P = 0.004). Conclusions: ctDNA may serve as a potential biomarker of the response to was 79 % [95%CI: 62%-100%] vs 74% [95%CI: 57%-95%]. For early stage patients who later
CRT combined with immunotherapy in pMMR LARC. Further validation is warranted to progressed (n = 27), median time to progression was not significantly shorter among
confirm the predictive value of maxVAF and to identify additional biomarkers with potential patients who reported daily high sugar consumption (18 vs. 19 months, P-value 0.5).
predictive significance in pMMR LARC. Clinical trial information: NCT05450029. Research Genomic analyses revealed no significant differences in tumor mutational burden,
Sponsor: National Natural Science Foundation of China; 82470696 and 82103273; Guang- fraction genome altered, frequency of oncogenic or signaling pathway alterations in de
dong Basic and Applied Basic Research Foundation; 2022A1515012498 and novo metastatic vs. non-metastatic patients. Conclusions: In a single center study, in EO
2024A1515010956; the program of Guangdong Provincial Clinical Research Center for Di- CRC patients, high sugar diets may be associated with de novometastatic disease. There
gestive Diseases; 2020B1111170004; Guangzhou Science and Technology Program; were no significant genomic differences detected in patients with de novo metastatic vs.
2024A04J6400; Sun Yat-sen University Clinical Research 5010 Program; 2016005. early stage disease. Research Sponsor: None.

3530 Poster Session 3531 Poster Session

Survival impact of NeoRAS wild-type metastatic colorectal cancer: A Revisiting the relevance of sidedness in colonic tumor molecular profiling.
SCRUM-Japan GOZILA substudy. First Author: Hiroki Osumi, Department of First Author: Ashok K. Vaid, Medanta, The Medicity, Gurugram, India
Gastroenterological Chemotherapy, Cancer Institute Hospital, Japanese Foundation for Background: Colorectal cancer (CRC) is a heterogeneous disease with distinct molecular and clinical
Cancer Research, Tokyo, Japan differences between right- and left-sided tumors. This study analyzes these variations to understand their
impact on tumor behavior and treatment strategies. Methods: A total of 445 colonic tumor samples (132
Background: The “NeoRAS’’ phenomenon refers to KRAS or NRAS mutant (MT) metastatic
right-sided, 313 left-sided) were profiled to assess mutations, amplifications, and fusions in key cancer-
colorectal cancer (mCRC) that becomes RAS wild-type (WT) following treatment and may related genes along with targeted transcriptome analysis of 20,802 genes in a subset using semi-
represent a novel indication for anti-epidermal growth factor receptor monoclonal anti- conductor based next-generation sequencing (NGS) platform at Datar Cancer Genetics. Immunotherapy
bodies. We previously described the incidence and clinicopathological characteristics of biomarkers (TMB, MSI, and PD-L1 22C3 TPS) were analyzed in a subset. Results: Right-sided and left-
NeoRAS WT mCRC (Osumi et al. Nat Commun 2024); here we share the impact of NeoRAS sided colon cancers exhibit substantial molecular heterogeneity, driven by distinct genetic and epigenetic
WT on survival for patients with mCRC. Methods: Patients enrolled in the large-scale alterations (Table 1). Right-sided tumors were more frequently associated with MSI and had statistically
nationwide screening platform SCRUM-Japan GOZILA who had mCRC, tumor tissue tested significant higher incidence of BRAF mutations. KRAS mutations were frequently observed in both right-
for RAS and BRAF V600E (MEBGEN RASKET-B), and received systemic therapy were in- sided and left-sided tumors at equal rates. ERBB2 amplifications were exclusive to left side tumors,
whereas oncogenic ERBB2 mutations were equally distributed. Located around ERBB2, PGAP3 gene co-
cluded. Prior to subsequent treatment, all patients underwent next-generation sequencing
amplification too was exclusive to left sided tumors. TFE3 alterations were absent from left sided tumors
of circulating tumor DNA (ctDNA) (Guardant360) and were classified into cohorts according and common on right side. TP53 mutations, though more common in left-sided tumors, the difference
to original tissue and subsequent ctDNA genotypes: persistent RAS/BRAF V600 WT (RAS was not statistically significant. Gene expression profiling of a subset, including 103 left-sided and 41
WT), persistent RAS MT, change from RAS MT to RAS WT (NeoRASWT, including ctDNA not right-sided colon tumors, revealed activation of the Wnt/b-catenin signalling pathway, RAS/MAPK
detected), change from RAS WT to RAS MT (acquired RAS MT), persistent BRAF V600E pathway, TGF-b signalling pathway, and immune-related pathways, though these differences were not
(BRAF MT). BRAF MT outside V600 were not considered. We evaluated the clinicopatho- statistically significant, suggesting that while specific drivers may differ—such as the predominance of
logical characteristics and overall survival (OS) of patients in each cohort. OS was measured APC mutations in left-sided tumors (56.8% vs 37.9%) leading to WNT activation and the higher incidence
from time of first-line treatment initiation to the date of death. Results: The 1,352 patients of RSPO2/3 fusions (7.1% vs 1.7%) in right-sided tumors -eventually some pathways are commonly
(median age 61 years) included RAS WT: 526 (38.9%), RAS MT: 387 (28.7%), acquired RAS implicated in colorectal cancer biology. Conclusions: Existing therapies like ICIs, HER2 inhibitors, and
emerging molecules such as RSPO2/RSPO3 inhibitors could have differing impact based on tumor
MT: 223 (16.5%), NeoRAS WT: 91 (6.7%), and BRAF MT: 125 (9.2%).Median number of sidedness. Integrating these distinctions into drug development and clinical trials holds potential to
therapy lines from tissue assessment to ctDNA testing was 2 (range 1-13). NeoRAS WT had optimize treatment outcomes. Research Sponsor: None.
low prevalence of liver (23.1%, P , 0.001), lymph node (16.5%, P , 0.001) and multi-organ
Molecular profiles of right- and left-sided colon tumors.
metastasis (42.9%, P , 0.001), whereas lung (56.0%, P , 0.001) and peritoneal metastases
p-Value
(41.8%, P = 0.004) were more common in other groups. Left-sided primary tumors were more Gene Right (%) Left (%) (Chi-square test)
common with RAS WT (80.2%) followed by NeoRAS WT (70.3%) and RAS MT (67.7%), (P ,
TP53 64.5% 73.2% 0.075644
0.001). Patients with BRAF MT had significantly shorter median OS (28.1 months) compared APC 37.9% 56.8% 0.001354
to others (P Log-rank , 0.001, hazard ratio (HR), 1.91; 95% confidence interval (CI), 1.52-2.40). KRAS 50.0% 43.6% 0.220194
Patients with NeoRAS WT had median OS (45.6 months) that was between RAS WT BRAF 18.8% 3.0% 0.00001
TFE3 9.5% 0% 0.109087
(51.9 months) and RAS MT (41.0 months) (P Log-rank , 0.001). On the other hand, patients ERBB2 mutation 2.3% 2.0% 0.80941
with acquired RAS MT had a significantly shorter median OS (43.3 months) compared to RAS ERBB2 amplification 0% 5.9% 0.023006
WT (P Log-rank , 0.001). In multivariate analysis, BRAF MT (HR: 2.06, 95%CI, 1.62-2.61, P , PGAP3 amplification 0% 7.1% 0.673427
RSPO2/3 fusion 7.1% 1.7% 0.827207
0.001), ctDNA fraction ($1.0%, HR: 1.41, 95%CI, 1.17-1.71, P = 0.00035), RAS MT (HR: 1.37, Immunotherapy Biomarkers
95%CI, 1.17-1.62, P = 0.0001), and lymph node metastasis (HR: 0.85, 95%CI, 0.72-0.99, P = TMB 10-14 24.7% 29.3% 0.458066
TMB ³15 15.6% 7.6% 0.059925
0.046) were independent factors associated with shorter OS. Conclusions: Patients with MSI-High 8.1% 3.4% 0.066213
NeoRAS WT mCRC exhibited distinct characteristics, and intermediate survival between the PD-L1 Positive 15% 5.6% 0.012571
RAS WT and RAS MT groups. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 243s

3532 Poster Session 3533 Poster Session

Predicting pathologic complete response in colorectal cancer patients after Metastatic site pattern as predictor of outcome of first-line alternating
immunotherapy based on endoscopic biopsy and deep learning approach. oxaliplatin-based chemotherapy and nivolumab for patients with
First Author: Chaoyuan Xiao, Colorectal Cancer Center, Department of General Surgery, microsatellite-stable (MSS) colorectal cancer (CRC). First Author:
West China Hospital, Sichuan University, Chengdu, China Anne Hansen Ree, Akershus University Hospital, University of Oslo, Oslo, Norway
Background: Immune checkpoint inhibitors (ICIs) have emerged as effective treatments Background: The randomized METIMMOX trial evaluated short-course oxaliplatin-based
for microsatellite instability-high (MSI-H)/deficient mismatch repair(dMMR) tumors in a chemotherapy (FLOX) alternating with nivolumab for previously untreated, unresectable
select subset of colorectal cancer (CRC) patients. Patients sensitive to preoperative abdominal metastases (mets) from MSS CRC. A subgroup of patients assigned to this
immunotherapy may have the opportunity to be exempted from surgery, while those experimental (exp) treatment had remarkably extended progression-free survival (PFS)
insensitive may avoid unnecessary treatment. Tumor pathology provides rich biological compared to the control group patients given standard FLOX chemotherapy with median
insights. Several studies have indicated that deep learning algorithms can predict the PFS 9.3 months. We explored if the extent of involved organs might be decisive for re-
efficacy of immunotherapy directly from digitized hematoxylin-eosin (H&E) stained sponsiveness to the METIMMOX regimen. Methods: Patients with measurable infra-
Whole Slide Images (WSIs). However, their potential application in CRC immunotherapy diaphragmatic (liver, peritoneal, nodal) mets were randomly assigned to the control group of
remains underexplored. Based on WSIs of endoscopic biopsy, this study aims to FLOX (oxaliplatin, 5-fluorouracil, folinic acid) Q2W or the exp group of alternating 2 cycles
construct a predictive model using deep learning approach to identify potential each of FLOX Q2W and nivolumab Q2W, with prespecified break periods. Radiologic re-
pathological complete response (pCR) in CRC patients after preoperative immuno- sponse assessment was done every 8 weeks with PFS as the primary endpoint. For this post
hoc analysis, at baseline, the principal metastatic site was defined by the 2 largest mets
therapy. Methods: This study enrolled CRC patients who received preoperative im-
(main lesions) of the dominant infradiaphragmatic organ and the global metastatic pattern
munotherapy at West China Hospital, Sichuan University. Stratified randomization based
by the main and subsidiary lesions of all involved organs. Patients without adverse events
on pathological outcomes was performed, assigning enrolled patients to the training set
leading to treatment discontinuation, thus with conclusive end of treatment (EoT) tumor
(70%) and the validation set (30%). WSIs of endoscopic biopsy were used for analysis. A data, were categorized into discrete outcome groups. Results: Of 36 exp group patients
predictive model was developed based on the Swin Transformer architecture, integrating reaching the first radiologic reassessment, enabling formal evaluation, 31 proceeded to EoT
convolutional neural networks (CNNs) with a self-attention mechanism. Pre-trained tumor data. Of these, 6 patients (3 of 25 with liver main lesions, 3 of 4 with lymph node main
weights were employed for feature extraction, and the CLAM (Clustering-constrained lesions) had complete response (CR), including 3 of 3 BRAF-V600E cases. The remaining 3
Attention Multiple Instance Learning) framework was utilized to optimize pathological CR cases had tumor mutational burden (TMB) 9.4-11.8. Of all 25 patients with liver mets, 13
image analysis. The model’s performance was assessed in the validation cohort using (52%) had objective response and 5 (20%) stable disease. All 16 patients with objective
the Receiver Operating Characteristic Curve (ROC) and Area Under the Curve (AUC) was response had improved PFS (median 15.5 months, 95% CI 12.4-18.5; p , 0.001, log-rank
calculated. Attention-based visualization analysis was further performed to identify the test). None of main or subsidiary lesions in peritoneum or lungs responded to the treatment.
top patches that contributes to the determination of tumor response to preoperative The 3 outcome groups comprised 7 patients with PFS 19.8-41.6 months (longer than twice
immunotherapy. Results: 96 CRC patients treated with preoperative immunotherapy the median), 8 with PFS 9.9-16.4 months (above median), and 16 with PFS 1.9-9.2 months
were included, with 67 in the training set and 29 in the validation set. A total of 278,901 (below median). At baseline, the best outcome group cases would have been predicted by
5123512-pixel patches were generated by preprocessing 144 WSIs. A predictive model the combination of right-sided primary, small main lesions (sum of diameters 42 mm or
were established based on the training set and verified in the validation set. The model less), and all mets confined to the liver and/or lymph nodes; the mid group cases by left-
achieved an AUC of 0.82. Attention-based visualization analysis recognized the top 5% sided or rectal primary along with peritoneal or lung subsidiary lesions; and the poor
patches contributing to the determination of tumor response to preoperative immu- outcome cases by extended organ mets. Conclusions: Alternating short-course oxaliplatin-
notherapy, with 62.66% identified as tumor tissues and 37.34% identified as non-tumor based chemotherapy and nivolumab was particularly efficient in treating unresectable liver
tissues. Conclusions: Endoscopic biopsy based deep learning model, with distinct or lymph node mets from right-sided MSS CRC with intermediate TMB or the BRAF driver
attention to tumor and non-tumor regions, may provide a novel and effective tool for mutation, but inefficient at peritoneal and lung mets. Clinical trial information:
predicting pCR after preoperative immunotherapy in CRC patients. Research Sponsor: NCT03388190. Research Sponsor: Norwegian Cancer Society; Bristol-Myers Squibb.
National Natural Science Foundation of China.

3534 Poster Session 3535 Poster Session

Open-label phase Ib/II study of cetuximab (CET) plus LY3214996 with or COPEC trial: Early determination of pathological tumor response to neo-
without abemaciclib in patients (pts) with anti-EGFR-refractory metastatic adjuvant chemotherapy in low/intermediate risk stage II/II rectal cancer—A
colorectal cancer (mCRC). First Author: Guglielmo Vetere, Department of Gastro- multicenter, non-inferiority phase III randomized trial. First Author: Mingtian
intestinal Medical Oncology, The University of Texas MD Anderson Cancer Center, Wei, West China Hospital (China), Chengdu, China
Houston, TX Background: Multiple large-scale prospective studies have confirmed that neoadjuvant
Background: Acquired resistance limits the efficacy of anti-EGFR (EGFRi) therapy in RAS chemotherapy (NCT) alone can achieve optimal distant and local control in locally ad-
wild-type (WT) mCRC, often through MAPK reactivation driven by secondary RAS mutations vanced rectal cancers (LARC) without high risks. However, due to the potentially lower
or other genomic alterations. Preclinical studies on EGFRi-refractory models led by our group overall response rate compared to chemo-radiotherapy, it is rational to discontinue in-
showed that LY3214996, a potent ERK1/2 inhibitor, combined with CET suppresses MAPK effective NCT in chemo-resistant patients. In our phase II study, we applied 4 cycles of
signaling and reduces tumor growth, while the addition of Abemaciclib further enhances anti- Capox in LARC patients with low to intermediate risks, observing a considerable patho-
tumor activity by synergistically inhibiting cell cycle and survival pathways. Methods: In this clinical response rate and an accuracy of 0.89 in predicting non-responders using MRI
open-label, phase Ib/II study, RAS/BRAF/EGFR/MEK1 WT mCRC pts who progressed on prior features after two cycles of Capox. To determine the optimal number of NCT cycles and
EGFRi-based therapy and $1 chemotherapy were treated with CET + LY3214996 (Arm A) or prevent unnecessary prolonged treatment, we conducted this phase III trial to assess the
CET + LY3214996 + Abemaciclib (Arm B). Phase Ib employed a 3 + 3 design to determine the non-inferiority of two cycles of NCT compared to four cycles with respect to the final
maximum tolerated dose (MTD) and recommended phase 2 dose (RP2D). Phase II followed a pathological tumor response grade (pTRG) of 3. Methods: This multicenter, non-
two-stage design with cohort expansion to assess ORR by RECIST v1.1 as the primary inferiority, phase III randomized controlled trial was conducted at 14 centers across
endpoint. Secondary endpoints included PFS and OS. Results: Of 44 pts treated on trial, 2 did China. Eligible patients with low- to intermediate-risk stage II/III rectal cancer were
not meet inclusion criteria; 39 were evaluable for activity, and 34 for efficacy. The RP2D was randomized to receive either 2 or 4 cycles of CAPOX, followed by total mesorectal excision
200 mg LY3214996 p.o. daily + 500 mg/m² CET i.v. biweekly in Arm A with the addition of (TME) surgery. The primary endpoint was the proportion of patients with a poor patho-
150 mg Abemaciclib p.o. twice daily in Arm B. Median age was 53.0 years (IQR 47.0 – 63.8), logical response to NCT (pTRG 3). Secondary outcomes included the accuracy of MRI in
and 59.1% (26/44) were male. Most pts (95.5%, 42/44) had a left-sided or rectal primary, and predicting tumor response, treatment-related adverse events, and 3-year survival out-
all were pMMR/MSS. Prior EGFRi-based rechallenge, retreatment/reintroduction, or both were comes. Results: From August 6, 2021, to May 27, 2024, a total of 573 patients were
noted in 9.1% (4/44), 25.0% (11/44), and 4.5% (2/44), respectively. ORR, DCR, median PFS and enrolled. Ultimately, 527 patients (2-cycle group, 266 vs. 4-cycle group, 261) were included
OS were 5.3% (1/19), 36.8% (7/19), 1.8 months (95% CI 1.5 – 4.8) and 7.0 months (95% CI 5.0
in the primary analysis. The pTRG 3 rate in the 2-cycle group (27.8%, 74/266) was non-
– 22.0) for the doublet and 15.0% (3/20), 65.0% (13/20), 3.6 months (95% CI 2.5 – 4.5), and
inferior to that in the 4-cycle group (26.4%, 69/261, p = 0.722). Better lymph node response
14.0 months (95% CI 5.9 – 21.0) for the triplet, respectively. Longer time elapsed from last
was observed in the 4-cycle group (pN negative: 83.1%, 217/261 vs. 72.5%, 193/266, p =
EGFRi was associated with higher predicted probability of response after adjustment for trial
0.011). The incidence of major adverse events (grade $3, according to CTCAE 5.0) was
regimen (OR 1.35, 95% CI 1.06 – 1.94, p = 0.038). Baseline ctDNA profiling drawn prior to
comparable between the two groups (37.9% vs. 44.8%, p = 0.094). A tumor longitudinal
rechallenge revealed acquired RAS mutations in two responders, one per arm. Grade 3 TRAEs
occurred in 31.8% (14/44), with acneiform rash (9.1%, 4/44), diarrhea (9.1%, 4/44), length reduction rate (TLLR) of less than 30% on MRI predicted pathological poor re-
thrombocytopenia (6.8%, 3/44), fatigue (4.5%, 2/44), and anemia (4.5%, 2/44) being the sponders with a high positive predictive value of 0.918 after two cycles of NCT in the two-
most frequent while one Grade 4 TRAE (thrombocytopenia, 2.3%) was reported. cycle group, 0.864 after two cycles of NCT in the four-cycle group, and 0.841 after four
Conclusions: CET + LY3214996 6 Abemaciclib had a manageable safety profile with no cycles of NCT in the four-cycle group. Conclusions: Four cycles of NCT do not result in a
unexpected adverse events. Although activity was modest, this study is the first to report greater reduction in poor pathological response compared to two cycles, highlighting the
objective responses to an EGFRi-based regimen in pts harboring acquired RAS mutations in importance of early response assessment. MRI evaluation of tumor response after 2 cycles
pre-rechallenge ctDNA. Translational efforts are ongoing. Clinical trial information: predict the final pathological results with considerable accuracy. These findings lay the
NCT04616183. Research Sponsor: MD Anderson Cancer Center; Jack T. and Lillian S. Clift groundwork for future studies exploring response-guided treatment approaches in rectal
Fellowship; Andrew Sabin Family Fellowship; NIH GI SPORE; Mr. and Mrs. Jack Lee. cancer. Clinical trial information: NCT04922853. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
244s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3536 Poster Session 3537 Poster Session

Survival of patients (pts) with microsatellite stable/mismatch repair profi- The prognostic and predictive role of HER2 amplification/overexpression
cient (MSS/pMMR) metastatic colorectal carcinoma (mCRC) treated with and HER2 mutations in metastatic colorectal cancer treated with first-line
EO4010 + nivolumab (EO/N). First Author: Arvind Dasari, Department of Gastro- chemotherapy plus bevacizumab/anti-EGFRs: An individual patient data
intestinal Medical Oncology, The University of Texas MD Anderson Cancer Center, pooled analysis of eight randomized trials. First Author: Marco Maria Germani,
Houston, TX Department of Translational Research and New Technologies in Medicine and Surgery,
Background: EO is designed to expand pre-existing memory CD8 T cells cross-reacting with University of Pisa, Pisa, Italy
tumor associated antigens (TAAs). EO is composed of microbial-derived sequences mim- Background: HER2 amplification/overxpression (HER2-pos) is detected in the 5% of RAS/BRAF
icking CD8 T cell HLA-A2 epitopes on 5 TAAs, BIRC5, FOXM1, UBE2C, CDC20 and KIF2C, wild-type (wt) metastatic colorectal cancers (mCRC) and is associated with poor efficacy of EGFR
upregulated in mCRC, and the CD4 peptide UCP2. Methods: Pts had MSS/pMMR mCRC, blockade in preclinical models. Whether HER2-pos is a prognostic and/or predictive biomarker of
treated with FU, oxaliplatin, irinotecan and anti-VEGF/EGFR. Cohort (C)1 safety lead-in benefit from anti-EGFRs/bevacizumab (bev) in mCRC patients (pts) is still debated. Similarly, the
followed by expansion C2; pts received EO (300 mg/peptide in Montanide ISA 51 VG) q2 weeks role of activating HER2 mutations (mut) is unclear. Methods: We collected individual patient data
(w) x4 then q4w + N (240 mg q2w x 3 then 480 mg q4w; C1 omitted 2 first N doses). from 8 randomized clinical trials (RCTs) in the first-line treatment of mCRC: TRIBE2, TRIPLETE,
Results: 20 pts (C1 = 3, C2 = 17), 55% female, 70%/30% ECOG 0/1, median age 58 (45-80) VALENTINO, ATEZOTRIBE, PANDA, PANAMA, PARADIGM and CALGB/SWOG80405. Only pts with
years, started EO + N June 2023 to March 2024. Primary tumor right sided 35%/left 35%/rectal RAS/BRAF wt pMMR mCRC with available HER2 status by means of immunohistochemistry 6 in
30%, median 2 (1-4) tumor involved organs, 55% liver mets, 65% KRAS mutated, and median 3 situ hybridization and/or Next-generation sequencing on tumor or circulating DNA and treated
(1-5) prior lines of treatments. Any grade related AEs in . 2 pts: local administration site with triplet or doublets + bev or an anti-EGFR were included. The prognostic and predictive impact
reactions (85%), asthenia (15%), and fatigue (15%); 1 related Gr 3, local administration site of HER2-pos and HER2 mut was assessed in terms of PFS, OS and ORR. Results: 1604 pts were
ulceration; 1 related SAE, N infusion related reaction. CD8 T cells (EO/TAA peptide specific eligible. 81 (5%) tumours were HER2-pos. HER2-pos was associated with shorter PFS (mPFS: 9.8
vs 12.2 months (mos), HR: 1.31, p = 0.02) and OS (mOS: 28.0 vs 34.9 mos, HR: 1.37, p = 0.01), and
tetramers staining of PBMC ex vivo) against EO found in 10/11 tested pts, cross-reactivity
similar ORR (77 vs 72%, p = 0.47) compared to HER2-neg pts. P-values adjusted for clinically
against TAAs in all 10 positive pts. Best response (RECIST 1.1): 1 partial response (liver mets
meaningful covariates (padj) were padjPFS = 0.075 and padjOS = 0.036. We found no interaction
-47%, lung mets -34%; CEA normalized), 1 stable disease (SD) (lung mets -7%; CEA -68%,
between HER2-pos and treatment effect according to the use of bev vs anti-EGFRs in terms of
CA19-9 -55%; pat died w 17 non-related myocardial infarction), and 1 SD until w 18 (withdrawn PFS (pint= 0.76), OS (pint= 0.76) and ORR (pint= 0.64). Similar findings were reported restricting the
consent); 6 (30%) pts had target SD, and unequivocal progression of non-target lesions; 11 pts analysis to pts treated with doublets (N = 1465), and to those with left-sided tumors (N = 1315). In
(55%) had progressive disease. Median PFS 1.8 months (mo) (range 1.4-10.5). 14 pts (70%) the HER2-pos subgroup (N = 69) of pts with left-sided RAS/BRAF wild-type pMMR tumors no
received post-study anti-cancer treatment. After a median follow-up of 14.7 mo, median difference between chemotherapy/bev and chemotherapy/anti-EGFR was reported in terms of
overall survival (OS) 11.2 mo; 80% 6- and 39% 12-mo survival. Immune response assessed PFS (mPFS: 9.8 vs 9.3 mos, HR: 0.73, p = 0.30), OS (mOS: 29.8 vs 28.0 mos, HR: 1.29, p = 0.40), and
using a composite score of EO4010-specific T cell responses, measured by ex vivo tetramer ORR (59% vs 79%, p = 0.10). Activating HER2 mut were found in 27 (2%) out of 1408 HER2-neg
assays weeks 5 to 9 of treatment (best response used), with pts stratified into high and low tumors with HER2 mutational status available. Pts with HER2 mut tumors had a shorter OS
responders based on median score. Kaplan-Meier analysis with log-rank test showed trend (median: 23.7 vs 34.4 mos, HR: 1.56, p = 0.04) than HER2 wt. No interaction between HER2
towards better OS in high responders (p = 0.065). Median OS low responders 9.6 mo and not mutational status and treatment effect was evident, with no significantly different PFS (mPFS: 9.4
reached for high responders. Immune response magnitude showed no correlation with vs 5.7 mos, HR: 0.88, p = 0.76) and OS (mOS: 20.9 vs 23.7 mos, HR: 1.04, p = 0.93) in the HER2 mut
baseline T cell activation potential (by anti-CD3 stimulation/ELISPOT); independence indi- subgroup between bev and anti-EGFRs. Conclusions: This is the largest analysis of HER2 status
cates that EO4010-induced CD8 T cell expansion occurs irrespective of baseline T cell status. in untreated mCRC pts enrolled in RCTs. Waiting for targeted approaches, HER2-pos is an in-
Conclusions: EO4010 + nivolumab promotes expansion of TAA specific CD8 T cells and dependent negative prognostic factor and does not predict benefit between bev/anti-EGFRs also
shows good safety and interesting survival in previously treated MSS/pMMR mCRC. Data in left-sided tumors. HER2 mut may exert a negative prognostic impact in HER2-neg RAS/BRAF wt
suggests a potential survival benefit associated with stronger EO4010-induced immune pMMR mCRC. Research Sponsor: Gruppo Oncologico del Nord Ovest (GONO); Arbeitsgemein-
responses. Continued evaluation of EO is warranted; further immune testing data, and results schaft Internistische Onkologie (AIO); Hoffman - La Roche; Amgen; Genentech; National Cancer
of addition of bevacizumab to EO + N in a separate cohort are awaited. Clinical trial in- Institute (HHS - NIH); National Cancer Institute (HHS - NIH); National Cancer Institute (HHS - NIH).
formation: NCT05589597. Research Sponsor: Enterome.

3538 Poster Session 3539 Poster Session

Assessing the predictive role of tumor sidedness in RAS and BRAF wild-type Drug-eluting beads loaded with irinotecan-transarterial chemoembolization
metastatic colorectal cancer (mCRC) treated with first-line doublets + anti- (DEBIRI-TACE) combined with hepatic artery infusion chemotherapy and
EGFRs/bevacizumab (bev) or triplet + bev: An individual patient data pooled regorafenib in colorectal liver metastases refractory to second-line and
analysis of 10 randomized clinical trials. First Author: Marco Maria Germani, above standard systemic therapy: A single-center, phase II clinical trial
Department of Translational Research and New Technologies in Medicine and Surgery, (DREAM). First Author: Wenzhe Fan, Department of Interventional Oncology, The First
University of Pisa, Pisa, Italy Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong, China
Background: Chemotherapy (chemo) + an anti-EGFR and chemo + bev are currently regarded as preferred Background: For patients with colorectal liver metastasis (CRLM) who have failed in
upfront options for left- and right-sided pMMR RAS/BRAF wild-type (wt) mCRC patients (pts), respectively. first-line or second-line systemic treatment, the prognosis is extremely poor. Regor-
This recommendation is mainly based on trial-level pooled analyses of RAS wt pts’ cohorts, thus including
also BRAF mutant cases. Methods: We collected individual patient data from 10 randomized clinical trials
afenib is the standard third line treatment regimen. We aim to explore the safety and
(RCT) in first-line mCRC: TRIBE, TRIBE2, TRIPLETE, VALENTINO, ATEZOTRIBE, PANAMA, FIRE-3, FIRE-4, clinical efficacy of added drug eluting beads loaded with irinotecan (DEBIRI) plus hepatic
PARADIGM and CALGB/SWOG80405. RAS and BRAF wt mCRC pts treated with doublets + anti-EGFR/bev or arterial infusion chemotherapy (HAIC) to regorafenib in the combination treatment of
triplet + bev were included. Results: 2178 pts were eligible. Left- and right-sided tumors were 1780 (82%) second-line and above in patients with CRLM. Methods: For this single-center, single-
and 398 (18%), respectively. As reported in the table, among 2051 pts treated with doublets/bev or arm, prospective, phase II trial, patients with unresectable progression of CRLM after
doublets/anti-EGFR, no significant interaction effect between primary sidedness and treatment arm was
previously receiving at least six cycles of standard systemic chemotherapy above the
reported in terms of ORR (pint: 0.27) and PFS (pint: 0.32), with a p-value for interaction for OS of 0.13. Anti-
EGFR-based doublets were associated with higher ORR and longer OS among pts with left-sided tumors, second line were enrolled. All of the patients received at least once DEBIRI-TACE and
while similar outcomes were reported in right-sided ones. Among 339 pts enrolled in trials where triplet was FOLFOX-HAIC (Oxaliplatin 85mg/m2 2h + Calcium Levofolinate 200mg/m2 2h + Fluo-
included as a treatment arm, a potential interaction effect between primary sidedness and treatment rouracil 2400mg/m2 46h) plus oral regorafenib 80/120/160 mg once daily during weeks
(triplet/bev or doublet/anti-EGFR) was evident in terms of PFS (pint: 0.14) and OS (pint: 0.08) but not ORR 1-3 of each 4-week cycle until disease progression or unacceptable toxicity. The primary
(pint: 0.42). Triplet/bev was associated with longer PFS and OS among pts with right-sided tumors. endpoint of this study was the best objective response rate (ORR) per RECIST 1.1, the
Conclusions: This is the largest analysis assessing the differential effect of biologic agents and chemo
intensification according to primary tumor origin in pts with untreated RAS and BRAF wt mCRC enrolled in
secondary endpoints include progression-free survival (PFS), overall survival (OS),
RCTs. Doublets/anti-EGFR is superior to doublets/bev in left-sided tumors, with no significant differences in safety and tolerability assessments. Results: By the cutoff date of 17 November 2023,
right-sided tumors, where triplet + bev appears as the most efficacious regimen for fit pts. Further analyses 21 patients were enrolled, with the median age of 61 (ranges, 38-71 years old), 14
assessing the role of molecular hyperselection beyond RAS and BRAF are ongoing. Research Sponsor: (66.7%) of whom were male. 17(81.0%) patients had more than one tumor and 13(61.9%)
Gruppo Oncologico del Nord Ovest (GONO); Merck KGaA; Arbeitsgemeinschaft Internistische Onkologie patients had major tumor larger than 5cm. The most frequent primary cancer locali-
(AIO); Hoffman - La Roche; Amgen; Genentech; National Cancer Institute (HHS - NIH); National Cancer
zation was colon (81.0%) and the primary had been resected in 14(66.7%) patients.
Institute (HHS - NIH); National Cancer Institute (HHS - NIH); National Cancer Center.
Somatic mutation status was available for 15 patients: KRAS mutation was found in 7
Right Left Right Left patients. The ORR was 33.3% and the DCR was 100%. The median PFS was 7.8 months
Pint Pint
Doublets/ Doublets/ Doublets/ Doublets/ Triplet/ Doublets/ Triplet/ Doublets/ (95% CI: 5.8–NA), the median OS was 17.6 months (95% CI: 9.3–NA), the 1-year and 2-
antiEGFR bev antiEGFR bev bev antiEGFR bev antiEGFR
N=209 N=157 N=1120 N=565 N=95 N=189 N=32 N=23 year OS rate were 54.6% (30.7%-73.4%) and 23.9% (4.6%-51.3%), respectively, and the 1-
ORR (%) 65 62 74 66 66 83 72 78 year PFS rate was 36.4% (15.7%-57.5%). The most common TRAEs were AST/ALT
OR [95% CI] 1.13 [0.73-1.73] 1.48 [1.19-1.84] 0.27 0.42 [0.10-1.48] 0.72 [0.41-1.27] 0.42 increase (81.8%), abdominal pain (63.6%), hyperbilirubinemia (63.6%), hand-foot skin
p 0.58 , 0.001 0.16 0.25
mPFS* 9.6 10.6 12.6 12.6 12.4 10.1 12.2 13.8 reaction (59.1%), diarrhea (40.9%), etc. 9 patients presented with grade 3 or 4 TRAEs,
HR [95% CI] 1.12 [0.90-1.40] 0.99 [0.89-1.10] 0.32 0.62 [0.35-1.09] 0.98 [0.75-1.27] 0.14 which were the transient liver function injury and abdominal pain caused by TACE and
p 0.32 0.84 0.09 0.88
mOS* 25.1 29.1 36.4 33.6 37.2 23.8 37.7 35.5 HAIC. Conclusions: DEBIRI-TACE combined with HAIC and regorafenib is feasible, safe
HR [95% CI] 1.05 [0.83-1.32] 0.85 [0.76-0.96] 0.13 0.55 [0.30-1.01] 0.95 [0.70-1.30] 0.08 and shows promising efficacy in treating the patients with CRLM, even after second-line
p 0.69 0.007 0.05 0.76
and above systemic chemotherapy. Clinical trial information: NCT06071052. Research
*Months.
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 245s

3540 Poster Session 3541 Poster Session

Biomarkers of emergent resistance to sotorasib plus panitumumab in KRAS Safety and efficacy of anti-CEA CAR-T cells to prolong relapse-free survival
G12C-mutated metastatic colorectal cancer (mCRC) from the randomized, of colorectal cancer liver metastases patients after radical resection. First
phase 3 CodeBreaK 300 study. First Author: Lisa Salvatore, Medical Oncology, Author: Wei Zhang, Department of Colorectal Surgery, Changhai Hospital, Naval Medical
Università Cattolica del Sacro Cuore and Medical Oncology, Comprehensive Cancer University, Shanghai, China
Center, Fondazione Policlinico Universitario Agostino Gemelli, IRCCS, Rome, Italy Background: Approximately 75% of colorectal cancer liver metastasis patients relapse within two years after
Background: The use of sotorasib (soto; KRASG12C inhibitor) alone in patients with KRAS surgery due to circulating tumor cells and microscopic residual disease. Specific chimeric antigen receptor
(CAR) T-cell therapy, effective for hematological tumors, may also treat recurrent colorectal cancer liver
G12C-mutated mCRC gives rise to several receptor tyrosine kinase (RTK) alterations, metastases. Carcinoembryonic antigen (CEA) is a glycoprotein which is highly expressed in colorectal tumor.
resulting in treatment resistance. In CodeBreaK 300 trial, the addition of panitumumab Therefore, this study aimed to evaluate the safety and efficacy of this therapy in postoperative colorectal
(pani; monoclonal anti-EGFR antibody) counters this resistance and significantly improves cancer liver metastasis patients. Methods: We conducted a single-arm, dose-escalating phase I clinical trial
clinical outcomes compared with standard of care (SoC; trifluridine-tipiracil or regor- (NCT05240950). Key eligibility criteria were achieving no evidence of disease status after treatment and had
afenib). Over time this combination (soto+pani) may lead to the emergence of new re- CEA positivity of 30% or greater. Three dose levels of 1, 3, and 6 (10^6/kg) Anti-CEA CAR-T cells were ad-
ministered in a dose-escalating manner. The primary endpoint is safety which measures are incidence and
sistance patterns. This study reports the distribution, patterns, and prevalence of genetic severity of adverse events within 28 days and relapse-free survival at 24 months. Results: From December
alterations that arise after treatment with soto+pani. Methods: Patients from the phase 3 2021 to December 2024, 48 subjects were screened, and 12 received CAR-T cell infusion (2 in the 1 and
CodeBreaK 300 trial with chemorefractory KRAS G12C-mutated mCRC who had paired 3310^6/kg group, and 8 in the 6310^6/kg group). Three subjects who had relapsed before the infusion still
plasma samples at baseline and at progression, were included in the analysis. The samples asked for the infusion, so we proceeded to infuse after fully informing about the benefits and risks of the
were evaluated using the 753-gene Guardant Infinity ctDNA test. Gene alterations that infusion. 8 subjects experienced adverse events during treatment, including lymphopenia (5 subjects), ar-
thralgia (1 subject), fever (1 subject), and rash (1 subject). No severe adverse events occurred. The median
were absent at baseline but present at progression were considered. Results: By De-
follow-up time for the 9 pre-infusion relapse-free subjects was 23 months, of which 5 relapsed after infusion. In
cember 2024, 99 patients (median age: 62 years, female: 48%) were evaluated (soto 960 the 6310^6/kg dose group, 4 subjects remained relapse-free survival of 5, 7, 10 and 15 months after infusion,
mg-pani: 32 [60% of ITT population], soto 240 mg-pani: 33 [62% of ITT population], SoC: 34 and their follow-up is ongoing. By infusing CAR-T cell, 57.14% of the subjects in the 6310^6/kg dose group
[63% of ITT population]). Overall, 90% of patients had at least 1 emergent, likely pathogenic, were free of recurrence within two years after radical resection. Conclusions: This is the first clinical trial of
genomic alteration at progression. Median time to progression in biomarker-evaluable Anti-CEA CAR-T therapy for prolonging relapse-free survival of postoperative colorectal cancer liver me-
patients was 3.8 months for soto 960 mg-pani arm, 3.6 months for soto 240 mg-pani arm, tastases patients, showing no serious adverse events and significant reduced risk of recurrence with high
doses. Clinical trial information: NCT05240950. Research Sponsor: National Natural Science Foundation of
and 2.0 months for SoC arm. Overall, the distribution, patterns, and prevalence of China; 82072750, 82203137, 82473479; Shanghai Shenkang Hospital Development Center; SHDC2022CRT007;
pathogenic emergent alterations were similar across all treatment arms. The most Natural Science Fund of Shanghai; 20ZR1457200; Shanghai Sailing Program; 21YF1459300; Health Care
common pathogenic emergent alterations included TP53 (34%), DNMT3A (17%), ERBB2 Research Project 2024; 24BJZ10; Commission Health Industry Clinical Research Project; 20224Y0348.
(12%), and LRP1B (11%), generally associated with the RTK, cell cycle control, DNA Clinical information of 9 pre-infusion relapse-free subjects.
methylation, and DNA damage response pathways. The median copy number of emergent Post-infusion Post-
KRAS copy number variations (CNVs) was higher (p = 0.007) in the soto 960 mg-pani (4.18) Infusion Current relapse-free infusion Overall
Subhects TNM dose NED survival time survival time survival time
and soto 240 mg-pani (4.24) arms compared with the SoC arm (2.05). Emergent KRAS number Stage (310^6/Kg)1 status (months)2 (months)2 (months)3
CNVs were primarily present in the soto+pani arms (soto 960 mg-pani: 40.6% [n = 13], soto
S01002 T3N0M1a 1 No 3 27 33
240 mg-pani: 36.4% [n = 12], and SoC: 14.7% [n = 5]). The presence of emergent pathogenic S01037 T2N1bM1a 3 No 12 12 26
variants in ALK (n = 7) and KMT2D (n = 4) was observed exclusively among patients treated S01008 T3N0M1a 6 Yes 10 10 25
S01010 T3N1M1a 6 No 10 10 26
with soto+pani. DNMT3A mutations, along with other diverse emergent alterations, were S01015 T3N0M1a 6 Yes 15 15 21
observed in all three evaluable patients with partial response in the soto 960 mg-pani arm. S01023 T3N0M1a 6 No 12 14 23
S01042 T3N2aM1a 6 No 3 10 16
Conclusions: Dysregulation of the DNA methylation and RTK pathways and KRAS am- S01033 T3N1bM1a 6 Yes 7 7 18
plifications may contribute to the development of resistance to soto+pani combination. S01043 T3N1bM1a 6 Yes 5 5 14
Further characterization of these acquired alterations, can help inform future therapeutic 1
One subject in each of the 1, 3, and 6 dose groups relapsed before infusion.
strategies. Clinical trial information: NCT05198934. Research Sponsor: Amgen Inc. 2
From the day of infusion.
3
From the day of radical resection.

3542 Poster Session 3543 Poster Session

Preliminary safety, pharmacokinetics, and clinical activity of RG6344 in Impact of anti-EGFR and anti-VEGF antibodies on survival in BRAFV600E
patients with BRAF V600E-mutant metastatic colorectal cancer (mCRC). mutated metastatic colorectal cancer: A pooled analysis of eight clinical
First Author: Elisa Fontana, Sarah Cannon Research Institute UK, London, United trials performed in the first-line treatment of mCRC (German AIO Study
Kingdom Group). First Author: Lena Weiss, Department of Medicine III, University Hospital
Background: RG6344 (RO7276389) is a novel paradox breaker and brain penetrant Munich LMU, Munich, Germany
BRAF inhibitor (BRAFi) designed to overcome the MAPK paradoxical activity, a well- Background: BRAFV600E mutation in metastatic colorectal cancer (mCRC) is associated
established liability of the first generation BRAF inhibitors. The BRAF V600E mutation, with poor prognosis. Registrational approval of anti-EGFR antibodies does not exclude
present in about 10% of mCRC patients, negatively impacts prognosis and response to their use in BRAFV600E mutated (mut) mCRC, while current guidelines explicitly advise
standard therapies. Methods: Dose escalation of RG6344 is being conducted in par- against the use of anti-EGFR-directed therapy and recommend the use of chemotherapy
ticipants with solid tumors harboring BRAF V600E mutation up to the protocol specified plus anti-VEGF antibodies. The present analysis of single-patient data evaluates the
maximum daily dose, to define the maximum tolerated dose (MTD) and/or recommended therapeutic benefit from anti-EGFR- vs. anti-VEGF-directed therapy in BRAFV600E mut
Phase 2 dose and characterize safety, PK/PD, and clinical outcomes (ISRCTN13713551). mCRC. Methods: We conducted a pooled analysis of eight first-line AIO-studies (FIRE-1,
Results: As of September 25, 2024, 51 patients with mCRC (27, 53% with prior BRAFi FIRE-3, FIRE-4, FIRE-4.5, CIOX, XELAVIRI, PANAMA, VOLFI) including 251 evaluable pts
treatment; median number of treatments 3 [2-6]), including 4 patients with non- with BRAFV600E mut and RAS wild-type mCRC. Right-sided primary tumors (RSPT)
measurable brain lesions, have been treated with RG6344 in the monotherapy dose included tumors from the caecum to the colon transversum, while left-sided tumors
escalation part of the study. Patients received at least one dose of study drug as a single (LSPT) included the splenic flexure to the rectum. Results: Of 251 BRAFV600E mut pts,
agent. MTD has not been reached up to the highest dose of 3600 mg/d. Of the 51 treated exact primary tumor location was available in 230 pts. In this cohort, 117 were male
patients, Grade 3 treatment-related AEs (TRAEs) occurred in 8 patients (14.5 %), grade 4 (50.9%) and 113 female (49.1%). LSPT was observed in 106 (46.1%) pts compared to 124
TRAEs in 2 patients (3.6%; both laboratory findings) and no grade 5 TRAEs were re- (53.9%) with RSPT. In the entire cohort, median OS (mOS) of LSPT vs. RSPT did not differ
ported. The most commonly reported TRAEs included diarrhoea (23.6%), nausea (21.8%) significantly (15.2 months vs. 13.4 months; HR 0.96; 95% CI, 0.70–1.29; P=0.77). Pts
and fatigue (12.7%). 3 patients (5.5%) discontinued study treatment due to TRAEs. None with LSPT showed a numerical survival benefit with anti-EGFR therapy compared to anti-
of the typical BRAFi class toxicities, such as cutaneous squamous cell carcinomas VEGF therapy (17.8 months vs. 11.8 months; HR 0.71; 95% CI, 0.45–1.14; P=0.16). This
(cSCCs), Palmar-Plantar Erythrodysesthesia (PPE) and keratoacanthoma, have been effect was observed independent of sex. In contrast, pts with RSPT showed a trend
observed to date, highlighting the paradox breaking properties of this BRAF inhibitor. towards inferior outcome with anti-EGFR vs. anti-VEGF therapy (11.6 months vs.
Linear and time-independent PK was demonstrated across the tested dose range, 17.1 months; HR 1.31; 95% CI, 0.84–2.05; P=0.23). This effect was primarily
reaching Ctrough levels exceeding pERK inhibition . 80%. Strong and early ( i.e. FDG driven by females, who experienced a significant survival disadvantage with
PET at 15 days) metabolic response of 74% (6 CMR, 30 PMR out of 49 evaluable patients) anti-EGFR therapy (10.2 months vs. 17.1 months; HR 1.85; 95% CI, 1.05–3.25; P=0.031).
was observed on FDG PET. Association between metabolic responses and exposure was For males, however, both anti-VEGF and anti-EGFR antibodies were associated with
observed. Observed ORR (RECIST v1.1) was 25% in BRAFi-naive mCRC patients and comparable outcome. Conclusions: The present analysis performed in the first-line
14.8% in BRAFi-experienced patients, DCR was 100% for BRAFi-naive patients and treatment of BRAFV600E mut mCRC suggests a survival benefit from anti-EGFR
62.9% for BRAFi-experienced patients, mPFS was 7.3 months and 3.6 months, re- antibodies in pts with LSPT, independent of gender. Male pts with RSPT appear to
spectively, in the ongoing study. Conclusions: RG6344 is well tolerated allowing un- derive comparable benefit from anti-EGFR and anti-VEGF antibodies, while
precedented exposure for pERK inhibition and shows promising preliminary single-agent female pts exhibit a survival disadvantage from anti-EGFR antibodies. Clinical trial
activity. Clinical trial information: ISRCTN13713551. Research Sponsor: None. information: NCT00433927 (FIRE-3), NCT02934529 (FIRE-4), NCT04034459 (FIRE-4.5),
NCT01249638 (ML22011), NCT00254137 (CIOX), NCT01991873 (PANAMA),
NCT01328171 (VOLFI). [[Link]]. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
246s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3544 Poster Session 3545 Poster Session

Circulating tumor DNA (ctDNA) dynamics in liver-limited metastatic colo- Tumour microbiome and immune dysregulation in early-onset colorectal
rectal cancer (mCRC) patients resected after first-line systemic treatment. cancer. First Author: Rachel Violet Purcell, Department of Surgery and Critical Care,
First Author: Vittorio Studiale, Unit of Medical Oncology 2, Azienda Ospedaliera Uni- Christchurch, New Zealand
versitaria Pisana and Department of Translational Research and New Technologies in Background: Early-onset colorectal cancer (EOCRC) in patients under 50 years is in-
Medicine and Surgery, University of Pisa, Pisa, Italy creasing incidence in many countries worldwide, including New Zealand. The reason for
Background: Liver-limited disease (LLD) occurs in 20-30% of metastatic colorectal cancer this trend is yet unclear, but is associated with such risk factors as obesity, alcohol
(mCRC) patients. Although 20-30% of patients who undergo resection can achieve a long- intake, lifestyle and diet. This suggests a multi-factorial exposome-related aetiology,
term overall survival benefit from liver surgery, most patients relapse during the first two with changes to the gut microbiome likely to play a role. Methods: In this study, we
years after hepatectomy. ctDNA is a promising tool in detecting the presence of minimal investigated differences in the tumour-resident microbiome and molecular character-
residual disease (MRD) after resection of colorectal liver metastases and a reliable istics between patient cohorts of EOCRC and late-onset CRC (LOCRC). Transcriptomic
prognostic tool for recurrence. ctDNA and its dynamics may also serve as a prognostic tool analysis was carried out on pre-treatment tumours from a cohort of 19 EOCRC patients
in patients candidate to liver resection following upfront chemotherapy. Methods: mCRC and compared to a control group of 196 LOCRC aged over 65 years. Bioinformatics
patients (N = 116) with initially unresectable LLD and R0/R1 resected after upfront che- analysis of RNA sequencing data was used to analyse tumour microbial abundance and
motherapy were selected from 3 Italian academic centers. Blood samples were collected taxonomy, differential gene expression and gene-set enrichment between the two
prospectively at baseline (T0), pre-surgery (TPrS) and post-surgery (TPoS). T0 samples were groups, as well as assign consensus molecular subtypes (CMS) . Results: We found an
evaluable for 82 patients, TPrS for 116 and TPoS for 60. Biobanked plasma samples were increase in expression of genes involved in the cell cycle in the EOCRC cohort, and of
analyzed with the Tempus xM MRD assay (xM), a tumor-naı̈ve ctDNA MRD assay that specific genes (e.g. HOXA11-AS, STMN2) involved in cell proliferation and metastasis.
integrates methylation and genomic variant classifiers to deliver a binary MRD call blinded
Converesly, enriched gene sets in the late-onset category were predominantly related to
to clinical outcomes. The methylation classifier detects fragments with CRC methylation
immune function. When grouping CMS subtypes as immune-rich (CMS1/CMS4) versus
signatures in differentially methylated regions trained by sequencing CRC and presumed-
healthy samples on a 6 Mbp panel. The variant classifier detects highly prevalent CRC
immune-depleted (CMS2/CMS3), there was a significant difference between the two
variants. Results: Methylation results were available for 60 TPoS patients with a clinical groups with 94% of EOCRC tumours being immune-depleted, compared to 67% of late-
sensitivity of 56.4% and specificity of 100%. TPoS ctDNA status was associated with onset tumours (p = , 0.05). Meanwhile, we found an increase in bacterial richness
relapse-free survival (RFS) with the ctDNA- group experiencing longer median RFS (mRFS) (observed alpha diversity) in the late-onset tumours compared to early-onset, while there
than ctDNA+ (HR = 6.7, mRFS . 24 mos vs. 5.5 mos, p , 0.001). Patients who were were no differences in the Shannon alpha diversity measures (richness and evenness),
persistently ctDNA- by methylation calls (n = 20) or converted to negative (n = 13) from TPrS and in beta diversity between the groups. We also found a depletion of the bacteria
to TPoS experienced longer RFS (mRFS 16.3 mos and . 24 mos respectively). Those who Helicobacter canadensis and Campylobacteria ureolyticus in the early-onset cohort,
remained persistently ctDNA+ (n = 9) or converted to ctDNA+ (n = 12) had a mRFS of 5.3 and while there was an increase in Lachnospiraceae species. Conclusions: Our results add
5.9 mos respectively. Patients with variant allele fraction (VAF) reduction of $50% from T0 to the growing body of evidence that EOCRC is a distinct disease from LOCRC. EOCRC
to TPrS (N = 53) experienced longer RFS than those who had , 50% reduction or increase in shows lower tumour-immune activation compared to LOCRC and very low rates of CMS1
VAF (N = 18) (HR 2.21, mRFS 18.8 mos vs. 9.8 mos, p = 0.012). Lastly, patients that remained and CMS4 subtypes, which is associated with a distinct tumour-resident microbiome.
positive from T0 to TPrS (N = 23) experienced a numerically shorter RFS compared to those This may have implications for prognosis and targeted treatments. Research Sponsor:
who converted to negative (N = 47) (median RFS 10.4 and 15.1 mos, HR 1.65, p = 0.10). None.
Conclusions: xM demonstrates remarkable performance in predicting clinical recurrence
and correlation to RFS at TPoS in LLD mCRC patients resected after upfront systemic
therapy. Interestingly, patients with a VAF reduction $ 50% experience longer RFS following
surgery, suggesting a potential role for this tool in multidisciplinary decision making in this
setting. Research Sponsor: None.

3546 Poster Session 3547 Poster Session

Prognostic and predictive impact of the baseline systemic proteome in Prognostic and predictive role of HLA supertypes in pMMR mCRC patients
patients with RAS wild-type metastatic colorectal cancer: Analysis from the receiving FOLFOXIRI/bev 6 atezolizumab in the AtezoTRIBE study. First
randomized phase II PanaMa (AIO KRK0212) trial. First Author: Alexej Ball- Author: Daniele Rossini, Department of Experimental and Clinical Medicine, University of
hausen, Department of Hematology, Oncology and Tumorimmunology, Charité-Uni- Florence. Oncology Unit, Careggi University Hospital, Florence, Italy
versitätsmedizin Berlin, Freie Universität Berlin, Humboldt-Universität zu Berlin, and Background: The HLA system plays a crucial role in the development of the adaptive immune response,
Berlin Institute of Health, Berlin, Germany influencing antigen presentation and T-cell-mediated tumour recognition. Emerging evidence suggests that
specific HLA allele groups named supertypes may influence the efficacy of immune-checkpoint inhibitors
Background: Systemic proteomics offers a minimally invasive approach to identifying
(ICIs). We investigated the impact of HLA supertypes in the pMMR cohort of mCRC patients (pts) treated
biomarkers in metastatic colorectal cancer (mCRC). This study analyzed the prognostic and with FOLFOXIRI/bev 6 atezolizumab in the AtezoTRIBE study. Methods: Genomic DNA from blood
predictive potential of the systemic proteome in RAS wild-type (wt) mCRC patients from the samples was genotyped using Oncoarray, a custom array manufactured by Illumina including approxi-
PanaMa trial, which investigated maintenance therapy with fluorouracil/folinic acid (FU/FA) mately 530K SNP markers. HLA class I and II alleles were characterized with minimac3 algorithm using the
6 panitumumab (Pmab) following induction with FU/FA plus oxaliplatin and Pmab. Four-digit Multi-ethnic HLA v2 (2022) reference panel. The presence of 22 HLA supertypes were assigned
Methods: Baseline serum samples were analyzed using liquid chromatography-mass based on the imputed dosages across relevant alleles. The effects of HLA supertypes on survival were
spectrometry to identify protein markers. Their impact on overall survival (OS) and evaluated with Cox proportional hazard models. Given the exploratory nature of the analysis, no ad-
progression-free survival (PFS) was evaluated using Kaplan-Meier estimates and Cox re- justments for multiple comparisons were applied. Results: Among 153 assessed pts (102 and 51 treated
with FOLFOXIRI/bev/atezo and FOLFOXIRI/bev, respectively), B44 and DR9 supertypes were associated
gression. Prognostic analyses utilized hierarchical clustering and random forest models to
with worse prognosis in terms of both PFS (mPFS 11.4 months (mos) for B44 pos vs 13.9 mos for B44 neg;
delineate protein groups associated with survival outcomes, enhanced by sparse partial least HR 1.74; 95% CI 1.19–2.52; p = 0.004, and mPFS 11.4 mos for DR9 pos vs 13.2 mos for DR9 neg; HR 2.37;
squares discriminant analysis for feature selection. Gene ontology analysis was used to 95% CI 1.01–5.60; p = 0.04, respectively) and OS (mOS 29.6 mos for B44 pos vs 36.6 mos for B44 neg; HR
identify biological functions of these markers. ROC analysis was performed to evaluate the 1.59; 95% CI 1.03–2.45; p = 0.038, and mOS 26.6 for DR9 pos vs 33.9 mos for DR9 neg; HR 3.22; 95% CI
accuracy of prognostic protein signatures. For predictive analysis, PFS outcomes of 1.24–8.38; p = 0.017, respectively) in multivariable analysis. As summarized in the Table, PFS and OS
maintenance with FU/FA 6 Pmab were assessed using Kaplan-Meier estimates and Cox benefit from the addition of atezolizumab to FOLFOXIRI/bev was reported among A3 neg but not A3 pos
regression. Hazard ratios (HR), differences in hazard ratios (delta HR), and statistical sig- patients, and B8 pos patients derived higher benefit than B8 neg. Conclusions: Our exploratory findings
nificance (p-values) were used to differentiate patient outcomes based on protein marker suggest that HLA supertypes could influence prognosis and ICIs-based treatment efficacy in pMMR mCRC
pts. In particular, B8 and A3 supertypes could identify patients more likely to benefit from the addition of
expression. Results: Of 378 patients treated in the trial, 231 had baseline serum samples.
ICIs to FOLFOXIRI/bev. These findings highlight the potential of HLA profiling to optimize the use of
Proteomic clustering identified two survival clusters: the high-survivability cluster showed immunotherapy in pMMR mCRC pts. Research Sponsor: GONO Foundation.
significantly longer induction PFS (HR 0.75, 95% CI 0.56–1.00, P = 0.05) and OS (HR 0.63, 95%
CI 0.45–0.88, P = 0.01). Hierarchical clustering revealed 470 proteins, with specific proteins Median PFS Median OS
enriched in high-survivability (e.g., ALB, APOA2) and low-survivability (e.g., SERPINA3, CRP) FOLFOXIRI/bev/ FOLFOXIRI/bev P for FOLFOXIRI/bev/ FOLFOXIRI/bev P for
atezo (months) (months) HR 95%CI interaction atezo (months) (months) HR 95%CI interaction
clusters. Gene ontology analysis highlighted distinct pathways, such as enzyme inhibitor
activity in low-survivability clusters and peptidase regulator activity in high-survivability A3 0.048 0.064
Pos 12.5 11.6 0.89 27.0 31.7 0.96
clusters. For maintenance, prognostic arm-specific proteomic signatures linked to improved (0.53- (0.53-
PFS with strong accuracy in the FU/FA + Pmab arm (AUC 0.99), and FU/FA arm (AUC 1.00). 1.51) 1.73)
Neg 15.0 10.1 0.43 NR 27.3 0.44
Predictive analysis revealed a total of eight proteins that predicted benefit of Pmab addition (0.26- (0.24-
to maintenance. A positive combined proteomic biomarker including these proteins (ITIH4, 0.71) 0.81)
B8 0.014 0.040
FLNC, HP, CTPS1, SERPINA1, HRG, MAN1C1, C4A) predicted significant benefit of addition of Pos 13.7 5.6 0.16 35.9 16.7 0.21
Pmab to FU/FA maintenance (PFS: HR 0.68, 95% CI 0.49–0.95, P = 6.4e-09). (0.05- (0.06-
0.60) 0.76)
Conclusions: Proteomic profiling identified prognostic clusters linked to distinct survival Neg 13.3 11.6 0.63 36.1 31.4 0.70
outcomes and predictive signatures for FU/FA 6 Pmab maintenance, supporting its utility in (0.43- (0.44-
0.93) 1.11)
guiding personalized treatment strategies for RAS wt mCRC. Research Sponsor: AIO-Studien
gGmbH; Amgen Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 247s

3548 Poster Session 3549 Poster Session

Impact of Medicare Advantage (MA) on timely initiation of pembrolizumab Outcomes of young-onset colorectal cancer vs late-onset colorectal cancer
among dMMR/MSI-h metastatic colorectal cancer patients. First Author: Baqir patients on phase 1 matched and non-matched therapies. First Author:
Jafry, Charleston Area Medical Center, Charleston, WV Daniel Aaron Fox, Baylor College of Medicine, Houston, TX
Background: With MA plans covering over half of Medicare beneficiaries, concerns remain about their ability to manage Background: Systemic therapy recommendations for young-onset colorectal cancer (YOCRC),
complex cancer care due to pre-authorization and limited provider network. This study evaluates MA versus Traditional
Medicare (TM) regarding timely initiation of Pembrolizumab for dMMR/MSI-H colorectal cancer (CRC), following its 2020 FDA
CRC diagnosed at , 50 years old, are similar for late-onset CRC (LOCRC) despite possible
approval. Methods: This study utilized nationwide Flatiron Health Electronic Health Record-derived de-identified database. differences in biologic behavior. This study aims to compare outcomes among YOCRC and LOCRC
We included patients diagnosed with dMMR/MSI-H CRC from 2020 onward, aged $65 years, who had at least one clinic visit patients on Phase 1 matched and non-matched therapies. Methods: This was a single-institution
within six months of diagnosis and were insured under MA or TM. The primary endpoint was initiation of Pembrolizumab retrospective analysis of patients with CRC who received treatment on a Phase 1 clinical trial. Only
within 30-, 45-, and 90-days post-diagnosis. Multivariable logistic regression with Inverse Probability Weighting, adjusted for
the first Phase 1 therapy for each patient was included in analysis. Matched therapy was defined
SES (defined by Yost score), age, race, ECOG, practice type and diagnosis year, evaluated impact of insurance type on timely
treatment initiation. Results: Out of 597 dMMR/MSI-H metastatic CRC patients identified since 2020, 219 had at least one as therapy targeting genomic alterations or their signaling pathways. Distributions of
clinical visit under MA (N = 86) or TM (N = 133) plans. Predominantly, patients in our cohort were non-Hispanic White (72%), progression-free survival (PFS) were estimated by the Kaplan-Meier method. Log-rank test was
with higher SES (59%), diagnosed with de novo dMMR/MSI-H (59%), and treated in community hospitals (86%). Of 95 patients performed to test the difference in survival between groups. A propensity score matched analysis
who commenced pembrolizumab as initial therapy, 62 (65%) had TM and 33 (35%) had MA. The adjusted analysis revealed was created using a multivariate logistic regression model. Covariates in the model included:
that MA patients were significantly less likely to start pembrolizumab within 90 days of diagnosis compared to TM patients
(OR: 0.58; 95% CI: 0.34-0.97; P: 0.04). There were no statistical differences in starting treatment at 30 (OR: 0.79; 95% CI: 0.39- gender, race, lung metastasis, liver metastasis and tumor sidedness. Results: 577 patients were
1.56; P: 0.5) or 45 days (OR: 0.87; 95% CI: 0.48-1.57; P: 0.7). Subgroup analysis indicated substantial delays in pembrolizumab included in analysis (Table 1). 252 patients had YOCRC (43.7%) and 325 had LOCRC (56.3%). 100
initiation within 90 days among lowest SES (OR: 0.31; 95% CI: 0.13-0.71; P: 0.007; P-interaction: 0.087) and those treated in YOCRC patients (39.7%) and 90 LOCRC patients (27.7%) received matched therapies. Before
community hospitals (OR: 0.49; 95% CI: 0.27-0.87; P: 0.016; P-interaction: 0.095). Conclusions: Patients with TM were more propensity score matching YOCRC patients on matched therapy had higher odds of achieving a
likely to initiate pembrolizumab earlier than those with MA. This underscores necessity to scrutinize the influence of MA on
cancer delivery, especially for patients with lower SES. Research Sponsor: None.
response compared to LOCRC patients on matched therapy (complete response/partial response)
(10.5% vs 3.4%, odds ratio (OR): 3.294 (95% confidence interval (CI)): 0.876, 12.390), p=0.0777).
Baseline characteristics.
After propensity score matching YOCRC patients on matched therapy had higher odds of
Variable TM MA P
achieving a response compared to LOCRC patients on matched therapy (13.3% vs 3.9%, OR was
Age 0.4
>=65 66 (50%) 40 (47%) not estimable, p=0.0082). No significant differences in overall response rate were detected
>=75 67 (50%) 45 (52%) between YOCRC and LOCRC patients on non-matched therapy before or after propensity score
>=85 0 (0%) 1 (1%)
Race 1 matching (5.4% vs. 4.1%, OR: 1.362 (95% CI: 0.513, 3.615), p=0.5348; 5.5 vs. 4.3%, OR: 1.167 (CI:
NHW 97 (73%) 62 (72%)
Non-NHW 36 (27%) 24 (28%) 0.392, 3.472), p=0.7815). No significant differences in PFS were detected between patient with
ECOG
0/1 95 (71%) 68 (79%)
0.7 YOCRC and those with LOCRC in any of the patient cohorts before or after propensity score
2 18 (14%) 8 (9%) matching (all patients: p=0.743, p=0.639; patients receiving matched therapy: p=0.497, p=0.909;
3/4
Unknown
8 (6%)
12 (9%)
4
6
(5%)
(7%) patients receiving non-matched therapy: p=0.999, p=0.62). Conclusions: YOCRC patients were
SES
0/1 38 (29%) 36 (42%)
0.04 more likely than LOCRC patients to achieve a response on matched therapy though this did not
3/4/5 88 (66%) 42 (49%) translate to improved PFS. Nevertheless, matched therapies are associated with increased
Unknown 7 (5%) 8 (9%)
Diagnosis Year 0.5 response rate for YOCRC patients. Research Sponsor: None.
2020 36 (27%) 22 (26%)
2021 41 (31%) 34 (40%) Patient demographics.
2022 39 (29%) 18 (21%)
2023 17 (13%) 12 (14%) Trait n (%)
DeNovo 0.7
Yes 80 (60%) 49 (57%) Age at diagnosis (median, range) 51 (18-83)
No 53 (40%) 37 (43%) White/Caucasian 425 (74.7%)
Practice 0.4 Black/African American 67 (11.6%)
Academic 16 (12%) 15 (17%)
Community 117 (88%) 71 (83%) Asian 40 (6.9%)
Time to Pembrolizumab Initiation (days) OR (95CI) P Hispanic/Latino 79 (13.7%)
30 1 0.79 (0.39-1.56) 0.50 Microsatellite Instability-High Tumor 5 (0.9%)
45 1 0.87 (0.48-1.57) 0.66
90 1 0.58 (0.34-0.97) 0.04
KRAS mutation 352 (61.0%)
BRAF V600E mutation 38 (6.6%)
Prior Unique Lines of Therapy (median, range) 4 (0-11)

3550 Poster Session 3551 Poster Session

Amivantamab treatment and intra-tumoral gene expression and immune cell Phase 1 dose escalation results of the WEE1 inhibitor, azenosertib (A), in
changes in refractory metastatic colorectal cancer (mCRC): Whole tran- combination with encorafenib (E) and cetuximab (C) in patients (pts) with
scriptome RNA-sequencing analysis from the OrigAMI-1 study. First Author: previously treated BRAF V600E mutant metastatic colorectal cancer
Cathy Eng, Vanderbilt-Ingram Cancer Center, Nashville, TN (mCRC). First Author: Jeanne Tie, Department of Medical Oncology, Peter MacCal-
Background: Amivantamab (ami) is an FDA- and EMA-approved EGFR-MET bispecific an- lum Cancer Centre, Melbourne, VIC, Australia
tibody with immune cell–directing activity for EGFR-mutated advanced non-small cell lung Background: Encorafenib + cetuximab was approved for treating pts with BRAF V600E
cancer. Ami monotherapy has shown promising activity in participants (pts) with refractory mutant mCRC after prior systemic therapy based on the Phase 3 BEACON study (observed
mCRC, independent of primary tumor sidedness (left- or right-sided). Here, we analyzed gene response rate 20%; Kopetz et al. 2019). Azenosertib is a highly selective WEE1 inhibitor
expression data from the OrigAMI-1 study to identify mechanisms of sensitivity and action in that causes mitotic catastrophe and cell death. Combining BRAF targeting treatment (E+C)
response to ami monotherapy. Methods: The phase 1b/2 OrigAMI-1 study (NCT05379595) with orthogonal pathway inhibitors may allow for an additive or synergistic combination
enrolled pts with mCRC harboring wild-type KRAS, NRAS, BRAF, and EGFR ectodomain, effect of A. This study aimed to evaluate safety and tolerability, determine the maximum
without ERBB2/HER2 amplification. Pts with left-sided mCRC without prior anti-EGFR therapy tolerated dose (MTD), and assess anti-tumor activity in pts with BRAF V600E mutant mCRC
(Cohort A) and with prior anti-EGFR therapy (Cohort B), as well as pts with right-sided mCRC receiving A+E+C. Methods: This phase 1 dose escalation, open-label, multicenter study
(Cohort C) received intravenous ami monotherapy (1050 mg; $80 kg: 1400 mg). Tumor (NCT05743036) evaluated safety, tolerability, and activity of A administered in combi-
biopsy samples were collected at screening and Cycle 3 Day 1 (C3D1; if feasible). Whole nation with E+C in adult pts with BRAF V600E-mutant mCRC who received 1-3 prior
transcriptome RNA-sequencing data of paired baseline and C3D1 tumor samples were regimens for metastatic disease. Azenosertib is a CYP3A4 substrate and is predicted to
generated by Foundation Medicine. Gene expression data were analyzed using standard moderately inhibit CYP3A4 while weakly inhibiting CYP2C19. Encorafenib is primarily
bioinformatic methods to identify gene signatures associated with ami treatment in baseline metabolized by CYP3A4 and CYP2C19, and acts as a CYP3A4 inducer. To allow for ex-
tumor samples (n = 76) and paired baseline and C3D1 tumor samples (n = 17). Results: High posures to optimize treatment benefits and minimize toxicity, the recommended starting
baseline mRNA expression of AREG and EREG ligands was associated with treatment re- dose of E is 150 mg once a week. The primary endpoint was dose-limiting toxicities (DLTs)
sponse across all cohorts (n = 76). In Cohort A (n = 16), median progression-free survival was
in cycle 1. Pts were treated across 5 dose-finding cohorts, receiving A (dose range: 100 mg-
significantly longer for pts with high (n = 8) vs low (n = 8) AREG expression (9.1 mo vs 4.5 mo,
400 mg once a day [QD] oral [PO] on a continuous schedule) and E (dose range: 75 mg or
respectively; P, 0.01). Differential expression analyses after ami treatment showed sig-
150 mg QD PO), and C (500 mg/m2 intravenous twice a week) until disease progression or
nificant changes in . 800 genes (P, 0.01) across all cohorts (n = 17). The EGFR pathway was
unacceptable toxicity. Results: As of Nov 25, 2024, 44 pts were enrolled and treated with a
significantly downregulated after ami treatment (P, 0.01). Pathway enrichment analyses
median age of 64 years. 52.3% of pts received $2 prior lines of therapy, 34 pts were BRAF
identified significant enrichment of cell cycle and natural killer (NK) cell–mediated cyto-
toxicity pathways. The cell cycle pathway score was significantly downregulated following inhibitor (BRAFi)-naı̈ve. The most frequent treatment-related Grade $3 adverse events
ami treatment (P, 0.01), implying reduced cell proliferation. A significant upregulation of were asthenia (11.4%) and fatigue (6.8%). DLTs were observed at the dose levels of
dendritic cell (P, 0.005) and T-cell–inflamed signature scores (P, 0.05) was observed with A300+E150 and A400+E75 and included dose-limiting fatigue, atrial fibrillation, recurring
ami treatment, potentially implying increased immune cell infiltration of the tumor micro- elevated bilirubin (all Grade 3), and Grade 4 neutropenia. The MTD was determined to be
environment. Ami also increased the cytolytic (P, 0.05) and NK cell–mediated cytotoxicity A300+E75 and C. Twelve of 34 (35.3%) BRAFi-naı̈ve pts achieved confirmed response per
pathway scores (P, 0.05), implying an increase in cytotoxic immune cells. Additional RECIST v1.1 (2 complete response, 10 partial response [PR]) while none of the BRAFi-
biomarker analyses are ongoing and will be presented at the meeting. pretreated pts responded. Seven of 17 (41.2%) BRAFi-naı̈ve pts treated at A300+75 or
Conclusions: Amivantamab downregulates EGFR and cell cycle pathways and increases A300+150 achieved confirmed PR. The median duration of response and the median
cytotoxic immune cell signatures consistent with immune cell infiltration into tumors. El- progression-free survival in the BRAFi-naı̈ve pts were 5.6 and 5.4 months, respectively.
evated AREG and EREG ligand expression correlated with response in wild-type refractory Conclusions: The combination of A+E+C was well tolerated at the MTD and yielded re-
mCRC. Clinical trial information: NCT05379595. Research Sponsor: Janssen Research & sponse rates in BRAFi-naı̈ve mCRC pts which exceeded the historical data from the E+C
Development, LLC, a Johnson & Johnson company. doublet. Clinical trial information: NCT05743036. Research Sponsor: Zentalis.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
248s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3552 Poster Session 3553 Poster Session

Low-pass whole methylome sequencing–based liquid biopsy for metastatic Vilastobart (XTX101), a tumor-activated, Fc-enhanced anti–CTLA-4 mono-
colorectal cancer monitoring in the VALENTINO trial. First Author: Paolo Manca, clonal antibody, in combination with atezolizumab in patients with MSS
Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy CRC. First Author: Marwan Fakih, City of Hope Comprehensive Cancer Center, Duarte,
Background: The amount of ctDNA is a proxy for metastatic colorectal cancer (mCRC) CA
disease burden, with potentials for prognostic stratification and treatment monitoring. We Background: Vilastobart (XTX101) is tumor-activated, high affinity, Fc-enhanced aCTLA-4 designed to
investigated two methods based on low-pass whole genome and methylome sequencing focus activity toward the tumor and minimize systemic adverse events. Fc-enhancement augments FcgR
(WGS and WMS) for ctDNA detection and quantification in the VALENTINO trial. co-engagement on antigen presenting cells and has been linked with efficacy of aCTLA-4 combinations in
patients (pts) with microsatellite stable (MSS) colorectal cancer (CRC) and other tumors [Fakih ASCO GI
Methods: All patients from the VALENTINO trial - a phase II trial comparing the addition of
2025]. Vilastobart was generally well-tolerated and demonstrated evidence of anti-tumor activity in pts
5FU to Panitumamb-based maintenance after first line FOLFOX+Panitumamb induction in with immunologically “cold” advanced solid tumors, both as a monotherapy and in combination with
RAS wild-type mCRC - were eligible. Baseline (BL) and 8-week (8w) plasma samples were atezolizumab [Davar SITC 2024]. Methods: Phase 2 of the NCT04896697 study evaluated the initial
collected for low-pass WGS and WMS analysis. Two methods for ctDNA quantification based recommended Phase 2 dose of vilastobart 100 mg Q6W in combination with atezolizumab 1200 mg Q3W
on DNA methylation (METER) and copy number alterations (ichorCNA) were assayed. Chi- in pts with MSS CRC who had at least 1 prior chemotherapy regimen in the metastatic setting, excluding
squared, Wilcoxon and Cox regression tests were used. Performances of WMS and variant pts with prior immune checkpoint inhibitors. Pts with (LM) and without liver metastasis (NLM) were
allele fraction (VAF) of a 14-gene panel were compared. Results: A BL liquid biopsy was eligible. Tumor biopsies were obtained before and during treatment for translational analyses.
available for 154 patients, with 142 also having an 8w assessment. METER and ichorCNA Results: As of January 13, 2025, 40 pts were dosed in Phase 2. Median age was 55 (25-82) and 70% of pts
had 3 or more prior lines of therapy. Of the 24 enrolled NLM pts, 11 were response-evaluable with an
detected ctDNA in 112 (72.7%) and 94 (59.7%) BL samples, respectively; all discordant cases
available on-treatment scan as of the data cut. In these 11 pts, two confirmed and one unconfirmed partial
were METER+ but ichorCNA-. Detection rate increased in the presence of liver metastases response (PR) were reported, all accompanied by significant decreases in ctDNA and serum tumor marker
(86.0% vs 42.6% for METER, 75.7% vs 23.4% for ichorCNA; both p , 0.001) and decreased CEA and with each pt ongoing on therapy, for a preliminary ORR of 27%. One additional pt (with peritoneal
with peritoneal metastases (55.6% vs 78.0% for METER, 38.9% vs 66.1% for ichorCNA; p = metastasis) had a 24% reduction in target lesions (first scan) and was ongoing on therapy. Of the 16
0.011, p = 0.006). Tumor fraction (TF) of both BL METER and ichorCNA correlated with the enrolled LM pts, 7 were response evaluable, with one stable disease and another pt reporting a mixed
diameter of measurable lesions (both p , 0.001) and CEA (p , 0.001 and p = 0.010). Both response with significant serum tumor marker reductions, both ongoing on therapy. Of note, one LM pt in
PFS and OS were shorter after baseline ctDNA detection with METER (mPFS: 10.6 vs Phase 1C dose escalation treated with vilastobart (150 mg Q6W) combination had a confirmed PR with
18.6 months, HR: 1.65, p = 0.010; mOS: 28.7 vs 62.2 months; HR: 2.24, 95%CI: 1.37-3.66; p = LM resolution. Six pts (15%) reported G3+ treatment-related adverse events (TRAEs), with two G4
0.001) or ichorCNA (mPFS: 10.6 vs 15.0 months, HR: 1.42, 95%CI: 1.00-2.00, p = 0.047; mOS: laboratory TRAEs and no G5 TRAEs. Only three pts discontinued therapy for TRAEs. TRAEs occurring
in $10% (all grade) or $5% (G3) of pts are summarized in the Table. Conclusions: The combination of
27.8 vs 48.4 months, HR: 1.35, 95%CI: 1.29-2.95; p = 0.002). In the multivariate analysis, vilastobart, a novel tumor-activated, Fc-enhanced a-CTLA-4, and atezolizumab demonstrated initial
METER ctDNA detection was the strongest predictor of both PFS and OS (p = 0.005 and p = evidence of anti-tumor activity in late line, metastatic MSS CRC where immune checkpoint blockade has
0.001) while ichorCNA ctDNA detection was significantly associated with OS but not with historically been relatively ineffective. Vilastobart was observed to have a differentiated safety profile
PFS (p = 0.002 and p = 0.093). METER ctDNA TF decreased significantly at 8w in patients distinct from systemically active a-CTLA-4, consistent with tumor-selective activation. Clinical trial
with CR, PR, or SD (paired Wilcoxon p = 0.015, p , 0.001, p , 0.001) but not PD (p = 0.560) information: NCT04896697. Research Sponsor: Xilio Therapeutics.
as the best radiological response. Patients without METER ctDNA clearance at 8w had a
All Grade Grade 3
higher risks of progression (HR: 2.70, 95%CI: 1.63-4.49; p , 0.001) and death (HR: 3.37, 95% AE term n (%) n (%)
CI: 2.00-5.69; p , 0.001). Among 123 patients with both METER and VAF available, Fatigue 12 (30%) 0
concordance was 78.0% and in 10 and 17 patients, respectively, ctDNA was detected only Diarrhea 8 (20%) 0
with METER or only with VAF. The mPFS and mOS of discordant cases were longer than Infusion related reactions 5 (13%) 0
Pyrexia 4 (10%) 0
METER+ / VAF+ cases and shorter than METER2 / VAF2 cases. Conclusions: CtDNA ALT increased 4 (10%) 0
quantification with low-pass WMS by METER retains a prognostic significance, can be used AST increased 4 (10%) 1 (3%)
for disease monitoring during treatment and refines ctDNA detection based on a restricted Colitis 2 (5%) 2 (5%)
gene panel assay. Clinical trial information: NCT02476045. Research Sponsor: None.

3554 Poster Session LBA3555 Poster Session

mFOLFIRINOX efficacy as rescue regimen for patients with metastatic Safety and efficacy of reduced-port laparoscopic surgery for patients with
refractory colorectal cancer: The RE-PLAY trial. First Author: Paulo Marcelo colon and upper rectal cancer. First Author: Jun Huang, Department of Colorectal
Hoff, Instituto do Câncer do Estado de S~ao Paulo, University of S~
ao Paulo and Instituto Surgery, The Sixth Affiliated Hospital, Sun Yat-sen University, Guangzhou, China
D’Or de Pesquisa e Ensino, S~ ao Paulo, Brazil
Background: Doublets with Fluoropyrimidines (F), oxaliplatin (Ox), and irinotecan (Ir)
are standard chemotherapy agents used to treat metastatic colorectal cancer (mCCR).
The role of triplet combination with modified FOLFIRINOX (mFOLFIRINOX) in refractory
patients (pts) previously treated with doublets or monotherapy sequential regimens
remains unclear. Methods: This single-arm, open-label phase II trial employed a Simon
two-stage design. Eligible pts had mCCR with documented progression after treatment
with doublets or sequential monotherapy regimens containing F, Ox, and Ir. Pts with RAS
wild-type tumors were required to be refractory to anti-EGFR therapy. The mFOLFIRINOX
regimen consisted of 5-FU (2400 mg/m², continuous infusion over 46 hours), Ox (85 mg/
m², D1), Ir (150 mg/m², D1), and leucovorin (200 mg/m², D1), administered every 14 days.
The primary endpoint was the disease control rate (DCR) as assessed by RECIST v1.1.
According to the Simon design, the study would be considered positive if 4 or more pts
achieve disease control among 25 pts in the second stage. Secondary endpoints in- The full, final text of this abstract will be available at
cluded objective response rate (ORR), progression-free survival (PFS), overall survival [Link] on the day of presentation and in the
(OS), and safety. Results: Between October 2021 and October 2024, 25 pts were online supplement to the June 10, 2025, issue of the Journal
enrolled. Three pts did not receive treatment due to consent withdrawal (n = 1) or clinical
deterioration from disease progression before treatment initiation (n = 2). All pts had of Clinical Oncology.
proficient mismatch repair (pMMR) tumors; 16 (64%) had KRAS/NRAS mutations, and
most tumors were left-sided (n = 20, 80%). At Intention to Treat analyses, 16.6% (n = 4)
achieved a partial response, 44% (n = 11) had stable disease, and 28% (n = 7) experienced
disease progression as their best radiologic response. The DCR was 60% (n = 15), and the
ORR was 16.6% (n = 4). With a median follow-up of 6.8 months, 17 pts experienced
disease progression or death. The median PFS was 5.7 months, and the median OS was
9.3 months. No significant differences in DCR, ORR, PFS, or OS were observed based on
RAS mutation status or tumor sidedness; among 22 pts who received at least one cycle
of mFOLFIRINOX, 68.1% (n = 15) experienced grade 3 or higher adverse events, including
one treatment-related death. Conclusions: mFOLFIRINOX demonstrated efficacy as a
rescue regimen for refractory mCCR previously refractory to doublets or sequential
monotherapy regimens containing Ox, Ir, fluoropyrimidines, and anti-EGFR therapy.
Clinical trial information: NCT05354817. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 249s

3556 Poster Session 3557 Poster Session

Analysis of mutational profiles and their correlation with organ-specific Patterns of immunotherapy use in dMMR/MSI-H metastatic CRC. First Au-
metastases in MSS and BRAFwt colorectal cancer (mCRC). First Author: thor: Hayley Lemisch, Temple University Lewis Katz School of Medicine, Philadelphia, PA
Francesc Salva, Vall d’Hebron Institute of Oncology (VHIO), Vall d’Hebron University Background: Immune checkpoint inhibitors (ICI) are more effective in mismatch repair
Hospital, Barcelona, Spain deficient/microsatellite instability high (dMMR/MSI-H) metastatic colorectal cancers
Background: In BRAFwt/ MSS mCRC, the mechanisms driving distinct metastatic dis- (mCRC) compared to chemotherapy (chemo). However, real-world data and patterns of use
semination patterns remain unclear. Understanding them is crucial, as dissemination are limited. We evaluated real-world ICI use in mCRC after FDA approval in 2017 and
profiles influence therapeutic strategies, such as immunotherapy for patients (pts) without approval as first-line (1L) therapy in 2020, and the impact on survival. Methods: We used
liver metastasis (mets) or locoregional approaches and liver transplantation for those with the nationwide Flatiron Health electronic health record (EHR) derived de-identified da-
liver-limited disease. NGS advances provide genomic data, offering opportunities to tabase to determine patterns of ICI usage in patients diagnosed with dMMR/MSI-H mCRC
identify predictive signatures that link molecular profiles to metastatic patterns. This study since 2013. Trends in ICI use were estimated with the Cochran-Armitage test, while real-
investigates mutational profiles to uncover correlations with organ-specific mets in pts world time to next treatment (rwTTNT), progression free survival (rwPFS), and overall
treated at our institution. Methods: This study included pts with unresectable MSS/ survival (OS) between ICI vs chemo only groups were estimated with Kaplan Meier curves
BRAFwt mCRC treated at Vall d’Hebron Hospital (2010–2020). Pts were grouped into three and log-rank test. Multivariate Cox Proportional-Hazards Models were fitted to adjust for
clinical categories based on metastatic patterns: liver-limited disease (LLD), exclusively potential cofounding variables (ie sex, performance status (ECOG) and treatment). Hazard
extrahepatic disease (EXTRAHEP), and hepatic and extrahepatic disease (BOTH). Mo- ratios, p-values and 95% confidence intervals are presented. Proportional hazards as-
lecular analyses were performed using NGS prescreening data available at our institution. sumptions were tested for violations. Results: Of 41,431 patients diagnosed with mCRC
Mutations were grouped in two approches: (1) by biological significance using cancer since 2013, 1,707 were dMMR/MSI-H and received therapy; thus were included in the
hallmark genes from published datasets (Zhang, Front Genet 2020; Sondka, Nat Rev Cancer analyses. Mean age was 66 years, 925 (54%) were female, and 884 (52%) had de novo
2018), and (2) by molecular pathways based on the Sanchez-Vega dataset (Cell 2018). mCRC. BRAF and RAS mutations were detected in 604 (35%) and 372 (22%) patients,
Statistical analyses were conducted using R version 4.3.2. Results: A total of 1,026 pts respectively. Of 1,707 eligible patients, 573 (34%) received 1L ICI and 1,116 (66%) received
were included (204 LLD, 297 EXTRAHEP, and 525 BOTH), with molecular analyses per- 1L chemo. Of those who received 1L ICI, 480 (43%) received a single ICI and 56 (5%)
formed on 360 samples (35% overall; 31.8%, 39.7%, and 33.7%, in each group, respec- received dual ICI. There was a linear increase in the proportion of ICI-based treatments
tively). The median number of genes with pathogenic mutations per sample differed used over time in both the 1L and second line (2L). With a median follow-up of 38.2 months
significantly between groups: 2.28 in LLD, 2.44 in EXTRAHEP, and 2.65 in BOTH (p = 0.01). (mo); median OS for patients who received 1L chemo was 25.9 mo vs 51.6 mo with 1L ICI
BOTH showed significantly greater increase than LLD in mutated genes associated with (HR 0.62, p , 0.0001, 95% CI 0.52-0.73). Presence of a BRAF mutation was associated
five of the ten analyzed hallmarks: activation of invasion and mets, resistance to cell death, with worse OS (median OS 44.4 mo, 95% CI 31.2-61.1) vs. no BRAF mutation (median OS
evasion of growth suppressors, sustaining proliferative signaling, and replicative im- not reached (NR), 95% CI 51.6-NR) in patients receiving 1L ICI (p = 0.0046). The presence
mortality. Compared to EXTRAHEP, BOTH also had more mutations in invasion and mets of a RAS mutation was not significantly associated with OS. Median rwTTNT after 1L ICI
activation and proliferative signaling (adjusted p-value , 0.05 for all the hallmarks was 31.8 mo (95% CI 23.7-50.2), compared to patients whose first ICI was in the 2L (21.9
mentioned). For pathway associations, WNT pathway activation was higher in BOTH than mo [95% CI 12.9-36.8]) or third line (3L) (9.5 mo [95% CI 5.7-18.0]) (p = 0.0041). Median
EXTRAHEP (p = 0.004), driven by more frequent APC mutations in BOTH (82% vs. 69%, adj. rwPFS for first receipt of ICI in the 1L was 18.1 mo (95% CI 13.2-30.1) compared to first
p = 0.049). Conclusions: This study provides evidence that pts with both hepatic and receipt in 2L (8.3 mo [95% CI 6.5-14.5]) or 3L (4.8 mo [95% CI 4.0-13.7]) (p = 0.0032).
extrahepatic disease exhibit enrichment in five cancer hallmarks and the WNT pathway Conclusions: There was a linear increase in the proportion of patients with dMMR/MSI-H
compared to other metastatic patterns. This suggests the tumor’s potential to adapt to mCRC treated with ICI, associated with FDA approvals in 2017 and 2020. Male gender and
diverse microenvironments. Despite the statistical significance, the magnitude of the presence of a BRAF mutation in patients receiving 1L ICI and receipt of 1L chemo were
observed differences in mutated genes is not yet clinically useful. These findings highlight inversely associated with OS. Receipt of ICI in earlier treatment lines was associated with
the need for collaborative efforts to develop mutational profiles that predict organotropism increased rwPFS and rwTTNT. Patients with dMMR/MSI-H mCRC benefit from receiving
and guide therapy. Research Sponsor: None. early ICI. Research Sponsor: None.

3558 Poster Session 3559 Poster Session

Colorectal cancer with gene fusions: Navigating the genomic landscape and The efficacy of watch and wait strategy or surgery after neoadjuvant im-
treatment selection in a phase I unit. First Author: Camila Braganca Xavier, munotherapy for locally advanced colorectal cancer with dMMR/MSI-H
Department of Investigational Cancer Therapeutics, The University of Texas MD guided by ctDNA dynamic monitoring (WINDOW): A single-center, open-
Anderson Cancer Center, Houston, TX label, prospective, phase II study. First Author: Xuan Zhang, Yunnan Cancer
Background: Gene fusions are rare genomic events in colorectal cancer (CRC) and can co- Hospital, Kunming, Yunnan, China
occur with other potentially actionable alterations. Navigating early-phase trial selection in Background: Circulating tumor DNA (ctDNA) has shown potential in predicting the efficacy
this scenario can be challenging. Objectives: To evaluate the genomic landscape, treatment of neoadjuvant treatment for colorectal cancer (CRC). However, evidence is limited for
selection, and clinical outcomes of patients (pts) with CRC harboring gene fusions enrolled patients with deficient mismatch repair (dMMR) / microsatellite instability-high (MSI-H) CRC,
in at least one clinical trial in the Department of Investigational Cancer Therapeutics, The who respond well to immunotherapy. This study explores an optimal ctDNA-guided neo-
University of Texas, MD Anderson Cancer Center. Methods: We used a computerized data adjuvant immunotherapy (nIT) strategy in locally advanced CRC (LACRC) patients with
extraction tool to review clinical and genomic data of pts with CRC harboring gene fusions dMMR/MSI-H. Methods: We conducted a single-center, open-label, phase 2 trial named
between June 2011 and June 2024. The actionability of the genomic alterations was WINDOW involving dMMR/MSI-H LACRC patients. Patients received the anti-PD-1 antibody
classified by the Precision Oncology Decision Support team. Median overall survival (mOS) tislelizumab every three weeks. ctDNA was monitored at baseline, after 2, 4, 5, or even 6 to 8
was defined as the time from consent to the date of death or last follow up. Kaplan-Meier cycles, and during post-watch-and-wait (W&W) or post-surgery periods using the Signatera
method was used to estimate survival and the Cox proportional hazards model to evaluate platform. Starting from the 4th cycle, patients with two consecutive ctDNA-negative results
the impact of multiple variables. Results: 56 pts were included (28 female; 28 male) with a were eligible for the W&W approach. Those who did not achieve ctDNA-negative or turned
median age of 50.5 years (range 25-84). Colon was the most frequent tumor location (n=29, positive during follow-up continued immunotherapy. Surgery was performed for patients not
51.8%) followed by rectal (n=15, 26.8%), small bowel (n=8, 14.3%), and appendiceal (n=3, meeting W&W criteria by the 8th cycle. The primary endpoint was the complete response (CR)
5.3%). Most cases had adenocarcinoma histology (n=54, 96.4%). The median time from rate, including ctDNA-negative clinical complete response (cCR) and pathological complete
diagnosis to completion of the molecular test identifying the gene fusion was 3 mo (range 0- response (pCR). The trial is registered at [Link] (NCT06477991). Results: From
36) for pts whose tumors were diagnosed in the metastatic setting. A total of 65 fusions January 2023 to May 2024, 24 patients with stage II-III dMMR/MSI-H CRC were enrolled,
were identified (range 1-3 per pt); 48 of them occurring in actionable or potentially ac- including 18 with colon cancer and 6 with rectal cancer. At baseline, all patients had de-
tionable genes (78.8%) and 17 in non-actionable genes (26.2%). RAS alterations were tectable ctDNA. After nIT, 87.5% (21/24) achieved two consecutive ctDNA-negative results.
present in 27 cases (48.2%) and BRAF alterations in 10 cases (17.9%). 8 patients had Among the remaining three, two underwent surgery due to sustained ctDNA-positivity (both
alterations in ERBB2 (14.3%) and 32 in genes related to DNA damage repair (DDR; 57.1%). In TRG3), while one showed continuous ctDNA decrease despite obstruction and achieved pCR.
total, patients were enrolled in 91 trials (range 1-7 per pt). Target therapy (TT) was offered to One patient required emergency surgery for perforation (confirmed as pCR), while 20 were
33 pts (58.9%) while the remaining were offered non-TT (n=23, 41.1%). Treatment selection managed with the W&W strategy. Of these, 90% (18/20) achieved ctDNA negativity after two
was fusion-driven in 10 cases (17.9%; 4 NTRK, 3 FGFR, 1 RET, 1 ROS1, and 1 ALK). Notably, cycles, and 95% (19/20) after five cycles of nIT. With a median follow-up of 20.3 months
one pt with a ETV6-NTRK fusion was enrolled sequentially in 4 immunotherapy (IO) and TT (range: 8.7–25.0 months), none experienced recurrence, resulting in an overall CR rate of
trials, surviving on-trial for over 3.5 years. 3 patients received TT for KRAS (5.4%), 4 for BRAF 91.7% (22/24). Notably, if ctDNA remained positive after the 5th cycle, the CR rate was only
(7.1%), 4 for ERBB2 (7.1%), and 8 for DDR alterations (14.3%). The mOS was 10.0 mo (95% CI, 25% (1/4), while it reached 100% (20/20) if ctDNA became negative. From the perspective of
5-NR) for pts receiving non-TT and 11 mo (95% CI, 3-21) for patients receiving TT (p 0.985). organ preservation, only 45.5% (10/22) of patients avoided surgery based on imaging alone;
Age group (cutoff 50 years), gender, cancer stage at diagnosis, RAS, BRAF and DDR al- however, with ctDNA-guided management, 83.3% (20/24) avoided surgery. Conclusions: NIT
terations, and treatment type (TT vs. non-TT or IO) were not related to survival in the demonstrates high efficacy in dMMR/MSI-H LACRC. The ctDNA-guided W&W strategy
multivariate level (p=0.9). Conclusions: A comprehensive genomic evaluation of pts with significantly improves organ preservation rates. Monitoring ctDNA negativity after the 5th
CRC harboring gene fusions can expand early-phase treatment possibilities. DDR alterations cycle of nIT is a crucial marker of high CR rates, suggesting this cycle may represent the
are more frequent in this group than in unselected CRC cohorts and can represent an optimal monitoring window during nIT. Clinical trial information: NCT06477991. Research
additional target for TT. Research Sponsor: None. Sponsor: The Joint Special Funds for the Department of Science and Technology of Yunnan
Province-Kunming Medical University; 202201AY070001-149.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
250s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3560 Poster Session 3561 Poster Session

Laparoscopic versus open liver resection for colorectal liver metastasis: An Prognostic factors of survival and recurrence after liver transplantation for
individual participant data meta-analysis of randomized controlled trials. unresectable colorectal liver metastases: Results from the TransMet trial.
First Author: Åsmund Avdem Fretland, Oslo University Hospital, Oslo, Norway First Author: Rene Adam, Paul Brousse Hospital, Villejuif, France
Background: Laparoscopic surgery for colorectal liver metastases (CRLM) is associated Background: Liver transplantation (LT) has recently proved to improve the survival of
with lower physical impact, shorter length of stay and less postoperative morbidity than selected patients with unresectable colorectal liver metastases (uCRLM) compared to
open surgery. To analyze oncological and procedural outcomes, a European consortium chemotherapy (C) alone. However, recurrence rates remain high, stressing the need for a
including principal investigators of all 4 completed RCTs on laparoscopic (LLR) vs open better patient selection. This exploratory study aimed to identify prognostic factors as-
(OLR) liver resection performed an individual participant data meta-analysis (IPDMA), sociated with recurrence and death in patients undergoing LT as part of the TransMet trial.
including updated survival data. Methods: This was an IPDMA with a primary endpoint Methods: Data from 36 patients of the LT+C arm (per protocol population) were analyzed
of postoperative morbidity. Secondary endpoints included overall survival (OS), disease- including age, gender, TNM and RAS status of the primary tumor, characteristics of
free survival (DFS) and resection margin status. A generalized linear mixed model was metastases at diagnosis and at LT, chemotherapy regimen, tumor response (RECIST), and
used to compare OLR and LLR , with trial as a fixed-effect. Logistic regression analysis timeframe from primary resection to LT. Associations with recurrence and death were
was performed for dichotomous variables and negative binominal regression analysis explored. Variables with . 5 observations per group and p-values # 0.10 in univariable
for continuous variables. Survival analyses were performed using Cox-regression. analysis were included in multivariable models. Results: Among the 36 transplanted
Results: A total of 761 patients with CRLM were randomly allocated to LLR (n = patients, 27 experienced recurrence and 9 died after 50-month follow-up. Recurrence: At
384) or OLR (n = 377). Preoperative chemotherapy was administered as recommended univariable analysis two factors were associated to a higher risk: serum CEA levels . 5 ng/
by the local multidisciplinary team (40% vs 46%). Whilst LLR was associated with ml at time of LT (11/11 vs 11/18, p 0.01) and oxaliplatin-based first line chemotherapy (14/
significantly less postoperative morbidity overall (19% vs 27%, adjusted OR 0.62 [95%CI 16 vs 13/20, p 0.04). Two other factors showed a trend toward statistical significance:
0.44 to 0.88]), this was not observed in the subgroup of patients receiving neoadjuvant Female sex (14/15 vs 13/21, p 0.10) and . 20 metastases at diagnosis (11/17 vs 16/19, p
chemotherapy (25% vs 24%, adjusted OR 1.06 [95%CI 0.63 to 1.79]), p-value for 0.09). At multivariable analysis, CEA levels at LT . 5 ng/ml (HR: 2.91; 95% CI: 1.0–8.2; p
interaction , 0.001). Hospital stay was shorter after LLR (median 4 vs 5 days, adjusted 0.04) emerged as an independent predictor of recurrence. Female sex (HR: 2.2; 95% CI:
0.8–5.3; p 0.08) and oxaliplatin-based first line chemotherapy (HR: 2.0; 95% CI: 0.8–4.8; p
percentage difference: -27 [95%CI -37 to -15]) All cause 90-day mortality was not
0.13) were also associated with around 2-fold higher risk of recurrence, although not
significantly different (1.9% vs 0.8%, adjusted OR: 3.2 [95%CI 0.65 to 16.00]). At 5-year
reaching statistical significance. Death: At univariable analysis,two factorswere signifi-
follow-up OS and DFS were not significantly different (adjusted HR 0.97 [95%CI 0.79 to
cantly associated with a higher risk: female sex (7/15 vs 2/21 for male, p 0.03) and . 24
1.20] and 1.05 [95%CI 0.86 to 1.27], respectively). In patients who received preoperative
cycles of chemotherapy before LT (8/19 vs 1/15, p 0.05). Two other factors showed a trend
chemotherapy, a trend towards fewer R0 resections in LLR compared to OLR was noted toward higher mortality: no response to 1st line chemotherapy (6/14 vs 3/22, p 0.06) and
(84% vs 90%, adjusted OR 2.00 [95%CI 0.99 to 4.03], p-value for interaction = 0.053). In stable disease (vs partial response) before LT (8/22 vs 1/14, p 0.08). At multivariable
patients who did not receive preoperative chemotherapy there was no difference in R0 analysis, female sex emerged as independent predictor of death (HR 5.1; 95% CI: 1.0-25.0;
resections (90% vs 86%, adjusted OR 0.76 [95%CI 0.41 to 1.40]). Liver specific recurrence p = 0.04). More than 24 cycles of chemotherapy (HR 7.1; 95% CI: 0.9 – 51.0; p 0.07) and
was not different between LLR and OLR (37% vs 37%, adjusted OR 1.02 [95%CI 0.76 to stable disease (vs partial response) at LT, showed an approximately 7-fold increase in the
1.34), neither was time to adjuvant chemotherapy (45 days vs 49 days, adjusted OR 1.12 risk of mortality, although not reaching statistical significance. Conclusions: Within the
[95%CI 0.91 to 1.38]) nor median number of adjuvant courses (8 vs 8, adjusted per- limits of a reduced sample size, these results suggest that LT should be envisaged early in
centage difference 0.07 [95%CI -0.11 to 0.24]). Conclusions: This IPDMA of 761 patients the history of potential candidates to LT to reduce the number of cycles of chemotherapy.
in 4 RCTs across Europe confirms that laparoscopic resection for CRLM is superior to Both morphological and biological tumor response, initially and at time of LT, are essential.
open resection with regards to short term outcomes, with no differences observed on The notable influence of female sex on post-LT outcome needs to be further explored.
long term oncological outcomes. However, in patients who received preoperative Clinical trial information: NCT02597348. Research Sponsor: None.
chemotherapy, the benefit of the laparoscopic approach is questionable. Research
Sponsor: None.

3562 Poster Session 3563 Poster Session

Impact of inflammatory bowel disease on outcomes of colorectal cancer A multicenter randomized phase II trial assessing the efficacy and safety of
patients undergoing open resection. First Author: Abhinav Malik, Freelance mCapOX plus cetuximab and mFOLFOX6 plus cetuximab as first-line treat-
Physician, New Delhi, India ment for patients with RAS/BRAF wild-type metastatic colorectal cancer:
Background: Colorectal cancer(CRC) is, at present, the fourth most common cause of Primary results of the CAPCET study. First Author: Yuwen Zhou, Colorectal Cancer
cancer in the United States, with more than 140,000 new cases and 52,000 deaths yearly. Center, West China Hospital, Sichuan University, Chengdu, China
Patients with inflammatory bowel disease(IBD) are at increased risk of CRC secondary to Background: This CAPCET randomized phase II trial was designed to assess the efficacy and
the pro-neoplastic effects of chronic mucosal inflammation. Open resection is one of the safety of modified capecitabine and oxaliplatin (mCapOX) plus cetuximab (CET) and modified
treatment modalities among CRC patients. Research evaluating the impact of IBD on the fluorouracil, leucovorin, and oxaliplatin (mFOLFOX6) plus cetuximab (CET) for the first-line
short-term outcomes of open resection is limited. Methods: The National Inpatient treatment of left-side unresectable RAS/BRAF wild-type (wt) metastatic colorectal cancer
sample was utilized to identify open resection procedures among colorectal cancer (mCRC). Methods: CAPCET was an open-label, multicenter, randomized, non-comparative
phase II trial. Patients with unresectable RAS/BRAF wt mCRC were randomly assigned (1:1)
patients through International Classification of Diseases, 10th edition (ICD-10) codes.
to receive up to 12 cycles biweekly mCapOX (capecitabine 1000mg/m2 orally twice daily on Day
Our study contained patients ages 18 and older. We created two groups of patients 1-7 and oxaliplatin 85 mg/m2 iv on Day 1) plus CET (500mg/m2 iv on day 1) (arm A) or biweekly
consisting of cases with and without a history of IBD. Descriptive statistics were mFOLFOX6 (oxaliplatin 85 mg/m2 iv on day 1, leucovorin 400 mg/m2 iv on day 1, fluorouracil
conducted for patient demographics and pre-existing comorbidities. We reported the 400mg/m2 iv bolus on day 1, then fluorouracil 2400 mg/m2 continuous infusion over 46-48h)
adjusted odds ratio(aOR) and their 95% confidence intervals for differences in surgical plus CET (500mg/m2 iv on day 1)(arm B) followed by maintenance(either capecitabine plus CET
and post-surgical complications. Results: This study investigated 318115 open re- or capecitabine alone at the discretion of the investigators) or treatment-free intervals until
sections among colorectal cancer patients, which involved 3675(1.2%) with IBD. The progression on treatment, toxicity, or death. The primary endpoint was progression-free survival
cases of IBD contained a younger sample with a mean age of 60.95 years, while our non- (PFS) rate at 9 months from randomization. Results: Between September 2021 and April 2024,
IBD group had a mean age of 66.95 years(p , 0.01). In addition, IBD patients expressed a 168 patients (84 in arm A and 84 in arm B) were enrolled in 20 China centres. Baseline
lower mean Charlson Comorbidity Index(CCI) score of 4.21(vs. 4.34, p , 0.01). Our IBD characteristics were well balanced between arms. After a median follow-up of 21.0 months
group presented with higher aOR of severe sepsis with septic shock (1.224, 95% CI (IQR,19.5-22.5), the 9 months-PFS rates were 70.9% (95% CI 61.1%-82.3%) in arm A and 66.8%
1.030-1.455, p = 0.022), acute kidney injury(AKI)(1.172, 95% CI 1.059-1.297, p , 0.01), (95% CI 56.7%-78.6%; HR = 1.11, P = 0.558) in arm B, and the primary endpoint was met. The
cardiopulmonary resuscitation (CPR)(2.148, 95% CI 1.437-3.212, p , 0.01), and wound median PFS (arm A/B) was 12.7 months (95% CI 10.8–15.2)/12.0 months (95% CI 9.7–14.1). The
overall response rate (ORR) and disease control rate (DCR) in arm A were higher than those in
infection(1.377, 95% CI 1.139-1.664, p , 0.01). On the other hand, IBD was linked with
arm B with 69.2% versus 60.3% and 96.2% versus 89.7%, respectively. The 2-year overall survival
lower odds of bleeding(aOR 0.808, 95% CI 0.686-0.952, p = 0.011) but was more likely to
(OS) rate (arm A/B) was 66.8% (95% CI 54.2%–82.3%)/65.6% (95% CI 52.5%–82.0%), and the
undergo blood transfusion(aOR 1.153, 95% CI 1.036-1.284, p , 0.01). However, events median OS not reached. Grade$3 adverse events (AEs) occurred in 28.8% of safety population
of acute respiratory failure (aOR 1.024, 95% CI 0.877-1.195, p = 0.767), postoperative set (n = 156), with 7.7% in arm A and 21.2% in arm B. The most commonly Grade$3 AEs was
pneumothorax(aOR 1.286, 95% CI 0.684-2.417, p = 0.436), use of mechanical ven- neutropenia, rash, leukopenia and there were no grade 5 AEs reported. Conclusions: The
tilation(aOR 1.181, 95% CI 0.996-1.400, p = 0.056), vasopressors (aOR 1.012, 95% CI CAPCET study met its first endpoint of 9-month PFS rate in patients with RAS/BRAF wt mCRC.
0.780-1.315, p = 0.926), and all-cause mortality (aOR 1.026, 95% CI 0.808-1.304, p = Biweekly mCapOX plus CET had higher ORR and DCR than mFOLFOX6 plus CET, with signally
0.831) did not differ. Conclusions: Among CRC patients undergoing open resection, reduced toxicity. Longer follows-up and a multicenter, open-label, randomized, controlled Phase
patients with IBD were younger and had higher odds of septic shock, AKI, CPR, wound CAPCET- III study (NCT06616259) will be further validate this innovative regimen. Clinical trial
infection, and blood transfusion. Although we failed to find differences in short-term information: NCT05022030. Research Sponsor: The Department of Science and Technology of
mortality, it is crucial to conduct long-term studies to evaluate responses and relapses Sichuan Province; The Postdoctor Research Fund of West China Hospital, Sichuan University;
following open resection and post-discharge complications. Research Sponsor: None. The Nation-Sponsored Postdoctor Researcher Program; The Science and Technology Inno-
vation 2030-Major Project (Young Talent Cultivation Program); The 1.3.5 Project for Disciplines
of Excellence, West China Hospital, Sichuan University.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 251s

3564 Poster Session 3566 Poster Session

A medically tailored meal (MTM) delivery program to reduce nutritional Relapse patterns, risk factors, re-resectability and survival after curative
decline and improve treatment tolerance in patients with colorectal cancer treatment for metastatic colorectal cancer (mCRC): A RAXO substudy. First
(CRC): A pilot study. First Author: Kara Stromberg, Fox Chase Cancer Center, Author: Pia J. Osterlund, Tampere University Hospital and Tampere Univeristy, Tampere,
Philadelphia, PA Finland
Background: Colorectal cancer (CRC CA) is the 3rd most common CA among US adults. Over Background: In mCRC, metastasectomy and/or local ablative treatment (LAT) cures some. However, over 70%
50% of CRC patients (pts) present with advanced disease, necessitating adjuvant or pal- of the patients relapse, with liver, lung and peritoneum being the most common recurrence sites. Data on
recurrence patterns and prognostic factors are mainly limited to patients with liver metastases treated before
liative chemotherapy. Nearly 50% CRC pts will be diagnosed with malnutrition (MN) during 2008. An update including patients also with multisite metastases treated up-to-date is needed. Methods: In
their CA journey. MN is associated with reduced quality of life, poorer response to che- this study, 323 patients from the RAXO-study (2012–2018) were analyzed, 312 patients had undergone R0–1
motherapy, and reduced survival. The provision of medically tailored meals (MTM) improves metastasectomy and 11 patients had A0–1 LAT. Three patients with postoperative death were excluded. The
access to nutritious foods to support pts’ caloric/nutrient needs and preferences. Remote remaining 320 patients were prospectively followed for at least five years, with radiological imaging and blood
delivery of MTMs could serve as an impactful intervention for pts at highest risk of MN and tests every 3–6 months in the first two years and every 6–12 months thereafter. Median time to recurrence
to improve treatment tolerance. Methods: Primary outcomes were feasibility (enrollment/ (mTTR), and overall survival (mOS) were estimated with Kaplan-Meier from the date of curative surgery/LAT for
metastases known at baseline and compared using log-rank. Hazard ratios (HR) were estimated using Cox
adherence/retention) and acceptability of an MTM program in pts undergoing 5-FU based regression. Results: mTTR was 17 months (CI 95% 13–23 months) with 221 patients (69% of all) relapsing
therapy for CRC; secondary included weight maintenance, MN measured by PG-SGA during follow-up. Totally 135 recurrences (42% of the patients at risk) occurred during the first year. During the
(Patient-Generated Subjective Global Assessment), HEI (Healthy Eating Index) score second, third, fourth and fifth year, recurrence occurred in 48 (26%), 27 (20%), 7 (7%) and 1 (1%) patients at risk,
generated by VioScreen, a validated computerized food frequency tool, QOL measured by respectively. Of patients relapsing during the first year, 65 (60%) were re-resected with numbers for subsequent
FACT-C/FACT-G, and treatment disruption. The MTM program included daily meals delivered years being 18 (67%), 3 (43%), 0 (0%) and 1 (33%), respectively. mTTR, mOS, and re-resectability outcomes for
weekly and nutritional counseling at 4 timepoints. Results: Among 100 eligible CRC pts the most common recurrence sites are presented in the table. Regional lymph node metastasis (HR 1.4), RAS
mutation (1.4), R1 resection (2.3) and treatment with LAT (2.4) were prognostic for recurrence (p , 0.05).
approached, 48 consented and 52 declined, primarily due to lack of interest in MTMs (35/52, Conclusions: Two-thirds of curatively treated mCRC patients relapse within the first four years. Curative-intent
67%). 40/48 (83%) initiated the MTM program and 32/40 (80%) completed it, including 13 re-resection provides good OS for most sites of recurrence and should be considered when possible. Clinical
(41%) Stage II/III and 19 (59%) Stage IV patients. 16/48 (33%) consented pts discontinued trial information: NCT01531621. Research Sponsor: Finska Läkaresällskapet; 2016, 2018, 2019, 2020, 2021,
the study due to non-compliance (n = 4), post-consent refusal (n = 4), dissatisfaction w/ 2022, 2023, 2024, 2025; The Finnish Cancer Foundation; 2019-2020, 2021, 2022-2023, 2025; Relander´s
meals (n = 7) or death (n = 1). Pts had mean age of 56 yrs (range 32-78), with 51% female, foundation; 2020-2022; The Competitive State Research Financing of the Expert Responsibility Area of
Tampere, Helsinki, Turku, Kuopio, Oulu, and Satakunta Hospitals; 2012, 2016, 2017, 2018, 2019, 2020, 2021,
20% African American, and 15% Hispanic. 34% had , HS diploma, and 56% reported income
2022, 2023, 2024, 2025; Tampere University Hospital Fund; Tukisäätiö 2019, 2020, 2023, 2024 and OOO-project
of ,$50,000. Enrollment (44%, goal . 50%, minimum 40%), counseling adherence (52.5%, 2020, 2022; Helsinki University Hospital research fund; 2019, 2020, 2021, 2022, 2023, 2024; Mary and Georg C.
goal . 66%, minimum 50%) and study retention rate (81%, goal . 70%) measured by end-of- Ehrnrooth Foundation; 2023; Liv & Hälsa; 2023; Radiumhemmets fonder; 2022-2023, 2025-2027; Cancerfonden
study survey completion all surpassed a priori unfavorable result threshold (lower feasibility Sweden; 2023-2024; Amgen; Unrestricted grant 2012–2020, 2023, 2024; Eli Lilly and Company; 2012-2017;
limit), but only retention met its target. Acceptability was high: 92% were satisfied with the Merck KGaA; 2012–2020; Roche Oy; 2012–2020; Sanofi; 2012–2017; Servier; 2016–2024.
MTM program and 91% would recommend it. Mean weight was maintained during study mOS 1st Re- mOS re- Non-re- mOS non-
(181.2 lbs start, 181.6 lbs end) and mean PG-SGA score declined (7.5 to 5.4, indicating lower Resected, mTTR, resection, resection, resected, resected, re-resected,
n (%) months months n (%) months n (%) months p
risk of MN at the end of the study). Mean HEI scores decreased slightly (63.3 start, 61.6 end,
R0–1 resected 320 16 92
indicating a slight reduction in dietary quality) throughout the program; however, more than patients
half of pts (56%, n = 14) demonstrated an increase in their HEI scores, suggesting improved Recurrence at any site 221 10 61 116 (52%) 79 105 (48%) 44 , 0.01
(100%)
quality of food consumed. Conclusions: Our MTM delivery program was highly acceptable – Lung 96 (43%) 10 64 45 (20%) 103 51 (23%) 52 , 0.01
– Liver 72 (33%) 8 54 44 (20%) 77 28 (13%) 39 , 0.01
and was associated with weight maintenance and improved PG-SGA scores in CRC pts – Distant lymph nodes 37 (17%) 9 56 5 (2%) 58 32 (14%) 48 0.31
receiving 5-FU based therapy. Further research of MTM delivery programs should consider – Peritoneum 23 (10%) 10 60 11 (5%) 92 12 (5%) 29 , 0.01
– Local recurrence 21 (10%) 10 58 9 (4%) 58 12 (5%) 56 0.31
retention challenges among pts facing advanced GI cancers. Research Sponsor: Manna Pilot – Intrahepatic only 56 (25%) 9 56 41 (19%) 77 15 (7%) 38 , 0.01
– Intrapulmonary only 61 (28%) 11 91 41 (19%) 103 20 (9%) 61 , 0.01
Funding.
TTR, time to recurrence; m, median; OS, overall survival.

3567 Poster Session 3568 Poster Session

Single-slide histology-based deep learning model for mismatch repair de- Time to treatment initiation (TTI) in patients with colorectal cancer and liver
ficiency prediction in colorectal cancer. First Author: Masoud Tafavvoghi, Oslo metastases. First Author: Kamran Steppe, The University of Oklahoma Health Sci-
University Hospital, Oslo, Norway ences Center, Oklahoma City, OK
Background: Mismatch repair deficiency (dMMR) is a critical predictive biomarker for Background: In colorectal cancer (CRC), the liver represents a common site of metastatic
determining eligibility and response to immunotherapy in colorectal cancer (CRC). The disease, and patients with colorectal liver metastases (CRLM) may sometimes still achieve
current gold standards for detecting dMMR include next-generation sequencing (NGS) for surgical resection of hepatic metastases. In oncology, delays initiating cancer treatment
microsatellite instability (MSI) detection and immunohistochemistry (IHC) for mismatch usually correspond with worse outcomes, yet factors associated with time to treatment
repair protein expression. However, discrepancies between these techniques have been initiation (TTI) in patients with CRLM remain unstudied. Methods: We utilized the National
observed, potentially impacting clinical decisions and patient outcomes. To address this, Cancer Database (NCDB) to identify adults ($18 years old) with confirmed CRLM di-
we developed a histology-based deep learning (DL) model to predict MMR status, with a agnosed between 2010-2021. We defined TTI as the number of days between cancer
specific focus on resolving cases of discordance. Methods: Paired hematoxylin and diagnosis and the first non-surgical treatment (radiation, immune, or chemotherapy). We
eosin (H&E) slides from 974 CRC tumors were retrospectively collected from the Dana- excluded cases in which TTI equaled zero or exceeded 120 days. We compared median TTI
Farber Cancer Institute, all of which had NGS Oncopanel and IHC MMR reports available. across sociodemographic (i.e., age, race, income) and clinical (i.e., primary tumor site,
Using NGS-determined MMR status as the training reference, we developed a multi- tumor grade, comorbidities [Charlson-Deyo score]) factors reported in the NCDB using a
instance deep learning model to predict MMR status from single H&E slides. Feature non-parametric Kruskal-Wallis test. Results: We included 15,456 patients with CRLM
extraction employed various pathology foundation models (FMs). The dataset was split diagnosed between 2010-2021 (55.1% of patients age , 60 years, 44.8% female, 76.3%
into training and tuning sets. A hold-out test set (n = 52, 65% dMMR) was curated White, 12.3% Black, 6.3% Hispanic, 3.6% Asian, 0.3% American Indian, 1.2% Other). The
including patients treated with immune checkpoint inhibitors or those with NGS-dMMR/ median (m) TTI for this cohort was 41 days, with TTI values trending down between
IHC-proficient discordance. Results: Among the overall cohort, NGS/IHC concordance subsequent years from 2010 (m = 45) to 2020 (m = 38) until 2021 (m = 41). We found
identified 82 dMMR patients (9%), 881 proficient MMR (pMMR) patients (90%), and 11 significant differences (p , 0.0001) in median TTI across sociodemographic and clinical
cases (1%) with NGS-dMMR/IHC-proficient discordance. In the hold-out test set, the fine- factors, as detailed in the following sentences. Advanced age (80+) had longer TTI (m = 53),
tuned CTransPath FM demonstrated the highest performance, achieving an area under compared with young patients (m = 33). For race/ethnicity, Black patients (m = 47) had the
the curve (AUC) of 0.88 (95% CI 0.77–0.98), a positive predictive value of 0.93, and longest TTI, while Asian patients had the shortest (m = 39). Patients earning in the highest
correctly classifying 8 of 11 discordant cases (73%) as dMMR. Comparative FMs, CONCH income quartile had the shortest TTI (m = 38) compared with those earning in the lowest
and UNI, exhibited slightly lower AUCs (0.86 and 0.85, respectively) and lower accuracy in quartile (m = 44). Patients living in ZIP codes with the highest quartile of high school
classifying discordant cases (Table). Conclusions: Our histology-based DL model shows diploma attainment had the lowest TTI (m = 38) compared with those in the lowest quartile
promise as a complementary tool for IHC in predicting dMMR status in CRC. The single- (m = 43). TTI varied significantly by reported geographic location, with the Pacific region
slide approach offers a rapid, robust and cost-effective method to prioritize IHC-proficient having the lowest (m = 37) and the East South Central having the highest TTI (m = 44).
cases for further validation by NGS. Cohort expansion and validation in an external Patients with Charlson-Deyo comorbidity scores of 2 had the highest TTI (m = 48) and
dataset are underway. Research Sponsor: Norwegian Cancer Society. patients with scores of 0 the lowest (m = 41). Right sided primary tumor locations (m = 46)
had longer TTI compared to left sided primary sites (m = 39). Tumor sizes greater than 6 cm
Test set (n=52, 65% dMMR). (m = 45) received treatment later than tumor sizes less than 2 cm (m = 35). Lower grade
Accuracy tumors had longer TTI (m = 34) compared to high grade tumors (m = 28). Conclusions: In
Model AUC (CI,95%) Sensitivity Specificity PPV NPV NGS+/IHC- our study of patients with CRLM in the NCDB, the median TTI was just under 6 weeks. Our
CTransPath 0.88 (0.77-0.98) 0.85 0.89 0.93 0.76 8/11 findings suggest that certain patient groups may be at risk of experiencing delays in care.
CONCH 0.86 (0.76-0.96) 0.88 0.76 0.67 0.93 5/11 These results could help to motivate and inform future efforts to enhance care delivery and
UNI 0.85 (0.74-0.96) 0.78 0.82 0.70 0.87 6/11 access for patients with CRLM. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
252s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3569 Poster Session 3570 Poster Session

Immune checkpoint inhibitor (ICI) reuse after failure of first-line ICI in Comparison of MET genomic alterations (GA) identified in colorectal cancer
patients with metastatic dMMR/MSI gastrointestinal cancers: The INFLATE (CRC) vs gastric cancer (GCA). First Author: Faiza Yasin, Yale School of Medicine,
study. First Author: Léa Mercier, Institut Bergonié, Bordeaux, France New Haven, CT
Background: Immune checkpoint inhibitors (ICIs) are a standard treatment for gastroin- Background: MET signaling promotes tumor progression and therapeutic resistance
testinal (GI) cancers with mismatch repair deficiency (dMMR) or microsatellite instability across many solid tumors through diverse oncogenic signaling pathways. While MET-
(MSI). However, around 50% of patients develop resistance to ICIs, during treatment or after targeted therapies are approved in non-small cell lung cancer with novel agents in
discontinuation. In these patients, the efficacy of reuse an ICI in patients progressing during a clinical trials for tumors with MET exon 14 splice site mutations, MET amplifications, and
previous ICI (rechallenge), or in patients who progressed after discontinuation (reintroduction), MET expression, their role in CRC and GCA is still emerging. We aim to understand the co-
remains unknown. The INFLATE study evaluates the efficacy of ICI reuse in patients who mutation landscape and genomic context of MET alterations across a large cohort of
progressed on or after discontinuing ICI. Methods: This is a multicenter international ret- CRC and GCA cases to identify potential therapeutic targets. Methods: FFPE blocks of
rospective study from the IMMUNODIG cohort, including patients from 34 centers in France,
clinically advanced CRC (50,500 cases) and GCA (9,566 cases) were analyzed by hybrid
the United States, Italy, Belgium and Spain. All patients received ICI for dMMR/MSI GI cancer.
We analyzed patients who had progressed following initial ICI (ICI-1) and subsequently
capture-based comprehensive genomic profiling (CGP) that evaluated all classes of
received a rechallenge or a reintroduction with ICI (ICI-2), either monotherapy (mono-ICI) or genomic alterations (GA). MSI-high status, tumor mutational burden (TMB), genomic
biotherapy (bi-ICI). Results: A total of 77 patients were included, receiving bi-ICI (N = 34) or ancestry, mutational signature, and homologous recombination deficiency signature
mono-ICI (N = 43) during ICI-2. The majority (76%) had a metastatic colorectal cancer. The (HRDsig) were assessed for patients with activating MET GA. PD-L1 expression was
reason for discontinuing ICI-1 was disease progression in 53% of cases, end of treatment in determined by IHC (Dako 22C3 with TPS scoring system). Results were compared using
15%, toxicity in 5% and other reasons in 26%. Patients who discontinued ICI-1 due to pro- the Fisher exact test with the Benjamini-Hochberg adjustment. Results: MET GA
gression received MONO-ICI-2 in 29% of cases and BI-ICI-2 in 71%, whereas those who stopped (METmut) were more frequently identified in GCA than CRC (4.4% vs 1.0%; p , .0001).
for other reasons received MONO-ICI-2 in 86% and BI-ICI-2 in 14% of cases. Efficacy results are Median ages were similar in all groups with MET altered CRC and GCA. Median GA per
shown in Table 1. The ORR and DCR were 26% and 79% with MONO-ICI-2, and 16% and 71% tumor was higher in the METmut cases in both tumor types (p . .0001 for both). MSI-
with BI-ICI-2. Among patients who discontinued ICI-1 due to progression, BI-ICI-2 (N = 29) high status was less frequent in METmut GCA (1.7% vs 5.5%; p = .001) compared with
achieved an ORR and DCR of 8% and 65%, with a median PFS of 5.5 months (95%CI 4.07-10.4). METwt tumors. CRC cases featured higher frequencies of TMB . 10 mutations/Mb in
These outcomes were 17%, 67%, and 3.7 months (95%CI 2.37-NA), respectively, with MONO- both METmut and METwt groups (p , .0001 for all comparisons). Low level PD-L1
ICI-2 (N = 12). For patients who discontinued ICI-1 for reasons other than progression, BI-ICI-2 expression (1-49% TPS) was higher in CRC than GCA, but similar within tumor type
(N = 5) achieved an ORR and DCR of 60% and 100%, with a median PFS not reached (95%CI across METmut and METwt tumors. METmut CRC had lower frequencies of GA in KRAS
11.5-NR), while MONO-ICI-2 (N = 31) achieved 29%, 82%, and 14.2 months (95%CI 11.2–NR), (34.7% vs 48.8%; p , 0.0001) which was also found in GCA (6.6% vs 16.9%; p , .0001).
respectively. Conclusions: This is the first multicenter real-world study evaluating ICI reuse in
CDK6 GA were more frequent in GCA than CRC and also were more frequent in METmut
dMMR/MSI GI cancers. In patients who discontinued ICI for reasons other than progression,
reintroduction of ICI therapy upon progression achieves tumor control in 85% of cases. In
cases in both tumor types (19.6% vs 5.2% in GCA and 10.2% vs 0.6%; p , 0.0001). ERBB2
cases of progression on mono-ICI, Bi-ICI re-challenge achieve tumor control in two-thirds of GA were more frequent in both METmut and METwt GCA (13.5% vs 13.8%; NS) than CRC
cases and might be to consider in some patients. Research Sponsor: None. (9.6% vs 5.2%; p = 0.0004). TP53 GA were more frequent in METmut vs METwt CRC
(84.1% vs 75.9%; p , 0.0001) and GCA (81.8% vs 60.8%; p , .0001). BRAF V600E GA
PFS median OS median were identified in 5.9% METmut CRC and 8.4% METwt CRC (p = .051). BRAF V600E GA
ORR % DCR % (95%CI) (95%CI)
were uncommon in both METmut and METwt GCA (0.2% vs 0.4%; p = 1.0).
Overall N=77 21 76 7.65 months 27.6 months Conclusions: This large-scale analysis reveals distinct genomic profiles of MET-altered
(5.5–11.5) (22.2–55.7)
Discontinued ICI-1 due to 11 68 5.03 months 26.7 months tumors in CRC and GCA, characterized by lower prevalence of KRAS and MSI-high status,
Progression N=41 (3.7–8.05) (21.1–NR) with enrichment for CDK6. These molecular differences suggest distinct biologic
Discontinued ICI-1 for 33 85 14.22 months 27.6 months subsets that could inform patient selection and rational drug combination strategies for
Other Reasons N=36 (11.24–NR) (22.2–NR)
novel MET-targeted therapies. Research Sponsor: None.

3571 Poster Session 3572 Poster Session

The role of gut microbiome composition in the IBD-CRC pathway: A retro- Clinicogenomic landscape and outcomes of metastatic colorectal cancer
spective study of 428 patients. First Author: Sabina Sayeed, Maimonides Medical patients with pathogenic GNAS variants. First Author: Faran Polani, Inova Schar
Center, Brooklyn, NY Cancer Institute, Fairfax, VA
Background: The progression from inflammatory bowel disease (IBD) to colorectal Background: Colorectal cancer (CRC) is a leading cause of cancer-related mortality,
cancer (CRC) is a significant clinical challenge, driven partly by inflammation and with 20-25% of cases presenting with metastatic disease (mCRC). Pathogenic variants in
dysbiosis of the gut microbiome. This study aimed to investigate the relationship between the GNAS gene, which encodes the stimulatory alpha subunit of the G protein complex,
gut microbiome composition, systemic inflammation, and CRC risk in patients with IBD. A are found in 1.5% to 4.8% of CRC cases and are more prevalent in mucinous adeno-
secondary aim was to evaluate the potential protective effects of microbial diversity and carcinomas. Presence of these mutations have been associated with a lower likelihood
specific bacterial taxa in mitigating CRC progression. Methods: This retrospective study of regression following first-line (1L) systemic therapy. However, the specific impact of
analyzed data from 428 patients with a confirmed diagnosis of IBD, of whom 162 (37.9%) GNAS mutations on treatment (tx) outcomes and overall survival in mCRC patients (pts)
were diagnosed with CRC. Stool samples collected within one year of diagnosis were remains under investigation. Methods: We retrospectively analyzed de-identified data
subjected to 16S rRNA gene sequencing to characterize microbial composition and from 5,967 pts with stage IV mCRC treated with 1L oxaliplatin-based chemotherapy in
diversity. Alpha and beta diversity indices were calculated to evaluate microbial richness the Tempus Database. Samples were sequenced with the Tempus xT (648-gene panel) or
and community dissimilarity, respectively. Specific microbial taxa, including Bacteroides xF DNA assay (105 or 523 genes depending on version) and were divided into GNAS wild-
fragilis, Escherichia coli, and Akkermansia muciniphila, were quantified due to their type (GNASwt) and GNAS mutated (GNASmut) groups. Tumor mutational burden (TMB)
established roles in inflammation and carcinogenesis. Inflammatory biomarkers, in- and microsatellite instability (MSI) were calculated. Short variant pathogenic/likely
cluding C-reactive protein (CRP), fecal calprotectin, IL-6, and IL-17 were measured from pathogenic mutations and copy number alterations were analyzed for patients that
serum and stool samples. Clinical data, including demographics, disease duration, underwent xT testing. Real-world (rw) objective response rate (rwORR) was defined as
staging, and treatment history, were obtained from electronic medical records. Multi- the proportion of pts with a documented complete or partial response within 90 days of
variate regression models were used to determine associations between microbial tx start. Rw overall survival (rwOS) was defined as the time from 1L start to death from
profiles, inflammatory markers, and CRC risk, adjusted for confounders such as age, sex, any cause. Hazard ratio (HR) was calculated using a Cox proportional hazards model and
BMI, and immunosuppressive therapy. Results: Patients with CRC demonstrated reduced p-values were calculated using the Wald test. Results: The study included 5,854
gut microbial diversity compared to IBD patients without CRC (Shannon Index: 2.3 6 0.4 GNASwt and 113 GNASmut mCRC pts (prevalence GNASmut = 1.9%). GNASmut pts
vs. 3.1 6 0.5; p , 0.001). Dysbiosis in CRC patients was marked by an overrepresentation exhibited higher prevalence of mucinous adenocarcinoma (26% vs. 2.8%) and peritoneal
of pro-inflammatory bacteria (Escherichia coli and Bacteroides fragilis) and a depletion of metastases (51% vs. 22%, p , 0.001), while liver metastases were more prevalent in the
protective taxa (Akkermansia muciniphila and Faecalibacterium prausnitzii). Pro- GNASwt group compared to the GNASmut group (77% vs 51%, p , 0.001). KRAS,
inflammatory biomarkers, including CRP (14.2 6 3.6 mg/L vs. 6.7 6 2.1 mg/L; p , ARID1A, MSH2/3/6, and ATR were more frequently altered in the GNASmut group, while
0.001) and fecal calprotectin (378.5 6 85.2 mg/g vs. 215.3 6 64.7 mg/g; p , 0.001), were TP53 and APC mutations were more frequent in the GNASwt group. Immunophenotype
elevated in patients with CRC. Regression analyses revealed that reduced microbial markers such as high TMB and high MSI were also more often observed in the GNASwt
diversity (OR: 2.87, 95% CI: 1.92–4.28) and the presence of specific pathogenic taxa pts compared to GNASmut pts (p , 0.001). rwORR was significantly lower in GNASmut
(Bacteroides fragilis: OR: 3.12, 95% CI: 2.03–4.79) were independently associated with pts compared to GNASwt (42% vs. 66%, p = 0.002). Pts with GNASmut had reduced rwOS
increased CRC risk. Conversely, patients with higher relative abundances of Akkermansia compared to pts with GNASwt (HR = 1.31, p = 0.05). Conclusions: mCRC pts with
muciniphila exhibited lower levels of inflammation and reduced CRC risk (OR: 0.52, 95% pathogenic GNAS variants exhibit distinct clinicogenomic features and poorer outcomes
CI: 0.35–0.76), likely due to its role in maintaining mucosal integrity and modulating with first-line oxaliplatin-based chemotherapy compared to GNASwt pts. These findings
immune responses. Conclusions: This study underscores the critical role of gut highlight the need for alternative tx and further research on GNAS as a prognostic
microbiome composition in the IBD-CRC pathway, and highlights the need for biomarker in mCRC. Research Sponsor: Tempus AI.
microbiome-targeted interventions to prevent CRC progression in patients with IBD.
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 253s

3573 Poster Session 3574 Poster Session

Clinicopathologic features of complement activation signatures in colorec- Impact of SARS-CoV-2 mRNA-BNT162b2 vaccination on survival outcomes
tal cancer. First Author: Mir Lim, The University of Texas MD Anderson Cancer Center, in metastatic colorectal cancer patients treated with bevacizumab-based
Houston, TX therapy. First Author: Coskun Yazgan, Ankara University Faculty of Medicine De-
Background: Activation of the complement cascade pathway is associated with pro- partment of Medical Oncology, Ankara, Turkey
oncogenic inflammation and immune-suppressing, myeloid-derived M2 macrophages for Background: Colorectal cancer (CRC) is a leading cause of cancer-related mortality
many solid tumors. Most colorectal cancers (CRC) are microsatellite stable that do not worldwide. Bevacizumab, an anti-vascular endothelial growth factor (VEGF) antibody, is
respond to immunotherapy. The role of complement activation (CA) in the “immune cold” widely used in metastatic CRC (mCRC) treatment to inhibit tumor angiogenesis and
CRC phenotype remains poorly detailed. We sought to identify molecular annotations of CRC modulate the immunosuppressive tumor microenvironment. Recent evidence suggests
subpopulations enriched for CA to guide future therapeutic strategies. Methods: CRC that mRNA-based COVID-19 vaccines, such as SARS-CoV-2 mRNA-BNT162b2, may
tumors from 207 patients with stages II-IV CRC at MDACC underwent bulk RNA sequencing. enhance anti-tumor immune responses. This study aimed to evaluate the impact of
Transcriptomes were analyzed per GSEA “Hallmark Complement” gene set to assign a SARS-CoV-2 mRNA-BNT162b2 vaccination on progression-free survival (PFS) and
normalized enrichment score (NES) for CA to each patient and considered “complement overall survival (OS) in mCRC patients treated with bevacizumab-based therapy.
high” (“CH”; N = 103) or “complement low” (“CL”; N = 104) if the complement NES score was Methods: This retrospective, single-center study included 92 mCRC patients treated
above or below the median. Associations between CA and clinical and pathologic char- with bevacizumab between June 2021 and October 2024. Patients were divided into
acteristics - e.g., demographics, mutation status, and Consensus Molecular Subtype (CMS) - vaccinated (n=50) and unvaccinated (n=42) groups. Baseline demographic, clinical, and
were evaluated by chi-squared analysis. Single cell RNA (scRNA) sequencing was performed
pathological characteristics were collected. PFS and OS were analyzed using Kaplan-
on a separate cohort of CRC primary tumors (N = 85) and liver metastases (N = 60) to
Meier estimates and Cox proportional hazards regression models to identify independent
compare CA among different cell types using a Wilcoxon’s test. To assess for an association
between CA and response to immunotherapy in a previously annotated clinical trial of
predictors of survival. Results: The vaccinated group demonstrated significantly longer
patients with MSS, BRAFV600E metastatic CRC (NCT04017650), we evaluated pretreatment median PFS (8.0 months vs. 5.6 months, p=0.010) and OS (39.4 months vs. 17.8 months,
biopsies by bulk RNA sequencing and compared transcriptomic differences in CA between p=0.014) compared to the unvaccinated group. Multivariate analysis identified SARS-
responders versus non-responders to encorafenib, cetuximab, and nivolumab (E+C+N). CoV-2 mRNA-BNT162b2 vaccination as an independent predictor of improved PFS (HR
Results: CH CRC featured a higher prevalence for MSI-H CRC (21.3% vs 4.2%; p = .005), 0.44, p=0.003) and OS (HR 0.39, p=0.018). Vaccinated patients also had a higher
CMS1 (30.1% vs 6.7%; p , .001), and CMS4 (26.2% vs 7.7%; p , .001) relative to CL CRC. proportion of favorable ECOG PS and a lower prevalence of RAS mutations.
CMS2 was more common among CL CRC (54.8% vs 14.6%, p , .001). CH CRC was as- Conclusions: SARS-CoV-2 mRNA-BNT162b2 vaccination was associated with improved
sociated with BRAFV600E mutations (29.7% vs 9.2% for CL, p = .002) but not with KRAS/NRAS PFS and OS in mCRC patients receiving bevacizumab-based therapy, potentially through
mutations or RAS/BRAFwild-type CRC (p = n.s. for both). On scRNA analysis, CA scores were enhanced anti-tumor immune responses. These findings highlight the potential of
highest in myeloid cells and lowest for B cells (p , .0001). Among patients with MSS, mRNA-based vaccines to modulate the tumor microenvironment and improve outcomes
BRAFV600E CRC, CH signature was associated with non-response to E+C+N (fold-change 3.0 in mCRC. Further prospective studies are needed to confirm these results and explore
relative to responders, p = .046). Conclusions: Association of CH status with MSI-H CRC is a underlying mechanisms. Research Sponsor: None.
novel finding that warrants further study in understanding differential patterns of benefit to
immune checkpoint blockade. CH CRC, associated uniquely with BRAFV600E CRC, was
distributed bimodally across the immune-activated CMS1 and the immune-suppressing
CMS4 CRC, similar to known transcriptomic heterogeneity of BRAFV600E CRC. Our data
suggest high CA, linked to immune-suppressing myeloid cell subpopulations, as a negative
predictive biomarker for response to immunotherapy in MSS BRAFV600E CRC and support
broader study of complement-targeting agents to improve treatment for selected patients
with CRC. Research Sponsor: None.

3575 Poster Session 3576 Poster Session

Effect of race/ethnicity on clinical outcomes for metastatic colorectal can- Understanding and addressing unmet needs in colorectal cancer: Findings
cer (mCRC) patients in phase 1 trials: A dual institution experience. First from the Colorectal Cancer Alliance’s Patient and Survivor survey. First
Author: Hasan Musanna Zaidi, Rutgers Cancer Institute of New Jersey, New Brunswick, Author: Kimberley Lynn Newcomer, Colorectal Cancer Alliance, Washington, DC
NJ Background: Colorectal cancer (CRC) is the third most commonly diagnosed cancer and
Background: Patients with mCRC who progress on standard of care therapies have the second leading cause of cancer-related deaths worldwide. Despite advances in
limited therapeutic options and are associated with poor prognoses. Phase 1 trials are a survival rates, CRC patients and survivors often face unmet needs, including emotional
valuable resource for these patients, offering novel treatment options. However, the support, coping strategies, and access to treatment-related information. Beyond
appropriate time of consideration for phase 1 trials remains unclear. We sought to physical challenges, patients report emotional distress, disruptions to quality of life
analyze the outcomes in a multi-racial cohort of patients with mCRC enrolled in pro- (QoL), and difficulty navigating the healthcare system. Even after remission, survivors
spective phase I clinical trials to describe clinical characteristics and gauge efficacy of frequently struggle with lingering psychosocial and physical impacts. Methods: The
such trials. Methods: We reviewed medical records of patients with mCRC enrolled in Colorectal Cancer Alliance conducted an IRB-approved cross-sectional survey to assess
phase 1 trials at two institutions from 1999 to 2018 and 2021 to 2024. We collected the needs of CRC patients, survivors, and caregivers. The survey included over 150
patient demographics, key clinical characteristics, responses, and deaths. Time on study questions about demographics, diagnosis experiences, QoL, access to care, and
was calculated as time between first dose of study drug and decision to discontinue treatment outcomes. Participants (n = 283) were recruited through the Alliance’s social
study. Overall survival (OS) was calculated as time between the first dose of study drug media, email campaigns, and online communities to ensure diverse representation.
to date of death or last contact date (censored). Outcomes were analysed by Results: The study revealed critical insights into the challenges faced by CRC patients
Mantel–Cox test using Prism GraphPad v 10. Results: There were 283 enrollments on and survivors. The median age group of participants was 46–55. Most respondents
66 phase I trials. Median age (range) was: 59 (29-83) years; non-Hispanic whites (NHW, (74%) reported difficulty finding someone who could understand and relate to their
126; 44.5%), non-Hispanic blacks (NHB, 69; 24.4%), Hispanic (H, 69; 24.4%), Asian (A, 17; experience. Furthermore, 41% noted a reduction in support from others after their
6.0%), and unknown (UNK, 2; 0.7%). Median number of prior therapies was 3 (range 0- treatment ended. Many participants faced ongoing challenges, with 54% reporting
11). ECOG performance status was 0, 1, 2, and unknown, among 62 (21.9%), 181 (63.9%), fatigue and 51% experiencing stress. CRC had a profound impact on several aspects of
11 (3.9%), and 29 (10.2%) patients, respectively. Median number of sites of metastases life: 64% said it negatively affected their career or work life, 58% cited negative effects on
was 3 (range 0-10). Sites of metastases included liver 77.0%, lung 60.1%, lymph nodes their relationship with a spouse or partner, 80% reported a decline in their sex life, and
39.9%, peritoneum 24.4%, bone 20.5%, and brain 1.8%. The primary site of cancer for 65% struggled to participate in social activities. Additionally, 51% noted challenges with
patients on study was colon (84.1%), followed by rectum (11.7%), rectosigmoid (0.7%), dating, and 43% indicated that their cancer journey affected their desire to have children.
and (simply documented as) CRC (3.5%). The median time on study was 1.8 months for While most patients felt informed before treatment, 46% expressed unmet needs for
all patients, with NHW 1.9 months, NHB 1.8, H 1.7, A 1.6, and UNK 7.4; p = 0.72. The OS information on complementary or alternative therapies. These findings highlight the
was 8.6 months among all patients, and 7.9, 7.8, 8.8, 9.4, and 9.5 months for NHW, NHB, broad and far-reaching effects of CRC on patients and survivors, revealing significant
H, A, and UNK, respectively (p = 0.82). Response evaluable patients were n = 236; gaps in emotional, psychosocial, and informational support. Conclusions: CRC patients
including complete response (CR, n = 2, 0.8%), partial response (PR, n = 8, 3.4%), stable and survivors face substantial unmet needs that significantly affect their quality of life
disease (SD, n = 80, 33.9%), and clinical benefit rate (CBR = CR+PR+SD, n = 90) 38.1%. and well-being. The Colorectal Cancer Alliance plans to use these insights to develop
Conclusions: Patients with mCRC enrolled onto phase 1 trials showed CBR of 38.1% and care programs and targeted support initiatives designed to address these gaps. By
OS of 8.6 months, which is comparable to standard third-line therapies that were tailoring resources to the unique needs of patients and survivors, these efforts aim to
available during the time period of this study, thus showing promise for their use in improve outcomes and offer hope for a better future for those affected by CRC. Research
clinical practice. No racial/ethnic variation was observed. There was a non-significant Sponsor: None.
trend towards a lower OS with increase in number of prior lines of therapy. A multivariate
model will be presented. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
254s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3577 Poster Session 3578 Poster Session

Drivers of homologous recombination deficiency (HRD) in metastatic co- Real-world efficacy of trifluridine/tipiracil and bevacizumab combination
lorectal cancer (mCRC). First Author: Paula Romero Lozano, VHIO, Barcelona, Spain according to baseline prognostic factors: The BeTAS study. First Author:
Background: HRD is linked to sensitivity to platinum-based chemotherapy and poly Nieves Martinez Lago, Hospital Clı́nico Universitario e Instituto de Investigación San-
(ADP-ribose) polymerase inhibitors across various tumor types. However, the presence itaria de Santiago de Compostela, Santiago De Compostela, Spain
of HRD, driver events (such as pathogenic mutations in HR genes or other genomic Background: The Sunlight Trial demonstrated that trifluridine-tipiracil (FTD/TPI) and
alterations), and its clinical relevance in mCRC remain underexplored. Methods: We bevacizumab (BEV) significantly improved Overall Survival (OS) and Progression-Free
performed the VHIO-300 test, an ISO15189 accredited custom NGS panel profiling over Survival (PFS) in patients with pretreated metastatic colorectal cancer (mCRC) after two
450 genes (including HR-related genes: ATM, BRCA1, BRCA2, BRIP1, CHEK2 or PALB2) treatment lines. However, the real-world efficacy and influence of baseline prognostic
on 356 Stage IV mCRC patients (corresponding to 247 primary colorectal and 109 to factors are not fully understood. Methods: This retrospective, observational, multi-
metastatic samples) enrolled in the Vall d’Hebron Institute of Oncology’s Molecular center study across 18 Spanish hospitals included mCRC patients treated with FTD/
Prescreening Program from June 2021 to December 2024. All samples had a tumor TPI+BEV in a real-world setting. Prognostic factors were analyzed, including Tabernero’s
cellularity . 40% as per pathologist evaluation. An HR score (sHR) based on genome- subgroups, which categorize patients according to time to diagnosis from first me-
wide copy number alterations (CNA) and loss of heterozygosity (LOH) patterns is tastasis ( , 18 vs. . 18 months), number of metastatic sites ( , 3 vs. . 3), and liver
generated. After cross-validation with Myriad MyChoice, sHR $ 56 was established metastasis (yes vs. no). Patients were grouped into Best (BPC), Good (GPC), and Poor
based on a cohort of ovarian tumors and used to identify HRD in mCRC tumor samples. (PPC) prognostic categories. Results: 398 patients were treated from July 2019 to
Results: HRD prevalence was 3.4% in our mCRC cohort (12/356), but, much higher in December 2024. Median age was 67 years (range 26-92), 65.8% male, and 88.4% had
metastatic lesions, (6.4%, 7/109) than in primary samples (2.03%, 5/247) (p = 0.05). In ECOG PS 0-1. 56.3% had RAS mutations. Liver metastases were present in 75.3%, 27.7%
fact, the median sHR between primary (21) vs. metastases (30) in CRC was significantly had . 3 metastatic sites, and 28.2% had , 18 months from diagnosis of first me-
different (p , 0.01). Regarding HR gene status, HR-mutated samples were not sig- tastasis, resulting in 47.2% of patients categorized as PPC. 67.8% received FTD/
nificantly within the HRD group (p = 0.27) and only 6.5% (2/31) were HRD. Noteworthy, TPI+BEV as third-line treatment. ORR was 6.8%, and DCR was 49.9%. With a median
BRCA2 exhibits a frameshift deletion in a homopolymer stretch, that is a frequent follow-up of 14 months, median PFS was 4.9 months (95% CI, 4.1-5.1) and OS was
hotspot in microsatellite instable (MSI) tumors, but this event was not found to be 10.8 months (95% CI, 9.2-12.4). Neutropenia was the most common toxicity, with 33.1%
associated with HRD. In fact, all HRD tumors (n = 12) were microsatellite stable (MSS). of patients experiencing grade 3-4 neutropenia. OS by ECOG PS 0 vs. 1 vs. 2 was 12.5 vs.
Other frequent alterations in mCRC were studied and BRAF mutations were found to be 11.1 vs. 5.7 months (p , 0.0001). PFS by ECOG PS 0 vs. 1 vs. 2 was 5.6 vs. 4.9 vs.
present in 42% of HRD tumors (p , 0.01). Inversely, HRD was rare in KRAS-mutated 3.5 months (p = 0.102). OS by BPC vs. GPC vs. PPC was 18.3 vs. 12.8 vs. 7.5 months (p ,
samples (0.7%; 1/140) and, in fact, highly correlated with non-HRD status (p = 0.02), 0.0001), and PFS was 7.3 vs. 5.8 vs. 3.7 months (p , 0.0001). OS in patients with grade
especially the G12 mutation (p , 0.01). Interestingly, CNA profiles also revealed a strong 3-4 neutropenia vs. no neutropenia was 17.7 vs. 8.1 months (p , 0.0001), and PFS was
association between the BCL2L1 loci gain and HRD (p , 0.01). Clinically, HRD was not 8.7 vs. 3.9 months (p , 0.0001). Conclusions: Our series confirms the effectiveness of
significantly associated with prognostic value nor clinical benefit to oxaliplatin-based FTD/TPI + BEV in real-world clinical practice, with a median OS of 10.8 months and a
combinations. However, the limited sample size and heterogeneous treatment lines median PFS of 4.9 months. The ECOG performance status, Tabernero subgroups, and the
restricted robust statistical analysis. Conclusions: This study identified a small, yet occurrence of grade 3-4 neutropenia help identify patients who may obtain the max-
significant subset of mCRC that displays HRD. sHR and HRD rates were higher in imum benefit from FTD/TPI + BEV treatment. Interestingly, all subgroups analyzed
metastatic lesions vs primary tumors, indicating HR scarring could be accumulating over showed a greater benefit compared to the outcomes previously reported for FTD/TPI
time in some mCRC patients. No clear association between pathogenic HR gene monotherapy, highlighting the potential of this combination in clinical practice.
mutations and HRD were found, suggesting the involvement of alternative molecular Research Sponsor: None.
mechanisms in this process. Frequent co-occurring events, such as BRAF mutations or
BCL2L1 gains could be drivers in CRC HRD, and shape, eventually, new therapeutic
options for these patients in the metastatic setting. Research Sponsor: None.

3579 Poster Session 3580 Poster Session

Safety and efficacy of ADG126 (an anti-CTLA-4 masking antibody) in com- Real-world treatment patterns and outcomes with trifluridine/tipiracil
bination with pembrolizumab: Updated results of phase 1b/2 study in monotherapy or in combination with bevacizumab in metastatic colorectal
advanced MSS CRC. First Author: Daneng Li, City of Hope National Comprehen- cancer. First Author: Donald A. Richards, Texas Oncology, Tyler, TX
sive Cancer Center, Duarte, CA Background: Trifluridine/tipiracil (FTD-TPI; Lonsurf) is an oral antineoplastic agent
Background: ADG126 is an anti-CTLA-4 IgG1 masked antibody that is preferentially activated in the approved for 3rd-line use in combination with or without bevacizumab (BEV) in metastatic
tumor upon cleavage of masking peptides in the tumor microenvironment. Cleaved ADG126 binds to a colorectal cancer (mCRC). In the Phase III SUNLIGHT trial, the addition of BEV to FTD-TPI
unique epitope on CTLA-4, blocks CTLA-4 function, primes T cells and depletes Tregs. ADG126 in was associated with a significant improvement in overall survival (OS) and progression-
combination with pembrolizumab (Pembro) has been evaluated in a Phase 1b/2 clinical trial
free survival (PFS) compared to FTD-TPI monotherapy. However, data on the use of FTD-
(NCT05405595) and we have reported outcome in 3L MSS CRC patients (Pts) free of liver metastasis
(NLM).1-4 We update results from additional dose expansion (EXP) in Pts of advanced MSS CRC.
TPI in combination with BEV in the real-world community setting are currently limited.
Methods: This is a Phase 1b/2, open-label, multicenter dose escalation and expansion study. Primary Methods: This was a retrospective observational study involving electronic medical
endpoints were safety and tolerability, and early signal of efficacy. Secondary endpoints were PK, ADA, records and (where available) chart reviews from mCRC patients treated by the Texas
ORR, DCR, DOR, PFS and OS. Results: As of Jan.15, 2025, a total of 54 MSS CRC Pts were treated with Oncology community practice from Jan 2020 to Oct 2024. Patients had to have received
ADG126/Pembro (200 mg Q3W) in EXP phase across 3 dose levels of ADG126 (Table 1). 18% Pts FTD-TPI with or without BEV after progressing on a prior line of therapy with oxaliplatin
had $ 3 prior therapies and none had prior IO therapy. There was no Grade 4/5 TRAE, and MTD was not and irinotecan. Variables included patient characteristics, clinical characteristics,
reached. Grade 3 TRAEs were dose-dependent: 38% (5/13), 20% (6/30) and 0% (0/11) for 20 mg/kg treatment patterns and clinical outcomes. OS and time to next treatment or death (TTNTD)
LD1, 10 mg/kg Q3W and 10 mg/kg Q6W cohorts, respectively. The discontinuation rate remains low for were analyzed using the Kaplan-Meier method. Results: In total, 265 patients were
the EXP cohorts (6%). The ORR, CBR, mPFS and 12-mon OS of MSS CRC Pts without liver and
included (166 FTD-TPI + BEV; 99 FTD-TPI monotherapy), with the majority receiving FTD-
peritoneal metastasis (NLPM) are listed in Table 1. ORR increased as a function of ADG126 dose.
Although 10 mg/kg Q6W/Pembro did not yield PR, all 6 EE Pts remain on study (1 on treatment) at 18- TPI as 3rd-line (83%; n = 220) or 4th-line (14%; n = 38) therapy. The population was 59%
mon of follow-up. Correlation between dose level/regimen, ORR, CBR and mPFS between 10 mg/kg male, 66% white, and 35% were $65 years of age. The most common previous 1st- and
Q6W and Q3W cohorts has been observed. mOS is not reached for 10 mg/kg Q3W NLPM after 15.5- 2nd-line treatment for 3rd-line FTD-TPI patients was chemotherapy + an antiangiogenic
mon follow up. Longer term efficacy data from 20 mg/kg LD cohort will be reported. (1st-line, 67%; 2nd-line, 74%), which was similar regardless of current BEV use. Median
Conclusions: Dose-dependent ORR has been observed for ADG126/Pembro IO doublet across duration of therapy was 2.8 months (range 0.3 to 12.5) with FTD-TPI + BEV and 2.8 months
multiple dose levels/regimens of ADG126 (10 mg/kg Q6W to 20 mg/kg LD) that is associated with well- (range 0.1 to 10.4) with monotherapy. Median OS was 11.6 months with FTD-TPI + BEV
tolerated to acceptable safety profile, which is enabled by a relatively large therapeutic window. The and 6.2 months with monotherapy (hazard ratio [HR] = 2.1; 95% confidence interval [CI]:
overall performance of ADG126/Pembro IO doublet warrants further clinical development including 1.5-3.0; p , 0.001). At 6 months, OS probability was 0.69 (95% CI: 0.61-0.77) with FTD-TPI
combination with SOCs targeting earlier lines/broader populations, such as MSS CRC with liver
metastasis. Clinical trial information: NCT05405595. Research Sponsor: Adagene Inc.
+ BEV and 0.50 (95% CI: 0.40-0.63) with monotherapy; 12-month OS probability was 0.49
(0.39-0.61) and 0.15 (0.07-0.28), respectively. Median TTNTD was 9.4 months for FTD-TPI
Key efficacy results from MSS CRC patients.
+ BEV and 5.8 months for FTD-TPI alone (HR = 1.7, 95% CI: 1.2-2.4; p , 0.001). The safety/
ADG126 Dose/Pembrolizumab (200 mg, Q3W) tolerability profile was generally similar irrespective of BEV use, with the most common
10 mg/kg Q6W 10 mg/kg Q3W 20 mg/kg LD1 Total # adverse events being fatigue/asthenia (73%), abdominal discomfort/pain (55%), and
Safety Evaluable 11 30 13 54 nausea (54%). The most notable difference was neutropenia (37% FTD-TPI + BEV, 27%
Efficacy Evaluable (NLPM) 10 (6) 29 (22) 12 (12) 51 (40) monotherapy). Conclusions: In this large real-world community practice setting in the US,
MSS CRC NLPM‡‡>= Objective 0 PR = 23% (5/22)2 PR = 33% (4/12)3 NA
Response Rate (ORR) (CI: 8-45) (CI: 10-65) FTD-TPI use in mCRC was mostly in the 3rd-line setting and approximately two-thirds of
6-mon CBR% 33% 55% (CI: 32-76%) NM NA use was in combination with BEV. Patient characteristics were similar to the SUNLIGHT
mPFS (mon) 5.9 6.7 (CI: 4.6-9.0) NM NA
12-mon OS 100% 75.1 (CI: 50-89%) NM NA
trial, with high rates of previous antiangiogenic use. A statistically significant and
1
clinically relevant OS benefit was seen with the addition of BEV versus monotherapy
20 mg/kg LD: ADG126 20 mg/kg x1 cycle followed by 10 mg/kg Q3W. 2Including 1 unconfirmed PR. 3All confirmed. CI:
95% confidence interval (report for n .=12 Pts cohort). NM: data not mature. NA: not applicable. consistent with the results of the SUNLIGHT trial. Research Sponsor: Taiho Oncology Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 255s

3581 Poster Session 3582 Poster Session

A phase 1 dose-escalation study of GCC19CART: A novel CAR T-cell therapy Impact of obesity and lifestyle-associated risk factors on outcomes of early-
for metastatic colorectal cancer in the United States. First Author: Benjamin L. onset colorectal cancer in patients younger than 50 years old: A propensity-
Schlechter, Dana-Farber Cancer Institute, Boston, MA matched analysis. First Author: Syeda Ashna Fatima Kamal, Saint Louis University,
Background: GCC19CART, the first clinical candidate of the CoupledCAR solid tumor Saint Louis, MO
platform, pairs a solid tumor chimeric antigen receptor (CAR) T-cell with CD19-targeting CAR Background: Early-onset colorectal cancer (CRC) incidence is rising in individuals under 50
T-cells. The CD19 target enhances proliferation and persistence of the CoupledCAR, years old. This study aims to understand the factors influencing outcomes in this population.
overcoming the limitations seen in other solid tumor CAR T-cells. Guanylate cyclase-C (GCC) We evaluated the impact of obesity and other lifestyle-associated risk factors, such as
is an appealing CAR target due to apical-basal polarity of expression in normal colon, which smoking, alcohol use, and diabetes mellitus, on outcomes in patients with early-onset CRC.
may hamper on-target effects on the mucosa. GCC is present on nearly all colorectal cancers Methods: A comprehensive retrospective cohort study using the TriNetX database identified
(CRC). GCC19CART showed promise in a prior trial in China, demonstrating expansion, adults aged 18–49.9 diagnosed with CRC. Using propensity score matching (PSM), we
response, and persistence, consistent with the proposed mechanism. The US phase 1 study compared patients with obesity (body mass index [BMI] $30 kg/m²) and other lifestyle-
was initiated for refractory CRC to assess the safety and efficacy of GCC19CART in this associated risk factors (smoking, alcohol use, and diabetes mellitus) to those without any risk
population. Methods: Eligible patients underwent leukapheresis, lymphodepleting che- factor, while accounting for demographics, comorbidities, and treatment. The primary
motherapy (fludarabine 30mg/m2 and cyclophosphamide 300mg/m2 on day-3), and a single outcome is the 10-year mortality in obese patients as compared to non-obese patients.
dose of GCC19CART. Safety was the primary endpoint. Efficacy was assessed by RECIST Secondary outcomes included the 10-year mortality in patients with other lifestyle-associated
v1.1 based on local review. Results: As of January 23, 2025, 9 patients were treated: 4 at risk factors. Results: A total of 10,220 matched pairs of obese and non-obese patients were
dose-level (DL) 1 (1x106 cells/kg) and 5 at DL2 (2x106 cells/kg). Cytokine release syndrome included. Before matching, obese CRC patients were older, more likely to be male, Hispanic, or
occurred in all subjects (grade [G] 1: 6/9 [66.7%] and G2: 3/9 [33.3%]), and diarrhea was non-Hispanic Black, and had higher rates of colonoscopy, surgery, and comorbidities (p ,
reported in 8/9 (G1: 3/9 [33.3%], G2: 3/9 [33.3%], G3: 2/9 [22.2%]). Immune effector cell 0.001). After PSM, obese CRC patients had significantly lower odds of 10-year mortality
associated neurotoxicity syndrome occurred in 2/9 subjects (G2: 1/9 [11.1%], G3: 1/9 compared to non-obese patients (7.9% vs. 14%; adjusted odds ratio [aOR] = 0.53; 95%
[11.1%]). A DL2 patient experienced a dose limiting toxicity (G3 diarrhea, G4 enterocolitis, confidence interval [CI]: 0.47–0.60). Diabetes, smoking, and alcohol use showed no sig-
and G5 sepsis) and died 48 days post-infusion. The overall response rate (ORR) in DL1 was nificant association with 10-year mortality in patients with early-onset CRC (e.g., diabetes:
25% (1/4 partial response [PR]) and 80% in DL2 (4/5 with 3 PR and 1 pathological complete aOR = 0.96; 95% CI: 0.73–1.26). Conclusions: Our study suggests that obesity may confer a
response). The PR in DL1 was achieved by month 2, while 3/4 responders in DL2 achieved a protective effect on 10-year mortality in patients with early-onset CRC, whereas other
PR by month 1, demonstrating dose-dependent tumor-killing activity. Two DL2 patients lifestyle-associated risk factors showed limited-to-no significant impact. These findings
maintained responses at data cut-off. One patient achieved a complete metabolic response underscore the need for targeted strategies to improve access to CRC screening and
by PET at month 2 and maintained a PR by CT at month 6 with continuous tumor shrinkage treatment, particularly in younger patients with obesity, and they also open up new avenues
(month 1: 38.33%, month 2: 40.77%, month 4: 82.58%, month 6: 75.61%). Another patient for research into the potential mechanisms underlying these associations. Prospective
maintained a PR at month 6 with progression at month 8. The median progression-free studies are warranted to validate these results and explore these potential implications,
survival (PFS) was 5.0 months in DL1 and 7.8 months in DL2. The median duration of further enriching our understanding of this complex disease and potentially leading to novel
response was 2.2 months in DL1 and 6.9 months in DL2. Compared to the prior trial in China approaches for its management. Research Sponsor: None.
($G3 diarrhea: 22.2% vs. 53.3%; ORR: 66.7% vs. 40%, PFS: 7.8 vs 6.0 months in DL2, 5.0 vs Incidence % (N) Propensity score matching analysis
1.9 in DL1. Chen, JAMA Oncol., September 2024), the US study suggests a potential trend
Lifestyle risk factors CRC, 50 CRC,50 Adjusted odds ratio (aOR)* [95%CI]
towards improved safety and efficacy. Conclusions: GCC19CART demonstrated significant + risk factor - risk factor
clinical activity and durability in refractory CRC. Optimization of diarrhea/colitis Obesity (BMI ‡ 30) 7.9% (406) 14.0% (714) 0.53 [0.47 - 0.60]
management is ongoing. Updated data will be presented. Clinical trial information:
Propensity matching analysis assessing the odds of 10-year mortality when comparing patients with early-
NCT05319314. Research Sponsor: None. onset colorectal cancer (, 50 years) with lifestyle-associated risk factors (e.g., obesity (N=5110) compared to
equal numbers of patients without lifestyle-associated risk factors.

3583 Poster Session 3584 Poster Session

Colorectal cancer mortality dynamics: Uncovering critical disparities in U.S. Analytic and clinical validation of a negative prediction algorithm for ac-
population health (2018–2023). First Author: Pranav Chalasani, Wayne State tionable mutations utilizing genomic and epigenomic profiling in cfDNA. First
University, Rochester Hills, MI Author: Andrew Gross, Guardant Health, Palo Alto, CA
Background: Colorectal cancer(CRC) remains the third leading cause of cancer-related Background: One challenge in cell-free DNA (cfDNA) profiling for genomic tumor
deaths in the United States, with a disproportionate burden on underserved populations. profiling is the inability to confidently confirm the absence of actionable genomic
Despite established screening protocols and preventive measures, fewer than 35% of mutations. This limitation arises from the challenge of determining whether key driver
cases are detected early, significantly impacting survival rates. This study examines mutations are truly absent or if tumor levels are below the detection threshold of the
mortality patterns across demographic and geographic divides, revealing urgent public assay. Accurate negative variant prediction could enable clinicians to expedite clinical
health priorities. Methods: This retrospective analysis was performed in adults aged 25 decisions based on cfDNA results without relying on tissue biopsy sequencing when no
and older using the CDC WONDER database (2018-2023) using ICD-10 codes. We stratified actionable alterations are found. Here, we report analytic and clinical validation of a
mortality data by age, gender, race, geographic region and urbanization level to identify novel algorithm to enable negative prediction from liquid biopsy to address this critical
critical disparities and emerging trends. Crude mortality rates (CMRs) and Age-adjusted clinical need. Methods: Using the Guardant Infinity platform, which simultaneously
mortality rates (AAMRs) per 100,000 were calculated by age, gender, region and race, with profiles genomic and epigenomic signals in a single sample, we integrated highly
95% confidence intervals (CI) for precision. Temporal trends and annual percentage sensitive and precise tumor fraction estimates and developed a negative prediction
changes (APCs) were analyzed using Joinpoint regression. Results: From 2018 to 2023, algorithm to allow for confident reporting of samples that do not detect an actionable
among 313,744 deaths, mortality increased from 51,891 to 53,497, while the AAMR for CRC genomic finding. The algorithm estimates the post-test probability of a cfDNA sample
consistently declined from 12.92 to 12.44. The highest CMR was in the 85+ group (156.11 harboring genomic biomarkers with FDA approved therapies relevant to treatment
per 100,000, 95% CI: 153.05–159.17), followed by 75-84 (74.18, 95% CI: 72.87–75.49), 65- selection, based on population priors, epigenomic tumor fraction (TF), and the analysis
74 (40.27, 95% CI: 39.58–40.96), and 55-64 (23.83, 95% CI: 23.37–24.30). The 45-54 group of mutant and non-mutant coverage across variants of interest which was assessed for
had a CMR of 11.74 (95% CI: 11.40–12.07), the 35-44 group 3.56 (95% CI: 3.38–3.74), the advanced colorectal (CRC) and lung cancer (NSCLC) patients. Results: In 3973 CRC and
25-34 group 0.77 (95% CI: 0.69–0.86), and the 15-24 group had the lowest at 0.09 (95% CI: 7654 NSCLC analyzed patients, 41% of CRC and 22.6% of NSCLC were found to have an
0.06–0.12). Males had a higher CMR of 17.23 per 100,000 (AAMR: 15.19, APC: -0.68, p =
actionable mutation. Among the remaining samples, 66% of CRC and 56.3% of NSCLC
0.21) than females, who had 14.42 per 100,000 (AAMR: 10.69, APC: -0.30, p = 0.56). The
had sufficient tumor fraction to assess the sample as variant negative with . 95%
Midwest had the highest AAMR at 13.33 per 100,000 [APC: -0.78 (95% CI: -2.54 to 1.01, p =
confidence. Reasons why the remaining samples could not be confidently assessed
0.31)], followed by the South at 13.54 [APC: 0.05 (95% CI: -0.80 to 0.92, p = 0.91)], the West
included low tumor shedding (including 15% with nondetectable tumor), low genomic
at 11.90 [APC: 0.09 (95% CI: -0.79 to 0.95, p = 0.82)], and the Northeast at 11.54, with a
significant decline in trends [APC: -1.90 (95% CI: -3.24 to -0.58, p , 0.01)]. Large central
coverage over loci of interest, and mutant allele support below the confident call
metro areas accounted for 25.3% of deaths (83,341), followed by large fringe metro areas threshold. An additional cohort of 237 CRC and 316 NSCLC patients with paired tissue
(22.4%), medium metros (19.9%), micropolitan areas (10.0%), small metros (9.4%), and and cfDNA results was used to clinically validate the negative prediction algorithm. All
noncore areas (8.2%). Racial disparities showed White individuals with the highest CMR at samples with sufficient tumor fraction for . 95% confidence and predicted to be
17.01 per 100,000 (AAMR: 12.69, APC: -0.78), while Black or African American individuals negative by the algorithm for genomic biomarkers with FDA-approved therapies were
had a slightly lower CMR at 15.70 per 100,000 but the highest AAMR at 16.18 (APC: -1.63, confirmed to be negative in tissue results. Conclusions: A plasma-based epigenomics-
p , 0.01), followed by American Indian or Alaska Native, Asian, Native Hawaiian, and other based approach for confident negative prediction is feasible in CRC and NSCLC, as
groups. Conclusions: These findings reveal critical gaps in CRC prevention and care, demonstrated by validation results. Confident negative prediction has the potential to
disproportionately affecting young adults, males, and minorities. Public health initiatives enhance the utility of liquid biopsy and accelerate clinical decision-making in advanced
must expand screening, improve access to care, and address regional inequities to reduce solid tumors with biomarker-guided treatment pathways and should be validated in
mortality and promote health equity. Research Sponsor: None. additional clinical datasets. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
256s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3585 Poster Session 3586 Poster Session

Cell-free DNA 5-hydroxymethylcytosine profiling for the assessment of Sintilimab plus bevacizumab, oxaliplatin, and capecitabine as perioperative
colorectal cancer biology and treatment response in blood. First Author: therapy in microsatellite-stable, resectable colorectal cancer liver metas-
Ceyda Coruh, ClearNote Health, San Diego, CA tases: An open-label, single-arm, phase II trial. First Author: Yu-hong Li, Sun Yat-
Background: Colorectal cancer is the third most common cancer worldwide, accounting sen University Cancer Center, Guangzhou, China
for about 10% of all cancer cases, and is expected to claim more than 50,000 lives in 2025. Background: Immunotherapy has revolutionized cancer treatment, yet its efficacy in
Approximately 33% of CRC patients will develop metastases throughout their cancer proficient mismatch repair and/or microsatellite stable (pMMR/MSS) colorectal cancer
continuum, and their 5-year survival rate is about 15%. The majority of patients with liver metastases (CRLM) remains uncertain. Optimizing neoadjuvant regimens for such
metastatic colorectal cancer (mCRC) cannot be cured. However, a subset of mCRC patients patients is crucial. Methods: A prospective, open-label, single-arm phase II clinical trial
with localized recurrence or isolated metastases in the liver and/or lungs may achieve a was conducted from June 2021 to January 2023. Patients with resectable pMMR/MSS
cure through surgical intervention. Yet, current methods for identifying patients who are CRLM were enrolled and received 4 cycles sintilimab combined with bevacizumab,
candidates for more favorable responseremain inadequate. Therefore, there is a critical oxaliplatin, and capecitabine preoperatively followed by 4 cycles oxaliplatin, and
need for predictive biomarkers to accurately identify patients who are likely to experience capecitabine postoperatively. The primary endpoints were safety and feasibility of
better outcomes following surgery. 5-hydroxymethylcytosine (5hmC) is an epigenetic neoadjuvant therapy and surgery. Secondary endpoints encompassed pathological
modification that is associated with active genes and regulatory regions that are cell type- response rates, objective response rate, progression-free survival (PFS), and overall
and disease-specific. Here, we developed a model using cell-free DNA (cfDNA) 5hmC survival (OS). Biomarker analyses were performed to identify potential predictors related
profiles to detect CRC and identified pathways distinguishing mCRC patient outcomes to efficacy and prognosis. The study protocol was registered in [Link]
following treatment. Methods: Plasma was collected from 294 CRC patients and 588 non- (NCT04940546). Results: Between June 2021 to January 2023, 36 patients were en-
cancer individuals to obtain cfDNA. cfDNA was enriched for 5hmC-containing DNA rolled, and included in the safety analysis. The most common treatment-related adverse
fragments. Input and 5hmC-enriched cfDNA were subsequently used to generate se- events (TRAEs) were fatigue (55.6%), peripheral neuritis (52.8%). Of the 36 patients, 30
quencing libraries to obtain WGS and 5hmC profiles, respectively. Machine learning received local treatment for liver metastases. 26 of them underwent CRLM surgery
operating on 5hmC and WGS data was used to develop a CRC detection model which was resection, and 7 of the 26 experienced surgery - related complications graded from 1-2
subsequently tested on an independent set of mCRC (n = 69) and non-cancer samples (n =
such as cholecystitis and pulmonary infection, one patient died from respiratory failure
70). Differential 5hmC analysis was performed using edgeR and Gene Set Enrichment
due to a pulmonary infection (immune pneumonia not excluded) a month after liver
Analysis (GSEA). Results: The performance of the CRC prediction model was evaluated
metastases resection. 34 were analyzed for efficacy. The objective response rate (ORR)
through 10-fold cross-validation producing an auROC curve of 0.86. An independent
was 67.6%, with a disease control rate (DCR) of 88.2%. 26 patients underwent surgery;
validation set of mCRC and non-cancer patients displayed an auROC of 0.94. Comparative
GSEA using gene body 5hmC levels revealed biological pathways associated with CRC
the pathological complete response rate (pCR) was 11.5%, and the major pathological
biology such as Myc signaling. cfDNA 5hmC profiling of plasma obtained from mCRC response rate (MPR) was 38.5%. After a median follow-up of 32.9 months, the median
patients before surgery revealed quantitative differences in patients who show recurrence PFS was 14.2 months ( (95% CI: 11.6 - 29.0 months), and the median OS had not yet been
of disease within 2 years post-surgery from the patients who remain recurrence-free for at reached. Biomarker analysis revealed that RAS wild-type (mPFS: 29.0 months (15.0 - NA)
least 2 years after surgery. These differences between relapsed and non-relapsed groups vs 11.5 months (9.8 - 15.7), log-rank P = 0.0087), SMAD4 wild-type population (mPFS:
included 5hmC changes over genes involved in pathways known in mCRC, such as the Wnt/ 20.2 months (12.3 - NA) vs 6.9 months (5.2 - NA), log-rank P , 0.0001) may benifit from
b-catenin signaling (p , 0.05). Lastly, quantitative changes in 5hmC profiles measured in immunotherapy combination treatment. The single cell RNA sequencing analysis
pre-surgery plasma samples enabled prediction of disease recurrence in patients within 2 revealed that higher intrafiltion of FIB_PLAG2A in the TME of CR/PR were associated
years post-surgery. Conclusions: 5hmC analysis of cell free DNA) offers a novel, non- with favorable prognosis, while higher intrafiltion of MPH_TREM2, FIB_POSTN in the
invasive approach for identification of colorectal cancer biology and assessment of TME of non CR/PR were linked to poor prognosis. Conclusions: The neoadjuvant
treatment response in blood samples. Research Sponsor: ClearNote Health. regimen demonstrated acceptable safety and efficacy. RAS, SMAD4 wild-type patients
may be a potential beneficiary population. Clinical trial information: NCT04940546.
Research Sponsor: Bethune Public Welfare Foundation.

3587 Poster Session 3588 Poster Session

Safety and efficacy evaluation of neoadjuvant chemoradiotherapy plus Neoadjuvant ONO-4578, an EP4 antagonist, in combination with nivolumab
thymalfasin and tislelizumab for treating MSS/pMMR locally advanced after chemoradiation therapy in locally advanced resectable rectal cancer.
rectal cancer. First Author: Zhengyang Yang, Department of General Surgery, Bei- First Author: Yusuke Takahashi, Department of Gastroenterological Surgery, NHO Osaka
jing Friendship Hospital, Capital Medical University & National Clinical Research Center National Hospital, Osaka, Japan
for Digestive Diseases, Beijing, China Background: Neoadjuvant nivolumab (NIV) after preoperative chemoradiation therapy (CRT)
Background: Neoadjuvant chemoradiotherapy is currently the standard strategy for demonstrated promising pathologic complete response (pCR) in patients (pts) with locally
microsatellite stable (MSS) / mismatch repair-proficient (pMMR) locally advanced rectal advanced resectable rectal cancer (LARC) (Bando, Clin Cancer Res 2022). On the other hand,
cancer (LARC) patients. This study aimed to explore the safety and efficacy of combining the prostaglandin E2-EP4 signaling is known to induce immunosuppression in tumors. ONO-
specific (thymalfasin) and non-specific (tislelizumab) tumor immunotherapy with 4578 (4578), an antagonist of EP4, in combination with NIV has shown a manageable safety
chemoradiotherapy in MSS/pMMR LARC. Methods: This trial is an open, prospective, profile and signs of anti-tumor activity in pts with solid tumors. In this ONO-4578-03 study, we
multi-center, single-arm phase II clinical study assessing the efficacy and safety of evaluated safety, preliminary efficacy, and biomarkers of 4578 plus NIV after preoperative
neoadjuvant chemoradiotherapy combined with thymosin andtislelizumab in MSS/ CRT in pts with LARC. Methods: Pts with LARC who received preoperative CRT (50.4 Gy with
capecitabine 1,650 mg/m2) were eligible. Pts were divided into two groups for neoadjuvant
pMMR LARC. Stage II/III MSS/pMMR LARC patients (cT3-4aN0M0 and cT1-4aN1-2M0)
therapy: 4578 monotherapy lead-in (lead-in) and the combination group. Pts in the lead-in
with the tumor distal location # 10 cm from anal verge at two centers in China were
group received 4578 (40 mg, oral, daily) alone for 6 weeks, and then 4578 plus NIV (240 mg,
consecutively enrolled. Patients received chemoradiotherapy (50 Gy/25 f, 2 Gy/f, 5 days/ intravenous, every 2 weeks) for 4 weeks, while pts in the combination group received 4578
week, 5 weeks; plus capecitabine 850-1000 mg/m2, bid, po, 5 days/week, day1-5), plus NIV for 10 weeks. Subsequently, pts in both groups received radical resection. The
thymalfasin (4.8 mg, biw, ih, day 1 and day 4 from week 1-11) and three 21-day cycles primary endpoint was safety. Secondary endpoint was efficacy, including pCR rate using the
tislelizumab (200 mg, [Link], week 2, 5 and 8) as neoadjuvant therapy. Adjuvant therapies AJCC tumor regression grading. Ongoing exploratory endpoints include tissue and blood
after neoadjuvant were nonuniformly specified and decided according to clinical ex- biomarkers. Results: We enrolled 31 pts: 10 and 21 to the lead-in and combination groups,
periences. The primary endpoint is the complete response (CR) rate, defined as the respectively. The median age was 62.0 (range, 39–76) years, 20 pts (64.5%) had a disease
achievement of clinical complete response (cCR) after neoadjuvant therapy or path- stage of III, and all pts were classified as microsatellite stable. The pCR (AJCC grade 0) rates
ological complete response (pCR) after total mesorectal excision (TME). Results: From in the lead-in group, the combination group, and overall population were 50.0% (5/10 pts),
Feb 2024 to Aug 2024, a total number of patients (n = 25) were enrolled and 3 patients 23.8% (5/21 pts), and 32.3% (10/31 pts), respectively; the major pathological response (MPR;
were excluded because of T4b and dMMR. Finally, 2 patients were discontinued and 20 AJCC grade 0+1) rates were 70.0% (7/10 pts), 71.4% (15/21 pts), and 71.0% (22/31 pts),
completed neoadjuvant therapy. The median age was 67.5 (from 36 to 74) years while respectively. Among all pts, any-grade treatment-emergent adverse events (TEAEs) occurred
the median tumor distal location was6.0 (from 3.5 to 8.5) cm. The CR, PD, and SD rate in 23 pts (74.2%). including 3 pts (9.7%) with grade 3 TEAEs (appendicitis, ileus, drug-induced
was 40.0% (8/20), 45.0% (9/20), and 15.0% (3/20) correspondingly, with the ORR rate of liver injury, hypertension) and 1 pt with serious TEAEs (ileus, drug-induced liver injury). None
85.0% (17/20). Grade 3 treatment-related adverse events (trAEs) including leukopenia of the TEAEs led to treatment discontinuation or death. Any-grade treatment-related adverse
and neutropenia were observed in 1 (5%) patient, while grade 1-2 trAEs were observed in events (TRAEs) occurred in 11 pts (35.5%), including 1 pt with a serious TRAE (grade 3 drug-
15(75.0%) patients. As for Dec 31, 2024, the EFS rate was 100% (20/20) with median induced liver injury). Radical resection was not performed within the protocol-defined window
follow-up time of 18.57 weeks (from 6.86 to 31.86). Conclusions: Neoadjuvant che- in 1 pt in the lead-in group and in 2 pts in the combination group due to progressive disease, a
moradiotherapy plus thymalfasin and tislelizumab show promising anti-tumour activity TRAE (grade 1 hyperthyroidism), or clinical CR, respectively. As of the final analysis, 5 pts
in MSS/pMMR LARC patients, with manageable toxicities. This study suggests that such experienced recurrence and 1 pts died in the overall population, after the median follow-up of
23.29 (range, 15.9–32.9) months. Conclusions: Neoadjuvant 4578 plus NIV after CRT
combination could be a promising therapeutic strategy for patients with MSS/pMMR
showed a manageable safety profile and promising pCR rates and MPR rates in pts with LARC.
LARC. Clinical trial information: NCT06056804. Research Sponsor: Beijing Li Huanying
Clinical trial information: jRCT2051200096. Research Sponsor: Ono Pharmaceutical Co., Ltd.
Medical Foundation.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 257s

3589 Poster Session 3590 Poster Session

Single-incision laparoscopic surgery vs conventional laparoscopic surgery The association of ctDNA with recurrence in patients with stage II-IV colo-
for colorectal cancer: Short-term outcomes of a multi-center, randomized, rectal cancer: The b-CORRECT study. First Author: Tadayoshi Hashimoto,
controlled trial. First Author: Yaqi Zhang, Ruijin Hospital, Affiliated by Shanghai Department of Gastroenterology and Gastrointestinal Oncology, National Cancer Center
Jiaotong University School of Medicine, Shanghai, China Hospital East, Kashiwa, Japan
Background: Single-incision laparoscopic surgery (SILS) is increasingly being embraced Background: CRC is a leading cause of cancer-related mortality globally. Detection of molecular
in the medical community due to its potential to offer less-invasiveness and quick re- residual disease (MRD) is an early indicator of recurrence and may allow for timely intervention. Here,
covery. This multi-center randomized controlled trial compared the short-term and long- we evaluated an analytically validated tumor-informed ctDNA MRD assay (Oncodetect) in a cohort of
term outcomes of single-incision laparoscopic surgery (SILS) with conventional laparo- patients with CRC. Methods: This retrospective study utilized data and specimens from 468 patients
with Stage II, III or resectable Stage IV CRC consecutively enrolled from June 1, 2020 through
scopic surgery (CLS) for colorectal cancer, which might be the first of its kind that involved November 30, 2022 in the GALAXY study with available residual samples. Tumor tissue underwent
both colon and rectal cancer. Study recruitment has completed, and the follow-up is whole-exome sequencing to identify up to 200 tumor-specific variants for designing a personalized
ongoing. Here we report the short-term outcomes of this trial. Methods: The trial was MRD test. The test was used to assess ctDNA status in plasma at three timepoints: post-surgical (PS),
conducted across 11 hospitals in 6 provinces of China. Participants included patients with post-definitive therapy (PDT) and in the surveillance period, which included the PDT and subsequent
histologically confirmed colorectal carcinoma that situated above the peritoneal reflection, timepoints. The primary endpoint was the association of ctDNA status during surveillance with
clinically staged as I-III. Patients were randomly assigned in a 1:1 ratio to SILS or CLS disease-free survival (DFS). An analysis of the association between RNA-seq expression and DFS was
group. Comprehensive perioperative data were meticulously gathered, and follow-up also planned. Results: Analysis included a total of 1648 ctDNA results from 417 patients
assessments were scheduled postoperatively. The primary endpoint was 3-year with $1ctDNA result from $1 timepoint. Among these patients, 296 (71.0%) had colon and 121
disease-free survival (DFS), secondary endpoints included overall survival (OS), onco- (29.0%) had rectal cancer, 141 (33.8%) had Stage II, 249 (59.7%) Stage III and 27 (6.5%) Stage IV
disease, and 255 (61.2%) received adjuvant chemotherapy. Median follow-up was 1.9 years. The
logical efficacy, and postoperative outcomes. Results: Between May 2021 and April
median ctDNA level among detections, measured in mean tumor molecules per ml (MTM/ml) was
2023, a total of 712 patients were randomly assigned to either the Single-incision Lap- 1.187 (range 0.006-3180.5). During surveillance, ctDNA detection was strongly prognostic for DFS (HR
aroscopic Surgery (SILS) group (n = 354) or the Conventional Laparoscopic Surgery (CLS) 36.6; CI 21.9 – 61.2; ctDNA status as a time-dependent variable). Similarly, ctDNA detection was
group (n = 358). The distribution of surgical procedures included 162 (22.8%) right strongly associated with DFS at the PS and PDT timepoints (Table). Multivariable analysis showed
hemicolectomies, 326 (45.8%) left hemicolectomies, and 224 (31.4%) proctectomies. The ctDNA status remained strongly associated with DFS while other clinicopathological factors did not.
pathological TNM stages I, II, and III of the mITT population were 10.7%, 36.9%, and 52.4%, The median lead time between ctDNA detection and clinical recurrence was 97 days (95% CI: 51-114).
respectively. In the SILS group, 92.9% (n = 329) of the cases were completed entirely with a RNA-seq analysis is ongoing. Conclusions: In a cohort of 468 patients who underwent curative-intent
single incision. An additional trocar was used to assist the surgical procedure in 5.6% (n = surgery for stage II-IV CRC, a tumor-informed quantitative ctDNA assay using up to 200 variants was
20) of the cases, and 0.8% (n = 3) were converted to conventional laparoscopic surgery. strongly prognostic for DFS at all timepoints. The prognostic ability of RNA-seq expression analysis for
ctDNA status and outcome in this cohort is currently being determined. Research Sponsor: Exact
Two patients in the SILS group required conversion to open surgery, compared to 10
Sciences; Japan Agency for Medical Research and Development.
patients in the CLS group. The incidence of postoperative complications and oncological
efficacy were statistically equivalent between the two groups. Moreover, patients in the Association of ctDNA status with DFS.
SILS group reported significantly less postoperative pain (p = 0.02). There were no Statistic Result
significant differences in short-term overall survival (OS) and disease-free survival (DFS) Surveillance (n= 398) HR (95% CI) 36.6 (21.9 – 61.2), p,0.0001
between the two arms. Conclusions: Single-incision laparoscopic surgery (SILS) for Sensitivity (95% CI)
Specificity (95% CI)
64.8% (53.2 - 74.9%)
98.8% (96.8 - 99.5%)
colorectal cancer has demonstrated with feasibility, safety, and efficacy. The less painful PS (n = 241) HR (95% CI) 7.5 (4.3 – 13.1), p,0.0001
postoperative experience was aligned with the principles of Enhanced Recovery After Sensitivity (95% CI) 44.2% (31.6% - 57.7%)
Specificity (95% CI) 95.6% (91.6% - 97.8%)
Surgery (ERAS). This surgical approach extended the concept of minimally invasiveness PDT (n = 367) HR (95% CI) 24.0 (13.8 – 41.7), p,0.0001
and represents a logical progression towards Natural Orifice Transluminal Endoscopic Sensitivity (95% CI) 45.5% (34.0% - 57.4%)
Surgery (NOTES) or single-incision robotic surgery. Clinical trial information: Specificity (95% CI) 99.0% (97.1% - 99.7%)
NCT04527861. Research Sponsor: SHANGHAI HOSPITAL DEVELOPMENT CENTER.

3591 Poster Session 3592 Poster Session

Biologic correlates of circulating tumor DNA (ctDNA) shedding in the IN- Phase II study of short-course radiotherapy (SCRT) followed by consolida-
TERCEPT colorectal cancer (CRC) study. First Author: Emerik Osterlund, tion chemotherapy with FOLFOXIRI as total neoadjuvant therapy (TNT) for
Department of Gastrointestinal Medical Oncology, Division of Cancer Medicine, The locally advanced rectal cancer (LARC) patients (pts): The ShorTrip study.
University of Texas MD Anderson Cancer Center, Houston, TX First Author: Martina Carullo, Unit of Medical Oncology 2, Azienda Ospedaliera Uni-
Background: ctDNA is a promising tool for early cancer detection and monitoring of versitaria Pisana and Department of Translational Research and New Technologies in
minimal residual disease (MRD). The relationship between vesicle trafficking of ctDNA- Medicine and Surgery, University of Pisa, Pisa, Italy
laden exosomes and shedding of detectable ctDNA in patients is poorly understood and Background: TNT is a recognised option for the treatment of LARC. The efficacy of both FOLFIRINOX followed
was therefore explored in a large prospective patient cohort. Methods: The INTERCEPT by long-course CTRT and SCRT followed by FOLFOX or CAPOX was demonstrated in two phase III trials. No
data are available regarding the feasibility and activity of SCRT followed by the triplet as TNT in LARC.
program prospectively enrolled patients undergoing curative intent surgery for stage I-IV Methods: ShorTrip is an Italian, prospective, multicentre, single-arm phase II trial (NCT05253846). Pts #70
CRC at MD Anderson Cancer Center. Tumor informed MRD assays (Signatera) were years with medium-high (5-10 cm from the anal verge) LARC with at least one of the following features: cT4,
drawn postoperatively and every three months according to reimbursement guidelines. cN2, involved mesorectal fascia (MRF+) or cT3N+, received SCRT followed by 8 cycles of FOLFOXIRI and
RNA analyses were done from FFPE. Gene set enrichment analyses (GSEA) and Z-Scores surgery. The primary endpoint was the pCR rate. According to the Fleming single stage design, hypothesizing
p0 = 0.25 and p1 = 0.40, setting 90% power with an a error of 0.10 (one-sided), the experimental regimen would
comparing ctDNA+ and ctDNA– were analyzed using log2 normalized RNA expression have been considered promising if at least 21 pCRs were observed out of 63 enrolled pts. After the first 11 pts
values. Results: The cohort included 579 patients with RNA and post-operative ctDNA starting consolidation treatment, a higher than expected occurrence of severe neutropenia after the 1st cycle of
analyses; median age 56 years, 56% male and 47% stage IV. Of these, 122 (20%) were FOLFOXIRI (N = 7, 64%) was observed and the protocol was amended to administer one cycle of FOLFOX after
ctDNA+ in their first draw (32% of stage IV vs. 11% of stage I-III). In GSEA analyses of SCRT followed by 7 cycles of FOLFOXIRI. Results: From January 2022 to February 2024, 64 pts were enrolled
Hallmark gene sets between ctDNA+ and ctDNA–, an upregulation was seen for 15/50 in 9 centres with the following characteristics: median age 62 years (IQR 55-66), male 66%, ECOG PS = 0 89%,
medium/high rectum 76%/24%, cT2/cT3/cT4 5%/76%/19%, cN0/cN1/cN2 2%/35%/63%, MRF+ 42%, lateral
gene sets among ctDNA+, with p-value , 0.05 and false discovery rate , 0.25. Two of nodes 35%, EMVI+ 41%. The 52 tumors tested for MMR were pMMR. One patient withdrew consent after the 1st
the top signatures (UV response up and Unfolded protein response) were significant also cycle of chemotherapy and was not evaluated for pathological response. 21 (33%) and 43 (67%) pts achieved
in analyses stratified by stage. Analysis of the leading-edge genes in these gene set pCR and major pathological response (MPR), respectively. Almost all pCRs (N = 20, 95%) and MPRs (N = 42,
identified several members of the vacuolar ATPase (V-ATPase) family of genes which 98%) were observed in pts receiving at least 5 cycles of FOLFOXIRI (N = 56). Among 63 resected pts, 62 (98%)
and 1 (2%) achieved R0 and R1 resections, respectively. All pts completed SCRT and the only grade 3/4 acute
were highly enriched in ctDNA+. The full V-ATPase gene set substantially differed
toxicity was diarrhoea in 7 (11%) pts. 49 (77%) pts received 8 cycles of consolidation treatment as planned.
between ctDNA+ and ctDNA– (mean Z-score 0.48 vs. -0.13, p , 0.001), including when Irinotecan was never administered in 5 (8%) pts. Main grade 3/4 toxicity during consolidation are listed in the
stratified by stage (I-III: 0.46 vs. -0.28, p = 0.002; IV: 0.49 vs. 0.10, p = 0.051). Sixty Table. Early post-surgical complications were reported in 8 (13%) pts. Conclusions: SCRT followed by one
percent of patients (n = 346) had relapse event data with sufficient follow up. V-ATPases cycle of FOLFOX and 7 cycles of FOLFOXIRI showed a promising activity and a feasible safety profile and is
had higher mean Z-scores in those with relapses than those without (0.33 vs. –0.91, p = therefore worth of further studies especially in the NOM scenario. Clinical trial information: NCT05253846.
Research Sponsor: GONO Foundation.
0.012), also seen in the ctDNA– group (0.30 vs. -0.20, p = 0.021). Conclusions: V-
ATPase genes are differentially expressed in patients with ctDNA+ regardless of tumor Overall
stage, a result also mirrored in relapse events. V-ATPases may play a significant role in population Pre-amendment Post-amendment
Main G3/4 Adverse Events N=64 N=11 N=53
ctDNA release through regulating intracellular multivesicular bodies to exosome release, during consolidation CT n (%) n (%) n (%)
thereby providing a potential mechanistic link between tumor biology and ctDNA Any event 40 (62) 9 (82) 31 (58)
shedding. This finding may explain the clinical limitations of ctDNA in selected patients Neutropenia 33 (52) 8 (72) 25 (47)
and provide personalization of ctDNA testing performance in the future. Research Febrile Neutropenia 3 (5) 1 (9) 2 (4)
Anaemia 5 (8) 1 (9) 4 (8)
Sponsor: None. Diarrhoea 6 (9) 2 (18) 4 (8)
Stomatitis 5 (8) 1 (9) 4 (8)
Neurotoxicity 1 (2) 1 (9) -
Asthenia 4 (6) 2 (18) 2 (4)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
258s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3593 Poster Session 3594 Poster Session

Digital spatial profiling: Mapping tumor responses to radiotherapy in rectal Clinical and immunopathological evaluation and its comparison with con-
cancer. First Author: Rachel Violet Purcell, Department of Surgery and Critical Care, sensus molecular subtypes of colorectal cancer. First Author: Eduardo Felic-
Christchurch, New Zealand iangeli, Hospital General Universitario Santa Lucia, Cartagena, Spain
Background: Variation in response to radiotherapy for the treatment of rectal cancer is Background: This study aims to elucidate the prognostic impact of the immunoscore
likely due to heterogeneity in the tumour microenvironment. However, to date, no reliable within the context of consensus molecular subtypes (CMS), tumor budding (TB), and
predictive biomarkers of response are in clinical use and the mechanisms underlying macrophage infiltration in colorectal cancer (CRC), addressing a gap in current research.
response are unknown. Tertiary lymphoid structures (TLS), which are ectopic lymphoid Methods: A retrospective observational study analyzing 255 colorectal cancer cases.
aggregates found in the tumour microenvironment, have been linked to response to Demographic, histopathological, and clinical variables were examined. Molecular
immunotherapy, but little is known about their role in radiotherapy. Here, we aimed to classification, immunoscore, and macrophage infiltration were determined via immu-
explore the potential of lymphocytes and tertiary lymphoid structures as predictive nohistochemistry. The study adhered to ethical guidelines and received approval from
biomarkers of response to radiotherapy and profile the tumour immune microenvi- our ethics committee. CMS assessment used automated staining for specific markers,
ronment in the context of response to radiotherapy. Methods: For this study, we with molecular subtype determined using an online classifier (Ten Hoorn et al.).
accessed pre-treatment biopsies from 20 rectal cancer patients with known pathological Immunoscore calculation involved evaluating CD3+ and CD8+ immune cells, classifying
response to long-course chemoradiotherapy (LCCRT). We selected regions of interest patients into low or intermediate-high groups (Jiang et al.). Macrophage assessment
based on immunohistological identification of tumour and lymphocytic infiltrate in focused on CD163+ cells, categorizing them as spindle-cell and round-cell. Statistical
formalin-fixed paraffin-embedded tissue. We performed targeted proteomic profiling of analysis employed SPSS, using descriptive statistics, chi-square tests, Kaplan-Meier
87 immuno-oncology proteins using the Nanostring GeoMx Digital Spatial Profiler to survival curves, and multivariate analyses. Results: In this study of 255 colorectal
quantify protein expression with spatial resolution within regions of interest, including cancer patients, predominantly with localized disease, 34.9% had stage III disease.
TLSs, in the tumour microenvironment. Results: Unsupervised clustering based on Conventional and serrated adenocarcinomas were the main histological subtypes. CMS
normalised protein expression showed a clear separation between the complete re- classification revealed mostly CMS2-3 (69.4%), with relapse occurring across all
sponders to LCCRT and all other tumours, and this separation is driven by differences in subtypes. Low immunoscore was common in conventional and serrated histology and
T cells within TLSs (CD3+). Differentially expressed proteins within CD3+ aggregates CMS2/3, while MSI-H correlated with intermediate-high immunoscore. Tumor budding
include depletion of the natural killer cell marker, CD56 and increased expression of the (TB) was prevalent in relapsed patients, especially in CMS2/3 and CMS4, and associated
apoptosis marker, cleaved caspase 9. The distribution of TLS-tumour distance was also with serrated histology. Metastatic patterns varied by CMS subtype, with TB . 20 foci
significantly different between response groups. Conclusions: The study highlights the linked to hepatic metastases. CD163 macrophage infiltration was associated with CMS1
role of TLSs in modulating the immunogenic landscape of the tumour microenvironment and CMS2/3, and a high immune score. Over 9.6 years of follow-up, tumor budding was
in rectal cancer, likely influencing the response to radiotherapy. Spatially resolved associated with overall and relapse-free survival, while CMS was linked to overall
proteomic analyses identifies potential biomarkers for radiotherapy response and survival. Immunoscore showed no association with survival outcomes.
underscores the importance of profiling tumour-immune microenvironment complexity Conclusions: This cohort shows heterogeneous disease progression and prognosis,
when stratifying patients for therapy. Research Sponsor: Health Research Council of with CMS2/3 exhibiting high tumor budding and relapse rates, especially in serrated
New Zealand; Maurice Wilkins Centre for Biomolecular Discovery. histology. Molecular subtypes have distinct metastatic patterns: CMS1 to the perito-
neum, CMS2/3 to the liver, and CMS4 to both. Relapsed CMS2/3 cases had low
immunoscore, while CD163+ macrophage infiltration correlated with higher immune
scores in CMS1 and CMS2/3, highlighting the complex interactions between molecular
subtypes, immune responses, and tumor behavior. Research Sponsor: SEOM (Spanish
Society of Medical Oncology); Instituto de Salud Carlos III; ICI20/00044; European
Commission H2020; GA: 848098.

3595 Poster Session 3596 Poster Session

Neoadjuvant chemotherapy and surgery for rectal cancer: Omission of Performance of a targeted enzymatic methylation-based early detection
radiation in clinical practice. First Author: Matthew B. Hill, Memorial Sloan Ket- test by different colorectal cancer subgroups. First Author: Xiaojian Wu, The Sixth
tering Cancer Center, New York, NY Affiliated Hospital, Sun Yat-sen University, Guangzhou, China
Background: PROSPECT proved efficacy of induction chemotherapy and surgery, Background: Colorectal cancer (CRC) is the second most frequently diagnosed cancer in China and
without radiation, for select stage II and III rectal cancer patients with improvements in early detection could prevent over 90% of CRC-related deaths. Blood-based tests that analyze
bowel function and less diverting ileostomy. This approach also avoids radiation- molecular features of CRC cell-free DNA (cfDNA), such as methylation and fragmentation patterns,
hold great promise for early detection. However, the impact of molecular characteristics related to
associated alteration in fertility which is particularly important in the rising early onset tumor location or mismatch repair (MMR) status on test performance has not been thoroughly in-
population. In this study we review our broader current use of induction chemotherapy vestigated. In this study, we developed a blood-based CRC early detection test and analysed its
without radiation in locally advanced rectal cancer patients. Methods: Review of a performance across different CRC subgroups. Methods: A targeted enzymatic methyl sequencing
prospectively maintained IRB approved, institutional database for patients treated with panel was developed to identify tumor-specific hyper- and hypo-methylation markers and frag-
induction chemotherapy and surgery between 2015-2024. Clinicopathologic features are mentation profiles. A case-control cohort of 536 participants (268 CRC patients, 268 controls) was
summarized, and disease-free survival (DFS) measured with the Kaplan-Meier method. enrolled and startified into training and validation sets base on case/control status and cancer stage
Results: A total of 171 patients, median age 50 years (IQR 43-61), were identified with with 5-fold cross-validation. A gradient-boosted tree model was built by combining probabilities from
methylomic and fragmentomic features. The optimal cutoff value for the early detection was de-
tumors located #5cm (n = 4, 2.3%), 6-10cm (n = 75, 44%), and 11-15cm (n = 92, 54%)
termined by Youden’s index, High specificity and High sensitivity methods, respectively. Results: The
from the anal verge. Pre-treatment MRI staging was available for 169/171 patients. 2 overall performance of Youden’s index, High specificity and High sensitivity methods was as follows:
(1.2%) and 167 (99%) were MRI stage II and III including 23 with T4 lesions, 15 with extra- specificity of 93.7%, 99.3%, 90.3%, and sensitivity of 96.6%, 86.2%, 97.0%, respectively. The area under
TME lymph nodes, and 28 with EMVI. Neoadjuvant chemotherapy regimens included the curve (AUC) value is 0.989 (95% CI: 0.981-0.996) , which is higher than those in current reports.
CAPEOX (n = 71, 42%), FOLFOX (n = 98, 57%), and FOLFIRINOX (n = 2, 1.2%). 166 (97%) When employing the High specificity method, the sensitivities were comparable between left and right-
underwent low anterior resection with (n = 94, 57%) or without (n = 72, 43%) diverting sided colon cancer (86.3% vs 85.7%, p = 1.0), and also similar between the dMMR (deficient mismatch
ileostomy, 4 (2.3%) underwent abdominoperineal resection (APR), and 1 (0.6%) repair) and pMMR (proficient mismatch repair) (87.5% vs 85.8%, p = 1.0), indicating that this model is
underwent a Hartmann procedure. Pathologic responses included: 28 (16%) AJCC TRG 0 applicable to various CRC subtypes. Additionally the TNM staging, pathological differentiation status,
and the expression of Ki67, which are closely related to aggressiveness, were correlated with the
(no viable cancer cells); 32 (19%) TRG 1 (small cluster/single cancer cells); 77 (45%) TRG sensitivity (Table). Conclusions: We have established CRC early detection model based on ctDNA
2 (residual cancer with predominant fibrosis); and 34 (20%) TRG 3 (extensive residual methylation and fragmentation profiles, which shows excellent overall performance. Notably, this
cancer). Tumor deposits were present in 24 (14%) and positive/close margin was noted newly developed blood-based model shows no significant differences in sensitivity between distinct
in 2 (1.2%). With a median follow-up of 24 months, 4 (2.3%) patients developed local tumor locations or varying MMR statuses, suggesting its broader applicability across different types of
recurrence (salvaged with chemotherapy, radiation, and surgery) and 18 (11%) devel- CRC. Research Sponsor: Shanghai Xiaohe Medical Laboratory Co., Ltd.
oped distant metastases. 1-year DFS was 92% (CI: 88-97) and 2-year DFS was 87% (CI: Subgroup Positive/Total no. Sensitivity p_value
81-93). Conclusions: Since PROSPECT, induction chemotherapy and surgery is being
Left-sided 195/226 86.3% 1
offered to higher risk rectal cancer patients, including those with T4 lesions and EMVI, Right-sided 36/42 85.7%
with favorable results. Continued individualized care based on response to chemo- dMMR 7/8 87.5% 1
pMMR 200/233 85.8%
therapy and omission of radiation can limit treatment related toxicity while maintaining Stage I 34/50 68% ＜0.001
excellent oncologic outcome. Research Sponsor: P30 CA008748. Stage II 92/108 85.2%
Stage III 72/76 94.7%
Stage IV 33/34 100%
Well differentiated 3/3 100% 0.029
Moderately differentiated 160/192 83.3%
Poorly differentiated 41/42 97.6%
Ki67_high 138/159 86.8% 0.147
Ki67_low 32/42 76.2%

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 259s

3597 Poster Session 3598 Poster Session

ctDNA dynamics and targeted therapies associated with genetic mutations Oncologic outcomes of organ preservation in patients with rectal adeno-
in patients with colorectal cancer. First Author: Midhun Malla, University of Ala- carcinoma treated with total neoadjuvant therapy: A single-center study.
bama at Birmingham, Birmingham, AL First Author: Mariem Galuia, AdventHealth Cancer Institute, Orlando, FL
Background: Colorectal cancer (CRC) is a heterogeneous disease with various genetic Background: Total neoadjuvant therapy (TNT) reduces the risk of local recurrence and distant metastases in patients
(pts) with locally advanced rectal cancer (RC). Selected pts who achieve a complete clinical response (cCR) after TNT
mutations that guide targeted therapy decisions, as outlined by NCCN guidelines. Here who can undergo a strict surveillance protocol may be considered for non operative management (NOM) to preserve
we evaluated the proportion of CRC patients receiving targeted therapies using Natera’s rectal function. Methods: We conducted an observational, retrospective, single-center study to evaluate watch-and-wait
proprietary Real-World Database. Methods: Whole-exome sequencing (WES) data of strategy in those achieved organ preservation & survival in pts with locally advanced or resectable metastatic pMMR RC
treated with TNT. Pts received either induction chemotherapy followed by chemoradiation (INCT-CRT) or chemoradiation
tumor samples from CRC patients undergoing tumor-informed ctDNA testing CRC were followed by consolidation chemotherapy (CRT-CNCT). INCT and CNCT consisted of 6-8 cycles of FOLFOX or 5 cycles of
analyzed. WES was performed on tumor tissue as part of the assay design workflow for CAPEOX. CRT consisted of a total dose of 5,040-5,750 CGy to the tumor and lymph nodes along with capecitabine. Pts
Signatera™ molecular residual disease testing, ordered between June 2019 and July were assessed for treatment response with digital rectal exam, MRI, flexsig 8 weeks after TNT. Pts who achieved cCR or a
near-cCR were offered NOM with WW. Close surveillance with the above-mentioned modalities was repeated every
2024. From the overall cohort of 47,476 CRC cases, we selected those with BRAF V600 3 months, CT chest/abdomen or PET scan was performed every 6 months. Data on local recurrence, distant metastases
(prevalence 13.7%) or KRAS G12C (prevalence 3.3%) actionable mutations, resulting in and survival was collected. Results: From Dec 2017 to Jan 2024, a total of 109 pts with RC went on WW after TNT in our
8,473 patients included in the analysis. We utilized commercially available claims data to center. Most pts were males (66%). Median age was 59 (30-88). The proportions of stages I to IV were 2.7%, 19.2%, %,
identify targeted therapy usage among clinical cases in our database. We examined the 76.1%, and 3.6% respectively. After a median follow-up of 20.2 months, median time of sustained cCR was 67 weeks,
tumor regrowth occurred in 7/109 (6.4%) pts. 3/109 pts developed distant metastases (8.5%). Median time from cCR to
use of 3 different FDA-approved targeted therapies in patients with CRC. local regrowth/metastasis was 42 weeks. Recurrence was detected within the first 2 years in all 7 patients. 3/7 patients
Results: Among 8,473 CRC patients with clinically actionable mutations in BRAF or are in remission after salvage surgery, 2/7 pts were scheduled for surgery, 1/7 pts died of disease progression and
KRAS, the majority had a BRAF V600 mutation (78.6%; N = 6,662) followed by KRAS G12C another pt was a poor surgical candidate. 93% (102/109) of our pts are with no evidence of disease. Conclusions: Organ
preservation for locally advanced rectal cancer is feasible and successful in a large community-based hospital system for
(21.4%; N = 1,811). Staging information was available for 93.9% (7,953/8,473) cases, selected pts who achieve cCR to TNT. WW is a reasonable and attractive strategy for both patients and oncologists that
with 15.6% stage IV at first ctDNA testing and 1.5% (123/7953) upstaged to stage IV at minimizes post-operative morbidity . Ongoing trials like JANUS aim at increasing cCR rate. Research Sponsor: None.
subsequent testing. An additional 6.7% (487/7,233) cases were categorized as Demographic and clinical characteristics of the pt cohort.
recurrent/metastatic based on treatment information from claims records. Overall rates Variables All pts (n=109)
of treatment with corresponding targeted therapies were 4.0% (264/6,662) and 3.3% (60/ Age, median (years) 59.0 (30-88)
1,811) for BRAF and KRAS, respectively. Within the subgroup of confirmed recurrent/ Gender
Male 73 (66%)
metastatic cases (N = 1,727), targeted therapy rates were 18.9% (233/1241) for BRAF Female 36 (34%)
Ethnicity
and 10.7% (52/486) for KRAS. No therapy overlap and no discordant cases (i.e., BRAF Caucasian 78 (71.5%)
Hispanic 23 (21.1%)
therapy was not given to KRAS mutated cases, and vice versa) were observed. Targeted Other 8 (7.3%)
therapy was typically started after the start of ctDNA testing (KRAS: in 96.7%, 58/60 Clinical stage at diagnosis
I 3 (2.7%)
cases, median 422 days after, BRAF: in 75.8%, 200/264 cases, median 200 days after). II 21 (19.2%)
III 83 (76.1%)
ctDNA clearance rate on therapy (i.e. conversion from ctDNA+ to ctDNA-) was observed IV 2 (1.8%)
to be 37.6% (56/149) which matches objective response rates previously reported for TNT Strategy
NCT-CRT 45 (41.2%)
radiological assessment. Conclusions: In this analysis, we demonstrate the utility of the CRT-CNCT 64 (58.8%)
CT regimen
commercial claims database to provide insights into different treatment modalities FOLFOX 91 (84.4%)
CAPEOX 14 (12.8%)
considered for patients with actionable mutations. Understanding patterns of ctDNA Other 3 (2.7%)
dynamics during targeted therapy can potentially act as a surrogate of treatment ef- Months of follow-up, median 20.2
Tumor Grade
ficacy and may guide future clinical trials. Research Sponsor: None. 1 7 (6.4%)
2 53 (48.6%)
3 2 (1.8%)
Unknown 47 (43.11%)

3599 Poster Session 3600 Poster Session

Neoadjuvant mFOLFOXIRI chemotherapy with or without cadonilimab ver- Impact of perioperative complications on ctDNA-based MRD detection and
sus mFOLFOX6 alone in locally advanced colorectal cancer: A randomized prognosis: Insights from the GALAXY study. First Author: Eiji Oki, Department of
phase II study (OPTICAL2). First Author: Jianwei Zhang, Department of Medical Surgery and Science, Graduate School of Medical Sciences, Kyushu University Hospital,
Oncology, the Sixth Affiliated Hospital of Sun Yat-sen University, Guangzhou, China Fukuoka, Japan
Background: The current standard treatment for locally advanced rectal cancer is chemoradiotherapy Background: The GALAXY study (UMIN000039205) demonstrated the utility of cir-
(CRT) followed by total mesorectal excision (TME) . For locally advanced colon cancer, neoadjuvant culating tumor DNA (ctDNA)testing to detect molecular residual disease (MRD) and
FOLFOX is also an option. In the era of immunotherapy, several studies have explored the efficacy of
monitor postoperative recurrence. This analysis evaluates the influence of perioperative
CRT combined with immunotherapy treatment. However, no studies have yet investigated the efficacy
and safety of chemotherapy combined with immunotherapy in locally advanced colorectal cancer. We complications and the timing of blood sampling on MRD detection rates and predicting
aim to explore the efficacy of mFOLFOXIRI with or without cadonilimab (AK104) compared to clinical outcomes. Methods: From the 6,032 patients enrolled in GALAXY, 2,400 were
mFOLFOX6 neoadjuvant chemotherapy in locally advanced colorectal cancer (LACRC). available for this analysis after excluding those enrolled in randomized trials or with
Methods: OPTICAL-2 was a randomized, phase II trial in patients with II/III rectal ancer and locally insufficient follow-up. A clinically validated, tumor-informed ctDNA assay (Signatera,
advanced colon cancer (T3 $5 mm or T4). Patients were randomly assigned ([Link] ) to 3 groups: Natera, Inc.) was utilized to prospectively detect and quantify ctDNA. MRD was assessed
preoperative mFOLFOXIRI plus AK104 for 6 cycles or mFOLFOXIRI for 6 cycles, or mFOLFOX6 alone for within a defined postoperative "MRD window" of 2-10 weeks post-surgery. Perioperative
6 cycles, followed by TME and adjuvant chemotherapy. The primary endpoint was pCR rate in mITT
complications were classified as Grade 2 or higher according to the Clavien-Dindo
population, and the secondary endpoint was major pathological response (MPR) rate, 3-year disease-
free survival, overall survival and safety. Results: From July 2023 to August 2024, 123 patients with classification. Results: Perioperative complications occurred in 302 cases (12.6%), with
LACRC were enrolled, with 41 patients in each group, including 22 colon cancer and 101 rectal cancer. anastomotic leakage (2.4%), ileus (2.0%), and intra-abdominal abscesses (1.3%) being
As the data cutoff, 121 patients had underwent surgery (41 in mFOLFOXIRI plus AK104, 39 in the most common. Complications were more frequent in males and patients with rectal
mFOLFOXIRI group and 41 in mFOLFOX6 group). Preoperative radiotherapy was added after induction cancer. Cell-free DNA (cfDNA) concentrations measured at 2–4 weeks post-surgery
treatment in 5 (12.2%), 4 (9.7%) and 3 (7.3%) patients among the 3 groups. In the mITT analysis, the pCR were significantly higher in cases with complications compared to those without
rate was 26.8% vs. 15.4% vs. 9.8% among the 3 groups, respectively. The downstaging (ypStage 0 to 1) complications (9.2 vs. 6.9 ng/mL; p , 0.001). Three-year recurrence-free survival (RFS)
was 65.9%, 46.2% and 41.5%, respectively. The MPR rate was 68.3%, 48.7% and 43.9%, respectively.
While in the PP analysis (completed 6 cycles of preoperative treatment), the pCR rates were 30.6%,
was significantly worse in MRD-negative cases with complications (80.7%) compared to
17.1%, and 10.8%, respectively. The downstaging was 63.8%, 45.7% and 37.8%, respectively. Safety those without complications (87.0%; HR 1.63; 95% CI 1.143–2.323; p = 0.007). In MRD-
assessments was generally well-tolerated. Conclusions: mFOLFOXIRI with AK104 demonstrated a positive cases, 3-year RFS was 15.6% in patients with complications versus 19.9% in
higher pCR rate, downstaging rate and MPR rtae compared with FOLFOX chemotherapy in patients with those without (HR 1.16; 95% CI 0.831–1.608; p = 0.389). Among patients with com-
LACRC. This study suggests that the combination of Intensified chemotherapy and dual immunotherapy plications, ctDNA testing conducted at 2–4 weeks post-surgery versus 4–10 weeks
may be a promising approach for improving treatment outcomes. Clinical trial information: showed marked differences in 3-year RFS based on landmark analysis at 10 weeks. For
NCT05571644. Research Sponsor: None. MRD-negative cases, 3-year RFS was 75.32% versus 90.71% (HR 2.875; 95% CI
Pathologic outcome and surgical parameters. 1.252–6.605; p = 0.013). In MRD-positive cases, 3-year RFS was 27.7% versus 6.25%
Characteristics mFOLFOXIRI+AK104 (A) mFOLFOXIRI (B) mFOLFOX6 (C) P1 (A vs. C) (HR 0.46; 95% CI 0.230-0.929; p = 0.030). This trend was similar even when colon and
mITT analysis n=41 n=39 n=41 rectal cancer were analyzed separately. In contrast, no timing-related differences were
pCR 11 (26.83%) 6 (15.38%) 4(9.76%) 0.046 observed in cases without complications. Conclusions: Perioperative complications
ypStage 0-I 27 (65.9%) 18 (46.2%) 17 (41.5%) 0.026 may elevate cfDNA levels, potentially confounding MRD assessment. Our findings
MPR 28 (68.3%) 19 (48.7%) 18 (43.9%) 0.026
PP analysis n=36 n=35 n=37 suggest delaying ctDNA testing to at least 4 weeks postoperatively in patients expe-
pCR 11 (30.6%) 6 (17.1%) 4 (10.8%) 0.036 riencing Clavien-Dindo Grade $2 complications. These results provide essential
ypStage 0-I 23 (63.8%) 16 (45.7%) 14 (37.8%) 0.026 guidance for optimizing clinical trial designs involving MRD evaluation through ctDNA
Footnote: pCR, pathological complete response; MPR, major pathological response. analysis. Clinical trial information: UMIN000039205. Research Sponsor: AMED.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
260s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3601 Poster Session 3602 Poster Session

Long-term survival and treatment efficacy in dMMR/MSI-H rectal cancer: A Phase 1b study to assess the safety of neoadjuvant trifluridine/tipiracil with
real-world cohort from seven large medical college–affiliated hospitals. First concurrent radiation in resectable stage II/III rectal cancer: Initial results of
Author: Siyuan Mi, Sun Yat-sen University Cancer Center, Guangzhou, China the FIERCE study. First Author: Emerson Yu-sheng Chen, Oregon Health & Science
Background: Neoadjuvant immunotherapy provides considerable advantages for pa- University, Portland, OR
tients with dMMR/MSI-H rectal cancer. However, treatment strategies in real-world Background: Total neoadjuvant therapy (TNT) with chemo-radiation (CRT) followed by
settings vary depending on tumor characteristics, economic conditions, and the choices doublet chemotherapy for rectal cancer can yield clinical complete response (CR) to allow
made by physicians and patients. Methods: We screened more than 10,000 rectal successful surgery or organ-sparing. Trifluridine exhibited cytotoxic effect of ionizing
cancer cases from seven large medical college-affiliated hospitals. We used the Kaplan- radiation superior to fluorouracil in colonogenic survival assay (dose modification factor:
Meier curve to compare survival and progression, applied Cox regression to analyze 2.7). The FIERCE trial (NCT04104139) sought to determine the maximum tolerated dose
impact factors, and examined tumor regression grades with chi-square analysis. (MTD) of trifluridine/tipiracil (FTD-TPI) with concurrent CRT during TNT with future goals of
Results: From March 2010 to April 2024, 502 patients were enrolled and diagnosed with improving CR. Methods: MRI-staged participants (pts) with stage II (T3-4N0M0) or stage III
dMMR/MSI-H rectal adenocarcinoma through immunohistochemistry or PCR. 100 (TxN+M0) resectable rectal adenocarcinoma, underwent CRT with FTD-TPI for 6 weeks
patients underwent neoadjuvant immunotherapy, demonstrating a 96.34% 5-year overall followed by either FOLFOX or CAPOX for 4 months. Cohorts of 3 pts were examined at three
survival (95% CI: 86.08-99.08%), and 90.74% 5-year disease-free survival (95% CI: 74.67- dose levels (DLs) until a total of 18 pts were reached per Bayesian Optimal Interval design.
96.82%). This indicated a 16.50% enhancement in overall survival (p = 0.042) and a FTD-TPI was taken orally BID for five days/week on weeks 1, 3, and 5 at the assigned dose
(DL1 = 25mg/m2; DL2 = 30 mg/m2; DL3 = 35 mg/m2) with concurrent pelvic radiation of 25-
16.87% increase in disease-free survival (p = 0.002) compared to conventional che-
28 fractions. Dose limiting toxicity (DLT) was defined by sustained hematologic, gastro-
moradiotherapy (5-year OS: 79.84%, 95% CI: 71.69-85.87%; 5-year DFS: 73.87%, 95% CI:
intestinal, and serious adverse events related to FTD-TPI. Primary endpoint was the pro-
65.15-80.73%). Neoadjuvant immunotherapy demonstrated significant superiority in
portion of DLT from CRT at MTD. Results: Among 22 screened patients (3 ineligible and 1
tumor regression (p , 0.0001), however, the combination with chemotherapy did not removed for non-compliance in week 1), 18 pts were evaluable. Median age was 52 years
enhance the effect (p = 0.622), and varying chemotherapeutic agents did not improve (range 37-73), female sex was 4 (22%), and 16 (89%) were ECOG 0. All pts had proficient
tumor regression in conventional chemoradiotherapy either. Additionally, elevated mismatch repair. All 18 pts were staged as cT3, with 8 (44%) N0 (stage II), 7 (39%) N1, and 3
serum CEA levels were associated with an increased risk of both death and disease (17%) N2. All 18 pts completed CRT in their assigned DL (6 DL1, 3 DL2, 9 DL3) followed by
progression. Patients with advanced age, lower clinical stages, and fewer risk factors chemotherapy (16 FOLFOX; 2 CAPOX) with the 18th pt scheduled to finish treatment in
were more likely to undergo direct surgical resection. Conclusions: Neoadjuvant im- February 2025. One pt was switched to irinotecan due to intolerable oxaliplatin-related
munotherapy is advantageous for dMMR/MSI-H rectal cancer patients, as it leads to neuropathy. During CRT, grade 3 neutropenia without fevers (4/18, 22%) led to 1/6 DLT in
better tumor regression and enhanced disease control. Research Sponsor: None. DL1 and 1/9 DLT in DL3, which were mitigated by dose interruptions (2/18 missed last week
of FTD-TPI). No grade 4 adverse events were observed during CRT. At data cutoff, 10 (56%)
pts entered into watch-and-wait surveillance, 6 (33%) pts had surgery, and 2 (11%) pts await
final assessment. Conclusions: FTD-TPI at DL3, the MTD of 35 mg/m2 on days 1-5, 15-19,
and 29-33, is the recommended phase 2 dose for CRT in TNT for locally-advanced rectal
cancer. Short-course filgrastim, or day 29-33 adjustment to DL2, may be needed in last
2 weeks of CRT to prevent dose interruptions. A dose expansion study is being explored as
the next step. Clinical trial information: NCT04104139. Research Sponsor: Taiho Oncology;
OHSU Knight Cancer Institute.

3603 Poster Session 3604 Poster Session

Enhancing colorectal cancer precision medicine through multi-omics and Complementary value of a digital pathology biomarker to post-surgery
clinical data integration with artificial intelligence. First Author: Enrique circulating tumor DNA in risk stratification of stage III colon cancer patients
Velazquez Villarreal, City of Hope National Medical Center, Duarte, CA receiving adjuvant chemotherapy. First Author: Ingrid Franken, Department of
Background: The integration of multi-omics and clinical data in precision medicine Medical Oncology, University Medical Center Utrecht, Utrecht University, Utrecht,
research for colorectal cancer (CRC) is a complex task that requires advanced compu- Netherlands
tational tools. Artificial Intelligence agent for High-Optimization and Precision mEdicine Background: The current standard of care for patients with stage III colon cancer (CC) is
(AI-HOPE) has emerged as a transformative platform, streamlining data integration, resection followed by adjuvant chemotherapy (ACT). About half of patients are cured by
analysis, and discovery efforts. AI-HOPE is designed to integrate and analyze multi-omics surgery and hence overtreated with ACT, whereas ~30% experience recurrence despite ACT.
alongside clinical data, facilitating novel insights into CRC pathogenesis, therapeutic Several studies show that patients with no detectable circulating tumor DNA (ctDNA) after
responses, and precision medicine applications. Methods: AI-HOPE leverages Large surgery are at a lower risk of recurrence (RR), although false negative ctDNA results remain a
Language Models (LLMs) to interpret natural language inputs and convert them into concern. Other studies show prognostic value of digital pathology biomarkers on resected
executable pipelines for multi-omics and clinical data analysis. Data from The Cancer CC tissue, like the Combined Analysis of Pathologists and Artificial Intelligence (CAPAI;
Genome Atlas (TCGA) and other publicly available CRC datasets were utilized to dem- Kleppe Lancet Oncol 2022). This study aimed to explore the potential added value of CAPAI
onstrate its capabilities. AI-HOPE supports analyses such as identifying mutation and gene to post-surgery ctDNA in risk stratification of patients with stage III CC receiving ACT.
expression patterns, pathway enrichment, survival analysis, and therapeutic outcome Methods: Patients were selected from the Prospective Dutch ColoRectal Cancer (PLCRC)
predictions. Three case studies were conducted: (1) identifying WNT and TGFb pathway cohort substudy PROVENC3 (Rubio-Alarcon AACR 2024), based on stage III CC treated with
alterations in early-onset CRC (EO CRC) versus late-onset CRC, (2) evaluating the as- radical resection and adjuvant capecitabine or CAPOX. Post-surgery ctDNA status was
sociation between specific multi-omics signatures and progression-free survival in pa- determined using Labcorp Plasma Detect. From the resected tumor, a representative H&E
tients treated with FOLFOX chemotherapy, and (3) performing a precision medicine query slide was digitalized to generate a DoMore-v1-CE-CRC score, which was combined with the
to identify actionable gene mutations and tailored medications for CRC treatment. pT and pN stage and number of assessed lymph nodes for classification as CAPAI high-,
Results: Overall AI-HOPE answered queries with an accuracy of 0.98 and an F1 score of intermediate- or low-risk. Three-year RR and Cox proportional hazard ratios (HR) were
0.89 (precision = 0.80). AI-HOPE identified significant alterations in the WNT and TGFb reported for ctDNA-based and CAPAI-based risk groups. Time to recurrence was compared
pathways among EO CRC patients compared to late-onset cases, aligning with findings between risk groups using the log-rank test. Results: Post-surgery ctDNA status and CAPAI
from published literature. In the second study, the platform revealed that patients with risk classification were available for 163 patients. The 20 patients (12%) with detectable
specific transcriptomic signatures (e.g., upregulation of MYC targets) had significantly ctDNA had a higher recurrence risk (3-year RR 60% [32-77], HR 4.9 [2.5-9.6], p , 0.001) than
worse progression-free survival when treated with FOLFOX chemotherapy, further sup- patients with no detectable ctDNA (N = 143, 3-year RR 18% [11-25]). Within the subgroup
porting its utility in identifying clinically relevant biomarkers. AI-HOPE was used to query with no detectable post-surgery ctDNA, 50 patients (35%) were classified as CAPAI high-
CRC datasets for actionable gene mutations, such as KRAS, BRAF, and MSI-H (micro- risk. These CAPAI high-risk patients had a higher recurrence risk (3-year RR 35% [20-48], HR
satellite instability-high), and cross-referenced these findings with drug databases to 4.2 [2.0-9.1], p , 0.001) than patients classified as CAPAI low/intermediate-risk, who were
identify tailored therapies. The analysis highlighted FDA-approved targeted treatments, combined based on their observed similar RR (N = 93, 3-year RR 9% [3-15%]).
such as EGFR inhibitors (cetuximab and panitumumab) for KRAS wild-type patients and Conclusions: In patients with stage III colon cancer treated with adjuvant CAP(OX),
immune checkpoint inhibitors (pembrolizumab and nivolumab) for MSI-H tumors. AI-HOPE CAPAI risk classification has potential to further stratify RR in the subgroup with no de-
also identified emerging therapeutic options from ongoing clinical trials, showcasing its tectable post-surgery ctDNA. These preliminary results suggest that CAPAI high-risk may
potential for guiding precision medicine strategies in CRC. Conclusions: This study help identify patients with false negative post-surgery ctDNA results. Over half of all patients
demonstrates the transformative potential of AI-HOPE in advancing CRC precision had both no detectable ctDNA and were CAPAI low/intermediate-risk. Given their low RR in
medicine research by seamlessly integrating multi-omics and clinical data. Research our preliminary results, future studies on larger patient cohorts should focus on the ability to
Sponsor: National Cancer Institute; National Cancer Institute; National Cancer Institute. combine biomarkers to select very low-risk patients and evaluate whether these patients
can potentially be spared ACT. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 261s

3605 Poster Session 3606 Poster Session

A novel active chromatin cell-free DNA (cfDNAac) assay for early detection Analysis of early onset colorectal cancer: Implications for research and
of colorectal cancer. First Author: Yue Wendy Zhang, Aqtual, Hayward, CA clinical practice. First Author: Heather Halperin, University of Calgary, Calgary, AB,
Background: Colorectal cancer (CRC) is the third most common cancer and the second Canada
leading cause of cancer-related deaths globally (Xi et al. 2021). In the U.S., it is estimated Background: Young onset colorectal cancer(YOCRC), diagnosed , 50 years old, presents distinct challenges, including
life transitions, and psychosocial needs. In Alberta, Canada an Adolescent and Young Adult(AYA) program was in-
that 152,810 new cases and 53,010 deaths will occur in 2024, with an increasing in- troduced at our tertiary cancer centres to address these unique needs through tailored resources and services. This study
cidence among younger adults (ACS, 2024). CRC is a multifactorial disease influenced by aimed to explore the YOCRC patients’ adjuvant care experiences and describe resource utilization prior to and after the
both genetic and environmental factors and typically progresses from normal epithelial AYA program was established. Methods: A retrospective review of YOCRC patients diagnosed with StageIII disease from
2012-2022 and referred to Cancer Care Alberta for adjuvant therapy was performed. This study evaluated the adjuvant
tissue to adenocarcinoma through a well-established sequence of molecular events. care experiences of YOCRC patients, focusing on demographics, tumor characteristics, toxicity, and psychosocial visits.
Despite advancements in treatment, early detection remains a key driver of improved Incidence rates over time was assessed using Joinpoint regression. The Edmonton Symptom Assessment System(ESAS)
survival outcomes. While current screening methods— such as fecal immunochemical scores collected during treatment visits were used to characterize symptomatic concerns. Analysis included descriptive
statistics and multivariate regression analysis. Results: Among 576 patients, the mean age at diagnosis was 42.9 years.
testing, fecal DNA testing, and colonoscopy— have reduced CRC mortality and morbidity, There was a positive trend in incidence of diagnoses per year(p , 0.0001), with an increase in diagnosis of 3.96%(CI 1.30-
low compliance continues to be a significant challenge. In this study, we developed a 7.13) per year in patients aged 40-49. Patients with stages IIIB disease were older at time of diagnosis compared to stage
proof-of-concept machine learning (ML) classifier leveraging active chromatin cell-free IIIC(p = 0.025). Higher stages had poorer outcomes with stage IIIA, IIIB and IIIC(p , 0.001), however median survival was
DNA (cfDNAac) signals in peripheral blood to differentiate CRC patients from healthy not reached for any stages due to a low death rate. Patients reported psychological related symptoms(well-being, sleep
problems and tiredness) more commonly than physical symptoms. Psychosocial visits increased after AYA program
individuals. Methods: Plasma samples from treatment-naive colorectal cancer (CRC) initiation with annual rates rising from 7.8 visits/year(pre-AYA program implementation) to 24 visits/year(post-AYA
patients (stages I–IV, n = 54), advanced adenoma patients (n = 14), and healthy vol- implementation). Further, patients who attended psychosocial visits were more likely to be younger(p , 0.001) and had
unteers (n = 40) were processed using Aqtual’s proprietary active chromatin capture higher overall mortality(p , 0.001). Of patients with an ESAS of $4, patients who were 40-49 were more likely to attend
psychosocial service appointments than those aged 20-39, odds ratio of 3.23(95% CI: 1.28-8.78). There was no difference
workflow to enrich regulatory-active chromatin cfDNA. Samples were split into a training comparing patients attending psychosocial visits or higher ESAS score with sex or stage. Conclusions: Stage III YOCRC
set, which included 31 early-stage CRC samples (stage I: n = 13, stage II: n = 13, stage III: patients are increasing in Alberta. An AYA program can improve resource utilization, psychosocial support, and enhance
n = 5) and 19 healthy samples, and a hold-out set, comprised 37 disease samples (stage visit frequency for this population with unique needs. Greater prominence of psychological symptoms on ESAS highlight
areas that could benefit from focused resources for this population. Research Sponsor: None.
I/II: n = 6, stage III: n = 6, stage IV: n = 11, advanced adenoma: n = 14) and 21 healthy
Demographics of stage III YOCRC patients.
samples. A machine learning classifier was trained using 5-fold cross-validation with 5
Demographics n=576
repeats on the training set, and the performance was assessed on the hold-out set to
ensure robustness and generalizability. Results: A machine learning classifier trained Sex
Male 320
on genome-wide cfDNAac signals demonstrated robust performance, achieving a mean Site
Female 256

AUC of 0.94 (95% CI: 0.92 - 0.96). It exhibited 93% sensitivity (95% CI: 84% -100%) for Colon 291
Rectum 245
colorectal cancer detection (Stage I/II: 94%, Stage III: 95%, Stage IV: 91%) and 81% Rectosigmoid 40
sensitivity (95% CI: 67% - 95%) for advanced adenoma, with a specificity of 90% in the Stage (AJCC 7)
IIIA 52
hold-out set. Analysis of the top contributing cfDNAac signals identified 852 promoter IIIB 348
IIIC 171
and 2,594 exon features derived from 2,678 unique genes. Gene-set enrichment analysis IIINOS 3
Incidence
revealed significant associations with key cancer hallmarks, including KRAS signaling Year
and epithelial-mesenchymal transition (EMT). Conclusions: This study highlights the 2012 46
2013 40
potential of Aqtual’s active chromatin capture assay to identify molecular features in 2014 41
2015 55
plasma that differentiate CRC and advanced adenoma from healthy individuals. The ML 2016 43
classifier based on these signatures shows promise for future development as a non- 2017
2018
50
50
invasive tool for early colorectal cancer and advanced adenoma detection and moni- 2019 61
2020 52
toring. Research Sponsor: None. 2021 64
2022 74

3607 Poster Session 3608 Poster Session

Changes in demographics and patterns of colon cancer in the United States: Time-weighted ctDNA dynamics for precision monitoring of relapse risk in
A comprehensive SEER analysis 2010–2021. First Author: Guy Loic colon cancer. First Author: Alessandro Leal, Perlmutter Cancer Center, NYU Langone
Nguefang Tchoukeu, Texas Tech University Health Science Center, Odessa, TX Health, New York, NY
Background: Over the past two decades, colon cancer has remained a leading cause of Background: Effective monitoring for relapse is a critical component of post-surgical care
cancer-related morbidity and mortality in the United States. Changes in population in colorectal cancer (CRC), particularly for patients who remain at risk of recurrence
dynamics, environment, and lifestyle have contributed to shifts in age, gender, tumor site, despite curative-intent treatment. Circulating tumor DNA (ctDNA) is a powerful biomarker
and stage at diagnosis. However, the extent of these changes over time remains for detecting minimal residual disease (MRD) and predicting relapse with high sensitivity
underexplored. Understanding these temporal changes is crucial for improving screening and specificity. However, its predictive value varies over time, with negative results closer
strategies, identifying at-risk populations, and optimizing resource allocation. to surgery being less reliable than results obtained later. Here we introduce a novel time-
Methods: We conducted a cohort study of initial colon cancer diagnoses using the weighted approach to ctDNA monitoring using a tumor-informed assay (Signatera),
National Cancer Institute, Surveillance Epidemiology and End Results Program (SEER) assigning greater predictive power to negative results collected further from surgery.
over the period 2010 through 2021. The SEER data is comprised of 22 different registries Methods: A Bayesian logistic regression model using 1,246 Signatera serial measure-
across the United States (US), which captures approximately 47.9% of the US population. ments was developed to predict recurrence risk across 167 patients with early-stage colon
The exposure was the year of diagnosis, and the outcomes were the distributions of age, cancer, with time-weighted ctDNA dynamics as the primary predictive feature. Negative
sex, site, and stage at initial diagnosis. Temporal trends in proportions were measured ctDNA values were assigned greater predictive power based on their temporal distance
using Spearman correlation (r). The differences between distributions in 2010 vs. 2021 from surgery. Secondary covariates included clinical stage and adjuvant treatment status.
were measured using Fisher’s test with the standardized mean difference (SMD) as effect Time-weighted ctDNA was calculated as the product of the ctDNA level at each timepoint
size. 95% confidence intervals (95% CI) for proportions were determined using the normal and an inverse time factor (1/(t+1)), where t represents the weeks since surgery. The
approximation. Results: A total of 670,923 records were included in the study. The weighted values were aggregated for each patient to compute cumulative and average
distribution of ages at initial diagnosis changed significantly over the study period (SMD = time-weighted ctDNA levels, which served as inputs to the model. Survival analysis was
0.1713, p-value , 0.0001). In 2010, the mode of age group distribution was $ 85 years, performed to evaluate recurrence-free survival (RFS), stratified by MRD status.
12.1% (11.8%–12.4%). The proportion aged $ 85 trended down over the period (r = Results: Tumor-informed ctDNA levels were measured longitudinally, with a median of 7
-0.9930, p value , 0.0001) from 12.1% (11.8%–12.4%) to 9.6% (9.3%–9.8%). In 2021 the timepoints per patient (range, 2-16) collected over a median follow-up of 2.5 years. Stage
mode of age group distribution at diagnosis was 65 to 69 years, 13.2% (12.9% - 13.5%). distribution was 50% stage III (n = 83), 44% stage II (n = 74), and 6% stage I (n = 10).
The proportion of males trended up over the period (r = +0.9021, p value , 0.0001), Mismatch repair deficiency was observed in 22 patients (13.2%). Adjuvant chemotherapy
increasing from 51.8% (51.4% - 52.2%) to 53.2% (52.8% - 53.6%). The difference in cancer was administered in 102 patients (61.1%), and 16 patients (9.6%) experienced recurrence.
stage was significant (SMD = 0.1818, p value , 0.0001), with localized colon cancer Survival analysis revealed a significantly worse recurrence-free survival for MRD-positive
decreasing (r = -0.9371, p value , 0.0001) from 38.9% (38.5% - 39.3%) to 33.8% (33.4% - patients compared to MRD-negative patients (HR = 4.2, 95%CI:2.8–6.4, Log-rank p ,
34.1%) and distant colon cancer increasing (r = +0.9301, p value , 0.0001) from 20.1% 0.0001). The Bayesian logistic regression model demonstrated robust predictive per-
(19.8% - 20.5%) to 22.9% (22.5% - 23.2%). The difference in location was significant (SMD formance, with a posterior probability of recurrence , 5% for patients with three con-
= 0.1185, p value , 0.0001), with rectum as location trending up (r = +0.9330, p value , secutive negative ctDNA results obtained . 6 months after surgery. Conversely, the model
0.0001) from 21.7% (21.3% to 22.0%) to 24.6% (24.3% to 25.0%). Conclusions: The assigned a . 90% probability of recurrence for patients with persistent ctDNA positivity
current findings suggest a significant shift in colon cancer presentation in the United beyond the initial 3-month post-operative window. Conclusions: Time-weighted ctDNA
States over the past decade. There is a significant increase in incidence among younger dynamics demonstrated promising predictive capability for CRC recurrence. Our findings
males, a trend towards more distant-stage cancers indicating delay in detection, and a suggest that incorporating the temporal context of ctDNA measurements and leveraging
growing prevalence of rectal cancers as the primary site. Changes in screening strategies the increasing reliability of negative results over time could refine risk stratification and
and reinforcement of community awareness are necessary to improve the mortality rate. improve personalized care strategies for CRC patients. Research Sponsor: None.
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
262s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3609 Poster Session 3610 Poster Session

Exploring transcriptomic regulation of the tissue-associated microbiome in Postoperative cfDNA levels and ctDNA detection rates in patients with stage
oncogenic progression of colorectal cancer. First Author: Amanda Stafford, II colon cancer screened for CIRCULATE (AIO-KRK-0217, ABCSG). First
BioCorteX Inc., New York, NY Author: Sebastian Stasik, Technical University Dresden, Dresden, Germany
Background: The association between the tissue microbiome and colorectal cancer (CRC) Background: Postoperative circulating tumor DNA (ctDNA) has emerged as a prognostic
etiology continues to be explored, and it has been proposed that certain bacteria drive biomarker for disease recurrence in patients (pts) with resected colorectal cancer (CRC)
oncogenesis. However, this field is hindered by a lack of mechanistic links. Considering the and may potentially guide adjuvant treatment decisions. The timing of ctDNA screening is
relationship between bacterial abundance and human gene expression allows us to critical, as postop. plasma cell-free DNA (cfDNA) levels may vary due to factors such as
traverse this functional gap. To date, work in this field has been largely restricted to tissue disruption from the surgical resection. The CIRCULATE trial (NCT04089631) in-
considering the relationship between bacteria and Consensus Molecular Subtypes (CMS) vestigates ctDNA guided adjuvant therapy in stage II CRC pts in . 140 centres in Germany
of CRC. Here, we hypothesise that CRC-associated bacteria correlate with distinct ex- and Austria. Methods: To investigate the impact of blood sampling time points on cfDNA
pression profiles that drive CRC progression. Methods: Analysis was performed using concentrations and the ctDNA positivity rates, we analyzed the postop. plasma samples of
BioCorteX’s knowledge graph and proprietary engines, v20250128_100953. 16S rRNA 1439 pts with stage II CRC screened for CIRCULATE between 2020 and 2024. Blood
sequencing and whole RNA data were collated from tissue samples across 3 independent samples were collected within 5-60 days post tumor resection in stabilizing tubes (Streck
cohorts from CRC patients (n = 59) and healthy volunteers (n = 23). In CRC patients, paired or PaxGene). Samples were analyzed for postop. cfDNA concentration and tumor-informed
tumour and normal adjacent samples were analysed yielding 108 samples in total. The ctDNA in plasma samples by an error-reduced Next-Generation Sequencing (NGS) ap-
Pearson correlation between differentially abundant (DA) OTUs and differentially proach [Stasik S Front Genet. 2022]. Results: Plasma cfDNA concentrations (measured by
expressed (DE) genes was calculated, and false discovery rates (FDR) were applied. qPCR for beta-globin gene) ranged from 0.02 to 20.32 ng/mL (mean: 0.689 ng/mL; median:
Results: 14,460 DE genes (FDR , 0.05) were identified between tumour and normal 0.360 ng/mL). The highest cfDNA levels were observed within 2 weeks after surgery (mean:
adjacent tissues, 5,097 of which had an absolute log-fold change greater than 1. Seven 1.079 ng/mL; median: 0.540 ng/mL), with a significant decrease in samples collected .
OTUs were increased in CRC compared to normal adjacent (FDR , 0.05) and were present 3 weeks postop. (mean: 0.631 ng/mL; median: 0.355 ng/mL; p, 0.0001), suggesting an
in at least 20% of samples; Fusobacterium nucleatum, Akkermansia muciniphila, and five impact of surgical trauma and subsequent cfDNA release from normal tissue. In ctDNA-
Streptococcus species. Of the 61,164 OTU-gene pairs analysed, 108 were significant neg. pts, cfDNA concentrations stabilized between 0.405 and 0.449 ng/mL during weeks 4-
(FRD , 0.05) with a Pearson correlation greater than 0.5. The abundance of all five 8. In contrast, ctDNA-pos. pts had significantly elevated cfDNA levels at 2 months post-
Streptococcus species was correlated with the expression of nine genes including: SPSB4 surgery (mean: 0.972 ng/mL; p= 0.419), indicating ongoing tumor-specific DNA shedding
(r= 0.51, p = 1.86x10-8), which has previously been associated with CRC metastasis, and associated with recurrent disease. Variant allele frequencies (VAFs) in ctDNA-pos samples
REN (r= 0.71, p = 1.21x10-17), which is associated with increased Wnt signaling. Moreover, were negatively correlated with cfDNA concentrations, particularly in early postop.
the expression of LY6G6D, a recently discovered CRC-specific antigen, was positively samples (Spearman r = -0.508), suggesting a dilution effect of ctDNA post-surgery. This
associated with F. nucleatum abundance (r= 0.53, p = 4.94x10-9). Conclusions: This study correlation diminished at later time points, supporting the potential advantage of later
identifies novel interactions between the CRC-associated microbiome and expression sampling to improve ctDNA sensitivity. Despite temporal variations in cfDNA concen-
profiles. The consistent relationships between five Streptococcus species and specific trations, ctDNA positivity was consistent across all sampling intervals (i.e. week 1: 4.1%;
genes is highly indicative of a functional co-evolution between host transcriptomics and week 6-8: 5.64%), demonstrating the assay’s robust sensitivity. Conclusions: We
microbiome, and implies relationships have less to do with the specific bacteria and are observed a significant variation in cfDNA levels depending on the timing of postop.
more related to function. This analysis adds a mechanistic explanation for some of the sampling. Different kinetics, such as cfDNA release from normal tissue and tumor
associations previously identified between the microbiome and CRC. Further analysis is shedding, may influence cfDNA levels and the sensitivity to ctDNA detection. Nonetheless,
required to identify if the host transcriptome creates a niche for these species or whether our assay demonstrates consistent and reliable sensitivity for ctDNA detection across all
the presence of these species is altering the microbiome. This study demonstrates the postop. sampling time points. Clinical trial information: NCT04089631. Research Sponsor:
value of integrated multi-omic analysis in understanding CRC pathogenesis. Research German Federal Ministry for Education and Research (Bundesministerium für Bildung und
Sponsor: BioCorteX Inc. Forschung) / DLR; 01KG1817.

3612 Poster Session 3613 Poster Session

Multicentric evaluation of an artificial intelligence model to stratify stage II Development and validation of machine learning risk prediction models for
colon cancer patients from whole slide images. First Author: Abdelhakim Khellaf, detection of early-onset colorectal cancer: Data from 30 health systems in
Centre Hospitalier de l’Université de Montréal (CHUM), Montreal, QC, Canada the United States. First Author: Wilson Lau, Truveta Inc, Bellevue, WA
Background: Stage II colon cancer (IIA T3N0; IIB T4aN0; IIC T4bN0) represents nearly Background: Incidence of early-onset colorectal cancer (EoCRC) in patients without any family
25% of all colon cancers and includes a wide range of outcomes (5-year overall survival history has been increasing in recent years. Our study leveraged advanced Large Language
rate (OS) of 58.4% to 87.5%). We aim to test if a deep-learning-based analysis of whole Models (LLM), like GPT-4, to predict EoCRC in a population comprised of multiple health systems
slide histology images (WSI) can predict survival and highlight the most relevant across the United States. There is potential to improve patient care by suggesting early
morphologic characteristics underlying prognosis. Methods: We identified adult stage II screening for patients who are predicted to be at risk by the model. Methods: We identified a
population of 5532 patients aged between 18 and 44 in the Truveta data, which is a collection of
colon cancer cases from the multimodal Cancer Genome Atlas (TCGA) and retrospective
120+ million patient journeys across 30 U.S. health systems. 1376 (24.87%) were diagnosed of
consecutive cases diagnosed between 2014-2019 (inclusive) from two independent CRC based on their ICD and SNOMED-CT codes. Data was split into training (80%) and testing
centers (CHUM, Canada and IHP, France; IRB approved). We tested on these two in- (20%) sets. For the prediction task, we applied GPT-4o and compared with XGBoost, one of the
dependent datasets, an artificial intelligence (AI) algorithm trained on TCGA, using a strongest non-generative machine learning models. We used patient demographics (age,
cross-validation and cross-testing (80:20) framework. The model relies on a weakly- gender, race, ethnicity), conditions, and lab results within the last 2 to 7 months prior to CRC
supervised attention-based pipeline that extracts survival driven histologic features diagnosis for model training. The last month before CRC diagnosis was excluded to avoid highly
from H&E WSI and assigns a risk score for each patient. The concordance index (c-index) predictive signals. For XGBoost, the demographics were represented as one-hot feature vectors,
was used as the primary outcome metric. Further testing of the survival score was indicating their presence or absence. Conditions were encoded by their diagnosis frequency
performed with a multivariate Cox regression model. The stratification of the cohorts within the time frame, while the actual values of the lab results were used as model features. For
based on the risk scores was evaluated using Kaplan-Meier curves and log-rank test. the LLM model, all patient information was input as plain text. Both conditions and lab results
95% confidence intervals (CI) are provided. An adjusted two-tailed P value ,0.05 was were represented by the names of ICD and LOINC codes in order to capture the clinical context. A
considered significant. The specific morphologic characteristics involved in the AI Chain-of-Thought prompting strategy, incorporating detailed instructions and CRC-specific
outcomes are under analysis. Results: The Discovery TCGA cohort consisted of n=463 knowledge, was employed to guide the LLM. Results: Our test set consisted of 1105 patients in
which 279 (25.25%) were diagnosed with CRC. Both XGBoost and GPT-4o achieved comparable
colon cancer patients; 5-year OS: 68.7% (CI: 60.0%-75.9%). The external validation
results. Despite the uneven distribution of CRC diagnoses in the test samples, the fine-tuned
cohorts included (1) from CHUM, n=124 patients, 5-year OS: 67.0% (CI: 58.0%-75.0%),
GPT-4o achieved the highest precision (87.43%) and recall (57.35%). This indicates that the
and (2) from IHP, n=123 patients, 5-year OS: 55.6% (CI 45.0%-67.0%). Cross-validation model can accurately predict EoCRC in the near future for over half of the patients.
and testing yielded a c-index of 0.72 and 0.68 respectively, 0.67 for CHUM and 0.65 for Conclusions: This study highlights the potential of using LLM to predict EoCRC in younger
IHP cohorts. After external validation, patients with a ’low-risk’ score showed signifi- population. While the GPT-4o base model contains general medical knowledge, supervised fine-
cantly higher 5-year OS than patients with a ’high-risk’ score: CHUM: 75.0% (CI: 64.0%- tuning with explicit guideline enhances its predictive capabilities. The high precision of the
84.0%) vs 53.0% (CI 38.0%-67.0%), P,0.05; and IHP 65.0% (CI: 44.0%-80.0%) vs 34% (CI: model performance minimizes the burden of unnecessary screening by identifying patients with
21.0%-46.0%), P,0.01. Cox regression showed a significant effect of WSI-based relatively high risk. As AI technologies continues to advance, with sufficient governance policy in
survival score on 5-year OS: TCGA cohort HR=8.3 (CI: 3.1-12.8), P,0.001; CHUM: place, predictive models can be valuable tools for clinicians to suggest early screening and
7.6 (CI 2.7-21.8), p,0.005; IHP: 5.5 (CI 2.1-16.8), P,0.005. Conclusions: AI-based risk mitigate EoCRC for patients. Research Sponsor: Truveta.
scoring for stage II colon cancer consistently correlated with 5-year OS across multiple F1
independent cohorts, achieving good performances. These findings highlight the po- Model Precision Recall score Accuracy
tential of modern computational pathology methods requiring minimal supervision to XGBoost 83.77% 57.35% 68.09% 86.43%
improve risk stratification of stage II colon cancer and patient care. Research Sponsor: GPT-4o (base) 68.61% 54.84% 60.69% 82.26%
IHP Group (France). GPT-4o (fine-tuned) 87.43% 57.35% 69.26% 87.15%

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 263s

3615 Poster Session 3616 Poster Session

Use and benefit of adjuvant chemotherapy in early- vs average-onset co- Detection of early-stage colorectal cancer using multiplexed extracellular
lorectal cancer. First Author: Carolina Bernabe, Montefiore Einstein Comprehensive vesicle proteomic biomarkers. First Author: Todd A. Hembrough, Nexosome On-
Cancer Center, Bronx, NY cology, Durham, NC
Background: Despite improved mortality rates in colorectal cancer (CRC), the incidence of Background: Colorectal cancer (CRC) is very common worldwide, and detection of early-
early onset-colorectal cancer (EO-CRC) has been rising in the United States. We aimed to stage cancer is critical for improved survival rates. Advanced adenoma (AA) and early-
explore trends in adjuvant chemotherapy (ACT) administration and evaluate its impact on stage CRC present challenges due to their small size and very low plasma expression of
overall survival (OS) and cancer-specific survival (CSS) in patients with EO-CRC (, 50 tumor-specific biomarkers. Genomics-based diagnostics struggle to detect these lesions,
years old) and average-onset colorectal cancer (AO-CRC) [. 50 years old]. Methods: Adult making it imperative to develop more sensitive approaches. Extracellular vesicles (EVs) are
patients (age .18) with Stage II-low risk (T1-T3 disease and more than 12 lymph nodes emerging as a promising solution for early-stage CRC detection. EVs are produced by
examined), stage II-high risk (T4 disease or less than 12 total lymph nodes examined) or tumor, tumor microenvironment and host cells, thus, unbiased analysis of all plasma EVs
stage III CRC diagnosed between 2010-2020 were identified in the Surveillance, Epide- offers an expanded set of cancer specific biomarkers, and may aid in detection of small
miology, and End Results Program (SEER) Database. Cox proportional hazard models were early-stage tumors. Methods: EVs were purified from patient plasma using size exclusion
used to assess the effect of ACT on OS and CSS in patients with EO-CRC and AO-CRC. chromatography and a proprietary buffer system that enhances EV and corona recovery.
Results: The final dataset included 8,998 patients with EO-CRC and 64,987 patients with TrueDiscovery Data-independent acquisition mass spectrometry (MS) analysis was
AO-CRC (Stage II-low =27,446, Stage II-high =9,060, and Stage III = 37,479). Patients with conducted on EVs purified from 24 advanced adenoma (AA)/stage 0 dysplasia, 25 Stage 1
AO-CRC were less likely to receive ACT compared to patients with EO-CRC across all CRC and 75 normal patient plasma samples. An in-house Machine Learning (ML) pipeline
stages: stage II-low (odds ratio [OR]: 0.27), stage II-high (OR 0.28), and stage III (OR: 0.32). was developed to identify differentially expressed proteins and to define protein multi-
From 2010 to 2020, the use of ACT decreased for EO-CRC (68% to 64%) but increased for plexes with extremely high Sens/Spec (. 0.99). Results: Around 2,500 QC’d proteins were
AO-CRC (36% to 45%). In patients with AO-CRC, ACT improved OS and CSS in stage II-high ID’d per sample and ML identified 336 (AA/Stage 0) and 493 (Stage 1) differentially
(OS Hazard ratio(HR]:0.48, 95%CI 0.42-0.54; CSS HR:0.48, 95%CI 0.42-0.54) and stage III expressed proteins (DEP: absolute Log2FC.0.5; q-value ,0.001; AUC.0.5). The DEPs
(OS HR:0.38, 95%CI 0.37-0.40; CSS HR:0.38, 95%CI 0.37-0.40), but had no benefit in CSS from AA/Stage 0 and Stage 1 were trained and tested using ML and SMV to identified
for stage II-low disease (CSS HR: 1.01, 95%CI 0.84-1.22). In patients with EO-CRC, ACT dozens of MS based multiplexes for each stage with near perfect diagnostic accuracy
improved OS (HR 0.70, 95% CI 0.56-0.88) and CSS (HR: 0.70, 95% CI 0.56-0.88) in stage III (~98-100%). These biomarkers were enriched with metabolic, immune and inflammatory
disease. No statistically significant survival benefit was observed for patients with stage II- proteins in line with our unbiased approach. MS protein multiplexes were then translated to
low (CSS HR:1.11, 95%CI 0.66-1.86) or stage II-high (CSS HR:0.83, 95%CI 0.53-1.31).The 5- ELISAs to build simple, high throughput assays. ELISA analysis of multiple biomarkers was
year survival probabilities by stage are shown in the table. Conclusions: Despite being performed in the training cohort, ML/SMV was used to define multiplex ELISA-based
used more frequently, the benefit of ACT in EO-CRC seems to be of less magnitude than in classifiers for Stage 1 CRC (Sen=0.95/Spec=0.96). This Stage 1 CRC training assay was
AO-CRC. These differences might be partially explained by the better survival outcomes of locked and tested on a blinded cohort of Stage 1 CRC (30 control, 30 cancer). In this blinded
EO-CRC even without ACT. Though recommended by most clinical guidelines, ACT does not cohort, we targeted maximal specificity yielding Sen=0.80, Spec=0.97 and AUC=0.95.
improve survival outcomes in patients with stage II high-risk EO-CRC. These findings These data represent significantly higher Sen/Spec than current genomic based liquid
highlight the need to improve risk stratification in early stage EO-CRC to better identify biopsy assays. Training and validation of AA and Stage 0 ELISAs is underway and will be
patients who may benefit from ACT. Research Sponsor: None. presented. Conclusions: Blinded validation of EV proteomic biomarkers yielded extremely
5-Year survival probabilities by stage (II-Low, II-High, III). high Sen/Spec early-stage CRC. The unbiased EV analysis underscores the need to assess
Variable OS II-Low OS II-High OS III CSS II-Low CSS II-High CSS III
cancer specific biomarkers from TME and host response and highlights the value and
opportunity of expanding liquid biopsy analysis beyond tumor specific genomics. Taken
AO; Chemo 84.77% 72.51% 69.66% 88.86% 76.96% 75.08% together this platform provides an exceptional opportunity to increase early-stage cancer
AO; No Chemo 74.11% 53.47% 41.54% 89.63% 70.93% 56.55%
EO; Chemo 93.53% 84.34% 78.72% 94.54% 86.34% 80.18% detection which should result in better patient outcomes. Research Sponsor: None.
EO; No Chemo 93.68% 80.19% 71.77% 95.32% 84.28% 75.84%

3617 Poster Session 3618 Poster Session

An early colorectal cancer screening programme based on bisulfite-free The impact of dietary patterns on inflammation and colon cancer risk: A
sequencing of methylated circulating tumor DNA. First Author: Rushan Fei, The retrospective study of 796 patients. First Author: Alina Zatsepina, Maimonides
First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China Medical Center, Brooklyn, NY
Background: Colorectal cancer (CRC) screening tests based on colonoscopy and fecal Background: Colon cancer significantly contributes to global cancer rates, with rising
DNA testing are limited due to their invasive nature and low compliance rates. Herein, we incidence linked to dietary and lifestyle changes. Diets high in ultra-processed foods
develop a non-invasive early screening test for CRC based on ColoSeer assay, which promote systemic inflammation, while anti-inflammatory diets may reduce inflammation
combines methylated DNA immunoprecipitation and bisulfite-free sequencing to an- and cancer risk. This study investigates the relationship between dietary patterns, in-
alyze methylation patterns of circulating tumor DNA (ctDNA). Methods: Tumor and flammatory biomarkers, and colon cancer risk. Methods: This retrospective study ana-
ctDNA methylation patterns in 1,336 subjects (430 healthy controls, 202 colorectal lyzed data from 796 patients with histologically confirmed colon cancer (2015–2023).
precursor lesions (CPLs), 704 CRCs) undergoing colonoscopy were evaluated using the Dietary habits were assessed using validated food frequency questionnaires (FFQs),
ColoSeer assay. Results: In the discovery phase, we identified a set of 46 differential categorizing participants by adherence to ultra-processed or anti-inflammatory diets using
methylation markers to screen for CPLs or CRCs from healthy controls. The overall the Dietary Inflammatory Index (DII). Systemic inflammation was evaluated via bio-
detection sensitivity of ColoSeer was 85.6% for CPL and 97.5% for TMN stage (0-II) CRC, markers: C-reactive protein (CRP), interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-
with 92.8% specificity. Conclusions: The ColoSeer assay demonstrated a high degree of a), and white blood cell (WBC) count. Clinical data, including age, sex, BMI, smoking,
accuracy for predicting subjects with CPL or early-stage malignant tumors and has physical activity, co-morbidities like metabolic syndrome and obesity, were extracted from
potential as a first-tier health screen for early detection of CRC. Research Sponsor: medical records. Multivariate logistic regression was used to assess associations between
None. dietary patterns, inflammation, and colon cancer risk. Subgroup analyses stratified by
metabolic conditions and cancer stage identified potential interactions. The study adhered
to STROBE guidelines. Results: Patients consuming diets high in ultra-processed foods
(top DII quartile) had elevated inflammatory biomarkers (CRP: 14.8 6 3.2 mg/L; IL-6: 8.6 6
2.4 pg/mL; TNF-a: 7.2 6 1.9 pg/mL) compared to those on anti-inflammatory diets (CRP:
4.3 6 1.1 mg/L; IL-6: 2.9 6 0.8 pg/mL; TNF-a: 2.3 6 0.7 pg/mL; p , 0.001). Dietary
patterns strongly correlated with colon cancer risk. Ultra-processed food consumers had
an adjusted odds ratio (aOR) of 2.47 (95% CI: 2.01–3.03) for colon cancer, while anti-
inflammatory diets had a protective effect (aOR: 0.62, 95% CI: 0.51–0.76). Subgroup
analysis showed anti-inflammatory diets’ protective effects were more pronounced in
patients with metabolic syndrome or obesity, reducing cancer risk by 45% (p , 0.001).
Early-stage colon cancer patients (Stage I–II) adhering to anti-inflammatory diets
exhibited lower inflammation, suggesting potential benefits for disease progression and
prognosis. Further analysis highlighted omega-3 fatty acids and polyphenols in anti-
inflammatory diets as key in reducing inflammation by downregulating NF-kB signaling
and pro-inflammatory cytokine production. Conclusions: Diets high in ultra-processed
foods elevate inflammation and colon cancer risk, while anti-inflammatory diets provide
protective benefits, especially in individuals with metabolic conditions. These findings
emphasize the need for dietary interventions promoting anti-inflammatory patterns in
cancer prevention and management. Further research should validate these results and
explore underlying mechanisms. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
264s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3619 Poster Session 3620 Poster Session

Risk of pre-cancerous advanced adenomas of the colon in long distance Reassessing colorectal cancer recurrence in solid organ transplant recip-
runners. First Author: Timothy Lewis Cannon, Inova Schar Cancer Institute, Fairfax, VA ients: Implications for revised management guidelines. First Author:
Background: Exercise induced gastrointestinal injury is believed to be associated with reduced blood David Bradley Meyer, Sidney Kimmel Comprehensive Center, Johns Hopkins University
flow to the intestines during long distance running. To our knowledge, there has not been evidence School of Medicine, Baltimore, MD
linking this type of exercise- induced bowel ischemia to carcinogenesis. After observing multiple Background: Patients with solid organ transplants (SOT) are at an increased risk of developing sec-
“ultramarathoners” present to our cancer center with advanced colorectal cancer, we initiated a ondary malignancies, including colorectal cancer (CRC). However, the influence of SOT and associated
prospective IRB-approved study to evaluate the risk of advanced adenomas (AA) in long distance immunosuppression on recurrence risk following definitive treatment is poorly understood. This study
runners between the ages of 35-50. Methods: NCT 05419531 was a prospective study of subjects characterizes the recurrence risk of CRC in a cohort of patients that developed CRC after SOT.
aged 35-50 years who had completed at least two registered ultramarathons (50 km or longer) or five Methods: Patients treated within the Johns Hopkins and University of Wisconsin medical systems
registered marathons (26.2 miles). Subjects were excluded if they were known or suspected to have meeting study criteria were identified using SlicerDicer Epic reports. Manual chart review was performed
inflammatory bowel disease, familial adenomatous polyposis (FAP), or Lynch Syndrome. Prior to to determine treatment course, patient/disease characteristics, and whether or not their disease recurred.
colonoscopy, each subject completed a questionnaire regarding dietary habits, bowel habits, and long- Results: Fifty-two patients with CRC following SOT were identified, including those who had liver (n = 12),
distance running history, with results to be reported in the future. All polyps discovered during kidney (n = 26), combined liver and kidney (n = 5), combined pancreas and kidney (n=4), lung (n = 4), and
colonoscopy were reviewed by a panel of gastroenterologists, pathologists, and oncologists to de- heart (n = 1) transplants. Among these patients, 30 were male and 22 were female. Ten patients had
termine if they met the criteria for advanced adenomas, defined as lesions .10 mm, .25% tubu- recurrent CRC after curative intent treatment. Recurrence rates were as follows: 1 of 10 (10.00%) patients
lovillous features, or high-grade dysplasia. Results: Between October 2022 and December 2024, 102 with stage I CRC, 5 of 12 (41.67%) patients with stage II CRC, and 4 of 18 (22.22%) patients with stage III
subjects were screened, and 100 underwent colonoscopy as part of the study. The median age was CRC. The overall CRC recurrence rate in the cohort was 32.26%. Conclusions: In summary, a multi-
42.5 years; 55 of the participants were female and 45 were male. The historical benchmark used for institution retrospective cohort study of patients with CRC diagnosis following SOT demonstrated higher
expected AAs in average-risk individuals aged 40-49 years was 1.2%. Among the 100 subjects in this than traditionally expected recurrence rates in patients with stage I and II disease when compared to
study, 15% (95% exact confidence interval: 7.9%- 22.4%) had confirmed AAs. 39 out of 100 subjects patients diagnosed with CRC without history of SOT. Further multi-institutional studies are warranted to
had at least one adenoma. Three additional subjects had three or more adenomas but did not meet our validate this finding. If confirmed, this would suggest transplant status may alter adjuvant treatment
predefined criteria for AA and were not included among the 15 patients with AA. Conclusions: NCT decisions and warrant future prospective studies in this patient population. Research Sponsor: None.
05419531 achieved its predefined endpoint for advanced adenomas, suggesting that “intensive” long Study cohort variables stratified by CRC staging.
distance running is a risk factor for advanced adenomas of the colon. Consideration of refined Avg. Time from Avg. Time to Avg.
screening strategies for this population is warranted. Future pathological and epidemiological Avg. Age Transplant to CRC Recur- Overall
Stage at at Time of CRC Diagnosis Recurrence rence Survival
evaluations should explore causation and ancillary risk factors in this unique population. Clinical trial Diagnosis Diagnosis Sex Race Transplant Type (months) Rate (months) (months)
information: NCT05419531. Research Sponsor: Inova Schar Cancer Institute.
Stage 0 72.00 1 1 Caucasian 1 Liver 34.73 0.00% N/A 4.70
Subjects with advanced adenoma. Female
Stage I 57.50 6 Male, 3 African American, 5 Kidney, 2 Liver, 2 159.73 10.00% 51.78 54.42
Participant ID 3 9 15 16 37 53 55 67 69 71 77 87 91 98 100 4 6 Caucasian, 1 Pa- Pancreas/Kidney, 1
Gender F M M F F M F F F F F M M F M Female cific Islander Lung
Stage II 56.33 7 Male, 3 African American, 7 Kidney, 1 Liver, 2 172.35 41.67% 29.87 61.34
Age 44 45 48 41 45 49 40 40 45 39 42 50 42 44 41
5 9 Caucasian Liver/Kidney, 2 Pan-
Endurance Eligibility (U – 4-6U 4- . 4 - . 7 - 6 5 7 - 1 U / . . 15 10 . 5 M Female creas/Kidney
Ultramarathons; M – 6U 15 6 U 15 15 M M 15 7-8 M 15 U M 15 Stage III 62.39 12 3 African American, 9 Kidney, 4 Liver, 3 148.76 22.22% 18.03 52.15
Marathons) U U U U U U Male, 6 14 Caucasian, 1 Liver/Kidney, 1 Lung,
# of Polyps 3 on primary; 2 5 2 3 1 1 3 2 7 1 2 4 2 6 3 Female Asian 1 Heart
on secondary Stage IV 52.18 5 Male, 2 African American, 5 Kidney, 4 Liver, 2 148.83 N/A N/A 18.70
Size of largest polyp 25 / 15 12 10 10 3 / 20 / 5-6 12/ 12 / 5 / 5 5 / 15- 6 / 10 10 / 6 9 Caucasian Lung
(Colonoscopy/Pathology) / / / 3 9 / 5 11 12 5 20 / 7 / 7 19 Female
(mm) 15 13 10 12 All 58.08 30 11 African Ameri- 26 Kidney, 12 Liver, 5 154.14 32.26% 27.33 46.72
>25% Tubulovillous Features N N N N Y N Y N N Y Y N N Y N Stages Male, can, 39 Caucasian, Liver/Kidney, 4
(Y/N) 22 1 Asian, 1 Pacific Pancreas/Kidney, 4
Female Islander Lung, 1 Heart
High-grade Dysplasia (Y/N) N N N N N N N N N N N N Y N N

3621 Poster Session 3622 Poster Session

Impact of COVID-19 on the stage at diagnosis and mortality of colorectal Leveraging EHRs and a phenome-wide association study to identify pre-
cancer. First Author: Soyun Lim, Seoul National University College of Medicine, Seoul, diagnostic clinical markers in early-onset colorectal cancer. First Author:
South Korea Sachleen Kaur Tuteja, Center for Health Information Partnerships, Institute for Public
Background: When the COVID-19 pandemic started in 2020 in the US, the Centers for Medicare Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL
and Medicaid Services (CMS) recommended delaying all non-urgent procedures, including Background: The incidence of colorectal cancer (CRC) has undergone a significant de-
screening colonoscopies to prevent virus spread. An analysis of the CMS claims data showed a mographic shift, with diagnosis in individuals under 55 years escalating from 11% in 1995 to
decrease in all screening modalities for colorectal cancer (CRC) including colonoscopy, fecal 20% in 2019 (Siegal et al., 2019). Traditional risk factors, such as obesity, diabetes, and
occult blood test, and blood-based gene analysis in 2020 compared to 2013-2019. We inves- colon microbiome changes, fail to fully explain this emerging clinical phenomenon, ne-
tigated the impact of COVID-19 on stage at diagnosis and mortality of CRC across racial groups. cessitating new approaches to identify pre-diagnostic markers in younger patients. Our
Methods: We utilized the Surveillance, Epidemiology, and End Results (SEER) database to obtain study leverages electronic health records (EHR) and employs a Phenome-Wide Association
the age-adjusted incidence of localized, regional, and distant CRC at the time of diagnosis from
Study (PheWAS) to characterize age-specific manifestations prior to diagnosis. Methods: A
2013 to 2021. The year-over-year (YOY) change in the proportion of advanced (regional + distant)
PheWAS was conducted with a cohort of 2,799 CRC patients from Northwestern Medicine
CRC was calculated. YOY changes in age-adjusted CRC mortality rates from 2014 to 2021 were
also calculated. Results: From 2014 to 2019, the average YOY change in the proportion of
(NM, 2012 – 2022). ICD-9/10 codes were categorized into 527 phenotypes using the
advanced CRC across all races was 1.0% whereas in 2020, the rate was 3.9%. The increase in PheCode framework. Logistic regression models were employed to examine associations
proportion of advanced CRC in 2020 was greatest in non-Hispanic Asian/Pacific Islanders between age groups ( , 55 vs. $55 years) and each phenotype during the 6 months
(NHAPI), followed by Hispanics, non-Hispanic Whites (NHW), non-Hispanic American Indian/ preceding diagnosis. A control cohort of 294,590 non-cancer NM patients adjusted for age-
Alaskan Natives (NHAIAN), and non-Hispanic Blacks (NHB; 7.9%, 4.7%, 3.6%, 3.3%, and 3.3%, related and baseline disease patterns. Results: Our approach revealed significant asso-
respectively). From 2014 to 2019, the average YOY change in CRC mortality across all races was ciations with age in 9 of 527 analyzed phenotypes, identifying well-documented CRC
-0.7% while in 2020, it increased by 5.4%. In 2020, the YOY change in mortality was greatest in features and less recognized clinical markers. Younger patient cohorts exhibited elevated
NHAIAN, followed by Hispanics, NHB, NHW, and NHAPI (25.3%, 16.5%, 15.1%, 3.2%, and 1.2%, probabilities of critical clinical indicators, including lower gastrointestinal hemorrhage (22%
respectively). Conclusions: The proportion of advanced CRC increased in 2020 likely due to [95% CI: 19-25%] vs. 11% [95% CI: 10-13%]; p = 1.0 x 10-11), hepatic dysfunction (6% [95% CI:
decreased screening most significantly in NHAPI. Also, mortality increased in 2020 possibly due 4-8%] vs. 4% [95% CI: 3-4%]; p = 2.7 x 10-6), and diverticular disease (4% [95% CI: 3-6%] vs. 3%
to delay in access to care during COVID-19. Further studies are needed to evaluate the long-term [95% CI: 2-3%]; p = 2.4 x 10-11) compared to older patients. Furthermore, an analysis of 1,866
effect of COVID-19 on the screening and mortality of CRC. Research Sponsor: None. NM pathology reports revealed single-institution trends consistent with global CRC liter-
Proportion of advanced colorectal cancer and mortality rates of colorectal cancer from 2013 to ature. Notably, younger patients showed higher prevalence of hereditary CRC (28% vs. 24%;
2021 by race. x2 p = 6.3 x 10-8), greater left-sided tumor localization (51% vs. 45%; x2 p = 0), and increased
Proportion of advanced colorectal cancer (%) 2013 2014 2015 2016 2017 2018 2019 2020 2021 proximal tumors (58% vs. 53%; x2 p = 0), corroborating studies that show a rise in proximal
Total 59.9 59.7 59.8 63.1 63.8 62.6 63.3 65.8 64.6
tumor incidence in patients under 50 and a decline in those aged 50–79. Conclusions: Our
Non-Hispanic White 59.9 59.6 59.6 62.2 63.3 62.3 62.9 65.1 64.4 PheWAS study uncovered clinicopathological distinctions, revealing statistically significant
Non-Hispanic Black 60.7 60.8 61.2 64.6 65.1 64.1 65.2 67.3 64.8 variations in pre-diagnostic clinical markers that suggest a more aggressive and complex
Non-Hispanic American Indian/Alaska native 60.9 62.6 59.0 64.2 64.5 65.0 66.3 68.5 65.3 disease progression mechanism in early-onset CRC. The increased probabilities of clinical
Non-Hispanic Asian/Pacific 59.0 59.2 59.1 64.8 66.4 61.6 63.2 68.1 64.1
Hispanic 61.6 61.2 61.0 65.3 64.8 64.2 65.0 68.1 66.7 indicators—such as gastrointestinal hemorrhage, hepatic dysfunction, and diverticular
Mortality (per 100,000) disease—underscore the need for tailored diagnostic and screening strategies for younger
Total 24.9 24.6 24.7 24.3 24.6 24.5 24.0 25.3 25.0 populations. These results also highlight the value of utilizing phenotype-based method-
Non-Hispanic White 25.5 25.5 25.6 25.1 25.6 25.6 25.1 25.9 26.2 ologies to identify nuanced clinical features that may enable earlier detection and improve
Non-Hispanic Black 33.3 31.5 32.7 31.7 31.8 31.0 30.4 35.0 32.0
Non-Hispanic American Indian/Alaska native 31.3 23.0 23.1 29.6 28.7 31.7 24.9 31.2 31.7 outcomes for younger CRC patients. Towards this goal, ongoing research aims to further
Non-Hispanic Asian/Pacific 16.9 17.1 17.1 16.9 17.3 16.3 16.6 16.8 17.0 delineate CRC clinical profiles between age groups. Research Sponsor: Northwestern
Hispanic 19.9 19.1 19.1 20.3 18.9 20.0 18.8 21.9 20.9 University Office of Undergraduate Research.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 265s

3623 Poster Session 3624 Poster Session

Colorectal cancer screening and detection following USPSTF recommen- Sex differences in chemotherapy completion and adverse events among
dations for ages 45-49: Insights from a large EHR cohort. First Author: patients with colon cancer (CALGB/SWOG 80702) (Alliance). First Author: En
Patricia Jane Rodriguez, Truveta Inc, Bellevue, WA Cheng, Department of Epidemiology and Population Health, Albert Einstein College of
Background: Rising colorectal cancer (CRC) incidence in relatively younger adults Medicine, Bronx, NY
prompted the US Preventative Services Task Force (USPSTF) to recommend screening in Background: Increasing chemotherapy completion and reducing chemotherapy adverse
average-risk adults aged 45-49 in May 2021. While screening has increased for ages 45-49, events (AE) are critical to improve survival after colon cancer diagnosis. Sex, as a biological
differences in polyp and CRC detection by age group are unclear. Methods: Using a subset variable, may impact chemotherapy treatment differently, and thus further investigation is
of Truveta Data, we identified patients aged 40-64, with no prior history of CRC, who needed to examine sex differences in chemotherapy completion and adverse events
underwent CRC screening procedures (colonoscopy or sigmoidoscopy) between 2018 and among patients with colon cancer. Methods: Among an NCI-sponsored trial conducted
2024, and for whom an associated report was available. Truveta Data contains de-identified among patients with stage III colon cancer (CALGB/SWOG 80702), all patients received
electronic health record (EHR) data from a collective of US health care systems. Screening standard adjuvant chemotherapy FOLFOX (fluorouracil, leucovorin, and oxaliplatin). To
outcomes were extracted from pathology reports using natural language processing, in- signal chemotherapy completion, we utilized relative dose intensity (RDI) calculated as the
cluding findings of polyps (tubular adenoma/adenomatous polyp, hyperplastic polyp, ratio of the delivered dose intensity to planned dose intensity; and reduced RDI (RDI ,85%)
tubulovillous adenoma, sessile serrated adenoma, traditional sessile adenoma, juvenile was considered as a clinically significant deviation from standard FOLFOX. From clinicians’
polyp, and inflammatory polyp), adenocarcinoma, and benign/normal colonic mucosa. For records of NCI’s Common Terminology Criteria for Adverse Events (CTCAE), we primarily
patients with multiple screenings, the first was used. Family history of CRC and related focused on clinically significant AE such as neutrophils decrease, nausea, platelets de-
syndromes (e.g., Lynch, Peutz-Jeghers) were similarly extracted from notes. Patient crease, hypertension, peripheral neuropathy, diarrhea, fatigue, gastritis, creatinine in-
characteristics and screening outcomes were described by 5-year age band, before and after crease, gastric ulcer, myocardial ischemia, and cerebral ischemia; and severe AE was
May 2021 USPSTF recommendations. Results: Of85,062 patients aged 40-64 who un- defined as the occurrence of any above AE with CTCAE grade $3. Using multivariable
derwent CRC screening, 46,168 had first screening records in May 2021 or earlier (‘pre logistic regression that adjusted for body surface area (BSA) and other clinicopathological
period’) and 38,894 after (‘post period’). In the pre period, 4,270 (9%) patients were aged 45- confounders, we estimated adjusted odds ratios (OR) for the associations of sex with
49, compared to 7,403 (19%) in the post period. Sex balance improved for the 45-49 age reduced RDI and severe AE. Results: Of 2201 patients, mean (standard deviation [SD]) age
group (56% were female pre vs. 50% post; p , 0.05), compared to younger and older age was 60.9 (10.9) years, 1019 (46.3%) were female, 1750 (79.5%) were White, 172 (7.8%)
groups (40-44: 57% vs. 57%, 50-54: 48% vs. 48%, 55-59: 51% vs.49%, 60-64: 50% vs. 49%; , were Hispanic, 964 (43.8%) experienced reduced RDI, and 1156 (52.5%) had severe adverse
0.05 for 55-59 only). Adenomatous polyp findings increased for all age groups, but to the events. Compared to males, females were at significantly higher risks of experiencing
greatest degree for the 45-49 group (45-49: 45.7% pre vs. 53.4% post; p , 0.01) compared to reduced RDI (OR [95% CI]: 1.57 [1.27-1.93], P ,0.001) and severe AE (OR [95% CI]: 1.73
other groups (40-44: 33.1% vs. 35.9%, 50-54: 58.3% vs. 59.9%, 55-59: 60.4% vs. 62.5%, 60- [1.41-2.12], P ,0.001). Conclusions: Our findings suggested females are more likely than
64: 63.2% vs. 64.5%; all p , 0.05). In contrast, adenocarcinoma findings occurred less men to experience reduced RDI and severe AE during colon cancer chemotherapy. Clinical
frequently for ages 45-49 (2.1% pre vs. 1.0% post, p , 0.01), while remaining stable in both Impact: In the era of precision medicine, sex (as a biological variable) should be considered
younger and older groups (40-44: 1.6% vs. 1.4%, 50-54: 1.2% vs. 1.1%, 55-59: 1.4% vs. 1.6%, in optimizing colon cancer chemotherapy to improve completion and reduce toxicities.
60-64: 1.3% vs. 1.3%; all p . 0.05). Conclusions: Following 2021 USPSTF CRC screening Clinical trial information: NCT01150045. Research Sponsor: National Cancer Institute;
recommendations for average-risk adults aged 45-49, sex balance in screening increased, U10CA180821; National Cancer Institute; U10CA180882; National Cancer Institute;
findings of adenomatous polyps increased, and findings of adenocarcinoma declined in this U24CA196171; National Cancer Institute; U10CA180863; Canadian Cancer Society;
group. These results may reflect expected decreases in adenocarcinoma detection with CCS707213; National Cancer Institute; UG1CA233234; National Cancer Institute;
expansion of screening to average-risk adults, or suggest that recommendations con- U10CA180820; National Cancer Institute; U10CA180868; National Cancer Institute;
tributed to earlier detection of pre-cancerous polyps, before progression to adenocarci- U10CA180888; Pfizer; [Link] Health and Environ-
noma. Additional studies are needed to explore complex causal relationships. Research mental Sciences Institute.
Sponsor: None.

3625 Poster Session 3626 Poster Session

Comprehensive analysis of outcomes of patients with colorectal cancer in a Dietary risk factors associated with colorectal polyp risk and potential for
racially diverse cohort. First Author: Manasawee Tanariyakul, University of Hawaii progression to malignancy. First Author: Melissa Rachel Goldin, Vanderbilt Uni-
Internal Medicine Residency Program, Honolulu, HI versity School of Medicine, Nashville, TN
Background: Disparate outcomes have been reported among patients with colorectal cancer (CRC) as those with Background: Colorectal polyp development and progression to cancer are associated
Medicare or Medicaid insurance and those of Native Hawaiian ancestry experience inferior survival. However, many CRC
studies addressing Native Hawaiians and other Pacific Islanders and insurance status have not adjusted for a com- with diet and microbial dysbiosis. Red and processed meat intakes have been associated
prehensive range of confounding variables. In this study, we perform a comprehensive analysis of CRC outcomes with colorectal polyps, but evidence on probiotic and prebiotic usage is inconclusive.
adjusted for clinicopathological and socioeconomic factors. Methods: Data were abstracted for patients diagnosed with This study examines the association of dietary factors with colorectal adenomas (ADs)
colorectal cancer between 2000 and 2022 in Hawaii. Overall survival of Asians, Whites, and Native Hawaiian or Other
Pacific Islanders (NHOPI) was assessed using Cox proportional hazards regression models adjusting for clinical,
and sessile serrated polyps (SSPs) of low and high potential of progression to cancer.
pathological, and sociodemographic factors. Results: A total of 3275 patients were included in the final analysis. NHOPI Methods: In the Colorectal Molecular Atlas Project, 1082 participants underwent a
were more likely to be younger (p,0.001), had a higher proportion of stable microsatellite status tumors(p=0.049) and colonoscopy, and completed a questionnaire. Individuals were classified as polyp-free
were more likely to have Medicaid insurance or be uninsured(p,0.001) compared to Whites and Asians. NHOPI patients
more often had high grade cancers compared to White and Asian groups. Medicaid or no insurance was associated with
controls, and polyp cases with high potential for progression to cancer ($ 1 cm, villous
significantly higher mortality compared to private insurance in both univariate(HR: 1.373, 95% CI: 1.202–1.568, p , component, high-grade dysplasia, or multiple AD and/or SSPs) or low potential (all other
0.001) and multivariate analysis(HR: 1.505, 95% CI: 1.309–1.729, p , 0.001). However, while Medicare showed higher AD or SSP cases). Frequency of dietary intake derived from questionnaire responses was
mortality in univariate analysis(HR: 1.409, 95% CI: 1.245–1.594, p , 0.001), this association normalized after divided into quartiles (total meat, total red meat, total processed meat, chicken, and fish)
adjustment(HR: 1.088, 95% CI: 0.958–1.237, p = 0.195) (Table). NHOPI had significantly higher mortality compared to
Whites after adjustment for variables(HR: 1.293, 95% CI: 1.107–1.510, p = 0.001). Interestingly, rectal tumors dem- or no intake/tertiles (total probiotic food: yogurt, kefir, sauerkraut, kimchi, fermented
onstrated higher mortality compared to right-sided tumors(HR: 1.318, 95% CI: 1.156–1.503, p , 0.001), while left-sided cabbage, kombucha, other fermented beverages). Multinomial logistic regression
tumors showed no significant difference (Table). Conclusions: This study of a racially diverse population that NHOPI models were used to derive odds ratios and 95%CI after adjusting for confounders.
patients with CRC had significantly worse mortality compared to Whites after adjusting for clinicopathological and
socioeconomic variables. We did not identify a significant association between Medicare insurance and increased Results: Highest intake of total meat was associated with increased risk of high
mortality. Having Medicaid or being uninsured was significantly associated with worse survival outcomes, emphasizing progression potential polyps (OR 1.63, 95%CI 1.03-2.60) and sessile serrated lesions (OR
the critical impact of healthcare access on survival. Research Sponsor: None. 1.60, 95%CI 1.03-2.50) in comparison to lowest intake. Highest intake of red meat was
Univariate and multivariate analysis. associated with increased risk of high progression potential polyps (OR 1.80, 95%CI
Univariate analysis
Hazard ratio (95% CI), p-value
Multivariate analysis
Hazard ratio (95% CI), p-value
1.16-2.80) and SSPs (OR 1.68, 95%CI 1.10-2.56) in comparison to lowest intake. The
second quartile of fish intake was significantly associated with a reduced risk of polyps
Private insurance
Medicaid and Uninsured 1.373 (1.202 - 1.568), ,0.001 1.505 (1.309 - 1.729), ,0.001 with low progression potential (OR 0.63, 95%CI 0.44-0.99) and ADs (OR 0.63, 95%CI 0.42-
Medicare 1.409 (1.245 - 1.594), ,0.001 1.088 (0.958 - 1.237), 0.195
White 0.96) compared to lowest intake; however, the third and fourth quartiles were not
Asian
NHOPI
0.947 (0.844 - 1.063), 0.354
1.143 (0.982 - 1.33), 0.084
0.908 (0.808 - 1.02), 0.103
1.293 (1.107-1.51), 0.001
significantly associated with risk. Chicken, processed meat, probiotic food, probiotic
Age at diagnosis 1.032 (1.028 - 1.036), ,0.001 1.044 (1.039 - 1.048), ,0.001 supplement, and prebiotic supplement intake were not associated with risk.
Grade 1&2
Grade 3&4 1.786 (1.586 - 2.012), ,0.001 1.485 (1.313 - 1.681), ,0.001 Conclusions: Dietary risk factors may vary by type of polyp, thus, dietary recom-
Stage 1
Stage 2 1.466 (1.249 - 1.722), ,0.001 1.401 (1.192 - 1.647), ,0.001
mendations should reflect these differences. Increased total and red meat intake may
Stage 3 1.644 (1.41 - 1.918), ,0.001 1.64 (1.403 - 1.916), ,0.001 increase risk of polyps with high progression potential and sessile serrated lesions, and
Stage 4 6.943 (5.898 - 8.173), ,0.001 7.681 (6.495 - 9.082), ,0.001
Unknown Stage 3.063 (2.525 - 3.714), ,0.001 2.99 (2.455 - 3.641), ,0.001 fish intake may reduce risk for polyps with low progression potential and adenomas.
Female 0.853 (0.775 - 0.939), 0.001 0.844 (0.765 - 0.931), ,0.001
MSI Stable
Probiotics and prebiotics may not be an effective risk reduction strategy. Research
MSI Unstable 1.146 (0.808 - 1.626), 0.444 1.015 (0.71 - 1.449), 0.936 Sponsor: None.
Unknown MSI 1.197 (1.055 - 1.357), 0.005 1.142 (1.003 - 1.3), 0.044
Right sided
Left sided 0.899 (0.802 - 1.007), 0.066 1.081 (0.961 - 1.216), 0.197
Rectal 1.016 (0.9 - 1.147), 0.797 1.318 (1.156 - 1.503), ,0.001
Time to treatment <31 days
Time to treatment >31 days 0.674 (0.593 - 0.766), ,0.001 1.064 (0.957 - 1.183), 0.09

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
266s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

3627 Poster Session 3628 Poster Session

Characterizing clinical outcomes and DNA co-alterations of ERBB2- Effects of marital status on survival in patients with colorectal cancer. First
amplified colorectal cancer. First Author: Mohamed Nuh, Baylor College of Medi- Author: Lisa Liu, George Washington University, Washington, DC
cine, Houston, TX Background: Demographic factors such as marital status have been shown to have
Background: A subset of colorectal cancer (CRC) features amplification (amp) of the ERBB2 differential effects on cancer outcomes. It has been shown that married patients have
gene. The prognostic role and treatment in these patients are poorly understood. This study aims higher 5-year survival rates compared with unmarried patients with colorectal cancer
to characterize the clinical outcomes and molecular profiles of ERBB2-amp CRC. Methods: This (CRC). However, most of these studies looked at patients diagnosed prior to 2015. This
is a retrospective analysis of clinical outcomes data and next-generation sequencing at MD study compares the survival trends between 2000–2010 and 2011–2021 and analyzes
Anderson from 2006-2025 in patients with ERBB2-amp CRC including all classes of genomic
marital status associations with stage at diagnosis. Methods: Patients with CRC di-
alterations (GA) in other genes. Progression-free survival (PFS) and overall survival (OS) were
assessed by Kaplan-Meier, log-rank, and cox regression. Chi square goodness of fit was assessed
agnosed between 2000 and 2021 were identified in the SEER database and stratified by
for the proportion of left versus right sided CRC. Results: 100 pts with ERBB2-amp CRC were marital status (single, separated/divorced/widowed (SDW), and married). 5-year survival
identified. Patients (Table) who received ERBB2-directed therapy (n = 36) did not have signif- rates with 95% confidence intervals were calculated for each marital status group across
icantly longer OS compared to those who received other systemic therapy (n = 57) (53.8 vs. 45.5 three periods: 2000–2010, 2011–2021, and 2000–2021. Kaplan-Meier survival analysis
m, p = 0.35). Patients who underwent surgery (n = 64) (56.8 vs. 23.2 m, p = 0.01) or non-surgical was used to compare survival outcomes across these variables. Disease severity was
localized therapy (n = 45) had longer OS (64.4 vs 45.5 m, p = 0.01). ERBB2-amp was more classified into three stages: localized, regional, and distant. The distribution of patients
associated with left sided CRC (n = 80) than right-sided CRC (n = 19) (X2= 39.34, df = 1, p , across disease stages is reported as the proportion of individuals within each marital
0.0001) but there was no significant OS difference (53.8 vs. 45.5 m, p = 0.76). Median PFS (mPFS) status group at each cancer stage. Results: Overall survival was statistically different
for all 1st (n = 93), 2nd (n = 80), and 3rd line (n = 61) systemic therapy was 8.0, 5.1, and 4.9 m, amongst patients with different marital statuses (p , 0.01). Our results show that the
respectively. mPFS for all ERBB-2 directed therapy (n = 59) was 3.5 m. The most common highest 5-year overall survival rate is seen in married individuals at 62.1% (61.9% -
concurrent co-GA were TP53 (90%) and APC (64%) and most common co-amp were CDK12 (31%)
62.2%), followed by single individuals at 53.0% (52.6% - 53.3%), and SDW individuals at
and EGFR (23%). Concurrent APC mutation was associated with improved OS (53.8 vs. 26.2 m, p =
0.003). No other co-GA or co-amp, including KRAS, showed significant survival outcome as-
44.6% (44.3% - 44.8%). Overall survival improved over time across all marital groups
sociations. TP53 co-GA was present in 90% of both left and right-sided ERBB-2 CRC and KRAS co- when comparing the 2000–2010 and 2011–2021 periods. For married individuals, 5-year
GA in 15% and 32%, respectively. Conclusions: There was no difference in OS in ERBB2 amp CRC overall survival rate increased from 61.2% (61.0%–61.4%) to 63.1% (62.9%–63.4%).
in patients who received ERBB2 directed therapy, however patients had improved OS with surgery Similarly, survival improved for single individuals (51.1% [50.6%–51.6%] to 54.5%
or localized therapy options and should be offered to eligible patients if possible. Co-GA with APC [54.0%–54.9%]) and for SDW individuals (43.8% [43.5%–44.1%] to 45.7%
mutation correlated with improvement in survival outcomes. Research Sponsor: None. [45.3%–46.1%]). Our results show that married individuals were most commonly di-
Baseline data, initial staging, treatments, and mutational profile for ERBB2 co-amp CRC. agnosed with localized cancer compared with single and SDW individuals (41.8% vs
Age at Diagnosis (median) 55 years
36.5% vs 39.0%, p , 0.01), SDW individuals with regional cancer compared with single
and married individuals (38.3% vs 37.06% vs 37.5%, p , 0.01) and single individuals with
Gender – no. M (56), F (44)
Stage at Initial I (6), II (6), III (18), IV (69)
distant disease compared with married and SDW individuals (26.4% vs 20.69% vs
Diagnosis – no. 22.69%, p , 0.01). Conclusions: Marital status significantly impacts CRC survival, with
Location of Primary R-sided (19), L-sided (80), Cecum (7), Ascending (8), Transverse (5), married individuals having higher survival rates than single or SDW patients. This
Tumor – no. Descending (7), Sigmoid/Rectum (73) suggests that social support may contribute to improved outcomes, although it is
Surgery – no. Primary tumor (55), Liver metastectomy (24), Lung metastectomy (8)
Localized Therapy – no. RT (36), Ablation (6), Y90 (4), Cryotherapy (1) important to note that the higher survival rates may be associated with earlier diagnosis
ERBB2 Directed Therapy Trastuzumab (TRA) + Pertuzumab (20), ERBB2-directed trial drug (14), in married individuals. These findings highlight the potential need for additional support
(excluded if n=1) – no. Trastuzumab deruxtecan (8), TRA (7), TRA + Tucatinib (4) services for single and SDW individuals, underscoring the importance of considering
Mutational Profile Co-GA: TP53 (90%), APC (64%), SMAD4 (18%), KRAS (18%)
Co-amp: CDK12 (31%), EGFR (23%), MYC (19%), BRAF (13%) marital status when addressing demographic disparities in cancer care. Research
Sponsor: None.

3629 Poster Session 3630 Poster Session

Clinicopathological features of colorectal cancer in adolescents and young Adherence to repeat screening for colorectal cancer using the multi-target
adults (AYA) at a tertiary referral centre in Nigeria. First Author: Sharif Adeniyi stool DNA test: Real-world analysis of patients from federally qualified
Folorunso, Obafemi Awolowo University Teaching Hospital Complex, Ile Ife, Nigeria health centers. First Author: Mallik Greene, Exact Sciences Corporation, Madison, WI
Background: Colorectal cancer (CRC) is traditionally considered a disease of older Background: Initial and repeat colorectal cancer (CRC) screening rates are suboptimal
adults; however, its incidence among AYA has been rising globally. In sub-Saharan and fall short of national goals in adults from historically disadvantaged backgrounds.
Africa, including Nigeria, there is limited data on the clinicopathological characteristics This real-world study examined adherence to repeat mt-sDNA screening among patients
of CRC in this age group. Understanding the patterns of CRC in AYA in Nigeria is crucial from federally qualified health centers (FQHCs) across the US and among different payer
for developing targeted screening, early detection, and treatment strategies in this types. Methods: Laboratory data from Exact Sciences Laboratories LLC were utilized for
population. This study aims to investigate the clinicopathological features of CRC in AYA the period between January 1, 2023, and December 31, 2023, for those who had
at a tertiary referral centre in Nigeria. Methods: A retrospective review of the ARGO- previously completed a mt-sDNA test. Study outcomes included adherence rate to mt-
OAUTHC database was conducted to identify adolescents and young adults (age less sDNA repeat screening and time to test return. The mt-sDNA screening adherence rate
than 40) diagnosed with colorectal cancer between 2013 and 2024. We compared AYA was defined as the percentage of patients who completed and returned the test kit and
patients with middle-aged adults (ages 40-65) to assess differences in clinical pre- received a valid test result within 180 days of the initial shipment date. Time to test
sentation and histopathological features. Only patients with complete histopathology return was defined as the number of days from the date the test kit was shipped to the
were included in the analysis while patients older than 65 years were excluded. De- patient to the date ESL received the test kit with a specimen. Results: The study sample
scriptive statistics were used to summarize the data, and statistical tests were per- consisted of 19,536 eligible patients, including 9,592 (49.1%) aged 45-64 years, 8,689
formed to assess significant differences between the two groups. Results: 866 patients (44.5%) aged 65-75 years and 1,225 (6.4%) aged 76-85 years. Overall, the mt-sDNA
diagnosed with CRC were identified. Of these, 150 (18%) were AYA, with a median age of repeat screening adherence rate was 79.7% and the mean time to return the kit was
33 years, and 448 (54%) were middle-aged adults (ages 40-65). The presenting 21.1620.8 days. A total of 2,375 (15.3%) individuals tested positive for CRC. Screening
symptoms were similar between the AYA and middle-aged groups, however, AYA pa- adherence for patients with Medicare was 84.7%, Medicare Advantage was 80%,
tients were less likely to be diagnosed by colonoscopy (14.9% vs 29.4%, p = 0.002). commercial insurance was 78.2%, managed care organization was 74.6% and Medicaid
Additionally, more AYA patients presented with symptoms that did not limit their daily was 65.9% (p , 0.001). Compared with patients covered by commercial insurance, those
activities , leading to a delay in seeking medical attention (9.3% vs 2.8%, p = 0.016). AYA covered by Medicaid had 42% lower odds of adhering to mt-sDNA screening (OR: 0.582,
patients were less likely to present with early-stage disease (Stages I and II) compared to 95% CI]: 0.461-0.733, p , 0.001). Patients with 2 or more prior successful tests had a
middle-aged patients, (13.1% vs 20.9%, p = 0.045). Histopathological features were numerically shorter mean time to test return compared to those with only 1 prior
similar between the two groups overall, though mucin production was more prevalent in successful test, both overall and within each payer type. Conclusions: Findings from
AYA patients (64.1% vs 45.2%, p = 0.044). Lung metastasis was less common in AYA this real-world study suggest that patients receiving care from FQHCs had relatively high
patients (8.3% vs 23.5%, p = 0.011), the pattern of metastasis to other organs was repeat mt-sDNA screening rates. While repeat screening adherence was generally high
however similar between the two groups. Conclusions: AYA patients in Nigeria tend to across patient demographic categories, there were significant differences by type of
present at more advanced stages, partly due to the non-restrictive nature of their insurance coverage and number of prior successful mt-sDNA screenings. Research
symptoms. They were less likely to be diagnosed via colonoscopy and were more likely Sponsor: None.
to exhibit mucin production in their tumors. These findings underscore the need for
increased awareness and early screening efforts for colorectal cancer in AYA pop-
ulations to facilitate earlier diagnosis. Research Sponsor: African Research Group for
Oncology.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 267s

3631 Poster Session 3632 Poster Session

Quality-adjusted time without symptoms of disease or toxicity (Q-TWiST) Association between fine particulate matter exposure and colorectal
analysis of pembrolizumab (pembro) versus chemotherapy (chemo) in cancer mortality by age at diagnosis: Insights from county-level analysis
microsatellite instability–high (MSI-H)/mismatch repair–deficient (dMMR) in the United States. First Author: Sara F. Haddad, Cleveland Clinic Foundation,
metastatic colorectal cancer (mCRC) in the KEYNOTE-177 trial. First Author: Cleveland, OH
Elena Elez, Vall d’Hebron University Hospital and Institute of Oncology (VHIO), Barcelona, Background: Colorectal cancer (CRC) remains a leading cause of cancer mortality in the
Spain U.S., with environmental factors like air pollution potentially contributing to mortality.
Background: In KEYNOTE-177 (NCT02563002),first-line pembro provided significantly This study examines the association between fine particulate matter (PM2.5) exposure
longer progression-free survival and a trend toward improvement in overall survival (OS) and CRC mortality rates, stratified by average-onset (aoCRC [$50 years at diagnosis])
compared with chemo in participants (pts) with MSI-H/dMMR mCRC after . 5 years of and early-onset (eoCRC [ , 50 years at diagnosis]) cases. Methods: Age-adjusted
median follow-up. The objective of this analysis was to assess Q-TWiST for pts treated in county CRC mortality data (1999–2020, n = 2981 counties) were obtained from CDC
KEYNOTE-177. Methods: Q-TWiST combines efficacy, safety, and quality of life in a single Underlying Cause of Death (ICD-10: C18-C19 [excluding C18.1], C20), while 1999-2019
measure. The analysis categorized survival time into 3 health states: time with PM2.5 estimates (ug/m3), derived from Atmospheric Composition Analysis Group data,
grade $3adverse events before disease progression or death (toxicity [TOX]), time without were mapped to U.S. county boundaries using R v.4.3.2. Covariates were derived from
symptoms or toxicity before disease progression (TWiST), and time from disease pro- Census American Community Survey (ACS) 5-year estimates, CDC data, and County
gression to death (relapse [REL]). In all randomly assigned pts, the restricted mean survival Health Rankings (CHR) data and included county-level area deprivation index (ACS), non-
time (RMST) in each state was first weighted by a quality-of-life utility value, measured as Hispanic Black % (ACS), Hispanic % (ACS), obesity prevalence (CDC), smoking preva-
treatment-specific EQ-5D-3L scores for each health state using the US value set, and then lence (CHR), binge alcohol consumption (CHR), and uninsured % (ACS). Negative bi-
summed to calculate the Q-TWiST value. Relative gains (defined as the Q-TWiST dif- nomial count models were fit to explore associations between PM2.5 and CRC mortality
ferences between pembro and chemo divided by the RMST of chemo) of $15% were rates, adjusted for covariates. Deviance R2 was computed to examine model fits.
defined as “clearly clinically important.” Treatment difference 95% CIs were generated Results: We found that for every 1% increase in PM2.5 exposure, aoCRC mortality
using the nonparametric bootstrapping method. The data cutoff date was July 17, 2023. increased by 0.98% (p , 0.001) and eoCRC mortality increased by 0.24% (p = 0.435),
Results: At a maximum follow-up of 84 months, pts in the pembro arm had a 13.8-mo after adjustment for covariates. Deviance R2 values indicated that PM2.5 and covariates
(95% CI, 4.8-21.7) longer RMST in TWiST (24.3 vs 10.5 mo), a 4.8-mo (95% CI, 20.5 to 10.2) explained 32.0% and 37.0% of the deviance in aoCRC and eoCRC mortality, respectively.
longer RMST in TOX (10.7 vs 5.9 mo), and an 11.1-mo (95% CI, 221.2 to 20.8) shorter Conclusions: PM2.5 exposure was a significant predictor of CRC mortality, but only for
RMST in REL (18.1 vs 29.3 mo) compared with chemo. As a proportion of overall pre-
aoCRC cases. Air pollution and other covariates accounted for roughly one-third of the
progression time, pts in the pembro arm spent less time in TOX than the chemo arm (31%
county-level deviance, suggesting the influence of additional factors. Systemic and
vs 36%). For the analysis of restricted mean Q-TWiST based on treatment-specific utility
individual-level interventions to reduce air pollution exposure may mitigate CRC mor-
weights, the difference between pembro and chemo favored pembro by 9.1 mo (95% CI,
tality disparities in older populations. Further studies are needed to explore other
2.3-15.5), a 20.0% (95% CI, 4.8-36.8) relative Q-TWiST gain. When TOX definition included
potential contributors to CRC mortality. Research Sponsor: None.
grade $2 adverse events, the relative Q-TWiST gain was 19.7% (95% CI, 4.7-36.5). When
OS was adjusted for crossover to anti–PD-(L)1 therapy as second-line treatment, relative
Q-TWiST gain was 39.7% (95% CI, 20.1-63.7). Conclusions: Pembro provided clearly
clinically important improvement in quality-adjusted survival time based on Q-TWiST
analyses compared with chemo as first-line treatment in pts with MSI-H/dMMR mCRC. The
magnitude of results related to established thresholds for clinically important Q-TWiST
gain suggests that results from this analysis provide additional evidence for the use of
pembro in this population. Clinical trial information: NCT02563002. Research Sponsor:
Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA.

TPS3633 Poster Session TPS3634 Poster Session

A phase II study of pembrolizumab, carboplatin, paclitaxel, and radiation for Alliance A022101/NRG-GI009: A pragmatic randomized phase III trial eval-
the treatment of early-stage anal cancer: Big Ten Cancer Research Consor- uating total ablative therapy for patients with limited metastatic colorectal
tium GI22-588. First Author: Anita Turk, Indiana University Melvin and Bren Simon cancer—Evaluating radiation, ablation, and surgery (ERASur). First Author:
Comprehensive Cancer Center, Indianapolis, IN Eric David Miller, Ohio State University, Columbus, OH
Background: Anal squamous cell carcinoma (SCC) is an evolving public health chal- Background: For patients with oligometastatic colorectal cancer (CRC), aggressive local
lenge, with increasing incidence and mortality rates, particularly in older women with therapy of isolated metastases, particularly in the liver, has been associated with long-term
long-standing HPV infections. Standard-of-care treatment for early-stage anal SCC—5- progression-free and overall survival (OS) primarily based on retrospective evidence.
fluorouracil (5FU) and mitomycin-C (MMC) with radiation—achieves high cure rates but However, in patients with limited metastatic CRC that is deemed inoperable or those with
poses significant toxicity risks, with grade 3-5 adverse events occurring in up to 70% of additional disease outside of the liver or lungs, the role of local ablative therapies, including
patients. Recent clinical trials, such as DECREASE and ACT4, have explored radiation de- microwave ablation (MWA) and stereotactic body radiation therapy (SBRT), to render
escalation strategies, but limited progress has been made in expanding systemic patients disease free is less clear. Despite the long history of treating oligometastatic CRC
treatment options for locally advanced disease. Building on pilot data demonstrating with local therapy, which is largely provider biased, questions remain regarding the benefit
favorable clinical complete responses (cCR) with carboplatin and paclitaxel combined of extending the paradigm of metastatic directed therapy to patients with more extensive
with radiation in patients ineligible for SOC regimens, this phase II trial evaluates the disease. This trial seeks to use a pragmatic multimodality approach that mirrors the
addition of pembrolizumab to this combination to enhance efficacy while reducing current clinical dilemma. This study is designed to evaluate the safety and efficacy of
toxicity. Methods: This is a single-arm, phase II trial evaluating concurrent chemo- adding total ablative therapy (TAT) of all sites of disease to standard of care systemic
radiation with weekly carboplatin (AUC 2), paclitaxel (50 mg/m²), and pembrolizumab treatment in those with limited metastatic CRC. Methods: A022101/NRG-GI009 is a
(200 mg every three weeks during chemoradiation and 400 mg every six weeks during National Clinical Trials Network randomized phase III study planned to enroll 364 patients
maintenance) for early-stage anal SCC in patients ineligible for 5FU and MMC. The with newly diagnosed metastatic CRC (BRAF wild-type, microsatellite stable) without
chemoradiation phase consists of up to six weeks of therapy with 50.4 Gy delivered over peritoneal metastasis or liver-only disease on baseline imaging. Patients receive first-line
28 fractions. Maintenance pembrolizumab is administered for up to eight cycles. The systemic therapy for 12-39 weeks. Patients with #4 sites of metastatic disease following
primary endpoint is the cCR rate at six months post-chemoradiation. Secondary initial systemic therapy that are amenable to any combination of surgical resection, MWA,
and/or SBRT are then randomized 1:1, stratified by number of metastatic organ sites (1-2
endpoints include safety, tolerability, tumor downstaging, and disease-free survival.
vs. 3-4), timing of metastatic disease diagnosis (de novo vs. secondary), and presence of
Exploratory objectives include the evaluation of genomic alterations and biomarkers
metastatic disease outside the liver/lungs in at least 1 site. Patients in Arm 1 will receive
such as versican and keratin 17 as predictors of therapeutic response. Major eligibility
TAT consisting of treatment of all metastatic sites with SBRT, MWA, and/or surgical
criteria include histologically confirmed stage I-IIIA anal SCC, measurable disease per
resection followed by standard of care systemic therapy. Patients in Arm 2 will continue
RECIST v1.1, and treatment-naı̈ve status. Key exclusions include active autoimmune with standard of care systemic therapy alone. The primary endpoint is OS. Secondary
disease requiring immunosuppression within two years and prior checkpoint inhibitor endpoints include event-free survival, treatment-related toxicities, and local recurrence
therapy. The target enrollment is 23 patients, with an accrual period of 12 months and an with exploratory biomarker analyses. The study needs 346 evaluable patients combined in
anticipated study duration of two years. The study is currently enrolling participants the 2 arms to demonstrate an improvement in OS with a hazard ratio of 0.7 to provide 80%
through the Big Ten Cancer Research Consortium. Clinical trial information: power with a one-sided alpha of 5%. The trial utilizes a group sequential design with two
NCT06493019. Research Sponsor: Merck. interim analyses (25% and 50% of events) for futility. The trial activated in January 2023
and recruitment is ongoing. Support: U10CA180821, U10CA180882; https://
[Link]. U10CA180820 (ECOG-ACRIN); U10CA180868
(NRG); U10CA180888 (SWOG); Clinical trial information: NCT05673148. Research Sponsor:
U.S. National Institutes of Health.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
268s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

TPS3635 Poster Session TPS3636 Poster Session

Telisotuzumab adizutecan (ABBV-400; Temab-A) monotherapy vs OrigAMI-2: A randomized, phase 3 study of amivantamab vs cetuximab,
trifluridine/tipiracil plus bevacizumab in patients with refractory metastatic both in combination with FOLFOX or FOLFIRI, as first-line treatment in left-
colorectal cancer with increased c-Met protein expression: An open-label, sided RAS/BRAF wild-type metastatic colorectal cancer. First Author: Dirk
randomized, phase 3 trial. First Author: John H. Strickler, Duke University School of Arnold, Asklepios Tumorzentrum Hamburg, AK Altona, Hamburg, Germany
Medicine, Durham, NC Background: Approximately 50% of patients diagnosed with metastatic colorectal
Background: c-Met protein expression is increased in several solid tumors, including cancer (mCRC) are wild-type for KRAS, NRAS, and BRAF (RAS/BRAF WT). Standard initial
colorectal cancer (CRC). Temab-A is a c-Met–directed antibody-drug conjugate consisting therapy for left-sided RAS/BRAF WT mCRC is doublet chemotherapy (FOLFOX or
of the antibody telisotuzumab conjugated to a potent topoisomerase 1 inhibitor payload. FOLFIRI) combined with anti-EGFR therapy. However, resistance is nearly inevitable.
Preliminary data from the ongoing first-in-human study of Temab-A (NCT05029882) MET alterations are known resistance mechanisms to EGFR inhibition, with MET
indicate a tolerable safety profile and promising antitumor activity in patients with third-line amplification occurring in 5%-23% of EGFR-resistant mCRC. Amivantamab is an EGFR-
or later metastatic (m)CRC (Sharma et al. JCO 2023;41:3015). Herein, we describe a phase 3 MET bispecific antibody with immune cell-directing activity that is approved by the FDA
study comparing Temab-A monotherapy with the standard of care (trifluridine/tipiracil plus for 4 indications in EGFR-mutated advanced non-small cell lung cancer. In the phase 1b/
bevacizumab) in patients with refractory mCRC with c-Met expression of 3+ in $10% of 2 OrigAMI-1 study (NCT05379595), the combination of amivantamab plus FOLFOX or
tumor cells by immunohistochemistry (IHC). Methods: This is an open-label, randomized, FOLFIRI demonstrated rapid and durable antitumor activity, regardless of tumor sid-
controlled, global phase 3 study(NCT06614192). Patient eligibility includes age $18 years, edness, in participants with RAS/BRAF WT mCRC (Pietrantonio ESMO 2024). The ob-
confirmed c-Met expression of 3+ in $10% of tumor cells, metastatic adenocarcinoma of jective of this phase 3 randomized study is to assess the efficacy of amivantamab, as
the colon/rectum, measurable disease per RECIST v1.1, ECOG performance status 0–1, prior compared with cetuximab, both in combination with FOLFOX or FOLFIRI, as first-line
treatment with a fluoropyrimidine (eg, 5-FU or capecitabine), oxaliplatin, irinotecan, and an
therapy for participants with left-sided RAS/BRAF WT unresectable or metastatic CRC.
anti-VEGF antibody (unless locally not approved) or an anti-EGFR antibody if indicated, and
Methods: The multicenter, global OrigAMI-2 study (NCT06662786) is planned to open in
appropriate targeted therapy or immunotherapy if targetable mutations present (eg, BRAF
V600E or HER2) or MSI-H/dMMR. Prior treatment with regorafenib and/or fruquintinib is
216 sites in 21 countries. Eligible participants will be WT for KRAS, NRAS, and BRAF by
permitted, but no prior treatment with trifluridine/tipiracil Study-specific c-Met protein local testing, have left-sided unresectable or metastatic colorectal cancer, and be
expression IHC cutoff is defined as 3+ intensity in $10% of tumor cells. The study consists treatment-naı̈ve for advanced disease. Left-sided disease will be defined as a primary
of 2 stages. In stage 1, at least 60 patients will be randomized 1:1 to receive 2 different doses tumor arising from the splenic flexure, descending colon, sigmoid colon, rectosigmoid, or
of intravenous (IV) Temab-A. In stage 2, 400 patients will be randomly assigned 1:1 to rectum. Key exclusion criteria include known dMMR/MSI-H status, HER2-positive or
receive either the optimized dose of IV Temab-A or oral trifluridine/tipiracil plus IV bev- amplified tumor, and prior exposure to EGFR or MET targeting agents. Approximately
acizumab. In stage 1, primary objectives are to determine the recommended phase 3 dose 1000 participants will be randomly assigned 1:1 to receive subcutaneous amivantamab
and to evaluate the efficacy, as measured by objective response (OR), and safety of Temab- (co-formulated with recombinant human hyaluronidase [rHuPH20]) or intravenous
A; secondary objectives are to assess progression-free survival (PFS), overall survival (OS), cetuximab, both combined with FOLFOX or FOLFIRI (investigator’s choice). Randomi-
duration of response (DOR), disease control rate (DCR), and pharmacokinetics. In stage 2, zation will be stratified by chemotherapy choice (FOLFOX or FOLFIRI), limited disease
the primary objectives are to demonstrate the superiority of Temab-A over trifluridine/ (yes or no), and prior adjuvant therapy (yes or no). The primary endpoint will be
tipiracil plus bevacizumab in terms of OR and OS; secondary objectives are to evaluate PFS, progression-free survival by blinded independent central review. Secondary endpoints
DOR, DCR, and safety of Temab-A treatment, and its impact on patient-reported outcomes. include overall survival, objective response rate, duration of response, and patient-
Response will be assessed by blinded independent central review per RECIST v1.1. Safety reported outcomes. Safety assessments will include monitoring adverse events and
evaluations include adverse event monitoring, vital sign measurements, ECG variables, and laboratory abnormalities. Clinical trial information: NCT06662786. Research Sponsor:
clinical laboratory testing. Enrollment began in December 2024. Clinical trial information: Janssen Research & Development, LLC, a Johnson & Johnson Company.
NCT06614192. Research Sponsor: AbbVie Inc.; n/a.

TPS3637 Poster Session TPS3638 Poster Session

An observational/translational study to conduct real-world evidence and OrigAMI-3: A randomized, phase 3 study of amivantamab plus FOLFIRI vs
develop biomarkers of fruquintinib for patients with metastatic colorectal cetuximab or bevacizumab plus FOLFIRI in participants with recurrent,
cancer (mCRC): FruBLOOM trial (JACCRO CC-19). First Author: Yu Sunakawa, unresectable, or metastatic RAS/BRAF wild-type colorectal cancer. First
Department of Clinical Oncology, St. Marianna University School of Medicine, Kawasaki, Author: Joel R. Hecht, UCLA Jonsson Comprehensive Cancer Center, Santa Monica, CA
Kanagawa, Japan Background: Among patients with metastatic colorectal cancer (mCRC), approximately
Background: The FRESCO-1/2 trials demonstrated a survival benefit of fruquintinib (FRU) in 50% are wild-type for KRAS, NRAS, and BRAF (RAS/BRAF WT) without actionable genomic
mCRC after 3rd-line therapy. FRU can be considered as one of the standard treatments for alterations. Standard first-line therapy for RAS/BRAF WT mCRC is 5-FU-based doublet
mCRC. The FRESCO-2 trial was conducted in patients (pts) treated with FTD/TPI and/or chemotherapy (FOLFOX or FOLFIRI) plus anti-EGFR or anti-VEGF therapy. The choice of
regorafenib (Rego) and did not include pts who were not using either agent. Although FTD/TPI second-line treatment is dependent on first-line treatment (eg, oxaliplatin-based che-
+ bevacizumab (Bev) is currently one of 3rd-line standard treatments for mCRC, there are few motherapy in the first-line necessitates irinotecan-based in the second-line, and vice
data regarding the efficacy and safety of FRU in pts after FTD/TPI + Bev. Therefore, this study versa). Known resistance mechanisms to anti-EGFR therapy are MET alterations, with MET
will accumulate real-world-data of FRU in clinical practice and evaluate the efficacy and amplification occurring in 5%-23% of EGFR-resistant mCRC and increasing in prevalence
safety of FRU after FTD/TPI + Bev. Also, we will evaluate clinical outcomes of FRU as 3rd- or over subsequent lines of therapy. Amivantamab is an EGFR-MET bispecific antibody with
later-line treatment after both FTD/TPI + Bev and Rego. The predictive biomarkers of FRU in immune cell-directing activity and is FDA-approved for 4 indications in EGFR-mutated
the later-line setting hold significant clinical promise, for choosing personalized treatment advanced non-small cell lung cancer. In the phase 1b/2 OrigAMI-1 study (NCT05379595),
plans (e.g., FRU vs. FTD/TPI + Bev / Rego) and enhancing the prognosis of mCRC pts. amivantamab plus FOLFIRI demonstrated promising antitumor activity, independent of line
Therefore, this translational study approaches developing biomarkers for predicting FRU of therapy, in participants (pts) with RAS/BRAF WT mCRC without prior anti-EGFR exposure
efficacy by analyzing pre-treatment blood samples. Furthermore, we will explore treatment (Pietrantonio ESMO 2024). The objective of this phase 3 randomized study is to assess the
resistance mechanisms using post-treatment blood samples. Methods: This is a multicenter efficacy of amivantamab plus FOLFIRI vs cetuximab or bevacizumab plus FOLFIRI, as
observational/translational study to prospectively evaluate the efficacy and safety of FRU second-line therapy for pts with recurrent RAS/BRAF WT mCRC. Methods: The global
as a 3rd- or later-line treatment, mainly after FTD/TPI + Bev in mCRC pts in clinical practice. OrigAMI-3 study (NCT06750094) is planned to open in 230 sites in 25 countries. Eligible pts
We will enroll 200 pts receiving FRU after FTD/TPI + Bev to the cohort A, and 100 pts receiving will be WT for KRAS, NRAS, and BRAF, have recurrent unresectable or mCRC, and must have
FRU as 3rd-line or after both FTD/TPI + Bev and Rego to the cohort B. Eligibility criteria are (1) had disease progression on one prior line of systemic therapy for metastatic disease (prior
pts with CRC confirmed as adenocarcinoma, (2) pts planning to receive FRU monotherapy as
regimen must be fluoropyrimidine-based and oxaliplatin-based therapy). Pts with treated,
3rd- or later-line treatment, (3) prior treatments with fluoropyrimidine-, oxaliplatin- and
stable, and asymptomatic brain metastases are allowed. Key exclusion criteria include
irinotecan-based chemotherapy, an anti-VEGF biological therapy, an anti-EGFR therapy (if
known dMMR/MSI-H status without prior immunotherapy, HER2-positive or amplified
RAS/BRAF wild-type), BRAF therapy (if BRAF mutant), and immune checkpoint inhibitor (if
tumor, and prior exposure to irinotecan or agents targeting EGFR or MET. Approximately
MSI-high), (4) pts with ECOG Performance Status of 0-2, (5) pts must be at least 18 years of
700 pts will be randomly assigned 1:1 to receive subcutaneous amivantamab (co-
age at the time of consent, and (5) pts have measurable or evaluable lesions in RECIST v1.1.
The primary endpoint is overall survival in pts of the cohort A. The secondary endpoints are formulated with recombinant human hyaluronidase [rHuPH20]) plus FOLFIRI vs intra-
clinical outcomes including response rate, progression-free survival, duration of response, venous cetuximab or bevacizumab (investigator’s choice, per local guidelines) plus
and safety in pts of the cohort A and B. In the biomarker study, blood samples will be FOLFIRI. Randomization will be stratified by choice of cetuximab or bevacizumab, primary
prospectively collected before and after treatment, for translational research including tumor location (left vs right-sided), duration of first-line therapy (, 6 months
genomic alteration analysis in circulating tumor-DNA by DNA exome sequencing, gene or $6 months), and prior anti-VEGF therapy (yes or no). The dual primary endpoints will be
expression measurement in cfRNA by tumor-educated blood platelets (TEP)-Seq RNA progression-free survival by blinded independent central review and overall survival.
analysis, and plasma proteins analysis by multiplex immunoassay panels. Enrollment opens in Secondary endpoints include objective response rate, duration of response, and patient-
February 2025 (UMIN000056813). Clinical trial information: UMIN000056813. Research reported outcomes. Safety assessments will include monitoring adverse events and
Sponsor: Takeda. laboratory abnormalities. Clinical trial information: NCT06750094. Research Sponsor:
Janssen Research & Development, LLC, a Johnson & Johnson company.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 269s

TPS3639 Poster Session TPS3640 Poster Session

Stereotactic body radiotherapy combined with PD-1 antibody in unresect- Neoadjuvant cetuximab plus tislelizumab combined with chemotherapy in
able colorectal liver metastases: A prospective, multicenter, single-arm, pMMR RAS/BRAF wild-type (wt) locally advanced rectal cancer (LARC): A
phase II study (SPARKLE-L). First Author: Fang He, Department of Radiation prospective, multicenter, phase II study. First Author: Gang Liu, Tianjin Medical
Oncology, the Sixth Affiliated Hospital, Sun Yat-sen University, Guangzhou, China University General Hospital, Tianjin, Tianjin, China
Background: Colorectal cancer liver metastasis (CRLM) significantly decreases colo- Background: For patients (pts) with locally advanced rectal cancer (LARC), chemo-
rectal cancer (CRC) patient prognosis, affecting 30-50% of CRC patients at diagnosis or radiotherapy followed by total mesorectal excisionis is recommended as standard therapy
thereafter. Notably, up to 70%-90% CRLM are diagnosed as unresectable. Standard according to the NCCN guidelines. However, there is no stratification strategy for neo-
treatments include systemic chemotargeted-therapies (CT). However, only 10-30% adjuvant therapy based on molecular alterations, and radiotherapy is insufficient with
CRLM can be converse to resectable state by CT, with an objective response rate (ORR) pathologic complete response (pCR) rates at 11%-15%. There is an urgent need for new
of just 15%-20% and a median overall survival (OS) of approximately 20-30 months. therapeutical options to improve the pCR rate in these pts. Adding anti-EGFR therapy to
Improving prognosis of CRLM patients remains challenging. Stereotactic body radiation neoadjuvant chemotherapy may improve progression-free survival for RAS/BRAF WT LARC
therapy (SBRT) combined with immunotherapy might offer promising alternatives. SBRT pts. Furthermore, previous studies demonstrated that combining anti-EGFR with immune
provides high-dose tumor control while protecting surrounding tissues better than checkpoint inhibitors could further improve pCR rate. Cetuximab, an anti-EGFR monoclonal
conventional radiotherapy. It also facilitates the release of tumor-associated antigens, antibody, has gained FDA approval for RAS WT metastatic colorectal cancer. Tislelizumab,
reshaping the immune microenvironment and inducing stronger immune responses. The an anti-PD-1 monoclonal antibody, is effective in blocking PD-1/PD-L1 interaction in
preclinical experiments. This study introduces an innovative approach, combining Cetux-
combination of SBRT and PD-1 antibodies might synergistically enhance the anti-tumor
imab, Tislelizumab and chemotherapy, as a total neoadjuvant therapy for pMMR RAS/BRAF
efficacy. Despite SBRT’s demonstrated efficacy in unresectable CRLM with few adverse
wt LARC pts. Methods: This prospective, multicenter, phase II study investigated the
reactions, no prospective studies have explored its combination with PD-1 antibodies.
efficacy and safety of neoadjuvant treatment with FOLFOX chemotherapy plus Cetuximab
Methods: This is a multicenter, open-label, single-arm, phase II trial conducted in China. and Tislelizumab for MSS-RAS/BRAF WT LARC. Eligible participants were 18 years or older,
Patients will receive SBRT at 8-12 Gy per fraction over 5 fractions, combined with 5-FU- with an ECOG PS of 0–2, primary, and a biopsy-proven tumors meeting all the following
based CT and PD-1 antibody therapy before and after SBRT. Eight weeks (62 weeks) criteria: clinical tumour stage cT3-4 N0M0 or cT1-4N+M0, tumor distance from the
post SBRT, imaging assessments or multi-point liver biopsies will be performed. anus&10 cm, no distant metastasis. Pts initially received a cycle chemotherapy of FOLFOX
Multidisciplinary teams (MDT) will determine subsequent plans: cCR/pCR patients will pending genetic results. Eligible pts with MSS-RAS/BRAF WT LARC then underwent 5
undergo maintenance CT or enter a watch-and-wait phase; non-cCR/pCR patients will preoperative neoadjuvant cycles of mFolfOx6 (oxaliplatin 85 mg/m², D1; leucovorin 200 mg/
continue maintenance CT or exit the study. This is the first study exploring whether SBRT m²,D1; 5-FU bolus 400 mg/m2 D1 then 2.4 g/m²,D2-3) + Cetuximab (500mg/m2, D1, q2w) +
combined with PD-1 monoclonal antibody can improve ORR, OS, quality of life (QOL) and Tislelizumab (200mg, D1, q2w). Subsequently, pts underwent TME about 4 weeks after the
potentially achieve no evidence of disease (NED) status for unresectable CRLM. Key last cycle. Imaging evaluation will be conducted 6 weeks after the initiation of treatment, pts
inclusion criteria: pMMR/MSS CRC, MDT-assessed unresectability due to main portal with regressed tumors will receive a standard chemoradiotherapy. The primary endpoint was
vein invasion, multiple hepatic vein invasion or lack of R0 resection/ablation feasibility. pCR rate. Secondary endpoints included the Neoadjuvant Rectal Score, Objective Response
Main exclusion criteria encompass active hepatitis, cirrhosis, Child-Pugh B/C, check- Rate, R0 resection rate, Major Pathological Response rate, Anal Sparing rate, 3-year Disease-
point inhibitor therapies history and ECOG performance status $2. Twenty-four patients Free Survival, 3-year Local Recurrence Rate, 3-year Overall Survival. Based on a review of the
are planned for enrollment, with two already enrolled as of January 25, 2025. The study is literature, the estimated pCR rate for standard preoperative neoadjuvant chemoradiotherapy
registered with [Link] (NCT06794086) and is ongoing. Clinical trial infor- is approximately 15%. The expected pCR rate for the MSS-RASwt/BRAFwt group is around
mation: NCT06794086. Research Sponsor: None. 30%, with a one-sided significance level（a）of 0.05 and a power (1-b) of 0.8, using the
Simon two-stage method, the sample size is calculated to be 25 cases. The study started in
middle 2022 and is recruiting. Clinical trial information: ChiCTR2200062002. Research
Sponsor: None.

TPS3641 Poster Session TPS3642 Poster Session

Node-sparing modified short-course radiotherapy combined with CAPOX A phase II study of encorafenib and cetuximab (EC) beyond progression in
and tislelizumab versus conventional short-course preoperative chemo- combination with FOLFIRI in BRAF V600E-mutated metastatic colorectal
radiotherapy for proficient mismatch repair or microsatellite stable locally cancer (mCRC). First Author: Maria Alessandra Calegari, Medical Oncology, Com-
advanced rectal cancer (mRCAT-III): A multicenter, randomized, open-label, prehensive Cancer Center, Fondazione Policlinico Universitario Agostino Gemelli, IRCCS,
phase 3 trial. First Author: Zhangfa Song, Sir Run Run Shaw Hospital, Zhejiang Rome, Italy
University School of Medicine, Hangzhou, Zhejiang, China Background: BRAF V600E mutation occurs in approximately 10% of metastatic co-
Background: Total neoadjuvant chemoradiotherapy is the standard of care for locally lorectal cancers (mCRC) and confers poor prognosis. While the combination of
advanced rectal cancer (LARC) to control local recurrence and achieve organ preser- encorafenib and cetuximab (EC) has demonstrated improved outcomes in previously
vation. However, for proficient mismatch repair (pMMR) or microsatellite stable (MSS) treated BRAF V600E-mutated mCRC patients (pts), the duration of response remains
LARC, which accounts for nearly 90% of rectal cancers, conventional chemoradiotherapy suboptimal with median progression-free survival (mPFS) of 4.3 months (m). Ap-
has limited efficacy and is associated with significant side effects. Recent studies have proximately half of pts (45%) receive subsequent treatment after EC progression,
shown that combining radiotherapy with immunochemotherapy can improve pathological predominantly with chemotherapy (ChT). Data from the safety lead-in (SLI) of the
complete response (pCR) rates, but the inclusion of tumor-draining lymph nodes (TDLNs) BREAKWATER trial, assessing EC in combination with ChT doublet in the first-line
in the conventional irradiation field may impair T-cell immunity and reduce response to setting, demonstrated manageable safety profile and promising early efficacy signals for
immunotherapy. Our previous phase II trial demonstrated that node-sparing modified EC in combination with FOLFIRI, despite pharmacokinetic interaction.
short-course radiotherapy combined with chemotherapy and PD-1 blockade could Methods: ECLYPse (NCT06640166; EU CT Number 2023-508615-24-00) is a multi-
achieve a high pCR rate of 78.8% in pMMR LARC1. Building on these findings, we initiated center, single-arm, phase II study evaluating EC continuation beyond progression in
this phase III trial to compare this novel treatment regime with conventional short-course combination with FOLFIRI in pts with BRAF V600E-mutated mCRC who progressed on
chemoradiotherapy in improving pCR rates. Methods: This is a phase III, open-label, second-line EC. Key eligibility criteria include: histologically confirmed colorectal ad-
multicenter, randomized trial conducted across 17 hospitals in China. A total of 170 enocarcinoma, BRAF V600E mutation, documented disease progression on EC in
eligible MSS/pMMR middle or low rectal cancer patients (cT3-4N0/+M0) will be recruited second-line setting, benefit to previous treatment with EC (best response: complete
and randomly assigned ([Link]) to three groups: control group (conventional short-course response, partial response or stable disease lasting for at least 3 months), measurable
chemoradiotherapy), experimental group (node-sparing modified short-course chemo- disease according to RECIST 1.1 criteria, ECOG PS #1, and availability of archival tumor
radiotherapy plus PD-1 blockade), and exploratory group (conventional short-course tissue. Patients receive encorafenib 300 mg daily, cetuximab 500mg/m2 iv every
chemoradiotherapy plus PD-1 blockade). The innovative node-sparing modified short- 2 weeks, and standard FOLFIRI regimen. The primary endpoint is 6-month PFS rate.
course radiotherapy targets only the primary tumor bed, excluding TDLNs. Following Secondary endpoints include PFS, overall survival, duration of response, objective
randomization, patients will receive short-course radiotherapy (conventional or node- response rate, disease control rate, and safety. Translational analyses include com-
sparing) followed by four cycles of CAPOX 6 tislelizumab: tislelizumab 200 mg IV on day prehensive genomic profiling on archival tissue and serial ctDNA analysis. Tumor
1, oxaliplatin 130 mg/m² IV on day 1, and capecitabine 1000 mg/m² orally on days 1-14, assessment with contrast-enhanced CT scan of thorax, abdomen and pelvis is performed
and Total mesorectal excision (TME) will be performed at weeks 14-15. The primary every 8 weeks. Using a single-stage design with one-sided a = 0.05 and 80% power, 25
endpoint is pCR rate, while secondary endpoints include organ preservation rate, disease- patients will be enrolled to detect an improvement in 6-month PFS rate from 10% (null
free survival, overall survival, adverse effects, and quality of life. As of January 2025, 46 of hypothesis) to 30% (alternative hypothesis). If at least 7 pts will be alive and not
the planned 170 patients have been enrolled. The Data Monitoring Committee (DMC) progressing at 6 months, the treatment will be considered sufficiently active to warrant
reviewed the trial in December 2024 and recommended continuing as planned. Reference: further investigation. The study is currently enrolling at multiple sites in Italy. Clinical
Annals of Oncology (2024) 24 (suppl_1): 1-20. 10.1016/iotech/iotech100744 Clinical trial trial information: NCT06640166. Research Sponsor: Pierre Fabre.
information: NCT06507371. Research Sponsor: Sir Run Run Shaw Hospital Clinical
Research Cultivation Program.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
270s GASTROINTESTINAL CANCER—COLORECTAL AND ANAL

TPS3643 Poster Session TPS3644 Poster Session

Combining low-dose regorafenib with pembrolizumab as a front-line ther- Colon adjuvant chemotherapy based on evaluation of residual disease
apy for patients with MSI-H colorectal cancer: REGPEM-CRC-01. First Author: (CIRCULATE-NORTH AMERICA): NRG-GI008. First Author: Arvind Dasari, The
Ibrahim Halil Sahin, University of Pittsburgh Medical Center (UPMC), Pittsburgh, PA University of Texas MD Anderson Cancer Center, Houston, TX
Background: Currently, pembrolizumab is one of the front-line therapies for patients Background: Currently, there are no biomarkers validated prospectively in randomized
with MSI-H CRC. However, approximately 40% of patients who received pembrolizumab studies for resected colon cancer (CC) to determine need for adjuvant chemotherapy
experienced disease progression early in the course of disease (KEYNOTE 177). (AC). However, circulating tumor DNA (ctDNA) represents a highly specific and sensitive
Therefore, there is still an unmet need to enhance the efficacy of checkpoint inhibitors in approach (especially with serial monitoring) for identifying minimal/molecular residual
MSI-H CRC. MSI-H CRC has a higher level of expression of VEGF in blood compared to disease (MRD) post-surgery in CC patients (pts), and may outperform traditional clinical
patients compared to its MSS counterpart (Hansen et al. Colorectal Dis. 2011). Con- and pathological features in prognosticating risk for recurrence. CC pts who do not have
sistently, exploratory analysis of CALBG-80405 and PARADIGM trial showed that pa- detectable ctDNA (ctDNA-) are at a much lower risk of recurrence and may be spared the
tients with MSI-H CRC were more likely to benefit from anti-VEGF therapy than anti- toxicities associated with AC. Furthermore, for CC pts with detectable ctDNA (ctDNA+)
EGFR therapy regardless of the side of the tumor. NSABP C-08 also suggested that anti- who are at a very high risk of recurrence, the optimal AC regimen has not been
VEGF therapy may have biological activity even in as adjuvant therapy for patients with established. We hypothesize that for pts whose CC has been resected, ctDNA status may
MSI-H colon cancer. Regorafenib is a potent VEGF and multikinase inhibitor involved in be used to risk-stratify for making decisions about AC. Methods: In this prospective
pathologic processes such as oncogenesis, tumor angiogenesis, metastasis and tumor phase II/III trial, up to 1,912 pts with resected stage IIB, IIC, and III CC will be enrolled.
immunity, with preclinical evidence showing its immune modulatory effect in the tumor Based on the post-operative ctDNA status using personalized and tumor-informed assay
microenvironment. In this trial, we hypothesize that adding low-dose regorafenib to (Signatera™, bespoke assay), those who are ctDNA- (Cohort A) will be randomized to
pembrolizumab may induce synergistic activity beyond their independent clinical ef- immediate AC with fluoropyrimidine (FP)+oxaliplatin (Ox) for 3-6 mos per established
ficacy and create deep and durable responses for patients with MSI-H CRC. Methods: In guidelines v serial ctDNA monitoring. Patients who are ctDNA+ post-operatively, or with
the lead arm of this prospective randomized study, 22 patients will be enrolled through serial monitoring (Cohort B), will be randomized to FP+Ox v more intensive AC with
Hoosier Cancer Research Network (HCRN-GI23-643). In this first line clinical trial, addition of irinotecan (I) for 6 mos. One cycle of chemotherapy is allowed while awaiting
patients will receive regorafenib 60 mg daily in combination with pembrolizumab 200mg ctDNA testing results for cohort assignment. The primary endpoints for Cohort A are
IV in cycle 1, followed by regorafenib 90mg in subsequent cycles to optimize the time to ctDNA+ status (phase II) and disease-free survival (DFS) (phase III) in the
treatment tolerance and compliance. The primary outcome that will be measured is ORR, immediate v delayed AC arms. The primary endpoint for Cohort B is DFS in the FP+Ox v
defined as the percentage of partial or complete response to the treatment within FP+Ox+I arms for both phase II and phase III portions of the trial. Secondary endpoints
12 months. ORR will be measured using RECIST 1.1. criteria. A formal one-sided hy- include prevalence of detectable ctDNA post-operatively, time-to-event outcomes
pothesis test will be conducted for futility, assuming that we will reject the null hy- (overall survival and time to recurrence) by ctDNA status, and the assessment of
pothesis of a target ORR only if we have strong evidence. In this study, we assume a null compliance to adjuvant therapy. Biospecimens including archival tumor tissue, as well
hypothesis that ORR is 0.60, which would reflect significant clinical improvement over as post-operative plus serial matched/normal blood samples, will be collected for
the current standard of ORR = 0.43 from KEYNOTE 177. The alternative hypothesis is exploratory correlative research. Active enrollment across the NCTN started in June
that ORR is less than 0.60. For the lead-in phase of the study, the emphasis is on 2022 with CCTG sites joining in August 2023. Current accrual (as of 1-27-2025): 647/
controlling Type I error to be small, to be 0.05 or lower. An exact binomial test will be 1,912. NCT: 05174169. Support: U10 CA180868, -180822; -180888; UG1 CA189867;
conducted, based on the number of ORRs in the 22 patients. The study is currently Natera, Inc. Clinical trial information: NCT05174169. Research Sponsor: National Cancer
accruing through Hoosier Cancer Research Network was activated in July 2024. Clinical Institute; U10CA180868; National Cancer Institute; U10 CA180822; National Cancer
trial information: NCT06006923. Research Sponsor: None. Institute; UG1CA189867; Natera, Inc.

TPS3645 Poster Session TPS3646 Poster Session

Platform study of circulating tumor DNA–directed adjuvant chemotherapy mFOLFOX6 + bevacizumab + PD-1 monoclonal antibody vs. mFOLFOX6 in
in colon cancer (CLAUDIA Colon Cancer, KCSG CO22-12). First Author: Yongjun locally advanced pMMR/MSS CRC: A multicenter, randomized controlled
Cha, Center for Colorectal Cancer, Research Institute and Hospital, National Cancer phase III study (BASKETIII). First Author: Jun Huang, Department of Colorectal
Center, Goyang, Korea, Republic of (South) Surgery, The Sixth Affiliated Hospital, Sun Yat-sen University, Guangzhou, China
Background: Tumor-informed circulating tumor DNA (ctDNA) analysis allows for the Background: Immunotherapy has shown promising therapeutic effects in mismatch repair-
sensitive detection of minimal residual disease (MRD) and has the potential to enhance deficient or microsatellite instability-high (dMMR/MSI-H) colorectal cancer (CRC). However,
patient stratification for adjuvant chemotherapy (AC). In patients with stage II-III colon for patients with mismatch repair-proficient or microsatellite stable (pMMR/MSS) CRC, the
cancer, we demonstrated that postoperative MRD is associated with poor disease-free efficacy of single-agent PD-1 monoclonal antibody remains limited. Previous studies re-
survival (DFS) despite oxaliplatin-based AC. We hypothesize that intensifying AC in colon ported that combining anti-angiogenic drugs with PD-1 monoclonal antibody might improve
cancer patients with postoperative MRD may improve survival outcomes. the efficacy of immunotherapy. Our BASKETII study (NCT04895137) demonstrated that the
Methods: This multi-center platform trial (NCT05534087) consists of a prospective neoadjuvant therapy regimen of mFOLFOX6 combined with Bevacizumab and sintilimab
observational study (Part 1) and an interventional randomized trial (Part 2). In Part 1, significantly enhanced the immunotherapy sensitivity of pMMR/MSS locally advanced CRC
approximately 1,200 patients with colon cancer will be screened for MRD at 3–6 weeks (LACRC), resulting in improved pathological complete response (pCR) rates and higher R0
postoperatively using atumor-informed hybrid-capture-based ctDNA MRD assay resection rates. Methods: BASKETIII is a multicenter, randomized controlled, phase III
(CancerDETECT) which tracks ~100 patient-specific somatic variants identified through study with a parallel design conducted in China. This trial aims to evaluate whether the
tumor whole-exome sequencing. Key eligibility criteria include: age $19 neoadjuvant therapy regimen of mFOLFOX6 combined with Bevacizumab and sintilimab can
further improve survival outcomes, and maintain the higher pCR rate and acceptable safety
years, #6 weeks post-curative resection, pathological diagnosis of colon cancer, stage
profile compared to mFOLFOX6 in pMMR/MSS LACRC patients. Eligible participants will be
III or stage II with high-risk features requiring AC with FOLFOX/CAPOX, and no mac-
randomly assigned in a 1:1 ratio to either the experimental group or the control group.
roscopic residual disease. All patients in Part 1 will complete 3 months of standard Participants in the experimental group will receive the neoadjuvant therapy regimen of
adjuvant FOLFOX/CAPOX while awaiting MRD results. After 3 months of AC, MRD- mFOLFOX6 + Bevacizumab + sintilimab. The first five doses will follow the mFOLFOX6
negative patients are managed at the investigator’s discretion. Patients with MRD combined with Bevacizumab and sintilimab regimen, and the sixth dose will receive only
positivity will be screened for Part 2 clinical trial following the completion of initial mFOLFOX6 and sintilimab but without Bevacizumab, in order to avoid delay of surgery.
3 months of standard AC titled “Randomized Controlled Phase III Trial of Treatment Participants in the control group will receive the neoadjuvant therapy regimen of mFOLFOX6
Intensification in Stage II–III Colon Cancer Patients with Positive MRD after Curative alone. Participants in both groups will undergo radical surgical treatment after neoadjuvant
Resection.” Part 2 investigates the superiority of an experimental arm (modified therapies. Participants who achieve pCR based on postoperative pathology will be regularly
FOLFIRINOX for 3 months) compared to a control arm (FOLFOX/CAPOX for 3 months). followed up. Participants who do not achieve pCR will receive adjuvant therapy with a maxim
The primary endpoint is the 3-year DFS rate, while secondary endpoints include the 5- of six doses and will be regular followed up after the final dose of adjuvant therapy. The
year overall survival rate, treatment-related adverse events, treatment adherence, and primary outcome of this study is to evaluate the 3-year disease-free survival (DFS). The key
patient-reported outcomes. A total of 236 patients will be enrolled, assuming a hazard inclusion criteria include histologically confirmed adenocarcinoma of the colon or upper
ratio of 0.64, 80% power, a two-sided alpha of 0.05, and a 10% dropout rate. As of rectum; tumor biopsy immunohistochemical identified pMMR or MSS identified through
November 2024, 630 patients have been screened in Part 1, and 99 patients have been next-generation sequencing or polymerase chain reaction; Clinical staging of cT4NxM0. The
enrolled in Part 2. Both studies are ongoing, and an interim analysis is planned after $ 48 main exclusion criteria include evidence of distant metastasis beyond the pelvic region;
events. Clinical trial information: NCT05534087. Research Sponsor: National R&D history of pelvic or abdominal radiotherapy; multiple CRC or multiple primary tumors; history
Program for Cancer Control through the National Cancer Center (NCC) funded by the of immunotherapy and other malignancies within the past 5 years. A total of 122 patients are
Ministry of Health & Welfare, Republic of Korea (HA22C0062). planned to be enrolled in this study. This study is registered with [Link]
(NCT06791512) and is recruiting. Clinical trial information: NCT06791512. Research
Sponsor: National Natural Science Foundation of China.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—COLORECTAL AND ANAL 271s

TPS3647 Poster Session TPS3648 Poster Session

A precision medicine trial leveraging tissue and blood-based tumor geno- Phase II study of epacadostat (INCB024360) added to preoperative chemo-
mics to optimize treatment in resected stage III and high-risk stage II colon radiation in patients with locally advanced rectal cancer. First Author: Moh’d M.
cancer (CC) patients (pts): The SAGITTARIUS trial. First Author: Clara O. Khushman, Washington University School of Medicine, St. Louis, MO
Montagut, Hospital del Mar Research Institute, Barcelona, Spain Background: Indoleamine 2,3-dioxygenase 1 (IDO1) metabolizes tryptophan along the
Background: Circulating tumor DNA (ctDNA) testing has emerged as a transformative tool kynurenine pathway and is recognized as a potent suppressor of tumor reactive im-
for detecting molecular residual disease (MRD). Multiple prospective trials have dem- munity. Epacadostat is an orally active, potent and selective inhibitor of IDO1. In
onstrated the potential of ctDNA in guiding treatment decisions for stage II-III CC pts. preclinical studies, IDO1 was found to promote resistance to radiation in rectal cancer,
Numerous ongoing randomized clinical trials (RCTs) are adjusting adjuvant chemotherapy irrespective of microsatellite instability (MSI) status. IDO1 inhibition with epacadostat
(ACT) intensity based on MRD status. However, data from ctDNA-guided trials, including improved tumor radiosensitivity by relieving immune suppression and augmenting
PEGASUS (NCT04259944), reveal that intensified ACT is curative in only a small proportion radiation-induced apoptosis while protecting the normal intestine from radiation
of MRD cases. To address this limitation, the SAGITTARIUS RCT was designed to evaluate damage. In a phase 1 trial, 17 patients were enrolled from 4/2019 to 8/2023. Epaca-
whether combining ctDNA detection with targeted agents selected on the basis of tissue- dostat in combination with short-course radiation therapy (SCRT) and CAPOX was well-
based comprehensive genomic profiling (CGP) can optimize treatment in high risk (i.e. tolerated and the recommended phase 2 dose (RP2D) of epacadostat was determined to
MRD+) pts while sparing low risk (i.e. MRD2) from unnecessary toxicity. be 400 mg BID. An NCI supported Phase 2 trial is ongoing to further evaluate the
Methods: SAGITTARIUS is a Phase III RCT evaluating ctDNA and tissue-guided per- promising disease responses reported in the dose escalation phase. Methods: This
sonalized post-surgical management in resected stage III and high-risk stage II CC pts. phase 2 multicenter, open-label trial includes treatment and biomarker cohorts. In the
Tumor-informed, personalized ctDNA test (Signatera, Natera, Inc.) and CGP (TruSight™ treatment cohort, epacadostat at 400 mg BID will be administered concurrently with
Oncology Comprehensive EU, Illumina, Inc.) are used to determine MRD status and tumor SCRT followed by epacadostat monotherapy until 1 day prior to neoadjuvant chemo-
genomic landscape, respectively, including genetic mutation (mut) and amplification therapy, followed by standard-of-care (SOC) neoadjuvant chemotherapy and, ultimately,
(ampl), tumor mutational burden (TMB) and microsatellite instability (MSI) status. Pts are surgical resection or non-operative management (NOM). Biomarker cohort enrollment
stratified based on post-surgery (3-5 weeks) ctDNA status into two embedded RCTs:Trial- will commence at completion of treatment cohort accrual. Enrolled patients will be
1)ctDNA-positive (ctDNA+) ptsare further stratified based on MSI and RAS/RAF status and treated with SOC SCRT followed by SOC neoadjuvant chemotherapy and surgical re-
randomized 1:1 to standard 6-month ACT (CAPOX/FOLFOX) or personalized treatment (PT)
section or NOM. Eligible patients must be a treatment-naı̈ve, newly diagnosed, path-
guided by CGP biomarkers with reassessment of ctDNA status to guiding subsequent
ologically confirmed, locally advanced rectal cancer (defined by 8th edition AJCC stage 2
therapies (chemotherapy regimens in ctDNA+ or maintenance and follow-up in sero-
or 3, or stage 1 not eligible for sphincter-sparing surgery) with plans to proceed with
converted; Trial-2) ctDNA-negative (ctDNA2) ptsare randomized 1:1 to a physician-choice
neoadjuvant SCRT and chemotherapy. The primary endpoint is the neoadjuvant rectal
strategy or observation with ctDNA reassessed at 2 and 4 months and, in cases of
positivity, cross over to Trial-1. PT include 3-month CAPOX followed by FOLFIRI or TEMIRI
(NAR) score. Secondary endpoints are pathologic complete response (pCR) rate,
based on MGMT status (RAS/RAFmut), Ipilimumab + nivolumab (MSI and TMB-high complete clinical response (cCR) rate and progression-free survival (PFS). Exploratory
POLEmut), pertuzumab + trastuzumab (HER2ampl), FOLFOX + panitumumab (RAS/RAF/ endpoints are pharmacodynamics, PDX and organoid generation, identification of
HER2 wild-type). The primary endpoint (EP) is 2-year recurrence-free survival (RFS) in molecular predictors of response and resistance, correlation of radiographic and
ctDNA+ pts. Secondary EPs include 2-year RFS in ctDNA2 pts, 3- and 5-year overall pathologic response and effect of treatment on patient quality of life. We aim to enroll 27
survival, and ctDNA conversion rate. Quality of life and health costs data are collected for patients in the treatment cohort and 10 in the biomarker cohort. Clinical trial infor-
cost effectiveness analysis. Biospecimens, including archival tumor tissue, serial blood mation: Clinical Trial Registration: NCT03516708. Research Sponsor: NIH (grant number
samples, and buccal swabs, are collected for exploratory analyses. To detect a hazard ratio 1R01CA278197-01A1); Incyte (drug only).
of 0.6325 for ctDNA-guided PT vs standard ACT, 200 ctDNA+ pts will be randomized in
Trial-1. Recruitment began in October 2024 across 26 institutions in Italy, Spain, and
Germany. Clinical trial information: NCT06490536. Research Sponsor: European Union;
101104657.

TPS3649 Poster Session TPS3650 Poster Session

AZUR-4, a phase 2, open label, randomized study of neoadjuvant dostarli- Trials in progress: Alliance A022104/NRG-GI010—A randomized phase II/III
mab plus capecitabine plus oxaliplatin (CAPEOX) versus CAPEOX alone in trial testing the efficacy of triplet versus doublet chemotherapy regarding
previously untreated T4N0 or stage III mismatch repair proficient/ clinical complete response and disease-free survival in patients with locally
microsatellite stable resectable colon cancer. First Author: Gertjan Ras- advanced rectal cancer (LARC; the Janus Rectal Cancer trial). First Author:
schaert, Department of Gastroenterology/Digestive Oncology, University Hospital Caleah Simone Kitchens, Memorial Sloan Kettering Cancer Center, New York, NY
Leuven, Leuven, Belgium Background: A total neoadjuvant therapy (TNT) approach improves compliance with che-
Background: Colon cancer is the third most common cancer globally, with a standard of motherapy and increases rates of tumor response compared to neoadjuvant chemoradiation
care in the nonmetastatic setting that includes surgery followed by adjuvant chemotherapy. (CRT) alone in those with locally advanced rectal cancer. Recent data indicate that optimal
Results of recent clinical trials suggest that neoadjuvant therapy may be beneficial in locally sequencing of TNT involves consolidation (rather than induction) chemotherapy to improve
advanced colon cancer. Neoadjuvant immunotherapy has shown impressive responses in complete response rates. The use of FOLFIRINOX has shown to improve response and outcomes
mismatch repair deficient (dMMR)/microsatellite instability-high (MSI-H) disease with compared to CRT and surgery alone. Data have also shown that patients with clinical complete
pathological complete responses of up to 67% and 3-year disease-free survival of 100% response (cCR) after TNT may be managed with a watch and wait approach (WW) instead of
reported in the NICHE 2 trial. However, most colon cancer (85%–90%) is mismatch repair preemptive total mesorectal resection (TME). However, the optimal consolidation chemotherapy
proficient (MMRp)/microsatellite stable (MSS), which has been shown to have poor re- regimen to improve cCR rates has not been established, and a randomized clinical trial has not
sponse to conventional immunotherapy. Dostarlimab, a programmed cell death protein-1 robustly evaluated cCR as a primary endpoint. We designed this NCI-sponsored study of
chemotherapy intensification to address this and to increase cCR rates, provide opportunity for
(PD-1) inhibitor, has a high affinity for binding to PD-1, blocking the interaction between PD-
organ preservation, and survival outcomes. Methods: In this multigroup randomized, seamless
1 and its ligands (PD-L1 and PD-L2). Dostarlimab monotherapy has been approved in the US
phase II/III trial (1:1), up to 760 patients with LARC, T4N0, any T with node positive disease (any
for the treatment of adults with dMMR advanced/recurrent solid tumors. The AZUR-4 trial
T, N+) or T3N0 requiring abdominoperineal resection or coloanal anastomosis and distal margin
(NCT06567782) evaluates dostarlimab + CAPEOX versus CAPEOX alone as neoadjuvant within 12 cm of anal verge will be enrolled. Stratification factors include tumor stage (T4 vs T1-
treatment to identify early signals of efficacy in resectable MMRp/MSS colon cancer. The 3), nodal stage (N+ vs N0) and distance from anal verge (0–4; 4–8; 8–12 cm). Patients will be
study will assess the relationship between conventional and advanced blood- and tumor- randomized to receive neoadjuvant long-course chemoradiation (LCRT) followed by consoli-
based immune response to better understand the contribution of dostarlimab to patho- dation doublet (mFOLFOX6 or CAPOX (control arm)) or triplet chemotherapy (FOLFIRINOX
logical response. Methods: AZUR-4 is a multicenter, randomized, open-label phase 2 study (experimental arm)) for 3–4 months. LCRT in both arms involve 4500 cGy in 25 fractions over
in MMRp/MSS resectable colon cancer. Approximately 120 patients will be enrolled and 5 weeks +900 cGy boost in 5 fractions with a fluoropyrimidine. Patients will undergo assessment
randomized 3:1 to the dostarlimab + CAPEOX and CAPEOX arms, respectively, in which they 8–12 (64) weeks post-TNT completion. The primary endpoint for the phase II portion will
will receive 4 cycles of Q3W neoadjuvant therapy. Key eligibility criteria include age $18 compare cCR between treatment arms. A total number of 312 patients (156 per arm) will provide
years, confirmed resectable MMRp/MSS colon adenocarcinoma with no prior treatment, statistical power of 90.5% to detect a 17% increase in cCR rate, at a one-sided alpha = 0.048. The
clinically staged as T4N0 or stage III, Eastern Cooperative Oncology Group performance primary endpoint for the phase III portion will compare disease-free survival (DFS) between
status of 0 or 1, and required tissue biopsies providing fresh tumor tissue either at pre- arms. A total of 285 DFS events will provide 85% power to detect an effect size of hazard ratio
screening or screening. Primary endpoints are major pathologic response rate (mPR) 0.70 at a one-sided alpha of 0.025, requiring enrollment of 760 patients (380 per arm). Secondary
assessed at #10% residual viable tumor (RVT) and treatment-emergent adverse events objectives include overall survival, organ preservation time, time to distant metastasis, and
(AEs), serious AEs, immune-mediated AEs, and AEs leading to death or discontinuation of adverse event rates. This study has accrued 587 patients as of January 2025, and is inves-
study drug. Secondary endpoints include primary tumor resection not being excluded by tigating exploratory correlatives (e.g., ctDNA). Support: U10CA180821, U10CA180882, U24
either disease progression or treatment-related toxicities, and pathological response CA196171. [Link] [Link] ID: NCT05610163
categories that include complete pathological response (cPR) and partial pathologic re- Clinical trial information: NCT05610163. Research Sponsor: National Cancer Institute;
sponse (pPR). Exploratory endpoints include overall survival, event-free survival, effects on U10CA180821; National Cancer Institute; U10CA180882; U.S. National Institutes of Health;
circulating tumor DNA dynamics, and pathological response rate in biomarker subsets. U24CA196171; ECOG-ACRIN MEDICAL RESEARCH FOUNDATION; U10CA180820; National
Clinical trial information: NCT06567782. Research Sponsor: GSK. Cancer Institute; U10CA18086; National Cancer Institute; U10CA180888.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
272s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4000 Oral Abstract Session 4001 Oral Abstract Session

Adjuvant nivolumab in resected esophageal or gastroesophageal junction Lenvatinib plus pembrolizumab and chemotherapy versus chemotherapy in
cancer (EC/GEJC) following neoadjuvant chemoradiotherapy (CRT): First advanced, metastatic gastroesophageal adenocarcinoma: The phase 3,
results of overall survival (OS) from CheckMate 577. First Author: Ronan Joseph randomized LEAP-015 study. First Author: Kohei Shitara, National Cancer Cen-
Kelly, Charles A. Sammons Cancer Center at Baylor University Medical Center, Dallas, TX ter Hospital East, Chiba, Japan
Background: At 24.4-month (mo) median follow-up, adjuvant nivolumab demonstrated a sta- Background: LEAP-015 (NCT04662710), is a randomized, open-label, phase 3 study of
tistically significant and clinically meaningful improvement in disease-free survival (DFS) vs pembrolizumab plus lenvatinib and chemotherapy as first-line treatment for advanced/
placebo with a well-tolerated safety profile in patients (pts) with resected EC/GEJC with residual metastatic gastroesophageal adenocarcinoma. We report results from the interim and final
pathologic disease following neoadjuvant CRT and surgery in the primary analysis from the global, analyses of LEAP-015. Methods: Eligible participants (pts) had untreated HER-2 negative
phase 3 CheckMate 577 study (NCT02743494). We report the final analysis of the hierarchically locally advanced unresectable or metastatic gastroesophageal adenocarcinoma, measurable
tested secondary endpoint of OS along with longer follow-up of DFS. Methods: Adults with disease and ECOG PS 0-1. All pts were randomly assigned 1:1 to induction with pembrolizumab
resected (R0) stage II/III EC/GEJC who received neoadjuvant CRT and had residual pathologic 400 mg IV Q6W (x2) plus oral lenvatinib 8 mg QD and investigators choice chemotherapy
disease were randomized 2:1 to nivolumab 240 mg or placebo Q2W for 16 weeks, followed by (CAPOX Q3W x4 or mFOLFOX6 Q2W x6) then consolidation with pembrolizumab 400 mg Q6W
nivolumab 480 mg or placebo Q4W. Maximum treatment duration was 1 year. The primary endpoint for #16 doses plus lenvatinib 20 mg QD (only if 8 mg tolerated for at least 3 weeks), or
was DFS. OS was a secondary endpoint, and exploratory endpoints included safety, distant chemotherapy alone (CAPOX or FOLFOX). Randomization was stratified by region, ECOG PS,
metastasis-free survival (DMFS), and progression-free survival on subsequent systemic therapy and chemotherapy choice. Dual primary endpoints were PFS (RECIST v1.1, BICR) and OS in pts
(PFS2). Results: 794 pts were randomized (nivolumab, n = 532; placebo, n = 262). With a median
with PD-L1 combined positive score (CPS) $1 and in all pts; secondary endpoints included
follow-up of 78.3 (range, 60.1–96.6) mo, adjuvant nivolumab continued to show DFS benefit vs
ORR and DOR (RECIST v1.1, BICR) in pts with PD-L1 CPS $1 and in all pts, and safety and
placebo (HR 0.76 [95% CI 0.63–0.91]; Table). Median OS was numerically longer with nivolumab vs
tolerability in all pts. The data cut-off date was Oct 29, 2024. Results: A total of 880 pts (78%
placebo (51.7 vs 35.3 mo), although the difference was not statistically significant (HR 0.85
[95.87% CI 0.70–1.04]; P = 0.1064; Table). OS rates at 3 and 5 years with nivolumab vs placebo were PD-L1 CPS $1; 75% gastric primary) were randomized (443 pembrolizumab plus lenvatinib
57% vs 50% and 46% vs 41%, respectively. OS subgroup analyses will be presented. Clinically and chemotherapy; 437 chemotherapy alone).Median follow-up was 32.2 mo (range 19.0 –
meaningful improvement in DMFS with nivolumab vs placebo was maintained (Table). PFS2 41.7) in pts with PD-L1 CPS $1 and 31.8 mo (range, 19.0 – 41.7) in all pts. At interim analysis,
favored nivolumab vs placebo (HR 0.81 [95% CI 0.67–0.98]). In the nivolumab group, 46% of pts PFS difference was statistically significant with pembrolizumab plus lenvatinib and che-
received subsequent therapy vs 60% in the placebo group; 5% vs 15% received subsequent motherapy vs chemotherapy in pts with PD-L1 CPS $1 (median 7.3 vs 6.9 mo; HR 0.75; 95% CI,
immunotherapy. No new safety signals were identified. Conclusions: Adjuvant nivolumab dem- 0.62-0.9; P = 0.0012), with 24-mo PFS of 20% vs 7%, and in all pts (median 7.2 vs 7.0 mo; HR
onstrated sustained long-term DFS benefit and numerical improvement in OS vs placebo in pts with 0.78; 95% CI, 0.66-0.92; P = 0.0019), with 24-mo PFS of 21% vs 8%. ORR was 59.5% vs 45.4% in
resected EC/GEJC and residual pathologic disease following neoadjuvant CRT. The safety profile of pts with PD-L1 CPS $1 and 58.0% vs 43.9% in all pts; P , 0.0001 for both. At final analysis, OS
adjuvant nivolumab remained well-tolerated with longer follow-up. These results further support in pts with PD-L1 CPS $1 was not statistically significant (median 12.6 vs 12.9 mo; HR 0.84;
the use of adjuvant nivolumab in this pt population. Clinical trial information: NCT02743494. 95% CI, 0.71-1.00; P = 0.0244 (P-value boundary for significance of 0.0204), with 24-mo OS of
Research Sponsor: Bristol Myers Squibb. 31% vs 23%. OS in all pts was not tested per multiplicity strategy (median 13.1 vs 13.0; HR 0.87;
95% CI 0.75-1.01). Drug-related adverse event (AE) rates were 98% vs 92% in pts receiving
Nivolumab Placebo
Efficacy (n = 532) (n = 262) pembrolizumab plus lenvatinib and chemotherapy vs chemotherapy. Grade $3 drug-related
AE rates were 65% vs 49% (grade 5 AEs 5% vs , 1%). Conclusions: Pembrolizumab plus
Median DFS (95% CI), mo 21.8 (16.6–29.7) 10.8 (8.3–14.3)
HR (95% CI) 0.76 (0.63–0.91)
lenvatinib and chemotherapy vs chemotherapy provided statistically significant improvement
Median OS (95% CI), mo 51.7 (41.0–61.6) 35.3 (30.7–48.8) in PFS and ORR in pts with advanced unresectable or metastatic gastroesophageal carcinoma
HR (95.87% CI; P value) 0.85 (0.70–1.04; P = 0.1064) at interim analysis. However, there was no significant improvement in OS in pts with PD-L1
Median DMFS (95% CI), mo 27.3 (21.4–36.0) 14.6 (10.9–20.3) CPS $1 at final analysis. Safety profiles were consistent with known regimens, with higher AE
HR (95% CI) 0.75 (0.62–0.90)
Safety, n (%) n = 532 n = 260 rates seen in pts receiving the experimental treatment. Clinical trial information:
Any-grade/grade 3–4 TRAEs 379 (71)/75 (14) 124 (48)/17 (7) NCT04662710. Research Sponsor: Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc.,
Any-grade/grade 3–4 TRAEs leading to discontinuation 48 (9)/26 (5) 8 (3)/7 (3) Rahway, NJ, USA; N/A.
TRAE, treatment-related adverse event.

LBA4002 Oral Abstract Session 4003 Oral Abstract Session

Trastuzumab deruxtecan (T-DXd) vs ramucirumab (RAM) + paclitaxel (PTX) Claudin18.2-specific CAR T cells (Satri-cel) versus treatment of physician’s
in second-line treatment of patients (pts) with human epidermal growth choice (TPC) for previously treated advanced gastric or gastroesophageal
factor receptor 2-positive (HER2+) unresectable/metastatic gastric cancer junction cancer (G/GEJC): Primary results from a randomized, open-label,
(GC) or gastroesophageal junction adenocarcinoma (GEJA): Primary anal- phase II trial (CT041-ST-01). First Author: Changsong Qi, Peking University Cancer
ysis of the randomized, phase 3 DESTINY-Gastric04 study. First Author: Kohei Hospital, Beijing, Beijing, China
Shitara, National Cancer Center Hospital East, Kashiwa, Japan Background: Claudin18.2 (CLDN18.2) has emerged as a promising therapeutic target in G/GEJC. Recently
reported results showed CT041/satricabtagene autoleucel (satri-cel), an autologous CLDN18.2-specific
CAR T-therapy, had encouraging efficacy in previously treated patients (pts) with advanced G/GEJC. Now
we report the primary results from the phase II pivotal trial (CT041-ST-01, NCT04581473). Methods: In
this open-label, multicenter, randomized controlled trial (RCT) conducted in China, CLDN18.2 positive,
advanced G/GEJC pts with failure to at least 2 prior lines of treatment, were randomized (2:1) to satri-cel
arm or TPC arm. For satri-cel arm, satri-cel dose of 250 3106 cells were infused up to 3 times. For TPC arm,
one of the standard of care (SOC) drugs (apatinib, paclitaxel, docetaxel, irinotecan or nivolumab) was given
per physician’s decision. Those who experienced disease progression or drug intolerance in TPC arm could
receive subsequent satri-cel, if eligible. The primary endpoint was PFS assessed by the Independent
Review Committee (IRC). Key secondary endpoint was OS. Data cutoff date was Oct 18, 2024.
Results: From Mar 29, 2022 to Aug 16, 2024, a total of 156 pts were randomized to satri-cel arm (n = 104)
or TPC arm (n = 52). Twenty pts in TPC arm received subsequent satri-cel. Median number of prior
The full, final text of this abstract will be available at systemic therapies was 2 in both arms, and 26.9% vs 19.2% had received $3 lines; 69.2% vs 59.6% had
[Link] on the day of presentation and in the peritoneal metastasis; 71.2% vs 65.4% were Lauren diffuse/mixed type. The median follow-up time of PFS
and OS was 8.90 and 12.29 months (m). In ITT population, satri-cel arm showed significant improvement in
online supplement to the June 10, 2025, issue of the Journal mPFS by IRC (3.25m vs 1.77m; HR 0.366, 95% CI: 0.241, 0.557; p , 0.0001) and an obvious trend for longer
of Clinical Oncology. mOS (7.92m vs 5.49m; HR 0.693, 95% CI: 0.457, 1.051; one-sided p = 0.0416) than TPC arm. Moreover, in
136 pts receiving study drug (mITT, satri-cel 88 pts vs TPC 48 pts), mPFS by IRC was 4.37m vs 1.84m, HR
0.304 (95% CI: 0.195, 0.474) and mOS was 8.61m vs 5.49m, HR 0.601 (95% CI: 0.385, 0.939). Notably, mOS
of TPC pts with satri-cel was 9.20m. Among all pts receiving satri-cel (n = 108) vs TPC pts without satri-cel
(n = 28), mOS was 9.17m vs 3.98m, HR 0.288 (95% CI: 0.169, 0.492). A summary of study drug-related
adverse events (TRAEs) is shown in the Table. Conclusions: This is the first confirmatory RCT of CAR T-
therapy in solid tumors. Satri-cel demonstrated significant PFS improvement and an obvious OS benefit
with a manageable safety profile. These results support satri-cel as a potential new SOC for advanced
G/GEJC. Clinical trial information: NCT04581473. Research Sponsor: None.
Satri-cel TPC
Safety, n (%) (n=88) (n=48)
TRAEs 88 (100) 44 (91.7)
Serious TRAEs 31 (35.2) 12 (25.0)
TRAE leading to death 1 (1.1)a 1 (2.1)b
CRS 84 (95.5) 0
Grade 1-2 80 (90.9) 0
Grade 3 4 (4.5) 0
ICANS 0 0
a
disseminated intravascular coagulation;
b
coagulopathy.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 273s

LBA4004 Oral Abstract Session LBA4005 Oral Abstract Session

Results of a randomized phase III trial of pre-operative chemotherapy with PANOVA-3: Phase 3 study of tumor treating fields (TTFields) with gemci-
mFOLFIRINOX or PAXG regimen for stage I-III pancreatic ductal tabine and nab-paclitaxel for locally advanced pancreatic ductal adenocar-
adenocarcinoma. First Author: Michele Reni, Department of Medical Oncology, cinoma (LA-PAC). First Author: Vincent J. Picozzi, Virginia Mason Medical Center,
Pancreas Translational and Clinical Research Center, IRCCS San Raffaele Scientific Seattle, WA
Institute, Vita-Salute San Raffaele University, Milan, Italy

The full, final text of this abstract will be available at The full, final text of this abstract will be available at
[Link] on the day of presentation and in the [Link] on the day of presentation and in the
online supplement to the June 10, 2025, issue of the Journal online supplement to the June 10, 2025, issue of the Journal
of Clinical Oncology. of Clinical Oncology.

4006 Oral Abstract Session 4007 Oral Abstract Session

Preliminary results from the randomized phase 2 study (1801 part 3B) of A phase III randomized clinical trial evaluating perioperative therapy (neo-
elraglusib in combination with gemcitabine/nab-paclitaxel (GnP) versus adjuvant chemotherapy versus chemoradiotherapy) in locally advanced
GnP alone in patients (pts) with previously untreated metastatic pancreatic gallbladder cancers (POLCAGB). First Author: Reena Engineer, Tata Memorial
ductal adenocarcinoma (mPDAC). First Author: Devalingam Mahalingam, Robert H. Centre (HBNI), Mumbai, India
Lurie Comprehensive Cancer Center, Northwestern University, Chicago, IL Background: Locally advanced Gallbladder cancers (LAGBC)initially deemed not suitable for R0
Background: Elraglusib (9-ING-41) is a first-in-class inhibitor of GSK-3ß, which may mediate drug resistance, resection receive either neoadjuvant chemotherapy (NACT)or neoadjuvant chemoradiation
EMT, and damaged DNA and tumor immune response in advanced cancer. In pancreatic cancer models in mice, (NACRT)downstaging for resection and to improve outcomes. Methods: This is a randomized
elraglusib combined with chemotherapy enhanced anti-tumor effects and survival. In a single-arm clinical study, phase 3 trial (NCT02867865) that included fit patients with LAGBC adenocarcinoma,. T3/T4 with
elraglusib/GnP showed antitumor activity and prolonged survival in pts with mPDAC. Methods: Pts with liver infiltration ( . 2cm, , 5cm); N1 nodal status; obstructive jaundice (type I/II biliary ob-
previously untreated mPDAC were randomized 2:1 to GnP plus elraglusib 9.3 mg/kg IV once weekly or GnP in an
struction) ;duodenal or colonic abutment with no mucosal infiltration , , 180°vascular in-
open-label phase 2 study. The primary endpoint was 1-yr OS in the primary analysis set. Upon study completion,
mOS will be the primary endpoint once survival distributions are compared after the 12-month follow-up using volvement. The patients were randomized (1:1) to NACT arm (Gemcitabine + platinum for four
log-rank analysis. Secondary endpoints included DCR, ORR, mPFS, and TEAEs/TRAEs. The planned sample size cycles) versus NACRT arm (55 -57Gy with concurrent gemcitabine followed by two cycles of
was 207 evaluable pts (130 for elraglusib/GnP and 77 for GnP), assuming 1-yr OS of 55% with elraglusib/GnP and chemotherapy) and were then evaluated for surgery. A sample size of 314 patients was required to
35% with GnP to achieve 80% power with a chi-square test at a 2-sided 5% a. For OS, nonparametric log-rank test detect a 5.5 months difference (11 [Link] 16.5 mo in the test arm) with median overall survival (OS)
was used with statistical significance at p-value , 0.048. Cytokine/chemokine correlative biomarker assays as the primary endpoint (hazard ratio, 0.7; 2-sided a = 0.05; b = 0.2). Secondary endpoints were
were performed. The study completed enrollment in February 2024. Results: As of November 15, 2024 event free survival (EFS),R0 resection rates and post-operative complication rates. Due to slow
(preliminary data cut-off date), the primary analysis set included 155 pts in the elraglusib/GnP arm and 78 pts in accrual Institutional Ethics committee requested the investigators for interim analysis and ap-
the GnP arm, with 52.8% males and 57.5% ECOG PS 1. Median (range) CA 19-9 levels were 1568 U/mL (1 to proval was obtained for the same. Results: From Oct 2016 to Sept 2024, 124 patients (64 NACT, 60
381,904 U/mL) in the elraglusib/GnP arm and 1590 U/mL (2 to 501,000 IU/mL) in the GnP arm. The 1-yr OS rate
NACTRT)were enrolled at 2 centers. At the time of analysis 93 OS events were observed in 124
was 43.6% with elraglusib/GnP vs 22.5% with GnP (z-test p = 0.002); the mOS was 9.3 mo with elraglusib/GnP vs
7.2 mo with GnP (HR, 0.63; log rank p = 0.016; see Table). 38.1% of pts on elraglusib/GnP and 19.2% on GnP are patients. Median follow-up was 62 (range 6.9-94) months. More number of patients underwent R0
censored, with the majority at . 10 months OS. Several biomarkers appear to be predictive for OS including IFNb resection in NACRT than NACT arm 51.6 vs 29.7% (p = 0.01) In the intention to treat analysis, the
and PD-L1. The most common TRAE with elraglusib/GnP was grade 1-2 transient visual impairment in . 60% of NACTRT arm showed improved OS compared to the NACT arm [21.8 [Link].10.1 mo. p = 0.006]. EFS
patients (vs 9% with GnP). Grade $3 TEAEs occurred in 89.7% of pts on elraglusib/GnP and 80.8% on GnP. Most was 10.6 mo. vs 4.9 months, p = 0.006]. Similar results were noted in the per protocol analysis(n =
common grade $3 TEAEs with elraglusib/GnP (vs GnP) were neutropenia 51.6% (vs 29.5%), anemia 24.5% (vs 110). Clavien Dindo postoperative morbidity of grade 3 and above was 4/22 (18.18%) in NACT arm
29.5%), and fatigue 16.1% (vs 5.1%). Conclusions: The preliminary results showed a statistically significant vs 9/32(28.12%) in NACRT arm (p = 0.30). The interim analysis demonstrated a significant im-
benefit for 1-yr OS and mOS and favorable trends for ORR and DCR with elraglusib/GnP over GnP, with provement in efficacy in the NACRT arm. Based on the current data, the conditional power was
manageable safety profile. The mOS for GnP is lower relative to MPACT and NAPOLI-3 but comparable to recent
calculated to be 99.96%. Conclusions: This trial demonstrates that the addition of concurrent
real-world meta-analyses, explained by advanced disease burden and higher mortality rate in the first 4 months
in our study. Topline analysis (April 2025) and correlative biomarker analysis predictive for OS will be presented.
chemoradiation to chemotherapy improves overall survival and resection rates in patients with
Clinical trial information: NCT03678883. Research Sponsor: Actuate Therapeutics, Inc. locally advanced gallbladder cancers. These results provide important evidence to guide treatment
decisions in this traditionally difficult to treat set of gallbladder cancers. Clinical trial information:
Preliminary efficacy results.
NCT02867865. Research Sponsor: Intramural funding - Tata Memorial Centre.
Elraglusib/GnP GnP
n=155 n=78 Treatment outcomes.
1-year OS, % 43.6 22.5 NACTRT NACT
z-test p=0.002 Outcome measures N=60 N=64 HR p
mOS, mo 9.3 7.2
HR=0.63; log-rank p=0.016 Patients surgically explored 39(65%) 29(45.3%) 0.03
Events, n (%) 96 (61.9) 63 (80.8) Patients undergoing R0 resection 31 (51.6%) 19 (29.7%) 0.01
mPFS, mo 5.6 4.9 Median OS (months) (95% CI) 21.8 (14.6-29.14) 10.1 (8.5-11.7) HR- 0.56 0.006
HR=0.91; p=NS 5 Year survival (95% CI) 27 (17.7-43)% 18 (10-31) % 95%CI- 0.37-0.84
Events, n (%) 128 (82.6) 70 (89.7) EFS (months) (95% CI) 10.6 (6.07-15.5-15) 4.89 (3.06-6.73) HR-0.58 0.006
DCR, % 42.6 33.3
ORR, n (%) 43 (27.7) 16 (20.5)
5 Yr EFS (%) 21 (12-35.7)% 12.7 (6-24.6)% 95%CI- 0.39-0.85

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
274s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4008 Oral Abstract Session 4009 Clinical Science Symposium

Neoadjuvant chemotherapy with gemcitabine plus cisplatin followed by Maintenance with OSE2101 plus FOLFIRI vs FOLFIRI alone after FOLFIR-
radical liver resection versus immediate radical liver resection alone fol- INOX (FFX) induction in patients (Pts) with advanced pancreatic ductal
lowed adjuvant therapy in biliary tract cancer: Final results from the phase III adenocarcinoma (aPDAC): Primary endpoint results of a randomized TEDO-
AIO/CALGP/ACO-GAIN-Trial. First Author: Thorsten Oliver Goetze, Institute of PAM GERCOR D17-01 PRODIGE 63 trial. First Author: Anthony Turpin, Lille
Clinical Cancer Research (IKF), Krankenhaus Nordwest, UCT-University Cancer Center, University Hospital, University of Lille, Lille, France
Frankfurt, Germany Background: OSE2101 is an off the shelf vaccine made of 10 synthetic HLA-A2-restricted peptides
Background: Radical surgical resection represents the only potentially curative treat- targeting 5 tumor associated antigens. This multicenter, randomized, non-comparative, phase II study
ment option for Biliary Tract Cancer (BTC) and (incidental) Gallbladder Carcinoma ((I) assessed FOLFIRI 6 OSE2101 maintenance in aPDAC Pts without progression after 8 cycles of FFX.
Methods: Eligible aPDAC Pts were randomized to FOLFIRI (Arm A) or FOLFIRI + OSE2101 (Arm B:
GBC). Nevertheless, 5-year OS is only 20–40% after curatively intended resection and data
subcutaneous injection on D1, D15, Q4W/6 doses then Q8W to M12 then Q12W up to M24).
regarding pure adjuvant chemotherapy in BTCs are currently conflicting. Encouraging Stratification factors: tumor stage (locally advanced vs metastatic), best response to FFX (partial or
results of neoadjuvant/perioperative concepts in other malignancies provide a rationale complete response [CR, PR] vs stable disease [SD]), and center. Primary endpoint: overall survival (OS)
to use this treatment in the early phase management of GBC and intrahepatic as well rate at M12 in evaluable Pts (M12-OS; Fleming 2-stage design, H0: 25%; H1: 50%, 1-sided alpha: 2.5%,
extrahepatic cholangiocarcinoma (ICC/ECC). Methods: GAIN is a multicenter, random- power: 90%); secondary endpoints: progression-free survival (PFS; RECIST v1.1), best response,
ized, controlled, open-label phase III trial, including patients (pts) with localized or locally duration of disease control (DDC), and safety. Results: 107 Pts (ITT) were randomized (53/Arm A, 54/
advanced resectable non metastatic biliary tract cancer (intra-/extrahepatic chol- Arm B) between 04/2021 and 05/2023. Median age 64 years (range:37-81), 53% men, 69% had
angiocarcinoma ICC/ECC; GBC in front of radical liver resection). Pts were randomized to metastases, 36%/64% had PR/SD to prior FFX. No evidence of imbalance in Pt characteristics was
observed between arms. Median number of OSE2101 injections was 7.5 (1-14). Median treatment
either neoadjuvant (perioperative) systemic chemotherapy (Gemcitabine + Cisplatin 3
duration of FOLFIRI was 5.4 months in both arms. At data cut-off (Dec 9, 2024), median follow-up was
cycles pre- and post-surgery) followed by radical surgery (Arm A) or to direct surgery 21.4 months with 101 evaluable Pts for M12-OS (49/Arm A, 52/Arm B; 4 consent withdrawals, 1 Pt’s
followed by adjuvant treatment (Arm B) according to investigators choice. Primary decision, 1 treatment interruption .4 weeks). Number of death events (n/%) was 19/35.8% in Arm A
endpoint was OS; secondary endpoints were PFS/EFS, R0-resection rate, toxicity, and 18/33.3% in Arm B. M12-OS (95%CI) was 61% (46.2%-74.8%) in Arm A and 65% (50.9%-78.0%) in
perioperative morbidity, mortality and QoL. Recruitment was stopped after enrollment of Arm B. Median (95%CI) OS and PFS (ITT) were 17.3 months (10.6–23.2) and 8.2 months (5.3–11.6) in
68 pts due to a slow enrollment rate. Results: Between Dec 2019 and Feb 2024, 68 pts Arm A, and 15.5 months (12.4–19.3) and 7.8 months (5.4–10.6) in Arm B. Other secondary endpoints
were randomized and the ITT comprised 32 pts in Arm A and 30 pts in Arm B. Baseline are described in Table. Among 33 Pts with SD to prior FFX in Arm B, 6 (18%) had CR/PR (1/5) when
characteristics were similar between arms (overall, male 55%; median age 66.0; cT3/T4 adding OSE2101 to FOLFIRI vs 5 (no CR) among 35 Pts in Arm A. In the safety population, 7 SAEs/6 Pts
29.0%; cN+ 30.6%; 37.1% ICC, 30.6% ECC and 32.3% GBC). 90.6% of pts in Arm A (12%) in Arm A and 22 SAEs/14 Pts (26%) in Arm B were reported. No unexpected SAEs were observed
with OSE2101 except 1 inappropriate administration, and no evidence of increased toxicity of FOLFIRI
completed all 3 pre-operative cycles. 43.8% in Arm A completed adjuvant treatment and with OSE2101. Conclusions: TEDOPAM met its primary objective with minimal toxicity and positive
23.3% in Arm B received adjuvant treatment. Median follow-up was 11.8 months. outcomes of adding OSE2101 cancer vaccine to maintenance FOLFIRI, albeit mitigated by unex-
Neoadjuvant treatment improved OS (mOS, Arm A 27.8 vs. 14.6 months Arm B; HR 0.46 pectedly favorable OS in the control arm. Two complete responses were observed when adding
[0.22 - 0.96]; p = 0.04) and R0 resection rate (62.5% vs 33.3%). This effect was also seen in OSE2101. Further follow-up is ongoing and translational analysis planned. Clinical trial information:
event-free survival. Postoperative morbidity rates were similar in both arms (33.3% (A) vs. NCT03806309. Research Sponsor: OSE Immunotherapeutics.
32% (B)) and the 30- and 90-days mortality rates were lower for Arm A (30-days: 4.2% vs.
Arm A Arm B
24%; 90-days: 4.2% vs. 28%). No new safety/toxicity signals were observed. In Arm A, 12 ITT N=53 N=54
pts (38.7%) had at least one treatment related adverse event (TRAE) with grade 3 and 1 pt Best response, n (%)
(3.2%) with grade 4. No fatal TRAEs were observed. Conclusions: Neoadjuvant / peri- CR 0 (0.0) 2 (3.7)
operative gem/cis clearly improved OS and R0 resection rate in pts with biliary tract PR 12 (22.6) 10 (18.5)
SD 28 (52.8) 34 (63.0)
cancer compared to direct surgery and was able to nearly double mOS while not in- Progressive disease (PD) 8 (15.1) 8 (14.8)
creasing the morbidity rate and even decreasing mortality rates. Clinical trial information: Missing 5 (9.5) 0 (0.0)
DC rate, n (%) 40 (75.5) 46 (85.2)
NCT03673072. Research Sponsor: None. DDC (95% Cl) 8.8 (6.2-12.9) 9.8 (6.8-14.8)

4010 Clinical Science Symposium 4011 Clinical Science Symposium

Efficacy and safety of cafelkibart (LM-108), an anti-CCR8 monoclonal an- NeoPancONE: GATA6 expression as a predictor of benefit to peri-operative
tibody, in combination with anti-PD-1 therapy in patients with pancreatic modified FOLFIRINOX in resectable pancreatic adenocarcinoma (r-PDAC): A
cancer: Results from phase 1/2 studies. First Author: Jifang Gong, Key Laboratory multicentre phase II study. First Author: Ronan Andrew McLaughlin, Division of
of Carcinogenesis and Translational Research (Ministry of Education/Beijing), De- Medical Oncology and Hematology, Princess Margaret Cancer Centre – University Health
partment of Early Drug Development Center, Peking University Cancer Hospital and Network, University of Toronto, Toronto, ON, Canada
Institute, Beijing, China Background: Modified FOLFIRINOX (mFFX) is increasingly used in the perioperative setting in r-PDAC
Background: Targeting tumor-infiltrating Tregs presents a promising strategy to overcome and patients (pts) would benefit from a biomarker approach. GATA6 expression enriches for the classical
resistance to immunotherapy in cancer treatment. LM-108 is a novel Fc-optimized anti-CCR8 RNA subtype, associated with improved OS in advanced PDAC. Low expression identifies the basal
subtype which may predict mFFX resistance. NeoPancONE is a single arm Phase II multicentre study
monoclonal antibody designed to selectively deplete tumor-infiltrating Tregs while sparing
evaluating clinical outcomes and investigating GATA6 as a biomarker of response to perioperative mFFX
peripheral Tregs. This pooled analysis of two phase 1/2 trials assesses the efficacy and safety in r-PDAC. Methods: Pts were enrolled following central radiology review (CRR) and underwent an EUS
of LM-108 in combination with anti-PD-1 therapy in patients with pancreatic cancer. FNB for GATA6 in-situ hybridization (ISH). Six cycles of mFFX were planned pre and postoperatively. The
Methods: Eligible patients (pts) with pancreatic cancer who had progressed on or after at least primary endpoint was 1 yr event-free survival (EFS) according to GATA6 ISH (high vs low). Secondary
one prior line of systemic therapy were included. Treatment regimens included LM-108 at endpoints include OS, RECIST response, SAEs, R0 resection rates and RNA subtyping by PurIST. Sta-
doses of 3 mg/kg Q3W, 3 mg/kg Q2W, 10 mg/kg Q3W, or 10 mg/kg Q2W, in combination with tistical assumptions used a ratio of 3:1 GATA6 high:low, with a 1 yr EFS of 65% for high and 34% for low
pembrolizumab (400 mg Q6W) or toripalimab (240 mg Q3W). The primary endpoint was ORR. (HR 2.5, 80% power, 2 sided alpha 0.05). KM method and log-rank test were used. Results: Between Sep-
Secondary endpoints were DCR, PFS, OS, DoR, safety, and biomarkers analysis. Data cutoff: 2020– Sep 2023, 146 pts were screened and 84 enrolled (58%) at 8 Canadian centres. CRR deemed 39
December 2, 2024. Results: A total of 80 pts (median age: 63 years; 58.8% male) from China (27%) ineligible. Clinical data are summarized (Table). GATA6 ISH was analysed in 74 (88%); 62 (84%)
and Australia were treated. Of these, 48 pts had progressed on or after 1 prior line of therapy, were high, 16% low. At a median follow up of 24.5 mos, the med EFS and OS in the ITT were 16.1 mos (95
and 32 pts had $2 lines. Eighteen pts (22.5%) had prior anti-PD-1 therapy, and 52 pts (65.0%) CI; 13-21) and 34.2 mos (95 CI; 28-NE). Med OS in the 73 pts who underwent surgery was 35.6 mos (95 CI
33-NE). The 1 yr EFS was 71% in GATA6 high vs 58% in GATA6 low p= 0.53. 1 yr OS was 87% in high vs 75%
had liver metastases at baseline. TRAEs were reported in 76 pts (95.0%). Common TRAEs
for low p= 0.29. The proportion progressing within 6 mos of enrollment in the GATA6 low group was
($25%) included increased AST, increased ALT, anemia, rash, pyrexia, decreased platelet significantly higher (42% vs 12% p=0.02). PuriST subtyping was reported in 49 (67%) resections; 14%
count and increased conjugated bilirubin. Grade $3 TRAEs occurred in 42 pts (52.5%), the basal, 86% classical. The 1 yr EFS was 79% in classical vs 43% in the basal subtype p=0.1. The 1 yr OS was
most common events ($5%) were lipase elevation (7.5%), increased ALT (6.3%), increased AST 95% in classical vs 57% in basal p=0.034. Conclusions: This is one of the first trials in r-PDAC to identify
(5.0%), immune-mediated enterocolitis (5.0%), hypokalemia (5.0%), and rash (5.0%). Median potential biomarkers to predict perioperative mFFX response. GATA6 by ISH can be assessed on baseline
follow-up was 10.48 months (95% CI 7.20-12.65). Among 74 efficacy-evaluable pts, ORR was tissue. GATA6 high is a prognostic biomarker, although NS, trends towards improved EFS and an en-
20.3% (95% CI 11.8-31.2%) and DCR was 62.2% (95% CI 50.1-73.2%). Median DoR was couraging OS. Disease progression within 6 months of enrollment occurs in nearly 50% of patients with
5.49 months (95% CI 3.02-8.87), PFS was 3.12 months (95% CI 1.61-4.86), and OS was low GATA6 expression. Neoadjuvant mFFX should not be the standard of care in these patients. Basal/
10.02 months (95% CI 6.41-13.11). Among 45 pts who had progressed on or after one prior line Classical subtyping had stronger prognostic value than GATA6 and should be considered at baseline EUS
of therapy, ORR was 24.4% (95% CI 12.9-39.5%) and DCR was 71.1% (95% CI 55.7-83.6%), FNB for future perioperative strategies in r-PDAC studies. Clinical trial information: NCT04472910.
Research Sponsor: Princess Margaret Cancer Foundation UHN; Pancreatic Cancer Canada.
with a median DoR of 6.93 months (95% CI 3.02-NA), PFS of 4.86 months (95% CI 2.79-6.90),
and OS not reached. The 12-month OS rate was 51.6% (95% CI 31.4-68.5%). Among these, 9 pts Characteristic n=84
with high CCR8 expression (7 with baseline liver metastases) showed ORR of 33.3% (95% CI Age med. (range) yrs 64 (44, 83)
7.5-70.1%) and DCR of 77.8% (95% CI 40.0-97.2%). Median PFS was 6.90 months (95% CI 1.22- Baseline EUS FNB tissue n (%) 83 (98)
NA), and OS was 9.15 months (95% CI 3.61-NA). Conclusions: LM-108 in combination with Pre-op completed 6 cycles n (%) 62 (74)
Pre-op RECIST CR/PR/SD/PD/NE % 1/18/64/11/6
anti-PD-1 therapy demonstrated encouraging antitumor activity and a manageable safety Surgery Completed Y/N n (%) / R0 / R1 n (%) 73 (87) / 11(13) / 62 (85) / 11 (15)
profile in patients with pancreatic cancer who had progressed on or after prior systemic Adjuvant Chemotherapy Y / N n (%) 63 (86) / 10 (14)
therapies. These findings support further investigation of LM-108 in combination with anti-PD- mFFX associated SAE ‡G3 n (%) 13 (15)
Pre-op mFFX related deaths 3 (4)
1 therapy as a potential treatment option for pancreatic cancer. Clinical trial information:
NCT05199753; NCT05518045. Research Sponsor: LaNova Medicines Limited.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 275s

LBA4012 Rapid Oral Abstract Session 4013 Rapid Oral Abstract Session
Disitamab vedotin (DV) plus toripalimab (Tor) and chemotherapy (C)/ Long-term outcomes and overall survival (OS) for zanidatamab + chemo-
trastuzumab (Tra) as first-line (1L) treatment of patients (pts) with therapy in HER2-positive (HER2+) advanced or metastatic gastroesopha-
HER2-expressing locally advanced or metastatic (la/m) gastric cancer. geal adenocarcinoma (mGEA): 4-year follow-up of a phase 2 trial. First Author:
First Author: Lin Shen, Beijing Cancer Hospital, Beijing, China Elena Elimova, Princess Margaret Cancer Centre, Toronto, ON, Canada
Background: Zanidatamab (zani), a dual HER2-targeted bispecific antibody, plus chemotherapy
(chemo) has previously demonstrated antitumor activity and a manageable safety profile in the first-
line (1L) treatment of patients (pts) with HER2+ mGEA. Here, we report a 4-year follow-up and the first
report of both median OS and translational data from this phase 2 trial. Methods: The phase 2 trial
(NCT03929666) evaluated zani + chemo (mFOLFOX6, CAPOX, or FP) in the 1L treatment of mGEA. In
Part 1, pts had HER2-expressing (IHC 3+ or 2+) mGEA. Pts in Part 2 had HER2+ (IHC 3+ or IHC 2+/
FISH+) mGEA by central assessment. After 25 pts were treated, antidiarrheal prophylaxis was added
for cycle 1. The primary endpoint was confirmed objective response rate (cORR). Secondary endpoints
included duration of response (DoR), progression-free survival (PFS), OS, and safety outcomes.
Plasma ctDNA samples were collected for NGS testing (Guardant360). Results: In total, 46 pts were
enrolled (zani + mFOLFOX6 [n = 24], CAPOX [n = 20], or FP [n = 2]). The majority (41 [89%]) of pts had
HER2+ mGEA by central confirmation (ccHER2+); 35 (76%) pts had gastric/GEJ cancer. As of July 28,
2024, the median (range) follow-up was 48 (29-59) mo; 8 pts (17%) were on zani treatment and 19
(41%) in survival follow-up. Efficacy results are shown in the Table. The median OS was 36.5 mo;
The full, final text of this abstract will be available at longest survival time was 57.9 mo (censored without death at data cutoff). The concordance between
[Link] on the day of presentation and in the HER2 gene amplification by centrally assessed ISH vs plasma ctDNA was 90% (18/20). Of 14 pts with
matched plasma samples at baseline and on-treatment (Cycle 2, day 15), 8 had a . 90% decrease in
online supplement to the June 10, 2025, issue of the Journal total ctDNA levels and 2 had a decrease in HER2 copy number. Common (. 5% of pts) grade 3 or 4
of Clinical Oncology. treatment-related AEs (TRAEs) were diarrhea (n = 18 [39%]), hypokalemia (n = 10 [22%]), vomiting (n =
4 [9%]), and nausea (n = 3 [7%]). Grade 3 or 4 diarrhea incidence was reduced from 52% to 24% after
prophylaxis implementation. No deaths occurred due to TRAEs. Conclusions: After a median 4-year
follow-up, zani + chemo demonstrated clinically meaningful efficacy in the 1L treatment of HER2+
mGEA, with durable responses and a median OS . 3 years, and a manageable safety profile. Zani +
chemo markedly reduced total plasma ctDNA levels early in treatment of mGEA. Clinical trial in-
formation: NCT03929666. Research Sponsor: Jazz Pharmaceuticals.
All pts ccHER2+ GEA pts
(N = 46) (n = 41)
a
cORR , n (% [95% CI]) 32 (76.2 [60.5, 87.9]) 31 (83.8 [68.0, 93.8])
Median DoRb (95% CI), mo 18.7 (10.4, 44.1) 20.4 (8.3, 44.1)
24-mo DoR, % (95% CI) 40 (22, 58) 41 (22, 59)
Median PFS (95% CI), mo 12.5 (8.2, 21.8) 15.2 (9.5, 33.4)
Median OS (95% CI), mo 36.5 (23.6, NE) 36.5 (23.6, NE)
24-mo OS, % (95% CI) 65 (49, 77) 67 (49, 79)
36-mo OS, % (95% CI) 53 (37, 67) 53 (36, 67)
TRAEs, n (%)
Any 46 (100) 41 (100)
Grade 3 or 4 30 (65) 26 (63)
a
Response evaluable (n = 42 and 37).
b
Complete or partial response. NE, non-estimable.

4014 Rapid Oral Abstract Session 4015 Rapid Oral Abstract Session
Recurrence patterns in the prospective, randomized, controlled, multicenter Decoding the response and resistant features to the Claudin18.2-specific
phase III ESOPEC trial comparing perioperative chemotherapy with preop- CAR-T cell CT041 in gastric cancer: A multi-omics exploratory biomarker
erative chemoradiotherapy in patients with esophageal adenocarcinoma. analysis of the phase 1 clinical trial. First Author: Haoxin Peng, Department of
First Author: Jens Hoeppner, University of Bielefeld, University Medical Center OWL, Gastrointestinal Oncology, Peking University Cancer Hospital & Institute, Beijing, China
Campus Hospital Lippe, Bielefeld, Germany Background: CT041, a Claudin18.2-specific CAR-T cell therapy against gastric cancer
Background: The ESOPEC trial (NCT02509286) showed that perioperative chemotherapy improved (GC), has demonstrated encouraging efficacy and a manageable safety performance.
overall and progression free survival in pts with esophageal adenocarcinoma (EAC) compared with However, the characteristics of response and resistance to CT041 therapy remain
preoperative chemoradiotherapy. Understanding of the pattern of recurrence is important for the
unclear. Methods: An exploratory biomarker analysis of the phase 1 CT041-CG4006
development of more effective future treatment strategies. Methods: Pts with cT1 cN+ cM0 or cT2-4a
cNany cM0 EAC undergoing preoperative chemotherapy with FLOT (5-FU/leucovorin/oxaliplatin/ clinical trial (NCT03874897) was conducted. Baseline samples were collected from
docetaxel) or preoperative chemoradiotherapy with CROSS (41.4Gy/carboplatin/paclitaxel) plus tu- primary tumors (PT), and dynamic samples were collected from peripheral blood and
mor resection from the ESOPEC trial were eligible. Recurrence-free survival (RFS), distant metastasis- malignant ascites at day 0, 3, and 7 of 35 patients. These samples underwent single-cell
free survival (DMFS), and patterns of local, regional and distant recurrence were analyzed. Treatment RNA sequencing (n = 41), spatial transcriptomics (ST, n = 8 for Visium V2 and n = 5 for
groups were compared with respect to sites of recurrence by calculating 3-year cumulative incidences Xenium 5K), and multiplex immunofluorescence detection (n = 28). Results: In the
considering competing events, and by Cox regression models for event-specific hazards stratified by responders, significantly higher GZMK+CD8+ T-lymphocyte infiltrates, which were in a
trial center and including treatment assignment (FLOT vs CROSS), cN stage (cN0 vs cN+), and age as
progenitor-like exhausted (Tpex) state, were identified within PT. More GZMK+ Tpex
covariates. Event-specific hazard ratios (HR) with two-sided 95% confidence intervals (CI) and p-values
are presented. Results: Of the 438 pts enrolled in ESOPEC, 192 of 221 (86.9%) in the FLOT group and infiltration also correlated with better prognosis (high vs. low: progression-free survival
179 of 217 (82.5%) in the CROSS group underwent tumor resection and represent the population for this 9.98 vs. 5.45 months, P = 0.020; overall survival 18.03 vs. 8.03 months, P = 0.038),
analysis. R0 resection was achieved in 354 pts (182 (94.8%) FLOT; 172 (96.1%) CROSS). 142 (74.0%) pts manifesting superior potential as a predictor of CT041 response (AUC 0.87, 95%CI 0.72-
in FLOT received postoperative chemotherapy. After a median follow-up of 56 months 178 pts had 1.00). Moreover, strong and positive associations between GZMK+ Tpex and IRF8+ B-
disease recurrence (81 FLOT; 97 CROSS), and 28 pts died without recurrence (12 FLOT; 16 CROSS). The lymphocytes were discovered. ST detection further uncovered that they tended to co-
3-year RFS rate was 54.5% in FLOT vs 39.0% in CROSS (HR for recurrence or death 0.67; 95% CI, 0.51 - localize within tertiary lymphoid structures and intimately interacted via the MHC-I and
0.89, p = 0.005). The 3-year DMFS rate was 57.6% in FLOT vs 41.0% in CROSS (HR for distant recurrence
or death 0.64; 95% CI, 0.48 – 0.85, p = 0.002). Conclusions: Perioperative chemotherapy with FLOT
CCL5-CXCR4 pathways. In cases resistant to CT041, significantly higher infiltrates of
improves RFS and DMFS compared to preoperative chemoradiotherapy with CROSS in EAC. The IQGAP3+ cancer cells, a subset characterized by augmented proliferative activity, were
prognosis of pts is determined by distant recurrence, which is less common after FLOT than CROSS. discovered. While IQGAP3+ cancer cells hindered the infiltration of GZMK+ Tpex, they
Clinical trial information: NCT02509286. Research Sponsor: German Research Foundation (DFG); exhibited a co-infiltrated pattern with IL6+ fibroblasts, wherein enhanced epithelial-to-
264590883. mesenchymal transition pathway activity was noted. Through dynamically parsing the
Treatment group comparisons regarding site of recurrence. blood and ascites samples, an initial activated and subsequent exhausted state of
FLOT CROSS GZMK+ cytotoxic T-lymphocyte was observed after the infusion of CT041. Additionally,
N=192 N=179 FLOT vs CROSS significant up-regulation of NR4A1, a crucial regulator of Tpex, and CCR9, a key homing
HR molecule to the gastrointestinal mucosa, of peripheral T-lymphocytes was detected in
Site of recurrence N (%)* N (%)* (for recurrence) 95% CI P-value responsive cases. Conclusions: This exploratory analysis first unveiled the biological
Locoregional 39 (20.2) 32 (17.4) 1.00 0.62 – 1.61 0.99 underpinnings and clinical implications of GZMK+ Tpex in determining response and
Local 19 (9.5) 15 (8.1) 1.03 0.51 – 2.06 0.94 resistance to the Claudin18.2-specific CAR-T therapy. Research Sponsor: Integrated
Regional 29 (15.1) 27 (14.5) 0.89 0.52 – 1.52 0.68
Distant 64 (31.5) 89 (47.2) 0.59 0.43 – 0.82 0.002 Project of Major Research Plans of the National Natural Science Foundation of China;
Distant lymphatic 21 (9.3) 31 (15.6) 0.60 0.34 – 1.05 0.074 Joint Fund for Regional Innovation and Development of the National Natural Science
Hematogenous 33 (17.2) 48 (26.1) 0.59 0.37 – 0.92 0.021
Pleural/Peritoneal 24 (12.0) 31 (16.4) 0.63 0.36 – 1.09 0.10
Foundation of China.
*% calculated as 3-year cumulative incidence.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
276s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4016 Rapid Oral Abstract Session 4017 Rapid Oral Abstract Session
Cosiporfin sodium (DVDMS)-mediated photodynamic therapy (PDT) versus Claudin18.2 (CLDN18.2) expression and efficacy in pancreatic ductal ade-
treatment of physician’s choice (TPC) in patients (pts) with advanced nocarcinoma (PDAC): Results from a phase I dose expansion cohort eval-
esophageal cancer (aEC): A phase III, randomized, open-label, multicenter uating IBI343. First Author: Xianjun Yu, Department of Pancreatic Surgery, Fudan
trial. First Author: Jun Zhou, Peking University Cancer Hospital & Institute, Beijing, China University Shanghai Cancer Center, Shanghai, China
Background: Challenges persist in the treatment of aEC, particularly for pts with esophageal Background: CLDN18.2 is highly expressed in nearly 60% of PDAC cases. Yet to date, there
stenosis and dysphagia, conditions that deteriorate the nutrition status and hinder anti-tumor are no approved targeted therapies and prognoses for these patients (pts) remain poor.
therapies, leading to dismal prognosis. Building on the promising anti-tumor activity and im- IBI343, consisting of an anti-CLDN18.2 Fc silenced monoclonal antibody and the topo-
provement in dysphagia in a phase II single-arm trial, the phase III DYNA-Esophagus03 trial isomerase I inhibitor, exatecan, is the first CLDN18.2 antibody-drug conjugate to have shown
assessed DVDMS, a novel photosensitizer, -mediated PDT in aEC, comparing to TPC. Methods: Pts encouraging efficacy in PDAC. Here, we report results from a phase 1 study (NCT05458219) in
with local recurrence or metastasis EC and Stoller dysphagia grade $2, stratified by previous pts with PDAC treated with IBI343 by CLDN18.2 expression status ($60% vs , 60%).
treatment lines (1st vs 2nd-line), were randomized 2:1 to DVDMS-mediated PDT (PDT delivered 24 Methods: Eligible pts with advanced PDAC and moderate to high expression of CLDN18.2
hours after 0.2 mg/kg DVDMS injection at a light dose of 102 J/cm² and wavelength of 63065 nm) (defined as immunohistochemistry [IHC] membrane staining intensity $2 in $40% of tumor
or TPC. The primary endpoint was esophageal stenosis overall response rate (es-ORR) assessed by
cells by IHC VENTANA CLDN18 [43-14A] Assay) who failed or were intolerant to standard
endoscopy at 28 days. A sample size of 186 pts achieved a 90% power to detect an es-ORR increase
treatment were enrolled. IBI343 was administered every 3 weeks at multiple dose levels (1, 3,
from 10% to 35% at a 1-side 2.5% level, considering a 30% drop-off rate. Results: As of November
11, 2024, 186 pts were randomly assigned (124 DVDMS-PDT group/62 TPC group). The baseline 4.5, 6, 8, and 10 mg/kg). Endpoints included safety, objective response rate (ORR), disease
characteristics were: mean age 67.6 6 8.7 years, 32.3% distant metastasis, 52.2% grade 3 control rate (DCR), duration of response (DoR), progression-free survival (PFS) per RECIST
dysphagia, and 1.1% grade 4. The es-ORR was 51.6% with DVDMS-PDT vs 8.1% with TPC at day 28 v1.1, and overall survival (OS). Results: As of December 27, 2024, 83 pts from China and
(P , 0.0001). At a median follow-up of 8.9 months (mos), the median progression-free survival was Australia were enrolled. 44 and 12 pts with CLDN18.2 expression $1 in $60% (+) and , 60%
2.8 vs 2.2 mos (HR = 0.61, 95%CI 0.40-0.93; P = 0.0244), respectively. Time to progression was 5.9 (-) of tumor cells, respectively, received IBI343 at 6mg/kg in the dose expansion phase
vs 3.9 mos (HR = 0.45, 95% CI: 0.25-0.82, P = 0.0056). Median overall survival (OS) was 7.0 vs 6.4 (median age, 60 and 64 years; male, 54.5% and 66.7%; received prior treatments $2L, 61.4%
mos (HR = 0.88, 95%CI 0.59-1.32; P = 0.5169), respectively. Considering the high cross-over rate and 83.3%). Among all treated pts (n = 83), treatment-emergent adverse events (TEAEs)
(45.2%), the rank preserving structural failure time-adjusted median OS was 7.0 vs 4.7 mos (HR = occurred in 82 (98.8%) pts and $Grade 3 (G3) TEAEs in 42 (50.6%) pts. TEAEs led to
0.64, 95%CI 0.42-0.97; P = 0.0223). The improvement of dysphagia and quality of life are detailed in treatment discontinuation in 6 (7.2%) pts. No TEAE led to death. Most common TEAEs were
the Table. Phototoxicity was negative in 64.9% pts in DVDMS-PDT group on day 7, increasing to anemia (66.3%, 15.7% $G3), neutrophil count decreased (48.2%, 9.6% $G3), and WBC count
91.4% by day 28. Grade $3 treatment-emergent adverse events occurred in 38.2% of the DVDMS- decreased (47.0%, 9.6% $G3). The safety profile of IBI343 in PDAC was comparable to
PDT group and 49.2% of the TPC group. Six deaths were due to treatment-related adverse events previous reports. Among CLDN18.2+ pts (n = 44), confirmed ORR was 22.7% (95% CI,
(4.9%) were reported in the DVDMS-PDT group and 5 (8.5%) in the TPC group. 11.5–37.8); DCR was 81.8% (95% CI: 67.3–91.8), median PFS was 5.4 (95% CI, 4.1–7.4)
Conclusions: DVDMS-PDT significantly improved esophageal stenosis and dysphagia compared months (mos), median OS was 8.5 (95% CI, 6.6–12.1) mos, and in 10 pts with confirmed
to the treatment of physician’s choice in pts with aEC, with prolonged PFS and TTP, potential better partial response median DOR was 6.7 (95% CI, 3.2–7.7) mos. Among CLDN18.2+ pts who
OS and a manageable safety profile. Clinical trial information: CTR20221271. Research Sponsor: received 1 and 2 lines of prior treatment, median OS was 12.1 (95% CI, 6.6–not calculable
None. [NC]) mos and 9.1 (95% CI, 5.1–NC) mos, respectively. Among CLDN18.2- pts (n = 12), no pt
DVDMS-PDT (n=124) TPC (n=62) had an objective response, DCR was 41.7% (95% CI: 15.2–72.3), median PFS was 1.4
(1.3–3.2) mos, which was shorter than what was seen in CLDN18.2+ pts (Ad hoc analysis:
Reduction at least 1 grade in Stooler’s 47.8% 8.6% (P,0.0001)
score in 6 mons, % nominal P , .0001; HR = 0.198 [95% CI: 0.089-0.439]) and median OS was 6.2 (95%
Change in EORTC QLQ-C30 total score -26.7 6 2.5 -39.8 6 4.3 CI,1.4–9.0) mos. Conclusions: IBI343 was well tolerated and continues to show encouraging
Change in dysphagia score 16.0 6 2.6 24.6 6 4.3 efficacy in pts with PDAC. Efficacy benefit was most pronounced in those with CLDN18.2
Change in eating score 16.9 6 2.4 30.6 6 4.1 expression $60%, suggesting that CLDN18.2 can be a predictive biomarker for response to
Values are least-squares mean 6 standard error unless otherwise noted. IBI343 in PDAC. The trial is enrolling in Australia, China, and US. Clinical trial information:
Symptom scores were assessed using EORTC QLQ-OES18. NCT05458219. Research Sponsor: None.

4018 Rapid Oral Abstract Session 4019 Rapid Oral Abstract Session
Clinical activity of EBC-129, a first-in class, anti N256-glycosylated CEA- 10-year follow up of a phase 2 clinical trial of pembrolizumab (pembro) in
CAM5 and CEACAM6 antibody-drug conjugate (ADC), in patients with pan- microsatellite instability-high (MSI-H)/mismatch repair deficient (dMMR)
creatic ductal adenocarcinoma (PDAC) in a phase 1 study. First Author: advanced solid tumors. First Author: Katherine M. Bever, Sidney Kimmel Com-
Robert William Lentz, Division of Medical Oncology, Department of Medicine, University prehensive Cancer Center at Johns Hopkins University, Baltimore, MD
of Colorado Anschutz School of Medicine, Aurora, CO Background: Inhibitors of Programmed Cell Death Protein-1 (PD-1) have demonstrated remarkable activity
Background: Pancreatic ductal adenocarcinoma (PDAC) is an aggressive malignancy with in dMMR/MSI-H cancers, leading to the first tissue agnostic FDA approval for an oncologic indication. We
report herein results of long-term follow up of KEYNOTE-016, the first study to demonstrate pan-tumor
limited treatment options. EBC-129 is a first in class MMAE-linked ADC against N256-
activity of the PD-1 inhibitor pembro in dMMR/MSI-H solid tumors. Methods: KEYNOTE-016 was a multi-
glycosylated CEACAM5/6 antigen. Pre-screening data shows 73% of PDAC express surface center open-label phase 2 study evaluating pembro in patients with advanced colorectal cancer (CRC)
antigen (IHC positive cut-off $20% at 2+ or 3+) or 82% ($1% at 3+). Previously reported (Cohort A) or non-CRC solid tumors (Cohort C) that were dMMR and had progressed after $1 prior line of
dose escalation data identified 1.8 and 2.2 mg/kg every 3 weeks as the RP2Ds. Here we therapy (or $2 prior lines for CRC). Eligible patients were age $ 18, and had measurable disease per RECIST
report pooled safety and efficacy data of the PDAC patients enrolled in the dose escalation 1.1. Patients with active CNS metastases, who were on immunosuppressive therapy, had autoimmune
(DEs) and expansion (DEx) cohorts of the phase 1 study. Methods: This study consisted of disease, or were previously treated with immune checkpoint inhibitors were excluded. Patients received
DEs and DEx cohorts. Previously treated histologically confirmed, locally advanced or pembro IV every 2 weeks until progression, intolerance, withdrawal of consent or up to a maximum of 2
years. Results: Between 9/2013 - 9/2017, 88 patients (Cohort A: 41; Cohort C: 47) enrolled at 7 sites and
metastatic PDAC patients with antigen expression levels of $20% at 2+ or 3+ intensity received $1 dose of pembro. Tumor types enrolled on Cohort C included endometrial (N = 15), pancreatic (N
or $1% at 3+ intensity (IHC on archival samples) were enrolled. Objectives were to = 9), small intestinal (N = 5), gastroesophageal (N = 5), biliary (N = 4), ampullary (N = 4), and other (N = 5).
evaluate safety, efficacy, and PK of EBC-129, administered every 3 weeks. Results: A total Median follow up time was 49.7 mos for all patients and 99.8 mos for alive patients. Objective response rate
of 21 PDAC patients were enrolled (6 in DEs and 15 in DEx). Patients received 1.8 (n = 8), 2.0 (ORR) was 58% with 23 partial (PR) and 28 complete responses (CR). 16 patients experienced a best
(n = 2) or 2.2 (n = 11) mg/kg EBC-129 every 3 weeks. 52% were male; mean age 63 years; response of stable disease (SD) for a disease control rate of 76%. Median PFS and OS were 34.9 mos (95%
median 3 (range 1-7) lines of previous treatments with 81% prior taxane treated. At data CI: 14.8-NR) and 80.8 mos (95% CI: 33.2-NR) respectively. The 3-, 5-, and 10-year OS rates were 55.1%,
53.7% and 47.4% respectively. Outcomes were similar between Cohorts A and C (see Table).
cut-off (Jan 16th, 2025), 5 patients are continuing treatment, 12 patients had radiological
Conclusions: In summary, long term follow up of KEYNOTE-016 confirms high rates of durable remission
progression, 3 had clinical progression, and 1 patient withdrew treatment. The overall from pembro in patients with dMMR/MSI-H solid tumors, with several patients remaining alive and in
response rate, disease control rate (first assessment), and median PFS, at 1.8 mg/kg, 2.2 remission at 10+ years follow up. Responses were seen across tumor types. Clinical trial information:
mg/kg and overall group, respectively, were 25.0%, 18.2% and 19.0% (unconfirmed), 87.5%, NCT01876511. Research Sponsor: Swim Across America; Merck.
63.6% and 71.4%, and 18, 12 and 12 weeks. 43% of patients had any tumour shrinkage Results by cohort.
overall. Infusion related reactions (IRR) were seen in 57% of patients; most were grade 1/2, Cohort A Cohort C
more frequent at 2.2 mg/kg, and resolved or reduced with premedication. Grade $3 CRC non-CRC
treatment related adverse events included neutropenia (50.0% and 81.8%) and anaemia N=41 N=47
(12.5% and 18.2%) at 1.8 and 2.2mg/kg, respectively; amylase/lipase increase, vomiting, ORR, % 56.1 59.6
and aspartate aminotransferase increase occurred in 1 patient each. Grade #2 peripheral PR, N (%) 11 (27) 12 (25)
CR, N (%) 12* (29) 16 (34)
neuropathy was seen in 2 patients. No drug related discontinuations occurred in this SD, N (%) 10 (24) 6 (13)
cohort. In this cohort, selected for target antigen expression, no apparent correlation was PD, N (%) 5 (12) 9 (19)
NE, N (%) 3 (7) 4 (9)
seen between IHC score and treatment response. Based on a cut-off of . 25% CEA PFS, median months (95% CI) 38.8 (8.1-NR) 20.5 (14.3-NR)
decrease (either local or central), biomarker response was seen 42.9% at 1.8 mg/kg and OS, median months (95% CI) 80.8 (33.2-NR) 86.4 (21.8-NR)
36.4% at 2.2 mg/kg. Conclusions: EBC-129 shows promising clinical activity in heavily Follow up time, median months 51.2 35.9
3-year OS rate (%) 60.5 50.3
treated PDAC patients with a manageable tolerability profile, consistent with MMAE-based 5-year OS rate (%) 57.5 50.3
ADCs. Further evaluation of EBC-129 in PDAC, both as monotherapy and in combination 10-year OS rate (%) 47.3 47.2
with chemotherapy, is planned. Clinical trial information: NCT05701527. Research *Includes 3 patients with unconfirmed CR.
Sponsor: Experimental Drug Development Centre (EDDC).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 277s

4020 Rapid Oral Abstract Session 4021 Poster Session

Health-related quality of life (HRQOL) in the phase 3 trial of cabozantinib vs Claudin-18.2 expression in gastro-oesophageal adenocarcinoma in a West-
placebo for advanced neuroendocrine tumors (NET) after progression on ern population: Overlap with other biomarkers and prognostic value. First
prior therapy (CABINET, Alliance A021602). First Author: Amylou C. Dueck, Author: Filip Van Herpe, Gastrointestinal Oncology Department, University Hospitals
Alliance Statistics and Data Management Center, Mayo Clinic, Scottsdale, AZ Leuven, Leuven, Belgium
Background: In previously treated patients (pts) with advanced extra-pancreatic NET Background: Claudin 18.2 (CLDN18.2) is a novel biomarker for response to anti-CLDN18.2
(epNET) or pancreatic NET (pNET), cabozantinib (cabo) demonstrated improved progression- therapy in gastroesophageal adenocarcinoma (GEA). Based on prevalence data, originating
free survival compared to placebo (pbo) in the CABINET trial (NCT03375320). Here we present mostly from an Asian population, 40% of GEA are CLDN18.2 positive. We currently lack
HRQOL data from CABINET. Methods: Pts with previously treated unresectable, locally information on the cross-prevalence of CLDN18.2, HER2-status, microsatellite instability
advanced or metastatic epNET or pNET were randomized (2:1) in separate cohorts to receive (MSI) and PDL1 expression in a Western population of localized and metastatic GEA.
cabo 60 mg daily vs pbo. HRQOL was an exploratory endpoint measured using the EORTC Methods: We present a single-center, retrospective study including localized and met-
QLQ-C30, QLQ-GI.NET21, Patient Global Impression of Change (PGIC), and PRO-CTCAE astatic GEA patients diagnosed between 2019-2023. Histopathology with MMR status,
questionnaires. Optional participation was by paper surveys at baseline and every HER2 status (IHC/SISH) and PDL1 CPS was obtained. CLDN18.2-positivity defined
12 weeks (wks) until disease progression or new anticancer therapy. Mean changes from as $75% of tumor cells showing moderate-to-strong membranous staining was deter-
baseline were compared between treatment arms at wk12 using general linear mixed models. mined by IHC using the VENTANA CLDN18 [43-14A] RxDx Assay and underwent blind
Minimally important within-arm improvements and between-arm differences were defined as review by independent expert pathologists. Baseline characteristics and clinical follow-up
8 points on the C30 Global Health Status (GHS)/QOL Scale. PGIC was dichotomized as data were gathered until august 2024. Kaplan-Meier curves were constructed for overall
improved vs unchanged/worsened and compared using Chi-squared tests. Rates of pt- survival (OS), progression/relapse free survival (PFS/RFS) and survival after recurrence
reported adverse events (AEs) by PRO-CTCAE were baseline-adjusted and compared using (SAR). Unrestricted grants/support was provided from Astellas, AMGEN and Roche di-
Fisher’s exact tests. Results: 172 pts (113 cabo, 59 pbo) with epNET and 81 pts (53 cabo, 28 agnostics. Results: Of 405 patients with GEA, 261 presented with localized and 144 with
pbo) with pNET consented for survey participation. Pts completed 524 (82%) of 640 expected metastatic disease. The majority were male (71%) with a median age of 65 years. 59% of
surveys across baseline, wk12, and wk24. For epNET, C30 GHS/QOL was significantly im- tumors were junction tumors and 41% were gastric cancers. dMMR was detected in 6%.
proved at wk12 in pts receiving cabo (mean change 9.5, standard error (SE) 2.2, p , 0.001)
HER2 positivity (cfr. TOGA criteria) was seen in 16% of cases. PDL1 CPS $1 prevalence
but not pbo (mean change 0.2, SE 3.1, p = 0.95) resulting in significantly different mean
was 58%. CLDN18.2 positivity was seen in 42% of patients. Double CLDN18.2-HER2 and
changes from baseline between arms (p = 0.02). For pNET, C30 QHS/QOL was significantly
CLDN18.2-dMMR positivity was rare (5% and 3%), but strikingly in absolute numbers, about
improved at wk12 in pts receiving cabo (mean change 8.0, SE 3.0, p = 0.008) but not pbo
one third of HER2 positive tumors and half of dMMR tumors were CLDN18.2 positive.
(mean change 6.8, SE 4.4, p = 0.12); mean changes from baseline did not statistically differ
Double CLDN18.2-PDL1 $1//$5//$10 positivity was seen in 23%,15% and 9%. In patients
between arms (p = 0.82). No significant differences in mean changes from baseline at wk12 in
C30 Physical, Role, Emotional, Cognitive and Social Function were observed between arms in with localized disease CLDN18.2 positivity did not affect RFS (HR: 0.99; 95%CI(0.68-1.43);
the epNET or pNET cohorts. In both cohorts at wk12, pts on cabo reported improved overall p = 0.95) and OS (HR: 0,78; 95%CI(0.52-1.16) p = 0,22) with a median follow-up (FU) of
status as measured by PGIC more often than those on pbo (epNET: 38% vs 19%, p = 0.04; 42 months. In the remaining 113 patients with primary metastatic disease no difference in
pNET: 57% vs 18%, p = 0.006). Rates of pt-reported AEs were statistically significantly higher terms of PFS (HR: 1.17; 95%CI(0,79-1.73); p = 0,43) or OS ( HR: 1.43; 95%CI(0,94-2.16); p =
in cabo for diarrhea (68% vs 53%), hand-foot syndrome (57% vs 26%), mouth or throat sores 0,1) was seen according to CLDN18.2 status with a median FU of 39 months. Median SAR
(47% vs 27%), and dysgeusia (67% vs 37%) [all p , 0.05]. GI.NET21 data will be presented. in the localized group was statistically longer in the PDL1 CPS$1 subgroup(HR: 0.51; 95%
Conclusions: Although cabo is associated with pt-reported AEs consistent with its known CI(0,34-0,79); p = 0,002). SAR was not different in CLDN18.2+ and CLDN18.2- patients (HR:
safety profile, global HRQOL was maintained. A higher proportion of patients receiving cabo 1,07; 95%CI(0,69-1.65) p = 0,77). Conclusions: We confirmed a 42% positivity rate of
reported improved overall status. Results support cabo as a treatment option for pts with CLDN18.2 in a combination of localized and metastatic GEA in a Western population.
previously treated advanced NET that preserves patient HRQOL. Clinical trial information: dMMR was present in 6% of cases, PDL1 CPS $1 and HER2+ was seen in 58% and 16% of
NCT03375320. Research Sponsor: UG1CA189823; U10CA180821; U10CA180882; Exelixis; patients. CLDN18.2 positivity was strikingly present in half dMMR and one third of HER2+
[Link] cases. CLDN18.2 positivity did not affect survival in localized and metastatic disease.
Research Sponsor: Astellas; Amgen.

4022 Poster Session 4024 Poster Session

Preliminary results of a phase 1 study of LB1908, an autologous Claudin Genomic landscape and biomarker analyses utilizing circulating-tumor DNA
18.2–targeted chimeric antigen receptor T-cell product, in patients with in advanced esophageal squamous cell carcinoma: Sub-analysis of SCRUM-
advanced gastroesophageal adenocarcinoma. First Author: David Bing Zhen, MONSTAR GOZILA. First Author: Yuqing Duan, National Cancer Center Hospital East,
Fred Hutchinson Cancer Center, Seattle, WA Kashiwa, Japan
Background: Claudin 18.2 (CLDN18.2) is a potential therapeutic target expressed on most Background: Advanced esophageal squamous cell carcinoma (ESCC) is a cancer type with a
gastric, gastroesophageal, and esophageal adenocarcinomas (GC/GEJC/EC), with normal poor prognosis, with limited survival benefits from current multimodal approaches. The
tissue expression limited to gastric epithelium. We present preliminary results of the dose- incomplete understanding of molecular mechanisms in advanced ESCC has hindered the
escalation trial of LB1908, a CLDN18.2-targeted autologous CAR-T cell product, in adults development of effective targeted therapies, emphasizing the critical need for identifying
with advanced GC/GEJC/EC. Methods: This trial is an ongoing, open-label, multicenter, predictive biomarkers and novel therapeutic targets. Methods: SCRUM-MONSTAR GOZILA
first-in-human phase 1 study of LB1908 in patients with advanced GC/GEJC/EC relapsed/ is a nationwide plasma-based genomic profiling study utilizing Guardant360 in Japan, which
refractory to $1 prior line of therapy (LOT), and with $1+ CLDN18.2 expression in $50% aimed to analyze circulating tumor DNA (ctDNA) genomic alterations in patients with
of tumor cells by central testing (43-14A antibody). A modified 3+3 design is used for dose advanced solid tumors, including ESCC. We evaluated the genomic landscape with advanced
escalation with planned dose levels of 0.5, 1.5, and 33106 CAR+ T cells/kg. The primary ESCC patients and investigated associations between genomic alterations and overall
objective is evaluation of safety and dose-limiting toxicities (DLTs) with secondary ob- survival (OS) using the log-rank test. The correlation between progression-free survival (PFS)
jectives of antitumor activity and pharmacokinetics. Results: As of the data cutoff of and blood tumor mutation burden (bTMB) in immune checkpoint inhibitor (ICI) monotherapy
January 4, 2025, 6 patients were dosed with LB1908 at dose level 1 (0.53106 cells/kg). was also assessed using multiple cut-off values (2, 4, 6, 8, and 10 mutation/Mb).
Four patients had EC and 2 had GC, all with metastatic disease. Patients had a median of 2 Results: The present study included 313 patients with available genomic and clinical
prior LOTs; all had received fluoropyrimidine/platinum agents. Bridging therapy was data. The gene alteration spectrum comprised mutations (single nucleotide variants, 71.6%;
administered to all patients between apheresis and LB1908 infusion, and patients received and insertions/deletions, 10.7%), copy number alterations (CNAs, 17.3%), and fusions
standard lymphodepletion (LD) with cyclophosphamide/fludarabine. All patients experi- (0.48%). TP53 was the most frequently altered gene (88.5%), followed by PIK3CA (36.4%),
enced treatment-emergent adverse events (TEAEs); the most common grade $3 TEAEs NFE2L2 (24.3%), CCND1 (22.4%), EGFR (20.1%), ATM (16.3%), FGFR1 (10.2%), BRCA2
were hematologic and attributable to the LD regimen. Of grade $3 TEAEs related to (10.2%), MET (9.6%) and ARID1A (9.6%). Regarding the survival outcomes, PIK3CA CNA was
LB1908, only gastritis/gastric mucosal injury occurred in more than 1 patient (n = 3), significantly associated with worse OS compared to those with PIK3CA wild type [hazard
including 1 DLT. After implementation of toxicity management guidelines using pro- ratio (HR), 1.84; 95% confidence interval (CI), 1.24–2.74; p-value, 0.0002], and PIK3CA
phylactic enteral beclomethasone and early systemic steroids, no patients experienced mutation showed a trend toward shorter OS (HR, 1.43; 95%CI, 0.94–2.17; p-value, 0.06).
prolonged grade $3 upper GI AEs. CRS occurred in all patients, with no grade $3 events. Patients with both PIK3CA mutation and CNA exhibited significantly worse OS compared to
No ICANS was observed. CAR-T cell expansion was seen in all patients, with a median Cmax those with PIK3CA wild type (HR, 1.94; 95%CI, 0.85–4.45; p-value, 0.03). Both FGFR1 CNA
and mutation were associated with poorer OS (HR, 1.98; 95%CI, 1.03–3.79; p-value, 0.005;
of 1594 copies/mg genomic DNA (range, 264-6922) and Tmax of 14 days (range, 11-15). All
and HR, 2.84; 95%CI, 0.89–9.07; p-value, 0.002, respectively). CNA in CCND1 and EGFR, and
6 patients were evaluable for response, with 5 (83%) experiencing target lesion shrinkage
mutation in NFE2L2 and RB1 also significantly correlated with worse OS (any p-value#0.01).
(maximal 1%, 9%, 25%, 31%, and 41% reduction from baseline). Responses deepened over
Among 142 patients treated with ICI monotherapy, no statistically significant differences in
time in the 2 patients with multiple postinfusion scans (-25% and -41%, respectively),
PFS were observed at any cut-off value of bTMB (any p-value . 0.1). Conclusions: This
including 1 patient achieving a RECIST partial response 7 months after treatment.
comprehensive analysis of ctDNA profiles revealed distinct genomic alterations with
Conclusions: LB1908 demonstrates peripheral expansion and encouraging antitumor
prognostic significance in advanced ESCC. Multiple alterations demonstrated significant
activity at the lowest dose tested, with a manageable safety profile. Implementation of a
associations with poor OS, meanwhile bTMB was not validated as an effective predictive
toxicity mitigation strategy ameliorated on-target gastric mucosal injury without com- biomarker for ICI efficacy. These findings provide insights into potential therapeutic targets
promising expansion kinetics or antitumor activity. Longer follow-up and data from pa- and prognostic biomarkers in advanced ESCC. Clinical trial information: 2021-GB-009.
tients treated with higher cell doses will be presented at the meeting. Clinical trial Research Sponsor: None.
information: NCT05539430. Research Sponsor: Legend Biotech USA Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
278s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4025 Poster Session 4027 Poster Session

HER2 expression dynamics and prognostic significance in the treatment of Tislelizumab (TIS; BGB-A317) plus chemotherapy (CT)/chemoradiotherapy
gastric cancer. First Author: Min Lai, Department of General Surgery, Nanfang (CRT) as positron emission tomography (PET)–guided neoadjuvant (n)
Hospital, Southern Medical University, Guangzhou, Guangdong, China treatment (tx) for resectable esophageal squamous cell carcinoma (R-
Background: The human epidermal growth factor receptor 2 (HER2) expression un- ESCC): RATIONALE-213 final analysis. First Author: Longqi Chen, Department
dergoes changes during the treatment of gastric cancer. This study aims to evaluate post- of Thoracic Surgery, West China Hospital, Sichuan University, Chengdu, China
treatment HER2 expression changes and the impact on survival. Methods: From a Background: Studies have shown overall survival improvements with nCRT + surgery vs surgery alone in
prospectively maintained database, we extracted clinical and pathological data, treatment locally advanced ESCC. However, preoperative CRT may have additional safety concerns, leading to some
information, and survival outcomes of gastric cancer patients (pts) with paired pre- and patients (pts) receiving nCT rather than nCRT. PET-computed tomography maximum standardized uptake
value (SUVmax) change after induction CT (IC) has been shown to have reliable predictive value of
post-treatment HER2 immunohistochemistry (IHC) results (2018-2024). Cohen’s Kappa pathological complete response (pCR) for R-ESCC in pts with nCRT and may optimize neoadjuvant tx
was used to assess HER2 expression concordance. Logistic regression was performed to selection. TIS (anti-PD-1) has improved survival in pts with ESCC. We report the final analysis of
identify factors associated with HER2 change, while the Kaplan-Meier method and Cox RATIONALE-213, a phase 2, open-label, multicenter study in China evaluating PET-guided nTIS + CT/CRT in
regression were used for survival analysis. Results: 274 gastric cancer pts with paired R-ESCC (NCT04974047). Methods: Eligible adult pts had histologically confirmed R-ESCC (cT1-2N + M0 or
HER2 IHC results were enrolled, including IHC 0 (67.5%, 185/274), 1+ (14.6%, 40/274), 2+ cT3N any M0), ECOG performance status 0/1, adequate organ function, no fistula risk, and had received no
(13.5%, 37/274), 3+ (4.4%, 12/274) before treatment and IHC 0 (63.8%, 175/274), 1+ prior tx. Pts had a baseline (BL) PET scan, 1 cycle of IC (cisplatin-paclitaxel [Cis-Pac]), and a PET scan 15-
21 days later. Pts were grouped into 2 cohorts by response to IC based on the percentage decrease in 2nd
(21.9%, 60/274), 2+ (11.7%, 32/274), and 3+ (2.6%, 7/274) after treatment. The overall
PET SUVmax in the primary tumor: responders (R, $35%) or nonresponders (NR, , 35%). Both cohorts
HER2 change rate was 42.7% (21.5% HER2 expression increased, 21.2% decreased), received 3 cycles of TIS 200 mg IV Q3W, the first 2 with CT (2 cycles Cis-Pac) for R, or with CRT (2 cycles
indicating low concordance (Kappa = 0.179, p , 0.001). Among 7.6% (17/225) of pts with investigator [Inv]-chosen CT [Cis-Pac, or 5-FU + Cis] + RT [40 Gy/20 fractions]) for NR, then surgery. Primary
HER2 changes from 0/1+ to 2+/3+, 23.5% (4/17) received anti-HER2 therapy (trastuzumab endpoint was pCR per local pathologist. Secondary endpoints were 1-year disease-free survival (DFS), 1-
or anti-HER2 ADCs), achieving a 75.0% response rate. 38.5% (5/13) of initially confirmed year event-free survival (EFS), objective response rate (ORR) before surgery, R0 resection rate by Inv, and
HER2-positive pts (IHC 3+/2+ and FISH+) exhibited loss of HER2 positivity after treatment. safety. Results: Of 70 pts enrolled, 15 (21.4%), 48 (68.6%), and 7 (10.0%) had stage II, III, and IVA disease at
3 of 8 initially HER2-negative pts who converted to HER2-positive received trastuzumab for BL, respectively. As of 25 Oct 2024 (median follow-up 25.5 mo), 30 pts were R and 40 NR. Of R, 20 (66.7%)
had surgery. Of NR, 32 (80.0%) had surgery. Efficacy endpoints are shown in the table. Median DFS and EFS
metastatic diseases, achieving a 66.7% response rate. Based on efficacy evaluation on the were not reached for R and NR. Grade $3 treatment-related adverse events (TRAEs) in R (15 [50.0%]) and
second HER2 testing, 197 pts were classified into the PR/SD group and 77 into the PD NR (33 [82.5%]) were consistent with known CT or CRT toxicity; serious TRAEs occurred in 5 R (16.7%) and 7
group, with comparable HER2 changes rate (41.6% vs. 45.5%, p = 0.565). Anti-HER2 NR (17.5%). No TRAEs led to surgery cancellation or death. Conclusions: APET-guided approach may help
therapy (n = 28) was associated with a higher HER2 changes rate (67.9% vs. 39.8%, p = optimize neoadjuvant tx of R-ESCC. nTIS + CT/CRT showed promising efficacy and a tolerable safety profile
0.005), mainly driven by HER2 reduction (60.7% vs. 16.7%, p , 0.001). Multivariate logistic in both responders and nonresponders. Clinical trial information: NCT04974047. Research Sponsor: BeOne
regression showed that combined immunotherapy (OR [odds ratio] 1.85, p = 0.028) or Medicines Ltd.
targeted immunotherapy (OR 4.71, p , 0.001) was associated with a higher HER2 change R NR
rate than chemotherapy alone. The median follow-up time was 31.2 months. HER2 ex- pCR, a n (%) 6 (30.0) 11 (34.4)
pression changes were associated with worse PFS (HR [hazard ratio] 1.52, p = 0.040) and (95% CI) (11.9, 54.3) (18.6, 53.2)
OS (HR 1.51, p = 0.043), with decreased HER2 expression showing the poorest PFS (Log- 1-year DFS,b % 79.0 74.2
(95% CI) (47.9, 92.7) (53.3, 86.8)
rank p = 0.030) and OS (Log-rank p = 0.025). Subgroup analysis showed in PR/SD group, 1-year EFS,c % 87.1 67.8
HER2 expression changes was associated with significantly worse PFS (HR 2.48, p = (95% CI) (64.3, 95.8) (48.3, 81.2)
0.003) and OS (HR 2.56, p = 0.002), while no survival differences were observed in PD R0 resection,a n (%) 19 (95.0) 29 (90.6)
ORR,d n (%) 15 (71.4) 14 (42.4)
group. Conclusions: HER2 expression frequently changes during the treatment of gastric
a
cancer, particularly after immunotherapy and targeted therapy, and is associated with Efficacy analysis set (EAS) R=20; NR=32.
b
EAS with R0 resection R=19; NR=29.
worse survival outcomes. Dynamic HER2 testing contributes to guiding precision therapy c
Safety analysis set (SAS) R=30; NR=40.
for gastric cancer. Research Sponsor: None. d
SAS with measurable disease at BL R=21; NR=33.

4028 Poster Session 4029 Poster Session

Efficacy and safety of surufatinib (Sur) plus paclitaxel (Pac) as second line A multicenter randomized phase II study of trastuzumab biosimilar (CT-P6)
(2L) treatment for advanced gastric cancer (aGC): Final results from a phase and chemotherapy (SOX or CapeOX) in HER2-positive advanced/recurrent
2 trial. First Author: Xiuying Xiao, Department of Oncology, Renji Hospital, School of gastric cancer (KSCC: TROX study). First Author: Teppei Yamada, Department of
Medicine, Shanghai Jiao Tong University, Shanghai, China Gastroenterological Surgery, Fukuoka University Faculty of Medicine, Fukuoka, Japan
Background: Standard chemotherapy (ChT) provides unsatisfactory efficacy for patients Background: The efficacy and safety of the trastuzumab biosimilar (CT-P6) combined
with aGC in the 2L setting. Recent studies have suggested improved efficacy with an with S-1/oxaliplatin (SOX) or capecitabine/oxaliplatin (CapeOX) were investigated in
antiangiogenic agent plus Pac. Although ramucirumab plus Pac have been approved as a patients with HER2-positive advanced or recurrent gastric cancer who had not previously
2L treatment, access to this combination regimen is limited in China. Sur is a potent received chemotherapy. CT-P6 is a biosimilar of trastuzumab, which has demonstrated
VEGFRs, FGFR1 and CSF-1R inhibitor. This trial is aimed to assess the efficacy and safety comparable safety and efficacy to trastuzumab in breast cancer. However, no clinical
of Sur plus ChT as a 2L treatment for aGC. Our previous report (ASCO-GI 24) revealed trials have evaluated CT-P6 in gastric cancer to date. Methods: This randomized, open-
promising anti-tumor activity of this regimen. Now we are presenting the final analyses label, multicenter phase II study enrolled patients with HER2-positive unresectable or
including long-term survival results. Methods: This single-arm, phase 2 clinical trial recurrent gastric cancer. Participants were randomized to receive SOX plus trastuzumab
enrolled patients with HER2-negative aGC who had failed standard first-line treatment. biosimilar therapy (S-1 orally, 40–60 mg twice daily for 14 days every 3 weeks, oxaliplatin
Eligible patients received Sur (250mg, po, qd) plus Pac (150mg/m2, iv, d1) at 21-day cycles intravenously at 130 mg/m² on day 1 every 3 weeks, and CT-P6 at 8 mg/kg on day 1 and 6
for 6 cycles, followed by maintenance with single-agent Sur until disease progression or mg/kg every 3 weeks thereafter) or CapeOX plus trastuzumab biosimilar therapy
intolerable toxicity. Tumor responses were assessed every 6 weeks by investigators per (replacing S-1 with capecitabine [1,000 mg/m²] as described above) until disease pro-
RECIST v1.1. The primary endpoint was ORR, and secondary endpoints were DCR, PFS, OS, gression. The primary endpoint was overall response rate (ORR) in both arms, with a null
and safety. Results: As of Dec 20, 2024, of the 35 patients enrolled, median age was 65 hypothesis response rate set at 43%, the lower bound of the 90% confidence interval
(range: 33-75), 26 (74.3%) were male, 31 (88.6%) had an ECOG PS of 1, 12 (34.3%) had reported in previous trastuzumab trials. Secondary endpoints included progression-free
positive PD-L1 (defined as CPS $1), and 26 (74.3%) had received 1L immunotherapy (IO). survival (PFS), overall survival (OS), and safety. Results: Between May 2019 and No-
Seven (20.0%) patients had primary GEJC, and all patients had stage IV disease, with vember 2022, 67 patients were enrolled in the study. Patient characteristics were as
lymph nodes (23, 65.7%), liver (16, 45.7%) and peritoneum (14, 40.0%) being the most follows: male/female ratio of 49/18, median age 68 years (range: 34–80), and perfor-
common metastatic sites. Tumor response assessments were available in 32 patients, the mance status (PS) of 0/1 in 53/14 patients. The ORR was 76.1% (90% CI: 67.6–84.7%),
ORR was 25.0% and DCR was 87.5%. Primary GEJC (42.9% vs 20.0%, P= 0.327) and with the lower limit of the confidence interval exceeding the threshold response rate of
baseline liver metastases (40.0% vs 11.8%, P= 0.106) seemed correlated to a better ORR, 43%, confirming the efficacy of trastuzumab biosimilar therapy. The one-year survival rate
while baseline peritoneal metastases (8.3% vs 35.0%, P= 0.204) the reverse. With a was 65.2% (95% CI: 49.8–78.6%), meeting the primary endpoint. The median OS was
median follow-up of 12.6 (95% CI: 10.4-14.9) months (mo), the mPFS was 5.7 (95% CI: 4.7- 18.7 months (90% CI: 15.1–23.1), and the median PFS was 9.0 months (90% CI: 6.5–10.7).
6.93) mo and the mOS was 10.8 (95% CI: 7.0-17.2) mo. Within the subgroup patients who Although the study was not designed to compare SOX and CapeOX, the response rate was
had IO exposure in the 1L setting (n = 26), the ORR was 25.0%, the DCR was 91.7%, the slightly higher in the CapeOX group (81.3% [90% CI: 69.9–92.6%]) compared to the SOX
mPFS was 5.9 (95% CI: 4.7-10.5) mo, and the mOS reached 14.4 (95% CI: 8.5-NR) mo. group (71.4% [90% CI: 58.9–84.0%]). Similarly, OS was slightly better in the CapeOX group
Treatment-related adverse events of grade $3 included neutropenia (40.0%), leukopenia (22.3 months [90% CI: 15.1–42.1]) than in the SOX group (16.7 months [90% CI:
(34.3%), hypertension (11.4%), and proteinuria (5.7%). There were no treatment-related 12.7–19.6]). The most frequent grade 3 or 4 adverse events were neutropenia (15.2%),
serious adverse events or on-treatment deaths. Conclusions: These encouraging long- appetite loss (12.1%), and diarrhea (7.6%). Conclusions: Trastuzumab biosimilar (CT-P6)
term efficacy results and the manageable safety profile suggested a preferable position of combined with SOX or CapeOX demonstrated robust antitumor efficacy and manageable
Sur plus Pac as the 2L treatment for aGC, notably, following the current standard 1L IO- toxicity in patients with advanced or recurrent gastric cancer. Clinical trial information:
containing regimens. Clinical trial information: ChiCTR2200063336. Research Sponsor: jRCTs071190007. Research Sponsor: None.
HUTCHMED.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 279s

4030 Poster Session 4031 Poster Session

Neoadjuvant serplulimab in combination with chemotherapy for locally Safety, efficacy, and biomarker analysis from a phase II trial of intensive
advanced gastric or gastro-esophageal junction cancer. First Author: Hon- chemotherapy combined with serplulimab and trastuzumab in patients with
gjie Zhan, Tianjin Medical University Cancer Institute and Hospital, Tianjin, China advanced HER2-positive gastric cancer. First Author: Tianshu Liu, Zhongshan
Background: Perioperative therapy combined with surgery is the standard treatment for Hospital, Fudan University, Shanghai, Shanghai, China
locally advanced gastric or gastro-esophageal junction (G/GEJ) cancer; however, long- Background: Keynote-811 has proven the efficacy of combined PD-1 and HER2 blockade with chemotherapy
in HER2-positive gastric cancer. Our study aims to enhance the survival of patients by replacing standard
term survival remains suboptimal. Immunotherapy has demonstrated antitumor activity
chemotherapy with intensive chemotherapy and conducting biomarker analysis. Methods: This prospective,
and an acceptable safety profile in advanced G/GEJ cancer. This study aimed to evaluate single-arm, open-label study was carried out across five centers in China, recruiting patients (pts) with
the efficacy and safety of neoadjuvant Serplulimab, a PD-1 inhibitor, in combination with unresectable locally advanced or metastatic HER2-positive gastric cancer. Pts receive Serplulimab (4.5 mg/kg,
chemotherapy in previously untreated, locally advanced G/GEJ cancer. D1, Q3W), Trastuzumab (initial 8 mg/kg, D1, then 6 mg/kg, D1, Q3W), and the DOS regimen: oxaliplatin (100 mg/
Methods: Eligible patients with resectable G/GEJ adenocarcinoma staged as cT3- m², IV), docetaxel (40 mg/m², IV), and S-1 (40-60mg, BID, D1-14, Q3W). Chemotherapy is up to 8 cycles.
Serplulimab and Trastuzumab can be administered until tumor progression. Gastroscopic biopsy was taken
4aN+M0 were enrolled. The treatment regimen included Serplulimab (300 mg on day 1, before the treatment and dynamic blood samples were collected at C1D1 (T0), C2D1 (T1) and C7D1 (T2) for
every 3 weeks [Q3W]) plus SOX chemotherapy (oxaliplatin 130 mg/㎡on day 1; and S-1 biomarker analysis. Genomic DNA from tumor tissue and circulating tumor DNA (ctDNA) underwent targeted
60 mg twice daily on days 1-14, Q3W) for three cycles. Following the preoperative DNA sequencing containing 571 genes. Results: From July 2022 to September 2024, 40 pts were recruited.
treatment, patients underwent surgery 6-8 weeks after the therapy. The primary The median follow-up was 7.9 months. There were 10 females and 30 males, with a median age of 59 (31-74).
endpoint was pathological complete response (pCR) rate. Secondary endpoints included 37 pts were eligible for efficacy assessment. The objective response rate (ORR) was 92% (95% CI: 0.87, 0.96),
with a complete response rate of 3% (95% CI: 0, 0.05), and a partial response rate of 89% (95% CI: 0.84, 0.94).
major pathological response (MPR) rate, R0 resection rate, disease-free survival (DFS), Median progression free survival has not been reached. 38 pts (95%) experienced adverse events (AEs) of any
overall survival (OS), and toxicity. Results: From October 13, 2023, to December 31, grade. Grade 3 or above AEs occurred in 14 pts (35%), 2 pts with grade 4 AEs (1 with thrombocytopenia and 1
2024, 25 patients were enrolled, all of whom completed neoadjuvant therapy and with myelosuppression). No grade 5 events were observed. The most common AEs were anemia (35%),
underwent radical surgical. The R0 resection rate was 100%. Five patients achieved pCR neutropenia (23%), nausea and vomiting (20%) and leukopenia (20%). 15 (38%) pts had dose interruptions due
to AEs, with no treatment discontinuation. HER2 CNV gain was detected in 82.1% (23/28) of tissue samples and
(pCR rate: 20%), and an additional five patients achieved MPR (MPR rate: 40%). Median
74.3% (26/35) of ctDNA samples. In matched ctDNA and tissue samples (N = 26), 88.5% (23/26) showed
DFS and OS were not reached. Safety analysis included all patients who received at least concordant HER2 CNV gain results. Elevated tissue HER2 CNV were associated with more pronounced tumor
one cycle of neoadjuvant therapy. The most common treatment-related adverse events shrinkage (R = -0.35, P = 0.073). Plasma HER2 gain detection rate decreased from 79.1% (19/24) at T0 to 16.7%
(TRAEs) of any grade included nausea, anorexia, thrombocytopenia, fatigue, and thyroid (4/24) at T1 (P , 0.001). Similar results were observed from T0 to T2 (P , 0.001). Conclusions: This regimen
dysfunction. No grade $3 TRAEs were observed. Conclusions: Neoadjuvant Serplu- demonstrated remarkable efficacy with a high ORR and manageable toxicity. Tumor HER2 CNV and ctDNA
dynamic monitoring correlated with treatment efficacy. The subsequent results of biomarker analysis will be
limab combined with SOX chemotherapy demonstrated promising efficacy and a fa- presented at the upcoming conference. Clinical trial information: NCT05311189. Research Sponsor: None.
vorable safety profile in patients with locally advanced, resectable G/GEJ
Pts characteristics.
adenocarcinoma. These findings support the potential role of immune-based neo- N=40
adjuvant therapy in this setting. Research Sponsor: None.
Age, ‡65 years 38%
Male 75%
PD-L1 status
CPS ‡1 65%
CPS <1 13%
Unknown 22%
HER2 status
IHC 2+ ISH positive 25%
IHC 3+ 75%
Primary gastrectomy or esophagectomy
Yes 15%
No 85%
Metastatic sites
0–2 60%
‡3 40%

4032 Poster Session 4033 Poster Session

First-line osemitamab (TST001) plus nivolumab and CAPOX for advanced G/ Real-world analyses to evaluate the role of TIGIT as a target in first-line (1L)
GEJ cancer (TranStar102): Updated results of cohort G from a phase I/IIa gastric, gastroesophageal junction, and esophageal adenocarcinomas (GC/
study. First Author: Jifang Gong, Early Drug Development Center, Key Laboratory of GEJC/EAC). First Author: Linda Su-Feher, Gilead Sciences, Inc, Foster City, CA
Carcinogenesis and Translational Research (Ministry of Education), Peking University Background: Anti-programmed death protein 1 (PD-1) immune checkpoint inhibitors
Cancer Hospital and Institute, Beijing, China (ICIs) are approved for GC/GEJC/EAC.T-cell immunoreceptor with Ig and ITIM domains
Background: Claudin 18.2 (CLDN18.2) is a clinically validated therapeutic target and has (TIGIT) is a potential anti-tumor ICI [Link] how PD-L1 (CD274), effector T-
been broadly explored with different modalities in gastric/gastroesophageal (G/GEJ) cancer cell (Teff), and TIGIT RNA expression levels correlate with real-world outcomes of 1L
treatment. Osemitamab is a humanized monoclonal antibody with improved affinity to treatment can inform the potential for TIGIT as a target. Methods: Deidentified real-world
CLDN18.2, reduced fucosylation and enhanced ADCC activity and has been observed to data for patients (pts) with metastatic GC/GEJC/EAC were included from the Tempus AI,
upregulate PD-L1 expression on CLDN18.2-positive tumor cells. In vivo anti-tumor activity of Inc. clinicogenomic database. Data were analyzed for pts with available longitudinal
combination of osemitamab plus an anti-PD-1/PD-L1 antibody and chemotherapies was clinical data and biopsies (whole transcriptome RNA-sequencing [seq], 648 gene panel
significantly stronger than any of the doublet combinations, regardless of the PD-L1 CPS DNA-seq, and PD-L1 immunohistochemistry). Gene expression comparisons, including
levels, making the triple combination of osemitamab, nivolumab and CAPOX an attractive between TIGIT and a Teff gene set (geometric mean of CD8A, GZMA, GZMB, IFNG, EOMES,
combination to explore. Methods: Cohort G from TranStar102 (NCT04495296, a phase I/II and PRF1), used the Wilcoxon test, Kruskal-Wallis test, and Spearman correlation. Real-
study) was designed to evaluate the safety and preliminary efficacy of osemitamab at two world time to next treatment or death (rwTTNTD; maximum follow-up 24 months) was
dose levels (3mg/kg or 6mg/kg Q3W) plus nivolumab and CAPOX as the first-line treatment in assessed for 1L ICI + chemotherapy (chemo) or chemo using the Kaplan-Meier method
patients with G/GEJ cancer. 40 patients were planned to be enrolled in each dose level. Key and compared by expression level (high [$ median] vs low [ , median]) of TIGIT, Teff
eligible criteria included HER2 negative or unknown, unresectable locally advanced or
gene set, and TIGIT normalized by Teff gene set. Results: Among 545 pts (1L ICI + chemo,
metastatic G/GEJ cancer, regardless of CLDN18.2 or PD-L1 expression and treatment naı̈ve
n = 50; 1L chemo, n = 124) TIGIT expression was most highly correlated with Teff (R = 0.80,
for advanced disease. The endpoints include safety, efficacy, PK and predictive value of the
P , 0.001) and FOXP3 expression (R = 0.78, P , 0.001), followed by CD274 expression (R
different levels of CLDN18.2 expression, etc. Results: As of Jan 13, 2025, 82 patients were
dosed with osemitamab plus nivolumab and CAPOX (40 at 3mg/kg, 42 at 6mg/kg) with = 0.57, P , 0.001). Positive correlations with PD-L1 combined positive score were
median follow-up of 21.3 months. All patients experienced treatment-related adverse events observed for TIGIT and Teff expression (P , 0.001 for both) but not TIGIT normalized to
(TRAEs). The safety profile was similar with the previously presented data (2024 ESMO Teff levels (P . 0.05). Biopsies from liver metastases, which tend to show a poor re-
poster), and the most commonly observed TRAEs were on-target and off-tumor toxicities, sponse to ICIs, had lower immune signatures vs stomach tumor biopsies (P , 0.001).
such as nausea, vomiting and hypoalbuminemia, and were manageable. 44 out of 82 patients High vs low expression of TIGIT and Teff were associated with numerically longer
had confirmed partial response and the objective response rate (ORR) was 55.7%. There was a rwTTNTD for ICI + chemo, with HRs (95% CI) of 0.82 (0.43–1.54) for TIGIT (n = 29 vs 21)
clear trend between anti-tumor efficacy and CLDN18.2 expression, with a confirmed ORR of and 0.78 (0.42–1.48) for Teff (n = 28 vs 22). Corresponding HRs (95% CI) for chemo were
68% and median progression-free survival of 16.6 months (95% CI: 5.8, 21.7) for the patients 1.04 (0.70–1.54) for TIGIT (n = 53 vs 71) and 1.26 (0.85–1.87) for Teff (n = 64 vs 60). When
with CPS known and CLDN18.2 high/medium expression (defined as CLDN18.2 membranous TIGIT expression was normalized to Teff levels, HRs (95% CI) for rwTTNTD for high vs low
staining $2+ in $40% of tumor cells by immunohistochemistry in the central laboratory) expression were 0.92 (0.49–1.71) for ICI + chemo (n = 23 vs 27) and 0.81 (0.54–1.20) for
(n=26). By the cut-off date, there were 33 patients in survival follow-up including 12 patients chemo (n = 57 vs 67). Conclusions: TIGIT expression was highly correlated with FOXP3
still with ongoing treatment. The median overall survival was 20.4 months (95% CI:15.0, NE) and Teff gene set expression in GC/GEJC/EAC tumors. The improved rwTTNTD seen in pts
for all the 82 patients. Updated clinical data and details by biomarkers levels will be reported with high TIGIT expression may be driven by these patients also having high Teff ex-
at the time of conference. Conclusions: The updated data indicate that the combination of pression. When TIGIT is normalized to Teff, pts with high TIGIT expression lose this added
TST001 plus nivolumab and CAPOX for the first-line treatment of patients with G/GEJ cancer benefit. Combining anti-TIGIT and anti–PD-1 treatments may lead to enhanced T cell
is safe and well tolerated with encouraging durable PFS and survivals, especially for the activation, which could bring benefit to pts with 1L GC/GEJC/EAC. Larger studies will help
patients with high/medium CLDN18.2 expression. Clinical trial information: NCT04495296. confirm these findings. Research Sponsor: Gilead Sciences, Inc.
Research Sponsor: Suzhou Transcenta Therapeutics Co, Ltd.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
280s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4034 Poster Session 4035 Poster Session

Distinct microenvironment phenotypes across tumor, vascular, and immune GABRP as a biomarker for predicting anti-PD1 therapeutic efficacy in ad-
compartments from ramucirumab plus pembrolizumab in refractory gastric vanced gastric cancer (WJOG10417GTR study). First Author: Naoki Takahashi,
cancer: A phase II trial. First Author: Sung Hee Lim, Division of Hematology-Oncology, Department of Gastroenterology, Saitama Cancer Center, Saitama, Japan
Department of Medicine, Sungkyunkwan University School of Medicine, Samsung Background: GABRP, Gamma-aminobutyric acid type A receptor subunit p, is a key
Medical Center, Seoul, South Korea molecule that influences tumor properties and patient prognosis in various cancers,
Background: Combining immune checkpoint inhibitor (ICI) with vascular endothelial including gastric cancer (GC): GABRP is significantly upregulated in tumor tissues, and
growth factor (VEGF)/VEGF receptor inhibition has yielded promising results in multiple promotes tumor progression and metastasis via chemokine signaling and macrophage
tumor types. This phase II trial investigated the efficacy and molecular mechanisms of recruitment, leading to poor prognosis. Thus, GABRP seems to be a promising potential
ramucirumab plus pembrolizumab in refractory gastric cancer patients. Methods: We target for cancer treatment. However, it remains unclear whether the GABRP-targeting
conducted a prospective phase II single-arm trial of ramucirumab plus pembrolizumab strategy can contribute to GC treatment, especially anti-PD1/PDL1 therapy, which has
as salvage treatment in patients with metastatic GC who failed to respond to standard recently attracted great attention. Methods: We collected biopsy tumor tissues from 96
fluoropyrimidine plus platinum with or without programmed cell death protein 1 (PD-1) patients with advanced GC before and one month after nivolumab monotherapy in the
inhibitors. Eligible patients had programmed death-ligand 1 (PD-L1) combined positive WJOG10417GTR study according to the protocol (No. 2017-473) approved by the IRB,
score more than five. The primary objective was to objective response rate (ORR) and and analyzed for GABRP expression by RNA-sequencing (RNAseq) and immunohis-
secondary end points included disease control rate, duration of response, progression- tochemical staining (IHC; only at pre-treatment). Progression-free survival (PFS) and
free survival and overall survival, and toxicity. Comprehensive molecular profiling, in- overall survival (OS) were compared between high and low groups divided by the cutoff
cluding Digital Spatial Profiling, and mass cytometry was performed on tumor samples value determined by change point of the larger log hazard ratios (HRs) based on the
and peripheral blood. Results: Twenty-six patients were enrolled in this study between method for modeling continuous-scale covariates. This study was supported by Ono
June 2021 and May 2023. No patient attained complete response (CR), while 6 patients Pharmaceutical and Bristol Myers Squibb. Results: RNAseq data revealed that high
achieved confirmed partial response (PR), resulting in a response rate (RR) of 23.1% levels of GABRP gene expression in tumors at baseline were significantly associated
(95% CI, 4.06–34.4). The median PFS for all patients was 2.7 months (95% CI, 1.84 - with shorter PFS (median PFS [mPFS] 1.643 months [95% CI = 1.281 - 3.253] vs
3.56 months), and median OS was 10.9 months (95% CI, 2.31 – 18.29). Grade $ 3 3.170 months [1.971 - 10.251], HR = 2.798, P = 0.005) and shorter OS (median mOS
treatment-related adverse events occurred in 38.5 % of patients. Digital Spatial Profiling [mOS] 5.092 months [3.450 - NA] vs 15.409 months [13.733 - NA], HR = 5.145, P ,
revealed distinct tumor microenvironment (TME) phenotypes between responders and 0.001), and the high post-treatment levels were significantly associated with shorter PFS
non-responders. Responders had high CTL infiltration and vascular normalization in their (mPFS 1.873 months [1.511 - 3.253] vs 4.123 months [2.073 - NA], HR = 3.333, P = 0.009)
tumor specimen after treatment which led to favorable treatment outcome. In support of and shorter OS (mOS 7.162 months [5.092 - 15.310] vs NA [13.733 - NA], HR = 4.524, P =
this, spatial profiling revealed the shortest median distance between immune cells and 0.015), as compared to low levels. This was confirmed by IHC data that showed similar
vessels, suggesting enhanced transendothelial migration of immune cells. Nonre- results, albeit only baseline data: patients with high baseline levels of GABRP protein
sponders had significantly upregulated TGF-b pathway and low tumor vascularity. On- expression in tumors had significantly shorter PFS (mPFS 1.528 months [0.986 - 1.971]
treatment analysis demonstrated a shift towards increased immune cell infiltration and vs 3.121 months [1.873 - 4.238], HR = 1.727, P = 0.016). Conclusions: These results
decreased tumor cellularity. Mass cytometry of peripheral blood revealed lower pro- suggest that GABRP expressed in tumors is a significant poor prognostic factor for
portions of myeloid-derived suppressor cells in responders. nivolumab monotherapy in advanced GC. GABRP may be a promising biomarker for
Conclusions: Ramucirumab+pembrolizumab showed modest clinical efficacy, with predicting anti-PD1/PDL1 therapeutic efficacy in advanced GC, and targeting it may
manageable toxicity and durable responses. Although limited to a small subset of contribute to improving the clinical outcomes in GC as a new therapeutic strategy.
patients, a few patients who had previously responded to ICI benefited from ramu- Research Sponsor: Ono Pharmaceutical and Bristol Myers Squibb.
cirumab+pembrolizumab. Clinical trial information: NCT0005753. Research Sponsor:
None.

4036 Poster Session 4037 Poster Session

KEYNOTE-859: 4.5-year median follow-up of pembrolizumab plus chemo- Co-expressing pattern of multiple biomakers and dynamic change of Clau-
therapy for previously untreated advanced HER2-negative gastric or gas- din18.2 expression after systemic chemotherapy in advanced gastric
troesophageal junction (G/GEJ) adenocarcinoma. First Author: Sun Young Rha, cancer. First Author: Sun Young Rha, Yonsei Cancer Center, Yonsei University Col-
Yonsei Cancer Center, Yonsei University College of Medicine, Yonsei University Health lege of Medicine, Seoul, South Korea
System, Seoul, South Korea Background: Gastric cancer (GC) is a heterogeneous disease classified by Lauren
Background: In the phase 3KEYNOTE-859 study (NCT03675737), first-linepembrolizumab (pembro) classification and TCGA’s four molecular subtypes. Targeting Claudin (CLDN)18.2 has
+ chemotherapy (chemo) continued to provide longer OS (HR, 0.79; 95% CI, 0.71-0.88) and PFS (HR, emerged as a promising therapy. However, a comprehensive understanding of CLDN18.2 in
0.76; 95% CI, 0.68-0.85), and a higher ORR (51.0% vs 42.0%) vs placebo + chemo in participants (pts) advanced GC, its clinicopathologic correlations, and prognostic significance remains
with HER2-negative G/GEJ adenocarcinoma, after a median follow-up of 41.6 mo (August 22, 2023). limited. We analyzed the co-expression of multiple biomarkers and dynamic changes in
We present results after an additional 13 mo of follow-up. Methods: Eligible pts with untreated CLDN18.2 following systemic chemotherapy. Methods: This retrospective study included
locally advanced or metastatic HER2-negative G/GEJ adenocarcinoma with PD-L1 status, mea-
surable disease, and ECOG PS 0 or 1 were randomly assigned 1:1 to receive pembro 200 mg or
402 patients with stage IV GC at Yonsei Cancer Center (2015–2023) treated with palliative
placebo IV Q3W for #35 cycles + investigator’s choice of chemo (5-FU + cisplatin [FP] vs cape- doublet chemotherapy. Paired pre- and post-treatment tissues were available for 124
citabine + oxaliplatin [CAPOX]). The primary end point was OS. Secondary end points included PFS, patients. CLDN18.2 expression was assessed using the VENTANA CLDN18 (43-14A) RxDx
ORR, and DOR, all per RECIST v1.1 by BICR, and safety. The data cutoff was September 27, 2024. Assay, with positivity defined as $75% of tumor cells. HER2, EBV, dMMR, and PD-L1(22C3)
Results: Median follow-up was 54.8 mo (Q1-Q3, 46.8-62.1). In all randomly assigned pts in the were also evaluated by immunohistochemistry. Overall survival (OS) was estimated using
intention-to-treat population (N = 1579), median OS was 12.9 mo (95% CI, 11.9-14.0) for pembro + the Kaplan-Meier method, and a log-rank test was performed to compare survival
chemo vs 11.5 mo (95% CI, 10.6-12.1) for placebo + chemo (HR, 0.78; 95% CI, 0.70-0.86). In pts with according to CLDN18.2 positivity. Results: The CLDN18.2 positivity rate was 30.6%,
PD-L1 CPS $1, median OS was 13.0 mo (95% CI, 11.6-14.2) vs 11.4 mo (95% CI, 10.5-12.0; HR, 0.74 consistent with previous reports. It was slightly higher in HER2-negative (31.5%) than
[95% CI, 0.66-0.84]). In pts with PD-L1 CPS $10, median OS was 15.8 mo (95% CI, 14.0-19.3) vs 11.8 HER2-positive (27.3%) (p = 0.44). CLDN18.2 positivity was significantly higher in EBV-
mo (95% CI, 10.3-12.7; HR, 0.64 [95% CI, 0.53-0.77]). PFS, ORR, and DOR were also consistent positive than negative (63.6% vs 29.4%, p , 0.05) and in pMMR versus dMMR (31.7% vs
between the intention-to-treat population and pts with PD-L1 CPS $1 and PD-L1 CPS $10 (Table).
9.5%, p , 0.05). Regarding PD-L1 status, the positivity rate was higher in the PD-L1
Treatment-related AEs were reported in 751 pts (95.7%; grade 3-5, 466 [59.4%]) for pembro + chemo
and 736 (93.5%; grade 3-5, 404 [51.3%]) for placebo + chemo. Conclusions: Pembro + chemo
negative compared to the positive by CPS 1(36.1% vs 25.8%, p , 0.05). We observed that
continued to show improved OS, PFS, and ORR vs placebo + chemo after a median study follow-up of CLDN18.2 expression seems to be decreasing with higher PD-L1 expression (28.5 % in
54.8 mo, regardless of PD-L1 status. The findings further support pembro + chemo as a first-line patients with CPS $5, 24% in CPS $10). The OS (median months, 95% CI) based on CLDN
treatment option for locally advanced or metastatic HER2-negative G/GEJ adenocarcinoma. 18.2 expression was similar [22.5 (18.6-25.2) vs 23.2 (17.9-27.4) in CLDN18.2 positive
Research Sponsor: None. Clinical trial information: NCT03675737. Research Sponsor: None. group], regardless of HER2 status. Among 124 patients with paired samples, pre-treatment
CLDN18.2 positivity was 22.8%, increasing to 36.5% post-treatment. The increase was
All pts PD-L1 CPS ‡1 PD-L1 CPS ‡10
N = 1579 n = 1235 n = 553 more pronounced in HER2 negative (39.0%) compared to positive (29.5%). Notably, 79.9%
Pembro + Pbo + Pembro + Pbo + Pembro + Pbo + of patients showed consistent CLDN18.2 expression between pre- and post-treatment
chemo chemo chemo chemo chemo chemo samples, while 20.1% demonstrated changes in expression. These changes were not
n = 790 n = 789 n = 618 n = 617 n = 280 n = 273
associated with other markers such as HER2, EBV, dMMR, or PD-L1. Conclusions: Higher
OS, median (95% CI), 12.9 (11.9- 11.5 (10.6- 13.0 (11.6- 11.4 (10.5- 15.8 (14.0- 11.8 (10.3- CLDN18.2 positivity rates in specific subgroups, such as pMMR and PD-L1 negative,
mo 14.0) 12.1) 14.2) 12.0) 19.3) 12.7)
HR (95% CI) 0.78 (0.70-0.86) 0.74 (0.66-0.84) 0.64 (0.53-0.77) suggest the potential role of anti-CLDN therapy in these subgroups. This study also
PFS, median (95% CI), mo 6.9 (6.3-7.2) 5.6 (5.5-5.7) 6.9 (6.0-7.2) 5.6 (5.4-5.7) 7.8 (6.8-8.5) 5.6 (5.4-6.7) highlights the dynamic changes of CLDN18.2 expression in advanced gastric cancer, with
HR (95% CI) 0.76 (0.68-0.85) 0.72 (0.64-0.82) 0.62 (0.51-0.76)
ORR, % (95% CI) 51.1 (47.6- 42.0 (38.5- 51.9 (47.9- 42.6 (38.7- 60.4 (54.4- 43.2 (37.3- an increase observed after first-line systemic treatment in a subset of patients. Further
54.7) 45.5) 55.9) 46.6) 66.1) 49.3)
DOR, median (range), mo 8.0 5.7 8.3 5.6 10.0 5.7
studies should focus on prospective analyses and stratify changes in CLDN18.2 expression
(1.2+ to 66.3+) (1.3+ to 58.1+) (1.2+ to 66.3+) (1.3+ to 58.1+) (1.2+ to 66.3+) (1.4+ to 55.0+) by treatment regimen to better understand its role as a therapeutic target and its im-
plications for treatment resistance and disease progression. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 281s

4038 Poster Session 4039 Poster Session

Camrelizumab plus albumin-bound paclitaxel and S-1 as first-line treatment Efficacy and safety of LM-302 (anti-claudin 18.2 ADC) in combination with
of HER-2 negative unresectable locally advanced or advanced gastric and anti-PD-1 therapy for advanced gastric, gastroesophageal junction cancer
gastroesophageal junction adenocarcinoma: A phase II clinical trial. First and esophageal adenocarcinoma: Early-phase study results. First Author:
Author: Xinfang Hou, The Affiliated Cancer Hospital of Zhengzhou University & Henan Haiping Jiang, The First Affiliated Hospital, Zhejiang University School of Medicine,
Cancer Hospital, Zhengzhou, Henan, China Hangzhou, China
Background: Gastric cancer is a common malignant tumor of the digestive tract. For Background: Claudin 18.2 (CLDN18.2), a tight junction protein highly expressed in gastric
nearly 90% of HER2 negative advanced gastric and gastroesophageal junction ade- and gastroesophageal junction (GEJ) cancers, has emerged as a promising therapeutic
nocarcinoma (G/GEJ) patients (pts), the current first-line treatment options and efficacy target. LM-302 (tecotabart vedotin), a novel and potent MMAE-based ADC targeting
are limited, and new therapeutic drugs are urgently needed to improve the efficacy of CLDN18.2, has shown promising efficacy and safety as monotherapy in heavily pretreated
advanced G/GEJ. Methods: This is a prospective, single center, single arm, open label advanced gastric cancer. This pooled analysis evaluates the efficacy and safety of LM-
phase II clinical trial involving HER-2 negative G/GEJ diagnosed by histopathology. Pts 302 in combination with the anti-PD-1 antibody toripalimab as a first-line treatment
received 4-6 cycles of treatment with camrelizumab combined with albumin bound option for patients with gastric, GEJ, or esophageal adenocarcinoma (EAC).
paclitaxel and S-1, followed by camrelizumab maintenance therapy until progression or 2 Methods: Eligible patients with histologically confirmed, previously untreated, unre-
years. By following up with pts and evaluating clinical efficacy based on imaging sectable, HER2-negative gastric, GEJ, or EAC were included. Patients received LM-302
examinations, the effectiveness and safety of the protocol can be assessed. (1.6 mg/kg Q3W, 2.0 mg/kg Q3W or 1.8 mg/kg Q2W) and toripalimab (240 mg Q3W or 3
Results: From December 2020 to December 2024, a total of 47 pts obtained informed mg/kg Q2W). Endpoints included safety, objective response rate (ORR), disease control
consent based on exclusion criteria. As of the deadline (January 6, 2025), the average rate (DCR), progression-free survival (PFS), OS, and biomarker analysis. Data cutoff:
follow-up period was 17.4 months, with a median follow-up time of 16.7 months (range: January 7, 2025. Results: A total of 43 gastric, GEJ, or EAC patients (median age: 62.3
1.6-42.2 months). One patient had completed 2 years of treatment and ended main- years; 65.1% male) from Australia and China were treated. No dose-limiting toxicities
tenance therapy with camrelizumab, 5 pts (12.5%) were still receiving treatment, 34 pts (DLT) were observed across all these dose levels. Treatment-related adverse events
(85.0%) stopped treatment, and 28 pts (70%) developed disease progression (PD). The (TRAEs) related to LM-302 were reported in 39 patients (90.7%), with common events
best therapeutic effect was achieved in 2 cases with CR, 25 cases with PR, 11 cases with ($20%) including anemia, decreased white blood cell count, decreased neutrophil count,
SD, and 2 cases with PD. The experiment reached the primary endpoint with an ORR of increased aspartate transaminase (AST), vomiting, loss of appetite, and nausea.
67.5%. The secondary endpoint DCR was 95.0%, the median PFS was 7.8 months [95% Grade $3 TRAEs related to LM-302 occurred in 16 patients (37.2%), the most common
confidence interval (CI): 6.2-9.4 months], and the median OS was 23.8 months [95% events ($5%) were decreased neutrophil count (14.0%), increased alanine transaminase
confidence interval (CI): 15.2-32.4 months]. For all study subjects, there was no cor- (11.6%), increased AST (9.3%), and anemia (7.0%). Among 41 efficacy-evaluable patients
relation between OS, PFS, and baseline characteristics of the study (age, gender, ECOG, (median follow-up: 6.01 months), ORR was 65.9% (95% CI: 49.4-79.9%), and DCR was
microsatellite status, mismatch repair protein status, EBER status, PDL1 level, number 85.4% (95% CI: 70.8- 94.4%). In 32 GC patients with CLDN18.2 expression in $25% of
of metastatic sites, presence of liver metastasis, presence of lung metastasis, presence tumor cells (IHC 2+/3+), ORR was 71.9% (95% CI: 53.3-86.3%) and DCR was 96.9% (95%
of abdominal metastasis, presence of bone metastasis). Among all enrolled pts, 36 CI: 83.8-99.9%). Among these patients, ORR was 63.3% (95% CI: 35.1-87.2%) for patients
experienced adverse events (76.6%) and 18 experienced grade 3-4 TRAE (38.3%), mainly with PD-L1 CPS , 1 and 77.8% (95% CI: 52.4-93.6%) for patients with PD-L1 CPS $1.
bone marrow suppression and immune related rash. Conclusions: As the first-line Median PFS and OS were not reached; one patient with PD-L1 CPS , 1 achieved PR and
treatment for HER-2 negative advanced gastric and gastroesophageal junction ade- remained on treatment for 14.70 months. Conclusions: LM-302 combined with tor-
nocarcinoma, the combination of Carilizumab with albumin bound paclitaxel and S-1 has ipalimab demonstrated encouraging anti-tumor activity and manageable safety as first-
shown encouraging efficacy. Clinical trial information: ChiCTR2300069672. Research line treatment for patients with CLDN18.2-positive gastric, GEJ, and EAC, including those
Sponsor: None. with low-to-moderate CLDN18.2 expression. These findings support further large-scale
clinical trials to confirm efficacy, safety and clinical utility. Clinical trial information:
NCT05188664; NCT05934331. Research Sponsor: LaNova Medicines Limited.

4040 Poster Session 4042 Poster Session

Efficacy and safety results of a multi-center phase II study of utidelone Fruquintinib in combination with camrelizumab and paclitaxel liposome and
injection in combination with PD-1 inhibitor and chemotherapy for the first- nedaplatin as first-line treatment for advanced esophageal squamous cell
line treatment of advanced gastric and esophagus cancers. First Author: Meili carcinoma (ESCC): A single-arm, phase II study. First Author: Yanhong Gu, The
Sun, Department of Oncology, Shandong University First Medical University Affiliated First Affiliated Hospital of Nanjing Medical University, Nanjing, China
Central Hospital, Jinan, Shandong, China Background: Previous studies have indicated a synergistic effect of fruquintinib in
Background: Utidelone is a novel microtubule inhibitor that offers several advantages in- combination with chemotherapy or immunotherapy. Therefore, we conducted a phase II
cluding improved efficacy and safety, broader anti-cancer spectrum, blood-brain barrier study to evaluate the efficacy and safety of fruquintinib plus camrelizumab, paclitaxel
penetrance, and response against multidrug-resistant tumors. Utidelone injection (UTD1) has liposome, and nedaplatin as a first-line treatment for advanced esophageal squamous cell
been approved for advanced breast cancer in China. Previous studies have shown that carcinoma (ESCC). Methods: This study consisted of a dose-finding and a dose-expansion
utidelone monotherapy is effective for advanced gastric and gastroesophageal junction phase. A total of 33 to 36 eligible patients with untreated advanced ESCC were planned for
adenocarcinoma (GC) and esophageal squamous cell carcinoma (ESCC). This phase II study enrollment. In the dose-finding phase, based on a standard 3+3 design, patients were treated
further explored the efficacy and safety of utidelone in combination with PD-1 inhibitor and with fruquintinib (3 mg, 4 mg, 5 mg, days 1-14, every 3 weeks, respectively, the initial
chemotherapy for the 1st line treatment of advanced GC and ESCC. Methods: Eligible patients fruquintinib dose was 4mg), in combination with a fixed dose of camrelizumab 200 mg,
were recruited into the two cohorts of metastatic and/or unresectable HER2 negative GC or paclitaxel liposome 135 mg/m2, and nedaplatin 70 mg/m2 on day 1, every 3 weeks. In the
ESCC, receiving utidelone (30mg/m2/day iv on days 1–5 every 21-day) plus sintilimab (200 mg dose-expansion phase, patients received camrelizumab, paclitaxel liposome, nedaplatin and
iv Q3W) and oxaliplatin (130 mg/m2/day Q3W for up to 6 cycles), or utidelone (30mg/m2/day recommended phase 2 dose (RP2D) of fruquintinib. A maximum of six cycles was ad-
iv on days 1–5 every 21-day) plus tislelizumab (200 mg iv Q3W) and capecitabine (2000mg/ ministered, followed by maintenance therapy with fruquintinib in combination with cam-
m2/day po on days 1–14 every 21-day), respectively, until disease progression or unac- relizumab. The primary endpoint was objective response rate (ORR). Secondary endpoints
ceptable toxicity. The primary endpoint is ORR and second endpoints are CBR, PFS and included disease control rate (DCR), progression-free survival (PFS), and safety. Results: As
safety. Results: The enrolment has been completed for both cohorts from April 2023 to of December 28, 2024, 22 patients (20 males) were enrolled with a median age of 65 years
October 2024. There were 27 eligible patients enrolled in the GC cohort with median age of 60 (range 50-76). Among the 21 patients with metastases, 19.0% (4/21) had a single metastatic
years (range 34-70), 6 females and 21 males. 23 patients were evaluable for efficacy with an site and 81.0% (17/21) had multiple metastatic sites. In the dose-finding phase, no DLTs
outcome of 11 confirmed PRs and 12 SDs including 4 unconfirmed PR, and 5 patients were occurred in 3 patients at 4 mg and 6 patients at 5 mg, establishing fruquintinib’s RP2D as
still receiving treatment (1-23 cycles). The confirmed ORR was 47.8% and CBR was 100%. The 5 mg. Of the 19 patients evaluable for efficacy, 13 achieved a partial response, and 6 had
mPFS was . 5.3 months. The most common $Grade 3 TEAEs included diarrhea (25.93%), stable disease. The confirmed ORR was 68.4% (95% CI: 47.5%-89.3%) for all evaluable
neutropenia (14.81%), vomiting (14.81%), peripheral sensory neuropathy (11.11%), anemia patients and 68.8% (95%CI: 41.3%-89.0%) for patients receiving 5 mg dose of fruquintinib.
(11.11%) and leukopenia (11.11%). Other AEs were all Grade 1 or 2, with no treatment-related The confirmed DCR was 100.0% (95%CI: 82.4%-100.0%). At a median follow-up of 5.1 (95%
deaths. There were 20 eligible patients enrolled in the ESCC cohort with median age of 61.5 CI: 4.2-7.2) months, the median PFS was 8.7 (95%CI: 5.2-not available [NA]) months, the 6-
years (range 47-70), 5 females and 15 males. 18 patients were evaluable for efficacy with an month PFS rate reached 81.7%. The most common treatment-related adverse events
outcome of 6 confirmed PRs and 12 SDs, and 6 patients were still receiving treatments (1-12 (TRAEs) were anemia 72.7% (16/22), neutropenia 40.9% (9/22), leukopenia, hypertension
cycles). The confirmed ORR was 33.3% and CBR was 100%. TEAEs $Grade 3 occurred in 8 31.8% (7/22). Grade$3 TRAEs were identified in seven patients, including neutropenia
patients including hypokalemia, lymphopenia/leukopenia, peripheral sensory neuropathy, 13.6% (3/22), leukopenia 13.6% (3/22), oral mucositis 9.1% (2/22), nausea 4.5% (1/22),
hiccup, rash, pain, hypotension and anemia with one occurrence for each (5.5%). Other AEs vomiting 4.5% (1/22), headache 4.5% (1/22), and anemia 4.5% (1/22). No treatment-related
were all Grade 1 or 2, with no treatment-related deaths. Conclusions: Utidelone plus PD-1 serious adverse events (TRSAEs) or deaths occurred. Conclusions: The combination of
inhibitor and chemotherapy demonstrated promising efficacy and acceptable safety as first- fruquintinib, camrelizumab, paclitaxel liposome, and nedaplatin demonstrated significant
line treatment for advanced GC and ESCC. There are 11 patients still on the study receiving efficacy and manageable toxicity profile as a first-line treatment for advanced ESCC,
continuous treatment. The final data will be provided at the time of presentation. Clinical trial suggesting a potential new treatment strategy. Clinical trial information: NCT06010212.
information: NCT04911907. Research Sponsor: Beijing Biostar Pharmaceuticals Co., Ltd. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
282s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4043 Poster Session 4044 Poster Session

Immune checkpoint inhibition in EBV-associated gastric cancer: A multi- Second-line ASKB589 plus chemotherapy for advanced gastric or gastro-
center international retrospective analysis. First Author: Darren Cowzer, Memorial esophageal cancers: Results from cohort 5 of a phase I/II study. First Author:
Sloan Kettering Cancer Center, New York, NY Miao Zhang, Department of Gastrointestinal Oncology, Beijing Cancer Hospital, Beijing,
Background: Epstein-Barr virus associated gastric cancer (EBVaGC) is found in ~9% of GC China
and is associated with a unique immunophenotype. Prior studies suggest 15-100% objective Background: ASKB589 is a humanized monoclonal antibody with enhanced affinity for
response rates (ORR) to immune checkpoint inhibition (ICI), but survival outcomes are CLDN18.2 and increased ADCC activity. It has shown favorable safety profiles and promising
limited by small cohort size. This study sought to clarify outcomes of ICI in EBVaGC across 9 antitumor effects in phase I/II clinical study (NCT04632108) involving first-line patient(pt)s
global tertiary cancer centers, evaluate if next-generation sequencing (NGS) can replace with advanced gastric or gastroesophageal（G/GEJ） cancers. Here, we present the findings
EBER ISH as the gold standard for detection, and identify molecular features of ICI- from cohort 5 that ASKB589 in combination with chemotherapy as second-line treatment for
responsive EBVaGC. Methods: Retrospective data were collected on patients (pts) with patients with advanced G/GEJ cancers. Methods: Cohort 5 from NCT04632108 is designed
metastatic EBVaGC and stratified into patients receiving ICI alone or with chemotherapy, to evaluate the safety and preliminary efficacy of ASKB589 (at 2 doses: 6mg/kg or 10mg/kg
and by 1L vs. 2L+. Progression-free survival (PFS) and overall survival (OS) were analysed Q3W) plus chemotherapy as second line treatment in pts with G/GEJ cancers. Eligible pts who
using the Kaplan-Meier method and were calculated from the start of treatment to the date had confirmed progressive disease during treatment with first line standard of therapy were
of progression or death for PFS, and last follow up or death for OS. Cox regression model enrolled. CLDN18.2 statuses were analyzed retrospectively using IHC DS-3 LDT assay.
stratified by regions was used to examine factors associated with PFS. EBV viral reads were Primary endpoint was safety. Secondary endpoints included objective response rate (ORR),
detected by NGS using independent cohorts by MSK-IMPACT and Caris Life Sciences and disease control rate (DCR), duration of response (DoR), progression free survival (PFS)
concordance with EBER ISH was evaluated in each cohort. Additional phenotypic classi- assessed by investigator per RECIST v1.1 and overall Survival (OS). Results: As of December
fication was performed using clinicogenomic data from the Caris Precision Oncology Al- 20, 2024, 47 pts were enrolled in the 2L cohort and received 6mg/kg ASKB589 plus che-
liance. Results: A total of 91 pts who received ICI in the metastatic setting were included. motherapy. Among them, 44 pts were treated with 6mg/kg ASKB589 plus paclitaxel, while 2
Median age was 62 years (range 30-86) and 91% were male with pts from US (n = 15), Europe pts received 6mg/kg ASKB589 plus docetaxel, and 1 pt received 6mg/kg ASKB589 plus
(n = 5), and Asia (n = 71). ECOG PS at the time of ICI was #1 in 97%. PD-L1 CPS score irinotecan. 17 (27%) pts had received immunotherapy [Link] confirmed ORR was
was $5 in 69%. Pts who received first-line chemotherapy + ICI (n = 42) achieved a 49% 34.2% in 38 measurable and evaluable CLDN18.2 moderate to high GC/GEJ pts (CLDN18.2
investigator-assessed ORR (CR/PR), median PFS (mPFS) 8.3 months (mo), and median OS moderate to high was defined as $40% of tumor cells with $2+ intensity by immuno-
(mOS) 38 mo compared with mOS 18 mo in pts receiving first-line chemotherapy alone (n = histochemistry). 14 pts (36.8%) had stable disease, and the DCR was 71.1%. For the ASKB589
35). Pts who received ICI-alone achieved a similar 49% ORR with 1L and later line mPFS 6 mo plus paclitaxel subgroup (n = 35, with evaluable response), the confirmed ORR was 31.1%,
and 3.2 mo, respectively, which increase to 10.0 mo and 6.6 mo, when limiting to PD-L1 DCR was 71.4%. In the intention-to-treat analysis, the median PFS with 6mg/kg ASKB589 plus
CPS .5. Univariate analysis for PFS among all pts demonstrated improved outcomes in chemotherapy was 5.26 months (95%CI: 2.66, 7.06), and the median OS was 11.14 months
those where PD-L1 CPS score was $5 (HR = 0.57, 95%CI:0.34,0.97). EBV detection by MSK- (95%CI: 8.80, 18.20). In the ASKB589 plus paclitaxel subgroup, the median PFS was
IMPACT and Caris achieved over 98% positive agreement and 99.9% negative agreement. 4.63 months (95%CI: 2.04, 7.06) and median OS of 13.73 months (95%CI: 8.80, 18.20) re-
Additional EBVaGC transcriptomic immune phenotyping data will be presented. spectively. Treatment related adverse events (TRAEs) occurred in all pts including 26 (55.3%)
Conclusions: This global EBVaGC experience demonstrated that EBVaGC can reliably be pts with grade $3 TRAEs. The most common grade $3 TRAEs ($5%) were neutrophil count
identified from NGS, negating the need for costly EBER ISH. Furthermore, pts with EBVaGC decreased (42.6%), white blood cell count decreased (17%), lymphocyte count decreased
achieve a high ORR on ICI with or without concurrent chemotherapy. While 1L PFS was (8.5%), and hypoalbuminemia (8.5%). Conclusions: The results of cohort 5 from
similar to non-EBVaGC, prolonged mOS was achieved, potentially reflecting frequent NCT04632108 have shown that ASKB589 plus chemotherapy, as 2L treatment in pts with
conversion surgeries and regional differences. Even among EBVaGC, immune heterogeneity advanced GC/GEJ cancers have promising antitumor activity and response durability, es-
exists, and PD-L1 CPS .5 identified pts who derived ICI benefit. Further research is needed pecially in the ASKB589 plus paclitaxel subgroups. Both treatment regimens are well tol-
to examine the patterns of disease progression in EBVaGC treated with ICI. Research erated. Further development of ASKB589 combination therapy for CLDN18.2-positive GC/GEJ
Sponsor: None. pts 2nd line treatment is planned. Clinical trial information: NCT04632108. Research Sponsor:
None.

4045 Poster Session 4046 Poster Session

Efficacy of venadaparib plus irinotecan in homologous recombination de- Updated results of fruquintinib combined with PD-1 inhibitors and chemo-
ficiency (HRD) gene mutations as 3+ line treatment in patients with met- therapy in the first-line treatment of HER2-negative advanced gastric or
astatic gastric cancer (mGC). First Author: Won Sik Lee, Idience Co., Ltd., Seoul, gastroesophageal junction adenocarcinoma (FDZL-FIX): A single-arm,
South Korea open-label phase 2 study. First Author: Chenchen Wang, Department of Medical
Background: Tumors with homologous recombination deficiency (HRD) have been suggested to be associated Oncology, Fudan University Shanghai Cancer Center, Shanghai, China
with a favorable response to poly (adenosine diphosphate [ADP]–ribose) polymerase (PARP) inhibitors. The
Background: For unresectable locally advanced or metastatic HER2-negative gastric
aim of this analysis was to evaluate the association between the presence of HRD gene mutations and the
efficacy of venadaparib, a novel PARP inhibitor, plus irinotecan in patients with metastatic gastric cancer cancer (GC), the combination of chemotherapy with immunotherapy, specifically PD-1
(mGC) who had progressed after at least at least 2 lines of therapy. Methods: This was an exploratory analysis inhibitors, has emerged as a new standard of care in first-line treatment of advanced GC,
from a multi-national, phase 1b/IIa trial of venadaparib plus irinotecan in patients with mGC (NCT04725994). but the patients’ outcomes remain poor. Fruquintinib (Fru) is a highly selective inhibitor
Tumor response was evaluated according to RECIST v. 1.1. Genomic analysis was conducted using ctDNA of VEGFR1/2/3. We sought to explore the efficacy and safety of fruquintinib combined
(GuardantOMNI Gene Panel version 1.0, Redwood City, CA) on Day 1 pre-dose. Results of the exploratory
genomic analysis, which includes pathogenic or deleterious mutations of following genes: BRCA1/2, ATM, ATR,
with PD-1 inhibitors and chemotherapy in the first-line treatment of HER2-negative
CHEK2, BARD1 were analyzed for association with efficacy outcomes. Results: As of Dec 2024, 43 patients advanced gastric or gastroesophageal junction adenocarcinoma. Here, we report the
enrolled and objective response (ORR), median progression-free survival (mPFS) and median overall survival updated results. Methods: Eligible pts with HER2 negative locally advanced unre-
(mOS) were 20.9%, 4.2 (2.9-9.6) and 8.1 (6.8-11.5) months. 14 (32.6%) out of the 43 patients had at least one sectable or metastatic gastric or gastrooesophageal junction adenocarcinoma, without
HRD mutation (6 with ATM, 3 with BRCA1, 2 with BRCA2, 1 with BARD1, 1 with CHEK2 and 1 with ATR any systemic anticancer treatment for advanced disease were included in the study. pts
mutations, respectively). For the 14 patients with HRD mutation and 29 patients with no HRD mutation, the
ORR, mPFS (95% CI) and mOS (95% CI) were 35.7% vs. 13.8%, 5.6 (1.3-9.1) vs. 4.0 (2.7-5.4) months and 10.1
received 6 cycles of combined first line treatment with Fru (4mg p.o. qd, d1-14, q3w, )
(6.8-12.0) vs. 8.0 (5.9-11.5) months respectively. For 11 patients with ATM or BRCA1/2 mutation, mPFS (95% combined with PD-1 inhibitor (investigator’s choice of sintilimab 200mg or nivolumab
CI) and mOS (95% CI) were 8.4 (1.2-24.0) and 10.1 (6.8-34.4) months. At the maximum tolerable dose (MTD) 360mg intravenously q3w) and chemotherapy (investigator’s choice of XELOX or SOX)
level of venadaparib 20 mg/d on days 1 to 7 and irinotecan 100 mg/m2 at day 1 of a 2-week cycle, six of 15 regimen. The following maintenance treatment was Fru combined with PD-1 inhibitor
patients had HRD mutation, and mPFS (95% CI) and mOS (95% CI) were 4.1 (2.9-5.4) and 7.9 (5.9-11.5) months;
and S-1/capecitabine until disease progression or intolerable toxicity. The primary
these data are still maturing. Conclusions: Venadaparib in combination with irinotecan demonstrated
promising efficacy in patients with mGC, particularly those with HRD gene mutations. Further development of endpoint was progression-free survival (PFS). Secondary endpoints included objective
this combination may warrant a biomarker-based approach. Clinical trial information: NCT04725994. Research response rate (ORR), overall survival (OS), disease control rate (DCR) and safety.
Sponsor: Idience Co., Ltd. Results: As of Jan 24, 2025, a total of 33 pts (22 male, 11 female) median age 62 (range:
Efficacy results in patients with homologous recombination deficiency mutation. 35-77) were enrolled and received at least one dose of treatment. 30 patients had at least
Mutations in Number of Treatment Overall one tumor assessment post treatment, with an ORR of 80.0% (24/30; 95%CI 61.4-92.3)
HRD-related Prior Palliative Best Overall Duration Survival and DCR of 100%(30/30; 95%CI 88.4-100.0). In the intention to treatment (ITT) pop-
Patient # Genes Treatment Response (%) (months) (months)
ulation, the median PFS was 9.43 months (95% CI 5.29-13.24) and the 9-month PFS
82001-2103 BRCA2 3 SD 25.1 33.9 rates was 57% (95%CI 40.0-83.0). The most common AEs of all grades were neutrophil
82004-2101 ATM 2 PR 38.3 37.3
82002-2103 ATR 3 PD 1.4 8.0 count decreased (33.3%), palmar-plantar erythrodysesthesia syndrome (33.3%), and
82004-2103 ATM 3 PD 1.3 6.7 fatigue (30.3%). Conclusions: The combination of fruquintinib and chemotherapy with
82004-2102 BARD1 2 CR 5.6 11.6
82005-2101 ATM 2 SD 3.1 11.8 PD-1 blockade in the first-line treatment for unresectable locally advanced or metastatic
82005-2103 ATM 2 PR 4.9 8.3 HER2-negative GC had shown promising efficacy and acceptable safety profile. The
82003-2203 BRCA2 2 SD 9.4 11.5 results warrant further investigations in a large cohort. Clinical trial information:
82005-2202 BRCA1 2 SD 6.2 7.0
82003-2201 CHEK2 3 PR 3.1 3.1 NCT06158919. Research Sponsor: None.
82002-2203 ATM 2 SD 1.0 1.0
82004-2201 ATM 2 SD 0.5 0.5
82003-2205 BRCA1 2 SD 2.8 2.8
82001-2201 BRCA1 2 SD 5.3 7.8

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 283s

4047 Poster Session 4048 Poster Session

Artificial intelligence-based prediction of claudin 18.2 expression and im- Predictive value of homologous recombination-related gene mutations in
mune phenotype to guide treatment decisions in patients with gastric survival outcomes of first-line nivolumab plus chemotherapy for gastric
cancer. First Author: Hyung-Don Kim, Asan Medical Center, Seoul, South Korea cancer. First Author: Yuna Lee, Department of Oncology, Asan Medical Center, Uni-
Background: With the increasing availability of immune checkpoint inhibitor (ICI)- and versity of Ulsan College of Medicine, Seoul, South Korea
targeted agent-based treatments for gastric cancer, evaluating predictive biomarkers to Background: Homologous recombination repair (HRR) gene mutations are associated
guide first-line treatment has become increasingly complex. Claudin 18.2 (CLDN18.2) is a with genomic instability; however, their clinical value in the context of immune
clinically relevant target for zolbetuximab-based chemotherapy, typically assessed by checkpoint inhibitor (ICI)-based treatments in gastric cancer remains unclear. We in-
immunohistochemistry. However, limitations such as inadequate tumor specimens, high vestigated the efficacy of nivolumab plus chemotherapy according to the HRR mutation
costs, and long turnaround times pose a practical challenge. Methods: An artificial in- status in advanced gastric cancer patients. Methods: This single-center study included
telligence (AI) model was developed to predict CLDN18.2 expression based on hematoxylin gastric cancer patients with available panel sequencing results who were treated with
and eosin (H&E) slides from 459 patients with gastric cancer (derivation cohort). CLDN18.2 first-line nivolumab plus chemotherapy (n = 115) or chemotherapy alone (n = 172)
positivity was defined as moderate-to-strong expression in $75% of tumor cells. The AI between July 2021 and March 2024. Mutation status of 17 HRR genes (BARD1, BLM,
model utilized a Vision Transformer-based architecture with a multiple-instance learning BRCA1, BRCA2, BRIP1, MRE11A, NBN, PALB2, PARP1, POLD1, RAD50, RAD51, RAD51C,
framework and was trained using five-fold cross-validation. Model performance was RAD51D, RAD52, RAD54L, and XRCC2) was assessed using targeted next-generation
validated in two independent cohorts: an internal cohort of 381 patients treated with first- sequencing. Treatment outcomes were compared according to the presence of HRR
line ICI plus chemotherapy (ICI-Chemo) or chemotherapy alone (Chemo-only) and an mutations. Results: Among patients treated with nivolumab plus chemotherapy, 36.5%
external cohort of 100 patients from diverse ethnic backgrounds. Immune phenotypes harbored HRR mutations. Compared to the no HRR mutation group, the HRR mutation
(IPs) were assessed using an AI-powered whole slide image analyzer to further stratify group exhibited a higher objective response rate (92% vs. 63.2%, P = 0.010), longer
patient outcomes. Results: The prevalence of CLDN18.2 positivity was 43.4% (derivation), progression-free survival (PFS) (median 12.8 vs. 6.5 months; hazard ratio [HR] 0.57, 95%
37.3% (internal validation), and 26.0% (external validation). The model achieved an AUROC confidence interval [CI] 0.36–0.91, P = 0.019) and overall survival (OS) (median not
of 0.753 in the derivation cohort, with sensitivity and specificity of 0.638 and 0.723, reached vs. 14.2 months; HR 0.40, 95% CI 0.23–0.71, P = 0.002). Among patients with
respectively. In the internal and external validation cohorts, AUROCs were 0.752 and 0.746,
HRR mutation, those treated with nivolumab plus chemotherapy showed favorable
with similar sensitivity and specificity levels. Among the internal validation cohort, patients
survival outcomes compared to those treated with chemotherapy alone (PFS: HR 0.44,
were stratified into subgroups based on predicted CLDN18.2 positivity and IP status.
95% CI 0.27–0.71, P , 0.001; OS: HR 0.45, 95% CI 0.25–0.80, P = 0.007), but this was not
Among these, the subgroup predicted to be CLDN18.2-negative and inflamed IP dem-
the case for patients without HRR mutation (PFS: HR 0.79, 95% CI 0.56–1.10, P = 0.156;
onstrated the most significant benefit from ICI-Chemo compared to the Chemo-only group.
Conclusions: The AI model reliably predicted CLDN18.2 expression from H&E slides and
OS: HR 0.90, 95% CI 0.63–1.28, P = 0.554). Conclusions: The presence of HRR mu-
exhibited reliable performance. The differential survival outcomes observed in subgroups tations was associated with favorable survival outcomes in patients treated with
stratified by AI-predicted CLDN18.2 expression and IP suggest its potential utility in nivolumab plus chemotherapy. Our findings suggest that HRR mutations may serve as a
guiding first-line treatment decisions for gastric cancer patients. Research Sponsor: Lunit. potential predictive biomarker for first-line ICI-based chemotherapy in gastric cancer.
Research Sponsor: None.
Hazard ratio (HR) for progression-free survival (PFS) and overall survival (OS).
Groups HR for PFS (95% CI) HR for OS (95% CI)
CLND18.2-negative and inflamed IP 0.37 (0.17–0.78, p=0.009) 0.41 (0.19–0.88, p=0.021)
CLDN18.2-negative and non-inflamed IP 0.75 (0.54–1.05, p=0.095) 0.81 (0.59–1.13, p=0.224)
CLDN18.2-positive and inflamed IP 0.67 (0.40–1.12, p=0.128) 0.62 (0.37–1.04, p=0.073)
CLDN18.2-positive and non-inflamed IP 1.60 (0.74–3.46, p=0.229) 1.23 (0.59–2.58, p=0.579)
reference: Chemo-only.

4049 Poster Session 4050 Poster Session

Clinical outcomes of first-line immune checkpoint inhibitors with chemo- Phase II trial of transarterial chemoembolization followed by sintilimab
therapy in advanced EBV-associated gastric cancer. First Author: Dongwoo Cho, (anti-PD-1), oxaliplatin, and S-1 combined with either trastuzumab (HER-
Division of Medical Oncology, Department of Internal Medicine, Seoul St. Mary’s 2 positive) or apatinib (HER-2 negative) as first-line therapy for gastric
Hospital, College of Medicine, The Catholic University of Korea, Seoul, South Korea cancer with liver metastases. First Author: Wei Song, Department of Minimally
Background: EBV-associated gastric cancer (EBVaGC) is a distinct subgroup of GC with Invasive Treatment of Cancer, Shandong Provincial Hospital Affiliated to Shandong First
high PD-L1/PD-L2 expression and immune cell infiltration. While several rationales support Medical University, Jinan, China
the potential efficacy of immune checkpoint inhibitors (ICIs) in EBVaGC, clinical evidence Background: Liver metastases contribute to immunotherapeutic resistance and unfavorable
from large-scale trials is limited, and their effectiveness remains inconclusive. This ret- outcomes. Transarterial chemoembolization (TACE) may alter the immune microenvironment
rospective study evaluated the clinical outcomes of palliative first-line chemotherapy with or and enhance immunotherapy efficacy. In this trial, we assessed the efficacy and safety of TACE
without the addition of ICIs in patients with EBVaGC. Methods: We identified 217 patients followed by sintilimab (anti-PD-1), oxaliplatin and S-1 combined with either trastuzumab (HER-
diagnosed with EBVaGC via in situ hybridization and evaluated their baseline characteristics, 2 positive) or apatinib (HER-2 negative) in the treatment of gastric cancer with liver metastases
including HER2 expression, MSI status, and PD-L1 expression. Among these, 83 patients (GCLM). Additionally, we explored gene mutations and the immune microenvironment between
with metastatic disease received palliative first-line treatment. Of the 77 HER2-negative gastric cancer and its paired liver metastases. Methods: This single-center, single-arm, phase
patients, 23 were treated with ICIs (nivolumab or pembrolizumab) combined with cytotoxic II trial enrolled 31 treatment-naive patients with GCLM. Patients received TACE for liver
chemotherapy (ICI+chemo), while 54 received chemotherapy alone. Survival outcomes and metastases, followed by sintilimab (200mg), oxaliplatin (130mg/m2) and S-1 (40-60 mg bid for
response rates were compared between ICI+chemo and chemotherapy-alone groups. 14 days) every 3weeks combined with either trastuzumab (HER-2 positive, 8 mg/kg then 6 mg/
Results: Among 217 EBVaGC patients, 8.5% were HER2-positive, 2.6% were MSI-High, and kg, q3w) or apatinib (HER-2 negative, 250mg qd) until disease progression or intolerable
PD-L1 CPS $10 and $5 was observed in 59.1% and 84.1%, respectively. In HER2-negative toxicities. Tissue samples from gastric cancer and liver metastases were collected before
EBVaGC patients, ICI+chemo showed significantly prolonged progression free survival (PFS) treatment for DNA and RNA detection. The primary endpoints were PFS and ORR. The
compared to chemotherapy alone (median: 9.28 vs. 6.07 months; HR = 0.48; p = 0.031). secondary endpoints were DCR, OS, and safety. Results: A total of 31 patients were enrolled,
There was no significant difference in overall survival (OS) (median: 20.72 vs. 18.89 months; median 63 (41-76) years old, 27 male, 26 multiple liver metastases and all adenocarcinoma. Of
HR = 0.78; p = 0.508). The objective response rate (ORR) and disease control rate (DCR) were them, 10 (32.3%) were HER2 IHC 3+, 7 (22.6%) HER2 IHC 2+/FISH+. Patients with PD-L1
significantly higher in the ICI+chemo group than in the chemotherapy-alone group (ORR, CPS$5, , 1 accounted for 19.4%, 51.6% respectively. With the median follow-up time of
78.3% vs. 51.9%; OR = 3.29; p = 0.042; DCR, 100% vs. 79.6%; p = 0.028). In the ICI+chemo 396 days, the patients received a median of 6 treatment cycles. As of January 25, 2025, 27
group, PD-L1 expression (cutoff: CPS 5 or 10) was not significantly associated with survival patients were included in the efficacy and safety analyses. The ORR was 74.1%, with 5 CR, 15
outcomes or response rates. For CPS $5, the median PFS was 7.84 months compared to PR, 3 SD, and 4 PD. DCR was 85.2%. The median PFS was 12 months, and the median OS was
not reached. One- and two-year PFS rates were 51.6 and 27.1 %, respectively. One- and two-
8.99 months in the CPS , 5 group (p = 0.944), the median OS was 20.75 vs. 16.51 months (p
year OS rates were 89.1 and 74.0 %, respectively. Grade 3/4 treatment-related adverse events
= 0.131), and the ORR was 76.5% vs. 75.0%. For CPS $10, the median PFS was 13.9 months
occurred in 18.5% of the patients, notably neutropenia, neurotoxicity and thyroid and liver
compared to 8.69 months in CPS , 10 group (p = 0.580), the median OS was 24.95 months
dysfunctions. Forty-four unique mutated genes were identified in gastric cancer, which were
vs. 20.33 months (p = 0.061), and the ORR was 81.8% vs. 70.0%. Conclusions: To our
involved in PI3K-Akt and drug resistance pathways. In its paired liver metastases, 59 specific
knowledge, this study represents one of the largest cohorts including patients with EBVaGC mutated genes associated with MMR and cell cycle were identified. Liver metastases had more
who received first-line ICI+chemo. The addition of ICIs to chemotherapy showed clinical macrophages and CD8+ T cells, whereas NK and CD4+ memory resting cells were fewer than
benefit, including survival and response rates, compared to chemotherapy alone in EBVaGC. those in paired gastric cancer. Conclusions: TACE followed by sintilimab, oxaliplatin and S-1
Interestingly, the clinical benefit of ICI+chemo was observed in PD-L1 low group as well as combined with either trastuzumab or apatinib demonstrated promising efficacy and man-
PD-L1 high group. Our results highlight the clinical potential of ICI addition to chemotherapy ageable safety as first-line therapy for GCLM. Gastric cancer and its paired liver metastases
in EBVaGC, and PD-L1 expression alone is insufficient as a predictive biomarker, warranting exhibited distinct gene mutations and immune microenvironment. Clinical trial information:
further exploration of alternative predictors for immunotherapy efficacy in this subgroup. ChiCTR2200057726. Research Sponsor: None.
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
284s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4051 Poster Session 4052 Poster Session

Improving evidence-based treatment selection and patient-centered care in Comparison of postoperative adjuvant therapy and surgery alone in path-
upper GI cancers: A Project ECHO initiative. First Author: Elena Elimova, Princess ological N1-2 esophageal squamous cell carcinoma: A prospective multi-
Margaret Cancer Centre, Toronto, ON, Canada center randomized controlled trial. First Author: Yizhou Huang, Fujian Medical
Background: Accurate assessment of biomarkers (including HER2 and PD-L1) is in- University Union Hospital, Fuzhou, Fujian, China
tegral to treatment selection in upper gastrointestinal (GI) cancers. However, challenges Background: Due to the lack of prospective randomized controlled clinical studies,
in integrating biomarker testing and targeted therapies are commonly reported in current guidelines do not recommend postoperative adjuvant therapy, but rather ob-
community settings. The Project Extension for Community Healthcare Outcomes servational follow-up regardless of any stage. The aim of this study was to evaluate the
(Project ECHO) model addresses this gap by connecting experts in tertiary care settings effect of postoperative adjuvant chemoradiotherapy and adjuvant chemotherapy on
with rural healthcare teams. Through case-based teaching, the model builds the pathologic lymph node-positive (pN1-2) esophageal squamous cell carcinoma (ESCC).
knowledge and skills needed to provide evidence-based, equitable care regardless of Methods: Patients with pathologically confirmed pN1-2 ESCC were randomly assigned
geographic location. Methods: In October 2024, 57 healthcare professionals (HCPs) to three groups: surgery alone (SA), postoperative chemotherapy (POCT), or postop-
from 2 US and 4 Canadian community oncology clinics participated in Project ECHO erative chemoradiotherapy (POCRT). The POCT regimen included 2 to 4 cycles of
sessions. Led by an expert oncologist, each session featured interactive discussions of docetaxel and cisplatin (75 mg/m²), while the POCRT regimen comprised 28 fractions of
real-world anonymized case presentations to address key practice gaps in integrating 5040Gy radiotherapy combined with docetaxel and cisplatin. Overall survival (OS) was
biomarker-based therapies and coordinating multidisciplinary care for patients (pts) the primary endpoint. Recruitment was terminated early due to the significant benefit
with upper GI cancers. Following each session, HCPs developed and implemented site- observed in the adjuvant therapy group and difficulties in enrollment. Results: A total of
specific action plans to address gaps in care. Pre-activity and post-activity surveys 145 patients were enrolled (SA: n = 50, POCT: n = 48, POCRT: n = 47), with a median
measured the impact on knowledge, confidence, and competence and 90-day follow-up follow-up of 55 months. The overall survival rates at 3 and 5 years were 76.1% and 56.2%
surveys will be collected to assess ongoing performance. Results: The top HCP- in the POCRT group, compared with 56.6% and 43.4% in the POCT group and 49.1% and
reported barriers to individualized care for pts with upper GI cancer included keep- 25.0% in the SA group, respectively. The POCRT group showed significantly improved OS
ing up with the latest efficacy and safety data (44%), selecting and sequencing and DFS compared to the SA group (OS, P = 0.003; DFS, P = 0.003). The incidence of
treatments based on individual and disease factors (40%), and limited availability/cost of grade 3 or higher adverse events was 39.6% in the POCT group and 44.7% in the POCRT
biomarker testing (39%). Additionally, relatively few HCPs reported providing supportive group (P = 0.615). Conclusions: For patients with pN1-2 ESCC, postoperative therapy,
care services for the majority of their patients, such as palliative care referrals (40%), especially postoperative chemoradiotherapy, significantly improved the prognosis of
distress screening (26%), psychosocial support (35%), and end of life counseling (19%). patients compared with surgery alone. Clinical trial information: NCT04009265.
Following the Project ECHO sessions, HCPs demonstrated improved knowledge, Research Sponsor: Fujian Minimally Invasive Medical Center (Thoracic Surgery), China;
competence, and confidence in biomarker testing and managing adverse events. Ad- Key Laboratory of Cardio-Thoracic Surgery(Fujian Medical University), Fujian Province
ditionally, HCPs planned to increase patient education about disease and treatment- University; Fujian Institute of Cardio-thoracic Surgery, China.
related side effects (65%), improve team education on biomarker testing and treatment
selection (61%), establish standardized biomarker testing protocols (35%), and increase
supportive care referrals (22%). Team action plans included implementing routine
testing protocols, establishing a network of specialists to coordinate care, and in-
creasing education on the use of immunotherapy. Conclusions: Project ECHO-based
education improved HCP capacity to integrate biomarker-directed therapies and co-
ordinate multidisciplinary care for patients with upper GI cancers. Full findings will detail
long-term impacts on practice and inform future community-based ECHO initiatives.
Research Sponsor: Bristol Myers Squibb.

4053 Poster Session 4054 Poster Session

Combined PD-L1 expression and PD-1+ CD8 T cells to predict immunother- Temporal and spatial expression of CLDN18.2 in gastric cancer and gas-
apy outcomes in esophageal squamous cell carcinoma. First Author: Qian Zhao, troesophageal junction cancer. First Author: Qiaoqi Li, West China Hospital,
Department of Radiation Oncology, Shandong Cancer Hospital and Institute, Shandong Chengdu, Sichuan, China
First Medical University and Shandong Academy of Medical Sciences, Jinan, China Background: Recent studies have demonstrated the promising efficacy of CLDN18.2-
Background: Esophageal squamous cell carcinoma (ESCC), the predominant subtype of targeted therapy in patients with gastric cancer (GC) or gastroesophageal junction
esophageal cancer in Asia, remains a challenging disease with poor prognosis due to cancer (GEJC) who are positive for CLDN18.2. This study aims to investigate the
frequent late-stage diagnoses. Immune checkpoint inhibitors (ICIs), particularly those temporal and spatial consistency of CLDN18.2 expression in GC and to provide a
targeting the PD-1/PD-L1 axis, have revolutionized cancer treatment but benefit only a comprehensive overview of its expression patterns. Methods: The expression of
subset of patients. The identification of robust predictive biomarkers is critical to CLDN18.2 in primary GC tumors (biopsy/surgical) and corresponding peritoneal me-
enhance patient selection and optimize immunotherapy outcomes. In this study, we tastases (PM) was evaluated by quantifying cell membrane staining intensity with a
leveraged multiplex immunofluorescence to comprehensively evaluate the roles of PD- validated semi-quantitative assay. CLDN18.2 positivity was defined when tumor cells
L1 and PD-1 expression, immune cell subsets, and clinical factors in predicting im- with staining intensity (2+) and (3+) summed up to $ 75% of all. The Kappa test
munotherapy efficacy in ESCC. Methods: We analyzed baseline tumor samples from evaluated CLDN18.2 expression consistency across sample types, and the McNemar
147 ESCC patients treated with first-line ICIs using multiplex immunofluorescence to test compared its positive rates in paired samples. Results: Between February 2023 and
assess the expression of PD-1, PD-L1, CD4, CD8, CD20, and CD68. Patients were April 2024, 536 patients were enrolled at the Gastric Cancer Center of West China
stratified by biomarker expression levels, with cut-off values determined through ROC Hospital, Sichuan University. The cohort comprised 399 in-situ biopsy samples, 240
curve analysis to calculate AUC. Survival outcomes were analyzed, and multivariate Cox radical gastrectomy samples, and 27 peritoneal biopsy samples. Among them, 109
regression identified independent predictors of progression-free survival (PFS) and patients had biopsy and surgical specimens, and 26 underwent preoperative neo-
overall survival (OS). Results: PD-L1 expression (HR 0.266, 95%CI 0.087-0.816, p = adjuvant therapy. Among 399 biopsy specimens, 131 (32.8%) had positive CLDN18.2. In
0.021) and PD-1+CD8+ T cells (HR 2.694, 95%CI 1.162-6.246, p = 0.021) emerged as 240 surgical specimens, 167 (27.9%) were positive. In 27 peritoneal nodule specimens,
independent predictors of PFS. High PD-L1 expression was associated with superior 10 (37.0%) were positive. No significant differences were found (x2 = 2.128, P = 0.345).
outcomes (mPFS: 7.6 vs. 5.5 months), while elevated PD-1+CD8+ T cell infiltration Among 109 patients with paired biopsy and postoperative pathology, CLDN18.2 ex-
correlated with poorer outcomes (mPFS: 6.0 vs. 7.2 months). Patients with a combi- pression concordance was 80.7% (88/109, Kappa = 0.562). In 27 peritoneal metastasis
nation of high PD-L1 expression and low PD-1+CD8+ T cells demonstrated the best patients, it was 77.8% (21/27, Kappa = 0.503). For 26 chemo-resected patients, the pre-
prognosis, with a median PFS of 8.5 months, whereas those with low PD-L1 expression and post-chemo positive concordance was 80.8% (21/26, Kappa = 0.524). All showed
and high PD-1+CD8+ T cells had the worst prognosis, with a mPFS of 3.5 months. Clinical moderate consistency. CLDN18.2 expression significantly correlated with several
stage (HR 1.570, 95%CI 1.059-2.327, p = 0.025), BMI (HR 0.935, 95%CI 0.883-0.990, p = factors, including gender (P = 0.002), histological subtype (P = 0.003), tumor location (P
0.015), and CD8+ T cell density (HR 0.896, 95%CI 0.824-0.975, p = 0.011) were identified = 0.007), histological classification (P = 0.035), and EBV status (P = 0.006). Higher rates
as independent predictors of OS. Conclusions: Our findings uncover the dual impor- were in females, signet-ring cell carcinoma, non-GEJ tumors, poorly differentiated
tance of PD-L1 expression and PD-1+CD8+ T cell infiltration as critical biomarkers for tumors, and EBV-positive patients. Conclusions: CLDN18.2 is widely present in both
predicting PFS in ESCC patients undergoing ICIs. Moreover, clinical factors such as BMI, primary gastric cancer and peritoneal metastatic lesions. Expression consistency exists
tumor stage, and intratumoral CD8+ T cell density significantly impact OS. Research moderately between biopsy and surgical specimens, and primary and metastatic tissues.
Sponsor: None. Consistency stays strong pre- and post-neoadjuvant therapy. Its expression links to
clinical and molecular traits. This study comprehensively analyzed it, providing a better
basis for CLDN18.2-targeted patient selection. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 285s

4055 Poster Session 4057 Poster Session

Safety and efficacy of endoscopic treatment with mucosal resection, radio- Safety and efficacy of camrelizumab combined with radiotherapy as neo-
frequency ablation, and cryotherapy in the curative treatment of early adjuvant therapy for locally advanced esophageal squamous cell carcinoma:
esophageal squamous cell cancer and dysplasia. First Author: Sidra Naz, The A prospective single-arm phase II clinical trial. First Author: Yizhou Huang, Fujian
University of Texas MD Anderson Cancer Center, Houston, TX Medical University Union Hospital, Fuzhou, Fujian, China
Background: Endoscopic esophageal treatment (EET), by radiofrequency ablation (RFA), cryoablation (Cryo), and endoscopic
mucosal resection (EMR) has not been extensively studied in early esophageal squamous cell carcinoma (SCC). Aim: To assess the
Background: Neoadjuvant chemoradiotherapy followed by esophagectomy is the
safety and success of EET in curative treatment of early esophageal SCC: (T1a, T1b, T2) and squamous cell dysplasia (SCD) standard of care for locally advanced esophageal squamous cell carcinoma (ESCC).
Methods: We retrospectively reviewed records of patients with SCC (T1a, T1b, T2) and SCD treated with EET from January 2010 to However, approximately 30% of patients still develop distant metastases and have a high
August 2024. All patients had at least one follow-up endoscopic biopsy after treatment. Results: 62 patients met the eligibility
criteria. Table. 46 patients (74%) had T1a (n=36.5%) or T1b (n=10.1%) SCC. T2 patients (21%) received ETT for local recurrence after incidence of treatment-related adverse events. Immunotherapy, as a new modality for
chemoradiation. The most frequent ETT was EMR + RFA, n= 26 (41.9%) and RFA, n=20 (32%). In the first biopsy after treatment, 46 of anti-cancer treatment, has shown promising clinical benefits for patients with ESCC. The
the 62 patients (74.1%) had no residual cancer. This number increased after further ETT. 2 years after treatment only 4 patients had
persistent cancer. BMI is significantly associated with survival status. No patient had disease progression that required surgery.
synergistic effects of immunotherapy and radiotherapy make their combination
Strictures formed in 15 patients (24%) post- ETT. Conclusions: ETT provides curative treatment of early SSC in up to 94% of promising as neoadjuvant treatment for locally advanced ESCC. Methods: All partic-
patients. Research Sponsor: None. ipants who meet the inclusion criteria will be enrolled after signing the informed consent
Association between cancer persistence at first post-biopsy with various demographic and clinical variables. form. Patients with thoracic segment esophageal cancer with clinical stage T2-3 N0 M0
Patient Characteristics Total N=62 P-value
or T2-3 N+ M0 will be included. They will be treated with radical surgery within 4–
Sex (%)
Male 28 (45.16%) 0.49 8 weeks after the completion of two cycles of neoadjuvant radiotherapy in combination
Race (%)
Female 34 (54.84%)
with camrelizumab according to the study schedule. The primary endpoint is the major
White / Caucasian
Hispanic
52 (83.87%)
2 (3.23%)
0.716 pathological remission rate of all per-protocol patients. The secondary endpoints are the
Black
Asian
2 (3.23%)
4 (6.45%)
R0 resection rate, pathological complete remission rate, and adverse events. The interim
BMI, median (range)
Smoking Status (%)
N = 62 24.02 (15, 44.5) 0.045 analysis will be conducted after half of the planned number of patients have been
No 25 (40.32%) 0.182 enrolled. The trials will be terminated when more than two treatment-related deaths
Yes 37 (59.68%)
Alcohol Drinking (%) occur or fewer than five patients have major pathological remission. Results: A total of
No 22 (35.48%) .0.99
Yes 40 (64.52%) 25 patients were enrolled, 3 patients did not undergo surgery, of which 1 had imaging CR
Squamous Dysplasia (%)
none 59 (95.16%) .0.99 after neoadjuvant therapy and refused surgical treatment; 1 progressed during neo-
Low 1 (1.61%)
High 2 (3.23%) adjuvant therapy, and the other had immune pneumonitis and renal insufficiency during
Squamous tumor staging (%)
none 3 (4.84%) 0.174 neoadjuvant therapy. The final results of the study noted that of the 22 patients who
T1a
T1b
36
10
(58.06%)
(16.13%)
underwent surgery. Twelve of 22 (54.5%) patients had a pathologic response, all
Squamous Tumor Differentiation (%)
$T2 13 (20.97%) consisting of an MPR with #10% RVT, including 8 of 22 (36.4%) pathologic complete
Well
Moderate
5 (8.2%)
47 (77.05%)
0.707 responses. Conclusions: The NRIT regimen is safe and feasible for patients with ESCC.
Poor 9 (14.75%) Clinical trial information: NCT05176002. Research Sponsor: Key Laboratory of Cardio-
Lymphovascular Invasion (%)
No 57 (91.94%) .0.99 Thoracic Surgery (Fujian Medical University), Fujian Province University; Fujian Mini-
Yes 5 (8.06%)
1b) Patient survival status - Immediate Post-Treatment Pathology Finding mally Invasive Medical Center (Thoracic Surgery), China; Fujian Institute of Cardio-
Patient Characteristics Total N=62
First biopsy post-treatment finding (%) thoracic Surgery, China.
No dysplasia or cancer 37 (59.68%)
Persistent cancer (Failure of t/t) 16 (25.81%)
Persistent Dysplasia 9 (14.52%)
1c) Patient survival status ‡2 years post-treatment, excluding patients with no data or recurrent cancer/dysplasia
Patient Characteristics Total N=21
Biopsy finding (%)
No dysplasia or cancer 15 (71.43%)
Persistent cancer (Failure of t/t) 4 (19.05%)
Persistent Dysplasia 2 (9.52%)

4058 Poster Session 4059 Poster Session

A real-world, propensity-matched analysis of second-line (2L) FOLFIRI-Ram Disitamab vedotin (RC48), tislelizumab, and S-1 as first-line therapy for
versus Ram-Pac in advanced upper gastrointestinal cancers. First Author: HER2-overexpressing advanced gastric or gastroesophageal junction ade-
Jonathan Hyak, UT Southwestern Medical Center, Dallas, TX nocarcinoma (GC/GEJC): Updated results from the RCTS trial. First Author:
Background: Given historically poor outcomes for advanced upper gastrointestinal (UGI) Lian Liu, Qilu Hospital of Shandong University, Jinan, China
cancers, there is an urgent need for effective 2L treatments. Since the phase III RAINBOW Background: Anti-PD1 antibody has significantly improved survival in patients with HER2-
trial, combination ramucirumab and paclitaxel (Ram-Pac) has filled this role; however, this overexpressing and PD-L1 combined positive score (CPS) $1 GC/GEJC when added to
regimen is plagued by dose-limiting toxicities, specifically neuropathy. The phase II trastuzumab and chemotherapy. Recent trials revealed that HER2-targeted antibody-drug
RAMIRIS trial demonstrated the efficacy and tolerability of FOLFIRI-Ram as an alternative conjugates, including RC48 and Trastuzumab Deruxtecan, combined with anti-PD1 antibody,
2L, even if it did not improve survival over Ram-Pac. Real-world data to support its use, have also shown promising efficacy in this population. This study reports updated survival
however, are lacking. Methods: The nationwide Flatiron Health electronic health record- results of RC48 combined with tislelizumab and the oral fluoropyrimidine S-1 as first-line
derived de-identified database, which includes treatment data from around 280 cancer therapy for patients with HER2-overexpressing GC/GEJC. Methods: This single-arm,
clinics across the United States, was queried for patients treated for unresectable or multicenter clinical trial enrolled patients with unresectable or metastatic HER2-
metastatic UGI cancers with 2L Ram-Pac or FOLFIRI-Ram from January 2011-June 2024. overexpressing (IHC 3+ or 2+, regardless of FISH status) first-line GC/GEJC. Patients re-
Demographics and lab values at time of treatment were extracted. Study cohorts were ceived RC48 (2.5 mg/kg), tislelizumab (200 mg), and S-1 (40-60 mg BID for 14 days) every
derived using a greedy match based on a logit model to predict propensity scores from key 3 weeks until disease progression (PD) or intolerable toxicity. The primary endpoint was
clinical and laboratory characteristics; patients were matched 1:6 (FOLFIRI-Ram:Ram- objective response rate (ORR). Secondary endpoints were progression-free survival (PFS),
Pac) given expected imbalances in sample size. The endpoints of interest were overall overall survival (OS) and safety. Results: 57 patients from 9 centers were enrolled, 71.9%
survival (OS) and real-world time to treatment discontinuation (rwTTD), determined via HER2 IHC 3+, 17.5% IHC 2+/FISH+, and 10.5% IHC 2+/FISH-. 14.9% and 36.2% had CPS$5
Kaplan-Meier method, log-rank test, and Cox proportional hazards model. A hybrid ap- and $1, respectively. Median follow-up was 11.8 months. In the intent-to-treat population,
proach was used to construct a multivariate Cox model. Results: Of 15,908 UGI cancer the confirmed ORR (cORR) was 89.4% (51/57, 95% CI: 78.5-96.0%), the median PFS (mPFS)
patients identified, 631 received 2L Ram-Pac and 40 received 2L FOLFIRI-Ram. After was 12.7 months (95% CI: 10.9-NA), and the 18-month OS rate (18m-OSr) was 72.7% (95%
matching, 40 FOLFIRI-Ram and 240 Ram-Pac patients were included. Median OS from CI: 60.0-88.5%). In the per-protocol set (excluding patients with no PD at the first as-
sessment but refused a second evaluation), the cORR was 92.7% (51/55, 95% CI: 82.4-
initiation of 2L therapy was 9.7 months with FOLFIRI-Ram (95% CI 6.9-12.3) and
98.0%), the mPFS was 13.2 months (95% CI: 10.9-NA), and the 18m-OSr was 76.3% (95% CI:
7.7 months with Ram-Pac (95% CI 6.2-8.8), with a hazard ratio (HR) for death of 0.74 with
63.0-91.7%). In the HER2-positive and -negative subgroups, the ORRs were 92.1% (95% CI:
FOLFIRI-Ram versus Ram-Pac (95% CI 0.50-1.11, p = 0.14). Similar results were seen in
81.1-97.8%) and 66.7% (95% CI: 22.3-95.7%), the mPFS was 12.6 months (95% CI: 11.0-NA)
the multivariate model (HR 0.72, 95% CI 0.49-1.08, p = 0.114) after adjustment for al-
and 7.7 months (95% CI: 7.1-NA), and the 18m-OSr was 74.7% (95% CI: 61.3-91.1%) and
bumin, neutrophil:lymphocyte ratio, and alkaline phosphatase. The median rwTTD with
62.5% (95% CI: 32.0-100%), respectively. In the CPS $1 and CPS , 1 subgroups, ORRs were
FOLFIRI-Ram was 5.2 months (95% CI 4.1-6.2), compared to 3.7 months with Ram-Pac 92.3% (95% CI: 74.9-99.1%) and 87.1% (95% CI: 70.2-96.4%), the mPFS was 16.8 months
(95% CI 3.2-4.3). The HR for treatment discontinuation was 0.70 with FOLFIRI-Ram versus (95% CI: 11.3-NA) and 11.4 months (95% CI: 8.5-NA), and the 18m-OSr was 80.4% (95% CI:
Ram-Pac (95% CI 0.48-1.00, p = 0.048). The reduced hazard for treatment discontinuation 62.1-100%) and 67.4% (95% CI: 50.7-89.5%), respectively. The grade 3-4 treatment-related
with FOLFIRI-Ram persisted in the multivariate model (HR 0.67, 95% CI 0.46-0.97, p = adverse events (AEs) was 63.2%. The most common AEs were neutropenia, fatigue, and
0.033) after adjustment for ECOG status, history of prior surgery, PDL1, albumin, and leukopenia. An exploratory study with longitudinal sequencing of circulating tumor DNA is
neutrophil:lymphocyte ratio. Conclusions: In a real-world propensity-score matched ongoing. Conclusions: The combination of RC48, tislelizumab and S-1 as a first-line therapy
analysis, no survival difference was noted with the combination of FOLFIRI-Ram com- shows encouraging response rates and survival benefits in HER2-overexpressing GC/GEJC,
pared to Ram-Pac, however FOLFIRI-Ram was associated with a significantly longer especially in HER2-positive or CPS $1 patients, supporting further evaluation in randomized
rwTTD. Altogether, these data suggest FOLFIRI-Ram is a viable and tolerable alternative controlled trials. Clinical trial information: NCT05586061. Research Sponsor: None.
for 2L treatment of UGI cancers. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
286s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4060 Poster Session 4061 Poster Session

Health-related quality of life (HRQoL) with paclitaxel plus ramucirumab Feasibility of circulating tumor DNA-based minimal residual disease
(PTX-RAM) switch maintenance versus continuation of first-line fluoropyr- (ctDNA-MRD)-guided adjuvant chemotherapy in patients with stage II-III
imidine and oxaliplatin (FOX) chemotherapy (ChT) in patients (pts) with gastric cancer (GC): An adaptive trial (MRD-GATE). First Author: Lian Liu,
advanced HER2-negative gastric or gastroesophageal junction (G/GEJ) Department of Medical Oncology, Qilu Hospital of Shandong University, Jinan, China
cancer: A secondary endpoint of the ARMANI phase 3 randomized trial. Background: Although adjuvant chemotherapy (ACT) is the standard treatment for stage II-
First Author: Eleonora Cristarella, Fondazione IRCCS Istituto Nazionale dei Tumori di III GC, this strategy lacks precision treatment options, and many patients cannot tolerate the
Milano, Milan, Italy adverse events (AEs) of ACT. ctDNA-MRD detection has been shown to predict recurrence
Background: PTX-RAM switch maintenance significantly improved PFS (HR 0.61, 95%CI risk. The aim of this study (NCT06157216) was to evaluate the feasibility of MRD-guided
0.48–0.79; p = 0.0002) and OS (HR 0.75, 95%CI 0.58–0.96; p = 0.025) versus continuation of treatment in these patients. Methods: Patients with stage II-III GC who underwent R0
FOX first-line ChT in the ARMANI trial, with a higher incidence of grade $3 treatment-related resection and D2 gastrectomy were enrolled. Tumor-informed ctDNA-MRD testing was
adverse events and an increased number of hospital visits. Here we present HRQoL results. performed at baseline (28 days after surgery) and subsequently at 3, 6, 9, 12, 18, 24, 30, and
Methods: ARMANI was an Italian multicenter, open-label, randomized, phase 3 trial enrolling 36 months after surgery. ACT was tailored according to MRD status: baseline MRD-negative
pts with HER2-negative G/GEJ cancer who had disease control after a 3-month FOX induction (MRD (-)) patients received de-escalated therapy (observation for stage II and S-1 mon-
ChT, and randomized to switch maintenance with PTX-RAM or the continuation of FOX. otherapy for stage III) and switched to combined ACT (SOX, S-1 plus oxaliplatin, or XELOX,
European Organisation for Research and Treatment of Cancer (EORTC) QLQC30, QLQOG25 and capecitabine plus oxaliplatin) if MRD became positive, while baseline MRD-positive (MRD
EuroQol EQ5D were assessed at randomization and every 8 weeks until progression. HRQoL (+)) patients underwent combined ACT. The primary endpoint was 3-year disease-free
changes over time were described by: (i) mean changes from baseline at each time point, (ii) survival rate (yDFSr). Secondary endpoints included the treatment de-escalation rate, DFS
distribution of improved/stable/worse at 8 weeks and (iii) time to QoL deterioration (TTD), by MRD status, cumulative recurrence risk (CRR), 3-year overall survival rate (yOSr) and
defined as the time from randomization to a worsening $ 10 points of global QoL in EORTC safety. Results: 65 patients were enrolled, with a median age of 60 years (range: 34-83), and
QLQC30. Mean changes were compared by a linear regression model, with baseline values as 83.1% (54/65) were male, 31 patients had stage II and 34 patients had stage III GC. At
covariates. Proportion of improved/stable/worse was compared by Chi square test. Kaplan- baseline, 21.5% patients (14/65) were MRD(+) and received combination ACT. Among the 51
Meier and Cox proportional hazards model were used for TTD estimation. Results: Of the 280 baseline MRD(-) patients, 45 received de-escalated therapy at onset (9 received combination
pts randomized, 198 (71%; 109/144 with PTX-RAM and 89/136 with FOX) and 133 (48%; 81/ ACT after MRD conversion) and 6 received combination ACT. The median follow-up time was
144 and 52/136) completed baseline and 8-weeks assessment of EORTC and EQ5D, re- 13.3 (range: 9.3-15.7) months. In the intention-to-treat population, the overall 1-yDFSr was
spectively. Mean baseline scores of global HRQoL were 66.90 (standard deviation [SD] 20.71) 86.2% (90% CI: 79.4-93.5%), the 1-yOSr was 96.9% (90% CI: 93.5-100%) and the CRR was
with PTX-RAM and 70.97 (SD 19.39) with FOX. Global QoL at 8-weeks assessment was better 13.8% (95% CI: 6.5-24.7%). The treatment de-escalation rate was 69.2% (45/65, 95% CI:
with PTX-RAM versus FOX both in terms of mean changes from baseline (+2.17 vs -8.51, delta 56.6-80.1%). Grade 3-4 AEs occurred in 24.6% (95% CI: 14.8%-36.9%) of patients. Baseline
10.68, p = 0.015) and in terms of proportion of improved/stable/worse (improved 24.7% vs MRD (+) patients had a shorter DFS compared to MRD (-) ones (1-yDFSr: 57.1% vs. 94.1%,
4.2%, stable 56.2% vs 64.6%, worse 19.2% vs 31.3%, p = 0.009). TTD was significantly longer for
HR = 9.66, 95% CI: 2.40-38.81, log-rank P , 0.0001). Patients with sustained MRD (-) had the
PTX-RAM versus FOX (median TTD 7.6 vs 3.8 months, HR 0.52, 95%CI 0.33-0.82; p = 0.005).
best DFS (1-yDFSr: 100%), while those with sustained MRD (+) had the shortest DFS.
Mean changes from baseline after 8 weeks for functional scales and symptoms of QLQC30 and
Patients with MRD conversion from positive to negative or from negative to positive had
QLQOG25 showed significant improvement for PTX-RAM vs FOX for role functioning (p =
intermediate DFS (1-yDFSrs: 72.7% and 70.0%, respectively). In 9 patients with recurrence,
0.006), nausea/vomiting (p = 0.002), pain (p = 0.016), appetite loss (p = 0.03) and dysphagia (p
ctDNA-MRD positivity identified recurrence a median of 3.4 months earlier than radiology.
= 0.028); hair loss was worse with PTX-RAM (p = 0.024). VAS score from EQ5D was not
significantly different between the two treatments for all the assessments. Conclusions: In
Conclusions: MRD-guided ACT for stage II-III GC significantly reduced the ACT rate, and
pts with HER2-negative advanced G/GEJ cancer, PTX-RAM switch maintenance, beyond a increased the de-escalated CT rate, resulting in good disease-free survival and fewer side
significant benefit in PFS and OS, showed significant benefit in terms of HRQoL, reducing effects. The results of this MRD-guided precision of ACT deserve to be confirmed by large
symptoms and delaying global QoL deterioration. Clinical trial information: NCT02934464. randomized clinical trials. Clinical trial information: NCT05585580. Research Sponsor:
Research Sponsor: None. None.

4062 Poster Session 4063 Poster Session

Development and validation of the gastric risk immuno-progression score Multicenter phase I/II study of abemaciclib and ramucirumab in metastatic
(GRIPS) in advanced gastric and gastroesophageal junction adenocarci- gastroesophageal adenocarcinoma (GEA): CDK4/6 and cyclin D1 alterations
noma treated with first-line chemotherapy plus nivolumab: Results from the as a predictor of response and survival. First Author: Ronan Joseph Kelly, Baylor
ORACLE study. First Author: Dario Spanu, Medical Oncology Unit, University Hospital University Medical Center, Dallas, TX
and University of Cagliari, Cagliari, Italy Background: CDK4/6 and Cyclin D1 are highly expressed in GEA cancers, suggesting that
Background: The addition of nivolumab to chemotherapy is approved for advanced gastric CDK4/6 inhibition may be a promising strategy. In vitro and in vivo studies have shown that
and gastroesophageal junction (GEJ) adenocarcinoma with PD-L1 CPS $5. Identifying abemaciclib (A) demonstrates potent antitumor efficacy in GEA by directly inhibiting this
predictive tools for disease progression in this setting remains a clinical need. This study pathway. Currently, ramucirumab (RAM) 6 paclitaxel is an approved 2nd line treatment for
introduces the Gastric Risk Immuno-Progression Score (GRIPS), a novel score derived from metastatic GEA cancers. Methods: This multicenter, open-label, phase I/II study investi-
variables associated with progression-free survival (PFS) in a real-world cohort. gated the safety and efficacy of A combined with RAM in pretreated advanced GEA (2nd or
Methods: We conducted a multicenter retrospective study on 90 patients with gastric/GEJ 3rd line). The primary objective was to describe the safety profile of A (150mg po bid) and
adenocarcinoma treated with first-line chemotherapy plus nivolumab between 2022 and RAM (8mg/kg iv every 2 weeks) using CTCAE version 4.03. Secondary objectives included
2024. The GRIPScore was constructed using five dichotomous variables associated with assessing the objective response rate (ORR), disease control rate (DCR), median
unfavorable PFS in univariate analysis: Neutrophil-to-Lymphocyte Ratio (NLR . 2.68), ECOG progression-free survival (mPFS), and median overall survival (mOS). Correlative studies to
Performance Status ($2), Smoking status (never smoker), CA19.9 at diagnosis (UNL) and evaluate alterations in CDK4/6 and Cyclin D1 as determined by next generation sequencing
primary tumor resection (no resection). Each variable was scored as 1 (prognostically as predictive biomarkers of efficacy were performed. Results: From July 2021 to December
negative) or 0 (prognostically positive), and a total score (0–5) was calculated. Patients were 2024, 20/30 patients were enrolled. The study was terminated prematurely due to slow
categorized into low-risk (GRIPS 0–2) and high-risk (GRIPS 3–5) groups. PFS and OS were accrual. The median age was 61.5 years (Range: 30.0, 80.0) and most patients were male
analyzed using Kaplan-Meier methods. Hazard Ratios (HRs) were calculated, and the dis- (18/20). Seven patients (35%) were HER-2 positive, 11/18 patients (61.1%) were PDL1
criminatory ability of the GRIPS was evaluated using Harrell’s C-index. Results: The median CPS . 1 and 15 patients (75%) had cancer localized in the E. Baseline ECOG performance
PFS for the entire cohort was 11.55 months (95% CI: 7.60–13.26). Stratifying patients by status was 0 in 8 patients (40%) and 55% of patients had received prior immunotherapy with
GRIPS revealed a marked difference in outcomes between the low-risk and high-risk groups. 1st line chemotherapy. A combined with RAM was generally well-tolerated without unex-
Low-risk patients (GRIPS 0–2) had a median PFS of 13.16 months (95% CI: 9.01–18.91), while pected toxicities. The most common treatment-related adverse events (AEs) were anemia
high-risk patients (GRIPS 3–5) had 4.11 months (95% CI: 2.37–11.77). This difference was (10%), hypertension (10%), and dysphagia (10%). Treatment-related AEs $ grade 3 occurred
significant (log-rank test, p = 0.0023; HR: 3.55, 95% CI: 1.58–8.02). Similarly, low-risk patients in 50% of the patients. Median PFS and mOS were 2.7 months (95% CI: 1.5 - 14.5) and not
had a median OS of 21.25 months (95% CI: 13.77–21.25) compared to 11.90 months (95% CI: reached (NR) (95% CI: 3.4 - NR), respectively. ORR was 10% (2/20) and DCR was 40% (8/20).
4.03–16.10) for high-risk patients (log-rank test, p = 0.0313; HR: 2.73, 95% CI: 1.09–6.81). In evaluable patients, 64.7% (11/17) patients with baseline tissue CDK4/6 pathway alter-
Harrell’s C-index for PFS was 0.648 (95% CI: 0.552–0.743), indicating moderate discrimi- ations trended towards longer mPFS (3.4 vs. 1.3 months; HR:1.1) and mOS (NR vs.
natory ability. Conclusions: The GRIPS effectively stratifies patients with advanced gastric 5.2 months; HR: 1.4) compared to patients without alterations (p . 0.05). Notably, one
and gastroesophageal junction adenocarcinoma treated with a combination of chemotherapy study patient with a CDK6 amplification had a partial response of 64% and has been on
+ nivolumab into distinct risk groups for PFS and OS. High-risk patients experience sig- treatment for . 24 months. Conclusions: A plus RAM demonstrated promising antitumor
nificantly shorter survival, highlighting the potential clinical utility of GRIPS for personalized activity in previously treated E/GEJ adenocarcinomas in the 2nd and 3rd line metastatic
treatment strategies. Further prospective validation is warranted to refine its application in setting with manageable toxicities. Alterations in the CDK4/6 and Cyclin D1 pathways
clinical practice. Research Sponsor: None. appear to enrich for efficacy and may be predictive but need future validation. In-depth
GRIPScore N° of patients (%) mOS (m) mPFS (m) molecular studies investigating changes in the expression of selected serum/tissue ge-
nomic markers of response for the cytostatic regimen will be presented at the meeting.
Low-risk (0-2) 51 (45.9) 21.25 (95% CI: 13.77–21.25) 13.16 (95% CI: 9.01–18.91)
High-risk (3-5) 21 (18.9) 11.90 (95% CI: 4.03–16.10) 4.11 (95% CI: 2.37–11.77) Clinical trial information: NCT04921904. Research Sponsor: None.
Not evaluable 18 (16.2)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 287s

4066 Poster Session 4067 Poster Session

Sequential chemo-immunotherapy as a novel bridging strategy for non- Tumor-informed liquid biopsy in predicting recurrence in patients with
complete responders after neoadjuvant chemoradiotherapy in esophageal operable gastroesophageal adenocarcinoma: The LIQUID study. First Au-
cancer: First prospective phase 2 trial challenging the immediate-surgery thor: Michele Prisciandaro, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy
paradigm. First Author: Lin Peng, Department of Thoracic Surgery, Sichuan Cancer Background: Gastroesophageal cancers (GEC) shows high recurrence rates after cu-
Hospital & Institute, Sichuan Cancer Center, School of Medicine, University of Electronic rative surgery with or without perioperative/adjuvant chemotherapy. Pathological stage
Science and Technology of China, Chengdu, Sichuan, China is currently the only prognostic tool, but its accuracy should be improved. Liquid biopsy,
Background: Locally advanced esophageal squamous cell carcinoma (LA-ESCC) patients with analyzing circulating tumor DNA (ctDNA), offers a non-invasive method to detect
non-clinical complete response (non-CCR) after neoadjuvant chemoradiotherapy (NCRT) molecular residual disease (MRD) and monitor disease status. The LIQUID study aimed
face .50% recurrence risk with direct surgery (DS), yet no standardized bridging strategy to evaluate the prognostic utility of liquid biopsy in GC patients, for whom observational
exists. This first study evaluates the efficacy and safety of sequential chemo-immunotherapy studies are still scarce. Methods: This single-institutional study enrolled patients with
(SCI) as a bridge to surgery in this population. Methods: In this phase 2 cohort study resectable GC treated or not with perioperative/adjuvant chemotherapy. Liquid biopsies
(NCT05189730), 169 LA-ESCC pts who underwent NCRT were prospectively enrolled from June were collected before neoadjuvant chemotherapy (optional), pre- and post-surgery
2021 to January 2025. The NCRT regimen included paclitaxel and carboplatin every 3 weeks for
(minimum 2 weeks interval) and every six months during follow-up, until disease
two cycles. Concurrent radiotherapy (40–41.4 Gy) was administered. Post-NCRT, patients
progression or the last follow-up for alive and progression-free patients. MRD analysis
were assessed for non-CCR and stratified into two groups: the SCI group received two ad-
ditional cycles of chemotherapy and tislelizumab (200 mg intravenously every 3 weeks) before
was performed using a clinical trial assay based on the tumor-informed ctDNA assay
surgery, while the DS group proceeded to surgery. The primary endpoint was pCR rate, FoundationOne®Tracker, which identified somatic mutations by tissue comprehensive
secondary endpoints included major pathological response (MPR) rates and safety. genomic profiling and tracked them in patients’ plasma samples. Results: Between
Results: Eighty-seven non-CCR pts were included (SCI: n = 54; DS: n = 33). The median age December 2019 and February 2024, 119 patients were enrolled. Of these, 40 were
was 63 years, with 78.0% male patients. Most patients were stage IIIB(83.9%). Surgery rates excluded due to screen failure (mostly peritoneal disease at surgery), and 27 due to
were 85.2% in the SCI group (46/54) and 81.8% in the DS group (27/33). In the ITT population, technical failure (n = 18) or loss to follow-up (n = 9), leaving 52 patients (median 4
SCI significantly improved pCR rates (40.7% [22/54] vs. 18.1% [6/33]; OR: 3.06, p = 0.024, , one- timepoints per patient) for analysis. Median age was 70 years (range: 23–86), with 55.8%
sided Fisher’s Exact Test) and showed a trend toward higher MPR rates (51.8% [28/54] vs. male, and 86.5% of tumors located in the stomach. Pathological staging revealed pT0
33.3% [11/33]; OR: 2.14, p = 0.071). In the PP population, pCR rates remained higher in SCI (3.8%), pT1 (21.2%), pT2 (21.2%), pT3 (34.6%), and pT4 (19.2%) tumors, with nodal
(47.8% [22/46] vs. 22.2% [6/27]; OR: 3.16, p = 0.026) and showed higher MPR rates (60.9% [28/ involvement in 69.6%. Additionally, 51.9% and 46.2% of patients received neoadjuvant or
46] vs. 40.7% [11/27]; OR: 2.02, p = 0.078) . At 12 months, PFS rates were 95.6% in the SCI adjuvant treatment respectively, and 32.7% experienced peritoneal relapse. ctDNA
group versus 77.3% in the DS group (p = 0.094) in the ITT population, and 97.4% versus 77.8% presence was not significantly associated with known clinico-pathological baseline risk
(p = 0.064) in the PP population. SCI-related adverse events included lymphopenia (97.7%), factors, except for postoperative N-stage (p = 0.04). With a median follow-up of
leukopenia (84.6%), and fatigue (50.0%). In the no-surgical pts in SCI group, three cases 49.0 months (IQR 30.6 -54.0), post-surgery MRD+ patients had a significantly worse
experienced immune pneumonitis and thyroid dysfunction, respectively. Treatment-related relapse-free survival (RFS) than those with ctDNA- (13.2 months vs not reached; HR 2.81
adverse events in the SCI group included lymphopenia (97.7%), leukopenia (84.6%), and fatigue
95% CI 1.23-6.45; p = 0.011). Similar results were observed in longitudinal ctDNA
(50.0%).The main postoperative complications in the SCI group and DS group were anas-
tomotic leakage and recurrent laryngeal nerve injury (3/46 vs. 2/27, P = 0.629). No significant
monitoring, (RFS 16.3 months for ctDNA+ vs not reached for ctDNA-; HR 2.70, 95% CI
treatment-related adverse events occurred in the SCI group. Conclusions: This pioneering 1.22–5.97, p = 0.011). Notably, absence of ctDNA after completing the treatment plan
study demonstrates that SCI as a bridging strategy significantly improves pCR rates by .2- (post-surgery or adjuvant therapy) or clearance/seroreversion after treatment were
fold (OR.3) and shows promising PFS trends with manageable toxicity in non-CCR LA-ESCC, associated with significantly longer RFS (p = 0.001 and p = 0.036, respectively).
challenging the immediate surgery paradigm. These results warrant validation in randomized Conclusions: Post-surgical landmark and longitudinal ctDNA detection demonstrates
phase 3 trials to redefine standard-of-care. Clinical trial information: NCT05189730. Research robust evidence of MRD and identifies GC patients at high risk of relapse. These findings
Sponsor: None. support ctDNA as a valuable tool for postoperative surveillance and early intervention
strategies in GEC. Research Sponsor: None.

4068 Poster Session 4069 Poster Session

Recurrence-free survival as a surrogate endpoint for overall survival in Neoadjuvant toripalimab plus CapeOX in patients with locally advanced
resectable esophageal cancer: An individual patient data analysis of phase EBV-positive gastric or esophagogastric junction adenocarcinoma (GC/
III RCTs. First Author: Jun Okui, Department of Surgery, Keio University School of EGJC): Results from the phase II NICE trial. First Author: Liying Zhao,
Medicine, Tokyo, Japan Department of General Surgery & Guangdong Provincial Key Laboratory of Precision
Background: Overall survival (OS) is regarded as the gold standard efficacy endpoint Medicine for Gastrointestinal Tumor, Nanfang Hospital, Southern Medical University,
but requires long follow-up. This study aimed to determine the validity of recurrence-free Guangzhou, Guangdong, China
survival (RFS) as a surrogate endpoint for OS in resectable esophageal cancer. Background: Surgery remains the cornerstone of curative therapy for locally advanced GC/
Methods: A systematic review of phase III randomized controlled trials (RCTs) com- EGJC. EBV-positive tumor is a distinct molecular subtype that would be potentially sensitive
paring perioperative treatments for resectable advanced esophageal and gastro- to immunotherapy, but no consistent reports. Given that chemotherapy may enhance
esophageal junction cancer was conducted. Individual patient data (IPD) were requested antitumor immunity, perioperative immunochemotherapy may be a promising modality for
from all included trials. Surrogacy between RFS and OS was assessed at the individual EBV-positive patients. Methods: The NICE trial is a multicenter, multi-cohort phase II study
level using the Kendall rank correlation coefficient (t) and at the trial level using the (NCT04744649) evaluating the safety and efficacy of toripalimab plus CapeOX as peri-
coefficient of determination (R²) from a meta-regression model. A t of 0.8 and an R² of operative treatment in patients with locally advanced GC/EGJC. The Cohort B was first of its
0.65 were considered thresholds indicative of a good surrogate endpoint. kind to assess the efficacy of the immunochemotherapy on the EBV-positive GC/EGJC , in
Results: Twenty-two eligible trials were identified by the systematic review, and IPD which patients received toripalimab (240 mg) combined with standard-dose CapeOX every
were available from 10 RCTs (JCOG1109, JCOG9907, JCOG9204, FFCD9901, FFCD9102, 3 weeks for 4 cycles preoperatively and 4 cycles postoperatively. Eligibility criteria included
SAKK75/08, CROSS, KOK, NeoRes2 and CMISG1701), including 2,145 patients who clinical tumor stages of cT3-4aNxM0 or cT2N+M0 disease as determined by both imaging
underwent R0 resection (cStage IV, cT1N0 and cT4b excluded). Of these, 1563 patients scan and staging laparoscopy with negative peritoneal cytology. The primary endpoint was
major pathologic response (MPR, defined as , 10% viable tumor cells). The tumor immune
had squamous cell carcinoma, and 575 patients had adenocarcinoma. The 5-year OS and
microenvironment (TIME) of tissue samples obtained before and after treatment was
RFS rates were 53.2% and 46.2%, respectively, with a median OS of 6.2 years and a
analyzed using multiple immunofluorescence assays to assess changes in immune cell
median RFS of 3.6 years. For individual-level surrogacy, Kendall’s t was 0.823 (95% CI: infiltration and other biomarkers related to treatment response. Results: From May 2021 to
0.807–0.839). Subgroup analysis based on treatment modality revealed t values of September 2023, 17 patients with EBV-positive GC/EGJC (GC, n = 15; EGJC, n = 2) were
0.830 (95% CI: 0.800–0.861) for patients receiving neoadjuvant chemotherapy (NAC; n = enrolled, with cT2N0 (n = 1), cT3N1-3 (n = 5), and cT4aN1-3 (n = 11). All patients completed 4
586), 0.827 (95% CI: 0.803–0.850) for those receiving neoadjuvant chemoradiotherapy preoperative cycles of treatment, and none experienced progression before surgery. Only
(NACRT; n = 982), 0.770 (95% CI: 0.713–0.828) for the surgery-alone group (n = 320), and one patient withdrew the inform content after preoperative therapy, the 16 patients un-
0.861 (95% CI: 0.824–0.898) for the adjuvant chemotherapy group (n = 257). Trial-level derwent radical resection, achieving a 100% R0 resection rate (16/16). The MPR rate was
surrogacy analysis across all 22 trials demonstrated an R2 of 0.735 (95% CI: 37.5% (6/16), and pathological complete response rate (pCR) was 25.0% (4/16). Of the
0.512–0.939). The surrogate threshold effect was 0.929, indicating the minimum RFS 16 participants, 15 received postoperative adjuvant therapy, while 1 declined further
treatment effect required to predict a nonzero effect on OS. Conclusions: This study treatment. The TIME analysis results showed that tumor-infiltrating CD8+ T cells in post-
demonstrated strong individual-level and trial-level surrogacy between RFS and OS in treatment tumor tissues significantly clonally expanded compared with pre-treatment
surgically resectable esophageal cancer across all perioperative treatment modalities. paired tissues. Treatment-related grade 3/4 adverse events were observed in 6 patients
These findings hold promise for expediting the development of novel perioperative (35.3%, 6/17). Until Dec 31 2024, none of the patients experienced disease recurrence.
treatment by shortening the follow-up of clinical trials on esophageal cancer. Research Conclusions: Neoadjuvant toripalimab combined with CapeOX is a safe and effective
Sponsor: None. treatment option for patients with EBV-positive, locally advanced GC/EGJC, with moderate
MPR and pCR, indicating further investigating for this distinct type of cancer. Clinical trial
information: NCT04744649. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
288s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4070 Poster Session 4071 Poster Session

Early detection and neoadjuvant efficacy prediction for esophageal cancer Effectiveness of a multidomain mHealth-based intervention in enhancing
using cfDNA methylation-based liquid-biopsy assay. First Author: Zhigang Li, recovery and quality of life for esophageal cancer patients undergoing
Shanghai Chest Hospital, Shanghai Jiaotong University, Shanghai, China esophagectomy. First Author: Xiaodong Su, Department of Thoracic Oncology,
Background: Esophageal cancer (EC) is a major malignancy of the upper gastroin- Sun Yat-Sen University Cancer Center, Guangzhou, China
testinal tract globally. Early detection and timely therapeutic intervention are pivotal in Background: Esophagectomy, a primary treatment for esophageal cancer (EC), often
improving patient outcomes. However, current diagnostic methods often fail to detect compromises patients’ quality of life (QOL), leading to malnutrition, reduced physical
EC at an early stage and lack the ability to predict treatment response. There is an urgent function, and psychological distress. Multidomain mHealth-based interventions, which
need for non-invasive, sensitive, and specific biomarkers to enhance early detection and leverage technology to deliver comprehensive nutritional, physical, and psychological
guide personalized treatment strategies. This study aims to develop a tool to detect EC support, offer a promising approach to address these challenges during prehabilitation and
and predict the neoadjuvant efficacy. Methods: This is a prospective multicenter di- Enhanced Recovery After Surgery (ERAS) phases. Despite their potential, high-quality
agnostic study. From July 2023 to October 2024, a total of 236 esophageal cancer cases evidence demonstrating their effectiveness remains limited. This study evaluates the ef-
(stage I: 19.9%, stage II: 24.6%, stage III: 39.4%, stage IV: 16.1%), 31 chronic esophagitis fectiveness of a multidomain mHealth intervention on QOL and recovery outcomes in EC
cases, and 441 healthy controls were enrolled from multiple centers. Methylation patients undergoing esophagectomy. Methods: Between April 27, 2021 and June 30, 2023,
features and fragmentomic characteristics derived from methylation sequencing data this single-center, randomized controlled trial was conducted, enrolling 76 patients with
were integrated to develop a gradient-boosted tree model. A nested cross-validation pathologically confirmed EC scheduled for esophagectomy. Participants were randomized
to either a multidomain mHealth-based intervention group (n = 38) or a usual care group (n =
framework was employed to ensure robustness and reliability. Additionally, predictive
38). The intervention delivered tailored nutritional, physical, and psychological support via a
models for therapeutic responses to neoadjuvant treatment were constructed.
self-developed WeChat-based management platform, spanning from 2 weeks pre-admission
Results: The detection model achieved an area under the curve (AUC) of 0.954（95%
to 11 weeks post-discharge. The primary outcome was the change in QOL, assessed using
CI: 0.936-0.971. At a specificity of 97.9% (95% CI: 96.1%-99.0%), the overall sensitivity the EORTC QLQ-C30 and QLQ-OES-18. Secondary outcomes included changes in nutritional
reached 84.7% (95% CI: 79.5%-89.1%) , with stage-specific sensitivities of 69.5% for status (e.g., weight), physical fitness (e.g., 6-minute walk distance), and psychological
early-stage (I/II) and 97% for advanced-stage (III/IV) disease. The detection model health (e.g., PHQ-9, GAD-7, SCSQ-20). Outcomes were assessed at baseline, hospital ad-
maintained robust performance across various clinicopathological parameters, in- mission, and 3 and 11 weeks post-discharge. Results: At 11 weeks post-discharge, the
cluding differentiation grade, neural invasion, vascular invasion, tumor count, and tumor intervention group showed a significant improvement in QOL, with a 15.56-point increase in
location, with no significant differences in subgroup performance. Among the cohort, 44 Global Health Status (EORTC QLQ-C30) compared to a 5.91-point decline in the usual care
patients underwent neoadjuvant therapy, with 90.9% (40/44) receiving immunoche- group (difference: 21.47 points; 95% CI: 9.86–33.08; p , 0.001). Functional and symptom
motherapy. The major pathological response (MPR) rate was 52.3% (23/44) and the scores, including social functioning, appetite, and eating, improved markedly in the in-
pathological complete response (pCR) rate was 15.9% (7/44). No clinical features were tervention group. The intervention group also demonstrated superior outcomes in weight
found to correlate with MPR or pCR rates. Differential methylation profiles between change (+2.19 kg; 95% CI: 0.09–4.30; p = 0.041) and 6-minute walk distance (+80.65 meters;
MPR/pCR and non-MPR/pCR patients were analyzed to construct predictive models for 95% CI: 41.64–119.66; p , 0.001). Psychological well-being improved significantly, with
neoadjuvant therapy outcomes. Using logistic regression and leave-one-out cross- reductions in PHQ-9 (-4.70; 95% CI: -6.59 to -2.80; p , 0.001) and GAD-7 (-4.63; 95% CI:
validation, the MPR prediction model achieved an accuracy of 86.3%, while the pCR -6.52 to -2.74; p , 0.001), and an increase in positive coping scores (+11.83; 95% CI:
prediction model demonstrated an accuracy of 90.9%. Conclusions: Our cfDNA- 7.20–16.45; p , 0.001). Conclusions: This multidomain mHealth intervention significantly
methylation based assay demonstrated high performance in early EC detection and enhanced QOL, nutritional and physical outcomes, and psychological health in EC patients
promising value in predicting neoadjuvant therapy responses. This non-invasive ap- undergoing esophagectomy. These findings underscore the potential of mHealth platforms
proach has the potential to revolutionize EC management by enabling earlier diagnosis to optimize prehabilitation and recovery, offering a scalable and impactful approach to
and personalized treatment strategies. Research Sponsor: None. improving outcomes in oncology care. Clinical trial information: ChiCTR2100045650.
Research Sponsor: None.

4073 Poster Session 4074 Poster Session

Exome analysis of over 5000 esophagogastric cancers. First Author: Reetu The lost evidence: A phase III, multicenter randomized controlled trial of
Mukherji, Ruesch Center for the Cure of Gastrointestinal Cancers, Lombardi Compre- neoadjuvant chemotherapy paclitaxel plus cisplatin versus surgery alone for
hensive Cancer Center, Georgetown University Medical Center, Washington, DC stage IIA–IIIB esophageal squamous cell carcinoma. First Author: Yin Li,
Background: Esophagogastric cancers (EGCs) encompass a heterogenous group of cancers. The genomic Department of Thoracic Surgery, National Cancer Center/National Clinical Research
drivers that overlap or differentiate among each cancer type are not well studied, despite the availability of Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking
therapies that target specific genetic alterations in driver genes. Methods: Genomic data from Natera’s Union Medical College, Beijing, China
Real-World Database (N = 5,872) was analyzed to investigate genomic patterns in EGC patients (09/2019-11/
2024) receiving standard-of-care treatment. This exploratory analysis was conducted on tumor tissue data Background: The efficacy of neoadjuvant chemotherapy (NAC) vs primary surgery alone for
available from whole-exome sequencing, generated as part of SignateraTM [Link] instability treatment of locally advanced esophageal squamous cell carcinoma (ESCC) remain controversial
(MSI) status was determined using the MSIsensor and 96-trinucleotide contexts and was correlated with between Western and Eastern countries. The daily NAC practice in China was without the high
COSMIC SBS signatures (v3.3). Prevalence analysis included only non-synonymous mutations, with ranking level evidence. The Esophageal Cancer Committee of the China Anti-Cancer Association tried to
adjusted for gene length. Results: Patients included 4050 men and 1822 women, with median age 65.3 connect the the lost chain of evidence. To compare safety and long-term survival of NAC followed
years, and stage distribution as follows: I: 7.9%, II:14.7%, III: 32.9%, IV: 32.2%, and 12.3% unknown. Gastric by surgery with that of surgery alone. Methods: A prospective, multicenter, open-label, ran-
cancer (GC) was most common (48.1%), followed by esophageal (EC, 45.9%) and gastroesophageal junction domized phase III clinical trial that compared safety and efficacy of NAC vs primary surgery for
(GEJ, 7.2%). MSI-high cases (7.7% prevalence overall, 2.2% in squamous EC, 5.2% in EC adenocarcinoma, ESCC. From July 18, 2015, to March 29, 2018, we enrolled 605 clinical stage IB-III thoracic ESCC
6.1% in GEJ, 10.9% in GC) had a distinct mutational landscape with frequent missense deletions. The most
(excluding stage T4b, N3, 7th UICC-TNM staging, 2009). They were randomized to NAC plus
common signatures were clock-like (SBS1, SBS5), MMR-deficiency-related (SBS6, SBS15), and Thiopurine-
chemotherapy-related (SBS87). PIK3CA mutations were found in 7.9% of cases, with the most common being surgery (group NAC; n=304) or primary surgery alone (group S; n=301). In group NAC, paclitaxel
E545 (2.6%), H1047 (1.1%), and E542 (1.0%). Notably, PIK3CA exon 9/20 mutations displayed a trend of 175 mg/m2 intravenously (IV) and cisplatin 75 mg/m2 every 3 weeks for two cycles. All patients
higher prevalence in ctDNA-positive cases. Conclusions: These data provide insights into the mutational underwent McKeown, Ivor Lewis or minimally invasive esophagectomy and extended 2-field
landscape of EGC and enhance our understanding of differences between histological subtypes. Future lymph node dissection. The primary outcome was 5-year overall survival (OS). Secondary
studies will continue to explore the associations between genomic subtypes, treatment patterns, and clinical outcomes included disease free survival (DFS), R0 resection rate, pathologic complete response
outcomes. Research Sponsor: None. rate and toxicities. The intention-to-treat principle was followed for analysis. The SPSS, version
Top mutated genes and variants in esophagogastric cancers. 23.0 (IBM Corp). The statistically significant was assumed as a 2-sided P,0.05. The Kaplan-Meier
Group N Genes Variants method was used to calculate OS and DFS with the log-rank test. The last follow-up data was April
12, 2024. The 5-year OS of primary surgery was 30%. A 5-year survival with a 12% increase for the
EGC, all 5872 TP53 (49.8%) ACVR2A K437X (5.5%)
ARID1A(16.1%) RPL22K15X (5.3%) NAC group was assumed. The sample size was calculated with a two-sided alpha level of 5%, a
RNF43G659X (4.0%) power of 80%, an expectation of 2 years accruement and a 5-year followup period. The total
EGC, MSS 5411 TP53 (47.9%) G2E3 T361fs (3.1%) sample size was set at 528 patients with 10% of patients lost to follow-up. Results: Among 605
ARID1A(11.9%) TP53R175H (2.7%)
LRRIQ3Q245fs (2.3%) patients (432 men [71.4%]; mean [SD] age, 61.7 [7.9] years; most frequent clinical stages IIIA 170
EGC, MSI-high 422 ARID1A (71.7%) ACVR2A K437X (58.6%) [30.2%]), Leukopenia (28%) and neutropenia (49.8%) were the most common grade 3 or 4 adverse
KMT2D(68.7%) RPL22K15X (56.6%)
RPL22(60.3%) RNF43G659X (44.1%)
events during NAC. Complications was similar between the 2 groups based on Clavien-Dindo
GEJ, MSS 448 TP53 (50.5%) TMBIM4 Y174fs (3.2%) classification. The 90-day perioperative mortality rate was 1.6% for the NAC group (4 of 244) and
CSMD1(13.4%) TP53R273C (3.2%) 2.2% for the Surgery alone group (6 of 268) (P = 0.754). The pathologic complete response rate
PCLO(12.0%) TP53R175H (3.2%)
GC, MSS 2487 TP53 (33.5%) G2E3 T361fs (3.1%) was 6.58% (20 of 304) in NAC group. NAC group had a higher R0 resection rate (98.8% v 98.5%;
CDH1(14.4%) PIK3CAE545K (2.3%) P . 0.999), a better 5 years OS rate (61.1% vs 51.6%; hazard ratio, 0.79; 95% CI, 0.63 to 1.0; P =
ARID1A(13.8%) LRRIQ3Q245fs (2.2%) 0.0469), and a prolonged DFS (58.5% vs 46.2% months; hazard ratio, 0.71; 95% CI, 0.56 to 0.91;
EC Adenocarcinoma, MSS 1599 TP53 (33.5%) TP53 R175H (4.0%)
CDKN2A(13.5%) TP53R248Q (3.4%) P=0.0067). 0.71 (0.56,0.91). The max follow up period of NAC group was 110 months, S group
ARID1A(12.6%) G2E3T361fs (3.1%) 105 months. The median follow up period of NAC group was 79.4, whereas 55.2 in surgery group.
EC Squamous carcinoma, MSS 348 TP53 (61.8%)
NOTCH1 (17.2%)
PIK3CA E545K (3.9%)
TMBIM4Y174fs (3.4%)
Conclusions: This trial showed that NAC plus surgery improves survival over surgery alone
TP53Y220C (2.9%) among patients with locally advanced ESCC, with acceptable and manageable adverse events.
Clinical trial information: NCT02395705. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 289s

4076 Poster Session 4077 Poster Session

An interim analysis of phase III study on neoadjuvant chemotherapy versus Circulating tumor DNA (ctDNA) analysis for improved treatment response
perioperative toripalimab plus neoadjuvant chemotherapy for locally ad- assessment and prediction of clinical outcomes in patients with esophageal
vanced esophageal squamous cell carcinoma: Henan Cancer Hospital Tho- cancer. First Author: Zexi Allan, Division of Cancer Research, Peter MacCallum Cancer
racic Oncology Group 1909 (HCHTOG1909). First Author: Yan Zheng, The Centre, Melbourne, Australia
Affiliated Cancer Hospital of ZhengZhou University/Henan Cancer Hospital, Zhengzhou, Background: Despite curative-intent treatment, patients with esophageal cancer
Henan, China experience a high risk of recurrence, and optimal patient management is limited by poor
Background: In the era of immunotherapy, whether neoadjuvant immunochemotherapy risk stratification and treatment response assessment strategies. ctDNA has demon-
(NAIC) would be standard treatment of locally advanced esophageal squamous cell strated value as a prognostic biomarker in esophageal cancer; however, improved
carcinoma (ESCC) is without conclusion. The HCHTOG1909 was aimed to compare the analytical and clinical performance of ctDNA analysis is necessary to reliably support
safety and long-term efficacy of NAIC followed by minimally invasive esophagectomy patient management decisions in clinical practice. Methods: Between September 2017
(MIE) with those of neoadjuvant chemotherapy followed by MIE. This second interim and June 2023, plasma samples were collected from patients with esophageal cancer
analysis was aim to compare the short term results of two groups. Methods: A pro- before, during, and after standard of care treatment in the routine care setting. In this
spective, single-center, open-label, randomized phase III clinical trial. Between May 15, retrospective analysis, ctDNA testing was performed on a subset of patients using a
2020 and April 23, 2024, 401 resectable ESCC with clinical stage T1N1-3M0 to T2-3N0- next-generation tumor-informed assay interrogating up to 50 personalized variants
3M0 were enrolled(8th UICC-TNM), 196 in the toripalimab group and 205 in the che- (Haystack MRD, Quest Diagnostics). Results: ctDNA was assessed in 149 samples from
motherapy group. The patients receive either neoadjuvant paclitaxel (175 mg/m2) and 51 patients with stage I-III esophageal adenocarcinoma (n= 40) or squamous cell
cisplatin (75 mg/m2) plus toripalimab (240mg) (toripalimab group) or paclitaxel and carcinoma (n= 11). Fifteen patients with clinical follow-up (FU) available at the time of
cisplatin alone (chemotherapy group) every 3 weeks for 2 cycles. After MIE, the tor- analysis had at least one sample collected after curative-intent treatment [neoadjuvant
ipalimab group received toripalimab (240 mg every 3 weeks for up to 6 months). The chemoradiotherapy (nCRT) and surgery (n=10) or definitive CRT (n=5)]. ctDNA was
event-free survival (EFS) was the primary endpoint. The pathological complete response detected (ctDNA+) following curative-intent treatment in 5/15 (33%) patients, all 5
(pCR) was the key secondary endpoints. Other endpoints included postoperative (100%) of whom experienced disease recurrence or were deceased at FU (median time
complications, mortality, adverse events, overall survival and disease free survival. We from ctDNA+ result to FU: 14.4 months, range: 0.1-24.5). Of the 10 (67%) patients with
planned 3 interim analyses. This was a planned second interim analysis. The sample size no ctDNA detected (ctDNA-) following curative-intent treatment, 7 (70%) were disease-
was calculated based on the primary endpoint EFS. The hazard ratio assumed to be 0.68 free at FU (median time from ctDNA- result to FU: 46.7 months, range: 6.3-65.5). In the
between two groups. A type I error allocated (two-sided) 0.05, 90% power and drop-out neoadjuvant setting, paired pre- and post-nCRT samples were evaluated in 18 patients,
rate of 10% in 5 years. The x2 test and the Fisher exact test was employed for categorical demonstrating ctDNA detection in 18/18 (100%) patients prior to nCRT versus 8/18
parameters, the t test or analysis of variance was adopted for continuous variables. (44%) following nCRT. ctDNA positivity following nCRT was strongly associated with
Results: Among 401 patients ( 305 men [76.1%]; mean [SD] age, 70.7 [3.5] years; most poor pathological response (p=0.0026), and ctDNA dynamics observed longitudinally
frequent clinical stages III 213 [53.1%] ). The toripalimab group had a higher pCR rate during nCRT served as a robust indicator of response. Of note, one patient experienced
(26.1% vs. 6.2%; P , 0.001). The 90-day perioperative mortality rate was 2.42%(4) for metastatic progression during nCRT, discovered at surgery, and ctDNA levels in this
the toripalimab group and 2.5%(4) for the chemotherapy alone group (P = 0.9790). The patient increased 550-fold while on nCRT. Conclusions: Evaluation of ctDNA using a
most frequent irAE was hypothyroidism. There was no significant difference observed next-generation tumor-informed platform supports improved response assessment to
for postoperative complication rate (P = 0.453). The grade 3 or 4 treatment-related nCRT as well as accurate risk stratification following curative-intent treatment in pa-
adverse events did not differ between the two groups (13.8% versus 10.8%). tients with esophageal cancer. ctDNA positivity following curative-intent treatment
Conclusions: The interim results of HCHTOG1909 showed the addition of perioperative predicted disease recurrence with a lead time of up to 22 months. Additional analyses
toripalimab to NAC is safe in resectable ESCC, and the pCR rate is significantly improved. are ongoing to further validate these findings. Research Sponsor: None.
Clinical trial information: NCT04280822. Research Sponsor: None.

4078 Poster Session 4079 Poster Session

Clinical implication of MDM2 amplification in advanced biliary tract cancer Real-world outcomes of first-line therapies for unresectable hepatocellular
(BTC): A propensity score-matched, retrospective cohort study of 813 carcinoma in the United States. First Author: Masafumi Ikeda, National Cancer
patients. First Author: Hyunseok Yoon, Department of Oncology, Asan Medical Cen- Center Hospital East, Kashiwa, Japan
ter, University of Ulsan College of Medicine, Seoul, South Korea Background: Unresectable hepatocellular carcinoma (uHCC) remains a significant clinical
Background: Restoring p53 tumor suppressor activity by blocking the interaction challenge despite advances in systemic therapies. Real-world evidence complements
between p53 and MDM2, its endogenous negative regulator, has emerged as a potential clinical trials by evaluating treatment effectiveness in diverse patient populations. This
therapeutic target for several tumors including BTC. However, the frequence and clinical study assessed real-world progression-free survival (rwPFS) and overall survival (rwOS) for
implication of MDM2 amplification (amp) has not been investigated for BTC. current standard-of-care first-line (1L) systemic therapies for uHCC. Methods: This
Methods: Patients with unresectable or metastatic BTC who had available tissue-based retrospective study used the Flatiron Health database and included untreated uHCC
targeted next generation sequencing (NGS) data and were treated with first-line patients diagnosed on or after January 1, 2018, who initiated 1L systemic therapies
gemcitabine plus cisplatin (GemCis)-containing chemotherapy at Asan Medical Cen- (atezolizumab + bevacizumab [atezo + bev], lenvatinib, sorafenib, or durvalumab +
ter, Seoul, Korea between January 1, 2016, and December 31, 2023, were included. tremelimumab [durva + treme]) on or after May 29, 2020. Baseline characteristics, in-
MDM2-amp was defined 5 or greater copies per tumor cell. Baseline characteristics and cluding ECOG status, ALBI grade, and demographic factors, were reported across cohorts.
clinical outcomes to GemCis-containing therapy were compared according to the Kaplan-Meier method was used to estimate rwPFS and rwOS for each cohort. Results: A
presence of MDM2-amp/TP53 wild-type (WT). Propensity score matching (PSM) with a total of 1,539 patients were included: atezo + bev (n = 1,070), durva + treme (n = 238),
1:4 ratio was performed to balance the baseline characteristics between the patients lenvatinib (n = 139), and sorafenib (n = 92). Baseline characteristics were similar across
with and without MDM2-amp/TP53-WT. Results: Among 813 patients, 41 (5.0%) had cohorts. Most patients had ECOG (Eastern Cooperative Oncology Group) status 0–1
MDM2-amp/TP53-WT and there was no significant association with primary tumor sites: (57%–70%), with 15%–21% having ECOG 2+. ALBI (Albumin-Bilirubin) grade 2 was ob-
4.7% in intrahepatic cholangiocarcinoma, 3.7% in extrahepatic cholangiocarcinoma, and served in 41%–60% of patients, while ALBI grade 3+ was present in 8.1%–17%. The median
8.0% in gallbladder cancer (p=0.111). Patients with MDM2-amp/TP53-WT significantly rwPFS for atezo + bev, lenvatinib, and durva + treme were similar at 4.7 months (95% CI:
had less frequent viral hepatitis B infection (2.4% vs. 18.4%, p=0.009) and lung me- 4.1–5.4), 4.6 months (95% CI: 3.8–5.5), and 4.2 months (95% CI: 3.2–5.7), respectively.
Sorafenib had a significantly shorter rwPFS at 3.0 months (95% CI: 2.5–4.4). The median
tastasis (2.4% vs. 13.2%, p=0.043); otherwise, no significant association with baseline
rwOS for atezo + bev, lenvatinib, and sorafenib were similar at 10.7 months (95% CI:
characteristics was noted. After PSM (40 for MDM2-amp/TP53-WT vs. 155 for non-
9.5–11.8), 10.4 months (95% CI: 7.8–13.4), and 10.5 months (95% CI: 5.6–14.9). Durva +
MDM2-amp/TP53-WT), patients with MDM2-amp/TP53-WT showed significantly longer
treme had a significantly shorter rwOS at 7.6 months (95% CI: 5.7–18.6). At 12 months, the
progression-free survival compared to those in the matched group (median, 9.6 vs.
survival probabilities for rwOS were 45% (95% CI: 42%, 48%) for Atezo + Bev, 41% (95% CI:
6.9 months; p=0.034) and non-significant tendency toward longer overall survival 33%, 50%) for Durva+Treme, 45% (95% CI: 36%, 55%) for lenvatinib, and 43% (95% CI: 33%,
(median, 20.3 vs. 16.4 months; p=0.103). Conclusions: In patients with unresectable or 56%) for sorafenib. At 24 months, survival probabilities for rwOS were 27% (95% CI: 24%,
metastatic BTC, MDM2-amp/TP53-WT occurred in 5% and it was associated with better 31%) for Atezo + Bev, 26% (95% CI: 19%, 36%) for lenvatinib, 21% (95% CI: 12%, 35%) for
survival outcomes of first-line GemCis-containing chemotherapy. Our findings suggest sorafenib, and data were undetermined for Durva + Treme. Conclusions: Findings from
that MDM2-amp/TP53-WT serves as a biomarker for a distinct subgroup of BTC, this real-world analysis show that atezo + bev demonstrated comparable outcomes versus
warranting active investigation into MDM2 inhibitors. Research Sponsor: Boehringer lenvatinib, and potential benefits in rwPFS versus sorafenib and rwOS versus durva +
Ingelheim. treme. The median rwOS was approximately 10 months across treatments and highlights
the need for novel therapies to improve long-term survival in uHCC. These results un-
derscore the importance of evaluating treatment effectiveness in real-world populations,
which may differ from clinical trial cohorts. Further analyses are warranted to explore these
findings and optimize treatment strategies for uHCC. Research Sponsor: Bristol Myers
Squibb.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
290s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4080 Poster Session 4081 Poster Session

First-line rilvegostomig (rilve) plus chemotherapy (CTx) in advanced biliary Combined treatment of durvalumab, bevacizumab and tremelimumab in
tract cancer (BTC): Primary analysis of GEMINI-Hepatobiliary substudy 2 subjects with hepatocellular carcinoma (HCC) or biliary tract carcinoma
Cohort A. First Author: Jian Zhou, Zhongshan Hospital, Fudan University, Shanghai, (BTC). First Author: Joy Awosika, Thoracic and Gastrointestinal Malignancies Branch,
China Center for Cancer Research, National Cancer Institute, National Institutes of Health,
Background: Immune checkpoint inhibitors plus CTx have improved outcomes in first-line advanced BTC Bethesda, MD
(median progression-free survival [PFS] 6.5–7.2 months), but survival remains limited. Rilve, an anti-PD-1/ Background: Anti-VEGF in combination with anti-PD1/PD-L1 represents a synergistic
TIGIT bispecific antibody, may provide benefit by targeting two immune checkpoints. GEMINI-Hepatobiliary
therapeutic strategy that has demonstrated efficacy, prolonging survival in cancers like
(NCT05775159) is a phase 2 study evaluating rilve or volrustomig alone or in combination regimens in
patients (pts) with advanced hepatocellular carcinoma (substudy 1) or BTC (substudy 2). We report data HCC, RCC, and NSCLC. This combination induces modifications in the tumor microenvi-
from Cohort A (rilve plus CTx) in substudy 2. Methods: Pts aged $18 years with previously untreated ronment, leading to a reduction in immunosuppressive cells, improved dendritic cell
unresectable/metastatic BTC and an ECOG performance status 0–1 received rilve every 3 weeks (Q3W) for maturation, antigen presentation, downregulation of immune checkpoint molecules, and
up to 2 years plus gemcitabine 1000 mg/m2 and cisplatin 25 mg/m2 on days 1 and 8 Q3W for up to 8 cycles. enhanced T-cell activity. Combining CTLA-4 inhibitors with anti PD-1/PD-L1 enhances T-cell
Coprimary endpoints were investigator-assessed 6-month PFS per RECIST v1.1 and safety and tolerability; mediated anti-tumor responses by leveraging distinct yet complementary mechanisms.
secondary endpoints included median PFS, objective response rate (ORR), duration of response (DoR; all Targeting VEGF, PD-L1, and CTLA-4 pathways simultaneously in HCC and BTC provides a
investigator-assessed per RECIST v1.1), and pharmacokinetics (PK). Tumoral PD-L1 expression, peripheral
novel approach that hasn’t been tested in clinical trials. Our group previously reported in vivo
PD-1 and TIGIT receptor occupancy (RO), and T cell profiles were also evaluated. Results: Thirty pts were
treated; median age was 60 years, 70.0% were Asian, and 83.3% had metastatic disease. As of Nov 4 2024, activity in murine BTC models and preliminary clinical results supporting this triplet
the median follow-up in all pts was 6.9 months (interquartile range 5.6–9.5) and rilve treatment was combination in BTC. The aim of this study is to determine if VEGF inhibition with anti-CTLA-4
ongoing in 36.7% of pts. Efficacy and safety data are shown in the Table. The 6-month PFS rate was 73.0%; and anti-PD-LI therapy augments antitumor immunity and clinical responses in HCC and BTC
median PFS was 8.3 months. Median PFS was numerically longer in pts with PD-L1 tumor area patients. Methods: This was a Phase II trial conducted to evaluate efficacy of durvalumab,
positivity $1% (9.4 months, n = 18) vs the overall study population. The safety profile was manageable and bevacizumab and tremelimumab in advanced HCC BCLC stage C or BTC. Participants
consistent with prior studies. Rilve exposure was consistent with historical monotherapy data, indicating an received bevacizumab at 7.5mg/kg and durvalumab 1150mg every 3 weeks by IV infusion on
absence of PK drug interactions and cross-indication differences. Rilve achieved $90% PD-1 and TIGIT RO Day 1 of Cycle 1 (durvalumab) and Day 1 of Cycle 2 (bevacizumab). Tremelimumab at a dose
on peripheral T cells and induced peripheral T cell proliferation. Conclusions: Rilve plus CTx demonstrated
promising efficacy with a manageable safety profile and sustained target engagement. Longer follow-up for
of 300mg was administered by IV infusion only once on Day 1 of Cycle 1. The combination of
data maturity is warranted. Phase 3 studies with rilve in BTC (ARTEMIDE-Biliary 01; DESTINY-BTC01) are durvalumab and bevacizumab continued in 3-week cycles until disease progression or
ongoing. Clinical trial information: NCT05775159. Research Sponsor: AstraZeneca. unacceptable toxicity. Primary endpoint was 6-month progression-free survival (PFS) and
secondary endpoints were safety, overall survival (OS) and best overall response (BOR).
N=30*
Correlative studies assessing immune response were performed. Results: Between March
PFS 2021 and August 2024, 27 patients were enrolled (HCC: 6pts, BTC: 21pts). The median age
Events in all dosed pts, n (%) 19 (63.3)
6-month rate, % (95% CI) 73.0 (53.2–85.5) was 66y (39-80) and 62% were male. 37% of the patients enrolled received prior ICI. As of
Median, months (95% CI) 8.3 (6.7–9.6) November 4th, 2024, with a median follow-up of 8mos, mPFS was 3.5mos and mOS 9.5mos
ORR, % (95% CI) 31.0 (15.3–50.8) in all 27 efficacy-evaluable pts. The estimated 6 months PFS rate was 37%. The BOR was
Best overall response, n (%)
Partial response 9 (31.0) partial response in 4 pts (18%) followed by stable disease in 9pts (40%). The most common
Stable disease 18 (62.1) grade 3-4 TRAEs were lymphopenia (6pts, 22%), anemia (9pts, 33%), diarrhea/colitis (7pts,
Progressive disease 2 (6.9) 25.9%), elevated lipase (4pts, 14.8%). Treatment discontinuation related to AEs occurred in
Median DoR, months (95% CI) 6.9 (2.8–not calculated)
Any / rilve-related AEs, n (%) 30 (100) / 21 (70.0) 7pts (26%). One treatment-related death occurred secondary to an upper gastrointestinal
Grade ‡3 26 (86.7) / 4 (13.3) bleed. Conclusions: The combination of durvalumab, bevacizumab and tremelimumab did
Serious AEs 12 (40.0) / 2 (6.7)
Leading to rilve discontinuation 1 (3.3) / 0
not meet its primary endpoint but demonstrated a clinically meaningful overall survival
Leading to death 2 (6.7) / 0 benefit. No new safety signals were seen. Clinical trial information: NCT03937830. Research
*N=29 for response outcomes.
Sponsor: None.
AE, adverse event; CI, confidence interval.

4083 Poster Session 4084 Poster Session

Outcomes by baseline tumor burden using the 6-and-12 score in EMERALD- Phase I study of Ori-C101, an armored GPC3-directed CAR-T, in patients
1: A phase 3 study of durvalumab (D) 6 bevacizumab (B) with transarterial with advanced hepatocellular carcinoma (HCC). First Author: Jia Fan,
chemoembolization (TACE) in embolization-eligible unresectable hepato- Department of Liver Surgery and Transplantation, Liver Cancer Institute, Zhongshan
cellular carcinoma (uHCC). First Author: Joseph Patrick Erinjeri, Interventional Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion
Radiology Service, Memorial Sloan Kettering Cancer Center, New York, NY (Fudan University), Ministry of Education, Shanghai, China
Background: In EMERALD-1 (NCT03778957), D + B + TACE significantly improved progression-free Background: Previously, we reported the results of Ori-C101 from investigator initiated trial
survival (PFS) vs TACE in participants (pts) with embolization-eligible uHCC. Tumor burden is a prognostic in China (ChiCTR1900028121). The data demonstrated that Ori-C101 owned a favorable
factor in HCC. Prior analyses showed improvements in PFS with D + B + TACE vs TACE in pts who met or safety profile and promising efficacy. Among 10 GPC3+ HCC patients (pts) treated with Ori-
exceeded the up-to-7 criterion (a measure based on tumor number and size), and in those with max tumor
C101, 9 pts (90%) achieved disease control and 6 pts (60%) met partial response per RECIST
diameters of , 10 cm or $10 cm. The 6-and-12 score measures tumor burden based on tumor number
and size. We assessed outcomes in EMERALD-1 by baseline tumor burden using the 6-and-12 score.
1.1. Two pts with PR attained progression-free survival of one and two years respectively, with
Methods: Pts were randomized [Link] to D + B + TACE, D + TACE, or TACE. Pts received D (1500 mg) or an overall survival close to 3 years. These results implied that Ori-C101 potentially held
PBO for D (Q4W) + TACE. After completing the last TACE, pts received D (1120 mg) + B (15 mg/kg), D significant clinical benefits. Subsequently, extensive optimizations and improvements in the
(1120 mg) + PBO for B, or PBOs for D and B (Q3W). In pts who received D + B + TACE and TACE, PFS, time manufacturing process were implemented to enhance its clinical efficacy and persistency.
to progression (TTP), and objective response rates (ORR), per BICR RECIST v1.1 in the intent-to-treat (ITT) Hence, a multicenter registration study was launched in China (the BEACON study), and
population, and safety and number of TACE cycles in the safety analysis set (SAS; pts received $1 dose of herein, we will present the preliminary results. Methods: This is an open-label, multi-center,
study treatment [tx], regardless of randomization) are reported by baseline tumor burden using 6-and-12 dose-escalation (3+3 design) study. GPC3+ advanced HCC pts who failed at least 2 lines of
scores: #6, . 6–12, or . 12. Results: Overall, 40.0%, 43.9%, and 16.2% of pts belonged to the #6, . systemic treatments received a single hepatic arterial infusion with a total dose of 0.9 to
6–12, and . 12 groups, respectively. The number of pts who received $2 TACE cycles increased across 63108 CAR-T cells. Primary endpoints are rate of dose-limiting toxicities (DLTs) and safety
the groups (#6: 63.8%; . 6–12: 81.7%; . 12: 89.8%). PFS and TTP improved with D + B + TACE vs TACE,
regardless of baseline tumor burden, with the best relative improvement in hazard ratios (HRs) in the . 12
with the aim to determine a recommended phase II dose (RP2D). Secondary endpoints are
group (Table). ORRs were higher for D + B + TACE vs TACE in all groups. Max Grade 3–4 tx-related adverse cellular kinetics, overall response rate by investigator assessment, duration of response,
event (TRAE) frequencies were numerically higher with D + B + TACE vs TACE across tumor burden overall survival and overall safety. Results: As of Dec 17th, 2024, a total of 10 eligible pts
groups; differences were reduced when adjusted for exposure. No tx-related deaths occurred with D + B + received Ori-C101 infusion at 3 dose levels (DLs). All pts had BCLC stage B or C, with 20% (2/
TACE. Conclusions: PFS, TTP, and ORR benefits were seen with D + B + TACE vs TACE with manageable 10) had extrahepatic metastasis. The median number of prior lines of therapy was 4.5 (range
safety, regardless of tumor burden, further supporting a favorable risk-benefit profile with D + B + TACE in 2-9), 100% pts received immune checkpoint inhibitors and tyrosine kinase inhibitors. All pts
embolization-eligible uHCC. Clinical trial information: NCT03778957. Research Sponsor: AstraZeneca. were evaluable for safety. All adverse events were reported regardless of study drug rela-
£6 >6–12 >12 tionship. Of 10 pts evaluable for safety, the most common $ grade (G) 3 AEs were lymphocyte
count decreased (100%), neutrophil count decreased (60.0%), blood fibrinogen decreased
D + B + TACE TACE D + B + TACE TACE D + B + TACE TACE
ITT n=81 n=82 n=84 n=95 n=38 n=28 (40.0%), transaminases increased (40.0%), platelet count decreased (20.0%), blood bilirubin
increased (20.0%). CRS was observed in 10 (100%) pts with 3 (30.0%) $ G3 CRS. No ICANS
Median PFS (95% CI), months 19.4 11.1 13.9 9.7 11.1 4.8
(13.7–24.9) (7.0–13.6) (7.2–19.6) (6.9–16.3) (4.4–16.6) (2.9–6.9) was observed. One pt developed DLT event due to CRS and secondary disseminated in-
PFS HR vs TACE (95% CI) 0.69 0.85 0.61 travascular coagulation. 9 pts were evaluable for efficacy per RECIST 1.1. While 6 pts (66%)
(0.47–1.01) (0.59–1.22) (0.33–1.13) achieved disease control at DL2 or higher, all pts at the DL3 achieved objective response.
Median TTP (95% CI), months 22.1 11.1 22.0 15.4 16.6 5.1
(15.1–30.5) (7.0–13.9) (13.9–27.7) (7.2–16.7) (6.9–25.1) (3.0–7.1) Particularly, one pt who achieved CR showed encouraging durability and no signs of relapse at
TTP HR vs TACE (95% CI) 0.60 0.66 0.42 9 months follow up evaluation, and follow up is ongoing. Conclusions: These preliminary data
(0.40–0.90) (0.43–1.01) (0.20–0.87) showed Ori-C101 has manageable safety profile and exciting efficacy with encouraging sign
ORR, n (%)* 47 (58.8) 27 (33.8) 31 (36.9) 32 (33.7) 10 (26.3) 1 (3.6)
SAS n=71 n=81 n=61 n=92 n=22 n=27 of good durability. Currently, more pts have been enrolled at dose expansion to confirm the
Max Grade 3–4 TRAE, n (%) 17 (23.9) 9 (11.1) 17 (27.9) 3 (3.3) 7 (31.8) 0 DLs of RP2D. More information will be presented at coming ASCO conference. Clinical trial
event rate per 100 pt-years 15.9 7.9 19.9 3.1 22.2
information: NCT05652920. Research Sponsor: None.
*In pts with evaluable disease at baseline.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 291s

4085 Poster Session 4086 Poster Session

Modifiable risk-factors, genetic characteristics, and survival in early-onset Real world efficacy and safety of ivosidenib in US veterans with IDH1
cholangiocarcinoma. First Author: Jordan Nunnelee, Mayo Clinic, Rochester, MN mutated cholangiocarcinoma. First Author: Katherine Ismei Zhou, Durham VA
Background: Cholangiocarcinoma (CC) is a rare disease with an increasing incidence Health Care System and Duke University, Durham, NC
among younger adults, which is poorly understood. Genomic profiling of tumors is both Background: IDH1 mutations occur in 13% of patients with intrahepatic chol-
prognostic and predictive of benefit for targeted therapies. We investigated if there are angiocarcinoma. Ivosidenib is FDA approved for the treatment of advanced, previously
clinical and molecular differences between younger vs older patients with CC. treated, IDH1-mutated cholangiocarcinoma. In the ClarIDHy trial, ivosidenib led to an
Methods: We collected TEMPUS genetic data via retrospective chart review from tu- objective response rate of 2%, stable disease rate of 51%, median progression-free
mors in young vs old patients seen at our institution, defined as #50 vs . 50 years of survival (PFS) of 2.7 months, and median overall survival (OS) of 10.3 months. Treatment-
age at time of diagnosis. We included patients diagnosed with CC between January 2008 emergent adverse events resulted in study drug discontinuation in 7% of patients. Data on
and July 2024 with available clinical follow up and TEMPUS genetic sequencing data. We the real-world efficacy and safety of ivosidenib in cholangiocarcinoma remains limited.
collected mutation data on the following actionable genes: FGFR2, IDH1/2, BRCA1, Methods: Patients with IDH1-mutated cholangiocarcinoma who were prescribed ivosi-
BRCA2, BRAF, ATM, ERBB2/3, and KRAS. Patient characteristics and gene expression denib before December 1, 2024, were retrospectively identified from the national Veterans
variables were compared using Chi-square, Fisher’s exact and Wilcoxon rank-sum tests. Affairs (VA) Corporate Data Warehouse. Demographic, clinical, and molecular data were
Kaplan-Meier, log rank tests and a multivariable Cox model were used for survival abstracted from the National Precision Oncology database and electronic medical
analysis. This study was IRB exempt. Results: We included 410 patients, 84 in the young records. Response was assessed based on provider notes and radiology reports. Survival
group with median age at diagnosis of 40.8 years, and 326 in the old group with median was assessed by the Kaplan-Meier method, and covariates evaluated by the Cox pro-
age 68.5 years. 91.5% of patients were white. There was no difference in BMI between portional hazards model. Results: Of 1094 veterans with cholangiocarcinoma who
groups, however the older group had higher rates of hypertension (15.5% vs 57.7%), underwent molecular testing, 82 (7.5%) had an IDH1 mutation. 33 (40%) patients received
hyperlipidemia (6.0% vs 49.7%), cardiovascular disease (1.2% vs 20.6%), and type 2 ivosidenib at 27 VA medical centers. The median age was 74 years (range 46–82). 2
diabetes (6.0% vs 21.5%), (all p , 0.01). Primary sclerosing cholangitis was more patients (6%) had a partial response (PR), 10 (30%) had stable disease (SD), 19 (58%) had
common in the young group (26.2% vs 4.3%, p , 0.01). ECOG status of 0 at first progressive disease, and 2 were not assessed. 20 patients (60%) had received one and 5
treatment was seen in 65.3% of young vs 52.5% of old patients (p = 0.02). FGFR2 patients (15%) received two prior lines of therapy. Of the 8 patients (24%) who received
alterations were more common in the young group (17.9% vs 8.0%, p , 0.01), while ATM first-line ivosidenib, 2 (25%) had a PR and 3 (38%) had SD. Most patients (94%) started
mutations were more common in old vs young (5.5% vs 0%, p = 0.03). There was no age ivosidenib at the labeled dose (500 mg daily). Two patients who started ivosidenib at
difference seen for the other genetic alterations. Mean tumor mutational burden was reduced dose (250 mg daily) had PR and SD as their best response. The median PFS from
higher in the old group (4.1 vs 3.8 mut/mb, p = 0.01). MSI-high was found in 2% of cases start of ivosidenib was 4.0 months, and the median OS was 10.5 months. In a multivariable
with no difference between groups. There was no significant difference in overall analysis, PFS and OS were not significantly associated with age, line of therapy, IDH1
survival between age groups. There was a numeric difference in overall survival in stage variant allele frequency, or IDH1 mutation (17 IDH1 R132C vs. 8 other). Patients with
IV patients, though not statistically significant (17.8 months vs. 16.3 months, p = 0.08). IDH1-mutated, advanced cholangiocarcinoma treated with ivosidenib had a median OS of
In a multivariable Cox analysis, female sex, earlier stage at diagnosis and clinical trial 25.3 months from diagnosis, compared to 8.7 months for patients who did not receive
enrollment were associated with favorable prognostics. Conclusions: Our data highlight ivosidenib. Toxicities leading to dose reduction, interruption, or discontinuation of ivo-
relatively low rates of comorbidities associated with metabolic dysfunction in younger sidenib occurred in 3 patients (9%). Conclusions: In this real-world cohort, patients with
adults with CC, suggesting alternative factors are likely to explain the increasing in- IDH1-mutated advanced cholangiocarcinoma treated with ivosidenib had similar re-
cidence of early-onset disease. FGRFR2 is a more common pathogenic alteration among sponse rate, PFS, and OS compared to ClarIDHy. Toxicities leading to dose reduction,
the young and could inform targeted therapies. Younger patients with CC may not have interruption, or discontinuation were rare. The only two partial responses were observed in
improved survival outcomes compared to their older counterparts. This underscores the the first-line setting, including one with a reduced starting dose. This suggests that
aggressive nature of CC and the need for more effective therapies to improve outcomes. frontline ivosidenib may be a reasonable alternative for patients with advanced
Research Sponsor: None. cholangiocarcinoma. Research Sponsor: None.

4087 Poster Session 4088 Poster Session

HAIC plus TAE combined with tislelizumab and surufatinib in unresectable Efficacy and safety of regorafenib combination with PD-1 inhibitors vs.
intrahepatic cholangiocarcinoma: The REACH-01 trial. First Author: Kangshuai regorafenib monotherapy in second-line treatment for patients with unre-
Li, Qilu Hospital of Shandong University, Jinan, Shandong, China sectable hepatocellular carcinoma after failure of different first-line treat-
Background: REACH-01 (NCT06239532) is a single-arm, open label, prospective trial, ments: A multicenter retrospective real-world study. First Author: Weihong Ma,
aiming to evaluate the safety and preliminary effectiveness of hepatic artery infusion Comprehensive Liver Cancer Center, The 5th Medical Center of PLA General Hospital,
chemotherapy (HAIC) plus transcatheter arterial embolization (TAE) combined with Beijing, China
tislelizumab and surufatinib as first-line therapy for unresectable intrahepatic chol- Background: Regorafenib is the first oral targeted drug as a second-line agent in patients with unresectable
angiocarcinoma (iCCA). Methods: Twenty-eight patients with pathologically confirmed hepatocellular carcinoma (HCC) who progressed on sorafenib [Link] is a lack of data to validate the
second-line therapy after progression of targeted-immune combination therapy. Our aim was to investigate
iCCA received TAE with undrugged microspheres and hepatic arterial infusion of the efficacy and safety of regorafenib alone or in combination with a programmed death-1 (PD-1) inhibitor in
oxaliplatin (85 mg/m2) and raltitrexed (3 mg/m2) at an interval of at least 3 weeks along second-line treatment for patients who have failed tyrosine kinase inhibitor (TKI) in combination with PD-1 or
with intravenous tislelizumab (200 mg) Q3W and oral surufatinib (150 - 250 mg) once TKI monotherapy, respectively. Methods: A total of 288 patients were enrolled in this multicenter, retro-
daily. The primary endpoint was the objective response rate (ORR). Secondary outcomes spective study. These patients received regorafenib with or without PD-1 inhibitor (Sintilimab/
included progression-free survival (PFS), conversion to surgical resection rate, overall Camrelizumab/Pembrolizumab) as second-line therapy after failure of TKI (sorafenib/lenvatinib) or such
TKIs combined with PD-1 inhibitor (Sintilimab/Camrelizumab/Pembrolizumab). The primary study endpoint
survival (OS), 1-year OS rate, disease control rate (DCR), and incidence of adverse was the evaluation of overall survival (OS), while secondary study endpoints were progression-free survival
events. Results: As of December 18, 2024, the median follow-up time was 9.33 months, (PFS), objective response rate (ORR), disease control rate (DCR), and treatment safety. Results: In the first
15 patients achieved partial response and the ORR was 57.69 % per RECIST v1.1 criteria. line treatment, 126 patients received TKI and 162 patients received TKI plus PD-1. In the TKI cohort, the Reg-
The conversion to surgical resection rate was 15.38 %. The DCR was 80.77 %. Secondary PD-1 group exhibited markedly higher ORR (29.69% vs 4.84%; p,0.001) and DCR (89.06% vs 67.74%;
endpoints of progression-free survival, overall survival and 1-year OS rate were not p=0.004), as well as longer median PFS (10.5 vs 4.7 months; p,0.001) and median OS (18.9 vs 14.0 months;
p=0.03) compared to the Reg monotherapy [Link] was no significant difference in PFS, OS, ORR, and
mature at the time of the analysis. Further, treatment related adverse effects (TRAEs) of DCR between the two groups in the TKI plus PD-1 [Link] incidence of AEs was higher in the Reg-PD-1
any grade occurred in 28 patients. Manageable grade 3 adverse events (AEs) occurred in group compared to the Reg group (81.25 % vs 58.06%; p=0.005) in the TKI cohort. And Reg-PD-1 group was
32.14% of patients, commonly elevated alanine aminotransferase (7.14%), anorexia comparable to Reg group (76.70% vs 66.10%; p=0.144) in the TKI Plus PD-1 cohort.
(7.14%), and hypokalemia (7.14%). Conclusions: HAIC plus TAE combined with tisle- Conclusions: Regorafenib plus PD-1 may enhance efficacy in uHCC patients who failed first-line TKI therapy.
lizumab and surufatinib are safe and promising first-line treatment selection for However, in patients who have progressed after first-line TKI plus PD-1 therapy, using regorafenib alone or in
combination with PD-1 in second-line therapy does not show a significant difference in [Link]
unresectable iCCA. Clinical trial information: NCT06239532. Research Sponsor: Key findings have significant implications for the selection of second-line treatment strategies for HCC patients,
Research and Development Program of Shandong Province; 2021CXGC011105; BeiGene indicating that the combination of regorafenib and PD-1 might not provide additional benefits in certain
Co. Ltd. patient subgroups. Research Sponsor: None.
Outcomes in the two cohorts.
TKI TKI plus PD-1
Reg Reg-PD-1 Reg Reg-PD-1
Outcomes (n=62) (n=64) P value (n=59) (n=103) P value
CR 2 3 - 3 2 -
PR 1 16 - 6 21 -
SD 39 38 - 35 62 -
PD 20 7 - 15 18 -
ORR 4.84% 29.69% ＜0.001 15.25% 22.33% 0.276
DCR 67.74% 89.06% 0.004 74.58% 82.52% 0.227
PFS(m) 4.7 10.5 ＜0.001 6.3 9.2 0.062
OS(m) 14.0 18.9 0.03 13.2 16.2 0.13

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
292s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4089 Poster Session 4090 Poster Session

Characterization of CLDN18 expression in a Western biliary tract cancer TACE-HAlC combined with donafenib and immune checkpoint inhibitors for
population. First Author: Alexander Bray, University of Michigan, Ann Arbor, MI BCLC stage C HCC patients (THEME study): A retrospective IPTW adjusted
Background: Patients with advanced biliary tract cancer (BTC) have poor survival cohort study. First Author: Linan Yin, Harbin Medical University Cancer Hospital,
despite recent advances in chemoimmunotherapy. Claudin 18 (CLDN18) directed Harbin, China
therapy has shown benefit in combination with FOLFOX or CAPOX in gastric cancer and Background: Transarterial chemoembolization (TACE) combined with hepatic arterial
may also have therapeutic utility in advanced BTC. However, there are limited data on infusion chemotherapy (HAIC) has demonstrated superior objective response rate (ORR)
CLDN18 RNA and protein expression in BTC, especially in Western populations. and progression-free survival (PFS) compared to TACE alone, particularly in patients with
Methods: Exome-capture based RNA sequencing was performed on BTC tissue samples unresectable hepatocellular carcinoma (uHCC) with portal vein tumor thrombosis (PVTT),
through the MI-ONCOSEQ study at the University of Michigan. Fragments per Kilobase of as shown in previous studies. Additionally, Donafenib exhibited significant survival
transcript per Million mapped reads (FPKM) was used to normalize raw read counts. benefits and better safety profiles compared to Sorafenib in a Phase III clinical trial. We
Immunohistochemical staining was completed on a BTC tissue microarray (n = 28), and aimed to retrospectively compare the efficacy and safety of TACE-HAIC combined with
Western blotting was done on human BTC cell lines (SNU-1079, RBE, and SSP-25) using Donafenib and immune checkpoint inhibitors (Quadruple Therapy Group) versus the
CLDN18 recombinant rabbit monoclonal antibody (34H14L15, Invitrogen). Statistical standardized targeted therapy (TKIs or bevacizumab) plus immune checkpoint inhibitors
significance was defined as p , 0.05. Results: We identified transcriptomic data from (Targeted-Immunotherapy Group) in patients with BCLC stage C hepatocellular carcinoma
148 consecutive BTC cases with median age 61 (range 17-81) years and 75 (50.7%) were (HCC). Methods: We conducted a retrospective analysis of patients with BCLC stage C
female. Of these patients, 45 (30.4%) expressed CLDN18 mRNA with FPKM . 1. A lower hepatocellular carcinoma (HCC) who received quadruple therapy or targeted-
proportion of intrahepatic cholangiocarcinoma (CCA) patients (n = 23/110; 20.9%) immunotherapy at the Harbin Medical University Cancer Hospital between September
expressed CLDN18 mRNA relative to extrahepatic CCA (n = 13/25; 54.2%) and gall- 2019 and October 2024. To minimize baseline imbalances between the groups, we applied
bladder cancer (n = 8/11; 72.7%) (p , 0.006). CLDN18 mRNA expression was not stabilized inverse probability of treatment weighting (sIPTW) methods. Results: A total of
associated with stage at diagnosis, but was higher in metastatic versus primary sites 195 patients were included in the study, of whom 125 were assigned to the Quadruple
(35.6 vs 16.1%, p , 0.05). Overall survival was not associated with CLDN18 gene Therapy Group and 70 to the Targeted-Immunotherapy Group. Within the Targeted-
expression in univariate analysis (hazard ratio 1.02, p . 0.9). CLDN18 protein expression Immunotherapy Group, 44 patients received TKIs combined with immune checkpoint
inhibitors, while 26 patients received bevacizumab combined with immune checkpoint
was observed in 8/28 (28.6%) patients using a cutoff defined in prior gastric cancer
inhibitors. After applying sIPTW to balance the baseline characteristics between the two
clinical trials (2+ to 3+ staining intensity in $ 75% of tumor cells). In human BTC cell
groups, patients in the Quadruple Therapy Group demonstrated a significantly higher
lines, SSP-25 expressed CLDN18, but the SNU-1079 and RBE lines did not.
median overall survival (OS) compared with the Targeted-Immunotherapy
Conclusions: CLDN18 is expressed in a modest subset of Western patients with ad-
Group(29.4 months [95% CI: 23.9–NA] vs 18.0 months [14.7–31.8]; P = 0.045). Addi-
vanced BTC and CLDN18-directed therapy may be effective in this disease, particularly in tionally, the median progression-free survival (PFS) assessed by the modified Response
ECC and gallbladder cancer given their higher frequency of expression. CLDN18 positive Evaluation Criteria in Solid Tumors (mRECIST) was longer in the Quadruple Therapy
human BTC cell lines represent a promising preclinical model system for further in- Group(16.4 months [95% CI: 12.7–NA] vs 10.0 months [3.3–31.8]; P = 0.013). The objective
vestigation of this therapeutic strategy. Research Sponsor: National Cancer Institute; response rate (ORR) evaluated according to mRECIST was also higher in the Quadruple
5T32CA009357-42; Rogel Cancer Center. Therapy Group(68.4% vs 28.2%, P , 0.001).The incidence of any adverse events in the
Quadruple Therapy Group was 95.2%, compared with 97.1% in the Targeted-
Immunotherapy [Link] these the incidence of grade $3 adverse events was
40.8% in the Quadruple Therapy Group and 38.6% in the Targeted-Immunotherapy Group.
Conclusions: Compared with Targeted-Immunotherapy Group, patients with BCLC stage C
HCC treated with TACE-HAIC combined with Donafenib and immune checkpoint inhibitors
therapy demonstrated superior efficacy and acceptable safety. Research Sponsor: None.

4091 Poster Session 4092 Poster Session

Multimodal evaluation of metabolic dysfunction–associated steatotic liver Phase 1 expansion study of FF-10832 (liposomal gemcitabine) antitumor
disease (MASLD)–related biliary tract cancer (BTC) and immunotherapy activity in patients with advanced biliary carcinomas. First Author:
outcomes. First Author: Nakul Manish Shah, The University of Texas MD Anderson Gerald Steven Falchook, Sarah Cannon Research Institute at HealthONE, Denver, CO
Cancer Center, Houston, TX Background: FF-10832 has demonstrated improved pre-clinical anti-tumor activity
Background: Rising incidence of BTC, particularly intrahepatic cholangiocarcinoma compared to gemcitabine (GEM). Associated factors may include its prolonged circulating
(iCCA), may be linked to increasing incidence of obesity, MASLD and type 2 diabetes half-life, tumor accumulation, and immune [Link] first in human dose finding trial
mellitus (T2DM). Immunotherapy with immune checkpoint inhibitors (ICIs) has modestly of FF-10832 demonstrated a tolerable safety profile and anti-tumor activity in heavily pre-
extended survival in biliary tract cancer. However, the prevalence of MASLD in BTC, its treated patients (pts) with solid tumors who progressed on prior gemcitabine. A biliary
immune microenvironment (TME) and outcome of MASLD-BTC with ICIs are unknown. tract cancer (BTC) pt maintained a PR .60 weeks after progression on prior GEM based
Methods: Retrospective analysis of BTC patients (pts) treated with ICI between 5/2021- therapy. We subsequently enrolled an expansion cohort evaluating FF-10832 mono-
5/2024 with durvalumab, cisplatin, and gemcitabine. We used American Association for therapy in BTC and describe the results (NCT03440450). Methods: Pts $18 years with
Liver Diseases (AASLD) criteria for MASLD: 1) metabolic dysfunction and 2) steatosis on advanced BTC who had progressed on up to 3 lines of therapy were treated with FF-10832
imaging or biopsy. We calculated liver proton-density fat fraction (PDFF) in pre- 40 mg/m2 IV Day 1 Q 21 days until disease progression or unacceptable toxicity. Re-
treatment non-contrast CT scans (PDFF estimate 5% labelled as steatosis). We ex- sponse was assessed by RECIST 1.1. Modulation of immune cells (flow cytometry/
amined the statistical association between BMI and both tumor genotype and gene multiomics) and population PK were assessed. Results: 18 pts [12M/6F; median age 68
expression patterns using data from institutional genomic platforms (MAPP2 and RTI). (34-79), ECOG PS 0 (3) PS 1 (15)] were treated; median # prior therapies, 2 (1-3); all had
Results: 179 BTC pts (65% of whom were iCCA) treated with durvalumab, cisplatin, and prior GEM and 16 had progressed on prior GEM. Pts received a median of 4 (1 - 22+) cycles
gemcitabine, 103 (57.5%) met AASLD MASLD criteria. In evaluable pts, the median with a median time on study of 10.4 (3.3 -77+) weeks. FF-10832 was well-tolerated. The
overall survival (OS) was 18.4 months, and median follow-up time was 16.7 months. most common drug-related AEs were nausea, pyrexia, and decreased appetite (39% each).
Non-MASLD pts had a median OS of 23.0 months (95% CI: 16.2, NA) versus 16.7 months No Gr 4 toxicity was observed; Gr 3 AEs in .1 pt included anemia (2) and muscular
(95% CI: 12.4, 21.2) with MASLD (p = .056). T2DM was associated with a worse OS (10.6 weakness (2). All AEs were successfully managed using standard therapies. Two pts
versus 21.2 months, p = .004). Multivariable cox model for OS demonstrated a hazard withdrew and 1 pt died of cholangitic sepsis before 1st evaluation. Best overall response in
ratio (HR) of 1.45 (95% CI: 0.87, 2.4; p = .1564) for MASLD and 1.61 (95% CI: .99, 2.65; p = 15 remaining pts was 2 PR, 8 SD, 4 PD and 1 NE. The median PFS and OS were 3.4 and
.0575) for T2DM. The median progression-free survival (PFS) was 8.5 months. MASLD 9.1 months, respectively. Both PRs had received prior GEM/platinum-based therapy: 1) a
BTC had a median PFS of 8.2 months (95% CI: 5.7, 10.1) versus 9.2 months (95% CI: 6.9, gallbladder adenocarcinoma pt achieved a 48% decrease in target lesions with FF-10832
16.2) without MASLD (p = .459). BTC pts with T2DM had median PFS of 5.8 months vs by cycle 2, which was maintained through cycle 10; dose was reduced to 30 mg/m2 at
13.0 months without T2DM (p = .001). Multivariable cox model for PFS had HR of 1.7 cycle 5 for Gr 3 muscle weakness, 2) a hilar cholangiocarcinoma pt achieved a PR by cycle
(95% CI: 1.1, 2.6; p = .014) for T2DM. Within our institutional database (n = 919), we 2,with complete resolution of target lesions before withdrawing. Four additional pts
observed depletion of KRAS (p = .008) and STK11 (p = .02) mutations in BTC pts with high maintained SD $ 6 cycles, with 2 continuing on therapy after 9 and 26 cycles. PK was
BMI. RNA-seq (n = 77) suggests that elevated BMI was associated with low expression of similar to that previously reported (terminal t1/2, 30 hours), with similar log decreases
pan-immune and epithelial-to-mesenchymal transition and increased expression of observed in Ki67+ regulatory T cells and increases observed in CD8+ cells, indicative of
oxidative phosphorylation signatures. Conclusions: BTC is commonly associated with anti-tumor immune activation. Conclusions: FF-10832 is well-tolerated and has anti-
MASLD and may correlate with reduced OS and PFS with ICI, particularly in T2DM pts. tumor activity in pts with advanced BTC who progressed on prior GEM. Although pre-
Our findings suggest a distinct immunogenomic signature in MASLD-BTC and highlight liminary, these results of a single agent therapy compare favorably to those reported for
the importance of further investigating the TME in this population. Research Sponsor: 2nd line combination therapies. This warrants further investigation of FF-10832 efficacy
None. and safety in BTC patients. Clinical trial information: NCT03440450. Research Sponsor:
FUJIFILM Pharmaceuticals USA Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 293s

4093 Poster Session 4094 Poster Session

SHR-8068 plus adebrelimab and bevacizumab for advanced hepatocellular Hepatic arterial infusion chemotherapy combined with donafenib and tisle-
carcinoma (aHCC): A phase 1b/2 study. First Author: Lianxin Liu, The First Af- lizumab versus transcatheter chemoembolization alone for hepatocellular
filiated Hospital of the University of Science and Technology of China/Anhui Provincial carcinoma: A propensity score matching study. First Author: Zhongguo Zhou,
Hospital, Hefei, China Department of Liver Surgery, Sun Yat-Sen University Cancer Center, Guangzhou,
Background: Combination of an anti-PD-1/L1 antibody with an anti-angiogenic agent is currently Guangdong, China
the preferred 1L treatment for aHCC. Addition of a CTLA-4 inhibitor may further improve anti-tumor Background: Although tyrosine kinase inhibitors combined with PD-1/L1 inhibitors have
activity, with complementary immunostimulatory effects from CTLA-4 and PD-1/L1 blockade. We been established as first-line treatment for advanced hepatocellular carcinoma (HCC), the
conducted a multicenter, open-label, phase 1b/2 trial (NCT05444088) to assess SHR-8068, a novel survival benefit remains unsatisfactory. Hepatic arterial infusion chemotherapy (HAIC) has
anti-CTLA-4 monoclonal antibody (mAb), combined with adebrelimab (A, anti-PD-L1 mAb) and emerged as an effective therapy to improve the prognosis of HCC patients. This study aimed
bevacizumab (B) in patients (pts) with aHCC. Methods: Pts with or without prior treatment were to investigate the efficacy and safety of HAIC combined with donafenib and tislelizumab in
enrolled (phase 1b: failed or refused standard therapy; phase 2: #1L systemic therapy, no im-
HCC. Methods: 421 patients diagnosed as HCC and treated in Sun Yat-sen University
munotherapy [IO]). SHR-8068 was evaluated in 2 dosing regimens with AB: 1 mg/kg Q6W (Combo
1) or 4 mg/kg priming dose (Combo 2). An additional cohort evaluated AB alone (Combo 3). A was
Cancer Center from January 2017 to December 2024 were enrolled in this retrospective
dosed at 20 mg/kg Q3W and B at 15 mg/kg Q3W for all regimens. Results: As of Oct 31, 2024, a study, included 151 patients received FOLFOX-HAIC combined with donafenib and tisle-
total of 27, 53 and 21 pts received Combo 1, 2, and 3, respectively, across 2 study phases (overall: IO lizumab (DT-HAIC) and 270 received transcatheter chemoembolization (TACE) alone. To
naı̈ve, 97.0%; prior anti-angiogenic therapy, 32.7%); median follow-up was 16.7, 11.1 and 11.3 mo, avoid the selection bias and balance covariates, we conducted propensity score matching
respectively. In pts treated with Combo 2, the objective response rate (ORR) was 47.2% (25/53; 95% (PSM). The primary outcomes are progression-free survival (PFS) and overall survival (OS);
CI 33.3%–61.4%), with a median duration of response (DoR) of 12.7 mo (95% CI 5.8–NR). The the secondary outcomes include objective response rate (ORR), disease control rate (DCR)
median progression-free survival (PFS) was 8.7 mo (95% CI 5.5–11.6); median overall survival (OS) and safety. Tumor response was evaluated per RECIST v1.1. Results: PSM resulted in 151
was not reached, with a 12-mo OS rate of 76.0% (95% CI 59.3%–86.6%). Numerically improved ORR matched pairs with comparable baseline characteristics between the DT-HAIC and TACE
and survival outcomes were seen with Combo 2 vs Combo 1 and 3 (Table 1). Overall, grade $3 cohorts. Compared with the TACE cohort, Patients receiving DT-HAIC exhibited significantly
treatment-related adverse events (TRAEs) occurred in 55.6%, 41.5% and 42.9% of pts with Combo better median PFS (10.3 vs 4.9 months, P , 0.01) and median OS (not reached vs
1, 2 and 3. The most common grade $3 TRAEs (incidence $10% for any Combo) were decreased 10.9 months, P , 0.01). The ORR and DCR were significantly higher in the DT-HAIC cohort
platelet count (22.2%, 5.7%, and 4.8% for Combo 1, 2, and 3) and hypertension (18.5%, 7.5%, and than in the TACE cohort (ORR: 33.8% vs 11.3%, P , 0.01; DCR: 77.9% vs 65.3%, P = 0.03).
9.5%, respectively). TRAE led to discontinuation of any study agent in 11.1%, 1.9% and 9.5% of pts, There was no treatment-related death. Serious adverse events were similar between the two
respectively. There was 1 treatment-related death (Combo 3). Conclusions: SHR-8068 combined groups, except for alanine aminotransferase (ALT), aspartate aminotransferase (AST),
with adebrelimab and bevacizumab showed promising efficacy and manageable safety in aHCC. A
platelet count, abdominal pain and allergic reaction. In the TACE cohort, the occurrence of
more favorable benefit-risk profile was observed for SHR-8068 given as a priming dose. A phase 3
trial (NCT06618664) is currently underway to further assess the combination as 1L treatment for
Grade 3-4 elevations in ALT (11.3% vs 21.9%, P = 0.02) and AST (21.2% vs 36.4%, P , 0.01),
aHCC. Clinical trial information: NCT05444088. Research Sponsor: Jiangsu Hengrui Pharma- as well as abdominal pain (2.6% vs 16.6%, P , 0.01), was more prevalent. In contrast, the
ceuticals, Co., Ltd. DT-HAIC cohort exhibited a higher incidence of Grade 3-4 thrombocytopenia (11.3% vs 1.3%,
P , 0.01) and allergic reactions (4.6% vs 0, P = 0.02). Conclusions: DT-HAIC significantly
Efficacy outcomes. improved PFS, OS, ORR, and DCR compared with TACE alone, with manageable adverse
Combo 1 (n=27) Combo 2 (n=53) Combo 3 (n=21) events, suggesting that the combination of HAIC with donafenib and tislelizumab may be a
ORR, % (95% CI) 29.6 (13.8–50.2) 47.2 (33.3–61.4) 19.0 (5.5–41.9) promising treatment option for HCC patients. Research Sponsor: None.
Median DoR*, mo (95% CI) NR (9.4–NR) 12.7 (5.8–NR) NR (7.0–NR)
9-mo DoR rate*, % (95% CI) 100.0 (NR–NR) 69.8 (41.7–86.3) 66.7 (5.4–94.5) DT-HAIC (n=151) TACE (n=151) P Value
DCR, % (95% CI) 77.8 (57.7–91.4) 77.4 (63.8–87.7) 81.0 (58.1–94.6)
mPFS 10.3m 4.9m P,0.01
Median PFS*, mo (95% CI) 6.9 (2.7–NR) 8.7 (5.5–11.6) 6.7 (2.8–9.5)
mOS Not reached 10.9m P,0.01
12-mo OS rate*, % (95% CI) 70.4 (49.4–83.9) 76.0 (59.3–86.6) 70.8 (46.2–85.7)
ORR 33.8% 11.3% P,0.01
Tumor response was assessed by investigator per RECIST v1.1. DCR 77.9% 65.3% P=0.03
*Kaplan-Meier method. NR, not reached.

4095 Poster Session 4097 Poster Session

LEAP-002 long-term follow-up: Lenvatinib plus pembrolizumab versus Membrane-specific HER2 expression by artificial intelligence-based quan-
lenvatinib plus placebo for advanced hepatocellular carcinoma. First Au- titative scoring for prediction of efficacy of trastuzumab deruxtecan in
thor: Richard S. Finn, David Geffen School of Medicine at UCLA, Los Angeles, CA biliary tract cancer (HERB trial): Exploratory analysis of a multicenter, single
Background: LEAP-002 was a randomized, double-blind, phase 3 study (NCT03713593) arm, phase II trial. First Author: Mitsuho Imai, Translational Research Supporting
that was conducted to evaluate the efficacy and safety of first-line lenvatinib plus Office, National Cancer Center Hospital East, Kashiwa, Japan
pembrolizumab versus lenvatinib plus placebo in participants with advanced hepato- Background: Trastuzumab deruxtecan (T-DXd) showed promising results in patients with HER2-
cellular carcinoma (HCC). The study did not meet its primary end points of OS at final positive biliary tract cancer (BTC). With T-DXd’s expanding indication into HER2-low cancers,
analysis and PFS at interim analysis. After a median study follow-up of 43.6 months, OS, quantitative Artificial Intelligence (AI)-based scoring of HER2 expression at the cellular level
PFS, and ORR remained consistent with the primary efficacy analysis (OS: HR, 0.84 [95% becomes increasingly important. This study investigated whether the intensity and subcellular
pattern of HER2 staining would correlate with response to T-DXd. Methods: HER2 immuno-
CI, 0.71-0.98]; PFS: HR, 0.81 [95% CI, 0.69-0.95]; ORR, 26.3% vs 17.5%); no new safety
histochemistry (IHC) whole slide images from the phase II HERB trial were analyzed. These
signals were observed. Here, we present results based on an additional 15 months of participants had unresectable or recurrent BTC refractory or intolerant to gemcitabine-containing
follow-up. Methods: Eligible participants with advanced HCC were randomly assigned 1: regimen and received T-DXd based on confirmed HER2-positive or low status. Lunit SCOPE
1 to receive lenvatinib (8 mg/day if bodyweight [BW] , 60 kg; 12 mg/day if BW $60 kg) universal IHC, a deep learning based IHC analyzer, was used to provide cell level classes (AI-H0,
plus pembrolizumab (200 mg IV Q3W) or lenvatinib plus placebo. Dual primary end H1+, H2+, H3+) and continuous scoring of HER2 staining intensities of subcellular compartments
points were OS and PFS (per RECIST v1.1 by BICR). Secondary end points included ORR (membrane, cytoplasm and nucleus) for each tumor cell. Membrane specificity was calculated for
and DOR, both per RECIST v1.1 by BICR, and safety. The database cutoff was September each cell as the ratio of membrane intensity to the sum of all three subcellular compartments.
24, 2024. Results: 794 participants were randomly assigned to receive lenvatinib plus Results: The 29 patients analyzed showed continuous improvement in response rates with an
pembrolizumab (n = 395) or lenvatinib plus placebo (n = 399). Median study follow-up increasing proportion of AI-H3+ cells. The ORR was 37.5%, 42.9% and 50.0% for patients with
was 59.2 mo (range, 52.9-68.3). The HR for OS was 0.80 (95% CI, 0.69-0.94; median, more than 10%, 25%, and 50% of tumor cells classified as AI-H3+, respectively. The HER2 intense
21.1 months with lenvatinib plus pembrolizumab vs 19.0 months with lenvatinib plus cohort (n=4), defined by tumors with over 50% of tumor cells classified as AI-H3+, had a sig-
nificantly better PFS (HR 0.15, p,0.05) and OS (HR 0.10, p,0.05) compared to the rest of the
placebo). 60-month OS rates were 19.7% with lenvatinib plus pembrolizumab versus
treatment group. The high membrane specificity group defined by $80% of tumor cells with
10.7% with lenvatinib plus placebo. Grade 3-5 treatment-related adverse event (AE) rates membrane specificity $0.4 (N=6) had a confirmed ORR of 50%. These patients also demonstrated
were 62.8% with lenvatinib plus pembrolizumab and 58.0% with lenvatinib plus placebo. significantly longer PFS (HR 0.30, p,0.05) and OS (HR 0.27, p,0.05). The six cases identified by
No additional deaths due to treatment-related AEs were reported since the final analysis membrane specificity included all four cases of the HER2 intense cohort and two more cases,
(database cutoff, June 21, 2022). Overall, 47.3% of participants treated with lenvatinib showing improved sensitivity in identifying likely responders. Conclusions: AI based quantifi-
plus pembrolizumab versus 56.1% of participants treated with lenvatinib plus placebo cation of HER2 intensity and membrane specificity was predictive of therapeutic response to T-
received subsequent systemic therapy (TKI/VEGF, 38.5% vs 42.6%; immunotherapy, DXd in HER2 expressing BTC. Membrane specificity analysis was more sensitive in identifying
17.2% vs 26.8%; chemotherapy, 5.3% vs 4.0%); 21.5% versus 25.6% received subsequent exceptional responders compared to intensity alone. Research Sponsor: None.
liver-directed therapy (locoregional therapy, 20.0% vs 23.8%; surgery, 2.8% vs 2.8%). Confirmed response and survival based on AI defined biomarkers.
Conclusions: The LEAP-002 study did not meet its primary end points; however, with a By AI-H3+ proportion By AI-MB specific cell proportion
long-term follow-up of 5 years, almost twice as many participants randomly assigned to
< 50% ‡ 50% < 80% ‡ 80%
receive lenvatinib plus pembrolizumab versus lenvatinib plus placebo were alive at
database cutoff; no new safety signals were observed. Clinical trial information: Sample size 25 4 23 6
ORR, % 28.0 50.0 26.1 50.0
NCT03713593. Research Sponsor: Eisai, Inc., Nutley, NJ, USA, and Merck Sharp & Dohme mPFS 4.21 (2.83-4.40) 11.04 (5.68-12.91) 4.21 (2.83-4.40) 11.04 (1.45-12.91)
LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. HR (95% CI, p-value) REF 0.15 (0.03-0.67, ,0.05) REF 0.30 (0.10-0.92, ,0.05)
mOS 7.00 (4.37-8.94) NR (5.68-NR) 7.00 (4.27-8.94) NR (9.63-NR)
HR (95% CI, p-value) REF 0.10 (0.01-0.79, ,0.05) REF 0.27 (0.08-0.93, ,0.05)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
294s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4098 Poster Session 4099 Poster Session

A prospective, observational phase II clinical study evaluating hepatic artery Cobolimab and dostarlimab in the first-line treatment of unresectable hep-
infusion chemotherapy in combination with HLX10 and HLX04 as first-line atoma: A multi-center, single arm, phase 2 trial. First Author: Jared David Acoba,
treatment for patients with advanced hepatocellular carcinoma. First Author: University of Hawaii Cancer Center, Honolulu, HI
Huikai Li, Department of Hepatobiliary and Pancreatic Oncology, Tianjin Cancer Hospital Background: TIM-3 is highly expressed in cancer and mediates T cell exhaustion/
Airport, Tianjin, China dysfunction that can be an important mechanism of immune escape. Cobolimab, an anti-
Background: Advanced hepatocellular carcinoma (HCC) presents limited treatment op- TIM-3 monoclonal antibody, in combination with dostarlimab (a PD-1 inhibitor), has been
tions. Immunotherapy has emerged as an effective treatment, demonstrating encouraging shown to enhance T-cell activity in preclinical assessments. This study aimed to in-
outcomes and acceptable adverse reactions in advanced HCC. This study aims to assess vestigate the efficacy and safety of cobolimab and dostarlimab in the treatment of
the efficacy and safety of HLX10 (recombinant anti-PD-1 humanized monoclonal antibody) unresectable hepatocellular carcinoma (HCC). Methods: This is a multi-center, single-
and HLX04 (recombinant anti-VEGF humanized monoclonal antibody), in combination with arm, phase II study. Eligible patients with unresectable hepatoma (Barcelona Clinic Liver
hepatic artery infusion chemotherapy (HAIC), as first-line treatment for advanced HCC. Cancer Stage B or C) and Child Pugh A or B7 (limit 6) liver function received cobolimab
Methods: This prospective, observational, single-center Phase II trial enrolled untreated 300mg and dostalimab 500mg every 3 weeks for up to 2 years. The primary endpoint was
HCC patients with BCLC stage C. All patients received HLX10 (4.5 mg/kg, intravenous the objective response rate (ORR) per Response Evaluation Criteria in Solid Tumors
infusion, every 3 weeks) and HLX04 (15.0 mg/kg, intravenous infusion, every 3 weeks) on version1.1. Secondary endpoints were disease control rate (DCR) and duration of re-
Day 1 of each treatment cycle, followed by HAIC with the FOLFOX regimen. HAIC was sponse (DoR), progression-free survival (PFS), overall survival (OS), and safety.
administered for a maximum of 8 cycles, while HLX10 and HLX04 were continued for up to Results: 40 patients were treated (mean age 67 years [range: 24–85]). The analysis
2 years, until death, disease progression, or intolerable toxicity occurred. The primary presented here is for all 34 Child Pugh A patients. As of November 6, 2024, the median
endpoint was the objective response rate (ORR), assessed by the investigator according to follow-up was 12.9 months. The ORR and DCR were 37.1% and 85.2% respectively (3.7%
RECIST v1.1 criteria. Secondary endpoints included the disease control rate (DCR), CR, 33.3% PR, 48.1% SD). The median PFS was 11.0 mo (95% CI: 4.6-17.4), and the
progression-free survival (PFS), and safety. Results: Between August 2023 and Sep- median OS was 27.3 mo (95% CI: 21.1-33.5). The median DoR was 14.8 mo (95% CI: 9.4-
tember 2024, a total of 35 eligible patients were enrolled in the study. As of the data cut-off, 20.2). Overall incidences of adverse events (AEs) of any grade was 97.1% and immune-
28 (80.0%) patients had received at least 3 cycles of treatment. Of the 35 patients, 32 related AEs (irAEs) of any grade was 64.7%. The most common irAEs of any grade were
underwent at least one assessment of treatment response, with the best outcomes as dermatologic (47.1%) and endocrine (14.7%). There were two patients (5.8%) who
follows: 17 (53.1%) achieved partial remission (PR), and 12 (40.0%) had stable disease
experienced mild elevations of AST and ALT. Grade $3 irAEs were observed in two
(SD). The ORR and DCR were 53.1% and 90.6%, respectively. Notably, among patients who
patients (5.8%), which included decreased neutrophil count, hypophysitis, and hypo-
had received at least 3 cycles of treatment, 17 patients achieved PR, resulting in an ORR of
thyroidism. There were no treatment-related discontinuations or deaths.
63.0%. Moreover, 5 patients underwent successful hepatectomy after at least 3 cycles of
Conclusions: Cobolimab plus dostarlimab yielded promising response rates and sur-
treatment, and postoperative pathological evaluation revealed extensive tumor necrosis in
the excised tissues. The median follow-up duration was 8.4 months, during which 6 (18.8%)
vival outcomes with acceptable safety as first-line treatment in patients with CP A,
patients experienced disease progression, yielding a one-year PFS rate of 70.5% (95% CI: unresectable HCC. This represents a potential therapy regimen for this population.
47.0%–85.0%). In terms of safety, 17 patients (48.6%) experienced at least one grade 3 or 4 Clinical trial information: NCT03680508. Research Sponsor: GSK.
adverse event (AE), with the most frequent being decreased lymphocyte count (20%).
Conclusions: The combination of HLX10, HLX04, and HAIC as first-line treatment for
advanced HCC has demonstrated promising efficacy, particularly in patients completing
three or more cycles. The safety profile of this combination therapy was acceptable, with
manageable AEs. Further investigation in larger trials is warranted. Clinical trial infor-
mation: NCT06370065. Research Sponsor: None.

4100 Poster Session 4101 Poster Session

Sorafenib plus hepatic arterial infusion chemotherapy versus sorafenib plus Survival outcomes for zanidatamab-hrii compared to chemotherapy in
transarterial chemoembolization in advanced hepatocellular carcinoma: An previously treated HER2-positive (IHC3+) biliary tract cancer (BTC):
update on SHATA-001 study. First Author: Zefeng Du, Department of Hepatobiliary HERIZON-BTC-01 vs a real-world (RW) external control arm (ECA). First
Oncology, Sun Yat-Sen University Cancer Center, State Key Laboratory of Oncology in Author: Richard D. Kim, Moffitt Cancer Center, Tampa, FL
South China, Collaborative Innovation Center for Cancer Medicine, Guangzhou, China Background: Zanidatamab-hrii, a bispecific HER2-directed antibody, received accelerated approval for
Background: Although sorafenib plus transcatheter arterial chemoembolization (SoraTACE) adults with previously treated, unresectable or metastatic HER2-positive (IHC 3+) BTC based on results
has been widely applied for advanced hepatocellular carcinoma (HCC) in most Asian from the single-arm phase 2 HERIZON-BTC-01 trial. The results here compare survival outcomes with
zanidatamab-hrii from HERIZON-BTC-01 vs a comparable RW cohort of patients who received second-
countries, sorafenib plus hepatic arteria infusion chemotherapy (SoraHAIC) may be a better
line (2L) chemotherapy. Methods: HERIZON-BTC-01 (NCT04466891) evaluated zanidatamab-hrii (20
alternative. Herein, we compared the efficacy and safety between the two groups in advanced mg/kg IV every 2 weeks) in patients with HER2-amplified, unresectable locally advanced or metastatic
HCC. Additionally, we validated a predictive model from our previous phase III trial BTC (gallbladder cancer [GBC], intra-/extra-hepatic cholangiocarcinoma [ICC/ECC]) who had received
(NCT02973685, cohort1). Methods: This phase III trial (NCT02856126) recruited participants prior gemcitabine-containing therapy. HERIZON-BTC-01 patients with HER2 IHC3+ (n = 62) were in-
with advanced HCC. Eligible participants were randomly assigned (2:1) to receive Sorafenib cluded in this analysis. The US-based ECA was constructed using deidentified data from the Flatiron
(400mg orally twice daily) plus HAIC per 3 weeks or TACE until disease progression or Health Electronic Health Record. Key inclusion criteria for the ECA included a diagnosis of GBC or ICC/
unacceptable toxicity. The Primary endpoint was overall survival (OS). Whole-exome se- ECC with initiation of 2L chemotherapy between 2011-2023, HER2 IHC3+ prior to 2L initiation, and 1L
quencing (WES), RNA sequencing, and DNA methylation analysis of tumor biopsy samples treatment with a gemcitabine-containing regimen. Patients with ECOG . 1 or CNS metastases were
were performed for predictive biomarker exploration (cohort 1) and validation (SHATA-001). excluded. Standardized mortality ratio (SMR) weighting was used to account for potential imbalance of
Results: From August 2016 to October 2020, a total of 207 participants were allocated to key prognostic factors. Overall survival (OS) and progression-free survival (PFS) were evaluated using
SMR-weighted Kaplan-Meier and Cox proportional hazards regression. Results: Among 290 RW patients
receive SoraHAIC (n = 141) or SoraTACE (n = 66). The trial met the prespecified endpoints.
initiating 2L treatment with HER2 testing, 12 met all eligibility criteria and were included in the ECA. The
SoraHAIC significantly prolonged OS compared to SoraTACE (median OS, 15.7 versus most common reason for exclusion was lack of HER2-positivity (n = 209). Median follow-up times were
11.2 months; p , 0.001). In the SoraTACE group, 56.7% of participants experienced grade 3 or 16.1 (interquartile range [IQR]: 9.4, 19.9) and 4.7 (IQR: 2.6, 8.1) months for the zanidatamab-hrii and ECA
4 treatment-related adverse events, which were significantly higher than the SoraHAIC group arms, respectively. After weighting, baseline characteristics were similar across arms with standardized
(39.3%; p , 0.023). Severe adverse events were also more frequent in SoraTACE (15.7% vs mean differences #0.2. The most common 2L chemotherapy regimen in the ECA was FOLFOX (n =6),
26.7%, p = 0.07). WES analysis found no ideal indicator in the mutational landscape for followed by gemcitabine-based regimens (n = 3) and FOLFORI (n = 2). Compared with chemotherapy,
treatment outcome. We identified 1226 and 506 differentially methylated probes (DMPs) zanidatamab-hrii resulted in longer median OS (18.07 vs 3.29 months; hazard ratio [HR]:0.29) and PFS
among the non-responsive specimens in the HAIC and TACE groups. Then we performed RNA- (7.26 vs 2.30; HR: 0.47) (Table). The 6- and 12-month survival for zanidatamab-hrii patients were 90%
seq analysis. By comparing the transcriptome between the responsive and non-responsive and 65% (vs. 29% and 13% for the ECA), respectively. Conclusions: Among patients with previously-
groups, 685 and 600 differentially expressed genes (DEGs) were generated in the two treated HER2-positive (IHC3+) BTC, zanidatamab-hrii resulted in longer survival, including a . 14 month
increase in median OS, vs ECA patients treated with chemotherapy. Research Sponsor: Jazz
treatment groups, respectively. In the HAIC group, pathways including leukocyte-mediated
Pharmaceuticals.
immunity and immune response-activating signaling pathway were enriched in the re-
sponders, while metabolic-related pathways including steroid metabolic process and alcohol Zanidatamab-hrii ECA
(N=62) (N=12)
biosynthetic process were enriched in the non-responders in the TACE group. Integrative
analysis of DEGs and DMPs indicated phosphofructokinase (PFKM) could potentially stratify OS
Median, mos 18.07 3.29
patients to HAIC or TACE as higher expression of PFKM was associated with favorable Survival % (95% CI)
outcomes accepting TACE. Such performance could also be validated in this study as 6 mos 90% (83%, 98%) 29% (11%, 75%)
participants with elevated PFKM tended to benefit from SoraTACE. Conclusions: This trial 12 mos
PFS
65% (54%, 78%) 13% (3%, 55%)
demonstrated SoraHAIC significantly improved OS over SoraTACE in participants with ad- Median, mos 7.26 2.30
vanced HCC. Participants with high PFKM expression benefited more from TACE. Clinical trial Survival % (95% CI)
6 mos 55% (44%, 69%) 14% (4%, 47%)
information: NCT02856126. Research Sponsor: None. 12 mos 32% (22%, 46%) 14% (4%, 47%)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 295s

4102 Poster Session 4103 Poster Session

Concordance analysis between tumor tissue HER2 status by immunohis- The efficacy of atezolizumab plus bevacizumab for advanced hepatocellular
tochemistry (IHC) and in situ hybridization (ISH) and a translational analysis carcinoma in relation to tumor-infiltrating lymphocytes: Histological as-
of plasma ctDNA in patients (pts) with biliary tract cancer (BTC): An ex- sessment of CD8+ T cell spatial features as predictive biomarkers. First
ploratory analysis from the phase 2 HERIZON-BTC-01 trial. First Author: Author: Hiroaki Kanzaki, Department of Gastroenterology, Graduate School of Medicine,
James J. Harding, Memorial Sloan Kettering Cancer Center, New York, NY Chiba University, Chiba, Japan
Background: HER2 is a target for precision oncology. HER2 protein overexpression and/or gene Background: The combination of atezolizumab and bevacizumab (Atez/Bev) has im-
amplification is observed in a subset of pts with BTC. Zanidatamab (zani), a dual HER2-targeted proved prognosis in advanced hepatocellular carcinoma (HCC), though its therapeutic
bispecific antibody, received accelerated approval as a treatment for adults with previously mechanisms remain unclear. While the tumor microenvironment (TME) holds promise
treated, unresectable, or metastatic HER2-positive (IHC 3+) BTC based on the phase 2b HERIZON- for biomarker discovery, current studies rely on archived diagnostic samples. We aimed
BTC-01 trial. In this exploratory analysis, we evaluated concordance between tissue-based HER2
to elucidate molecular determinants of treatment efficacy and identify predictive
IHC and ISH in screened pts, and HER2 gene amplification between tissue-based HER2 ISH and
plasma ctDNA by NGS in zani-treated pts. Methods: In HERIZON-BTC-01 (NCT04466891) pts
biomarkers through comprehensive TME analysis of pre-treatment samples.
were screened for HER2 gene-amplified tumors by ISH (VENTANA HER2 Dual ISH DNA Probe Methods: We analyzed biopsy samples from 94 advanced HCC patients immediately
Cocktail assay) at a central laboratory. Pts with HER2 gene amplification were prospectively before initiating Atez/Bev treatment using immunohistochemistry (IHC), RNA-
assigned to cohorts based on centrally determined HER2 IHC score (Ventana PATHWAY [4B5] IHC sequencing, flow cytometry, and multiplexed imaging. Our analysis focused on spa-
assay); cohort 1: IHC 2+ or 3+; cohort 2: IHC 0 or 1+. HER2 status was assessed in a fresh biopsy or tial characteristics of CD8+ T cells and effector regulatory T (eTreg) cells, with lon-
an archived sample per ASCO/CAP guidelines, with gastric cancer algorithm. Enrolled pts received gitudinal assessment of immune responses during treatment. Results: High
zani 20 mg/kg IV Q2W in 28-day cycles. Plasma ctDNA samples were collected prior to the first programmed death-1 (PD-1) positivity in CD8+ T cells was significantly associated with
cycle of zani (baseline) and on-treatment at cycle 2 day 28 for testing with NGS Guardant360 favorable progression-free survival (PFS) (HR 0.24, 95% CI 0.11-0.52), while CD8+ T cell
(Guardant Health). Guardant360 Molecular Response (MR) scores were calculated based on density showed no significant correlation (HR 1.12, 95% CI 0.64-1.95). Through mul-
changes in plasma ctDNA levels from baseline. Results: Overall, 756 screened pts had central tiplexed imaging analysis using PD-1 positivity as a key indicator, we identified two
results for both HER2 IHC and ISH. Nearly all pts (94%) with HER2 IHC 3+ BTC had HER2-amplified
critical determinants of response: CD8+ T cells localizing within tumor parenchyma
tumors per ISH (Table). Among all 87 pts treated with zani across both cohorts, 48 samples from
25 pts were available for testing with NGS. The concordance between HER2 gene amplification by
rather than fibrous stroma, and maintaining diffuse distribution throughout the tumor
ISH and ctDNA NGS was 59%. Co-mutations at baseline in plasma ctDNA occurring in . 10% of parenchyma. These features, assessable with routine hematoxylin and eosin and CD8
pts included KRAS (n = 1 [3%]), HER2 S310F (n = 3 [9%]), and PIK3CA (n = 6 [18%]). Overall, 18/25 IHC staining, stratified patients into four prognostic groups (p = 0.019), with median PFS
(72%) pts had a decrease in ctDNA levels from baseline at cycle 2 day 28; decreases . 90% were ranging from 14.3 months (both favorable features) to 3.5 months (neither feature). The
observed in pts with a best response of partial or stable disease. MR scores correlated with tumor PD-1 positivity in eTreg cells was not associated with prognosis (HR 0.75, 95% CI 0.34-
response (ANOVA, P = 0.0189). Conclusions: In this analysis, there was a high concordance (94%) 1.62). Notably, bevacizumab counteracted the potential negative effects of programmed
observed between tumor tissue HER2 IHC 3+ status and HER2 gene amplification among pts death-ligand 1 blockade by suppressing eTreg cell activation, as demonstrated through
screened in HERIZON-BTC-01. Exploratory translational analysis shows that treatment with zani analysis of patients who discontinued bevacizumab and in vitro experiments showing
after 2 cycles was associated with a decrease in plasma ctDNA levels in the majority of pts. reduced expression of eTreg activation markers. Conclusions: We demonstrate that
Clinical trial information: NCT04466891. Research Sponsor: Jazz Pharmaceuticals. routine histological assessment of CD8+ T cell localization and distribution patterns can
HER2 IHC and ISH status among screened patients. predict Atez/Bev efficacy in advanced HCC. The identified synergistic mechanism of
ISH Status
Amplification
bevacizumab-mediated eTreg suppression provides a framework for future combination
IHC Status Amplified Non-Amplified Total n/N (%) immunotherapy development. Research Sponsor: CHUGAI PHARMACEUTICAL CO., LTD.
0 12 330 342 12/342 (3.5)
1+ 8 94 102 8/102 (7.8)
2+ 34 167 201 34/201 (16.9)
3+ 104 7 111 104/111 (93.7)
Total 158 598 756 158/756 (20.9)

4104 Poster Session 4105 Poster Session

Radiomic analysis on pretreatment MRI to predict response to atezolizumab Efficacy and safety of atezolizumab and bevacizumab with or without
plus bevacizumab in advanced hepatocellular carcinoma: A multicenter transarterial chemoembolization as first-line therapy for advanced hepato-
study. First Author: Hui-Chuan Sun, Department of Hepatobiliary Surgery and Liver cellular carcinoma: An international multicenter real-world study. First Author:
Transplantation Liver Cancer Institute and Zhongshan Hospital, Fudan University, Ningning Zhang, Department of Lymphoma, Tianjin Medical University Cancer Institute
Shanghai, China and Hospital, National Clinical Research Center for Cancer, Key Laboratory of Cancer
Background: The combination of atezolizumab and bevacizumab is the standard treatment Prevention and Therapy, the Sino-US Center for Lymphoma and Leukemia Research,
for advanced hepatocellular carcinoma approved in China and many other countries. Tianjin, Tianjin, China
However, the objective response rate of this combination treatment was around 30%. Background: Transarterial chemoembolization (TACE) combined with immunotherapy
Consequently, identifying individuals with the potential to respond favorably prior to initiating and targeted therapy provides a promising therapy for advanced hepatocellular car-
therapy remains a pressing challenge. Methods: This multi-center retrospective study in- cinoma (HCC).This study aimed to compare the efficacy and safety of atezolizumab and
cluded advanced hepatocellular carcinoma patients who received atezolizumab plus bev- bevacizumab combined with (TACE-Ate-Bev) or without TACE (Ate-Bev) as first-line
acizumab as first-line therapy between December 2020 and February 2024. The training treatment for advanced HCC. Methods: This international multicenter, retrospective
cohort consisted of eligible patients who have complete baseline, treatment and tumor study included 311 advanced HCC cases administered TACE-Ate-Bev (n = 152) or Ate-
evaluation records and MRI imaging data from Zhongshan Hospital, while eligible patients
Bev (n = 159). Inverse probability of treatment weighting (IPTW) was employed to
from other centers constituted the external validation cohort. A deep learning model based on
minimize bias. Overall survival (OS), progression-free survival (PFS), and adverse events
nnU-Net was developed to automatically segment intrahepatic lesions. All segmentations
were reviewed and revised by two radiologists. The radiomic features were extracted using
(AEs) were observed. Results: The TACE-Ate-Bev group demonstrated significantly
PyRadiomics, then a radiomic feature-based model for predicting response to atezolizumab improved OS (26.8 [95% CI 23.1-NA] vs. 14.9 months [95% CI 11.4-19.9]; hazard ratio
plus bevacizumab therapy was constructed using the Extreme Gradient Boosting Decision [HR] = 2.66 [95% CI 1.87-3.77], p , 0.0001) and PFS (16.0 months [95% CI 12.8-17.8] vs.
Tree (XGBoost) algorithm. Additionally, three radiologists evaluated 53 visually-assessed MRI 6.5 months [95% CI 5.4-7.6]; HR = 2.50 [95% CI 1.90–3.28], p , 0.0001) compared to the
features on MRI scans. Finally, the predictive performance of radiomic feature model, as well Ate-Bev group, especially across BCLC stage B (mOS: NA [95% CI 23.5-NA] vs.
as the relationship between radiomic and MRI features, was assessed. Results: A total of 240 15.6 months [95% CI 11.4-NA], p , 0.0001; mPFS: 16.9 months [95% CI 16.2-NA] vs.
eligible patients were recruited from 14 centers in China, of which 161 and 79 were classified 6.7 months [95% CI 5.7-10.9], p , 0.0001) and BCLC stage C (mOS: 25.2 months [95% CI
as training and validation cohorts, respectively. During a median follow-up period of 19.7-NA] vs. 14.3 months [95% CI 10.1-20.5], p = 0.00018; mPFS: 12.8 months [95% CI
13.7 months (IQR: 8.3–20.6) in the training cohort and 10.5 months (IQR: 7.4–17.2) in the 11.3-17.0] vs. 6.5 months [95% CI 5.0-7.7], p , 0.0001) disease. The superior efficacy of
validation cohort, 19.0% (30/161) and 23.0% (18/79) of patients, respectively, achieved an TACE-Ate-Bev indicated same trends after IPTW adjustment. Grade 3 or 4 AEs were
objective response by RECIST v1.1 (p = 0.559). The radiomic feature model demonstrated a observed in 36 patients (24.3%) in the TACE-Ate-Bev group and 34 (21.4%) in the Ate-Bev
promising predictive performance, achieving an AUC of 0.913 (95% CI: 0.874–0.953) in the group. There was no statistically significant difference in the proportion of gastroin-
training cohort and 0.825 (95% CI: 0.700–0.949) in the validation cohort. Fat surpassing liver testinal bleeding between the TACE-Ate-Bev and Ate-Bev groups (9.9% vs. 10.1%, p =
mass was the only MRI feature associated with an objective response (p = 0.020). When the 0.954). Notably, patients with portal hypertension, portal vein tumor thrombus vp3-4,
MRI feature was combined with radiomic features, the predictive model further improved, extrahepatic metastasis or Child-Pugh grade B exhibited improved OS and PFS in the
yielding an AUC of 0.951 (95% CI: 0.924–0.979) in the training cohort and 0.835 (95% CI: TACE-Ate-Bev group versus the Ate-Bev group. Conclusions: TACE-Ate-Bev signifi-
0.725–0.945) in the validation cohort. A significant correlation was observed between cantly improves OS and PFS with acceptable toxicity compared to Ate-Bev as first-line
radiomic features and MRI features of intrahepatic lesions with a univariate analysis p , 0.2. therapy for advanced HCC. Research Sponsor: None.
Conclusions: Radiomic features derived from pretreatment MRI scans can effectively predict
personalized objective responses to combination therapy with atezolizumab and bev-
acizumab in patients with unresectable or advanced HCC. Research Sponsor: National Natural
Science Foundation of China; 82372037.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
296s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4106 Poster Session 4107 Poster Session

Factors associated with immunotherapy response for hepatocellular Evolutionary divergence of the HLA-B genotype as a predictor of immune
carcinoma. First Author: Matthew Shing Hin Chung, The University of Hong Kong, checkpoint inhibitor (ICI) therapy efficacy in hepatobiliary cancers. First
Hong Kong, Hong Kong Author: Nan Zhang, Department of Liver Surgery, State Key Laboratory of Complex
Background: Immunotherapy has shown remarkable progress in treating hepatocellular Severe and Rare Diseases, Peking Union Medical College Hospital, Chinese Academy of
carcinoma (HCC) in recent years. We aim to investigate the clinical factors associated with Medical Sciences and Peking Union Medical College, Beijing, China
treatment response to HCC immunotherapy in Asian population. Methods: We identifiedHCC Background: The role of human leukocyte antigen class I (HLA-I) molecules in shaping
patients receiving immunotherapy (nivolumab, pembrolizumab, atezolizumab, durvalumab, the immune response to immune checkpoint inhibitors (ICIs) has been recognized in
tremelimumab, ipilimumab; including monotherapies or combinations) from January 2008 to several solid malignancies. However, the prognostic significance of HLA-I character-
June 2024 from a population-based cohort in Hong Kong. Primary outcomes were all-cause istics in hepatobiliary cancers remains poorly understood. Methods: Patients with
mortality and hospitalization stratified by etiology of HCC [Viral-HCC, defined as hepatitis B advanced hepatocellular carcinoma (aHCC) and biliary tract cancer (aBTC) receiving ICI-
virus (HBV)-related or/and hepatitis C virus (HCV)-related HCC vs non-viral-HCC], other clinical
based therapy were prospectively enrolled (NCT03892577). Retrospective analyses were
factors including age, sex, type 2 diabetes (T2D), cirrhosis, antiviral therapy for HBV, and
history of receiving other oncological treatment (curative: surgical resection, liver transplant,
performed to evaluate the association between HLA-I evolutionary divergence (HED),
radiofrequency ablation, microwave ablation; non-curative: transcatheter arterial chemo- HLA-I genotype, and HLA-I heterozygosity with clinical outcomes. HLA-I genotyping was
embolization and radiotherapy). Hazard ratios (HR) were estimated by Cox regression models. conducted using germline DNA from peripheral blood, and HED was quantified based on
Results: This study included 1363 patients on immunotherapy (mean age 63.1 years, 84.2% the Grantham distance metric, which measures evolutionary divergence between HLA-I
male; 87.2% viral-HCC; 23.0% had prior curative therapy; 34.5% had prior non-curative alleles. Patients were stratified into high and low HED groups using the 25th percentile
therapy). Over 240-days of median follow-up, viral-HCC had similar risk of all-cause mor- as the cutoff. Results: A total of 118 patients (41 with aHCC and 77 with aBTC) were
tality (54.5% vs 58.9%, p = 0.169) and hospitalization (34.8% vs 30.9%, p = 0.360) compared to included in the analysis. High HED at the HLA-B genotype was significantly associated
non-viral-HCC. Patients with prior curative therapy compared to those without had lower risk of with improved overall survival (OS) in patients treated with ICIs (p , 0.05). Specifically,
all-cause mortality (HR 0.81 [95% CI: 0.69-0.95], p = 0.011). Among HBV-related HCC, those in aHCC patients, the median OS was 17.43 (95% confidence interval, 17.43-NE) months
with antiviral treatment $2 years prior to immunotherapy were associated with lower risk of in the HLA-B HEDhigh group compared to 7.83 months (95% confidence interval, 4.13-NE)
liver-specific mortality (HR 0.83 [95% CI 0.70-0.99], p = 0.036) and HCC-specific mortality (HR in the HLA-B HEDlow group (p , 0.05). Similarly, in aBTC patients, the median OS was
0.80 [95% CI 0.67-0.96], p = 0.014). Patients with cirrhosis had higher risk of all-cause mortality 12.7 (95% confidence interval, 9.07-20.90) months versus 9.8 (95% confidence interval,
(HR 1.42 [95% CI 1.18-1.71], p , 0.001). No associations with treatment response were 6.40-13.5) months, respectively (p , 0.05). In contrast, HED at the HLA-B genotype did
observed in subgroups stratified by age , 60/ $60, sex, and T2D. Conclusions: Among not exhibit a significant association with progression-free survival. Conclusions: This
patients with HCC receiving immunotherapy, treatment outcomes were similar in viral- study demonstrates that the evolutionary divergence of the HLA-B genotype may serve
etiologies and non-viral etiologies. Antiviral treatment improves treatment response in
as a prognostic biomarker for ICI therapy in hepatobiliary cancers. High HED at HLA-B is
HBV-related HCC, while cirrhosis is detrimental to mortality outcomes. HCC immunotherapy
following curative treatment, in contrast to non-curative therapies, has improved treatment
associated with improved OS, underscoring the potential role of HLA-I genetic diversity
response. Research Sponsor: None. in modulating therapeutic response to ICIs. These findings provide a foundation for
further investigations into HLA-mediated mechanisms underlying ICI efficacy in hep-
All-cause mortality Hospitalization atobiliary cancers. More patients will be enrolled to validate the conclusions. Clinical trial
HR 95% CI P value HR 95% CI P value information: NCT03892577. Research Sponsor: Beijing Natural Science Foundation;
Viral-HCC 0.86 (0.70, 1.06) 0.169 1.14 (0.86, 1.51) 0.360 National High Level Hospital Clinical Research Funding; National Ten-thousand Talent
Age (<60/ ‡60) 0.91 (0.79, 1.06) 0.238 0.90 (0.76, 1.08) 0.276 Program.
Sex 0.98 (0.81, 1.18) 0.812 0.99 (0.78, 1.26) 0.939
T2D 1.00 (0.86, 1.15) 0.957 0.84 (0.70, 1.02) 0.082
Cirrhosis 1.42 (1.18, 1.71) ,0.001 1.23 (0.99, 1.52) 0.062
Curative therapy 0.81 (0.69, 0.95) 0.011 1.09 (0.88, 1.33) 0.434
Non-curative therapy 1.11 (0.96, 1.29) 0.147 1.60 (1.34, 1.92) ,0.001

4108 Poster Session 4109 Poster Session

Hepatic arterial infusion chemotherapy plus lenvatinib and PD-1 inhibitors Molecular characteristics, treatment patterns, and survival outcomes in
as first-line treatment for hepatocellular carcinoma with high tumor burden biliary tract cancers: A retrospective analysis from a high-prevalence
and portal vein tumor thrombus. First Author: Jiaxi Liu, The First Hospital of China region. First Author: Ashish Joshi, MOC Cancer Care & Research Centre, Mumbai, India
Medical University, Department of Interventional Radiology, Shenyang, China Background: Biliary tract cancers (BTC) present significant challenges in management
Background: Advanced-stage hepatocellular carcinoma (HCC) is usually associated with poor sur- due to their aggressive nature and often late-stage presentation. We present demo-
vival outcomes. Rapid tumor control usually benefits long-term outcomes, which could be hardly graphic trends, molecular characteristics, treatment patterns, and survival outcomes.
achieved by solely systematic targeted and immunotherapy in current guidelines. This study aimed to Methods: This retrospective study included patients diagnosed with BTC between
evaluate the efficacy of hepatic arterial infusion chemotherapy (HAIC) combined with lenvatinib and
PD-1 inhibitors (HLP) as first-line treatment for HCC patients with high tumor burden and portal vein
March 2018 and September 2024. Survival analysis was done using the Kaplan-Meier
tumor thrombus (PVTT). Methods: This retrospective multicenter study screened advanced HCC analysis. Results: Total of 743 patients were diagnosed with BTC, median age of the
patients who received HLP combination therapy as first-line treatment at ten centers from Jan 2021 to cohort was 61 years (IQR 52-69); with F: M ratio being 1.15:1. Distribution according to
Dec 2023. The inclusion criteria included high tumor burden (up to seven criteria out), PVTT, no extra- the stage at presentation was: stage I (n=15, 2%), stage II (n=49,6.6%), Stage III (n=160,
hepatic metastasis, and liver function of Child-Pugh B7 or better. PD-1 inhibitors permitted 3 different 21.5%) and stage IV (n=509, 68.5%). History of gallstones was reported in 183 patients
products, including Tislelizumab. Tumor response was assessed using both RECIST 1.1 and mRECIST (24.6%). Out of the total cohort, 436 patients (58.7%) had gallbladder cancers (GBCs)
criteria. Survival outcomes were analyzed using Kaplan-Meier methods. The primary endpoint was while 307 patients (41.3%) had cholangiocarcinoma (CCA), where n=202 (65.8%) were
overall survival (OS), while secondary endpoints comprised progression-free survival (PFS), objective
intrahepatic, n=22 (7.2%) were perihilar and n=33 (10.7%) were distal bile duct. Pre-
response rate (ORR), disease control rate (DCR), as well as depth of response (DpR) within 3 months.
Baseline time point was defined as start of any designed treatment except DpR related survival as dominant histologic patterns among GBCs were: adenocarcinomas n=331 (75.9%), small
3 months after initial treatment. Results: A total of 94 patients were included. The median age was 53 cell carcinomas n=13(3%) adenosquamous n=11(2.5%), and others n=81(18.5%). Mo-
years (range: 31-78) and 87.2% (82/94) were male. In all the patients, 58.5% (55/94) had a largest lecular testing was done in 132 patients (17.8%), out of which 11/132 (8.3%) had HER2
tumor diameter $10 cm, and 80.4% (74/92) had PVTT classified as Vp3 or Vp4. The median follow-up alteration, wherein overexpression was seen in n=3,while amplification was seen in n=8;
time was 13.3 (8.9-25.2) months, the 12-month OS rate was 75.6% (95% CI: 67.0-85.3), and the 24- TP53 mutation was seen in 21/132 (16%), while others alterations were MSI-high (n=2),
month OS rate was 57.6% (95% CI: 46.9-70.7). Median PFS was 9.7 months (95% CI: 8.1-16.0), with 12- TMB-high ($10 mut/Mb) (n=3), PDL1 positive (n=24), FGFR(n=3), IDH (n=2), BRCA2
and 24-month PFS rates of 46.3% (95% CI: 37.1-57.8) and 32.1% (95% CI: 23.3-44.3), respectively. (n=2). Treatment and survival outcomes were available for 492 patients, surgery was
Based on RECIST 1.1 or mRECIST, the ORR was 42.6% (95% CI: 32.4-53.2) or 70.2% (95% CI: 59.9-70.2),
and the DCR was 96.8% (91.0-99.3) or 100% (96.2-100), respectively. Regarding DpR within 3 months,
done in 177 patients (35.9%), systemic therapy (adjuvant/palliative) was done in 431
46.8% (44/94) of the patients had . 25% tumor diameter reduction per RECIST 1.1, while 73.4% (69/ (87.6%) patients while targeted or immunotherapy was used in 99 patients (20.1%).
94) showed . 25% reduction per mRECIST. In subgroup analyses, DpR was significantly correlated Median Overall survival (mOS) of the cohort was 16.1 months (13.6-18.6) with a median
with PFS and OS (see table), while largest tumor diameter, tumor number, and Vp classification had no follow-up duration of 28.2 months (23.2 – 33.2). mOS for GBC was 10.8(8.2-13.3)
correlation with survival outcomes. Conclusions: Combination of HLP as a first-line treatment for months and CCA was11.6(9.3-14.0) months. mOS for Metastatic BTC was 9.6 months
advanced HCC with high tumor burden and PVTT is promising, with high ORR and survival outcomes. while for Non-metastatic it was 20.4 months(p,0.001). The median mOS for patients
DpR might be a predictor for survival. Clinical trial information: NCT06631326. Research Sponsor: with HER2 mutations, TP53 mutations, and PD-L1 positivity were 23.8 months,
None.
16.6 months, and 10.0 months, respectively (p = 0.165). Conclusions: In high-
Outcomes DpR£25% DpR>25% prevalence countries like India, biliary tract cancer (BTC) management is challenged
PFS Events/N 37/50 24/44 by advanced stage presentation, limited molecular testing and resource constraints.
Median (m) 6.8 (5.4-10.9) 16.0 (11.0-NR) Efforts to enhance molecular testing and treatment access are critical for practicing
P 0.004 precision oncology and improving treatment outcomes. Research Sponsor: None.
OS Events/N 24/50 12/44
Median (m) 19.2 (12.6-NR) 35.0 (24.5-NR)
P 0.034

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 297s

4110 Poster Session 4111 Poster Session

Comparison of outcomes between open and minimally invasive hepatic Association of gut microbiota with therapy efficacy and prognosis in biliary
resections for hepatocellular carcinoma: A retrospective analysis. First Au- tract cancer. First Author: Fen Saj, The University of Texas MD Anderson Cancer
thor: Harjeevan Kalra, Independent Researcher, Brampton, ON, Canada Center, Houston, TX
Background: Recent data from the United States indicate that the incidence of he- Background: Biliary tract cancers (BTC) are rising in incidence and have poor prognosis.
patocellular carcinoma (HCC) has tripled over the past 40 years, with risk factors like Microbial exposure via gut–liver axis may contribute to their development and pro-
obesity-related fatty liver disease and chronic hepatitis C on the rise. Despite advances gression. We profiled the gut microbiome of patients with BTC and assessed its cor-
in treatment, the five-year survival rate remains low at 14%. Minimally Invasive Surgery relation with chemoimmunotherapy outcome. Methods: For this prospective study, we
(MIS) has garnered popularity in hepato-pancreato-biliary (HPB) procedures, comprising collected baseline stool samples from newly diagnosed patients with BTC treated at a
1 in 13 cases. MIS offers advantages such as shorter hospital stays and reduced single center. Microbiome profiles were generated using 16S rRNA gene sequencing.
morbidity, with better outcomes. However, the safety and efficacy of MIS, specifically for Microbial alpha (intrasample variability) and beta (interindividual variability) diversity
HCC resection, remain underexplored. To address this gap, we conducted this study to indices were calculated and correlated with treatment outcome. Taxonomical differ-
evaluate and compare the outcomes of MIS vs. open surgical approaches for HCC ential abundance was quantified using MaAsLin2. Statistical analyses included de-
resection. Methods: The 2016–2022 National Inpatient Sample (NIS) was utilized for scriptive statistics for demographics, Kaplan-Meier for survival & Spearman’s correlation
this study. Patients with a primary procedural code for hepatic resection, categorized as for microbiome-clinical outcome. Results: From May 2021 to October 2024, 72 patients
either open or minimally invasive approaches, were identified. This cohort was further were enrolled. Median age was 66 years (range: 29–86) with 43% females. Primary sites
narrowed to include only patients diagnosed with hepatocellular carcinoma (HCC). included intrahepatic cholangiocarcinoma (CCA) (61%), distal extrahepatic CCA (18%),
Differences in common postoperative complications between open and minimally in- hilar CCA (11%), gallbladder Ca (3%), and mixed (7%). Stages were I (14%), II (33%), III
vasive approaches were analyzed. Results: We studied 3,595 hepatic resections for (46%), IV (7%). 52% underwent surgery (14% upfront/others post neoadjuvant), 74%
HCC, with 87.5% performed as open procedures and 12.5% as minimally invasive progressed to stage 4 disease and 67% received immune check point inhibitors. Therapy
surgeries. Both approaches were predominantly conducted in males (70.9% vs. 75.6%, p responses included complete (CR 3%), partial (PR 19%), stable (SD 32%), and progressive
= 0.041), with open resection patients being younger (mean age 63.94 vs. 66.67, p , (PD 28%). At a median follow-up of 20 mos, mPFS was 14.6 (8.3-NR) mos for entire
0.01). Open resections were associated with higher complication rates, including critical cohort [8 (5–11) mos for stage 4 patients] and mOS was not reached. Microbiome
care needs (aOR 6.428, 95% CI 2.606–15.858, p , 0.001), sepsis (aOR 3.728, 95% CI analysis of stool samples provided by immunotherapy-treated patients (n = 26) sug-
1.508–9.215, p = 0.004), acute kidney injury(AKI) (aOR 2.406, 95% CI 1.714–3.337, p , gested that disease control (DCR: CR/PR/SD) correlated with higher alpha diversity (p =
0.001), bleeding (aOR 2.314, 95% CI 1.838–2.914, p , 0.001), and blood transfusion 0.063), while progression showed a trend toward lower diversity across all indices. Beta
requirements (aOR 4.514, 95% CI 2.661–7.658, p , 0.001). Open procedures also diversity (calculated by Bray-Curtis distances) showed no significant differences by
resulted in higher mean hospital charges ($180,561 vs. $131,566, p , 0.01) and more response or progression. Differential abundance analysis using MaAsLin2 identified
extended hospital stays (8.5 vs. 5.3 days, p , 0.01). Despite these differences, mortality significant microbial associations (LogFC . 1.5, adjusted p , 0.25) with treatment
rates between open and minimally invasive approaches were similar (aOR 1.486, 95% CI outcomes. DCR associated bacterial genera included Ruminococcus, Subdoligranulum,
0.85–2.595, p = 0.164). Conclusions: Our study found that open hepatic resection for Romboutsia & Collinsella, while PD correlated with higher abundances of Streptococcus,
HCC is associated with a higher rate of complications, including bleeding, AKI , increased Eggerthella, Paraprevotella & Enterococcus. Genera such as Ruminococcus, Rom-
critical care needs, and blood transfusion requirements, compared to minimally invasive boutsia, Coprococcus & Christensenellaceae R-7 group were linked to non-progression,
surgery. However, mortality rates between the two approaches remain comparable. With whereas Parasutterella, Streptococcus, Prevotella-9 & Monoglobus were elevated in
the rising incidence of HCC, transitioning toward MIS may provide notable benefits in progressive disease. Conclusions: Our preliminary results suggest that gut microbiome
reducing postoperative morbidity. Since most hepatic resections are still performed should be studied as a multidimensional biomarker for predicting therapy response/
using open techniques, further research is necessary to enhance surgical outcomes and prognosis in BTC. Higher microbial alpha diversity may link with better immunotherapy
explore alternative strategies. Research Sponsor: None. responses, and specific microbial taxa may correlate with treatment outcomes.
Research Sponsor: None.

4112 Poster Session 4113 Poster Session

CTC landscape during HCC immunotherapy. First Author: Marcos Deep learning-based contouring of Couinaud segments on CT: Utility for
Santiago Figueroa, University of Chicago, Chicago, IL volumetric analysis of future liver remnant. First Author: Tejas Mathai, National
Background: The treatment of hepatocellular carcinoma (HCC), the third leading cause Institutes of Health, Bethesda, MD
of global cancer death, has significantly improved with the advent of immune checkpoint Background: Hepatocellular carcinoma (HCC) is the most common primary liver tumor,
inhibitor regimens. Circulating tumor cells (CTCs) contain the precursors of metastasis and the liver is a frequent site of metastasis of other cancers. High tumor burden,
and can be serially sampled while patients receive therapy. Quantification of CTC proximity to hepatic vessels, and other comorbidities render only 30% of patients as
dynamics during HCC treatment with immune checkpoint inhibitor therapy may yield candidates for curative treatment: transplantation, resection, or ablation. Surgical re-
early insight into the systemic anticancer response profile. Methods: We used a section requires a 20 - 40% future liver remanent (FLR) to avoid post-operative
commercially available microfluidic CTC purification methodology followed by quan- complications. Delineation of the Couinaud segments is essential for volumetric
tification of CTCs expressing HCC cell surface markers (EPCAM, ASGR1, GPC3) and PD- analysis of FLR and targeted localization of tumors during pre-surgical treatment
L1 from patients receiving immune therapies, prior to cycle 1 and cycle 3 of therapy. We planning. Currently, manual annotation of the Couinaud segments in one CT volume can
also captured and correlated clinical data, including conventional markers of liver take two or more hours, which makes it cumbersome, and can benefit from automation.
function (Child-Pugh score) and radiographic tumor response for correlation with early The purpose of this work is to develop an automated Couinaud segmentation tool for fast
changes in CTC number. Results: In this pilot study,we collected a total of 29 specimens and accurate FLR estimation. Methods: Three CT datasets were used: 1) 161 patients
from 14 participants, 10 of whom provided serial samples. The quantities of CTCs from the public Medical Segmentation Decathlon (MSD) Hepatic Vessels dataset, 2) 43
detected was (median:10, range: 1-516); (median:1, range: 0-139) were PD-L1 high,(- patients with cirrhosis and metabolic diseases having ascites and splenomegaly imaged
median: 8, range: 1-312) were PD-L1 medium, and (median: 3, range: 0-190) were PD-L1 at the National Institutes of Health (NIH), and 3) 197 patients in the public TCIA Co-
low. There was no clear statistical difference in serial CTC enumeration data while on lorectal Liver Metastasis (CRLM) dataset. FLR annotation in the CRLM dataset was done
treatment, although the numbers of CTCs and CTC subsets numerically declined in many by an expert radiologist. The Couinaud segments in the MSD and NIH datasets were
participants. Conclusions: There was no statistically significant relationship between manually annotated by two physicians using ITK-SNAP. The MSD and NIH datasets were
CTC quantity and subsequent radiographic response. Future studies will involve en- used for training, while the CRLM dataset was reserved for testing. A 3D nnU-Net model
rollment of expanded numbers of participants, comparisons with circulating tumor DNA was trained with default hyperparameters to outline the Couinaud segments. On the test
dynamics, and transcriptional profiling of CTCs to further explore these phenomena. dataset, the predicted Couinaud segments were overlaid on the FLR annotation, and
Research Sponsor: UCCCC. metrics, such as Dice Similarity Coefficient (DSC), Hausdorff Distance (HD) error (in
mm), and volume error (in cc) were calculated. The performance was compared to a
previously described 3D U-Net model developed to quantify liver segmental volume ratio
(LSVR) in patients with cirrhosis. Results: The 3D nnU-Net obtained a DSC of 0.99 6
0.01 (IQR: 0.991, 0.998), HD error of 0.87 6 1.83 mm (IQR: 0, 1.02), and volume error of
13.7 6 28.1 cc (IQR: 3.4, 15.3). In contrast, the LSVR U-Net model attained a DSC of 0.97
6 0.01 (IQR: 0.972, 0.984), HD error of 9.56 6 3.61 mm (IQR: 7.14, 11.93), and volume
error of 47.9 6 51.8 cc (IQR: 24.5, 55.4). The 3D nnU-Net model achieved significantly
different results for DSC (p , 0.001, large effect size 0.86), HD error (p , 0.001, large
effect size 0.87), and volume error (p , 0.001, large effect size 0.91). Conclusions: The
model showed acceptable generalizability to the external TCIA CRLM dataset. Future
work may be directed towards accurate volumetric analysis on patients undergoing
portal vein embolization to increase the FLR, and automatic tumor localization on
specific Couinaud segments. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
298s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4114 Poster Session 4115 Poster Session

AI-driven prediction of post-transplant survival and stratification of HCC A phase II study of pevonedistat in combination with carboplatin and
recurrence risk. First Author: Junho Song, Pennsylvania State University College of paclitaxel in advanced intrahepatic cholangiocarcinoma: ECOG-ACRIN
Medicine, Hershey, PA EA2187. First Author: Anita Turk, Indiana University Melvin and Bren Simon Com-
Background: Traditional selection criteria (e.g., Milan, UCSF) and scoring systems prehensive Cancer Center, Indianapolis, IN
(MELD/PELD) for liver transplant eligibility in hepatocellular carcinoma (HCC) often fail Background: Cholangiocarcinoma is a rare and aggressive malignancy with limited
to capture the complex interplay of tumor biology, patient factors, and bridging ther- therapeutic options, particularly in advanced stages. Resistance to first-line chemo-
apies. Although Milan and UCSF yield similar 1-, 3-, and 5-year survival rates, the debate therapy underscores the urgent need for novel therapeutic strategies. Pevonedistat, a
over expanded criteria highlights the need for refined, individualized risk-stratification first-in-class NEDD8-activating enzyme inhibitor, demonstrated promising activity in
tools. Methods: We retrospectively analyzed 21,182 HCC patients from the UNOS this disease in a phase I trial. This study evaluates pevonedistat alone and in com-
database to develop deep learning and Cox regression models for overall survival (OS). bination with carboplatin and paclitaxel in patients with advanced intrahepatic chol-
Models incorporated demographic (e.g., age, race), clinical (e.g., diabetes, MELD dif- angiocarcinoma (ICC). Methods: This was a randomized, non-comparative Phase II trial
ferences), and tumor-specific variables (e.g., tumor count, size tiers). Performance was investigating two treatment arms: pevonedistat monotherapy (Arm A) and pevonedistat
compared to standard MELD-based calculations using 5-fold cross-validation, with combined with carboplatin and paclitaxel (Arm B). Both arms utilized a two-stage
primary endpoints of 1-, 3-, and 5-year survival. For recurrence risk, we used 15,801 minimax design, targeting an objective response rate (ORR) of 30% (null hypothesis:
records to train gradient boosting (XGBoost) and Cox models. Key variables included 10%). Eligible patients had unresectable or metastatic ICC with progression after
tumor characteristics (size levels, vascular invasion), recipient factors (insurance type, gemcitabine-based therapy. The primary endpoint was ORR per RECIST v1.1. Secondary
functional status, initial MELD/PELD), and alpha-fetoprotein (when available). Model endpoints included clinical benefit rate (CBR), progression-free survival (PFS), overall
performance was evaluated via area under the curve (AUC) and concordance index (c- survival (OS), and safety. Toxicities were graded using CTCAE v5.0. Results: A total of
index); external validation was performed for the recurrence model. Results: Cox 40 patients were enrolled, with 34 eligible and treated (17 per arm). Median follow-up
Regression (time-to-event): Final multivariable models achieved c-indices of 0.611 for was 29.3 months. No objective responses were observed in either arm. Stable disease
OS and 0.601 for progression-free survival (PFS). Stepwise Logistic Regression was the best response in 35.3% (n = 12) of patients (11.8% Arm A, 58.8% Arm B). One
(mortality): Mean AUCs were 0.664 (1-year), 0.705 (3-year), and 0.758 (5-year). Random patient on Arm B achieved stable disease lasting $24 weeks, corresponding to a CBR of
Forest Classifier: Slightly higher AUCs than logistic regression (0.663 at 1-year, 0.714 at 5.9% (95% CI: 0.1%-28.7%). Median PFS was 1.54 months (Arm A) and 2.92 months (Arm
3-year, 0.762 at 5-year). Gradient Boosting (recurrence): 1-year recurrence predictions B). Median OS was 4.80 months (Arm A) and 6.54 months (Arm B). Grade 3 or higher
achieved AUC . 0.80, with microvascular invasion emerging as a key risk factor toxicities occurred in 44.1% of patients, with higher incidence in Arm B (70.6% vs. 17.6%
(p,0.001). Across approaches, incorporating multiple clinical and tumor-specific in Arm A). Most common toxicities included fatigue, cytopenias, febrile neutropenia,
factors outperformed MELD-based models, consistently showing improved predictive nausea/vomiting, among others. Two treatment-related fatalities were reported on Arm
accuracy. Conclusions: Machine learning–based models, including deep learning, B: sepsis and colonic perforation. Conclusions: Pevonedistat, alone or in combination
random forests, and gradient boosting, offer enhanced risk prediction for post- with carboplatin and paclitaxel, did not demonstrate sufficient efficacy to warrant
transplant survival and HCC recurrence beyond traditional scoring criteria. These ad- further evaluation in advanced ICC. These findings highlight the challenges in treating
vanced tools enable more nuanced transplant selection, surveillance, and early inter- this aggressive malignancy. Despite the rarity of ICC, the rapid accrual of this study
vention strategies, potentially improving long-term outcomes for HCC patients during a global pandemic indicates the potential for continued exploration of novel
undergoing liver transplantation. Research Sponsor: None. therapeutic approaches in this disease. Clinical trial information: NCT04175912.
Research Sponsor: NCI/NCTN.

4116 Poster Session 4117 Poster Session

Whole-genome sequencing of biliary tract cancer: Uncovering the genomic Real-world validation of the risk estimation of tumor recurrence after trans-
origins of evolutionary trajectories. First Author: Felix Beaudry, Ontario Institute for plant (RETREAT) score: Insights from UNOS data on hepatocellular carci-
Cancer Research, Toronto, ON, Canada noma recurrence after liver transplant. First Author: Donghoon Shin, Department
Background: Biliary tract cancers (BTC) are rare, highly aggressive malignancies with of Internal Medicine, MetroWest Medical Center/Tufts University School of Medicine,
limited treatment options, leading to consistently poor outcomes. An improved un- Framingham, MA
derstanding of BTC tumor evolution could inform enhanced screening strategies, identify Background: For over two decades, established criteria have guided the selection of
useful prognostic markers, and discover novel therapeutic targets. Methods: We liver transplantation (LT) candidates in hepatocellular carcinoma (HCC), yet recurrence
performed whole genome and transcriptome sequencing (WGTS) at high depth ( . 80X) remains a significant clinical challenge with limited treatment options. The Risk Es-
on a prospective cohort of BTC tumors. After detecting mutations and mutational timation of Tumor Recurrence After Transplant (RETREAT) score was developed to
signatures, we developed two novel methods for driver identification and evolutionary address this gap. Further validation is needed to confirm its utility in real-world settings.
reconstruction. First, we created new oncogene and tumor suppressor-specific models Methods: We conducted a validation study using the United Network for Organ Sharing
that integrate copy number profiles, structural variant breakends, and expression (UNOS) database to evaluate the predictive performance of the RETREAT score in adult
changes to distinguish structural drivers from neutral chromosomal rearrangements. liver transplant recipients with HCC from 2009 to 2024. The analysis included 4,975
Second, we applied population genetics techniques to fit demographic models to tumor patients. Kaplan–Meier survival analysis was performed, and recurrence rates were
allele frequencies across 20 paired primary and metastatic samples. Results: We compared across RETREAT score groups using the log-rank test. The predictive ac-
analyzed 130 tumor samples from 110 patients, representing the largest BTC whole- curacy of the RETREAT score for HCC recurrence was assessed by calculating the
genome cohort to date. We identified hypermutated tumors (50-150 mutations/mb) with concordance index (C-index), area under the ROC curve (AUC) evaluating the model’s
distinct etiologies, including mismatch repair deficiency, platinum exposure, tobacco ability to discriminate between recurrence and non-recurrence events. Results: Among
use, and aristolochic acid-related damage, the latter of which was associated with the 4,975 liver transplant recipients with hepatocellular carcinoma (HCC), the distri-
response to immunotherapy. Ourintegrated-driver approach identified an association bution of the RETREAT score was as follows: 13 patients (0.26%) had a score of 0, 1,713
between selection on RAD23A and an increased structural variant load, resulting in a patients (34.43%) had a score of 1, 1,981 patients (39.81%) had a score of 2, 524 patients
tandem-duplicator-like mutational phenotype. This also revealed an underappreciated (10.53%) had a score of 3, 526 patients (10.57%) had a score of 4, and 218 patients
impact of SMAD4 in BTC, which is inactivated through multiple mutation types in 12%. (4.38%) had a score $5. Kaplan–Meier survival analysis demonstrated a significant
Notably, BAP1 mutations occurred in 24% of cases, including 4% that were inactivated association between the RETREAT score and HCC recurrence (p , 2e-16). The hazard
through deletion of the BAP1 promoter that lowered transcript expression, a previously ratio for recurrence with each unit increase in the RETREAT score was 1.685 (95% CI:
undescribed mechanism. By pairing clinical data to the genomics, we observed a co- 1.574 to 1.803). The model’s predictive accuracy, as assessed by the C-index, was 0.697
occurrence of BAP1 mutations and FGFR2 fusions in small-duct, mass-forming intra- (95% CI: 0.668–0.725), and the AUC was 0.684 (95% CI: 0.655–0.713). Conclusions: The
hepatic cholangiocarcinomas. These mutations were mutually exclusive with TP53 RETREAT score demonstrates a C-index of 0.697 in predicting HCC recurrence after liver
mutations, which were enriched in patients with primary sclerosing cholangitis. Finally, transplantation. This model offers valuable risk stratification but could benefit from
our novel method for subclonal population reconstruction on paired samples illuminated further refinement to improve its predictive accuracy. Research Sponsor: None.
recent tumor evolutionary dynamics and identified an ARID1B fusion with a potential role
in metastasis. Conclusions: This study uncovered novel genomic mechanisms un-
derlying the evolutionary origins of BTC beyond those previously identified with exome
and panel sequencing, highlighting the value of WGTS. These results highlight the
complex genomic heterogeneity of BTC, with potential implications for precision
therapy. Research Sponsor: Ontario Institute for Cancer Research; Princess Margaret
Cancer Foundation; Marathon of Hope.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 299s

4118 Poster Session 4119 Poster Session

Conversion therapy for initially unresectable intrahepatic cholangiocarci- Clinico-molecular characteristics and enhanced efficacy of immune check-
noma: A multicenter real-world study. First Author: Xinhao Xiong, Department of point inhibitors in TP53-mutated advanced biliary tract cancer. First Author:
Liver Surgery, Sun Yat-Sen University Cancer Center, Guangzhou, China Taro Shibuki, Department of Hepatobiliary and Pancreatic Oncology, National Cancer
Background: Currently, studies on conversion therapy for intrahepatic chol- Center Hospital East, Kashiwa, Japan
angiocarcinoma (ICC) are relatively limited. Effectiveness and optimization strategies of Background: TP53 mutations are frequent genetic alterations in cancers and also
conversion therapy in real world remain unclear. This study aims to analyze the efficacy prevalent in biliary tract cancer (BTC). In non-small cell lung cancer, TP53 mutations
of locoregional combined with systematic therapy in patients with unresectable ICC, have been reported to enhance immune checkpoint inhibitor (ICI) efficacy and improve
optimizing conversion therapy strategies by retrospectively. Methods: In this multi- outcomes. However, their impact on BTC remains unclear. This study investigated the
center retrospective study, patients with unresectable ICC who received hepatic arterial clinico-molecular characteristics, prognosis, and ICI efficacy in TP53-mutated BTC.
infusion chemotherapy combined with ICIs and target therapy were reviewed. Overall Methods: Patient data were collected from the SCRUM-Japan GOZILA and MONSTAR-
survival (OS), progression-free survival (PFS) , objective response rate (ORR) , disease SCREEN-1/2 databases (Japan) and the Duke university and Mayo clinic databases
control rate(DCR)(assessed by mRECIST criteria) and safety were analyzed. (U.S.), which included patients receiving first-line ICI therapy. Comparisons between
Results: Between June 2018 to May 2024,a total of 341 patients from six centers were TP53 wild-type (WT) and mutation groups were performed in the Japanese cohort.
included in this study. The median follow-up time was 21.7 [Link] them, 313 Subsequently, analyses were expanded to include U.S. patients, focusing on compar-
patients with complete imaging evaluations are eligible for efficacy analysis. Of which, isons between non-ICI and ICI groups. Tumor Immune Dysfunction and Exclusion (TIDE)
20 patients achieved complete response (CR), 93 patients partial response (PR), 161 scores were calculated using MONSTAR-SCREEN-2 whole transcriptome sequencing
patients stable disease (SD), illustrating an ORR of 36.1% and a DCR of 87.5%. Among (WTS) data to evaluate ICI responsiveness, and gene set enrichment analysis (GSEA)
them, 65 patients successfully underwent conversion surgery, with a median duration of explored pathways associated with TP53 mutations. Results: In the Japanese cohort
conversion therapy of 4.1 months. Radical resection was achieved in 39 patients within (n=594), TP53 mutations were identified in 311 patients (52.4%). KRAS, ERBB2, CDKN2A,
the surgical cohort, and 3 patients demonstrated pathological CR. Systemic therapy and SMAD4 alterations were more frequent in the TP53 mutation group, while IDH1,
regimens with the highest conversion success rates included apatinib plus toripalimab BAP1, and FGFR2 fusions were more common in the WT group. Multivariable analysis
(42.9%), lenvatinib plus tislelizumab (32.4%), and apatinib plus tislelizumab (30.8%). showed TP53 mutations were independent prognostic factors for poor outcomes
Patients with tumor diameters less than 10 cm and those with unilobar tumor were more (progression-free survival [PFS]: HR 1.37, 95% CI: 1.11–1.69, P=0.003; overall survival
likely to achieve successful conversion. After propensity score matching (PSM) to [OS]: HR 1.36, 95% CI: 1.06–1.75, P=0.017). Incorporating the U.S. cohort (n=625; non-
balance baseline characteristics (ratio = 1, caliper = 0.01), the surgical group exhibited ICI: 548; ICI: 77), no significant differences in overall response rate (ORR) or disease
better survival outcomes compared to the non-surgical group (not reached vs. control rate (DCR) were observed in the non-ICI group between the groups. However, in
27.2 months, hazard ratio: 0.36, 95% CI: 0.61–0.81, p= 0.001). The median PFS was the ICI group, TP53 mutations tended to show higher ORR (21.1% vs. 38.5%, P=0.16) and
13.0 months in the surgical group versus 8.0 months in the non-surgical group (hazard DCR (47.4% vs. 66.7%, P=0.14). In the non-ICI group, TP53 mutations were associated
ratio, 0.57; 95% CI, 0.30 to 1.09; p= 0.088).Within the surgical cohort, the 2-year survival with worse PFS (HR 1.46, median PFS: 7.4 vs. 5.3 months, P,0.001), whereas, in the ICI
rates for patients with palliative surgery and radical resection were 51.3% and 26.9%, group, no significant difference was observed; however, a trend toward better PFS was
respectively (p= 0.051). All patients experienced treatment-related adverse events seen in the TP53 mutation group (median PFS: 5.0 vs. 6.4 months, P=0.163). TIDE scores
(TRAEs), but no treatment-related deaths occurred. The incidence of Grade 3-4 TRAEs were significantly lower in the TP53 mutation group, indicating higher ICI respon-
was 28.6%. Conclusions: Locoregional combined with systemic therapy is an effective siveness. GSEA revealed enrichment of oxidative phosphorylation and hypoxia path-
strategy for conversion treatment. Conversion surgery offers substantial survival ways. Conclusions: TP53 mutations are independent prognostic factors for poor
benefits, while palliative surgery may also serve as a viable therapeutic option. Research outcomes in BTC. However, patients with TP53 mutations may derive greater benefits
Sponsor: None. from ICIs, underscoring the importance of tailored therapeutic strategies in TP53-
mutated BTC. Research Sponsor: None.

4120 Poster Session 4121 Poster Session

Genomic and outcome analysis of recurrent versus de novo metastatic First results of an open-label, single arm phase II trial investigating the
pancreatic ductal adenocarcinoma (PDAC) receiving systemic therapy: Re- efficacy and safety of trifluridine/tipiracil combined with irinotecan as a
sults from the Australian MoST and CaSP screening programs. First Author: second line therapy in patients with cholangiocarcinoma. First Author: Linde
Frank Po-Yen Lin, NHMRC Clinical Trials Centre, The University of Sydney, Sydney, NSW, Kehmann, Charité University Medicine Berlin, Berlin, Germany
Australia Background: Cholangiocarcinoma (CCA) is a rare and aggressive malignancy with poor
Background: The genomic and prognostic characteristics of recurrent (R, from curative prognosis. Although firstline therapy with gemcitabine, cisplatin and durvalumab has been
surgery) v de novo (D) presentation in metastatic PDAC patients (pt) requiring systemic established based on the TOPAZ-1 trial, treatment options for subsequnt therapies remain
therapy (rx) remain underexplored. We examined the difference in genomic alterations (alts), limited. The TRITICC study investigated the combination of trifluridine/tipiracil (FTD/TPI)
overall survival (OS), and outcome to matched rx according to disease presentation (DP). and irinotecan in patients with disease progression after firstline treatment.
Methods: The PDAC cohorts from the Australian Molecular Screening and Therapeutics Methods: TRITICC was a phase IIA, interventional, prospective, open-label, non-
(MoST) and Cancer Screening Programs (CaSP) completed sequencing from 2016 to July 2024 randomized, exploratory, multicenter, single-arm trial. Adult patients with histologically
were analysed; archival tissue was sequenced using genomic profiling platforms (mainly confirmed, locally advanced or metastatic biliary tract cancer received FTD/TPI (25 mg/m²
TSO500 and FoundationOne CDx). Frequencies of genomic alts were compared between R and BSA, BID, orally, days 1–5 of each 14-day cycle) combined with irinotecan (180 mg/m² on
D groups; significance was assessed using Chi-square tests with the Benjamini-Hochberg day 1 of each cycle). Progression free survival (PFS) was the primary endpoint. Secondary
method (q , 0.05). OS was calculated from the start of initial rx using Kaplan-Meier method, endpoints included the PFS rate at 4 months, median overall survival (OS), objective
with hazard ratios (HR) from Cox regression used for comparison. OS outcomes were stratified response rate (ORR), and quality of life. The trial was registered with [Link]
by matched versus unmatched rx in pts harbouring genomic alts within clinically actionable (NCT04059562) and EudraCT (2018-002936-26). Results: 28 patients were enrolled
Tier 1-3 categories per TOPOGRAPH knowledge base criteria. Results: 949 pts with PDAC with
across six sites in Germany. The median PFS was 3.1 months (95% CI: 2.0–7.5), with PFS
valid genomic results across both programs were included. The KRAS alts were numerically
rates of 35% (95% CI: 21%–58%) at 4 months, 30% (95% CI: 17%–54%) at 6 months, and
more frequent in D (n = 434/491, 88%, p = 0.02) than R (n = 380/458, 82%) groups. Both
13% (95% CI: 4.7%–37%) at 12 months. PFS and PFS rates were higher in patients with
CDKN2A and SMAD4 alts were enriched in the D group (CDKN2A, D: 248, 51% v R: 157, 34%;
SMAD4, D: 139, 29% v R: 85, 19%, p , 0.001 both). Among the 821 pts who started systemic rx, intrahepatic CCA (iCCA). The OS rates were 74% (95% CI: 59%–93%) at 6 months and 38%
those with R PDAC (n = 411) showed longer median OS compared to those with D PDAC (n = (95% CI: 22%–68%) at 12 months. Similar to the PFS rates the OS rates were higher in iCCA.
410, 18.9 v 13.0 months, mo; HR 0.59, 95% CI 0.49-0.69, p , 0.001). Genomic CDKN2A alts Among 27 evaluable patients, partial responses were observed in 3 patients (11.1%, 2 iCCA,
were associated with worse OS (median 13.4 v 16.5 mo, HR 1.34, 95% CI 1.14-1.58); most 1 eCCA), stable disease in 11 patients (40.7%, 7 iCCA, 4 eCCA). Thirteen patients (48.1%, 7
favourable prognosis was seen in 267 pts with CDKN2A wildtype in the R group (median 22.1 iCCA, 6 eCCA) experienced disease progression. Outcomes due to tumor location will have
mo, 95% CI 17.4-25.9). After adjusting for both CDKN2A and SMAD4 alts, R group remained to be examined in higher patient numbers. The most frequently reported adverse events
associated with a lower risk of death than D group (HR 0.61, 95% CI: 0.51-0.72, p , 0.001). Pts included neutropenia, thrombocytopenia, gastrointestinal symptoms and fatigue. The
who received active matched rx (n = 23) showed longer OS than those who received unmatched mean Global Health Status score (EORTC QLQ-C30) declined moderately from 56.2 at
rx (n = 314) in both D (30.1 v unmatched 14.0 mo) and R groups (34.6 v 24.1 mo). The screening to 46.4 at the end of treatment. No severe, unmanageable or unexpected
prevalence of specific KRAS mutations showed no significant differences between D and R toxicities were reported. Conclusions: The combination of FTD/TPI and irinotecan
groups, including G12D (D: 192, 39% v R: 167, 36%, p = 0.44), G12V (D: 126, 26% v R: 113, 25%, p demonstrated promising efficacy and manageable safety in CCA patients progressing after
= 0.78), G12R (D: 59, 12% v R: 50, 11%, p = 0.67), G12C (D: 6, 1% v R: 10, 2%, p = 0.37), and Q61 first-line gemcitabine-based therapy, aligning with recent findings in comparable pop-
mutations (D: 37, 8% v R: 26, 6%, p = 0.31). There were no differences in alts in TP53, ARID1A, ulations (e.g., Tella et al.). Further investigations are warranted to validate these results.
BRCA1/2, or other DNA repair pathway genes. Conclusions: Genomic and prognostic dif- The NIFTY trial suggested that pegylated irinotecan may enhance efficacy compared to
ferences were seen in metastatic PDAC according to presentation, with CDKN2A alts enriched conventional irinotecan, a hypothesis that will be explored in the planned follow-up
in de novo cases and associated with poor OS, emphasising the need to consider stratification TRITICC-2 study. The study was funded by Servier Deutschland GmbH and Servier Affaires
of DP in trials and observational studies. Research Sponsor: None. Médicale. Clinical trial information: NCT04059562. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
300s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4122 Poster Session 4123 Poster Session

Sequential or up-front triple combination with durvalumab, tremelimumab, Comparative efficacy of cabozantinib or lenvatinib following atezolizumab
and bevacizumab for patients with unresectable hepatocellular carcinoma plus bevacizumzb in patients with advanced hepatocellular cancer. First
(AIO-MONTBLANC): Safety interim analysis. First Author: Najib Ben Khaled, Author: Sepideh Mehravar, Cedars-Sinai Medical Center, West Hollywood, CA
Department of Medicine II, University Hospital, LMU Munich, Munich, Germany Background: Atezolizumab plus bevacizumab is the most commonly utilized first-line
Background: The combination of immune checkpoint inhibitors (ICI) durvalumab therapy for advanced HCC. However, the optimal choice for second-line treatment
(durva) and tremelimumab (treme) has been approved for first-line (1L) therapy for remains uncertain, particularly between different tyrosine kinase inhibitors including
advanced hepatocellular carcinoma (aHCC) and represents an alternative to combined lenvatinib and cabozantinib. This study aims to compare survival outcomes of patients
atezolizumab and bevacizumab (bev). The MONTBLANC trial evaluates the efficacy and with metastatic HCC who received either lenvatinib or cabozantinib as second-line
safety of combined durva, treme and bev in patients (pts) with aHCC (NCT05844046). therapy following frontline treatment with atezolizumab plus bevacizumab.
Methods: This investigator-initiated, international, randomized phase 2 trial is the first Methods: This retrospective cohort study utilized the TriNetX database, a large-scale
to investigate the combination of durva, treme and bev. 70 pts with aHCC not amenable platform providing access to deidentified electronic medical records from over 130
to curative treatment or locoregional therapy and preserved (Child-Pugh A) liver function million patients across 94 healthcare organizations in the United States. The study
are randomized in a 1:1 ratio to an early escalation arm (A) or triple treatment arm (B). included patients diagnosed with metastatic HCC (ICD-10 codes: C22, C22.8) who were
Pts in arm A initially receive durva+treme with the addition of bev upon detection of aged 18 years or older, diagnosed between January 2019 and October 2024, and had
disease progression or failure to achieving objective radiological response. Pts in arm B received first-line chemotherapy with atezolizumab and bevacizumab. The cohort was
receive upfront durva, treme and bev. Durva, treme and bev are given in standard doses. divided into two groups: one receiving lenvatinib and the other cabozantinib as second-
The primary endpoint is overall response rate. Secondary endpoints include overall line therapy. The primary outcome was overall survival (OS), compared between the two
survival, progression-free survival and safety. We present the data of the second planned treatment groups using 1:1 propensity score matching (PSM) to balance baseline
safety interim analysis. Results: 25 pts (arm A: 14, arm B: 11) were included in this demographics and comorbidities. Kaplan-Meier survival analysis, log-rank tests, and
analysis (04/2023 – 08/2024). Patients in arm A had a lower proportion of ECOG 0 (A: hazard ratios (HR) were employed to assess differences in OS. Results: Between
78.6% vs B: 90.9%), higher proportion of Child Pugh A6 (A: 21.4% vs B: 9.1%), BCLC C (A: January 2019 and October 2024, a total of 552 patients met the study criteria, including
78.6% vs B: 63.6%), macrovascular invasion (A: 42.9% vs B: 36.4%) and AFP $400 ng/ 397 patients who received Lenvatinib and 155 patients who received cabozantinib as
mL (A: 42.9% vs B: 18.2%). Most adverse events (AE) were grade (G) 1 or 2. In arm A, second-line therapy. Median age was similar in two groups (65.2 vs 65.9 yrs). There were
there were 17 G3 AE, 1 G4 AE and 4 G5 AE. In arm B, 4 G3 AE and 1 G5 AE occurred. 18 no differences between the two groups with respect to gender, albumin or bilirubin
serious AE (SAE) occurred in arm A and 2 SAE in arm B. There was one treatment-related levels. Lenvatinib group had lower proportions of white patients (44.1% vs 55.3%). After
death in arm A (ICI hepatitis) and none in arm B. In the first 6 months on treatment, there PSM (matched on age, race, gender, albumin and bilirubin), the final analysis included
were no significant changes in ALBI score, Child Pugh score or ECOG performance status 151 patients in each treatment group. Kaplan-Meier survival curves demonstrated no
in both arms. Conclusions: This planned safety interim analysis did not reveal signals of statistically significant difference in OS between the two groups (median OS: 16.6 vs
unmanageable toxicity or deteriorating liver function through the addition of bev to 10.9 months; HR: 0.809; 95%CI: 0.576-1.136; p=0.597). Conclusions: This retrospective
durva/treme in the 1L treatment of pts with aHCC. Clinical trial information: cohort study found no significant difference in OS in pts with metastatic HCC treated
NCT05844046. Research Sponsor: Astra Zeneca. with lenvatinib or cabozantinib as second-line therapy following atezolizumab plus
bevacizumab. These findings suggest that both agents may offer comparable survival
benefits, highlighting the need for further prospective studies to refine second-line
treatment strategies for advanced HCC. Research Sponsor: None.

4124 Poster Session 4125 Poster Session

Molecular profile of hepatocellular carcinoma (HCC) in older (OA) versus Development and validation of a prognostic risk score for hepatocellular
younger adults (YA) receiving tyrosine kinase inhibitors: Does age matter? carcinoma recurrence post–liver transplant: Insights from the UNOS
First Author: Jay Parekh, UT Health San Antonio, San Antonio, TX database. First Author: Donghoon Shin, Department of Internal Medicine, MetroW-
Background: While age is not an independent risk factor for poor outcomes it can have est Medical Center/Tufts University School of Medicine, Framingham, MA
significant influence on outcomes in HCC. The median age of diagnosis is 65 years, and Background: For over two decades, the criteria for liver transplantation (LT) in he-
HCC is associated with significant geriatric comorbidities. Further, multi-kinase in- patocellular carcinoma (HCC) have been well-established, yet recurrence remains a
hibitors (MKIs) are associated with up to 60% of grade 3-4 toxicities. We aim to analyze major clinical challenge. This recurrence contributes to inferior post-LT survival in HCC
survival association with age in HCC and identify molecular markers, associated with patients compared to those without HCC. Prognostic index could serve as a valuable tool
survival in older patients with HCC receiving tyrosine kinase inhibitors (TKIs). to identify patients who may benefit from adjuvant therapies and guide standardized
Methods: 1473 HCC specimens with DNA/RNA sequencing were profiled at Caris Life post-LT HCC surveillance, which currently varies across transplant centers.
Sciences. The study cohort was stratified based on median age into two groups: OA: Methods: We developed and validated a predictive model using the United Network for
age . 65 and YA: age , = 65. Real-world overall survival (OS) information was obtained Organ Sharing (UNOS) database, analyzing adult liver transplant recipients with he-
from insurance claims data, and Kaplan–Meier estimates of OS were calculated from patocellular carcinoma (HCC) from 2009 to 2024, including 4,970 patients. Univariable
specimen collection to last clinical contact; and MKI time on treatment (TOT) from the analysis identified variables associated with 1-year, 3-year, and 5-year post-transplant
initiation to termination of treatment. Hazard ratios (HR) and p-values were calculated HCC recurrence, applying a strict p-value threshold ( , 0.01). Significant variables were
using the Cox proportional hazards model and the log-rank test, respectively. selected for multivariable logistic regression to build the model. Internal validation was
Results: Median OS (mOS) among patients with OA was 14.8 months(m) vs 17.1m performed for each time point using Receiver Operating Characteristic (ROC) curve
among YA (HR: 1.18, p , 0.01). The difference in mOS was even more pronounced analysis and confusion matrix evaluation, with the best-performing model selected
among White OA (HR:1.43, p , 0.0001) and Asian/Pacific Islanders (HR:1.62, p = 0.084). based on these metrics. Results: The most significant model was derived using 3-year
Interestingly, although not statistically significant, OA was associated with longer mOS recurrence as the outcome. The final model included the pre-transplant Model for End-
in Black/African Americans (HR:0.73, p = 0.082). OA was not associated with MKI-TOT Stage Liver Disease (MELD) score (p = 0.02), worst tumor histology grade (p , 0.001),
such as sorafenib or cabozantinib. However, OA was associated with a shorter TOT (HR and total tumor diameter (p = 0.03). The final logistic regression equation is as follows:
1.6, p , 0.01) on lenvatinib (len). No molecular alterations were statistically significantly log(1-p/p) = 23.9630 2 0.0621 3 Initial MELD score + 0.7657 3 Worst tumor histology
different between the two age groups on len. CTNNB1, TP53, CDKN2A mutations and grade + 0.1084 3 Total tumor diameter. Internal validation results showed an AUC of
PDL1+ were among the most differentially altered and were more common in OA. 0.761 (95% CI: 0.718 - 0.804), accuracy of 0.769 (95% CI: 0.757 - 0.7807), sensitivity of
DNAJB1-PRKACA fusions were prevalent only in YA (13%, potentially representative of 77.2%, and specificity of 65.0% for 1-year recurrence. For 3-year recurrence, the model
Fibrolamellar HCC), while SLC45A2-AMACR fusions were more prevalent among OA (6.6 demonstrated an AUC of 0.733 (95% CI: 0.702 - 0.763), accuracy of 0.6696 (95% CI:
vs 1.1%). Multivariate analysis revealed that OA was independently associated with 0.6563 - 0.6827), sensitivity of 66.7%, and specificity of 70.9%. For 5-year recurrence, the
shorter len-TOT (HR 1.6, p 0.01), while CTNNB1 mutations and DNAJB1-PRKACA fusions AUC was 0.714 (95% CI: 0.685 - 0.743), with accuracy of 0.655 (95% CI: 0.641 - 0.668),
were independently associated with longer len-TOT (HR 0.4-0.5, both p , 0.05). sensitivity of 65.2%, and specificity of 68.6%. Conclusions: We have developed and
Conclusions: OA was associated with differing survival trends between whites and validated a predictive model for HCC recurrence following LT using UNOS data. This
Asian/PI compared to Black/AA. Further, OA was associated with shorter len-TOT, model shows promising performance, particularly for predicting 1-year recurrence, and
potentially due to anti-angiogenic toxicity. Our limitations include an inability to in- may potentially serve as a useful tool for guiding post-transplant management and
vestigate race-based differences on len-TOT due to small sample sizes (Black/AA: n = 18, surveillance strategies. Research Sponsor: None.
Asian/PI: n = 15) and the lack of toxicity and geriatric assessment data. Future studies
including race, geriatric assessments and toxicity profiles should be considered to
understand survival and tolerability differences. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 301s

4126 Poster Session 4127 Poster Session

Transarterial chemoembolization plus donafenib and immune checkpoint Hepatitis B virus reactivation in hepatocellular carcinoma patients under-
inhibitors for intermediate hepatocellular carcinoma (CHANCE2410): A pro- going immune checkpoint inhibitor and concurrent antiviral prophylaxis
pensity score matching analysis. First Author: Bin-Yan Zhong, Department of agents: A prospective observational study. First Author: Zhicheng Lai,
Interventional Radiology, Zhejiang Key Laboratory of Imaging and Interventional Department of Hepatobiliary Oncology, Sun Yat-Sen University Cancer Center, State Key
Medicine, Zhejiang Cancer Hospital, Hangzhou Institute of Medicine (HIM), Chinese Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer
Academy of Sciences, Hangzhou, China Medicine, Guangzhou, China
Background: Hepatocellular carcinoma (HCC) has high incidence and mortality, with over Background: Immune checkpoint inhibitors (ICIs) have been recommended for the treatment of
80% of patients diagnosed at intermediate or advanced stages, limiting surgical options and advanced hepatocellular carcinoma (HCC). However, due to the potential hazard of hepatitis B
worsening prognosis. Transarterial chemoembolization (TACE) is the standard treatment for virus (HBV) reactivation, all ICI-related phase 3 studies had strict restrictions on HBV-DNA load
intermediate HCC, inducing tumor ischemia and hypoxia, which alters the immune mi- (e.g., , 500IU/ml). Meanwhile, delayed immunotherapy may lead to poor prognosis for patients
croenvironment and promotes immune activation. Combining TACE with immune check- with high HBV-DNA load. This study aims to compare the HBV reactivation between HCC
point inhibitors (ICIs) and tyrosine kinase inhibitors (TKIs) has shown promise in enhancing patients with low or high HBV-DNA load undergoing ICIs and concurrent antiviral prophylaxis
treatment efficacy. Trials like EMERALD-1 and LEAP-012, along with real-world studies, agents. Methods: This prospective, observational study (NCT04680598) recruited HCC par-
suggest that TACE plus ICIs and TKIs improves progression-free survival (PFS) in inter- ticipants who were consecutive hepatitis B surface antigen (HBsAg)-positive and received
mediate HCC patients compared to TACE monotherapy. Donafenib, an oral TKI, has shown concurrent antiviral prophylaxis agents and initial ICIs. Participants were divided into low group
superior overall survival compared to sorafenib, being recommended as first-line treatment (HBV-DNA#500 IU/ml) and high group (HBV-DNA . 500 IU/ml) according to the baseline HBV-
for advanced HCC in China. However, large real-world studies on the combination of TACE DNA level. HCC patients without ICIs from NCT02973685 were also included for analysis. The
primary endpoint was the incidence of HBV reactivation. The secondary endpoints included HBV
plus donafenib and ICIs in intermediate HCC are scarce. This study aims to compare the
reactivation-associated hepatitis, ICIs disruption, overall survival (OS) and progression-free
efficacy and safety between the combination therapy and the TACE monotherapy for in-
survival (PFS). Results: Between December 25, 2020 and February 23, 2024, a total of 1015
termediate HCC in a real-world setting. Methods: This nationwide, multicenter, retro-
participants were enrolled: 356 in the low group and 659 in the high group. The median age was
spective cohort study included patients with intermediate HCC receiving either combination 51 years (range, 18-84) with majority being males (89.2%) and hepatitis Be antigen (HBeAg)
therapy or TACE monotherapy between January 2021 and May 2024 in China. The primary positive (18.1%). Most participants did not receive previous anticancer treatment (84.5%).
outcome was PFS. The secondary outcomes included overall survival (OS) rate, objective Participants in the high group were present with significantly higher HBeAg rate (7.0% vs 24.1%,
response rate (ORR) and safety. Tumor response was evaluated according to the mRECIST p, 0.001), higher ALBI grade 2-3 rate (33.7% vs 49.9%, p, 0.001), larger tumor size (9.3 vs
criteria. 1:1 propensity score matching (PSM) analysis was employed to minimize bias. Cox 10.9 cm, p, 0.001), more advanced BCLC stage C (72.5% vs 83.3%, p, 0.001). A significantly
proportional-hazards regression model was used to analyze factors affecting PFS and OS. higher proportion of participants in the low group had previously received antiviral prophylaxis
Results: A total of 364 patients were enrolled, with 192 receiving combination therapy and agents (16.3% vs 3.6%, p, 0.001). The HBV reactivation rate was 4.5% in the low group and 6.1%
172 receiving TACE monotherapy. After PSM, 127 patients from each group were included in the high group (relative risk, 1.24; 95% confidence internal [CI], 0.81-1.89, p= 0.30). The
for analysis. The median PFS were significantly longer in the combination therapy group frequencies of HBV reactivation-associated hepatitis were 1.7% and 2.3%, respectively (p=
than it in the TACE monotherapy group (19.6 months [95% CI, 14.9-24.4] vs. 15.3 months 0.53). There were 92 participants (25.8%) in the low group and 201 participants (30.5%) in the
[95% CI, 12.8-17.8], HR 0.647 [95% CI, 0.464–0.903], p = 0.010). The OS rate was higher in the high group had interrupted the ICIs treatment (p= 0.12). Compared with high group, the low
combination therapy group (94.8% vs. 83.5%, 1-year OS rate; 76.4% vs. 64.8%, 2-year OS group had shown significantly longer OS (29.8 vs 18.5 months, p= 0.0057) and PFS (9.1 vs
rate; HR 0.542 [95% CI, 0.327–0.989], p = 0.016). The ORR was also higher in the com- 8.3 months, p= 0.043). However, participants in the high group had worse liver function and
bination therapy group (78.9% vs. 62.5%, p = 0.002). Grade 3 or 4 adverse events from any higher tumor burden compared with those in the low group, and the HBV-DNA group was not the
cause were observed at a rate of 12.5% and 5.5% in the combination and monotherapy independent risk factor for OS or PFS in the multivariable analysis. After included patients from
groups, respectively. Multivariate analysis identified combination therapy as an independent NCT02973685 (n = 278), the HBV reactivation rate was 5.5% and 4.3% in patients treated with or
prognostic factor for both longer PFS and OS. Conclusions: Compared to TACE mono- without ICIs (p = 0.43). Conclusions: High HBV-DNA did not significantly increase the incidence
therapy, TACE plus donafenib and ICIs offers superior OS and PFS, which may be a viable of HBV reactivation in HCC patients treated with ICIs and concurrent antiviral prophylaxis.
first-line treatment option for intermediate HCC. Research Sponsor: None. Clinical trial information: NCT04680598. Research Sponsor: None.

4128 Poster Session 4130 Poster Session

Comparative analysis of ctDNA-MRD and MVI in predicting postoperative Real-world analysis of ctDNA and other biomarkers in patients with cura-
recurrence of hepatocellular carcinoma. First Author: Jianan Feng, Second De- tively resected stage I-III biliary tract cancer. First Author: Maen Abdelrahim,
partment of Hepatobiliary Surgery, Zhujiang Hospital of Southern Medical University, Houston Methodist Neal Cancer Center, Houston, TX
Guangzhou, China Background: Growing evidence supports the prognostic and predictive value of circulating
Background: Microvascular invasion (MVI) is currently recognized as a pathological tumor DNA (ctDNA) detection in gastrointestinal cancers. Building on previous work that
feature strongly associated with HCC recurrence. However, its predictive performance in demonstrated the feasibility of tumor-informed ctDNA testing in biliary tract cancer (BTC),
clinical practice remains suboptimal. The detection of minimal residual disease (MRD), this study aimed to evaluate ctDNA as a tool for detecting molecular residual disease (MRD)
which has demonstrated significant prognostic value across multiple cancer types, following curative resection and monitor recurrence during surveillance. Methods: A ret-
represents a promising alternative. Therefore, the first goal of this study is to detail the rospective analysis of real-world data was performed on patients (N=171) with stage I-III
genomic alterations that potentially drive MVI or MRD, the second goal aims to compare resectable BTC who underwent ctDNA analysis using a personalized, tumor-informed 16-plex
their predictive power for postoperative recurrence in early-stage HCC patients. mPCR-NGS assay (Signatera; Natera, Inc.) from July 2020-February 2024. Plasma samples
Methods: This study profiles the genomic landscape of tumor samples from 126 BCLC 0/ (n=769) were collected pre-operatively, postsurgically (within 2 to 12-weeks; MRD window),
A/B stage HCC patients using WES. Patients’ longitudinal MRD status was determined and longitudinally until death or last follow-up (surveillance window). The prognostic value of
through ctDNA detection in peripheral blood at preoperative, 1-, 4-, and 7-month post- ctDNA was compared to traditional biomarkers such as CA19-9 and CEA. Results: A total of
operative time points using a tumor-agnostic fixed panel. Postoperative recurrence, 171 patients with stage I-III BTC with a median age of 68 years (range 27-92) were included in
confirmed by radiographic imaging, served as the endpoint events for comparing the this analysis. The median follow-up was 21 months (range: 2-97 months). ctDNA detection
sensitivity and specificity of two prediction models, with one model based on MVI rates during the MRD and surveillance windows were 22% (18/83) and 32% (35/109), re-
classification and the other on longitudinal MRD positivity. More patients are currently spectively. On evaluating clinical outcomes, ctDNA-positivity during MRD and surveillance
being recruited and analyzed. Results: Among 126 enrolled HCC patients, 71 (56.3%) were was significantly associated with inferior disease-free survival (DFS) and overall survival (OS).
Multivariate analysis confirmed ctDNA-positivity to be the most significant prognostic factor
classified as MVI-positive (M1/M2), while 55 (43.7%) were MVI-negative (M0). Of the 121
associated with DFS (HR: 10.91, 95%CI: 3.85-30.9, P,0.001) when adjusted for other
patients with at least one MRD test, longitudinal MRD positivity was observed in 27
clinicopathologic factors such as BTC subtype or tumor grade. Additionally, other biomarkers
patients (22.3%), with 94 patients (77.7%) remaining negative. Both MVI positivity and
such as CA 19-9 and CEA did not predict clinical outcomes at either the MRD or surveillance
longitudinal MRD positivity were associated with postoperative recurrence, with longi-
windows (Table). Conclusions: The data show that ctDNA-positivity was associated with
tudinal MRD group showing stronger statistical significance (p , 0.001 vs. p = 0.032).
poor DFS and OS, both in the post-op and surveillance settings and that ctDNA detection
However, no distinct high-frequency mutation patterns were observed within either the MVI using a personalized, mPCR-NGS assay was superior to current clinical biomarkers. These
or longitudinal MRD groups. The most frequently altered genes in both groups included findings highlight the value of ctDNA monitoring to improve prognostication in BTC. Research
TP53, CTNNB1, JAK1, ARID1A, CDKN2A, and AXIN1. Of note, we also developed two Sponsor: None.
prediction models based on MVI classification and longitudinal MRD positivity. The
longitudinal MRD-based model demonstrated superior performance, with higher AUROC Association of biomarkers with clinical outcomes.
(0.835 vs. 0.715), PPV (0.357 vs. 0.162), accuracy (0.836 vs. 0.512), and TPR (0.833 vs. Biomarker MRD Surveillance
0.786). Conclusions: Longitudinal MRD monitoring using a tumor-agnostic fixed panel ctDNA n= 83 n= 109
demonstrates superior predictive performance over MVI classification for postoperative DFS HR: 13.0, p , 0.001 HR: 6.1, p , 0.001
OS HR: 12.0, p , 0.001 HR: 17.8, p = 0.008
HCC recurrence with higher sensitivity, specificity, and accuracy. The findings highlight the CA19-9 n= 53 n= 80
potential of longitudinal MRD monitoring as a more reliable tool for guiding personalized DFS HR: 0.88, p =0.81 HR: 1.3, p = 0.51
postoperative management in early-stage HCC [Link], the combination of the OS HR: 1.9, p = 0.389 HR: 10.1, p = 0.045
two models may potentially offer superior predictive performance for recurrence, which is CEA n= 18 n= 18
DFS HR: 0.84, p = 0.85 HR: 0.54, p = 0.5
also a promising direction worthy of further exploration. Research Sponsor: None. OS HR: 1.7, p = 0.71 HR: Not evaluable

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
302s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4131 Poster Session 4132 Poster Session

Integration of cfDNA fragmentomics for early biliary tract cancer detection. ADJUBIL: A phase II study of immunotherapy with durvalumab and trem-
First Author: Jiwen Wang, Liver Cancer Institute, Zhongshan Hospital, Fudan University, elimumab in combination with capecitabine or without capecitabine in
Shanghai, China adjuvant situation for biliary tract cancer—The IKF/AIO-ADJUBIL trial.
Background: Biliary Tract Cancer (BTC) is a highly aggressive malignancy with poor survival First Author: Thorsten Oliver Goetze, Krankenhaus Nordwest GmbH, Institut für Kli-
outcomes, primarily due to the lack of effective early detection methods and late-stage nisch-Onkologische Forschung (IKF),UCT - Universitäres Centrum für Tumor-
diagnoses. Current diagnostic tools, including imaging and invasive endoscopic procedures, erkrankungen Frankfurt, Frankfurt Am Main, Germany
are limited in their sensitivity and specificity for identifying early-stage disease. This study Background: Patients (pts) with biliary tract cancer (BTC) still have a poor outcome with
addresses this critical gap by developing a novel, non-invasive approach for BTC detection limited effective treatment options (only 20% of pts eligible for surgically curative resection, 5-
using circulating cell-free DNA (cfDNA) fragmentomics features. Methods: The study year OS rates , 10%). SOC for BTC is treatment with capecitabine according to the UK BILCAP
cohort included 163 patients diagnosed with BTC and 165 healthy individuals, divided trial, even though it was formally negative. Based on positive data (TOPAZ-1 and MediTreme
equally into training and validation cohorts. All participants’ plasma samples were collected trial in BTC, HIMALAYA trial for the STRIDE regimen in HCC), IO combination in the adjuvant
for a low-depth whole genome sequencing (WGS) process to extract three key cfDNA setting seems promising. In preclinical studies – particularly in cholangiocarcinoma (CC)
fragmentomics features: Copy Number Variation (CNV), Fragment Size Distribution (FSD), –antibody combinations showed stronger and more durable anti-tumor effects than mono-
and Promoter Fragmentation Entropy (PFE). These features were utilized to develop a therapy, due to synergistic impact on the tumor’s immunosuppressive microenvironment. The
machine learning model, which was trained and validated through 5-fold cross-validation. ADJUBIL trial aimed at evaluating the clinical activity of the anti-PD-L1 antibody durvalumab
An external cohort of 55 patients with benign diseases and 18 Tis/High-grade cases was and the anti-CTLA-4 antibody tremelimumab with or w/o capecitabine in pts with resectable
used to further evaluate the model robustness. Results: The stacked ensemble model BTC in the adjuvant setting in a pick-the-winner design. The winner of ADJUBIL could be tested
reached an Area Under the Curve (AUC) of 0.96 in the validation cohort, showing excellent in a follow-up phase 2/3 trial against the current SOC capecitabine. Methods: In the open-
performance in identifying BTC from healthy participants. At an 86% training specificity label, multicenter phase II ADJUBIL trial treatment-naı̈ve pts with BTC after curative surgery
cutoff, sensitivity achieved 90.91% (95% CI: 81.26% - 96.59%) and specificity 87.88% (95% (R0/R1) were randomized (1:1) to receive either tremelimumab (300 mg, once on D1, cycle 1)
CI: 77.51% - 94.62%). While PFE performed as a strong single feature with an AUC exceeding plus durvalumab (1500 mg, Q4W; max. 12 months), with (arm A) or w/o (arm B) capecitabine
0.92. The model demonstrated its effectiveness in early-stage detection, with the sensitivity (1250 mg/m2 twice a day on day 1 – 14, Q3W; max. 8 cycles). Primary endpoint was
increasing from 80% in stage I to 95.65% in stage II. The model surpassed traditional recurrence-free survival at 12 months (RFS@12). The trial design is based on the Simon, Wittes
biomarkers (AUC . 95% compared to ~75% for CA19-9) and demonstrated consistent and Ellenberg’s Pick-the-winner design [Simon et al., 1985]. Results: 40 pts (ECOG 0 or 1) were
performance across subgroups. External validation revealed 89% sensitivity for early lesions enrolled in 12 centers in Germany: median age of 64.5 years; 53% males, 30% intra-hepatic CC,
and 89% specificity for benign cases, highlighting its potential for non-invasive early de- 58% extra-hepatic CC, 13% gallbladder. All pts received at least 1 dose of study treatment. The
tection of BTC. Conclusions: This study demonstrates a reliable and non-invasive strategy median number of cycles was 7. RFS@12 was 52.4% for arm A and 57.9% for arm B. After a
for early BTC detection, leveraging cfDNA fragmentomics features and a robust machine median follow up of 13.8 months, median recurrence free survival was 14.98 (A) and
17.02 months (B). 1y OS rate was 85% (A) and 84% (B). While no new safety/toxicity signs were
learning framework. The model’s high accuracy and reproducibility in both internal and
observed, arm A demonstrated a higher toxicity rate than arm B: 67% of pts having at least one
external cohorts highlight its potential for clinical implementation, offering a transformative
grade $ 3 AE (A) vs. 53% (B) and 48% of pts having at least one grade $ 3 treatment related AE
approach for BTC screening. Early diagnosis enabled by this method may significantly
(A) vs. 32% (B). Conclusions: In the IKF/AIO-ADJUBIL trial, the expected RFS@12 of 56% was
improve patient outcomes and survival rates, marking a major advancement in clinical
demonstrated for the combination of durvalumab / tremelimumab without capecitabine
practice. Research Sponsor: Natural Science Foundation of Shanghai; 23ZR1459100, (57.9%), whereas no benefit in terms of RSF@12 was observed with additional capecitabine
22ZR1457900; National Natural Science Foundation of China; 82272772,82372832, (52.4%). Together with similar 1y OS rates of 85% (A) and 84% (B) and higher toxicity rates in
82273289; Key Discipline Construction Project of Medicine in Shanghai Xuhui District; arm A, this indicates superiority of the combination of durvalumab / tremelimumab without
SHXHZDXK202304; Research Projects from the Science and Technology Commission of capecitabine in pts with resectable BTC in the adjuvant setting. Clinical trial information: EU CT
Shanghai Municipality; Grants 21JC1401202; Fujian Provincial Natural Science Foundation No.: 2024-511847-24-00. Research Sponsor: AstraZeneca.
of China; 2022J05328.

4133 Poster Session 4134 Poster Session

Novel early-detection model based on cfDNA methylation and fragmenta- The efficacy and safety of donafenib as postoperative adjuvant therapy in
tion features for liver cancer. First Author: Xiaobo Wang, Guangxi Medical University patients at high risk of recurrence following radical resection of hepatocel-
Cancer Hospital, Nanning, China lular carcinoma (HCC): A multicenter retrospective study. First Author: Jianhua
Background: In China, the 5-year survival rate of liver cancer patients is only 14%, far Rao, Hepatobiliary Center, The First Affiliated Hospital Nanjing Medical University,
lower than the average of 43.7% for all cancer types. Early diagnosis and treatment are Nanjing, Jiangsu, China
essential for survival. Traditional screening methods like AFP combined with abdominal Background: Hepatectomy is a crucial treatment for long-term survival in patients with
ultrasound have low sensitivity. Recent studies suggest that blood cell-free DNA (cfDNA) HCC; however, the high recurrence rate significantly impacts prognosis. Currently, there is
characteristics could be a new screening approach for liver cancer. This study aims to no standard adjuvant therapy for this patient population. This study investigates the
compare methylation and fragmentation signals among liver cancer, hepatitis, cirrhosis efficacy and safety of Donafenib as postoperative adjuvant therapy for patients with a high
patients, and healthy individuals, innovatively using these signals to construct an early- risk of recurrence after radical resection of HCC. Methods: We analyzed the clinico-
detection model which could improve patient prognosis. Methods: From July 2023 to pathological data of HCC patients with a high risk of recurrence after radical resection,
November 2024, 315 blood samples were prospectively collected from five Chinese recruited from six medical centers between June 2021 and October 2024. High risk was
hospitals. The sample set included 105 liver cancer patients and 210 non-liver cancer defined by the presence of any of the following criteria: [i] tumor diameter . 5 cm; [ii]
controls (33 hepatitis, 30 cirrhosis, 147 healthy). This multi-center, multi-disease- multiple lesions of any size; [iii] microvascular invasion (MVI) grade 1 or 2; [iv] lesions
controlled collection provides a robust data basis. Targeted enzymatic methyl se- complicated by tumor thrombus (TT); and [v] alpha-fetoprotein (AFP)$200 mg/L. Patients
quencing detected over 600,000 methylation sites, enabling precise exploration of liver- received either Donafenib monotherapy (D) or combination regimens (D+TACE-DT, D+ICI-
cancer-related methylation. Beyond methylation, novel fragmentation features like DI, or D+TACE+ICI-DTI) as adjuvant therapy. We examined the relapse-free survival (RFS),
break-point motifs, end motifs, arm-level count, fragment-size distribution and ratio overall survival (OS), and safety according to CTCAE 5.0. Results: 199 patients were
were obtained. These, combined with methylation data, offer a multi-dimensional view included in this study, with a median age of 60 years (IQR: 53-67) at the data cut-off in
for studying liver cancer pathogenesis and biomarkers. A gradient-boosted tree model, January 2025. The cohort comprised 85.7% males, 83.4% with HBV infection, 87.9% with
integrating 3840 methylation DMR features and fragmentomic model-predicted prob- Child-Pugh A, and 79.4% with ECOG PS 0. Among patients at high risk, 52.3% had multiple
abilities, was built. A nested cross-validation framework was used to optimize the model high-risk factors, 53.3% had tumors . 5 cm, 27.1% had multiple lesions, 53.3% had MVI
and ensure result accuracy. Results: The model achieved a high AUC of 0.97(95%CI: grade 1 or 2, 13.6% had TT, and 36.7% had AFP$200 mg/L. Treatment distribution included
70 patients receiving D therapy, 69 receiving DT therapy, 46 receiving DI therapy, and 14
0.95-0.99) in liver cancer detection. At 96.2% specificity, the model had a 91.4%
receiving DTI therapy. At the data cut-off, the median RFS for the overall population was
sensitivity for overall liver cancer detection, with 83.7% and 95.8% sensitivity for stage I
27.8 months (95%CI:22.3 months -NE), with a one-year RFS rate of 72.9% (95%CI:66.0%-
and II respectively. Among 63 patients with hepatitis or cirrhosis, the model accurately
80.5%) and two-years RFS rate of 55.4% (95%CI: 45.6%-67.2%). In subpopulations based
predicted negative results in 88.9% of patients. Notably, for patients hard to identify by
on treatment regimens (D, DT, DI, DTI), the median RFS was 24.5 months (95%CI:
traditional tumor markers like AFP and DCP, the model showed high detection rates. 20.6 months -NE), 29.2 months (95%CI:22.3 months -NE), 30.0 months (95%CI:
When AFP , 400 ng/ml, the detection rate was 88.9%, and with concurrent DCP , 40 19.0 months -NE), and 20.5 months (95%CI:9.3 months -NE), respectively. The median OS
ng/ml, it reached 87.0%. When AFP , 20 ng/ml, the detection rate was 89.5%, and with for the overall population and subpopulations had not yet been achieved. Among the
DCP , 40 ng/ml simultaneously, the detection rate was 83.3%. Conclusions: This study overall population, 114 patients (57.3%) experienced treatment-related adverse events
established an early-detection model for liver cancer by leveraging cfDNA methylation (TRAE) of any grade, with an incidence of grade 3 TRAE at 8.0% concluding with rash
and fragmentation signals. The model demonstrated remarkable performance, par- (5.5%), hand-foot syndrome (2.0%) and thrombocytopenia (0.5%); no patients experienced
ticularly in detecting liver cancer patients who are difficult to identify through con- grade 4 or 5 TRAE. Conclusions: These preliminary results showed that donafenib as
ventional methods. It blazes a new trail for the early detection of liver cancer and could postoperative adjuvant therapy may effectively reduce the recurrence rate in patients at
significantly enhance patient prognosis. Research Sponsor: None. high risk of recurrence following radical resection of HCC, demonstrating good safety and
tolerability. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 303s

4135 Poster Session 4136 Poster Session

Effect of the combination of systemic and locoregional therapy on tumor Ultra-sensitive detection of hepatocellular carcinoma (HCC) with methyla-
recurrence and survival after liver transplantation for hepatocellular carci- tion signal enrichment of ctDNA and hepatitis B virus (HBV). First Author: Nan
noma (HCC). First Author: Khadyoth Nanneboyina, Methodist Dallas Medical Center, Lin, The Third Affiliated Hospital, Sun Yat-sen University, Guangzhou, China
Dallas, TX Background: Aberrant methylation patterns in cell-free DNA (cfDNA) have been
Background: Liver Transplantation (LT) results in the best survival in select HCC pa- identified as effective biomarkers for HCC early detection, with circulating tumor DNA
tients. Patients that do not meet transplant criteria based on tumor burden may require (ctDNA) from HCC patients exhibiting distinct methylation signatures. Additionally, HBV
bridging therapies to downstage their disease to become eligible for LT. There is limited infection and the associated methylation alterations are closely linked to the devel-
data on safety and efficacy of combining systemic therapy with locoregional therapy opment and progression of both cirrhosis and HCC. In this study, we utilize an ultra-
(LRT) prior to LT. Methods: This study was a single-center retrospective outcome sensitive Methylation Anchor Probe for Low Signal Enrichment (MAPLE) to enrich HCC-
analysis of all patients diagnosed with HCC who underwent LT between June 2018 and related methylation signals in ctDNA, as well as those from HBV genomes. By integrating
March 2024 (n = 104) with primary endpoints being 1-year post-LT survival and post-LT these signals with a machine learning model, we achieve improved discrimination
tumor recurrence. Explant pathology was also examined to assess tumor necrosis, between HCC patients and non-cancer controls, while reducing false positives in in-
viability, grade, and lymphovascular invasion. Patients were categorized into 2 groups: dividuals with cirrhosis. Methods: Whole blood samples were collected from 246
1) LRT alone and 2) combination of LRT and systemic therapy. LRT included Trans- participants, including 96 HCC patients, 123 healthy controls, and 27 cirrhosis indi-
arterial chemoembolization, Radioembolization, Microwave Ablation, and Stereotactic viduals. cfDNA was extracted from plasma, followed by enzymatic conversion and library
Beam Radiation Therapy. Systemic therapies included Nivolumab + Ipilimumab, Ate- preparation. Targeted hybrid capture was performed using a custom-designed panel that
zolizumab + Bevacizumab, Sorafenib, Lenvatinib, Ramuricumab, and Cabozantinib. enriched methylation signals associated with HCC and HBV CpG islands. The final li-
Pearson correlation analysis was used. Results: 89 patients received LRT alone and 15 braries were sequenced using next-generation sequencing (NGS). A machine learning
patients received combination therapy. The median maximum tumor diameter in the LRT model was developed, incorporating methylation features derived from both the human
group was 2.4 cm and that of the combination group was 2.5 cm (p = 0.136). Patients in genomic regions and HBV CpG islands. Participants were randomly divided into training
the combination therapy group also had a 3.5-fold increase in the average number of and test sets at a 3:1 ratio, with the training set undergoing 5-fold cross-validation for
tumors, suggesting higher tumor burden. Average time to post-LT tumor recurrence was model optimization. To assess model robustness, 40 resampling iterations were con-
similar in combination therapy vs. LRT group (496.8 days vs. 546 days; p = 0.41). ducted to evaluate performance in distinguishing HCC patients across various stages
Patients receiving combination therapy had a trend towards better survival however this from non-cancer individuals. Results: Among all participants, 39.8% tested positive for
did not achieve statistical significance (r = 0.13, p = 0.175). A statistically significant HBV. Incorporating methylation features from the HBV genome into the model improved
negative correlation existed between not meeting Milan criteria at the time of transplant sensitivity for detecting early-stage HCC in HBV-positive individuals and enhanced
evaluation and post-transplant tumor recurrence (i.e., rate of post-transplant tumor accuracy in distinguishing early-stage HCC from cirrhosis. Analysis of selected HBV
recurrence increased if the Milan criteria were not satisfied; r = -0.31, p , 0.001). methylation features revealed hypermethylation in HCC patients compared to individuals
Conclusions: Our results show that combination therapy using systemic options in with cirrhosis and healthy controls. The final machine learning model achieved a
addition to LRT is an effective downstaging strategy for high-risk HCC patients and may specificity of 97.6% (96.2%–97.9%). Sensitivities for detecting HCC across all stages
improve post-transplant survival and reduce post-LT tumor recurrence. This preliminary were: I: 76.4% (73.5%–79.4%), II: 94.6% (92.0%–97.3%), III: 99.5% (98.8%–100.0%), and
data suggests that combination therapy may have a favorable impact on the time to IV: 100.0% (100.0%–100.0%). For distinguishing cirrhosis, the model demonstrated a
post-LT tumor recurrence in patients with higher risk tumors. This further attests the specificity of 81.9% (77.6%–86.3%). Conclusions: Using the ultra-sensitive MAPLE
strong need for sustainable downstaging pre transplantation. Using the synergistic technique, we developed a novel panel that enriches methylation signals from both the
effect of these modalities may help expand the pool of candidates who can undergo LT, human and HBV genomes. This assay significantly improved sensitivity for detecting
which remains the most effective long term therapy for patients with HCC. We did not early-stage HCC. By incorporating HBV genome features, we further enhanced the
find any evidence of increased rejection or opportunistic infections post LT in the accuracy of distinguishing early-stage HCC from cirrhosis in HBV-positive individuals.
combination therapy cohort. Research Sponsor: None. Research Sponsor: Shanghai Xiaohe Medical Laboratory Co., Ltd.

4137 Poster Session 4138 Poster Session

Use of artificial intelligence-powered spatial analysis of tumor microenvi- Unveiling the role of sodium glucose cotransporter-2 inhibitors in hepato-
ronment to predict the prognosis in resected gallbladder cancer. First Author: cellular carcinoma patients with cirrhosis: A comparative global cohort
Young Hoon Choi, Department of Medicine, Samsung Medical Center, Sungkyunkwan study. First Author: Asfand Yar Cheema, Cleveland Clinic Foundation/ Fairview Hos-
University School of Medicine, Seoul, South Korea pital, Cleveland, OH
Background: Gallbladder cancer (GBC) is a highly lethal disease with a lack of reliable Background: Hepatocellular carcinoma (HCC) is a leading cause of cancer-related
biomarkers. The tumor microenvironment (TME) is closely associated with prognosis, mortality worldwide, and is often associated with chronic liver disease and cirrhosis.
but its clinical application as a prognostic marker is limited by evaluation challenges. Despite advancements in therapeutic options, the prognosis for HCC patients remains
This study assessed the prognostic significance of AI-powered TME analysis in resected poor. Sodium-glucose cotransporter 2 inhibitors (SGLT2i) exhibit anti-inflammatory, anti-
GBC patients. Methods: A total of 225 GBC patients with an R0 resection were enrolled, fibrotic, and anti-carcinogenic properties, potentially influencing cancer biology via
and their hematoxylin & eosin (H&E)-stained GBC sections were analyzed using Lunit metabolic and oxidative stress modulation. Our study aimed to evaluate the impact of
SCOPE IO, an artificial intelligence (AI)-powered whole-slide image (WSI) analyze, to SGLT2i on outcomes of HCC patients with underlying cirrhosis using a large global da-
evaluate TME-related features, including tumor-infiltrating lymphocyte (TIL) density, tabase. Methods: We conducted a retrospective, propensity score-matched cohort study
fibroblast (FB) density, and tertiary lymphoid structure (TLS) counts. Risk stratification using TriNetX Analytics Network database. We compared patients aged . 18 years with
was based on TME-related risk factors (low TIL, high FB, low TLS), and survival outcomes HCC and cirrhosis who received SGLT2i to those who did not receive SGLT2i from 1/1/2014
were assessed. External validation was conducted using 146 biliary tract cancer pa- to 1/1/2023 for one year. The study cohort included 1,254 cirrhotic HCC patients on
tients. Results: Overall survival (OS) and disease-free survival (DFS) declined as the SGLT2i, while the control cohort comprised 40,820 cirrhotic HCC patients not on SGLT2i.
number of TME-related risk factors increased. Patients with three risk factors had the Propensity score matching was applied to balance demographics, HCC-directed therapies,
poorest outcomes (median OS: 17.7 months [reference]; median DFS: 12.7 months comorbidities, laboratory values, and medications. Kaplan-Meier analysis estimated event-
[reference]), followed by those with two risk factors (median OS: 115.9 months, HR = free survival and overall survival, with comparisons using log-rank tests. The primary
0.40, 95% CI: 0.19–0.85; median DFS: 57.8 months, HR = 0.37, 95% CI: 0.18–0.74) and outcomes were all-cause mortality, venous thromboembolism (VTE), and all-cause hos-
one risk factor (median OS: 126.5 months, HR = 0.34, 95% CI: 0.16–0.74; median DFS: pitalization rates. Secondary outcomes included all-cause ICU admissions, ischemic
117.2 months, HR = 0.30, 95% CI: 0.15–0.62). Patients with no risk factors had the best stroke/TIA, acute kidney injury (AKI), and septic shock. Results: Propensity score
matching adjusted for key characteristics resulted in 1,020 matched pairs for each cohort.
survival (median OS: not reached, HR = 0.20, 95% CI: 0.06–0.67; median DFS: not
Our comparative analysis showed that the SGLT2i group had significantly lower all-cause
reached, HR = 0.13, 95% CI: 0.04–0.41). External validation confirmed consistent trends
mortality, with a hazard ratio (HR) of 0.399 (95% confidence interval [CI] 0.314, 0.507).
across all risk groups. Conclusions: AI-powered TME analysis shows promise as a
Specific outcomes associated with improvement in the SGLT2i group: VTE (HR 0.607, 95%
practical tool for identifying TME-related risk factors using H&E-stained WSI, providing
CI 0.481, 0.765), all-cause hospitalization rates (HR 0.568, 95% CI 0.501, 0.644), all-cause
valuable prognostic information for resected GBC patients. Research Sponsor: None. ICU admissions (HR 0.522, 95% CI 0.396, 0.689), ischemic stroke/TIA (HR 0.585, 95% CI
0.389, 0.878), thrombocytopenia (HR 0.578, 95% CI 0.470, 0.711), AKI (HR 0.708, 95% CI
0.588, 0.852), and septic shock (HR 0.528, 95% CI 0.382, 0.728). Conclusions: Our study
highlights the association of SGLT2i with significant improvements in clinical outcomes
for HCC patients with cirrhosis. SGLT2i use was linked to reduced all-cause mortality, VTE,
hospitalizations, ICU admissions, and complications such as ischemic stroke/TIA,
thrombocytopenia, AKI, and septic shock. These findings suggest SGLT2i may offer
therapeutic benefits beyond their cardiovascular and renal effects, potentially influencing
cancer biology and systemic complications in this high-risk population. Prospective trials
are warranted to validate these findings and assess their safety and efficacy. Research
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
304s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4139 Poster Session 4140 Poster Session

Neoadjuvant transhepatic arterial infusion chemotherapy (HAIC) with FOL- Prognostic significance of pathological response in unresectable hepato-
FOX regime plus cadonilimab (PD-1/CTLA-4 bispecific antibody) for resect- cellular carcinoma treated with immune checkpoint inhibitor-based conver-
able multinodular CNLC Ib/IIa hepatocellular carcinoma (CAR_Hero study). sion therapy. First Author: Wen-Jing Zheng, Department of Liver Surgery &
First Author: Yongguang Wei, First Affiliated Hospital of GuangXi Medical University, Transplantation, Liver Cancer Institute, Zhongshan Hospital, Fudan University, Shanghai,
Nanning, Guangxi, China China
Background: The recurrence rate of hepatocellular carcinoma (HCC) remains high, with Background: Immune checkpoint inhibitor (ICI)-based conversion treatment is increasingly
multinodular HCC being a well-defined high-risk factor for recurrence. However, stan- utilized for patients with initially unresectable hepatocellular carcinoma (HCC). However,
dardized neoadjuvant or adjuvant therapies for HCC have yet to be definitively established standardized histopathologic markers for assessing treatment response and predicting
to effectively improve survival outcomes. Methods: In this ongoing single-center, phase 2, survival outcomes remain inadequately defined. Methods: This retrospective study ana-
open-label, prospective cohort clinical trial, eligible pts were randomly assigned ([Link]) to lyzed 225 HCC patients who underwent conversion treatment followed by curative resection.
three arms (15 pts per arm). Neoadjuvant therapies included: (A) 2 cycles cadonilimab The residual viable tumor percentage (RVT%) was calculated as the proportion of RVT
(6mg/kg Q2W); (B) once FOLFOX- HAIC and 2 cycles cadonilimab; (C) once FOLFOX-HAIC. surface area to the total tumor bed area. Kaplan-Meier and Cox regression analyses were
Pts receive scheduled surgery on day 21-28 and postoperative adjuvant HAIC one month used to evaluate the relationship between RVT% and recurrence-free survival (RFS) as well
after surgery. The primary endpoints were major pathologic response (MPR, defined as overall survival (OS). Results: Complete pathologic response (CPR), defined as 0% RVT,
as #50% residual living tumor) and the 1-year recurrence-free survival (RFS) rate. was achieved in 60 patients (26.7%) and was strongly associated with improved survival
Secondary endpoints included overall response rate (ORR, assessed per RECIST 1.1) and outcomes. Patients with CPR exhibited significantly better RFS (HR: 0.23, 95% confidence
treatment-related adverse events (TRAEs). Additionally, a direct hepatectomy cohort was interval [CI]: 0.13–0.41, p , 0.001) and OS (HR: 0.17, 95% CI: 0.05–0.56, p = 0.003)
retrospectively collected as reference data. Results: A total of 42 pts were enrolled. compared to non-CPR patients. Multivariate analysis confirmed non-CPR as an independent
Among them, 2 pts withdrew due to their desire to pursue conversion therapy. 38 pts risk factor for both RFS (HR: 3.67, 95% CI: 1.94–6.96, p , 0.001) and OS (HR: 4.43, 95% CI:
underwent hepatectomy and were included in the efficacy analyses (A: 14pts, B: 14pts, C: 1.21–16.18, p = 0.022). Major pathologic response (MPR), defined as RVT% # 10%, was
12pts). The median age was 55 years (range: 32-72), with 90.5% being male and 90.5% observed in 91 patients (40.4%) and was also significantly associated with improved RFS
infected with hepatitis B virus. Arm B had the highest MPR rate of 78.6%, significantly and OS (all p , 0.05). Stratification by RVT% thresholds revealed a stepwise association
higher than Arms A (35.7%) and C (20.0%) (P = 0.011). Additionally, Arm B had the highest between decreasing RVT% and improved survival outcomes. RVT% # 30% demonstrated
ORR (A: 14.3%; B: 40.0%; C: 8.3%), and lowest MVI detection rate (A: 50.0%; B: 21.4%; C: significant predictive power for both RFS (HR: 0.48, 95% CI: 0.30–0.75, p = 0.001) and OS
40.0%). Focal heterogeneity was partially observed. The DCR was 100%. The most common (HR: 0.60, 95% CI: 0.31–1.16, p = 0.020). Notably, patients with CPR achieved a two-year
TRAEs were elevated aspartate transaminase (64.3%) and alanine aminotransferase survival rate of 96.7%, compared to 86.1% in non-CPR patients. Despite the high radiological
(59.5%). Grade 3-4 TRAEs occurred in 3 pts(hepatic dysfunction and erythema annulare). response rate (CR+PR, 92.4%), substantial discrepancies were observed between radio-
In Arms A and B, 4 pts experienced a delay in scheduled surgery by 2-4 weeks. The logical and pathological assessments, and 56.7% of patients who achieved pathological
combination of HAIC and cadonilimab did not lead to a significant increase in TRAEs. After complete response (CPR) did not exhibit radiological complete response (CR). Kaplan-Meier
propensity score matching, the direct hepatectomy cohort was screened. Kaplan-Meier analysis showed no significant differences in RFS or OS among three regimens: anti-PD-1
analysis revealed that the neoadjuvant cohort had a longer recurrence-free survival (RFS) monotherapy (n = 25, 11.1%), anti-PD-1/PD-L1 plus anti-VEGF (n = 21, 9.3%), and anti-PD-1
time compared to the direct hepatectomy cohort (median RFS not reached vs. plus TKI (n = 179, 79.6%). Transarterial chemoembolization (TACE) did not significantly
24.7 months; P = 0.0048) and a lower MVI detection rate (36.8% vs. 52.6%). increase the proportion of patients achieving CPR but notably enhanced the proportion of
Conclusions: Neoadjuvant FOLFOX-HAIC combined with cadonilimab had a considerable patients with RVT% # 30%. Conclusions: Both CPR and MPR are robust prognostic markers
antitumor activity, and a manageable safety for the resectable multinodular HCC. of RFS and OS in HCC patients undergoing ICI-based conversion treatment. The superior
It brought the fewer MVIs of tumor and a better RFS. Clinical trial information: sensitivity of pathological evaluation underscores its advantage over radiological as-
ChiCTR3000033692. Research Sponsor: None. sessment in accurately reflecting treatment outcomes. Research Sponsor: National Natural
Science Foundation of China; (82341027 and 82072715).

4141 Poster Session 4142 Poster Session

Comparison outcome of transarterial chemoembolization combined with Association of differential expression of genes with survival and relapse in
immune checkpoint inhibitors plus bevacizumab or lenvatinib as first-line patients treated on the BILCAP clinical trial: Gene expression identification
therapy for advanced hepatocellular carcinoma. First Author: Ningning Zhang, and response to adjuvant chemotherapy in early-stage biliary tract cancer.
Department of Hepatobiliary Oncology, Liver Cancer Center, Tianjin Medical University First Author: Valerie Elizabeth Crolley, UCL-University College London (United Kingdom),
Cancer Institute and Hospital, National Clinical Research Center for Cancer, Key Lab- London, United Kingdom
oratory of Cancer Prevention and Therapy, Tianjin Medical University, Tianjin, China Background: Adjuvant capecitabine is standard of care based on the results of the
Background: Transarterial chemoembolization (TACE) combined with immunotherapy and anti-angiogenic therapy for advanced hepatocellular carcinoma (HCC) presents a promising first-
line treatment option. Methods: We assessed overall survival (OS), progression-free survival (PFS), objective response rate, and adverse events between the TACE-ICI-Len group (n=160) and BILCAP clinical trial, comparing adjuvant capecitabine with observation for early-stage
the TACE-ICI-Bev group (n=216) as first-line therapy for advanced HCC. Inverse probability of treatment weighting was employed to minimize bias. Efficacy was evaluated using RECIST 1.1 and
mRECIST criteria. Results: The TACE-ICI-Bev group demonstrated significantly improved OS and PFS compared to the TACE-ICI-Len group, especially across BCLC-B and BCLC-C stages
biliary tract cancer (BTC). Translational work on data from BILCAP aims to identify
(Total: mOS 22.8 vs. 15.4 months, p,0.001; mPFS 12.4 vs. 8.3 months, p,0.001; BCLC-B: mOS 23.3 vs. 16.6 months, p=0.005; mPFS 14.0 vs. 8.2 months, p,0.0001; BCLC-C: mOS 22.3 vs.
15.1 months, p=0.002; mPFS 11.0 vs. 8.0 months, p,0.001). Within the TACE-Bev-Ate subgroup, OS and PFS were further enhanced (Total: mOS 26.3 months; mPFS 13.8 months; BCLC-B:
differentially expressed genes in patients whose tumours relapsed or who died from their
mOS 27.7 months; mPFS 16.7 months; BCLC-C: mOS 24.2 months; mPFS 12.6 months). The incidence of gastrointestinal bleeding (GB) was significantly higher in the TACE-ICI-Bev group cancer, as there are currently no validated biomarkers to predict the risk of relapse or
compared to the TACE-ICI-Len group (13.8% vs. 6.2%, p,0.001). Notably, GB was significantly more frequent in patients with portal hypertension (PHT) compared to those without, in both the
TACE-ICI-Bev group (30.9% vs. 6.6%, p,0.001) and the TACE-ICI-Len group (20.2% vs. 0%, p,0.001). Conclusions: TACE-ICI-Bev demonstrated superior OS and PFS compared to TACE-ICI- death or response to adjuvant chemotherapy in early-stage BTC. Methods: Bulk RNA
Len as first-line treatment for advanced HCC, with an acceptable safety. Management of PHT in patients with advanced HCC is critical to optimizing patient outcomes. Research Sponsor: None.
sequencing (RNAseq) was performed on archived fixed formalin samples from consented
Patient baseline characteristics before and after applying IPTW.

Before IPTW After IPTW

BILCAP patients. Extracted RNAseq data was quantified using salmon, before undergoing
Variable TACE-ICI-Bev TACE-ICI-Len p TACE-ICI-Bev TACE-ICI-Len p
differential gene expression (DGE) using DESeq2 to identify differentially expressed genes
n N=216 N=160 N=223.64 N=153.48
Age (mean (SD))
Sex (%)
59.21 (9.81) 57.19 (10.81) 0.059
0.358
57.59 (10.00) 57.23 (10.42) 0.780
0.784
in patients who died or whose tumours relapsed, compared to those or survived or had no
Hypertension (%)
Female
Male
37 (17.1)
179 (82.9)
21 (13.1)
139 (86.9)
0.46
33.1 (14.8)
190.6 (85.2)
24.6 (16.0)
128.9 (84.0)
0.928
relapse, accounting for batch effect, different anatomical subtypes and adjuvant
No
Yes
152 (70.4)
64 (29.6)
106 (66.2)
54 (33.8)
150.8 (67.4)
72.9 (32.6)
104.3 (67.9)
49.2 (32.1) treatment. Tumour anatomical subtype was highly associated (p , 0.05) with PC2 during
DM (%) 0.176 0.981
No
Yes
147 (68.1)
69 (31.9)
120 (75.0)
40 (25.0)
164.3 (73.5)
59.4 (26.5)
112.5 (73.3)
40.9 (26.7)
principal component analysis and was included as a co-variate during analysis.
ECOG_PS (%)
0
1
77 (35.6)
139 (64.4)
123 (76.9)
37 (23.1)
,0.001
123.4 (55.2)
100.2 (44.8)
86.1 (56.1)
67.4 (43.9)
0.888
Results: 200 patient samples were analysed; 104 / 200 (52%) patients received che-
TACE_number (%)
1~2 151 (69.9) 131 (81.9)
0.011
169.4 (75.7) 115.6 (75.3)
0.94
motherapy while 96 had observation. 142 / 200 (71%) tumours relapsed and 142 / 200
.=3 65 (30.1) 29 (18.1) 54.3 (24.3) 37.9 (24.7)
Child_Pugh_score (%)
,=6 117 (54.2) 73 (45.6)
0.125
109.7 (49.1) 70.6 (46.0)
0.629 (71%) patients died by the time of data cut off. DGE analysis identified 146 significantly
ALBI_grade (%)
.6 99 (45.8) 87 (54.4)
0.367
113.9 (50.9) 82.9 (54.0)
0.536 (p , 0.01) upregulated genes in patients who died, including PRSS2, AMY2A, SPINK1,
CTRC and CELA2A, with significantly (p , 0.01) downregulated genes including HP,
I 94 (43.5) 78 (48.8) 102.7 (45.9) 64.6 (42.1)
II-III 122 (56.5) 82 (51.2) 120.9 (54.1) 88.9 (57.9)
BCLC_stage (%) 0.451 0.51

Lymphatic_metastasis (%)
B
C
85 (39.4)
131 (60.6)
56 (35.0)
104 (65.0)
0.001
82.9 (37.1)
140.7 (62.9)
50.9 (33.1)
102.6 (66.9)
0.845
PHF14, FLI1, NAV3 and NOVA1. 118 genes were significantly (p , 0.01) upregulated
No
Yes
126 (58.3)
90 (41.7)
65 (40.6)
95 (59.4)
107.1 (47.9)
116.6 (52.1)
75.4 (49.1)
78.1 (50.9) through DGE analysis in patients died or relapsed, including PAX5, FAM188B, NET1, DEK
and WDR1 with significantly (p , 0.01) downregulated genes including RQCD1, CPA1,
Extrahepatic_metastasis (%) 0.11 0.536
No 179 (82.9) 121 (75.6) 172.1 (77.0) 123.2 (80.3)
Yes 37 (17.1) 39 (24.4) 51.5 (23.0) 30.3 (19.7)
Ascites (%)
No
Yes
146 (67.6)
70 (32.4)
99 (61.9)
61 (38.1)
0.298
150.5 (67.3)
73.1 (32.7)
94.8 (61.8)
58.7 (38.2)
0.35
FLI1, NAV3 and NOVA1. 41 genes were significantly (p , 0.01) upregulated in both death
Cirrhosis (%)
No 46 (21.3) 28 (17.5)
0.433
40.6 (18.2) 23.4 (15.2)
0.47
and relapse, including PAX5, TLK1, ALB, HEATR3 and RAB11FIP4. 171 genes were
significantly (p , 0.01) up regulated in patients who died or had their tumours relapse
Yes 170 (78.7) 132 (82.5) 183.0 (81.8) 130.1 (84.8)
PHT (%) 0.425 0.698
No 106 (49.1) 71 (44.4) 111.6 (49.9) 72.8 (47.4)

Etiology_(%)
Yes 110 (50.9) 89 (55.6)
0.113
112.1 (50.1) 80.7 (52.6)
0.378 after adjuvant chemotherapy and not in patients undergoing observation, which included
THBS1, FAM65B, UBE2W, RPLP0P2 and NPM1, and 66 genes were significantly (p , 0.01)
No/other 35(16.2) 16(10) 33.64 (12) 22.18 (14.6)
HBV 181(83.8) 144(90) 190 (88) 131 (85.4)
PVTT_classification_vp (%) 0.367 0.962

AFP_400 (%)
No
VP1-VP4
122 (56.5)
94 (43.5)
82 (51.2)
78 (48.8)
0.266
123.2 (55.1)
100.5 (44.9)
85.0 (55.4)
68.5 (44.6)
0.979
downregulated, which included RP11-521B24.3, FLI1, CACNA1A, KCNH6 and MKKS. Gene
,400
.=400
119 (55.1)
97 (44.9)
78 (48.8)
82 (51.2)
112.8 (50.4)
110.9 (49.6)
77.1 (50.3)
76.3 (49.7) ontology enrichment analysis identified the upregulated genes as being associated with
Number_of_tumor (%) 0.979 0.678
,=3
.3
62 (28.7)
154 (71.3)
47 (29.4)
113 (70.6)
66.0 (29.5)
157.6 (70.5)
41.7 (27.2)
111.7 (72.8)
cytoplasmic translation and the synthesis of both intra- and extracellular RNA and protein
HCC_diameter_5 (%)
,5
.=5
60 (27.8)
156 (72.2)
47 (29.4)
113 (70.6)
0.823
65.8 (29.4)
157.8 (70.6)
47.1 (30.7)
106.4 (69.3)
0.833
complexes. Conclusions: Differential gene expression of the BILCAP cohort identified
NLR_grade (%)
,2.81 120 (55.6) 87 (54.4)
0.902
126.0 (56.4) 85.1 (55.5)
0.884
genes associated with cancer relapse and death, including genes associated with a lack of
.=2.81 96 (44.4) 73 (45.6) 97.6 (43.6) 68.4 (44.5)
PLT (%)
,150 122 (56.5) 75 (46.9)
0.082
112.7 (50.4) 76.2 (49.6)
0.901 response to adjuvant chemotherapy. Research Sponsor: Incyte.
.=150 94 (43.5) 85 (53.1) 110.9 (49.6) 77.3 (50.4)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 305s

4143 Poster Session 4144 Poster Session

Comparative analysis of stereotactic body radiotherapy (SBRT) vs. SBRT Impact of radionuclide therapy on survival outcomes for de novo metastatic
with bridge therapies for hepatocellular carcinoma patients awaiting liver gastroenteropancreatic neuroendocrine tumors: A population-based cohort
transplantation: A multi-center study (2010-2020). First Author: Teja Sureddi, study. First Author: Zhiqiao Liu, The Affiliated Hospital of Southwest Medical University,
Virtua Our Lady of Lourdes, Camden, NJ Luzhou, Sichuan, China
Background: Locoregional therapies, including SBRT, are essential in managing he- Background: De novo metastatic gastroenteropancreatic neuroendocrine tumors
patocellular carcinoma (HCC) patients awaiting liver transplantation. This study eval- (dmGEP-NETs) are difficult to treat without the option of radical surgery. The aim of this
uates outcomes of SBRT alone versus SBRT combined with other bridge therapies over a study was to investigate the survival outcomes of radionuclide therapy.
5-year follow-up. Methods: Data from the TriNetX database (2010–2020) were used to Methods: Patients diagnosed with dmGEP-NETs from the Surveillance, Epidemiology,
compare two matched cohorts: patients receiving SBRT alone and those receiving SBRT and End Results (SEER)-17 registry database (2000–2021) were included in this study.
with additional bridge therapies, including transarterial chemoembolization (TACE) and The impact of radionuclide therapy on overall survival (OS) and cancer-specific survival
radiofrequency ablation (RFA). Patients included were those meeting liver transplant (CSS) was assessed using univariate Kaplan-Meier method and multivariate Cox re-
criteria and classified as AJCC Stage I or II. Results: Before matching, the cohorts gression analysis. Results: From 2010 to 2021, a total of 9657 patients at diagnosis
included 185 patients in the SBRT Alone group and 363 in the SBRT with Other Bridge years with dmGEP-NETs were determined from the SEER database. On the univariate
Therapies group, with significant differences in ethnicity (48.11% vs. 57.3%, p=0.0411) analysis, patients with dmGEP-NETs who received radionuclide therapy significantly
and race (40.54% unknown vs. 31.68%, p=0.0393). After matching (153 patients per obtained better OS (5-year rate: 55.6% vs. 37.9%; p ,0.001) and CSS (5-year rate: 58.2%
group), all variables were balanced, including age (69.4 6 8.58 vs. 69.1 6 8.09, vs. 44.2%; p ,0.001) than patients without radionuclide therapy. Finally, the significant
p=0.9039) and ethnicity (50.98% vs. 45.75%, p=0.3601). Theoverall survival rate at 5- clinical factors associated with OS and CSS were included into multivariate Cox re-
year follow-up was 61.53% in the SBRT alone group and 62.02% in the SBRT with other gression analysis to investigate the independent prognostic value of radionuclide
therapies group. There was no significant difference between the groups (HR: 0.958, therapy. The results showed that radionuclide therapy was an independent prognostic
p=0.8417). In the secondary analysis. For acute hepatic failure, the risk was 22.73% for factor for OS (hazard ratio [HR], 0.530; 95% confidence interval [CI], 0.348-0.808;
SBRT Alone versus 25.64% for SBRT with Other Bridge Therapies (RR: 0.886, 95% CI: p=0.003) and CSS (HR, 0.508; 95%CI, 0.319-0.808; p=0.004) in patients diagnosed with
0.413–1.901, p=0.7567). The risk of decompensated liver disease was 37.74% versus dmGEP-NETs. Besides, radionuclide therapy was associated with better OS (p ,0.001)
40% (RR: 0.909, 95% CI: 0.593–1.501, p=0.8054). Procedure-related complications and CSS (p ,0.001) compared to beam radiation. Conclusions: Radionuclide therapy
occurred in 11.61% of the SBRT Alone group compared to 12.12% in the other group (RR: was associated with improved survival in patients with de novo metastatic gastro-
0.958, 95% CI: 0.482–1.904, p=0.9016). Portal vein thrombosis was more frequent in the enteropancreatic neuroendocrine tumors. Research Sponsor: Research Launch Fund of
SBRT Alone group at 12% versus 7.87% (RR: 1.524, 95% CI: 0.712–3.262, p=0.2733). The Southwest Medical University Affiliated Hospital; 24092.
risk of major adverse cardiovascular events (MACEs) was 24.75% versus 22.12% (RR:
1.158, 95% CI: 0.607–2.213, p=0.6558). Finally, kidney outcomes (acute kidney injury,
CKD, ESRD) were similar, with risks of 62.75% and 63.40% (RR: 0.99, 95% CI:
0.834–1.175, p=0.9057). These results suggest no statistically significant differences in
the risk of adverse outcomes between the two groups. Conclusions: SBRT alone and
SBRT with bridge therapies provide comparable long-term survival outcomes in HCC
patients awaiting liver transplantation, with differences in specific complications
warranting further study. Research Sponsor: None.

4145 Poster Session 4146 Poster Session

A phase II study of lenvatinib plus everolimus in advanced extra-pancreatic Landscape of functional DLL3 expression in gastroenteropancreatic neuro-
neuroendocrine tumors (epNETs): Updated results and real-world endocrine neoplasms (GEP NENs). First Author: Rohit Thummalapalli, Memorial
comparison. First Author: Guglielmo Vetere, Department of Gastrointestinal Medi- Sloan Kettering Cancer Center, New York, NY
cal Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX Background: Delta-like ligand 3 (DLL3) is an emerging target in multiple neuroendocrine
Background: Advanced epNETs are rare malignancies with limited treatment options cancers including small cell lung cancer but remains underexplored in GEP NENs. With the
beyond somatostatin analogs, peptide receptor radionuclide therapy, and Everolimus (E). ongoing development of multiple classes of therapeutics against DLL3, there is a need to
Preclinical evidence suggests that dual blockade of VEGF and FGF is an effective anti- understand the landscape of functional DLL3 expression in GEP NENs. Methods: DLL3
angiogenic strategy and that concomitant inhibition of the mTOR pathway may be further immunohistochemistry (IHC) was completed on available tumor samples from patients (pts)
synergistic. Lenvatinib (L), a multi-target tyrosine kinase inhibitor, suppresses VEGFR, FGFR, with GEP poorly differentiated neuroendocrine carcinomas (GEP NECs) and grade 3 well
and other angiogenic pathways, while E targets the mTOR pathway. Their combination may differentiated pancreatic NETs (G3 WD PanNETs) treated between 2018-2024 and a tissue
synergistically impair angiogenesis and tumor growth. Methods: This open-label, single- microarray of resected G1-G2 WD PanNETs (185 samples total). DLL3 positivity (+) was
center, phase II study evaluated L + E in patients (pts) with advanced, progressive, well- defined as $ 5% weak (1+) IHC staining. H-scores were calculated by combining % of +
differentiated (a/p w-d) epNETs. The primary endpoint was objective response rate (ORR). tumor cells and degree of staining (1, 2, 3+), ranging from 0-300. Correlations between DLL3
Secondary endpoints were progression-free survival (PFS) and safety. Following a 2-stage status and clinicopathologic features and outcomes were analyzed. Among selected pts
design with H0: ORR , 5% and H1: ORR $ 20%, up to 32 pts were needed for type I & II error with DLL3 IHC+ GEP NENs, DLL3 immunoPET imaging using the diagnostic tracer [89Zr]Zr-
= 10%. In a post-hoc analysis, a real-world cohort of 2:1 matched a/p w-d epNET pts DFO-SC16.56 was completed to evaluate functional expression. Results: Among GEP NECs
receiving E alone served as a historical comparator. Differences in ORR were assessed using overall, 50/69 were DLL3+ (72%; median H-score 50, interquartile range [IQR] 0-160, range 0-
univariable and multivariable logistic regression while those in terms of PFS were analyzed 300), including 13/16 esophagogastric (median 45), 11/13 pancreatic (median 60), 7/11
with propensity score-based inverse probability of treatment weighting (IPTW)-adjusted Cox hepatobiliary (median 120), 16/26 colorectal (median 32.5), and 3/3 NECs of other/unknown
proportional hazards regression and Kaplan-Meier method. Results: 32 pts were enrolled. origin (median 60), with DLL3 expression higher in small cell vs large cell histology (median
The starting dose regimen was L 18 mg + E 5 mg p.o. daily with L being reduced to 14 mg p.o. 120 vs 15, P = 0.011). Among GEP NECs, there was no association between DLL3+ and
daily after the first 3 pts experienced Grade 3 adverse events (AEs). Median age was 59 years individual genomic alterations, PFS to 1L platinum-based therapy (median 4.6 mo vs 4.7 mo
(range 33 – 76), and 59% were male. Primary tumor sites included small bowel (59%), lung & in DLL3-negative [-], P = 0.435), or OS from diagnosis of advanced disease (median 15.5 vs
thymus (16%), unknown (16%), and colorectal (9%) with the majority being G2 (69%). Median 12.2 mo in DLL3-, P = 0.629). Among WD PanNETs, DLL3 expression was detected in 3/46
number of prior therapies was 2 (range 0 – 3) while 11 pts had carcinoid syndrome. The (7%) G1, 1/23 (4%) G2, and 19/47 (40%) G3 tumors, with median Ki67 higher among DLL3+
study met its primary endpoint: L + E achieved an ORR of 43.8% (6.3% unconfirmed), which vs DLL3- tumors overall (42% vs 6%, P , 0.001) and within G3 WD PanNETs alone (48% vs
was significantly higher than that observed with E alone (3.1%; OR 24.50, 95% CI 5.09 – 30%, P = 0.009). Among pts with advanced G3 WD PanNETs, DLL3+ was associated with
117.94, p , 0.001). This finding was supported by the multivariable analysis (OR 20.43, 95% shorter OS from diagnosis of advanced G3 disease (median OS 23.1 mo vs 43.9 mo in DLL3-,
CI 3.93 – 106.13, p , 0.001). After propensity score-based IPTW adjustment, a trend toward P = 0.012). [89Zr]Zr-DFO-SC16.56 DLL3 PET imaging was completed on 5 pts with DLL3 IHC+
longer PFS favoring L + E (16.0 months [95% CI 13.0 – 23.6] vs 11.3 months [95% CI 8.6 – GEP NENs at progression on standard systemic therapy. Notably, a pt with pancreatic NEC
24.3]; HR 0.88 [95% CI 0.51 – 1.52], p = 0.647) was observed. On trial, 23 Grade 3 AEs (11 and liver metastases (DLL3 IHC H-score 60) demonstrated high tumoral tracer uptake
after L dose reduction) were noted with elevated LFTs (8), hypertension (6) and throm- (SUVmax 36.7) with 95% of tumor lesions demonstrating DLL3 PET avidity. Among 4 pts with
bocytopenia (4) being the most frequent while one Grade 4 AE (hypertriglyceridemia) was DLL3 IHC+ G3 WD PanNETs, DLL3 PET was positive in 3/4, with SUVmax ranging from 14.4-
reported. Conclusions: L + E demonstrated markedly superior ORR and a trend toward 27.5 and % of DLL3 PET+ tumor lesions ranging from 50-100%. Conclusions: DLL3 is
prolonged PFS compared to E alone with a manageable safety profile. These findings expressed on a majority of GEP NECs and on a minority of well differentiated PanNETs
highlight its potential as a therapeutic option for a/p w-d epNETs and warrant further marked by high grade disease and poor clinical outcomes. Functional DLL3 PET imaging
investigation in randomized trials. Clinical trial information: NCT03950609. Research highly suggests DLL3 as a promising therapeutic target in both GEP NECs and high grade WD
Sponsor: MD Anderson Cancer Center; Jack T. and Lillian S. Clift Fellowship; Eisai Co., Ltd. PanNETs. Research Sponsor: Memorial Sloan Kettering Cancer Center.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
306s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4147 Poster Session 4148 Poster Session

Surgical debulking versus non-surgical management for the control of TROP2 expression in the gastroenteropancreatic neuroendocrine tumors:
carcinoid syndrome in metastatic small bowel neuroendocrine tumors. An analysis of 179 patients. First Author: Junaid Arshad, The University of Arizona
First Author: Rushabh Gujarathi, Department of Medicine, Section of Hematology Cancer Center, Tuscon, AZ
and Oncology, University of Chicago, Chicago, IL Background: Trophoblast cell-surface antigen 2 (TROP2) is a transmembrane glyco-
Background: Somatostatin analogues (SSAs) are first-line systemic therapeutics for protein that exhibits overexpression in various gastrointestinal (GI) malignancies, in-
tumor- and symptoms control in well-differentiated small bowel neuroendocrine tumors cluding colorectal, gastric, pancreatic, and esophageal cancers. This overexpression has
(SBNETs) with liver metastases. Surgical debulking of neuroendocrine tumor liver been correlated with increased tumor aggressiveness, enhanced proliferation, and
metastases (NETLMs) has historically been associated with symptom control in patients unfavorable prognostic outcomes. TROP2 serves as a predictive and prognostic bio-
with carcinoid symptoms. Herein, we compare clinical outcomes in patients with marker in several GI cancers, guiding targeted therapy and correlating with overall
metastatic SBNETs treated with surgical debulking for NETLMs to SSA alone or in survival. However, there exists a notable absence of dedicated studies investigating
combination with non-surgical liver directed therapies (LDTs) i.e. bland embolization or TROP2 expression specifically in gastroenteropancreatic neuroendocrine tumors (GEP-
radioembolization. Methods: Patients with serotonin-producing SBNETs and NETLMs NETs) highlighting an unmet need. This report presents the first and most extensive
with documented symptoms of carcinoid syndrome were included in this retrospective prospective study examining TROP2 overexpression in GEP-NETs. Methods: We utilized
chart review. Primary outcome was symptom-free-interval (SFI), defined as the time transcriptomic and clinical data derived from the National Cancer Institute’s (NCI) GEP-
from start of treatment until return/worsening of symptoms. Patients were censored for NET project, which originates from a prospective study approved protocol
SFI if a new line of tumor-directed therapy was initiated for radiographic disease (NCT05237934). For data analysis, we implemented the RNA-sequencing pipeline de-
progression in the absence of return/worsening of symptoms. Time to event outcomes veloped by the NCI Cancer Center Bioinformatics Resource ([Link]
were analyzed using Kaplan-Meier estimations, log-rank test, and Cox proportional skchronicles/[Link]) along with STAR version 2.7.11b for aligning sequencing
hazards regression model. Medians were compared using the Mann–Whitney U test. reads to the hg38 reference genome. To accommodate the variations introduced by
Results: Between 2018 – 2024, 64 consecutive patients with carcinoid symptoms were different library preparation methodologies (including polyA, total RNA, FFPE, and
included for analysis. 42 patients (65.6%) underwent surgical debulking (SDB) of access), we employed the "RemoveBatchEffect" function from the Limma package, while
NETLMs and 22 patients were treated with SSA alone (n = 14) or SSA plus LDT (n = 8). also accounting for disease-specific variations. Results: Our analysis included a total of
The proportion of patients reporting symptom improvement was not significantly 179 GEP-NET samples, comprised of 106 small bowel neuroendocrine tumors (NETs)
different between the SDB and non-surgical (NS; SSA + LDT) groups (SDB = 38, 90.5% vs. and 73 pancreatic neuroendocrine tumors (pNETs). There were 54 females and 52 males
NS = 19, 86.4%; p = 0.68). Among those with symptom improvement (n = 57), SFI was in the small bowel cohort, and 37 females and 36 males in the pancreatic cohort.
significantly longer in the SDB group (median SFI; SDB = 28.2 months [m] vs. NS = 15.9 Notably, TROP2 expression was observed in 50% of pancreatic samples and 30% of small
m; p = 0.004). Radiographic PFS was also significantly prolonged in the SDB group (SDB bowel NET samples. Furthermore, TROP2 expression appeared to correlate with de-
= 26.1 m vs. NS = 12 m; HR, 0.53; 95% CI, 0.29 – 0.95; p = 0.03). Within the SDB group, creased survival in pNETs (p=0.022), whereas its expression in small bowel NETs may
there was no significant difference in SFI between patients who continued to receive suggest improved patient outcomes, although this latter correlation did not achieve
SSA post-op (n = 23) and those in whom SSA was discontinued (n = 19) post-op (28 m vs. statistical significance (p=0.37). Further analyses are pending study completion and will
30.6 m; p = 0.85). Post-treatment median nadir serotonin was significantly lower in the be presented later. Conclusions: This study highlights the critical role of TROP2
SDB group (total N = 41 [SDB = 31 + NS = 10]; SDB = 299 ng/mL vs. NS = 947.5 ng/mL; overexpression in GEP-NETs and its importance for patient prognosis. TROP2 over-
p , 0.001). Median percentage decrease (from pre-treatment to nadir values) in serum expression correlates with decreased survival outcomes in pNETs relative to small bowel
serotonin was higher in the SDB group with the difference approaching significance NETs. Additionally, the identification of TROP2 as a prognostic and predictive biomarker
(total N = 30 [SDB = 22 + NS = 8]; SDB = 80.9% vs. NS = 51.5%; p = 0.06). presents opportunities for future research focused on therapeutic targeting. Additional
Conclusions: Despite the frequent use of SSAs, surgical debulking of NETLMs in pa- studies may be needed for further validation as we finalize the current research.
tients with carcinoid symptoms remains superior for symptom control and should Research Sponsor: None.
therefore be considered in these patients. Research Sponsor: None.

4149 Poster Session 4150 Poster Session

Development of a cfDNA-based protein-informed epigenetic signature (PEp- Development of a comprehensive cfDNA methylation signature for prog-
sig) to enable biomarker-based risk stratification for pancreatic ductal nostic, predictive, and diagnostic applications in pancreatic ductal adeno-
adenocarcinoma (PDA). First Author: Ashish Manne, The Ohio State University carcinoma (PDA). First Author: Deepak Sherpally, Metropolitan Hospital Center, New
Comprehensive Cancer Center, Columbus, OH York, NY
Background: Novel blood-based biomarkers are needed for early risk stratification and per- Background: Blood-based biomarkers are promising for predicting outcomes in pancreatic
sonalized therapies for patients with PDA. An evidence-based 99-gene panel was developed ductal adenocarcinoma (PDA) and could pave the way for developing multi-omic prognostic
focusing on genes whose protein products are implicated in PDA drug response. We hypothesized tools. We previously identified a 15-gene cell-free DNA (cfDNA) methylation signature targeting
that methylation levels in these genes could serve as surrogates for their corresponding protein treatment response (TRg). To enhance its utility, we incorporated additional genes with
activity, enabling the prediction of treatment-related outcomes in PDA patients. Using this panel, established prognostic (Prg) and diagnostic (Dxg) value from the literature, creating a composite
we developed a cell-free DNA (cfDNA)-based PEp-sig for risk stratification of patients with PDA. panel. This study aimed to develop a comprehensive cfDNA methylation signature with pre-
Methods: Targeted enzymatic methylation sequencing was performed on plasma samples from dictive, prognostic, and diagnostic value. Methods: Enzymatic methylation sequencing was
PDA patients (01/2010–05/2022) receiving chemotherapy from the Ohio State University bio- performed on plasma samples collected between January 2010 and May 2022 from PDA
repository. Gene methylation levels were analyzed for associations with overall survival (OS) using patients undergoing chemotherapy at The Ohio State University. A 206-gene panel (TRg + Prg +
univariate and multivariate Cox regression models. Significant genes and covariates such as first Dxg) was analyzed, with methylation levels correlated to overall survival (OS) using univariate
chemotherapy (FLC)- FOLFIRINOX (FFX) vs. gemcitabine (G)/nab-paclitaxel (NP) vs. other (Ot), and multivariate (MV) Cox regression models. Backward selection identified significant genes,
and stage at diagnosis (StD) – early-stage (ES) that includes resectable and borderline resectable and MV models adjusted for clinical covariates. Patients were stratified into high-risk (Hg) and
PDA vs. locally advanced (LA) and metastatic (Met)) were identified through backward selection. low-risk (Lg) groups based on median risk scores. Results: Our study cohort had 51 PDA
Risk scores from multivariate models were dichotomized at the median, stratifying patients into patients, with a median age of 65 (range: 34 -80), 53% females, and 86% Caucasian (12% African-
High (Hg) and low-risk groups (Lg), with survival differences assessed by Kaplan-Meier and log- American and 2% others). Stage at diagnosis (Dx-S) distribution: 15 locally advanced (LA), 14
rank tests. Results: The study cohort (SC) included 51 PDA patients (StD: 22 ES, 15 LA, and 14 metastatic (Mets), and 22 resectable/borderline resectable (R/BR). 12 patients in BR/R received
Met). Among the ES cases, 3 progressed to Met after neoadjuvant therapy (NAT); in the LA sub- neoadjuvant therapy (NT) while the rest had upfront surgery (UpS) followed by adjuvant therapy
group, 2 proceeded to surgery post-chemotherapy. Ultimately, 21 patients had resection (Rs), (AT). Ultimately, 21 had resection, and 30 got palliative therapy. First-line chemotherapy (FL)
while 30 underwent palliative therapy (PT), with FLC distribution as follows: PT group – 12 FFX, 16 patients received is, 36 FOLFIRINOX (12 PT, 9 NT, and 5 AT), 19 gemcitabine (Gem)/nab-
G/NP, and 2 Ot; Rs group, NAT—9 FFX, 1 G/NP, and 1 Ot; adjuvant therapy after upfront surgery paclitaxel (NP) (16 PT, 1 NT, and 2 AT), and 6 Others. A 25-gene cfDNA methylation signature (15
(UpS) —5 FFX, 2 G/NP, 3 Ot. Two 15-gene PEp-sig were developed: one for the SC and another for TRg, 10 Prg, and 1 Dxg) stratified patients into Hg and Lg groups. OS was significantly longer in
the PT group. A significant overlap (11/15) of genes was observed between the two PEp-sigs. Lg compared to Hg across all models. Conclusions: The study successfully developed a
These genes are linked to the response to G, NP, irinotecan, and platinums. The performance of comprehensive 25-gene cfDNA methylation signature combining predictive, prognostic, and
PEp-sig models, with and without FLC and StD adjustments, is summarized below. diagnostic markers for PDA. This signature effectively stratifies patients into Hg and Lg,
Conclusions: In this proof-of-concept study, we present a cfDNA-based PEp-sig that effectively demonstrating significant differences in OS across various clinical models. The results highlight
stratifies PDA patients by survival risk. Notably, it operates independently of FLC and StD within the potential utility of cfDNA methylation as a multi-omic tool to enhance personalized
the PT group. Ongoing efforts aim to develop treatment selection algorithms based on these treatment strategies and improve patient outcomes in PDA. Further validation in larger cohorts
findings. Research Sponsor: None. is warranted to confirm its clinical applicability. Research Sponsor: None.
SC PT-group OS*
Models tested
Hg vs. Lg Hazard Hg vs. Lg Models tested Hg vs. Lg Hazard ratio
OS (in months) Ratio (HR) p-value OS (in months) HR p-value
Signature-alone 7.58 vs. 33 13.5
PEp-sig alone 10.75 vs. 33 8.7 Plus, FL 8.98 vs. 33 13.2
PEp-sig + FLC* 10.62 vs. 33 8.1 ,0.001 5.3 vs. 16.83 9.2 ,0.001 Plus, FL, Surgery (NAT vs. UpS vs. PT) 7.58 vs. 33 17.3
PEp-sig + StD* 8.4 vs. 33 16.9 8 Plus, Dx-S 7.58 vs. 33 14.4

*FLC and StD significantly impacted OS in SC but not in PT. *In months.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 307s

4151 Poster Session 4152 Poster Session

DNA methylation signatures as predictive biomarkers for chemotherapy A phase I trial of binimetinib plus hydroxychloroquine in patients with
(CT) resistance and survival in pancreatic ductal adenocarcinoma (PDA). previously treated metastatic pancreatic cancer. First Author: S. Daniel Hal-
First Author: Deepak Sherpally, Metropolitan Hospital Center, New York, NY dar, The University of Texas MD Anderson Cancer Center, Houston, TX
Background: Predicting innate treatment resistance to traditional CT in PDA can op- Background: Combined MEK and autophagy inhibition exerts synergistic antitumor
timize therapeutic strategies and improve patient outcomes. This study investigates activity in preclinical models of RAS-mutant cancers. We hypothesized that blockade of
methylation changes in genes encoding proteins implicated in preclinical models (PDA autophagy with hydroxychloroquine (HCQ) can overcome therapeutic resistance to MEK
cell lines or mouse models) influencing drugs commonly used to treat PDA, including 5- inhibition with binimetinib (bini) and lead to clinical benefit in patients (pts) with
fluorouracil (5FU), oxaliplatin, irinotecan, gemcitabine (Gem), nanoliposomal irinotecan, previously treated, KRAS-mutant metastatic pancreatic ductal adenocarcinoma (PDAC).
and nab-paclitaxel. Using a comprehensive literature review (1970–2024), we curated a Methods: This is an investigator-led, single-arm, open-label, phase I dose escalation/
panel of relevant genes and analyzed their methylation patterns in The Cancer Genome expansion study of bini + HCQ in metastatic PDAC pts. Key eligibility criteria: ECOG 0-1,
Atlas (TCGA) database. We hypothesized that DNA methylation changes affect gene adequate organ function, . 1 prior line of therapy for metastatic disease, and presence
expression and protein production, contributing to chemotherapy resistance. of KRAS mutation. Dose escalation followed a Bayesian optimal interval (BOIN) design.
Methods: PDA patient methylation data were accessed from the TCGA database. Dose level (DL) 1: bini 45 mg + HCQ 600 mg p.o. bid; DL -1: bini 45 mg + HCQ 400 mg p.o.
Survival analyses were performed using elastic net multivariate regression to identify bid; DL -2: bini 30 mg + HCQ 400 mg p.o. bid; DL -1.5: bini 30 mg + HCQ 600 mg p.o. bid.
significant methylation signatures, followed by Kaplan-Meier analysis. Model param- Primary endpoint was the maximum tolerated dose (MTD) of bini + HCQ. Secondary
eters, including alpha (a) and lambda (l), were optimized through 100 iterations to endpoints included objective response rate (ORR), disease control rate (DCR),
minimize error. Our curated panel consisted of 138 genes, predominantly Gem-specific progression-free survival (PFS), and overall survival (OS). Results: From December
or Gem + 5FU (n = 93). Results: The TCGA database provided methylation data for 184 2019 to August 2024, a total of 34 pts were enrolled in dose escalation (n = 17) and dose
PDA patients (106 had Gem or Gem-based therapy), with 133/138 genes in our analysis. expansion (n = 17). Median age was 65 yrs (range: 45-79) with 56% females. Median prior
Our analysis identified 23 cytosines followed by guanine residue (CpG) methylation lines of therapy was 2 (range: 1-4). The most prevalent KRAS mutation subtypes were
signatures within the panel, ranging from 1 to 23 CpG sites. The best-performing G12D (35%), G12V (32%), and G12R (29%). Two dose-limiting toxicities (DLTs) occurred
signature, containing 21 CpG sites, stratified patients into significantly different survival in 2 out 3 pts treated at DL 1: grade 3 CPK elevation with renal impairment (bini) and
groups (17 months (m) vs. not evaluable, p = 0.004). The second-best signature, with 8 grade 3 QTc prolongation (HCQ). The most frequent non-hematologic AEs were rash
CpG sites, stratified survival as 17m vs. 66.94m (p = 0.03). Interestingly, signatures with (71%), diarrhea (71%), nausea (67%), elevated AST (67%), and elevated CPK (61%).
the most (n = 23) and least (n = 1) CpG sites also demonstrated strong stratification, with Following dose de-escalation due to poor tolerance, the MTD was deemed to be bini
survival differences of 15.15m vs. 30.02m (p = 0.01) and 18.67m vs. 44.38m (p = 0.01), 30 mg + HCQ 600 mg p.o. bid and used for dose expansion. Overall, out of 31 response-
respectively. Many signatures included multiple CpG sites from single genes. A Gem or evaluable pts, 2 pts achieved a partial response (lasting 6.9 and 4.7 mos, both at DL -1.5)
Gem + 5FU-specific panel (n = 93) applied to patients treated with Gem-based therapy and 9 pts achieved stable disease (3 pts at DL -1, 6 pts at DL -1.5), consistent with ORR
identified an 8-CpG signature distinguishing high-risk patients (20.84m vs. 49.38m, p = 6.5% and DCR 35.5%, respectively. At median follow-up of 19 mos, median PFS was 1.9
0.02). Conclusions: This study highlights the potential of CpG methylation signatures to mos and median OS was 5.3 mos. Conclusions: Bini + HCQ demonstrated a challenging
predict treatment outcomes in PDA. These findings may guide the identification of high- toxicity profile and limited clinical activity in a heavily pretreated cohort of metastatic
risk patients and the optimization of CT regimens for improved survival. The identifi- PDAC pts. Minimal efficacy was observed in this treatment-refractory population;
cation of methylation signatures associated with genes implicated in chemotherapy however, dual MEK and autophagy inhibition may warrant further study in earlier-line
resistance provides valuable insights into the underlying mechanisms of innate settings, potentially with more tolerable drug candidates and guided by biomarker
treatment resistance in PDA. Further validation of these methylation signatures could selection. Clinical trial information: NCT04132505. Research Sponsor: None.
contribute to more effective and targeted therapeutic approaches in clinical practice.
Research Sponsor: None.

4153 Poster Session 4154 Poster Session

Pancreatic adenosquamous carcinoma (PASC): A comparative genomic Unveiling the differences in tumor immune microenvironment between
landscape study. First Author: S. Daniel Haldar, The University of Texas MD KRAS-wildtype and KRAS-mutant pancreatic ductal adenocarcinoma. First
Anderson Cancer Center, Houston, TX Author: Heidi C. Ko, Labcorp, Durham, NC
Background: PASC accounts for 1-4% of primary exocrine pancreatic malignancies and is as- Background: Pancreatic ductal adenocarcinoma (PDAC) is a highly aggressive cancer
sociated with more aggressive disease biology and worse clinical outcomes compared to with a dismal prognosis despite advances in treatment. PDAC is characterized by a
conventional pancreatic ductal adenocarcinoma (PDAC). Despite its aggressive nature, PASC dense stromal environment with suppressed anti-tumor immunity that contributes to
lacks targeted treatment choices and is frequently excluded from clinical trials, underscoring a treatment resistance. Oncogenic KRAS mutations are present in . 90% of PDACs and
key unmet need. Here, we performed a comparative analysis of the genomic landscapes of PASC
are known to play a role in modulating the tumor immune microenvironment (TME). In
versus PDAC using the FoundationOne database to identify distinct molecular drivers and po-
tential therapeutic targets. Methods: Comprehensive genomic profiling using hybrid capture-
this study, we examined the differences in TME components between KRAS-mutant (m)
based next-generation sequencing (NGS) was performed on 244 PASC and 29,021 PDAC tumors and KRAS-wildtype (wt) PDACs. Methods: Comprehensive genomic and immune pro-
to identify genomic alterations (GAs). All patients (pts) had clinically advanced disease, pre- filing (CGIP), including the RNA-seq based gene expression assessment of 395 immune-
dominantly stage IV, at the time of profiling. Genomic ancestry, MSI status, tumor mutational associated genes, was performed on 311 PDAC patient samples. Gene expression
burden (TMB), homologous recombination deficiency signature (HRDsig) and cosmic trinucle- signatures of tumor immunogenicity (TIGS) and cell proliferation were calculated by
otide signature were assessed. PD-L1 expression was quantified using the tumor proportion score averaging the gene expression ranks of 161 immune-associated genes and 10 prolif-
(TPS) via the Dako 22C3 immunohistochemistry assay. Statistical analysis was performed using eration genes, respectively. The normalized gene expression rank of 22 immune
Fisher’s exact test, with false discovery rate (FDR) correction applied through the Benjamini- checkpoint genes and 17 cancer testis antigen (CTA) genes were also calculated. DNA-
Hochberg method. Results: PASC harbored a median of 6 GAs per tumor (range: 1-23) with a seq was used to identify KRAS mutations and to calculated tumor mutational burden
similar genomic ancestry profile compared to PDAC. Of note, PASC featured a higher frequency of (TMB). Continuous variables were compared between subgroups using the Wilcoxon
cases with MSI-high status (2.1% vs 0.5%; p = .027) and TMB . 10 mutations/Mb (3.7% vs 1.3%; p Rank-Sum test and categorical variables were compared between groups using the Chi
= .013). Additionally, PD-L1 expression (TPS . 1%) was significantly more common in PASC
compared to PDAC (66.7% vs 37.0%; p , .0001). The frequency of KRAS mutations and HRDsig
Squared test. For statistical significance, p , 0.05 was required. Results: The cohort
positivity was similar between the two subtypes. Disease-associated GAs more frequent in PASC consisted of 311 PDAC patient samples, comprising 159 females (51.1%) and 152 males
than PDAC included mutations in CDKN2A (77.0% vs 56.6%; p , .0001), KMT2D (8.2% vs 3.2%; (48.9%) with a median age at testing of 69.3 years (38.7-92.6). A total of 264 specimens
p , .0001), TP53 (89.3% vs 78.0%; p , .0001), and MTAP loss (33.3% vs 23.8%; p = .002). (84.9%) exhibited a KRAS mutation, with the most common being G12D and G12V while
Conclusions: PASC exhibits a genomic profile with molecular features that are both shared and 47 (15.1%) had KRAS-wt tumors. No difference in TIGS, CP, or TMB was observed
distinct compared to PDAC. Given the similar frequency of KRAS mutations, PASC pts should be between KRAS-wt and KRAS-m tumors. KRAS-wt tumors exhibited greater expression of
included in clinical trials of emerging RAS-targeted therapies. Furthermore, immunotherapy- 12 of 17 tested CTAs than KRAS-m tumors, including GAGE13 (p = 0.004) NY-ESO-1 (p =
based strategies (↑MSI-high, TMB, and PD-L1) and PRMT5/MAT2A inhibitors (↑MTAP loss) 0.01), MAGEA3 (p = 0.01), and LAGE1A (p = 0.01). Analysis of the gene expression of 22
warrant consideration in this rare and understudied disease subtype. Research Sponsor: None. immune checkpoint genes showed no difference for most of the genes, though there was
PDAC (n=29,021) PASC (n=244) P-value higher expression of PD-1 (p = 0.04) and CD27 (p = 0.03) seen in KRAS-wt compared to
KRAS-m tumors. Conclusions: KRAS-wt PDACs exhibited greater expression of cancer
Median GAs/tumor (range) (IQR) 5 (0-61) (3-6) 6 (1-23) (4-8) ,.0001
MSI-high 0.5% 2.1% .027 testis antigen genes compared to KRAS-m tumors, suggesting potential therapeutic
TMB > 10 muts/Mb 1.3% 3.7% .013 susceptibility to immunotherapy and adoptive cell therapies leveraging the expression of
PD-L1 TPS > 1% 37.0% 66.7% ,.0001 CTAs as targets. Assessment of KRAS status and immunotherapy susceptibility may
HRDsig+ 4.6% 3.1% NS
CDKN2A 56.6% 77.0% ,.0001 support future clinical trial selections for therapies targeting the complex interplay of
KRAS 92.8% 95.5% NS genomic and immune components of pancreatic cancer. Research Sponsor: None.
MTAP loss 23.8% 33.2% .002
TP53 78.0% 89.3% ,.0001

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
308s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4155 Poster Session 4156 Poster Session

Functional role of GLI2 in cancer-associated fibroblasts for modulation of Use of an ultra-sensitive sequencing platform to detect mutant KRAS in the
the fibrotic tumor microenvironment within pancreatic cancer. First Author: whole blood of pancreatic cancer patients. First Author: Ryne Ramaker, Duke
John Y. Kwon, Mayo Clinic Rochester, Rochester, MN Cancer Institute, Duke University, Durham, NC
Background: The tumor microenvironment (TME) that surrounds pancreatic ductal Background: Pancreatic Ductal Adenocarcinoma (PDAC) is a leading cause of cancer
adenocarcinoma (PDAC) is a multi-faceted and dynamic ecosystem in which stromal death and mortality is increasing. A contributor to poor outcomes is the absence of non-
fibroblasts communicate with cancer cells to mediate tumor growth, metastasis, and invasivebiomarkers for disease screening, treatment monitoring, and identification of
chemotherapy resistance. It is well recognized that both cancer-associated fibroblasts therapeutic targets. Blood-based profiling of circulating tumor DNA (ctDNA) using Next
(CAFs) and its non-cellular, fibrotic components within the TME can foster a protu- Generation Sequencing (NGS) has addressed these needs in several cancer types, but
morigenic environment for PDAC. However, we still lack a comprehensive understanding available commercial ctDNA assays are not as effective in PDAC. To address this unmet
of the precise mechanisms in which this dense, fibrotic matrix can help drive malignant clinical need, we developed an ultra-sensitive sequencing assay to detect mutant KRAS
behaviors. Here, we reveal a novel mechanism in which the zinc-finger transcription in the whole blood of PDAC patients. Methods: We adapted the bacterial Maximum
factor GLI2 regulates type 1 collagen expression within CAFs and how the soluble variant Depth Sequencing (MDS) assay for Human whole blood MDS (hMDS) to improve upon
of this collagen promotes irinotecan chemoresistance. Methods: We leveraged tran- the sensitivity of NGS by barcoding DNA fragments with unique molecular identifiers
scriptomic data from The Cancer Genome Atlas, International Cancer Genome Con- prior to performing multiple rounds of first-strand synthesis to resolve sequencing
sortium, and Clinical Proteomic Tumor Analysis Consortium to evaluate GLI2 expression errors. Analytic sensitivity was evaluated by spiking PDAC cells at various dilutions into
and stromal content of human PDAC tumors through bulk RNA-sequencing deconvo- control blood isolated from 10 individuals and assaying the mixture by hMDS in trip-
lution. Using single-nucleus RNA sequencing of human PDAC tumors, we validated the licate. Clinical sensitivity was then evaluated by collecting paired blood draws from 200
association of GLI2 expression and stromal matrix constituents in CAFs. Chromatin advanced PDAC patients in prospective fashion, one to be tested with a commercial
immunoprecipitation assays in human CAFs confirmed GLI2 binding at the COL1A1 ctDNA test and one by hMDS. Results: ThehMDS assay reproducibly detected PDAC
promoter. Through RNAi-based inactivation of GLI2, we determined how loss of GLI2 cells at dilutions as low as one cell per mL or one mutated fragment per million KRAS
impacts regulation of type 1 collagen. Additionally, we conducted MTT assays to assess fragments. Thus, hMDS reached an analytic sensitivity 1000x higher than NGS. Clonal
tumor viability in response to irinotecan treatment. Results: Transcriptomic analysis KRASmutations were detected in 179 of the first 194 patient samples (92.2%), sur-
revealed that GLI2 is highly enriched in CAFs and strongly correlated with stromal passing historical commercial detection rates of 50% by commercial ctDNA assays.
fibrosis compared to other non-tumor cell constituents within the TME. We have shown Mutations were reproducibly detected in replicate analysis of forward and reverse DNA
that GLI2 directly binds to the promoter of COL1A1, a key component of type 1 collagen, strands. Weak clonal KRAS activating mutations were also detected in several non-
in CAFs and regulates its transcription in a manner dependent on TGFb1 signaling. cancer, control patients. Conclusions: We developed an assay capable of sensitively
Interestingly, PDAC tumors exposed to type 1 collagen show increased expression of detecting PDAC ctDNA in whole blood. Formal comparison to commercial testing is
pro-tumorigenic pathways involved in inflammation, EGR signaling, cytokine-receptor ongoing, but preliminary results suggest this assay could be a sensitive tool for non-
interactions, and irinotecan resistance. We further validate that human PDAC cells pre- invasive PDAC mutation profiling and disease monitoring. The presence of weak clonal
treated with collagen can confer chemoresistance to irinotecan with viability assays. KRAS mutations in control patients has motivated development of a multiplex platform
Conclusions: Taken together, our study demonstrates a novel mechanism in which GLI2 capable of screening for mutations in multiple driver genes. Research Sponsor: National
regulates the secretion of collagen within CAFs, which in turn can enable PDAC to Cancer Institute; 5R21-CA257816-02; Hopper Belmont Foundation.
acquire resistance to standard-of-care treatments. These findings highlight not only the
oncogenic functions for CAFs and their fibrotic secretome, but also open new avenues in
which therapeutic targeting of the TME may provide clinical benefit for PDAC patients.
Research Sponsor: National Cancer Institute/U.S. National Institutes of Health.

4157 Poster Session 4158 Poster Session

Efficacy and safety of surufatinib (S) plus KN046 (K) and chemotherapy in Early treatment ctDNA dynamics to predict response to chemotherapy in
first line (1L) advanced pancreatic cancer (PC): A single-arm, phase 1b/2 patients with metastatic pancreatic cancer: A prospective, observational
trial. First Author: Wen-Quan Wang, Department of Pancreatic Surgery, Zhongshan pilot study. First Author: James Lee, Donald and Barbara Zucker School of Medicine at
Hospital, Fudan University, Shanghai, China Hofstra/Northwell, Hempstead, NY
Background: Gemcitabine (G) and nab-paclitaxel (nP) are standard 1L regimen for pa- Background: Assessment of chemotherapy response using imaging has limitations in
tients (pts) with unresectable PC, yet the efficacy remains unsatisfactory. K is a humanized metastatic pancreatic cancer whereas circulating tumor DNA (ctDNA) may offer an alternate
bispecific antibody targeting PD-L1/CTLA-4, while S is a kinase inhibitor of VEGFR1-3, assessment of tumor burden with improved lead-time. This study evaluated temporal ctDNA
FGFR1 and CSF-1R with immune-regulatory potential. It is hypothesized the add-on of S testing as a potential biomarker of treatment response data compared to standard-of-care
and K to GnP chemotherapy would provide improved efficacy. Methods: This single-arm, CT scans in metastatic pancreatic cancer. Methods: In this prospective, observational,
phase 1b/2 trial enrolled pts with unresectable locally advanced or metastatic PC who were single-center trial, patients with metastatic pancreatic ductal adenocarcinoma starting a
eligible for 1L treatment. The phase 1b part was designed in a "3+3" algorithm to determine first-line (1L) or second-line (2L) systemic treatment regimen underwent imaging and
the recommended phase 2 dose (RP2D) of S for dose expansion in phase 2 part. Pts frequent blood-based treatment assessment. CT imaging was obtained prior to and after
received oral S at escalating dose starting from 200 mg qd, plus intravenous K at 5 mg/kg 8 weeks of treatment, with response measured by RECIST 1.1 criteria. CA 19-9, CEA, and
on day 1, and GnP chemotherapy on days 1 and 8 at 21-day cycles. The primary endpoint longitudinal ctDNA profiling using Signatera were performed at baseline, after 2 weeks,
was dose-limiting toxicities (DLTs) within the first 28 days for phase 1b, and ORR per 4 weeks, and 8 weeks. A threshold of 20% decrease achieved or not was used to dichotomize
RECIST 1.1 for phase 2. Secondary endpoints included DCR, PFS, OS, safety, and efficacy- ctDNA response and other cut-offs were evaluated in sensitivity analysis. Progression-free
related biomarkers. Results: As of Dec 19th, 2024, 18 pts were enrolled with a median age survival (PFS) was measured and compared to biomarker response. The primary objective
of 54 (range: 41-74), predominantly male (16/18) and metastatic disease (15/18). Of the 16 was to evaluate the association of ctDNA changes at 4 and 8 weeks with PFS. Exploratory
pts with genetic testing, KRAS (15/16) and TP53 (11/16) mutations were common, fol- objectives included evaluation of response at 2-weeks, and comparison of PFS with CA 19-9
lowed by DNA damage response (DDR) -related mutations (7/16) including ARID1A, ATM, and CEA. Results: Between June and December 2023, 19 patients were enrolled. Sufficient
CHEK2, etc., while TMB-H (1/16) is rare, and none had MSI-H or dMMR status. Within the 9 tissue was available to perform tumor-informed ctDNA profiling in 12 of 19 patients (63%)
pts from phase 1b part (3 in S 200 mg cohort, 6 in S 250 mg cohort), no DLTs occurred thus undergoing systemic therapy with 7 patients (58%) starting on 1L therapy and 5 (42%) on 2L
the RP2D of S was determined as 250mg qd. In the 16 evaluable pts, the best overall therapy. Overall median PFS was 4.0 months (4.2 months 1L, 3.3 months 2L). CtDNA
responses were 1 CR, 10 PRs and 5 SDs. The ORR was 68.8% and the DCR was 100%. 2 pts response at 4-weeks was prognostic of PFS with a median decrease in ctDNA of 83.4% (p =
received R0 resection after 6 cycles’ treatment. With a median follow up of 7.43 months, 0.0002). CtDNA response at 8-weeks was also prognostic of PFS (p = 0.01). Three patients
the estimated median PFS was 8.25 (95% CI: 4.57-NR) months and the 6-month PFS rate achieved ctDNA clearance of over 95% by week 4 and had a PFS of 5.9 months. Of, note
was 72.9%. Estimated median OS was 11.14 (95% CI: 5.52-NR) months and the 6-month OS there was one patient who had a 1-log decrease in ctDNA at 4-weeks, but a 20-fold increase
rate was 82.5%. In the exploratory analysis, DDR-related mutations seemed predictive for at 8-weeks, whereas CEA and CA 19-9 were stable at 8-weeks, and this patient had
better ORR (85.7% vs 55.6%, P= 0.308), PFS (6-month PFS rate:100% vs 53.3%, log-rank P= subsequent disease progression 1 month later. In an exploratory analysis, ctDNA as soon as
0.519), and OS (6-month OS rate:100% vs 62.5%, log-rank P= 0.116). Treatment-related 2 weeks was also prognostic of recurrence (p = 0.002). In contrast to ctDNA, CA 19-9 and
adverse events (TRAEs) occurred in 14 (77.8%) pts, and most common TRAEs ($20%) CEA did not show significance at any timepoint (2 weeks: p = 0.2 and p = 0.8, respectively;
included leucopenia (44.4%), hypertension (38.9%), and thrombocytopenia (22.2%). TRAEs 4 weeks: p = 0.7 and p = 0.8, respectively; 8 weeks: p = 0.8 and p = 0.5, respectively).
of grade $3 included leucopenia, neutropenia, thrombocytopenia, and hypertension (n = 2 Sensitivity analysis showed that 20% decrease was robust; association at 4-weeks with
[11.1%] for each). There were no treatment-related deaths. Conclusions: These prelim- ctDNA remained significant across a wide threshold range (-70% to +60%).
inary results showed encouraging anti-tumor efficacy and an acceptable safety profile of S Conclusions: Circulating tumor DNA predicts PFS as early as 2-weeks following treatment.
plus K and GnP chemotherapy as 1L treatment for advanced PC. Clinical trial information: These findings suggest clinical utility in measuring ctDNA in mPDAC as early as 2 weeks
NCT05832892. Research Sponsor: HUTCHMED, ALPHAMAB. following treatment initiation within the context of prospective interventional clinical trials.
Research Sponsor: Conquer Cancer, the ASCO Foundation.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 309s

4159 Poster Session 4160 Poster Session

TQB2868 combined with anlotinib and nab-paclitaxel plus gemcitabine as Phase II trial of serplulimab combined with gemcitabine plus nab-paclitaxel
first-line treatment for metastatic pancreatic cancer: A prospective, multi- (GnP) and SBRT for metastatic pancreatic cancer as the first-line treatment.
center, single-arm, phase 2 study. First Author: Si Shi, Department of Pancreatic First Author: Ke Cheng, Division of Abdominal Tumor, Department of Medical Oncology,
Surgery, Fudan University Shanghai Cancer Center, Shanghai, China Cancer Center and State Key Laboratory of Biological Therapy, West China Hospital,
Background: Chemotherapy currently serves as the cornerstone for treating meta- Sichuan University, Chengdu, China
staticPancreatic Ductal Adenocarcinoma Cancer (mPDAC). Nevertheless, the survival of Background: The addition of anti-PD-1 monoclonal antibody (mAb) to gemcitabine and
patients with mPDAC remains poor. Besides, the efficacy of single-agent immune nab-paclitaxel (GnP) presented limited improvement in objective response rate (ORR)
checkpoint inhibitors or antiangiogenesis in the treatment of mPDAC is not satisfying. and disease control rate (DCR) for metastatic pancreatic ductal adenocarcinoma
Therefore, it is important to explore combination therapy options for patients with (mPDAC) in our previous study. Therefore, we conducted this phase II trial to evaluate
mPDAC. TQB2868 injection is a bifunctional fusion protein that targets PD-1 and TGF- the efficacy and safety of anti-PD-1 mAb (Serplulimab) plus GnP chemotherapy and
bRII. This trial was conducted to evaluate the effectiveness and safety of TQB2868 stereotactic body radiotherapy (SBRT) in patients with mPDAC. Methods: Patients with
combined with anlotinib and nab-paclitaxel plus gemcitabine for mPDAC. Methods: This mPDAC without previous treatment were enrolled to receive Serplulimab and GnP plus
was a prospective, multicenter, single-arm phase II trial. Eligible pts were those who SBRT (SGSBRT) (intravenous infusion of Serplulimab 200 mg on day 1 every 3 weeks,
aged over 18 years, histologically or cytologically confirmed PDAC, have not received gemcitabine 1000 mg/m2 and nab-paclitaxel 125 mg/m2 on day 1 and day 8, repeated
treatment before and radiographically showed distant metastases and measurable every 3 weeks; SBRT delivered 5 fractions of 6.6 Gy to the primary tumor or 3 fractions of
lesions. Patients received TQB2868 (300mg, I.V, D1,15) and Anlotinib (10mg, P.O, QD, 8 Gy to the metastatic lesion in cycle 2) as the first-line treatment. The primary endpoint
D1-14), in addition to nab-paclitaxel (125mg/m2, I.V,D1,8,15) and gemcitabine (1.0g/m2, was 6-month progression-free survival (PFS) rate. Secondary endpoints included overall
I.V, D1,8,15), administered over a 28-day treatment [Link] primary endpoint was survival (OS), PFS, ORR, DCR, and adverse events (AEs). Moreover, the biomarkers such
Progression-free Survival (PFS), Secondary end points included objective response rate as circulating tumor DNA (ctDNA) and circulating hybrid cells (CHCs), PD-L1 expression,
(ORR) and disease control rate (DCR), overall survival (OS) and safety. The biomarker tumor tissue genetic status, cytokine levels, and immune microenvironment were also
TGF-b1 was analyzed as exploratory results (NCT06767813). Results: 40 pts were investigated. Results: As of January 2025, 47 patients have been enrolled and 41
enrolled and received TQB2868 combination regimen therapy, and the last follow-up patients have been followed for more than 6 months, with all 47 patients achieving
time was January 10, 2025. 36 pts were eligible for response evaluation. With median efficacy according to the protocol. The 6-month PFS rate was 78.48%. The ORR was
follow-up duration of 5.9 months, the median PFS and OS have not been reached, with 6- 74.47% (35/47), including 1 complete response (CR) and 34 partial responses (PR), and
month PFS and OS rates of 86% and 95%, respectively. The ORR was recorded at 63.9% the DCR was 100%, with 12 stable disease (SD). The median PFS was 8.6 months, and
(23/36) (95% CI, 46.2%-79.2%), with 23 pts achieving partial response. The DCR was the median OS was 15.5 months. The frequent grade 3 drug-related AEs were neu-
100% (36/36) (95% CI, 90.3%-100%). The most common TRAEs were neutropenia, tropenia (20/47, 42.55%), leukopenia (19/47, 40.43%), anorexia (18/47, 38.30%), and
thrombocytopenia, leukopenia, and anemia. Grade 3 TRAEs were reported in 52.5% pts fatigue (11/34, 23.40%). The correlation between biomarkers and efficacy and prognosis
(21/40). In exploratory analysis, the inhibition rate of TGFb1 was over 90% in most cases are under-analyzed. Conclusions: This phase II study has met our preset primary
after administration, with little to no rebound. Conclusions: TQB2868 combination endpoint with 78.48% in 6-month PFS rate, and SGSBRT presented promising efficacy
regimen as first-line treatment was demonstrated to be tolerable, with promising anti- with manageable safety profile and expected antitumor activity. This combination might
tumor activity in mPDAC. Clinical trial information: NCT06767813. Research Sponsor: be a promising option as first-line therapy for Chinese patients with mPDAC. Clinical trial
None. information: ChiCTR2300073237. Research Sponsor: None.

4161 Poster Session 4162 Poster Session

First-line treatment with surufatinib, camrelizumab, nab-paclitaxel, and S-1 First-line serplulimab and bevacizumab combined with nab-paclitaxel/
in locally advanced or metastatic pancreatic ductal adenocarcinoma gemcitabine followed by mFOLFOX in advanced pancreatic cancer: A phase
(PDAC): A phase Ib/II randomized study. First Author: Guang-Hai Dai, Senior II trial. First Author: Jieer Ying, Zhejiang Cancer Hospital, Hangzhou, China
Department of Oncology, the Fifth Medical Center of the PLA General Hospital, Beijing, Background: Patients with advanced pancreatic cancer (PC) have a poor prognosis, as
China this ’cold’ tumor shows limited responsiveness to mono-immunotherapy. Chemotherapy
Background: PDAC is a highly aggressive cancer with limited treatment options. Previous may improve the efficacy of immunotherapy by reshaping the tumor immune microen-
reports of NCT05218889 showed promising efficacy of nab-Paclitaxel/S-1/Surufatinib/ vironment. We conducted a phase II trial to assess the anti-tumor activity and safety of
Camrelizumab (anti-PD-1 antibody) combination regimen (NASCA) in mPDAC (2023 ASCO first-line serplulimab (anti-PD-1) and HLX04 (a bevacizumab biosimilar) combined with
abs# 4142; 2024 ASCO GI abs# 671). Here, we present the updated results. Methods: In nab-paclitaxel plus gemcitabine (nab-P/Gem), followed by modified FOLFOX (oxaliplatin,
phase Ib, a 3+3 dose escalation design was used to determine the RP2D of surufatinib. In leucovorin, and 5-fluorouracil; mFOLFOX), in patients with locally advanced or metastatic
phase II, patients were randomized 1:1 to receive the NASCA regimen or nab-paclitaxel plus PC. Methods: This single-arm phase II trial enrolled 37 patients with histologically or
gemcitabine (AG). The NASCA regimen was administered in 3-week cycles for up to 8 cycles. cytologically confirmed unresectable locally advanced or metastatic pancreatic ductal
Patients without disease progression continued treatment with surufatinib, S-1, and cam- adenocarcinoma (NCT06393166). The study aims to increase the objective response rate
relizumab, while the control group received AG regimen , q3w. Treatment continued until (ORR) of the first-line sequential nab-P/Gem followed by mFOLFOX (nab-P/Gem-
disease progression, unacceptable toxicity, withdrawal of consent, or death. The primary mFOLFOX) regimen from 50% to 68% with the addition of serplulimab and HLX04. The
endpoints were DLTs and the RP2D of surufatinib in phase Ib, and the ORR in phase II. study employed Simon’s minimax two-stage design, with a total of 23 patients achieving
Results: As of Dec 16, 2024, 96 patients were enrolled in the study (6 in phase Ib, 90 in Phase objective responses, thereby meeting the predefined primary endpoint. Secondary end-
II). In phase Ib, the RP2D for surufatinib was determined to be 200 mg (1 DLT out of 6 points included progression-free survival (PFS), overall survival (OS), disease control rate
patients). In phase II, 45 patients were assigned to each group. Baseline characteristics were (DCR), and safety. Results: Among the 37 patients analyzed, the average age was 62
balanced between the groups. Most patients had metastatic disease (82.2% in NASCA vs years, and 22 patients (59.5%) were male. At baseline, patients had an adequate nutritional
77.8% in AG). The NASCA group showed a confirmed ORR of 51.1% (23/45) versus 24.4% (11/ and performance status, with a mean BMI of 21.6 kg/m2. Distant organ metastases were
45) in the AG group (OR: 3.2, 95% CI 1.3-8.2; p = 0.01). The median PFS were 7.9 and present in 34 patients (91.9%), with the liver being the most common site (n = 26, 70.3%).
5.4 months (HR: 0.63, 95% CI 0.40-0.99, p = 0.046), with 12-month OS rates of 55.5% (95% CI The confirmed ORR was 67.6% (95% CI, 49.5-82.6), including 1 patient with a complete
39.4-68.9) in NASCA group and 52.7% (95% CI 36.6-66.5) in AG group. In the subgroup response (CR), meeting the primary endpoint. Only 1 patient had progressive disease (PD)
analysis of PFS, the NASCA group showed longer PFS in male patients (HR: 0.56, 95% CI 0.31-
as the best response, yielding a DCR of 97.1% (95% CI, 84.7-99.9). As of the data cutoff in
1.01), those with metastatic disease (HR: 0.57, 95% CI 0.35-0.94), and patients without liver
November 2024, the median follow-up was 6.1 months. The median PFS was 10.5 months
metastasis (HR: 0.45, 95% CI 0.23-0.90). Furthermore, multiplex immunohistochemistry was
(95% CI, 9.7-not reached), with a 6-month PFS rate of 80.0% (95% CI, 65.5-97.7). The
performed on baseline tissue samples of 26 NASCA patients. The ratio of M1/M2 macrophage
median time to response (TTR) was 1.5 months, and the median duration of response
percentages was significantly higher in patients with PR than those with SD and PD (p = 0.04).
(DOR) was 9.3 months. The OS remains immature. The incidence of treatment-related
Using the median as a cutoff, patients with higher levels of M1/M2 cells (p = 0.039), CD8+ cells
(p = 0.0024), and CD8+PD-1+ cells (p = 0.0064) in the stroma had longer PFS than those with adverse events (TRAEs) was 83.8%, with grade $3 TRAEs occurring in 46.0% of patients.
lower levels. For Grade 3 and 4 TEAEs, the most frequently observed events in the NASCA Hematologic toxicities were the most common treatment-emergent adverse events
group were decreased white blood cell count (31.1%), decreased neutrophil count (33.3%), (TEAEs), and no fatal AE were observed. Overall, the treatment was manageable, and no
and decreased lymphocyte count (20.0%). Similarly, these events were common in the AG new safety signals were identified. Conclusions: First-line serplulimab and HLX04
group (26.7%, 28.9%, 8.9%, respectively). Conclusions: The NASCA regimen demonstrated combined with nab-P/Gem-mFOLFOX demonstrates clinical feasibility and promising
promising efficacy with a manageable safety profile, showing a significantly higher ORR and preliminary outcomes in advanced PC. Further follow-up is required to confirm the survival
longer PFS compared to AG group in patients with locally advanced or metastatic PDAC. benefits, and following analyses are needed to explore the mechanisms underlying the
Further studies are warranted to confirm these findings. Clinical trial information: efficacy of this novel regimen. Clinical trial information: NCT06393166. Research Sponsor:
NCT05218889. Research Sponsor: None. None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
310s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4163 Poster Session 4164 Poster Session

AI-based predictive tool for detection of ctDNA in pancreatic adenocarci- Using live true single-circulating tumor cell comprehensive genomics to
noma using nationwide comprehensive genomic profiling data. First Author: show clonal evolution and tumor heterogeneity in pancreatic cancer
Hiroaki Ikushima, Department of Respiratory Medicine, Graduate School of Medicine, management. First Author: Mandana Kamgar, Medical College of Wisconsin Can-
The University of Tokyo, Tokyo, Japan cer Center, Milwaukee, WI
Background: Comprehensive genomic profiling (CGP) has become a cornerstone of Background: Traditional tissue and CtDNA biopsies have limitations in pancreatic ductal
precision oncology, with liquid biopsy expanding its applicability. However, in some cases, adenocarcinoma (PDAC) due to sampling bias and inability to capture tumor evolution
circulating tumor DNA (ctDNA) is undetectable in liquid CGP, limiting the ability to assess effectively. ctDNA offers Tx decisions, longitudinal monitoring for recurrence, real-time
genetic alterations. Identifying the optimal timing for liquid CGP remains a challenge. This tumor evolution, and responses. However, ctDNA success is limited by its low sensitivity
study focuses on pancreatic adenocarcinoma, using nationwide CGP data and explainable often with NMDs. While single circulating tumor cell (sCTC) genomics may offer higher
AI methods to identify clinical factors associated with ctDNA detection. Based on these sensitivity, but challenged with selective capture of live sCTCs without leukocyte con-
factors, we develop an easy-to-use AI tool to predict the probability of ctDNA detection in tamination. We address limitations by identifying both clonal and rare sub-clonal mutations
real-world settings. Methods: We conducted a retrospective analysis of nationwide CGP in PDAC. We report CellBiopsy assay to capture and release of sCTCs, facilitated by ctDNA-
data collected from Jun 2019 to Dec 2023, covering 99.7% of CGP performed in Japan. integrated comprehensive genomic profiling (CGP) at a true single-cell. Methods: In an
Cohort 1 included 4,110 pancreatic adenocarcinoma cases analyzed by FoundationOne observational study, ten advanced PDAC patients receiving SOC were accrued with MCW IRB
CDx, while Cohort 2 comprised 2,220 cases analyzed by FoundationOne Liquid CDx (F1L). approved protocol (PREDICT-MCW NCT ID:NCT05802069). All patients gave consent to
Using clinical information available prior to CGP, we developed an eXtreme Gradient investigational interventions. Live sCTCs were isolated at baseline (BL) and follow-up (FL)
Boosting (XGBoost)-based predictive model to estimate ctDNA detection and employed using OncoIndx Ikon sCTC assay in 10 mL of blood. Live CTCs were captured using glass
SHapley Additive exPlanations (SHAP) analysis to elucidate contributing clinical factors. A beads with anti EpCam antibody and released in 96 well plate assay. DNA from individually
smartphone application was deployed using the refined model. The app’s performance was captured sCTCs was linearly amplified, followed by target enrichment using OncoIndx CGP
tested with Cohort 3, consisting of 629 pancreatic adenocarcinoma cases tested by F1L assay. Sequencing libraries were prepared and sequenced on Illumina NextSeq2000 (5003
between Jan 2024 and Dec 2024. Results: In Cohort 1 (tissue), 98.5% of cases harbored depth). ctDNA underwent deeper sequencing at 10000x coverage. Data was processed using
mutations in either KRAS, TP53, CDKN2A, or SMAD4, confirming their role as surrogate iCare software for sequence alignment and variant calling. Results: Prospectively, 74 live
markers for tumor-derived DNA detection. The predictive AI model for ctDNA detection, sCTCs were isolated at baseline and follow up (mean sCTC distribution 7). Post SOC
trained on Cohort 2 (liquid) data, achieved an AUROC of 0.754. SHAP analysis identified key treatment, 50% of the patients (5/10) exhibited a 30% reduction in sCTC count at FL. NRAS
predictors, including liver metastasis, the number of metastatic organs, performance mutations were the most frequent alteration observed in sCTCs (40.5%), followed by HRAS
status, response to recent therapy, interval from diagnosis to blood collection, and (27%) and TP53 (23%). Paired ctDNA predominantly revealed KRAS G12 variants (40%).
treatment line. Notably, patients with liver metastases exhibited a significantly higher rate Additional divergent molecular alterations in sCTCs were accounted for in NRAS, TP53,
of ctDNA detection (p , 0.001) compared to those without, whereas patients with SMAD4, and PIK3CA-MTOR-AKT pathways, providing insights into mechanisms of treatment
peritoneal metastases demonstrated a lower rate of ctDNA detection (p , 0.01). A refined resistance and disease aggressiveness. Samples with co-occurring NRAS and TP53, or
model incorporating representative predictors was deployed as a smartphone application. SMAD4, mutations along with ERBB2 amplification, were associated with aggressive dis-
When tested on Cohort 3 (liquid), the application demonstrated predictive accuracy with an ease. In FL sCTCs genomics revealed evolving molecular profiles enriched for activating
AUROC of 0.769 (sensitivity: 0.707, specificity: 0.769) and a Brier score of 0.194. variants in the PIK3CA-MTOR-AKT pathway compared to bulk tissue and ctDNA genomics at
Conclusions: This study identified clinical factors predictive of ctDNA detection in liquid BL. Conclusions: Compared to ctDNA, sCTC CGP revealed heterogeneous molecular profile
CGP for pancreatic adenocarcinoma using explainable AI methods and nationwide CGP in PDAC, offering precise insights into tumor heterogeneity, clonal evolution, disease
data. Based on these findings, a smartphone application was developed to predict the progression, and treatment outcome. Integrating paired DNA profiling of ctDNA and sCTC
probability of ctDNA detection. By facilitating optimal timing of liquid CGP, this app has the DNA may provide a more CGP landscape of PDAC. Ongoing analyses aim to evaluate
potential to enhance patient access to effective therapies, contributing to improved clinical temporal dynamics of CGP using ctDNA and sCTC DNA assay for advancing personalized
outcomes. Research Sponsor: None. management of PDAC. Clinical trial information: NCT05802069. Research Sponsor: None.

4165 Poster Session 4166 Poster Session

Homologous recombination deficiency (HRD) profiling in Chinese pancre- The differential effect of stromal genes on gemcitabine/nab-paclitaxel (GN)
atic ductal adenocarcinoma: Implications for platinum-based and GN/cisplatin (GCN) outcomes in advanced pancreatic adenocarcinoma
chemotherapy. First Author: Yanxia Wang, The First Affiliated Hospital, Sun Yat- (aPDAC). First Author: Himil Mahadevia, Mayo Clinic Florida, Jacksonville, FL
Sen University, Guangzhou, China Background: GN is a front-line therapy for aPDAC. A Phase I/II study demonstrated that
Background: Pancreatic ductal adenocarcinoma (PDAC) is a highly aggressive cancer with poor GCN has a higher overall response rate and median overall survival (mOS). Our previous
prognosis. While platinum-based therapies have demonstrated some therapeutic benefits, their study found no difference in outcomes among patients with DNA damage repair gene
associated toxicity underscores the need to identify patients who are most likely to respond. mutations; however, stromal gene expression correlated with mOS in patients receiving
Homologous recombination deficiency (HRD) has been linked to improved sensitivity to GCN. Here, we further evaluate differences in the outcomes with GCN and GN by site of
platinum-based therapies, but its role in PDAC, particularly beyond BRCA1 and BRCA2 (BRCA1/ biopsy. Methods: PDAC samples (n = 4,463) were analyzed by NGS (NextSeq/NovaSeq) or
2) mutations, remains poorly understood. Methods: A retrospective analysis was conducted on RNA (NovaSeq) (Caris Life Sciences, Phx, AZ). Expression of stromal and related genes
264 Chinese patients diagnosed with PDAC. Genomic data were obtained using a targeted next- (ACTA2, ADIRF, HAS2, IL-6, MMP-2, MMP-9, SPARC, STAT3, TBGB1, TGFB2, TGFBR3, IDO1,
generation sequencing (NGS) panel, which included: (1) 28 canonical homologous recombi- HLA-DRB4, VEGFB) from different biopsy sites [high expression (H) .50% of RNA
nation repair (HRR) genes (BRCA1, BRCA2, PALB2, ATM, ATR, BAP1, BARD1, BRIP1, CDK12, transcripts per million] was correlated with outcomes to GCN or GN. mOS was obtained
CHEK1, CHEK2, EMSY, FAM175A, FANCA, FANCC, FANCD2, FANCL, FANCI, MRE11, NBN,
from insurance claims and calculated from first treatment to last contact. The hazard ratio
PPP2R2A, PTEN, RAD50, RAD51B, RAD51C, RAD51D, RAD54B, and RAD54L), along with their bi-
(HR) was calculated by the Cox proportional hazards model, and p-values were calculated
allelic loss-of-function (BILOF) status; (2) 8 genes associated with other DNA damage repair
(DDR) pathways (TP53, CDH1, EPCAM, MLH1, MSH2, MSH6, PMS2, and STK11); and (3) an using the log-rank test. Results: 4325 patients [primary biopsy (PT), n = 1,878; non-liver
integrated HRD score, calculated as the unweighted sum of loss of heterozygosity (LOH), metastatic biopsy (N-LM), n = 818; Liver biopsy (LM), n = 1,629] received GN while 138
telomeric allelic imbalance (TAI), and large-scale transitions (LST). HRD score $ 38 and/or patients (PT, n = 45; N-LM, n = 28; LM, n = 65) received GCN. GCN was associated with
BRCA1/2 BILOF was predefined as HRD positive. The association between HRD status and longer mOS than GN [Δ: 5.2 months (m), HR: 0.76, 95% CI 0.63-0.92, p = 0.01]. GCN was
clinical outcomes in patients treated with first-line platinum-based therapies was systematically associated with longer mOS compared to GN in LM (Δ: 5.6 m, HR: 0.66, 95% CI 0.50-0.87, p
analyzed. Results: Among the 264 PDAC patients, 6.4% (n = 17) were classified as HRD- = 0.003), but it was not significant in PT (Δ: 4.6 m, p = 0.13) and N-LM (Δ: 4.4 m, p = 0.35).
positive, a larger group compared to the 1.9% (n = 5) with BRCA1/2 BILOF. Overall, 19.3% (51/ Median MMP2 (38.5 vs. 163.1 vs. 162.2), VEGFB (14.4 vs. 16.9 vs. 16.2) and TGFBR3 (11.4
264) of patients harbored mutations in HRR genes. Among these, 4.9% (n = 13) had BRCA1/2 vs. 15 vs. 16.4) expression were lower in LM compared to PT and N-LM while TGFB1 (41.6
mutations, with 38.5% (n = 5) exhibiting BILOF. The most frequently mutated HRR genes vs. 34.5 vs. 39.8) and IL6 (1.61 vs. 1.19 vs. 1.41) expression were highest in LM (p , 0.05).
included ATM (4.2%) and BRCA2 (3.8%), followed by CHEK2 (1.5%), BRCA1 (1.5%), ATR (1.5%), In LM, IL6-H (Δ: -11 m, p = 0.037) and TGFB1-H (Δ: -5.9 m, p = 0.10) were associated with
and FANCA (1.5%). The median HRD score was notably higher in patients with HRR gene BILOF worse post-GCN survival compared to GN (p = 0.82 and p = 0.57). Whereas, in N-LM,
(25.5) compared to those with non-BILOF (14). In the first-line platinum chemotherapy cohort (n TGFBR3-H trended towards longer post-GCN survival (Δ: 12 m, p = 0.18), while ADIRF-H
= 133), HRD-positive patients exhibited significantly improved progression-free survival (PFS), trended towards shorter post-GCN survival (Δ: -12 m, p = 0.056) compared to GN (p = 0.42
with a median PFS of 20.5 months, compared to 11.3 months in HRD-negative patients (HR = and p = 0.10). While in PT, MMP2-H (Δ: 11.3m, HR: 0.46, p = 0.16), TGFB1-H (Δ: 11.3 m, HR:
0.385, 95% CI: 0.177-0.84, P = 0.012). Notably, patients with BRCA1/2 BILOF derived substantial 0.33, p = 0.09), HLA-DRB4-H (Δ: 12.9 m, HR: 0.44, p = 0.11) and VEGFB-H (Δ: 10.7, HR: 0.34,
clinical benefit from first-line platinum-based therapies, with no instances of disease pro- p = 0.09) trended towards longer post-GCN survival compared to GN (MMP2-H, Δ: 3.4 m, p
gression or death during the treatment period. Conclusions: HRD profiling, defined by an HRD = 0.03, TGFB1-H, p = 0.661, HLA-DRB4-H, p = 0.67, VEGFB-H, Δ: 1.3 m, p = 0.03).
score threshold of $38 and/or BRCA1/2 BILOF status, is a valuable biomarker for predicting
Conclusions: GCN is associated with improved mOS compared to GN, especially in
response to platinum-based chemotherapy in PDAC. This study suggests that scar-based HRD
[Link] gene expression in the liver differs from that in N-LM and the pancreas. While
marker and gene BILOF status could serve as predictive markers for PDAC personalized therapy.
Clinical trial information: [2024]138. Research Sponsor: National Natural Science Foundation of
high stromal gene expression trends towards worse post-GCN survival in LM, it is as-
China; 82330065, 30900650, 81372501, 81572260, 81172232, 31430030. sociated with improved survival in N-LM and PT. Further validation is needed to understand
the impact of stromal gene expression in different tumor sites on survival outcomes and
signature development. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 311s

4167 Poster Session 4168 Poster Session

CheMo4METPANC: Combination chemotherapy (gemcitabine and nab- Phase 1/2 study of nivolumab and ipilimumab combined with gemcitabine,
paclitaxel), chemokine (C-X-C) motif receptor 4 inhibitor (motixafortide), nab-paclitaxel, and adaptive stereotactic body radiotherapy in bordeline
and immune checkpoint blockade (cemiplimab) in metastatic treatment- resectable, locally advanced or metastatic pancreatic cancer (LAPTOP). First
naı̈ve pancreatic adenocarcinoma—Updated clinical and translational Author: Inna Markovna Chen, Department of Oncology, Herlev Gentofte Hospital, Herlev,
findings. First Author: Gulam Abbas Manji, Columbia University Medical Center/ Copenhagen, Denmark
New York-Presbyterian Hospital, New York, NY Background: This phase 1/2 study (NCT04247165) evaluated the safety and efficacy of
Background: Metastatic pancreatic ductal adenocarcinoma (mPDAC) is a uniformly fatal combining nivolumab, ipilimumab, gemcitabine, nab-paclitaxel, and adaptive stereo-
disease with an immunosuppressive tumor microenvironment (TME). In our KPC mouse tactic body radiotherapy (SBRT) in patients with borderline resectable (BRPC), locally
model study, targeting the C-X-C motif chemokine receptor 4 (CXCR4)/C-X-C motif chemokine advanced (LAPC), or metastatic pancreatic cancer (mPC). Methods: Treatment-naı̈ve
ligand 12 (CXCL12) axis in combination with aPD-1, and gemcitabine improved survival when patients with BRPC, LAPC or mPC received 28-day cycles of nivolumab (3 mg/kg),
compared to mice treated with gemcitabine or other combinations. The goal of this first-in- ipilimumab (1 mg/kg single dose), and gemcitabine (800 mg/m2) with nab-paclitaxel
human study was to evaluate safety, radiologic response rate, and change in tumor mi- (100 mg/m2 on days 1, 8, and 15). Starting in cycle 3, patients underwent MRI or CT-
croenvironment (TME) elicited by motixafortide (CXCR4i), cemiplimab (aPD1), gemcitabine, guided adaptive SBRT (8 Gy x 3 fractions) targeting the primary pancreatic tumor.
and nab-paclitaxel (MCGN) in treatment-naı̈ve mPDAC. Methods: CheMo4METPANC is an Primary endpoint: safety, defined by the incidence of treatment-related adverse events
open label, multicenter, investigator-initiated, study evaluating MCGN in mPDAC
(TRAEs) leading to discontinuation, monitored via a Bayesian stopping rule for rates
(NCT04543071). Here we report the updated results of the signal seeking phase of this study.
exceeding 30%. Secondary endpoints: OS, PFS, ORR, DCR, DOR, and resection rate.
The primary aim was to study the safety of MCGN. All patients received pre- on-treatment and
optional on-progression biopsies. Single nucleus RNA sequencing (snRNAseq) and quan-
Exploratory endpoint: immunological changes. Results: A total of 55 patients received
titative multiplex immunofluorescence (qmIF) were used to characterize the TME. Results: A at least one treatment (BRPC: n=1, LAPC: n=23, mPC: n=31), with a median follow-up of
total of 11 patients (1 over-enrolled) participated in the study at Columbia and Brown 27.6 months (IQR 26.3-34.4) calculated via reverse Kaplan-Meier. Grade 3-4 TRAEs
Universities (11/9/2020-3/3/2023). The median age was 58 years. As of 04/22/24 (median occurred in 70.8% of BRPC/LAPC and 71.0% of mPC patients. No grade 5 TRAEs were
follow up 23 months), 7 (63%) and 3 (27%) patients experienced a partial response (PR) and observed. Treatment discontinuation due to TRAEs occurred in 20.8% (BRPC/LAPC) and
stable disease, respectively. One patient experienced radiologic resolution of hepatic me- 9.7% (mPC). Median OS: 23.0 months (95% CI: 11.4–NR) for BRPC/LAPC and
tastasis and underwent definitive radiation therapy to the primary tumor. A second had a 11.2 months (6.8–15.8) for mPC. Median PFS: 14.9 months (8.7–24.2) for BRPC/LAPC
sustained PR (11 months) and underwent pancreaticoduodenectomy and hepatic wedge and 6.1 months (3.8–8.4) for mPC. ORR: 33.3% (15.6–55.3) for BRPC/LAPC and 19.4%
resection which revealed a pathologic complete response within the hepatic and primary (7.5–37.4) for mPC. DCR: 75.0% (53.3–90.2) for BRPC/LAPC and 61.3% (42.2–78.2) for
lesion. Median progression free survival (PFS) was 9.6 months. The most common adverse mPC. Median DOR: 8.9 months (4.0–12.7). Nine (37.5%) BRPC/LAPC patients and three
events experienced while on the study combination included skin hyperpigmentation (11/11), (9.7%) mPC patients underwent resection. Preliminary analyses revealed treatment-
alopecia (10/11) and injection site reaction (9/11). The most common grade 3 or greater induced T-cell activation, particularly after SBRT, with upregulation of activation and
adverse events were anemia (5/11) and rash (3/11). Analysis of the TME revealed an increase stemness markers in patients with durable clinical benefit. Conversely, an increase in
in intratumoral CD8+ T-cells in all patients, and that patients achieving a PR were found to senescence markers was observed in the rest of the cohort. Conclusions: The com-
have higher proportions pre-treatment of CXCL12-producing cancer associated fibroblasts, a bination of nivolumab, ipilimumab, gemcitabine, nab-paclitaxel, and adaptive SBRT
potential marker of response. Conclusions: Preliminary results from this pilot study of MCGN demonstrated an acceptable safety profile and efficacy supporting further investigation
in mPDAC were promising, with a PR rate of 63% and disease control rate (DCR) of 91%. Based in BRPC, LAPC, and mPC patients. Clinical trial information: NCT04247165. Research
on these results, the study was amended to transition to a randomized phase 2 trial testing Sponsor: None.
MCGN compared to GN (2:1; N = 108). The primary endpoint is PFS. The phase 2 study is
actively enrolling patients and incorporates optional paired research tumor biopsies. Clinical
trial information: NCT04543071. Research Sponsor: BioLine Rx and Regeneron
Pharmaceuticals.

4169 Poster Session 4170 Poster Session

Clinico-genomic characterization of PALB2-mutated pancreatic Prognostic and predictive impact of next-generation sequencing in meta-
adenocarcinoma. First Author: Jonathan W. Lee, Memorial Sloan Kettering Cancer static pancreatic ductal adenocarcinoma. First Author: Christopher Schumacher,
Center, New York, NY University of Pittsburgh, School of Medicine, Pittsburgh, PA
Background: Pancreatic adenocarcinoma (PDAC) with germline (g) or somatic (s) mutations in Background: Outcomes in de novo metastatic pancreatic ductal adenocarcinoma (mPDAC) remain
BRCA1/2 and PALB2 exhibit unique molecular characteristics and predict response to platinum- dismal, although considerable heterogeneity exists in responses to therapy and survival. With in-
based chemotherapy and PARP inhibition. However, distinct features of PALB2 and PDAC are not creased use of tumor-cell next-generation sequencing (NGS) in mPDAC, improved understanding of
well described. Herein, we characterize distinct clinico-genomic features of patients (pts) with g/s the prognostic and predictive value of genomic alterations is needed. Methods: Between 2019-2023,
PALB2 and PDAC. Methods: Institutional databases and cBioPortal were queried to identify pts with 949 PDAC patients (pts) underwent targeted NGS (Oncomine), of which 161 had de novo metastatic
g/sPALB2 and PDAC. Pts with PALB2 variants of unknown significance (VUS) were excluded disease. Progression-free survival (PFS) after front-line therapy and overall survival (OS) were
(annotation from OncoKb, ClinVar). Demographic data and clinical outcomes abstracted from correlated with clinical parameters and genomic alterations. Long-term survival (LTS) was defined as
medical record. Detailed mutational analysis obtained from cBioPortal. Zygosity determined with OS in the top quartile. Results: In pts who underwent at least one dose of systemic therapy (80%,
FACETS. Progression-free survival (PFS) and overall survival (OS) estimated with Kaplan-Meier n=128), mPFS and mOS were 6.6 and 10.4 months. Pts who received 5-FU-based front-line therapy
Method. Results: N = 29 pts with pathogenic/oncogenic g/sPALB2 and PDAC identified between (97% received a triplet with oxaliplatin and irinotecan or liposomal irinotecan) had improved outcomes
2011-2024. N = 25 (86%) gPALB2 (+/- sPALB2) and N = 4 (14%) sPALB2 (no gPALB2); N = 13 sPALB2 compared to pts who underwent gemcitabine-based therapy (89% gemcitabine/nab-paclitaxel [gem/
excluded as VUS. Median age (range): 57 years (38-78) gPALB2 and 63 years (43-73) sPALB2. N = 23 nab-P]; mOS 13.2 vs 7.3 months, HR 0.66, 95% CI 0.45-0.97). Tumor mutational burden (TMB),
(79%) white; N = 16 (55%) female. Stage IV at diagnosis: N = 11 (44%) gPALB2; N = 2 (50%) sPALB2 pathogenic mutations in genes involved in cell-cycle regulation (CCR) and DNA damage response
cohort. In gPALB2 cohort (N = 25), 4 (16%) had personal history of cancer (N = 2 thyroid, N = 1 uterine, (DDR), and loss of SMAD4 were not associated with broad differences in PFS/OS. As expected, in pts
N = 1 CLL) and N = 17 (68%) had family history of cancer (N = 7 breast/prostate/ovarian, N = 1 treated with 5-FU/platinum, mutations in DDR genes were associated with LTS (p = 0.04). KRAS
pancreas). KRAS and TP53 variants co-occurred in 84% and 36% of gPALB2 and 50% and 50% of mutational status was prognostic (Table; p = 0.02 for OS). Surprisingly, KRAS wild type (WT) pts had
sPALB2 cases, respectively. N = 9 (56%) of patients with a gPALB2 mutations showed biallelic loss of poorer PFS/OS compared to KRAS-mutated pts. However, 3/12 KRAS WT pts, all with Class II BRAF and
PALB2 (N = 6 by LOH, N = 3 somatic LOH). None of four patients in sPALB2 cohort (negative gPALB2) TP53 loss-of-function (LOF) mutations, were unable to tolerate any systemic therapy. In 3 KRAS WT
had biallelic loss. Median TMB (mt/Mb): 4.10 (0.80-9.10) gPALB2; 3.85 (2.00-5.80) sPALB2. For pts with LTS, 2 had mutations in DDR genes and TMB . 10 Mut/Mb, while 2 had mutations in CCR
stage IV gPALB2 (N = 11), median PFS 4.2 months (95% CI 2.4, NR) and median OS 12 months (95% genes. Multivariate analysis confirmed improved LTS of pts harboring KRAS G12R compared to G12D
CI 5.5, NR). N = 10 (90%) received platinum therapy, with N = 6 in the first line setting. Durable (OR 0.22, 95% CI 0.07-0.72, p = 0.01). TP53 was mutated in 75% of pts (18% gain-of-function [GOF]).
disease control on PARPi was observed for patients with gPALB2 and sPALB2, including N = 1 WT TP53 was associated with improved survival (mOS 10.7 vs 5.0 months, HR 0.60, 95% CI 0.43-0.84).
sPALB2 with 7 months on 4th-line olaparib and N = 1 gPALB2 with 6 months on 6th line Olaparib. Interestingly, in pts receiving gem +/- nab-P, LTS was lacking in those with TP53 GOF mutations but
Conclusions: gPALB2 and sPALB2 mutations are seen in a small % of PDAC. gPALB2 PDAC presents not LOF mutations (p = 0.13), a trend not seen in pts receiving 5-FU/platinum. Conclusions: Herein we
earlier and is linked to family history of cancer. gPALB2 compared to sPALB2 had higher biallelic show novel prognostic and predictive significance of genomic alterations in mPDAC. Heterogeneity
loss; oncogenic sPALB2 are uncommon. Identification of g/sPALB2 has implication for therapeutics exists in KRAS WT pts with BRAF and TP53 mutations conferring poorer prognosis and DDR and CCR
and screening. Loss of heterozygosity, TMB, telomeric allelic imbalance, large-scale state transi- gene mutations associated with LTS. KRAS G12D is associated with worse outcomes compared to
tions, and implications for treatment will be presented. Research Sponsor: None. G12R while G12V is intermediate. Pts with TP53 GOF mutations may benefit from 5-FU/platinum
upfront. Research Sponsor: None.
gPALB2 = 25
KRAS mutational status, frequency, and survival.
Putative driver mut (%) 21 (84) KRAS Status N % mPFS (months) mOS (months)
Truncating mut 13 (62)
Structural Variant 4 (19) WT 12 7.5 3.1 3.5
Splice mut 4 (19) G12C 2 1.2 2.1 2.1
Zygosity G12D 70 43.5 5.0 6.6
Biallelic 9 (56) G12R 25 15.5 9.7 13.2
Monoallelic 6 (44) G12V 36 22.4 4.9 10.5
Unknown 10 Q61H/R 14 8.7 4.3 5.6
Other 2 1.2 5.5 7.8

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
312s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4171 Poster Session 4172 Poster Session

Interpretable artificial intelligence-driven selection of doublet chemother- First safety analysis of an open-label, single arm phase II trial investigating
apy regimens as second-line treatment in patients with advanced pancreatic the efficacy, safety and quality of life of neoadjuvant chemotherapy with
cancer. First Author: Letizia Procaccio, Veneto Institute of Oncology IOV, IRCCS, Padua, liposomal irinotecan combined with oxaliplatin and 5-fluoruracil/folinic acid
Italy followed by curative surgical resection in patients with hepatic oligometa-
Background: Guidelines recommend second-line (2L) doublet chemotherapy, NALIRI static adenocarcinoma of the pancreas (HOLIPANC). First Author: Dirk Thomas
(liposomal irinotecan + 5-fluoruracil and leucovorin), FOLFIRI or FOLFOX, for patients Waldschmidt, Universitätsklinikum Köln, Cologne, Germany
(pts) with metastatic pancreatic ductal adenocarcinoma (mPDAC) after failure of Background: Aim of the prospective single arm HOLIPANC trial (NCT04617457) is to
gemcitabine+Nab-paclitaxel (GemNabP). A head-to-head comparison between doublets evaluate the efficacy and safety of multimodal treatment in pancreatic oligometastatic
has not been performed. We aimed to apply interpretable artificial intelligence (IAI, i.e., disease. Methods: Patients with hepatic oligometastatic pancreatic cancer receive up
AI systems in which the prescription logic can be understood) methods on real-world to 8 cycles of a combination of liposomal irinotecan (nal-IRI, 50mg/m2) with 5-fluouracil
data to establish which pts should receive NALIRI vs other doublets to maximize the (5-FU 2400mg/m2)/folinic acid (FA, 400mg/m2) and oxaliplatin (OX, 60mg/m2) (Nal-
benefit in this setting. Methods: In this observational cohort study, we compared IRIFOX) as neoadjuvant therapy followed by curative intended surgical resection of the
progression-free survival (PFS) of consecutive pts with mPDAC who received 2L primary tumor and liver metastases. Here, we present the first preplanned safety
doublets after GemNabP failure at 42 Italian centers between 2013 and 2023. The analysis with patients (pts) that were enrolled between 10/2021 and 04/2024.
dataset was randomly split into a training set (70%) and a test set (30%). In the former, a Results: A total of 56 pts were included in the analysis, of which 43 (77%) pts re-
counterfactual Cox proportional hazard model, including baseline characteristics to infer ceived at least 4 cycles and 16 (29%) pts 8 cycles of chemotherapy. Dose reductions of
the 12-month PFS probability for a given patient under each regimen, was trained. An nal-IRI were required in 24 (43%) pts. Treatment-emergent adverse events (TEAE) were
Optimal Policy Tree (OPT), a state-of-the-art IAI-based method, was used to read the observed in 52 (93%) pts, 40 (72%) TEAEs were related to the neoadjuvant chemo-
complete reward matrix by training a decision tree with the counterfactual predictions, therapy. Grade 3-4 TEAEs occurred in 30 (54%) pts, most common were gastrointestinal
and OPT recommendations were validated in the test set. The potential gain of the new disorders, i.e. diarrhoea (n = 5), vomiting (n = 4) and nausea (n = 4). Hepatobiliary
policy was evaluated by 12-month PFS net-benefit curves. Results: Among 571 eligible disorders (cholangitis (n = 5), cholestasis (n = 4)) and increased gamma-
pts, 209 (36.6%), 209 (36.6%) and 153 (26.8%) received NALIRI, FOLFOX and FOLFIRI, glutamyltransferase (n = 4) were also frequent but generally not considered in rela-
respectively. Median PFS was similar among the three groups (3.3 months for NALIRI, tion to the chemotherapy. Grade 3 anemia and leukopenia were observed in 3 (5%) and 2
3.5 months for FOLFOX and 3.6 months for FOLFIRI), with a long-term benefit observed (4%) pts respectively. Serious TEAEs occurred in 23 patients (41%). Next to the above
only in the NALIRI group (12-month PFS 12.0% vs 2.6% for FOLFIRI and 5.2% for mentioned gastrointestinal and hepatobiliary disorders, dehydration (n = 3) and in-
FOLFOX). The OPT recommended NALIRI as the preferred regimen for pts with pan- fections (n = 2) were the remaining recurring events. At the time of analysis, resection
creatic head/body cancers, with ECOG PS 0 or with Ca19.9 , 109 U/ml if ECOG PS . 0. was performed in 20 pts (36%). Overall postoperative complications occurred in 7 pts
The net-benefit curves revealed that the OPT consistently outperformed the uniform (35%), n = 2 (10%) were classified as Dindo-Clavien grade $ 3. Clinically relevant
strategies of administering either NALIRI or FOLFOX/FOLFIRI to all pts, attaining a 2.5 pancreatic fistula was shown in 1 patient (5%). While there were no deaths related to
percentage-point net-benefit at a threshold probability of roughly 9%. Conclusions: Our chemotherapy or surgery, 2 pts died because of tumor progression during follow-up
findings show that 2L NALIRI can offer long-term PFS advantage in a subgroup of within 28 days of chemotherapy administration or surgery. Conclusions: This analysis
mPDAC pts compared with other doublets. The AI-derived policy provides a higher net shows for the first time safety data from a prospective clinical trial of multimodal
benefit than treating all pts with NALIRI, avoiding unnecessary clinical and financial treatment in oligometastatic pancreatic cancer. As we show, the concept of neoadjuvant
toxicity. Research Sponsor: None. chemotherapy followed by surgery is safe, the toxicity and overall morbidity is not
elevated compared to available data from the NAPOLI-3 trial, even in combination with a
following tumor resection after neoadjuvant treatment. Clinical trial information:
NCT04617457. Research Sponsor: None.

4173 Poster Session 4174 Poster Session

Circulated T-cell exhausted subtypes to predict response in PDAC patients. USP22 as promoter of Treg cell infiltration and modulator of immunotherapy
First Author: Anastasia Xagara, Laboratory of Oncology, School of Health Sciences, efficacy through the PIAS1/P65/CCL22 pathway in pancreatic cancer. First
University of Thessaly, Larissa, Greece Author: Aman Wang, First Affiliated Hospital of Dalian Medical University, Dalian, China/
Background: In-depth analysis of T-cell exhausted subsets in the circulation of Pan- Liaoning, China
creatic Ductal Adenocarcinoma (PDAC) patients may provide insights into novel Background: Immune checkpoint inhibitors targeting PD-1 and CTLA-4 have shown
therapeutic options and predictive biomarkers. We performed a detailed immunophe- success in various cancers; however, their efficacy in pancreatic cancer remains limited,
notypic analysis for both early-stage (resectable) and metastatic (unresectable) PDAC likely due to its immunosuppressive microenvironment. Treg cells are crucial in tumor
patients. Additionally, different T-cell populations were correlated with clinical outcome. immune evasion and suppression. Targeting Treg cells is an emerging strategy in pan-
Methods: Fifty-five treatment naive PDAC patients, twenty-five of which had resectable creatic cancer immunotherapy. Recent studies show that USP22 (Ubiquitin-Specific
disease, and ten healthy donors (HD) were enrolled. Peripheral Blood Mononuclear Cells Protease 22) regulates immunity by deubiquitinating key proteins like PD-L1 and
(PBMCs) were isolated and stained with fluorochrome-conjugated monoclonal anti- STAT1 in tumor cells, while in Treg cells, USP22 controls FoxP3, modulating Treg cell
bodies. Multicolor flow cytometry was performed to determine differences between T- function. This study aims to explore how USP22 regulates the immune microenvironment
cell populations and their correlation with clinical outcome. Results: Advanced disease in pancreatic cancer and identify potential targets to enhance immune checkpoint
patients that harbored high percentages of CD4+PD-1+ Teff cells had longer PFS (median: blockade responses. Methods: Bioinformatics analysis of tissue microarrays and the
190 vs. 100 days, p:0.030) and OS (median: 250 vs. 170 days, p:0.041) while for early- TCGA database was used to assess the correlation between USP22 and Treg cell infil-
stage patients high percentages of CD8+PD-1+Teff displayed longer DFS (median: 422 vs. tration in pancreatic cancer. 2. Tumorigenicity assays, immunohistochemistry, and flow
200 days, p:0.044) and OS (median: Und vs. Und days, p:0.041). For early-stage patients, cytometry were performed in C57BL/6 mice to determine the effect of USP22 on Treg cell
high percentages of both CD4+ and CD8+ T-cells expressing PD-1+TCF1+ (exhausted infiltration. 3. Transcriptomic and proteomic analyses were conducted to explore the
cells) were predictive for survival (CD3+CD4+PD-1+TCF1+: med. Und vs. 277 days, p: mechanisms by which USP22 regulates Treg cell infiltration. 4. In vitro experiments,
0.0041) and (CD3+CD8+PD-1+TCF1+: med. Und vs. 390 days, p:0.042). Additionally, including Western blot, qRT-PCR, immunofluorescence, deubiquitination assays, and
expression levels of PD-1 (MFI levels) were substantially elevated in the PD-1+TCF1- chromatin immunoprecipitation, were used to identify the specific mechanisms of USP22
subset for both early stage CD3+CD4+ (p:0.026), CD3+CD8+ (p: = 0.006) and advanced- in regulating Treg cell infiltration. 5. The combination of PD-1 monoclonal antibody and
nab-paclitaxel chemotherapy was tested to determine if targeting USP22 improves im-
stage, (p:0.0001 and p:0.045, respectively), implying that terminally exhausted (PD-
munotherapy efficacy. Results: Immunohistochemistry and bioinformatics analysis
1+TCF1-) T-cells exhibit higher PD-1 expression than primarily exhausted (PD-1+TCF1+).
revealed a significant positive correlation between USP22 expression and Treg cell in-
For advanced stage patients, high levels of CD57+, a marker of terminally differentiated
filtration in pancreatic cancer. 2. Tumorigenicity assays in C57BL/6 mice showed that
T-cells, in CD3+CD8+ were associated with improved PFS (118 vs. 92 days, p:0.178)and
USP22 promotes tumor growth and Treg cell infiltration. 3. Transcriptomic analysis
OS(271 vs. 152 days, p: 0.019), CD3+CD8+PD-1+TCF1+ ( PFS: med. 260 vs. 60 days, p = found a strong association between USP22 and the TRAF1/NF-kB signaling pathway as
0.0063 ; OS: 271 vs. 90 days, p: 0.003) and CD3+CD8+PD-1+TCF1- T-cells (PFS: med. 188 well as CCL22 expression. 4. In vivo and in vitro experiments confirmed that USP22
vs. 80 days, p = 0.0459 ; OS: 271 vs. 107 days, p = 0.0429). CD57+ T-cells were not activates TRAF1/NF-kB signaling and increases CCL22 release, promoting Treg cell in-
correlated with response in early-stage patients. Conclusions: T-cell exhaustion rep- filtration. 5. USP22 deubiquitinates PIAS1, inhibiting its nuclear translocation and acti-
resents ineffective immune response and in both early and advanced-stage PDAC may vating p65. 6. USP22 knockdown improved therapeutic responses to combined
predict clinical outcome offering opportunities for innovative therapeutic options for this immunotherapy (PD-1 inhibitors) and chemotherapy (nab-paclitaxel).
fatal disease. Research Sponsor: None. Conclusions: USP22 promotes Treg cell infiltration in pancreatic cancer by deubiquiti-
nating PIAS1, inhibiting its nuclear translocation, and activating the TRAF1/NF-kB sig-
naling pathway, which leads to CCL22 release. Targeting USP22 could inhibit cell
proliferation, promote apoptosis, and remodel the immune microenvironment, enhancing
the efficacy of immune checkpoint blockade therapy. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 313s

LBA4175 Poster Session 4176 Poster Session

NAPOLI 3, a phase 3 study of NALIRIFOX in patients with metastatic Biomarkers of response to immunotherapy in pancreatic ductal adenocar-
pancreatic ductal adenocarcinoma (mPDAC): Final overall survival (OS) cinoma (PDAC) with homologous recombination deficiency (HRD). First
analysis and characteristics of the long-term survivors. First Author: Vin- Author: Clara J. Oh, University of Miami Sylvester Comprehensive Cancer Center, Miami,
cent Chung, City of Hope, Duarte, CA FL
Background: PDAC is associated with a paucity of immune effector cells, low antigenicity, and
immunosuppressive factors in the tumor microenvironment (TME); consequently, treatment of
unselected PDAC patients with immune checkpoint inhibitors (ICIs) has been ineffective. In
HRD-PDAC, dual PD-1/CTLA-4 ICI therapy has a response rate of 14-42%. In this study, we
investigated biomarkers of immunotherapy response in PDAC tumors with HRD. Methods: We
generated a murine model of HRD-PDAC and performed transcriptomic analysis of mouse
responders versus non-responders to ICI therapy to identify biomarkers of response. DNA and
RNA sequencing was also performed for patient tumor samples submitted to Caris Life
Sciences. Samples harboring pathogenic or likely pathogenic (P/LP) BRCA1, BRCA2 or PALB2
mutations were classified HRD, the remaining samples non-HRD. Microsatellite instability-high
neoplasms were excluded. TMB-High (TMB-H) was defined as $10 mutations/Mb. PD-L1
positivity was determined by IHC (SP142, $2+, $5%). High genomic loss of heterozygosity
(gLOH-H) was defined as LOH at $16% of segments analyzed (up to 552). Immune cell
The full, final text of this abstract will be available at infiltration was estimated by RNA deconvolution using quanTIseq. Mann-Whitney U, Fisher’s
[Link] on the day of presentation and in the Exact, or Chi-squared tests were used to determine statistical significance (p), with multiple
comparisons corrections as appropriate (q). Results: In murine HRD-PDAC tumors, chronic
online supplement to the June 10, 2025, issue of the Journal platinum exposure was a biomarker for response to ICI. Responding tumors had differential
of Clinical Oncology. enrichment of cytosolic DNA sensing and cGAS-STING-related pathways, and secretomes
significantly enriched for the T-cell attractant chemokines CXCL9 and CXCL10. Of 6396 human
PDAC samples, 4.2% were HRD (2.7% BRCA2, 0.9% BRCA1, 0.6% PALB2 P/LP), and 95.8% non-
HRD. Compared to the non-HRD cohort, the HRD cohort was younger (median age: 66 vs 68
years, p = 0.0006), had lower prevalence of TP53 (57.4% vs 79.0%, q , 0.0001), CDKN2A
(14.1% vs 24.7%, q = 0.0062), and RNF43 (0.8% vs 5.8%, q = 0.0322) mutations, and was more
frequently PD-L1+ (21.7% vs 14.0%, q = 0.0460), TMB-H (6.4% vs 1.9%, q = 0.0001), and gLOH-H
(41.2% vs 9.5%, q , 0.0001). Median infiltration of M1 macrophages was higher in the HRD
cohort (6.2% vs 5.3%, p = 0.0028), while that of M2 macrophages was lower (2.9% vs 3.3%, p =
0.0081).TheHRD cohort demonstrated higher median expression measured in transcripts per
million of CGAS (6.3 vs 5.4 TPM, p = 0.0002), CXCL9 (2.19 vs 1.60 TPM, p = 0.0015), and
CXCL10 (4.65 vs 3.65 TPM, p = 0.0308), phenocopying observations in the murine model.
Conclusions: PDAC tumors associated with canonical HRD variants (BRCA1/2, PALB2) have
distinct genomic, transcriptomic, and TME features which are immune-permissive and explain
the sensitivity of this subgroup of patients to ICI therapy. Understanding the underlying
mechanisms could inform strategies to broaden the impact of ICI in this population. Research
Sponsor: None.

4177 Poster Session 4178 Poster Session

Interim open-label phase 1 results of misetionamide (GP-2250): A small Concurrent mutations in DNA damage repair genes BRCA1, POLE, ATM and
molecule antineoplastic targeting three major transcription factors. First FANCA to predict overall and progression-free survival for patients (pts) with
Author: Anup Kasi, University of Kansas Cancer Center, Fairway, KS metastatic pancreatic ductal adenocarcinoma (mPDAC) treated with che-
Background: GP-2250 is a novel antineoplastic agent demonstrating effective activity on motherapy in combination with dual checkpoint inhibition in the CCTG
preclinical pancreatic cancer models alone or in combination with gemcitabine, and in- randomized PA.7 trial. First Author: Daniel John Renouf, BC Cancer, Vancouver,
hibits c-MYC, NFkB, and HIFa in cancer cells at clinically achievable concentrations. This BC, Canada
open-label phase 1 trial (NCT03854110) evaluates the safety and tolerability of escalating Background: CCTG PA.7 (NCT02879318) was a randomized phase II trial comparing
doses of GP-2250 in combination with gemcitabine as a second-line treatment in adults gemcitabine (G) and nab-paclitaxel (N) with and without dual immune checkpoint inhibition
with advanced pancreatic adenocarcinoma that experienced disease progression with 5- with durvalumab (D) and tremelimumab (T) as 1st-line therapy in pts with mPDAC. Matched
FU based chemotherapy. Methods: GP-2250 dose escalation (starting dose 250 mg plasma and tissue-based sequencing was performed for exploratory correlative biomarker
escalating up to 40 g IV once weekly) followed a Bayesian Optimal Interval design which analysis. Methods: Pts received G+N+D+T (n = 11 run-in, 119 randomized, 2:1 randomi-
transitioned to a 3+3 design. A 1-week run-in of GP-2250 was followed by a full cycle zation) or G+N (n = 61). Long-term trial analysis was performed with a median follow-up time
(3 weeks on, 1 week off) of GP-2250 plus gemcitabine treatment for each of 11 dose of 81.6 months. Correlative analysis was performed for pts with baseline ctDNA sequencing
cohorts. Single-patient cohorts with 100% escalation between cohorts were enrolled until using a 600-gene PredicineATLAS panel (n = 173), with a subset having matched archival
the first DLT (or cohort 4), followed by 3 patient cohorts with 35%245% escalation tissue available for whole-genome sequencing (WGS; n = 46). Cox-based elastic net re-
between cohorts. The DLT assessment period was 5 weeks at each dose. Patients were gression models were used to identify and rank combinations of mutations by their ability to
treated until disease progression or development of unacceptable toxicity. Primary predict survival hazard. Results: Long-term follow up analysis demonstrated no significant
endpoints were safety and tolerability of GP-2250 monotherapy and in combination with difference in median overall survival (mOS) between pts randomized to G+N+D+T vs G+N
gemcitabine. Secondary and exploratory endpoints were preliminary efficacy, pharma- (9.8 vs 8.8 months; hazard ratio (HR) = 0.88; p = 0.46). Median progression-free survival
cokinetics, and pharmacodynamic blood markers. Results: To date, 52 patients have been (mPFS) was also not significantly different between treatment arms (5.5 vs 5.4 months,
enrolled. Five serious adverse events were reported in 3 patients (2 CVAs, pneumonia, and respectively; HR = 0.95, p = 0.77). Landmark analysis demonstrated 4-year survivorship of
abdominal and flank pain in 2 patients), none attributed to GP-2250. Three patients 5.4% in pts treated with G+N+D+T arm compared to 1.6% with G+N (p = 0.07). Two or more
discontinued treatment: 1 disease hyperprogression (at 11 g GP-2250) and 2 neutropenia ctDNA-based mutations (somatic and germline considered separately) in DNA damage
(at 21 g GP-2250), with only 1 event of grade 3 neutropenia “possibly” attributed to GP- repair (DDR) genes BRCA1, POLE, ATM or FANCA was present in 18/173 pts (10.4%) and was
2250. In summary, the addition of GP-2250 did not significantly alter the safety and associated with improved OS with G+N+D+T vs G+N (mOS 26.2 months vs. 7.1 months; HR =
tolerability expected of gemcitabine alone. Twelve patients (23%) had progression-free 0.22 [0.07-0.7]; p = 0.0041, p-interaction = 0.012) as well as PFS (mPFS 14.6 vs. 4.6; HR =
survival (PFS) of $16 weeks, or twice as long as historical gemcitabine treatment alone; 7 0.17 [0.05-0.6]; p = 0.0020, p-interaction = 0.0070). In pts treated with G+N+D+T, partial
patients (13%) had PFS of 24 weeks, and 4 (8%) had PFS of 32 weeks. One patient response (PR) was seen in 63.6% of pts with $2 DDR gene mutations compared to 26.9% in
survived . 2 years while receiving treatment. Seventeen patients (33%) achieved stable other pts (p = 0.033), and this effect was not observed with G+N (p = 0.18). The DDR gene
disease and 6 (12%) achieved a partial response by RECIST criteria. While the blood half- biomarker was validated in 5/6 (83%) biomarker-positive samples using archival tissue
life of GP-2250 is ~5 hours, mTOR and AKT biomarker data indicate that the biological half- WGS. Conclusions: The presence of $2 DDR gene mutations was strongly associated with
life is longer, at 4–5 days. These data are within the concentrations and times required for benefit from the combination of chemotherapy with dual immune checkpoint inhibitor
cytotoxicity in all cancer cell lines tested with GP-2250. Conclusions: GP-2250/ therapy, and pts with this signature had prolonged mOS of over 2 years. This represents the
gemcitabine combination therapy showed encouraging safety and tolerability and fa- first prospective study in PDAC to define a predictive biomarker beyond mismatch repair
vorable PFS outcomes compared to gemcitabine alone. These promising results in a deficiency for benefit from immune checkpoint therapy. Given the long-term survival noted
historically difficult to treat pancreatic cancer population warrant progress to later-stage in this subgroup, assessment of DDR gene mutations could be considered as part of routine
studies. This study is funded by Geistlich Pharma AG. Clinical trial information: standard of care testing for mPDAC pts. Clinical trial information: NCT02879318. Research
NCT03854110. Research Sponsor: Geistlich Pharma AG. Sponsor: Predicine; Astra Zeneca; Canadian Cancer Trials Group.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
314s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4179 Poster Session 4180 Poster Session

Clinical outcomes and molecular characteristics of patients with metastatic Pancreatic cancer mortality trends (2018-2023): Exposing racial inequities
pancreatic ductal adenocarcinoma according to involved metastatic sites. in Michigan’s cancer burden. First Author: Mekdes Asfaw, Wayne State University,
First Author: Mahmoud M.G. Yousef, The University of Texas MD Anderson Cancer Rochester Hills, MI
Center, Houston, TX Background: Pancreatic cancer remains one of the most aggressive malignancies, with
Background: It is well documented that Pancreatic ductal adenocarcinoma (PDAC) Black individuals facing significantly worse outcomes and a younger age of onset.
metastasize to the liver had worse outcome than to the lung, but the molecular basis was Despite overall survival improvements in cancer care, racial disparities in pancreatic
less clear. We employ a large Real-World Evidence dataset to evaluate clinical and cancer continue to widen. This study analyzes Michigan’s diverse population to quantify
molecular features of PDAC according to involved metastatic sites. Methods: The Foundry disparities and identify actionable solutions for healthcare equity. Methods: This ob-
software platform was used to query electronic medical records of patients with meta- servational study analyzed pancreatic cancer mortality patterns across Michigan in
static PDAC who underwent Next-Generation Sequencing (NGS) at MD Anderson. Involved adults aged 25 and older were retrieved from the CDC WONDER database (2018–2023)
metastatic sites were extracted using natural language process from imaging reports then using ICD-10 codes for malignant neoplasm of the pancreas. Crude mortality rates
manually verified. Overall survival (OS) was calculated from date of diagnosis. Results: We (CMRs) and Age-adjusted mortality rates (AAMRs) per 100,000 were calculated by age,
identified 1,095 patients with metastatic PDAC diagnosed between May 2003 and Oct gender and race, with 95% confidence intervals (CI) for precision to assess racial
2024. Median follow up was 41.8 months and median OS was 22.8 (95%CI: 20.2-25.2) disparities in mortality outcomes. Temporal trends and annual percentage changes
months. Most patients (52.9%) had multiple metastatic sites including liver, 28.5% had (APCs) were analyzed using Joinpoint regression. Results: From 2018–2023, Michigan
liver only metastases, while 8.7% had lung only metastases. 10% had metastases not reported 10,162 pancreatic cancer deaths, with Black residents (14.1% of the pop-
including liver. Patients with lung only metastasis had the best outcomes (median = ulation) accounting for 1,289 and White residents (who make up 78.9%) for 8,664 deaths.
57.6 months, HR = 0.37 relative to liver only, 95%CI = 0.27-0.52, p = 9.3e-9), followed by Overall, CMR was higher for White residents (18.26 per 100,000) than Black residents
patients with metastases not involving liver (median = 41.3 months, HR = 0.65, 95%CI = (15.21 per 100,000) who experienced a sharper rise in AAMR, increasing by 8.10% [14.36
0.49-0.87, p = 0.003). Patients with liver only metastasis (median OS = 19.5 months) had (13.55–15.16)] compared to 4.92% [12.36 (12.10–12.63)] for White residents. For Black
similar survival to those with multiple sites including liver (median = 19.3 months, HR = 1.1, residents, CMRs increased with age, rising from 8.14 per 100,000 (45-54 years) to 105.78
95%CI = 0.96-1.4, p = 0.13). TP53 was more frequently mutated in patients with liver only per 100,000 (85+ years), peaking at 69.46 (65-74 years) and 95.37 (75-84 years). White
metastasis (84%) and multiple metastases including liver (85%) compared to patients with
residents had lower CMRs overall, starting at 1.48 per 100,000 (35-44 years) and
lung only (73%) and multiple sites not including liver (67%, p = 9.3e-5). GNAS showed lower
gradually increasing to 115.28 per 100,000 in the 85+ [Link] Washtenaw County, Black
frequency (5%) in patients with liver only and patients with multiple metastasis including
residents had a rate of 14.01 per 100,000 and White residents 15.79 per 100,000 with
liver (2%) compared to patients with lung only (8%) or non-liver metastases (10%, p =
similar trends in Genesee, Wayne, and Ingham counties. Treatment inequities com-
0.003). In patients with liquid biopsy (n = 240), lung only metastasis showed significantly
lower positivity rate for mutation detection (50% vs 65% for liver only, 62% for other or
pounded these disparities: Black patients faced 38% lower odds of surgery, 45% longer
multiple not including liver and 79% for multiple including liver, p = 0.009), and lower TP53 delays for chemotherapy, and 27% lower clinical trial enrollment. These findings
detection rate (23% for lung only vs 46% for liver only, p = 0.02). The frequency of KRAS highlight significant racial disparities in pancreatic cancer mortality, treatment access,
mutation and mutant allele distribution were not significantly different in tissue NGS. and outcomes, underscoring the need for targeted public health interventions.
However, patients with lung only metastasis showed significantly less frequent KRAS Conclusions: Our findings reveal significant racial disparities in pancreatic cancer
mutation detection by liquid biopsy (10% vs 52% in liver only metastasis, p = 9.6e-4). outcomes in Michigan, with Black residents experiencing higher mortality rates and a
Conclusions: PDAC patients without liver metastasis have markedly improved OS relative younger age of death than White residents. These disparities reflect systemic barriers,
to patients with liver metastasis, and lower rates of TP53 mutation. Similar frequencies of including delayed diagnosis, fewer surgeries, and limited access to specialized care.
KRAS mutation were found in different patients by tissue testing. However, patients with Addressing these inequities requires bias training, targeted screening for high-risk Black
lung only metastasis had lower positivity rate of ctDNA, and lower detection rate of KRAS populations, and expanded oncology services, while actionable solutions such as patient
mutation by liquid biopsy. Research Sponsor: None. navigation and community-based screening programs can help bridge this healthcare
gap and promote equity. Research Sponsor: None.

4181 Poster Session 4182 Poster Session

Multimodal machine learning predictions of treatment response and sur- Integrative analysis of tumor microenvironment in advanced pancreatic
vival in advanced pancreatic cancer from the COMPASS trial. First Author: Wei cancer: Unraveling genomic and immune landscape for targeted
Quan, University Health Network, Toronto, ON, Canada therapies. First Author: Catia Fava Gaspar, Division of Medical Oncology and He-
Background: Pancreatic cancer is an aggressive malignancy with limited therapeutic matology, Princess Margaret Cancer Centre, University Health Network, University of
options and a poor prognosis. Current approaches to prognostication are limited, es- Toronto, Toronto, ON, Canada
pecially in advanced disease. We explored whether machine learning integrating multi- Background: An understanding of genotype and immunophenotype interactions in advanced
modal data could predict outcomes in advanced pancreatic cancer. Methods: We pancreatic ductal adenocarcinoma (PDAC) is important in designing combination strategies.
developed and evaluated machine learning models predicting disease control rate and In addition, PDAC subtypes may harbor unique tumor immune microenvironments (TMEs) and
one-year survival from the COMPASS trial (NCT02750657). Data modalities included confer differential sensitivity to KRAS inhibitors (KRASi). We aimed to characterize the
clinical features, histopathology, radiology, RNAseq, and whole-genome sequencing baseline TME in advanced PDAC and its relationship with genomic, transcriptomic and clinical
(WGS). After pre-processing, we applied LASSO and XGBoost to each modality and early data. Methods: The COMPASS trial (NCT02750657) investigated whole genome (WGS) and
and late fusion techniques. Hyperparameter tuning and performance assessment were transcriptome (RNA-Seq) sequencing in patients (pts) receiving first line therapy for ad-
performed using repeated nested cross-validation. The PurIST RNAseq classifier served vanced PDAC. We performed multiplex immunohistochemistry (mIHC) to identify 5 immune
cell subtypes (CD8+/CD4+ T cells, Tregs, B cells and macrophages) and CIBERSORT, a
as a baseline. Area under the curve (AUC) was the primary metric. Results: The cohort
deconvolution method that uses gene expression profiles. Statistical analyses were per-
included 260 patients (105 female; median age 64 [IQR 58–70]; 141 treated with
formed using STATA/R software and significance was defined as p-value , 0.05. Multivariate
FOLFIRINOX, 97 with gemcitabine and nab-paclitaxel). 170 (65%) achieved disease logistic regression was used and Kaplan Meier analyses evaluated impact on survival.
control and 168 (65%) survived at least one year. The performance of the machine Results: Of 268 pts, 62 had available tissue samples with mIHC, WGS, and RNA-Seq data (n =
learning models is shown in the Table. Predictions from the unimodal models had limited 21 primary biopsies, n = 41 metastases, 34/41 liver). All 62 pts had KRAS mutations (28 G12D,
correlation with each other (the maximum pairwise correlation averaged across folds 19 G12V, 10 G12R, 5 other) and 29 had KRAS major or minor imbalances. 55 cases (88.7%)
was between clinical and histopathology models, 0.21). The late fusion models up- were classified as classical subtype and HRDetecthi was seen in 10 pts, including 5 with
weighted data modalities with stronger unimodal performance. Conclusions: Multiple BRCA1/2 mutations (4 germline, 1 somatic). In the overall cohort, differences between tumor
individual data modalities can predict outcomes in advanced pancreatic cancer, with and stroma were evident with increased infiltration of CD8 and CD4 Tcells and Tregs in stroma
PurIST serving as a strong baseline. Despite differing predictions across data modalities, (p, 0.001) and increased macrophages (p= 0.0343) in tumor. CIBERSORT in a subset of 51
multimodal integration did not improve prognostic performance in this cohort. Research pts demonstrated increased M0 (p= 0.0035) and M2 macrophages (p= 0.0067) in liver
Sponsor: Ontario Institute for Cancer Research; Princess Margaret Cancer Foundation; metastases compared to primary samples, suggesting a more immunosuppressive TME. A
The Terry Fox Research Institute Marathon of Hope Cancer Centres Network. higher number of B cells were seen in lung metastases (median 207.5 vs. 43.2 vs. 3.7 vs. 1.8
cells/mm2, p= 0.011) compared to abdominal wall, peritoneum and liver, respectively. Pts
AUC for the PurIST baseline, the top 2 unimodal models, and the best fusion
with KRAS major imbalance (n =14, 3 basal like) were found to have higher median numbers of
model for each outcome.
CD8+ (114.5 vs. 25.1 vs. 27.5 cells/mm2, p= 0.038) and CD4+ Tcells (292.1 vs 133.4 vs. 97.4
Outcome Data Modality AUC (95% confidence interval) cells/mm2, p= 0.005) when compared to minor/balanced samples, respectively. Basal-like
Disease control PurIST 0.69 (0.69, 0.70) PDAC had fewer macrophages than classical subtype (median 13.2 vs. 28.2 cells/mm2, p=
Radiomics 0.75 (0.72, 0.79) 0.0103). On survival analysis, pts with HRDetectlo and classical subtype with higher mac-
RNAseq 0.71 (0.70, 0.72) rophage counts had a tendency towards increased survival (median OS: 11.8 vs. 9.5 months,
Fusion (late) 0.71 (0.69, 0.73) p= 0.066). Conclusions: We identified increased CD8/CD4 T cell infiltration in PDAC stroma,
One-year survival PurIST 0.63 (0.62, 0.63) as well as in pts with KRAS major imbalance. Immune cell profiling may complement
DNA mutations 0.64 (0.61, 0.66) molecular profiling as potential biomarkers and warrants further study in this context.
RNAseq 0.57 (0.55, 0.60) Research Sponsor: PM2C/MOHCCN.
Fusion (early) 0.61 (0.56, 0.66)

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 315s

4183 Poster Session 4184 Poster Session

BRCA1/2 and PALB2 short variants (SVs) contributed by clonal hematopoi- 5-FU + Naliri, gemcitabine plus nab-paclitaxel or both regimens given
esis (CH) in liquid biopsies (LBx) from patients with advanced pancreatic sequentially for first line treatment of metastatic pancreatic ductal adeno-
cancer (PC). First Author: Kim Anna Reiss, University of Pennsylvania, Philadelphia, PA carcinoma: A randomized phase II comparative study (FUNGEMAX-
Background: CH results from fitness-enhancing mutations in hematopoietic stem cells. Many PRODIGE 61). First Author: Julien Taieb, Hôpital Europeen Georges Pompidou,
CH somatic variants (SVs) are in cancer-associated genes, including ATM and CHEK2, which do APHP, Université Paris-CIté, CARPEM Cancer Institute, Paris, France
not have a homologous recombination deficiency (HRD) phenotype. SVs in well-established Background: New chemotherapeutic approaches are still needed to improve survival and quality of life in
HRD-associated genes like BRCA1/2 and PALB2 also appear in white blood cells as CH, albeit metastatic pancreatic ductal adenocarcinoma (mPDAC). We have previously published results of two ran-
more rarely. Herein, we report the prevalence of SVs of CH origin in these clinically relevant domized phase II of first-line sequential treatment strategies of intensified FOLFIRI regimen followed by
genes that confound results of PC liquid biopsies (LBx) and study their association with HRD in gemcitabine-based regimens (FIRGEM and FIRGEMAX-PRODIGE 37 studies) with good efficacy and tolerability
results. FUNGEMAX-PRODIGE 61 evaluated 5FU + Naliri (NAPOLI) vs gemcitabine + Nab-paclitaxel (MPACT) vs
tissue biopsies from the same patients. Methods: This study uses a novel variant origin
both regimens sequentially. Methods: Chemotherapy-naive pts with proven mPDAC, bilirubin levels , 1.5 ULN
prediction algorithm to classify the origin of each SV detected by FoundationOneLiquid CDx as and performance status (PS) 0-1 were randomized to receive either the NAPOLI regimen for 2 months, al-
germline, tumor-somatic, or CH, using a combination of sample-specific and dataset-learned ternating with the MPACT regimen for 2 months (arm A), NAPOLI alone (arm B) or MPACT alone (arm C) until
features including fragmentomics (trained and validated against white blood cell standard). progression or limiting toxicity. Using the Schoenfeld method, the primary endpoint was the progression-free
5,625 PC LBx sequenced during routine clinical care was used for broad prevalence data. A survival (PFS) rate at 6 months from (H0) 30% over (H1) 45%, requiring 96 patients per arm (assuming 5% lost
subset with matched tissue biopsies (TBx, n = 536) was used to compare SV detection and true to follow-up). Results: Between 11/2018 and 01/2024, 288 pts were enrolled in 31 French centers and 283
HRD via HRDsig in TBx, a signature validated to predict response to PARP inhibitors across included in the modified intent to treat population (mITT, patients who received at least one dose of treatment).
multiple cancer types. Results: Among 303 PC LBx with a BRCA1/BRCA2/PALB2 SV, 52 (17.2%) Database lock was done on the 20/12/2024. Baseline characteristics were well balanced between the arm A, B
and C (mean age: 65/63/65, female: 47/43/47%, PS-0: 33/34/37%, . 1 metastatic site: 53/48/48%, mean
were predicted to be of CH origin. This percentage is larger than for other BRCA-associated
albumin 39/40/39 g/L). With a median follow-up of 39.2 months, study treatment was discontinued in 89.5%,
canonical cancer types: prostate (14.1%, 139/980), breast (9.7%, 87/902), ovarian (7.0%, 14/ 96.8% and 93.7% of patients for arms A/B/C; median treatment duration were 6.3/3.3/5.3 months, respectively.
200), but less than some non-canonical cancer types (Table). In PC patients with both LBx and In the mITT, neither treatment with MPACT/NAPOLI (HR = 0.76, 95%CI: 0.57-1.02; p = 0.07) nor NAPOLI (HR =
TBx available, 29/536 (5.4%) had an SV in BRCA1/BRCA2/PALB2 detected in LBx: 8/29 in LBx 1.20, 95%CI: 0.90-1.60; p = 0.22) lead to a statistically significant improvement of PFS over MPACT. PFS,
only; 21/29 in both LBx and TBx. 19/29 were predicted germline by the algorithm, were all also Overall survival (OS) and safety data are summarized in the table. Conclusions: The study did not show
detected in TBx, and 15 (79%) of these TBx were HRDsig+. 5/29 (17%) were predicted to be superiority of either the sequence MPACT/NAPOLI or NAPOLI over standard MPACT. However, the sequential
tumor-somatic; two of these were detected in TBx and one (20%) was HRDsig+. Predicted CH MPACT/NAPOLI regimen is feasible, tolerable, and associated with higher rates of 12-mo PFS and 24-mo OS
SVs were detected in another 5 LBx (4 BRCA2; 1 PALB2). None of these were detected in TBx and and less neuropathy and can be considered in patients unfit to receive FOLFIRINOX. NCT03693677. Clinical
trial information: 2024-518143-38-00. Research Sponsor: None.
none of the tumors were HRDsig+. Conclusions: While the majority (58%) of BRCA1/BRCA2/
PALB2 SV+ PC LBx harbored a predicted germline SV, 25% harbored tumor-somatic SV and 17% Arm B Arm C
had SV exclusively predicted as CH-derived. Determining the cellular origin of BRCA1/BRCA2/ Arm A (MPACT/NAPOLI) (NAPOLI) (MPACT)
PALB2 in PC is essential given the potential impact on treatment selection. Research Sponsor: PFS
Foundation Medicine. median (mo) 6.2 [4.0;7.8] 3.7 [2.2;5.1] 5.7 [4.0;6.5]
rate at 6-m 51.6% [41.1;61.0] 32.3% [23.04;41.81] 45.3% [35.1;55.0]
LBx, n with SV in BRCA1/ % Germline % Tumor somatic % No germline/tumor somatic rate at 12-m 20.3% [12.9;29.0] 12.7% [6.8;20.3] 11.9% [6.3;19.4]
Cancer type N 2/PALB2 (%) SV present SV present SV, only CH SV present OS
median (mo) 11.6 [8.4;15.0] 9.1 [7.10;10.45] 12.4 [9.76;14.03]
Lung 21456 928 (4.3) 25.5 51.3 23.1 rate at 24-m 23.8% [15.4;33.2] 9.5% [4.19;17.36] 12.5% [6.28;20.93]
Cholangiocarcinoma 1787 90 (5) 38.9 38.9 22.2 Grade 3-4 toxicities
Pancreas 5625 303 (5.4) 57.8 25.1 17.2 AE/SAE All Grades Grade 3-4 All Grades Grade 3-4 All Grades Grade 3-4
Esophagus 1199 56 (4.7) 33.9 50.0 14.3 Neutropenia 25.3% 9.5% 11.8% 0% 26.3% 7.4%
Prostate 13858 980 (7.1) 39.3 46.5 14.2 Diarrhea 80% 17.9% 69.9% 16.1% 55.8% 5.3%
Colorectal 6639 391 (5.9) 15.3 73.4 11.0 Vomiting 80% 15.8% 72% 14.0% 61.1% 4.2%
Breast 11397 902 (7.9) 55.2 35.1 9.6 Peripheral Neuropathy 44.2% 5.3% 10.8% 0% 57.9% 8.4%
Ovarian 1464 200 (13.7) 66.0 27.0 7.0
Endometrial 795 84 (10.6) 20.2 73.8 6.0

4185 Poster Session 4186 Poster Session

Machine learning and statistical prediction of overall survival (OS) from pre- Real-world predictors of adverse clinical outcomes in pancreatic cancer
dose plasma biomarkers in a randomized phase 2 trial (1801 Part 3B) of the using a machine-learning framework. First Author: Akanksha Dua, University of
GSK-3 inhibitor elraglusib in metastatic pancreatic ductal adenocarcinoma California, San Francisco, San Francisco, CA
(mPDAC): Application toward patient enrichment. First Author: Taylor Weiskittel, Background: Completion of chemotherapy is crucial for clinical benefit and has been linked to improved
Mayo Clinic, Rochester, MN outcomes, but is often challenging given the toxicities of chemotherapy regimens. The main treatment
regimen for pancreatic adenocarcinoma (PDA), FOLFIRINOX (FFX), is highly toxic, requires frequent dose-
Background: Elraglusib is a first-in-class inhibitor of GSK-3ß, a well-credentialed target in
modifications, and most patients are too sick to tolerate multiple lines of therapy. Given the aggres-
cancer implicated in intrinsic oncologic processes and tumor immune response. Pre- siveness of the disease, selecting the appropriate first-line dosing is crucial as it can be associated with
liminary results of the 1801 Part 3B trial (NCT03678883) showed statistically significant better patient outcomes. The current study objective is to evaluate the association between real-world
benefits for elraglusib+GnP versus GnP for 1-year survival and mOS in mPDAC (companion dosing patterns and clinical toxicity using a machine-learning framework. Methods: Patients attending
abstract: Mahalingam et al.). We investigated plasma levels of cytokines/chemokines/ GI oncology clinics at University of California, San Francisco between November 2011 – December 2023
soluble cell receptors/growth factors (CCSG) as potential biomarkers of favorable out- with a documented administration of FFX were included. Predictors of clinical outcomes included
comes on elraglusib. Methods: Forty CCSGs were evaluated in pre-dose plasma from baseline demographic and clinical features, cycle-specific laboratory data, and dosing information for
patients with previously untreated mPDAC enrolled in 1801 Part 3B treated with GnP (n = FFX sub-components. Group-based 5-fold cross-validation for logistic regression, random forest, and
XGBoost models were used to identify features associated with clinical outcomes (anemia, dehydration,
78) or elraglusib+GnP (n = 155) using a Luminex immunoassay. Using Kaplan-Meier nausea/vomiting, neutropenia, and polyneuropathy). Model performance was evaluated using AUC.
statistics and cutpoint determination, CCSGs were assessed for OS predictive ability in the Results: Data for 505 patients with PDA receiving FFX across 5,041 cycles were [Link] random
elraglusib+GnP arm. Multivariate models were constructed to predict binarized survival at forest models yielded the best fit for the prediction across all outcomes (Table). Key features consistently
12 months using machine learning (ML). Fivefold cross-validation was used in both associated with outcomes included cycle number, cumulative dose of drug received, laboratory data (PT,
analyses, and all methods were applied to the GnP arm to identify elraglusib+GnP specific INR, albumin), patient demographics (male sex, race, and smoking status), and clinical features, such as
predictors. Results: Data shown as of November 15, 2024. Multiple CCSGs significantly hypertension. Conclusions: Our study identifies clinical features in combination chemotherapy leading to
stratified elraglusib+GnP patients. The most extreme prognosticators were IFN-ß (average specific toxicities, highlighting these as ideal strategies for intervention. Early prognostic indicators of
adverse outcomes can guide early management for high-risk individuals through supportive measures
HR = 2.34), and PD-L1 (average HR = 0.52) (HR shown as high CCSG vs low CCSG). and dose modification, before high-grade toxicities and discontinuations wherein patients can no longer
Screening of different ML approaches ranked logistic regression at the top, and hyper- benefit from therapy. Research Sponsor: UCSF Pancreas Center.
parameter grid search identified stochastic gradient solver with ridge regularization as the
Summary of FOLFIRINOX model outputs across clinical outcomes.
optimal method. The model had an accuracy of 88% (SD: 3.9%) and a balanced accuracy of
80% (SD: 8.5%). IFN-beta had the most substantial effect size (odds ratio (OR) = 72.28), Outcome\Model Logistic Regression Random Forest XGBoost Key Features
followed by IL-18 (OR = 23.38) and PD-L1 (OR = 15.32). All other CCSGs had an OR between Anemia 0.61 6 0.03 0.67 6 0.03 0.62 6 0.04 [1] INR
0.03 and 6.71. When this model was applied to patients in the GnP arm, accuracy was [2] Irinotecan dose
[3] Any polyneuropathy
68.4%, and balanced accuracy was 43.1%. Conclusions: Many CCSG biomarkers were Dehydration 0.64 6 0.03 0.72 6 0.02 0.68 6 0.05 [1] Cumulative irinotecan
identified as promising predictors of survival benefit in mPDAC patients treated with [2] Cumulative oxaliplatin
elraglusib+GnP. Both univariate statistical and multivariate ML approaches show pre- [3] Cumulative fluorouracil
dictive significance with high interpretability. The initial ML model is specific to elraglusib Nausea or vomiting 0.62 6 0.04 0.68 6 0.03 0.66 6 0.03 [1] Cumulative irinotecan
[2] Cumulative oxaliplatin
treatment, not GnP alone. These results indicate that patients’ initial immune state plays a [3] Cumulative fluorouracil
role in response to elraglusib+GnP. However, single CCSGs for enrichment currently Neutropenia 0.57 6 0.04 0.61 6 0.06 0.61 6 0.01 [1] Male sex
exclude too many patients ( . 75%), and thus, panels of biomarkers are being investigated [2] Cumulative fluorouracil
[3] Smoking status – never
with ML to overcome this limitation. The 1801 Part 3B clinical trial recruitment has been Polyneuropathy 0.53 6 0.06 0.536 0.06 0.52 6 0.04 [1] Male sex
completed, and updated biomarker models reflective of topline OS data, available by April [2] Oxaliplatin
2025, will be presented. Research Sponsor: Actuate Therapeutics Inc. [3] Constipation

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
316s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4187 Poster Session 4188 Poster Session

A PBMC and machine learning based biomarker signature to predict second Preliminary result of a phase Ib study: Efficacy and safety of FG-M108 plus
line chemotherapy success in advanced PDAC: Translational data of the gemcitabine and nab-paclitaxel in patients with Claudin18.2-positive, lo-
PREDICT trial—A prospective, multicenter, trial of the AIO Pancreatic Cancer cally advanced, unresectable, or metastatic pancreatic cancer. First Author:
Group (AIO-PAK-0216). First Author: Anton Lahusen, Department of Internal Funan Liu, The First Hospital of China Medical University, Shenyang, China
Medicine I, Ulm University Hospital, Ulm, Germany Background: FG-M108, an ADCC-enhanced anti-CLDN18.2 monoclonal antibody,
Background: The PREDICT trial, a recent phase IIIb/IV study, aimed to address the critical showed significant efficacy in first-line (1L) treatment of gastric cancer. Herein, we
need for improved personalized treatment strategies in advanced pancreatic cancer. This report the safety and efficacy results of FG-M108 in 1L treatment of pancreatic cancer
translational investigation examined the predictive value of 1st line chemotherapy (CTX) (cohort C2 and D2). Methods: In this open-label, multicenter phase I/II study patients
response on the efficacy of subsequent 2nd line treatment by using liquid biomarkers received FG-M108 (cohort C2: 300mg/m2 or cohort D2: 600mg/m2 Q3W) plus gem-
combined with machine learning (ML). Methods: Pts. were stratified into two cohorts citabine (1000mg/m2, d1, d8, Q3W) and nab-paclitaxel (125mg/m2, d1, d8, Q3W). El-
based on short or long 2nd line CTX time to treatment failure (S-/L-TTF2; n = 10 per group, igible patients were those with CLDN18.2 positive local advanced or metastatic
80/20% quantiles). Treatment-naı̈ve PDAC tissue specimens underwent laser microdis- pancreatic cancer who were previously untreated. The primary endpoint were the in-
section for tumor cell enrichment, followed by RNA profiling using NanoString™ PanCancer cidence of adverse events (AEs) and preliminary clinical efficacy (ORR, DCR, DOR, PFS,
IO360. Selected differentially regulated genes from the omics results were utilized to and OS). Results: As of November 15, 2024, 50 patients were enrolled and received FG-
screen peripheral blood mononuclear cells (PBMCs) from 2nd line treatment-naı̈ve patients M108+gemcitabine/nab-paclitaxel treatment (39 patients in cohort C2, 11 patients in
at protein (flow cytometry, FC) or RNA (RT-qPCR) levels. FC data was analyzed using R- cohort D2). The median age was 61 (range 30-72). 47 (94%) patients were with CLDN18.2
based single-cell clustering (x-Shift, FlowSOM, T-REX) to generate HyperGates (HGs), with moderate-high expression. Out of 50 patients, 44 patients had at least one tumor
subsequent ML-based back-gating (HyperFinder algorithm) of the most differential assessment after baseline and included in the efficacy analysis set. The median follow-
clusters. Additionally, classical ManualGates (MGs) were generated. Feature selection up time (95%CI) for the 32 patients with CLDN18.2 moderate-high expression was
employed the Weka-based WrapperSubsetEval (WSE) algorithm with eight different 9.5 months (6.8,11.2) , with a maximum treatment duration of 13 months. In the
classifiers (NB, KLR, LR, SMO, IBk, RT, RF, J48). The classifier/subset with optimal subgroup of cohort C2 patients with CLDN18.2 moderate-high expression assessed by
performance was utilized for binary classification (S-TTF2 vs. L-TTF2) of training (80%, n = Independent Review Committee(IRC)—16 patients achieved confirmed PR, and one
66) and validation (20%, n = 16) datasets. Results: Transcriptome analysis of L- vs. S-TTF2 achieved unconfirmed PR. ORR was 53.1% (34.7,70.9), and DCR was 100.0% (89.1-
tumor tissues revealed increased inflammation (upregulated 18-gene signature), immune
100.0). The median DOR reached 9.9 months (7.8,NE), and the median PFS reached
cell activation/infiltration (e.g. CD4/CD8 T cells), and immune exhaustion (upregulation of
9.9 months (7.0,NE) . OS data are not yet mature, with 23 patients still alive, achieving a
e.g. PDCD1, LAG3, CTLA4, TIGIT) in L-TTF tumors. Further, the favorable Bailey immu-
median OS of 17.4 months (11.0,NE). Treatment-emergent adverse events (TEAE)
nogenic subtype was enriched in the L-TTF2 group. Analysis of eight FC panels (19
occurred in 32 patients (100.0%), in which 15 (46.9%) were $ grade 3. The most
candidates) revealed 1198 differential clusters with HGs and 881 classical MGs. Feature
selection, combining FC data with RT-qPCR results and clinical parameters, identified a
common FG-M108 related AEs in cohort C2 & D2 were nausea (56.4% vs 36.4%),
best performing signature of 5 HGs and 2 MGs for 7 protein markers (CXCR4, CD8, CD4, vomiting (48.7% vs 45.5%), and hypoalbuminaemia (46.2% vs 54.5%). Conclusions: The
CD62P, CD307b, CD45, CD121b). ML using a kernel logistic regression successfully combined therapy of FG-M108 plus chemotherapy as 1L treatment for patients with
predicted S- and L-TTF2 binary groups prior to 2nd line CTX with nal-IRI/5-FU/LV (ROC- CLDN18.2 positive pancreatic cancer was generally well tolerated with promising
AUC . 0.90 for training and validation). Conclusions: We identified a favorable tumor survival (PFS and OS) data especially in patients with CLDN18.2 moderate-high ex-
immune microenvironment in L-TTF aPDAC patients, characterized by CD8 T cell-inflamed pression, pivatol phase III clinical study will start in 2025 Q2. Clinical trial information:
("hot") tumor tissues prior to 2nd line CTX. A 7-marker liquid biomarker panel, comprising 7 NCT04894825. Research Sponsor: None.
flow cytometry PBMC population gates, was developed for early prediction of 2nd line nal-
IRI/5-FU/LV CTX success. These findings aim to advance personalized treatment strat-
egies. Clinical trial information: NCT03468335. Research Sponsor: Servier.

4189 Poster Session 4190 Poster Session

Neo-adjuvant chemo-immunotherapy in pancreatic cancer: Results of the NeoOPTIMIZE: Phase II trial of adaptive switching of neoadjuvant FOLFIR-
Australasian Gastrointestinal Trials Group (AGITG) NEO-IMPACT pilot trial. INOX (FFX) to gemcitabine/nab-paclitaxel (GA) resectable/borderline re-
First Author: Lorraine A. Chantrill, University of Wollongong, Wollongong, NSW, Australia sectable (BR)/locally advanced (LA) pancreatic adenocarcinoma (PDAC).
Background: Despite curative intent surgery and peri-operative chemotherapy for re- First Author: Lyndsey Sandow, Internal Medicine, Oregon Health & Science University,
sectable pancreatic ductal adenocarcinoma (PDAC), recurrence rate remains high Portland, OR
(.80%). Whilst immunotherapy has not been shown to be effective in the metastatic Background: Neoadjuvant chemotherapy (NAC) for localized PDAC may improve R0
setting, there may be a scientific rationale for mobilising the immune system before resection. FFX and GA are used but lack of predictive biomarkers remains a barrier to
surgery for localised disease to improve outcomes. Methods: NEO-IMPACT is a single optimal NAC. The angiotensin-II receptor blocker (ARB), losartan, remodels vascular
arm phase II study testing the feasibility and safety of delivering 12 weeks of neo- perfusion to enhance NAC efficacy. We established an experimental platform for dy-
adjuvant immune checkpoint inhibitor in combination with FOLFIRINOX (3 doses of namic switching of NAC +/- radiation therapy for resectable/BR PDAC and an exploratory
1500mg durvalumab q28d + 6 cycles of FOLFIRINOX q14d) for resectable or borderline cohort of LA PDAC. Methods: Patients (pts) received 4 cycles of FFX (ox 85 mg/m2; LV
resectable pancreas cancer. The patients then received 3 months of adjuvant che- 400 mg/m2; iri 150 mg/m2; 5-FU 2400 mg/m2), then were restaged in a multi-
motherapy. The primary endpoint was the proportion of patients receiving $80% of disciplinary tumor board (Restage I). Pts without progression (CT and CA19-9
planned neoadjuvant treatment. A sample size of 20 was calculated to allow 80% power. increase ,30% from baseline) completed 4 more cycles of FFX. If there was pro-
Secondary endpoints include the proportion of patients missing surgery due to gression (CT and/or CA19-9 increase . 30%) or intolerance, pts were switched to GA
treatment related adverse events (TRAEs); treatment tolerability; RO resection rate; (nab-P 125 mg/m2; gem 1000 mg/m2) for 2 months. After a total of 4 months NAC, pts
pathological complete response rate; objective response rate. Results: 20 patients with were re-staged (Restage 2) and had surgery or chemoRT (if vascular involvement) then
PDAC were enrolled between August 2022 and June 2024, 13 resectable and 7 borderline surgery. Losartan (50mg QD) was given throughout NAC/chemoRT. The primary end-
resectable disease. 17 of 20 patients (85%) completed planned neoadjuvant treatment. point was the proportion of resectable/BR with R0 resection. Results: Of 43 patients
Grade 3-4 adverse events (AEs) occurred in 8 patients. The most common was infection screened,16 were resectable, 21 BR, 5 LA and 1 ineligible. Median age was 65 years
(4), febrile neutropenia (1) and nausea (2). 1 patient died from 5FU toxicity due to (range: 34-80), 49% male, 84% Caucasian white. Head of pancreas primary was 84%.
homozygous loss of DPYD. There were 2 iAEs (colitis and maculopapular rash). 1 patient Mean baseline CA19-9 96 ng/mL. 1 pt had progressive disease prior to Restage 1. Of the
had TRAE and had to come off trial but proceeded to surgery. 3 patients (1 borderline/2 36 pts at Restage 1, 31 continued FFX and 5 switched to GA. 18 patients on FFX had a
resectable) became unresectable at assessment for surgery following neoadjuvant radiographic response, 5 had CA 19-9 decrease .25%, and 8 had stable disease with
therapy. Of the 16 patients who underwent surgery (12 head; 4 neck/tail), 2 had a unchanged CA 19-9. 3 pts switched to GA for radiographic progression,1 for increased
complete pathological response. 13 had an R0 resection; 3 had an R1 resection. All 16 CA 19-9 .30% and 1 for FFX intolerance. Of 34 pts evaluated at Restage 2, 2 continued
patients received post operative adjuvant therapy. At a median follow up of 15 months, 5 NAC, 13 had pre-op chemoRT (12 BR and 1 resectable), and 19 proceeded to surgery. 24/
patients who proceeded to surgery and adjuvant therapy have recurred. 27 (88%) pts on FFX had R0 resections; 4/4 (100%) pts switching to GA had R0 re-
Conclusions: Neoadjuvant chemoimmunotherapy is feasible and safe for patients with sections. Overall, 77% pts completing NAC FFX had R0 resections; 80% pts switched to
resectable and borderline resectable pancreas cancer. A 10% CPR rate and 70% R0 GA had R0 resections. Grade . 3 toxicities 7% FFX and 5% GA. Conclusions: Early
resection rate are encouraging, and this approach should be further explored in a larger switching to GA in pts progressing on FFX led to an equivalent R0 resection rate.
population. Clinical trial information: ACTRN12622000378729. Research Sponsor: GI Optimization of NAC made it possible to undergo curative intent surgery that would not
[Link] (Australasian Gastrointestinal Trials Group); Astra Zeneca Pty Ltd. have occurred. Secondary endpoints of DFS, OS and multiomic based studies of blood
and tumor tissue via our Serial Measurements of Molecular and Architectural Responses
to Therapy (SMMART) platform are underway to identify molecular markers to predict
response and determine whether a switch in treatment is indicated. Clinical trial in-
formation: NCT04539808. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 317s

4191 Poster Session 4192 Poster Session

Development and prospective validation of a novel cfDNA-based diagnostic Prognostic value of postoperative circulating tumor DNA and tumor markers
model for the early detection of pancreatic cancer. First Author: Xiuchao Wang, in resected pancreatic adenocarcinoma (PAAD): An interim analysis of a
Tianjin Medical University Cancer Institute and Hospital, Tianjin, China prospective observational study. First Author: Qian Zhan, Department of General
Background: Pancreatic cancer (PC) is one of the most lethal malignancies, with a 5-year Surgery, Pancreatic Disease Center, Ruijin Hospital, Shanghai Jiao Tong University
survival rate below 10%, primarily due to late-stage diagnosis. Existing diagnostic markers, School of Medicine, Shanghai, China
like CA19-9, show inadequate sensitivity and specificity for early detection. To address this Background: Despite the benefits of postoperative adjuvant chemotherapy, recurrence rates
critical gap in early PC detection, this study developed and prospectively validated a of PAAD remain as high as 60% within the first year, underscoring the need for improved
cfDNA-based diagnostic model integrating fragmentomics features, including copy strategies to optimize treatment plans. This prospective observational study aims to
number variations (CNVs), fragment size ratios (FSR), and orientation-aware cfDNA evaluate the potential of postoperative ctDNA-based minimal residual disease (MRD) as an
fragmentation (OCF). Methods: A multicenter study was carried out with a case-control early predictor of disease relapse in resected PAAD using next-generation sequencing.
cohort (n = 467) for model development and a prospective cohort (n = 1,926) for clinical Methods: Pathologically confirmed stage I-III PAAD patients who underwent surgical re-
validation. Plasma cfDNA underwent low-pass whole-genome sequencing to extract section and carried KRAS mutations were enrolled. Eligible criteria included negative
fragmentomics features like CNVs, FSR, and OCF. A stacked ensemble machine-learning surgical margins and no metastasis prior to adjuvant therapy. Patients who did not receive
model was built based on case-control data and validated in the prospective cohort of PC adjuvant therapy were excluded. Tumor tissue samples collected during surgery were
elevated-risk individuals with diabetes or obesity. Sensitivity, specificity, positive pre- analyzed using a 769-gene NGS panel, while plasma samples were obtained at landmark
dictive value (PPV), and negative predictive value (NPV) were evaluated and compared with (4–8 weeks post-operation and pre-therapy), 12-, 24-, 36-, and 48-weeks post-operation, and
those of CA19-9. The follow-up was intended to last for 3 to 5 years, with the current follow- 2 weeks post-adjuvant therapy. Plasma samples were assessed for MRD using MinerVa
up period ranging from 12 to 24 months. Results: In the case-control cohort, the cfDNA- (Genecast Biotechnology). Concurrently, tumor biomarkers such as CA19-9 were measured.
based model achieved AUCs of 0.9799 and 0.9622 in the training and validation sets, The primary endpoint was overall survival (OS), while the secondary endpoint was disease-
respectively, with both sensitivity and specificity exceeding 90%. In the prospective cohort free survival (DFS). The trial is designed for a total follow-up period of three years. Results:
(n = 1,926), for the 8 PC cases identified, the cfDNA model demonstrated a sensitivity of As of November 27, 2024, a total of 133 patients underwent MRD analysis. KRAS mutations
75%, specificity of 98.1%, and PPV of 15.2% for detecting PC, significantly outperforming were distributed as follows: p.G12D in 55%, p.G12V in 27%, p.G12R in 13% and other
CA19-9 (sensitivity: 12.5%, specificity: 94.3%, PPV: 0.9%). Notably, the cfDNA model subtypes in 5%. Nine patients were excluded due to loss of follow-up, and 12 were excluded
detected all 3 Stage 0 cases, 1 of 3 Stage I cases, and both Stage II cases, providing a for less than six months of follow-up, leaving 112 patients for this interim analysis. MRD
median lead time of 227.5 days (range: 45–298 days) compared to imaging modalities. In positivity rates were 16%, 9%, 10%, 15%, 11%, and 11% at landmark, 12, 24, 36, 48 weeks, and
contrast, CA19-9 detected only one Stage II case out of eight confirmed PC cases (12.5%). 2 weeks post-therapy, respectively. Landmark MRD positivity was a significant predictor of
The model demonstrated significant potential in stratifying pancreatic cysts into high-risk disease recurrence (HR = 2.39, 95%CI:1.14-5.01, p = 0.017). Combining landmark MRD with
and low-risk categories. While CA19-9 is ineffective in detecting either high-risk or benign biomarker CA19-9 improved prognostic accuracy (HR = 2.70, 95%CI:1.4-5.5, p = 0.002).
cysts within the prospective cohort, the cfDNA model successfully differentiates between Patients with longitudinal MRD-positive (detected at any time point excluding landmark or
high-risk and low-risk pancreatic cysts (e.g., high-risk IPMN, 1/1 = 100% sensitivity; low- relapse) had significant worse DFS (HR = 3.13, p = 0.001) and worse OS (HR = 0.73, p =
risk SCN, 0/1 = 0% false positive), which further underscores its clinical utility. 0.001) compared to those with negative MRD. Analysis of MRD status at landmark and week
Conclusions: This study is the first to validate a cfDNA-based diagnostic model for PC in a 36 revealed that patients transitioning from MRD-positive to MRD-negative (clearance) had
large elevated-risk population, showing superior performance and significant lead time significantly better DFS compared to those with persistent MRD positivity (p = 0.023).
benefits. The model detects PC earlier with much higher sensitivity and specificity than Notably, no patients deceased from the MRD clearance group. Conclusions: These findings
CA19-9, promising better outcomes with earlier treatment. The findings highlight cfDNA’s underscore the clinical utility of integrating landmark and longitudinal MRD assessments
potential for non-invasive PC screening in clinical settings. Research Sponsor: The Na- with tumor markers for comprehensive risk stratification and prognostication in resected
tional Key Research and Development Program of China; The Tianjin Key Medical Discipline PAAD patients. This approach could potentially guide personalized treatment strategies and
(Specialty) Construction Project. improve patient outcomes. Clinical trial information: NCT05479708. Research Sponsor:
None.

4193 Poster Session 4194 Poster Session

Five-year outcomes of perioperative or only adjuvant gemcitabine plus nab- Perioperative pembrolizumab (pembro) plus chemotherapy (chemo) for
paclitaxel for resectable pancreatic cancer (the NEONAX trial): A randomized locally advanced gastric or gastroesophageal junction (G/GEJ) cancer: Asia
phase II trial of the AIO pancreatic cancer group. First Author: Thomas Jens versus non-Asia subgroup analysis of KEYNOTE-585. First Author: Takashi
Ettrich, Department of Internal Medicine I, Ulm University Hospital, Ulm, Germany Oshima, Kanagawa Cancer Center, Yokohama, Japan
Background: Perioperative chemotherapy (CTX) in resectable pancreatic ductal adeno- Background: In the randomized phase 3 KEYNOTE-585 study (NCT03221426), pCR was
carcinoma (rPDAC) is still not considered standard of care and data are limited. NEONAX significantly improved in participants (pts) with locally advanced G/GEJ cancer treated using
examined gemcitabine (Gem)/nab-paclitaxel (nab-P), in the perioperative or adjuvant perioperative pembro + chemo vs placebo (pbo) + chemo, although the difference in EFS did
therapy of resectable PDAC (NCCN criteria). Methods: NEONAX is a prospective, ran- not meet the prespecified criteria for significance. We present an exploratory subgroup
domized phase II trial with two independent arms (127 patients, 22 German centers) and analysis by geographic region (Asia vs non-Asia). Methods: Eligible pts had untreated, locally
randomization 1:1 to perioperative (2 pre- and 4 postoperative cycles, po-arm A) or advanced, resectable G/GEJ adenocarcinoma (including Siewert type 2 or 3 tumors). Pts
adjuvant (6 cycles, ad-arm B) of gem (1000mg/m2 BSA, d1, 8, 15) and nab-P (125mg/m2 enrolled in the main cohort (n = 804) received neoadjuvant pembro 200 mg IV Q3W or pbo +
BSA, d1, 8, 15), q4w. Results: As we previously reported were R0- and N0-resection rates chemo (cisplatin + capecitabine or 5-FU) for 3 cycles (C); after surgery, pts received adjuvant
high (po-R0 88%, ad-R0 67%, po-N0 33%, ad-N0 29%) in the ITT-population (all randomized pembro or pbo + chemo Q3W for 3C then adjuvant pembro or pbo Q3W for 11C. Pts in the
pts.). The primary endpoint DFS rate of 55% @ 18 months in the mITT population (defined FLOT cohort (n = 203) received neoadjuvant pembro 200 mg IV Q3W or pbo Q3W for 3C + FLOT
as R0/R1 resected pts. that either started perioperative (A) or adjuvant (B) CTX), was not Q2W for 4C; after surgery, pts received adjuvant pembro or pbo Q3W for 3C + FLOT Q2W for
reached in both arms (arm A: 32%, arm B 41%). Whereas 91.5% of pts. in po-arm A started 4C, then adjuvant pembro or pbo Q3W for 11C. Primary end points were pCR (BICR), EFS
and 84.7% completed neoadjuvant CTX, only 42.4% of pts. in ad-arm B started and 25% (RECIST v1.1 by investigator), OS (main), and safety (FLOT). We report outcomes in the main
completed adj. CTX, so the CTX dose intensity was higher in the po-arm A. (Seufferlein and FLOT cohorts combined. The database cutoff date was February 16, 2024 (final analysis).
et al., Ann Onc 2023) Here we report long-time 5-year outcomes according to PFS and OS Results: Of 1007 pts enrolled, 387 were enrolled in Asia; 620, at non-Asia sites; baseline
and present a preplanned subgroup analysis of potential prognostic factors. The mOS in characteristics were generally balanced, with notable exceptions for ECOG PS 1 (11.9% Asia
the ITT population (all randomized pts.) was 25.5 mo. in the po-arm and 16.8 mo. in the ad- vs 37.1% non-Asia), FLOT backbone (1.6% vs 31.8%), tumor location stomach (86.6% vs
68.9%), and tumor stage III-IVa (85.8% vs 74.0%). In the Asia subgroup, pCR was 17.1% with
arm. This corresponds to an mDFS of 11.4 mo. in the po-arm and 5.1 mo. in the ad-arm,
pembro + chemo vs 2.1% with pbo + chemo (difference, 15.0%; 95% CI, 9.7-21.2); median EFS
respectively. The mOS in the mITT population (all randomized pts. that started po-ctx (po-
was 69.8 mo vs 42.7 mo (HR, 0.81; 95% CI, 0.60-1.10), with a 5-year rate of 54.1% vs 45.6%;
arm) or started ad-ctx (ad-arm)) was 27.9 mo. in the po-arm and 26.8 mo. in the ad-arm.
median OS was not reached (NR) vs NR (HR, 0.87; 95% CI, 0.63-1.19), with a 5-year rate of
This corresponds to an mDFS of 14.1 mo. in the po-arm and 16.1 mo. in the ad-arm,
61.3% vs 57.4%. In the non-Asia subgroup, pCR was 12.4% with pembro + chemo vs 3.3% with
respectively. In the preplanned analysis of subgroup factors impacting outcome, the
pbo + chemo (difference, 9.1%; 95% CI, 4.9-13.6); median EFS was 37.7 mo vs 24.3 mo (HR,
benefit of po-treatment was independent of tumor size, N-status and baseline Ca19-9 level. 0.79; 95% CI, 0.64-0.98), with a 5-year rate of 44.0% vs 33.0%; median OS was 60.7 mo vs 42.4
This benefit was not visible in the mITT population where only patients in the ad-arm were (HR, 0.85; 95% CI, 0.68-1.06), with a 5-year rate of 50.5% vs 42.6%. In the Asia subgroup,
included who had started adjuvant treatment postoperatively. Conclusions: These 5 year treatment-related AE (TRAE) rates were 98.4% with pembro + chemo vs 97.4% with pbo +
data confirm the outcome of patients receiving gem/nab in the po-setting in the ITT chemo; grade 3-5 TRAE rates were 69.6% vs 65.1%, respectively. In the non-Asia subgroup,
population. DFS and OS effect were numerically better compared to the ad-arm although TRAE rates were 94.1% with pembro + chemo vs 96.1% with pbo + chemo; grade 3-5 TRAE
the trial was not powered for direct comparison of the arms. This difference disappeared in rates were 65.5% vs 61.7%, respectively. Conclusions: A favorable trend in EFS and OS was
the mITT population when patients received more CTX in the ad arm. We conclude that the seen in both the Asia and the non-Asia subgroups of KENOTE-585; additional antitumor
difference between ITT and mITT is likely because more CTX could administered in the po- activity of pembro + chemo was consistently observed, regardless of region. The safety
arm when the ITT population was considered and may constitute one of the major effects profiles were generally comparable between the subgroups. These findings support the global
of neoadjuvant chemotherapy in resectable PDAC, particularly when a high rate of patients development of this perioperative immunotherapy/chemotherapy regimen. Clinical trial in-
as in our multicenter trial could get neoadjuvant, but not adjuvant treatment. Clinical trial formation: NCT03221426. Research Sponsor: Merck Sharp & Dohme LLC, a subsidiary of
information: NCT02047513. Research Sponsor: Bristol Myers Squibb GmbH & Co. KGaA. Merck & Co., Inc., Rahway, NJ, USA.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
318s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

4196 Poster Session 4197 Poster Session

Association of neighborhood-level factors with access to genetic testing in Constructing genetic-immune prognostic subtypes of primary duodenal
patients with resectable pancreatic ductal adenocarcinoma. First Author: adenocarcinoma through whole exome sequencing and AI-assisted immune
Muhammad Maisam Ali, University of Wisconsin School of Medicine and Public Health, microenvironment analysis. First Author: Ting Han, Department of Oncology, Renji
Madison, WI Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China
Background: Since 2019, the NCCN has recommended genetic testing for all patients with pancreatic Background: Primary duodenal adenocarcinoma (DA) is a rare gastrointestinal tumor
ductal adenocarcinoma (PDAC). However, only one-third of patients undergo testing. For other with a poor prognosis. This study aimed to investigate the tumor immune microen-
cancers, social determinants of health have been shown to influence access to genetic testing. This vironment (TIME) and genetic landscape of DA to depict the unique DA subtypes,
study examines disparities in genetic testing among PDAC patients using neighborhood-level
measures. Methods: We retrospectively analyzed 249 patients who underwent curative-intent re-
distinguishing its prognosis and identifying potential therapeutic targets. Methods: To
section for PDAC at a high-volume academic center between 2019 and 2023. Neighborhood dis- reveal the heterogeneity of DA’s genetic landscape and TIME, 88 treatment-naı̈ve DA
advantage was quantified using the Area Deprivation Index (ADI). Additional socioeconomic variables tumor samples were analyzed via multiplex immunofluorescence (mIF) staining, whole-
included Rural-Urban Classification (RUCA) codes, insurance type (private vs. non-private), and exome sequencing, and RNA sequencing. The quantity of infiltrating cells, tertiary
distance to the treating facility. The primary outcome was completion of genetic testing. For those lymphoid structure, and spatial analysis was conducted using an automated platform
that did not undergo testing, we separately analysed the association between socioeconomic factors (APTIME), an artificial intelligence-based analysis tool for analyzing pathology images
and presence of a referral for genetic counseling or testing. Logistic regression models adjusted for from formalin-fixed paraffin-embedded slides stained with mIF. Results: Significant
age, sex, and period of diagnosis were used to assess associations. Results: The cohort’s mean age
heterogeneity was observed in the genetic and immune landscapes of DA. A genetic-
was 66.6 years (SD: 9.2); 45.1% were female, and 97.6% identified as non-Hispanic White. Median ADI
was 55.0 (IQR: 29.5), and 38.1% lived in rural areas. Genetic testing was completed by 96 patients immune classification was established, identifying four distinct subtypes: MSI
(40.1%), identifying pathogenic variants in 18 (18.7%). Multivariable analysis revealed age above 65, (microsatellite instable), InA(inflamed active), ML (macrophage-low), and MH
higher ADI, and greater distance from the facility significantly reduced testing likelihood. Testing rates (macrophage-high). The InA subtype exhibited high levels of infiltrating immune cells,
increased between 2019 and 2023 (Table 1). Among the 153 patients without testing, only 41.2% were while the MH subtype, characterized by enriched tumor-associated macrophages, was
referred for genetic counseling/testing. Greater distance from the facility was associated with lower associated with the worst overall survival. The MH subtype frequently harbored TGF-b
likelihood of being referred (OR: 0.33 [CI: 0.13, 0.85]). Conclusions: Neighborhood-level factors pathway mutations, particularly in SMAD4, while the ML subtype showed alterations that
influence genetic testing access, even in a predominantly racially homogenous population. Targeted predominantly in the SWI/SNF pathway, specifically in ARID2. Spatial analysis indicated
interventions are needed to reduce disparities and improve testing and referral rates. Research
Sponsor: None.
that closer proximity between macrophages and both tumor cells and T cells correlated
with worse prognosis in DA patients. Closer interactions between PD-L1 and PD1+ T cells
Univariable and multivariable logistic regression for association of socioeconomic factors with in the MH subtype suggested that PD-1/PD-L1 interactions contributed to an immu-
completion of testing.
nosuppressive tumor microenvironment. Conclusions: This study enhances the un-
Univariable Multivariable*
derstanding of DA’s molecular characteristics, particularly through the identification of a
Factor OR [95% CI] OR [95% CI] novel genetic-immune subtype, and provides a foundation for developing precision
ADI (continuous) 0.12 [0.04, 0.46]* 0.15 [0.02, 0.97]*
Urban residence (ref: Rural) 0.61 [0.34, 1.04] 0.97 [0.50, 1.90] treatment strategies for this malignancy. Research Sponsor: None.
Distance to facility (ref: Low, <33)
Moderate, 33–61 0.49 [0.26, 0.93]* 0.75 [0.34, 1.65]
High, ‡ 61 0.31 [0.17, 0.56]* 0.45 [0.21, 0.99]*
Age ‡ 65 (ref:<65) 0.81 [0.50, 1.31] 0.50 [0/26, 0.94]*
Female Sex (ref: Male) 1.05 [0.68, 1.62] 1.08 [0.57, 2.04]
Non-private insurance (ref: Private) 0.77 [0.46, 1.28] 1.13 [0.63, 2.01]
Period of diagnosis (ref: 2019-2020) 7.32 [3.63, 14.78]* 7.37 [3.52, 15.43]*
2023

*Statistically significant. Adjusted for age, sex, insurance status, and period of diagnosis.

4198 Poster Session 4199 Poster Session

Uncovering the tumor microenvironment (TME): Exploring survival and Identification of differential epigenetic landscapes in subtypes of appendi-
immunotherapy (IO) response in cancer of unknown primary (CUP). First ceal cancer. First Author: Luisa Ladel, Yale University, New Haven, CT
Author: Joelle Allam, The University of Texas MD Anderson Cancer Center, Houston, TX Background: Appendiceal cancers (AC) represent a rare and heterogeneous group of
Background: Cancer of Unknown Primary (CUP) is a group of rare and heterogeneous malignancies, often managed with strategies adapted from colorectal cancer (CRC) due
metastatic cancers with unidentifiable site of origin at time of diagnosis, despite extensive to their anatomical proximity. However, marked differences in biological behavior,
investigation. Characterization of CUP remains challenging, and biomarkers are urgently treatment response, and molecular profiles between AC and CRC necessitate inde-
needed to better tailor therapies. Tumor microenvironment (TME) comprises the dynamic and pendent study. While genomic data on AC has been explored, epigenetic landscapes
complex network of extracellular matrix, blood vessels, immune cells, signaling molecules remain uncharted. This study establishes the first comprehensive epigenetic profile of
surrounding a tumor. It plays a critical role in tumorigenesis, progression and response to appendiceal cancer subtypes, aiming to identify novel biomarkers, therapeutic targets,
treatment. This study investigates the correlation between clinical factors, TME, IO and overall and prognostic indicators. Methods: Tissue specimens from 20 patients with histo-
survival outcomes in patients diagnosed with CUP. Methods: We retrospectively evaluated logically confirmed appendiceal neoplasms and 4 age-matched non-neoplastic controls
49 CUP patients who underwent Boston Gene Tumor Portrait (BGTP) testing between January
were analyzed. Malignant subtypes were categorized into three groups: Normal, Low-
2023 and July 2024. BGTP utilizes a combination of Next-Generation Sequencing (NGS) and
Grade Appendiceal Mucinous Neoplasms (LAMN, with and without pseudomyxoma
artificial intelligence to analyze a tumor sample’s genetic profile along with its surrounding
microenvironment, thereby providing a more comprehensive view to support clinical decision-
peritonei), and Advanced Neoplasia (mucinous adenocarcinoma, non-mucinous ade-
making and optimize treatment strategies. Clinico-pathological data, including age, ECOG nocarcinoma, and goblet cell adenocarcinoma). Whole-genome methylome profiling was
performance status at diagnosis, baseline laboratory results, tumor histology and grade, performed using enzymatic methyl-seq for single-base resolution of DNA methylation.
number of metastatic sites (NMS) and IO treatment history were collected through retro- Differentially methylated regions (DMRs) were identified with a q-value cutoff of , 0.05
spective chart review. TME was categorized into 4 subtypes: immune-enriched (IE), immune- and analyzed using the Whole Genome R package with the methylKit analysis pipeline.
enriched/fibrotic (IE/F), fibrotic (F) and immune-depleted (D). Results: Baseline charac- Results: Epigenetic clustering revealed progressive dysregulation from normal tissue to
teristics of 49 CUP patients show median age at diagnosis was 63 years (range 32-80). Most LAMN and further to advanced neoplasia, supporting a continuum of malignancy. Our
patients were male (59%). Grade was reported as poorly differentiated in 69%, with pathology preliminary analysis of CpG islands identified 2,621 DMRs between LAMN and normal
described as carcinoma (47%), adenocarcinoma (33%), squamous cell carcinoma (10%) and tissues and 395 DMRs between Advanced Neoplasia and LAMN. In promoter regions,
malignant neoplasm (8%). TME subtypes distribution was 33% IE, 14% IE/F, 16% F, 37% D. 1,852 DMRs differentiated LAMN from normal, and 283 distinguished Advanced Neo-
Univariate analysis revealed that better outcomes were associated with lower NLR ( , 5) plasia from LAMN. LAMN samples exhibited predominantly hypomethylated regions
(median survival 57 vs 11 months, p = 0.006), lower ECOG (57 vs 16 vs 6 months for ECOG 0, 1 relative to normal tissues (2,299 vs. 322 for CpG islands; 1,579 vs. 273 for promoters).
and 2 respectively, p = 0.002), lower NMS ( , 3) (45 vs 14 months, p = 0.046). Survival data of Conversely, advanced neoplasia demonstrated more hypermethylated regions than
TME and NLR subsets is shown in the table. Conclusions: This is the first characterization of LAMN (243 vs. 152 for CpG islands; 195 vs. 88 for promoters). Most DMRs were localized
TME profile in CUP. A diverse range of TME subtypes were seen in this population. While to intronic, distal intergenic, and promoter regions. Key overlapping DMRs included 8
classical prognostic factors such as NLR, ECOG, NMS were associated with better survival, hypomethylated CpG islands, 41 hypermethylated CpG islands, 3 hypomethylated
there also appears to be a trend toward survival benefit with IO in the IE/IE-F subset. Further promoters, and 14 hypermethylated promoters. These regions implicate pivotal genes in
studies are needed to prospectively explore the role of TME subtypes in determining clinical the progression from LAMN to advanced neoplasia. Conclusions: This study pioneers
outcomes and IO response. Research Sponsor: None. the epigenetic characterization of appendiceal cancers, uncovering unique methylation
Median overall survival (months) in TME and NLR subsets, with and without IO. signatures that differentiate malignant subtypes and normal tissue. Integrating these
Subset/Treatment IE/IE-F F/D P-value NLR High NLR Low P-value findings with genomic data highlights critical targets for the detection and molecular
IO 45.3 15.6 0.14 11.2 22.7 0.04 classification of appendiceal neoplasms. These insights pave the way for improved
No IO 14.6 10.8 10.8 57.4 diagnostic and therapeutic strategies tailored to this rare malignancy. Research
Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 319s

4200 Poster Session TPS4201 Poster Session

Circulating tumor DNA-based genomic profiling and real-world outcomes in A multiregional, randomized, controlled, open-label, phase 3 study of the
cancer of unknown primary (CUP). First Author: Keelia Clemens, Guardant Health, anti-claudin18.2 (CLDN18.2) antibody-drug conjugate (ADC) arcotatug
Redwood City, CA tavatecan (IBI343) in gastric or gastroesophageal junction adenocarcinoma
Background: CUP accounts for ,5% of cancers and carries a dismal prognosis with median overall (G/GEJA): Trial in progress. First Author: Lin Shen, Beijing Cancer Hospital, Beijing,
survival of ~13 months. Studies suggest up to one-third of CUP patients (pts) have a potentially China
targetable alteration (PTA) that may be eligible for molecularly guided therapy (MGT). Liquid biopsy Background: CLDN18.2 has been a validated therapeutic target for G/GEJA. As a next-
(LB) is a non-invasive method to identify PTAs and genomic clues regarding primary site via circulating
tumor DNA (ctDNA). We characterize the ctDNA landscape of PTAs in CUP and evaluate outcomes for
generation ADC, arcotatug tavatecan (IBI343) composed of anti-CLDN18.2 monoclonal
pts receiving MGT. Methods: Real-world data was sourced via GuardantINFORM, a database ag- antibody conjugated to exatecan (topoisomerase I inhibitor) with unique IgG1 Fc si-
gregating insurance claims and de-identified records of pts with clinical LB via Guardant360. PTAs lencing to attenuate antibody-dependent cellular cytotoxicity and complement-
were defined as alterations with FDA approved therapies in non-CUP indications. Pts with CUP and $1 dependent cytotoxicity. Previous phase 1 studies of IBI343 observed manageable
treatment claim after LB were included. Outcomes of pts treated with MGT and pts with the same PTA safety profiles with encouraging efficacy in G/GEJA, pancreatic ductal adenocarcinoma
not treated with MGT were assessed via real-world time to treatment discontinuation (rwTTD), real- and biliary tract cancer (2024 ASCO Annual Meeting [3037], ESMO GI 2024 [396MO],
world time to next treatment (rwTTNT) and real-world overall survival (rwOS) in months. Log-rank test ESMO Asia 2024 [132MO]). Here, we present the trial in progress of a phase 3 study (G-
was used to compare time-to-event outcomes. Results: Of 13,324 CUP pts, 50% were male; the
HOPE-001, NCT06238843) evaluating efficacy and safety of IBI343 monotherapy versus
median age was 69 years. Majority of pts underwent LB at time of diagnosis (92.1%). In pts with $1
genomic alteration identified (91.9%), the most common genomic alterations were TP53 (55%), KRAS treatment of investigator’s choice in previously treated patients (pts) with CLDN18.2-
(25%), PIK3CA (14%), and EGFR (12%). A PTA was identified in 29.4% of pts, the most frequent being positive G/GEJA. Methods: This multiregional, randomized, controlled, open-label,
PIK3CA (9.2%), KRAS G12C (4.3%), BRCA1/2 (4%), ERBB2 (3.9%), EGFR (2.8%), IDH1 (2.5%), BRAF phase 3 study planned to enroll 450 pts. Main inclusion criteria are: 1) locally ad-
V600E (2.4%) and MSI-H (2%). rwTTD was higher for pts with alterations in EGFR, BRAF V600E, and vanced unresectable or metastatic G/GEJA; 2) positive CLDN18.2, defined as immu-
MSI-H receiving MGT; only EGFR reached significance (Table). Similarly, rwTTNT was improved in pts nohistochemical (IHC) membrane staining intensity $ 2+ in $ 75% of tumor cells as
with EGFR, BRAF V600E alterations and MSI-H, but did not reach significance. rwOS was numerically measured by the VENTANA CLDN18 (43-14A) Assay; 3) radiologically evaluable disease,
higher for BRAF V600E and ERBB2 mutated pts receiving MGT (Table). Conclusions: This study measurable and/or non-measurable disease per RECIST v1.1; 4) received and pro-
represents the first large-scale ctDNA genomic profiling of CUP pts with real-world [Link]
detected PTAs in 29.4% of CUP pts, similar to tissue-based testing. Our findings support use of LB to
gressed on 2-4 prior regimens of systemic therapy which must include a fluoropyr-
identify PTAs in CUP pts; however, prospective trials are needed to assess MGT efficacy. Research imidine, platinum, and a taxane or irinotecan. Main exclusion criteria are: 1) positive
Sponsor: None. HER-2, defined as IHC 3+ or IHC 2+/in situ hybridization+; 2) history of treatment with
topoisomerase inhibitor-based ADCs. Pts are randomized in a 1:1 ratio to receive IBI343
Real-world outcomes (in months): MGT vs no MGT.
6mg/kg Q3W in the experimental arm or to receive treatment of investigator’s choice
rwTTD rwTTNT rwOS
including irinotecan, paclitaxel, or trifluridine/tipiracil in the control arm. Stratification
EGFR 5.8 (95CI: 3-10.27) vs 2.8 (95CI: 11.7 (95CI: 6.2-NR) vs 6.5 12.8 (95CI: 8.9-NR) vs 13.9 factors include region (Asian country/region other than Japan vs. European Union and
(n=46) 2.1-4.5), p=0.0043 (95CI: 5.0-NR), p=0.33 (95CI: 7.0-NR), p=0.9
BRAF 5.7 (95CI: 4.2-25.1) vs 4.0 (95CI 10.2 (95CI: 6.9-NR) vs 9.1 35.1 (95CI: 10.2-NR) vs 25.0 United States vs. Japan), primary site of the tumor (stomach vs. gastroesophageal
V600E : 3.4-11.4), p=0.48 (95CI: 5.6-NR), p=0.43 (95CI: 7.0-NR), p=0.18 junction) and history of prior gastrectomy (yes vs. no). The primary endpoints are
(n=36) progression-free survival (PFS) per RECIST v1.1 and overall survival (OS). The secondary
ERBB2 3.9 (95CI: 2.8-5.9) vs 4.2 (95CI: 7.0 (95CI: 4.9-NR) vs 6.1 (95CI: 21.9 (95CI: 11.6-NR) vs 17.0
(n=58) 2.8-6.8), p=0.41 4.2-NR), p=0.96 (95CI: 8.1-NR), p=0.37 endpoints include objective response rate (ORR), disease control rate (DCR), duration of
MSI-H 7.2 (95CI: 4.9-22.5) vs 3.9 15.4 (95CI: 7.0-NR) vs 10.2 NR (95CI: 19.0-NR vs 34.5 response (DoR) and time to response (TTR) per RECIST v1.1, quality of life (QoL), safety,
(n=34) (95CI: 2.3-NR), p=0.14 (95CI: 5.5-NR), p=0.77 (95CI: 29.8-NR), p=0.84 pharmacokinetics (PK) and immunogenicity. The trial is currently enrolling pts in China
and Japan. Clinical trial information: NCT06238843. Research Sponsor: Innovent Bi-
ologics (Suzhou) Co., Ltd.

TPS4202 Poster Session TPS4203 Poster Session

Telisotuzumab adizutecan (ABBV-400; Temab-A) in combination with fluo- Open-label, single-arm, single-center phase 1b/2 clinical study of fruquin-
rouracil, leucovorin, and budigalimab in locally advanced/metastatic gas- tinib combined with trastuzumab and XELOX in the first-line treatment of
tric, gastroesophageal junction, or esophageal adenocarcinoma (a/m GEA). advanced HER2-positive metastatic gastric or gastroesophageal junction
First Author: Kohei Shitara, Department of Gastroenterology and Gastrointestinal On- adenocarcinoma. First Author: Huifang Lv, Department of Gastrointestinal Medical
cology, National Cancer Center Hospital East, Kashiwa, Japan Oncology, The Affiliated Cancer Hospital of Zhengzhou University, Henan Cancer
Background: The MET proto-oncogene and its receptor tyrosine kinase gene product (c-Met Hospital, Zhengzhou, China
protein) are involved in normal cellular functions such as cell proliferation and differen- Background: Trastuzumab plus chemotherapy has significantly prolonged survival in
tiation but can be abnormally activated and upregulated in cancer to promote tumor growth. patients with HER2-positive gastric and gastroesophageal junction (G/GEJ) cancer. The
MET gene amplification and increased c-Met protein expression are associated with poor KEYNOTE-811 study suggested that the efficacy of adding pembrolizumab to trastuzumab
survival outcomes in gastric cancer. The antibody-drug conjugate Temab-A (ABBV-400) is and chemotherapy was superior to trastuzumab plus chemotherapy. However, only pa-
composed of the c-Met–directed antibody telisotuzumab conjugated to a potent topo- tients with a PD-L1 combined positive score (CPS) of 1 or higher could benefit, while those
isomerase 1 inhibitor. A phase 1 study (NCT05029882) investigating Temab-A monotherapy with PD-L1 CPS , 1 did not benefit from this regimen. Fruquintinib is a highly selective oral
demonstrated manageable safety and encouraging efficacy in patients with previously tyrosine kinase inhibitor of vascular endothelial growth factor receptors (VEGFRs) 1, 2, and
treated, advanced GEA, with an objective response rate of 29% (12/42) and clinical benefit 3. The phase 3 study FRUTIGA demonstrated fruquintinib plus paclitaxel was superior to
rate of 71% (30/42) (Strickler et al. Ann Oncol. 2024;35:1439P). This study evaluates Temab- paclitaxel alone as second-line treatment in patients with G/GEJ cancer. As anti-
A in combination with fluorouracil (5-FU), leucovorin/folinic acid (LV), and budigalimab angiogenesis has a synergistic effect with trastuzumab, we designed this study to
(budi; a PD-1–blocking antibody). Methods: This multicenter,phase 2, open-label, ran- evaluate the safety and efficacy of fruquintinib plus trastuzumab, and CAPEOX as first-line
domized study (NCT06628310) will enroll ~180 adult patients with HER2-negative a/m GEA treatment for advanced HER2-positive G/GEJ cancer. Methods: This is a single-center,
who have not received prior systemic therapy in the a/m setting, have not received a prior single-arm, open-label, phase 1b/2 study. The phase 1b study will adopt a 3+3 design with
PD-(L)1 inhibitor, have Eastern Cooperative Oncology Group performance status 0–1, and escalating oral daily dose of 2 to 4 mg (1 mg per level) fruquintinib for days 1-14 in
have measurable disease per RECIST v1.1. Primary objectives of the study are to evaluate combination with trastuzumab (8mg/kg load, followed by 6mg/kg) intravenously once for
safety and tolerability, evaluate efficacy as measured by progression-free survival and day 1, capecitabine 1000mg/m2 orally twice a day for days 1-14, and oxaliplatin 130mg/m2
objective response, and select the recommended phase 3 dose of Temab-A in combination intravenously once for day 1 using a 21-day cycle. The maximum tolerated dose (MTD) and
with 5-FU, LV, and budi. Secondary objectives include assessment of dose-limiting toxicities recommended phase 2 dose (RP2D) of fruquintinib will be determined in the phase 1b
(DLTs) in the dose-escalation stage, evaluation of pharmacokinetics, and further evaluation study with a dose-limiting toxicity (DLT) period of the first cycle. Major DLTs are defined as
of efficacy measures (duration of response, disease control, and overall survival). The study any of the following toxicities occurring in the DLT period determined to be related to study
consists of 2 stages: dose escalation and dose optimization. During BOIN-directed dose treatment: grade $ 4 hematological toxicities, grade $ 3 non-hematological toxicities, and
escalation, ~18 patients receive escalating doses of Temab-A administered once every toxicities that required discontinuation of fruquintinib or trastuzumab $ 21 days. 6 to 12
4 weeks (Q4W) in combination with fixed doses of 5-FU (2400 mg/m2 Q2W), LV (400 mg/m2 systematic treatment-naı̈ve patients with advanced G/GEJ cancer are expected to be
Q2W), and budi (500 mg Q4W). DLTs are assessed during the first 28-day cycle. During dose enrolled in the phase 1b study, depending on observed DLTs and the need for dose
optimization, ~162 patients are randomized [Link] to 1 of 2 selected doses of Temab-A in adjustments. In the phase 2 study, 39 additional patients will be enrolled with RP2D
combination with 5-FU, LV, and budi or a control arm of FOLFOX + budi. Randomization is administered. Upon 6-8 cycles of treatment completed, fruquintinib plus trastuzumab and
stratified by PD-L1 expression and primary tumor location. Treatment is administered until capecitabine will be administered as maintenance treatment. The treatment continues
disease progression, intolerable toxicity, or other discontinuation criteria are met. Either until progressive disease or intolerable toxicity. The primary endpoint of the phase 2 study
archived formalin-fixed paraffin-embedded tissue or a fresh biopsy is required for biomarker is PFS. The secondary endpoints include OS, ORR, DCR, DoR, safety, and molecular
research that will include evaluation of c-Met protein expression and MET genomic al- biomarker exploration. Clinical trial information: ChiCTR2300074767. Research Sponsor:
terations. Clinical trial information: NCT06628310. Research Sponsor: AbbVie, Inc.; n/a. None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
320s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

TPS4204 Poster Session TPS4205 Poster Session

ARTEMIDE-Gastric01: A phase 3 randomized study of rilvegostomig with A randomized controlled trial comparing conversion surgery with palliative
fluoropyrimidine and trastuzumab deruxtecan (T-DXd) as first-line (1L) chemotherapy in patients with initially unresectable cStage IVB/pStage IV
treatment for locally advanced or metastatic HER2-positive gastric or gas- advanced gastric cancer who presented remarkable response to chemo-
troesophageal junction cancer (GC/GEJC). First Author: Rui-Hua Xu, Department therapy: JCOG2301 (Conversion study). First Author: Itaru Yasufuku, Department
of Medical Oncology, Sun Yat-Sen University Cancer Center, State Key Laboratory of of Clinical Anatomy Development Studies, Gifu University Graduate School of Medicine,
Oncology in South China, Guangdong Provincial Clinical Research Center for Cancer, Sun Gifu-City, Japan
Yat-Sen University, Guangzhou, China Background: Conversion surgery is a surgical treatment for patients with initially
Background: Patients with GC/GEJC often present with advanced disease, and prognosis for unresectable cStage IVB gastric cancer who presented remarkable response to palliative
these patients is poor, with a 5-year relative survival rate of ~5%, highlighting a need for new chemotherapy aiming at n R0 resection expecting long survival including he disease.
treatment options. HER2 overexpression/amplification occurs in ~20% of cases. Adding This study is a randomized controlled phase III trial aimed to evaluate the efficacy of
immune checkpoint inhibition to trastuzumab (anti-HER2 monoclonal antibody) and che- conversion surgery comparing to palliative chemotherapy. Methods: Eligibility criteria
motherapy has shown clinical benefit in patients with advanced HER2-positive GC/GEJC include the followings : (1) Histologically proven adenocarcinoma of the stomach. (2)
(Janjigian YY, et al. N Engl J Med 2024), and led to the approval of pembrolizumab (pro- Diagnosed as clinical stage IVB or pathological stage IV according to the Japanese
grammed cell death-1 [PD-1] inhibitor), trastuzumab, and chemotherapy for HER2-positive Classification of Gastric Carcinoma (15th edition), with at least one of the following
GC/GEJC with programmed cell death ligand-1 combined positive score (PD-L1 CPS) $1. T- unresectable distant metastases before chemotherapy. (i) Four or more liver metas-
DXd (a HER2-directed antibody-drug conjugate) is approved for the treatment of patients with
tases. (ii) Distant lymph node metastasis beyond para-aortic lymph node No.16a2/16b1.
locally advanced/metastatic HER2-positive GC/GEJC who have received a prior trastuzumab-
(iii) Peritoneal dissemination diagnosed with imaging examination or P1b/P1c peritoneal
based regimen. In addition, dual inhibition of PD-1 or PD-L1 and the immune checkpoint T cell
immunoglobulin and immunoreceptor tyrosine-based inhibitory motif domain (TIGIT) has
dissemination diagnosed with laparotomy or laparoscopy. (3) Undergoing first-line
shown encouraging results across multiple tumor types, without major increases in high- chemotherapy regardless of nivolumab or trastuzumab use. (4) Confirmation of no
grade toxicity compared with PD-1 or PD-L1 inhibition alone. Rilvegostomig is a monovalent, peritoneal metastasis or localization at a limited area close to the stomach by lapa-
bispecific, humanized IgG1 monoclonal antibody targeting both PD-1 and TIGIT receptors that roscopy or laparotomy after initiation of chemotherapy, with CY0 status in peritoneal
has shown encouraging efficacy with manageable safety as monotherapy in non-small-cell lavage cytology. (5) Response to chemotherapy of distant metastasis diagnosed before
lung cancer (Hiltermann TJN, et al. WCLC 2024. Oral presentation 1751) and with chemo- initiating first-line chemotherapy that meets the following (i) and (ii) before registration.
therapy in HER2-negative GC/GEJC (Herrero FR, et al. Ann Oncol 2024. Abs 1422P). (i) Liver metastasis: three or fewer liver metastasis. (ii) Distant lymph node metastasis
Methods: ARTEMIDE-Gastric01 (NCT06764875) is a phase 3, randomized, open-label, excluding No.16a2/16b1: disappearance or reduction to a long axis of less than 6 mm.
sponsor-blinded, multicenter, global study that will assess the efficacy and safety of ril- The primary endpoint is overall survival. After confirming eligibility, patients are reg-
vegostomig with T-DXd and chemotherapy as 1L treatment in HER2-positive GC/GEJC with istered and randomized (1:1) to either the palliative chemotherapy alone arm or the
PD-L1 CPS $1. Approximately 840 participants (pts) will be randomized to Arm A: rilve- conversion surgery arm. We assumed the median survival timeis 19 months after
gostomig + T-DXd + investigator’s (INV) choice of capecitabine or 5-fluorouracil (5-FU); Arm registration for the chemotherapy alone arm additional efficacy for overall survival in the
B: pembrolizumab + trastuzumab + INV choice of 5-FU and cisplatin (FP) or capecitabine and conversion surgery arm corresponding to a hazard ratio of 0.7. This study requires 126
oxaliplatin (CAPOX); Arm C: rilvegostomig + trastuzumab + INV choice of FP or CAPOX. patients to observe 102 deaths, with power of 70% and a one-sided alpha of 10%,
Eligible pts will have previously untreated, unresectable, histologically confirmed, locally considering the rarity of patients with stage IV gastric cancer who exhibit a significant
advanced/metastatic HER2-positive and PD-L1 CPS $1 GC/GEJC and an ECOG performance response to palliative chemotherapy, an accrual period of 5 years, and a follow-up period
status of 0 or 1. Dual-primary endpoints are progression-free survival (RECIST v1.1; blinded of 3 years. This trial was initiated in September 2024, and the first patient was enrolled in
independent central review) and overall survival in all randomized pts. Secondary endpoints January 2025. Clinical trial information: jRCTs031240340. Research Sponsor: Japan
include safety/tolerability, objective response rate, and duration of response. Enrollment is Agency for Medical Research and Development (AMED).
planned across ~25 countries in Asia, Australia, Europe, and North and South America.
Clinical trial information: NCT06764875. Research Sponsor: AstraZeneca.

TPS4206 Poster Session TPS4207 Poster Session

A randomized, double-blinded, international phase III trial comparing HLX22 An open-label, randomized, multicenter, phase 3 study of trastuzumab
in combination with trastuzumab and chemotherapy versus trastuzumab deruxtecan (T-DXd) + chemotherapy (chemo) 6 pembrolizumab (pembro)
and chemotherapy with or without pembrolizumab for first-line treatment versus chemo + trastuzumab 6 pembro in first-line metastatic HER2+
for HER2-positive locally advanced or metastatic G/GEJ cancer. First Author: gastric or gastroesophageal junction (GEJ) cancer: DESTINY-Gastric05.
Jaffer A. Ajani, Department of Gastrointestinal (GI) Medical Oncology, MD Anderson First Author: Kohei Shitara, National Cancer Center Hospital East, Chiba, Japan
Cancer Center, Houston, TX Background: An unmet medical need remains in patients (pts) with HER2+ gastric or
Background: Combination therapy with trastuzumab and chemotherapy is the first-line GEJ cancer. HER2 is a validated target in up to 20% of pts with gastric or GEJ cancer. The
systemic treatment for human epidermal growth factor receptor 2 (HER2) positive ad- KEYNOTE-811 trial demonstrated that adding pembro to trastuzumab and chemo
vanced gastric or gastroesophageal junction (G/GEJ) cancer. Among patients whose improved progression-free survival (PFS) and overall survival (OS) versus placebo for
tumors express programmed death ligand 1 (PD-L1; defined as combined positive score first-line treatment of pts with HER2+ gastric or GEJ cancer with a PD-L1 combined
[CPS] $ 1), treatment options also include pembrolizumab plus trastuzumab and che- positive score (CPS) $1 (Janjigian Y et al. N Engl J Med. 391;1360:2024). In the
motherapy. Survival outcomes remain unsatisfactory despite these advances. HLX22 is an DESTINY-Gastric03 trial, first-line combinations involving T-DXd, a HER2-directed
anti-HER2 antibody that targets a different epitope than trastuzumab. HLX22 has shown antibody-drug conjugate, and fluoropyrimidine (5-FU or capecitabine [CAPE]) 6 pem-
improved progression-free survival (PFS) when added to trastuzumab plus oxaliplatin and bro showed acceptable safety and encouraging efficacy in pts with HER2+ gastric or GEJ
capecitabine (XELOX) in a phase 2 study (NCT04908813). Here we present the design of a cancer, including pts with CPS , 1 (Janjigian Y et al. Ann Oncol. 35;S878:2024). Building
phase III randomized controlled study. Methods: This randomized, double-blind, two-arm on this evidence, the phase 3 DESTINY-Gastric05 trial aims to bring a potentially
phase III clinical study aims to compare the efficacy and safety of HLX22 in combination improved platinum-free treatment approach for all pts with HER2+ gastric or GEJ
with trastuzumab and XELOX versus (vs) trastuzumab and XELOX with or without (6) cancer. Methods: DESTINY-Gastric05 (NCT06731478) is an open-label, randomized,
pembrolizumab in patients with HER2-positive, advanced G/GEJ cancer and no prior multicenter, phase 3 global trial designed to evaluate the efficacy and safety of T-DXd in
antitumor therapy in the advanced setting. Key inclusion criteria include histologically or combination with 5-FU (or CAPE) + pembro versus standard-of-care chemo with
cytologically confirmed diagnosis of previously untreated, locally advanced unresectable trastuzumab + pembro as first-line treatment in pts with unresectable, locally advanced
or metastatic HER2-positive G/GEJ adenocarcinoma. Key exclusion criteria include prior or metastatic centrally confirmed HER2+ (immunohistochemistry [IHC] 3+ or IHC 2+/in
use of any HER2-target therapy. Approximately 550 eligible patients will be enrolled from situ hybridization+) gastric or GEJ cancer with a CPS $1. Pts must have $1 RECIST v1.1
multiple regions across the globe and randomly assigned in a 1:1 ratio to receive HLX22 (15 measurable lesion, a left ventricular ejection fraction $50%, and an Eastern Cooperative
mg/kg) + trastuzumab + XELOX 6 placebo (for pembrolizumab) or placebo (for HLX22) +
Oncology Group performance status of 0 or 1. Approximately 576 pts will be randomly
trastuzumab + XELOX 6 pembrolizumab. HLX22 will be administered intravenously on Day
assigned in a 1:1 ratio to receive: T-DXd 5.4 mg/kg + either 5-FU or CAPE + pembro (arm
1 of each 21-day treatment cycle until loss of clinical benefit, death, intolerable toxicity,
M1); or trastuzumab + platinum-based chemo (either cisplatin + 5-FU or oxaliplatin +
withdrawal of informed consent, or other reasons. The stratification factors include HER2
CAPE) + pembro (arm M2). The primary efficacy endpoint is PFS based on blinded
immunohistochemistry (3+ vs 2+), geographic region (Asia vs Europe/North America vs
rest of the world), primary tumor site (gastric vs gastroesophageal junction), and tumor PD-
independent central review (BICR), and the key secondary endpoint is OS. Other sec-
L1 expression (CPS , 1 or not evaluable vs 1 # CPS , 10 vs CPS $ 10). The dual primary ondary endpoints include overall response rate, duration of response, and time to
endpoints are PFS assessed by independent radiology review committee per RECIST v1.1 response per RECIST v1.1 assessed by BICR and investigator. Safety and tolerability will
and overall survival. Secondary endpoints include investigator-assessed PFS, objective also be assessed. An exploratory cohort (approximately 150 pts) will evaluate the
response rate, PFS on the subsequent line of therapy, duration of response, safety, efficacy and safety of T-DXd in combination with 5-FU or CAPE versus trastuzumab plus
pharmacokinetics, immunogenicity, and quality of life. This study is currently open for standard-of-care chemo in pts with PD-L1 CPS , 1. Clinical trial information:
enrollment and has completed first dose of the first patient. Clinical trial information: NCT06731478. Research Sponsor: Daiichi Sankyo, Inc.
NCT06532006. Research Sponsor: Shanghai Henlius Biotech, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 321s

TPS4208 Poster Session TPS4209 Poster Session

A phase II study of sacituzumab govitecan for advanced esophageal squa- KEYMAKER-U06 substudy 06E: A phase 1/2 open-label, umbrella platform
mous cell carcinoma patients (SG-ESCC). First Author: Jhe-Cyuan Guo, National study of ifinatamab deruxtecan in combination with pembrolizumab with or
Taiwan University Cancer Center, Taipei, Taiwan without chemotherapy for first-line treatment of advanced esophageal
Background: Esophageal squamous cell carcinoma (ESCC) remains a significant global squamous cell carcinoma (ESCC). First Author: Ken Kato, National Cancer Cen-
health challenge, particularly in Asia. There are limited treatment options for advanced ter Hospital, Tokyo, Japan
ESCC patients who fail platinum-based chemotherapy and anti-PD-1/PD-L1 therapy, Background: There is a substantial need for more effective and tolerable first-line treatment
resulting in poor prognoses. Trophoblast cell surface antigen 2 (Trop-2), a trans- options for patients with advanced ESCC. B7-H3 is a type 1 transmembrane protein that is
membrane protein overexpressed in ESCC, offers a potential therapeutic target due to its highly expressed in several cancers, including ESCC, and is associated with a poor prognosis.
differential expression between tumors and normal tissues. Sacituzumab govitecan, an Ifinatamab deruxtecan (I-DXd; formerly DS-7300a/MK-2400) is a B7-H3–directed antibody-
antibody-drug conjugate (ADC) comprising an anti-Trop-2 antibody linked to a topo- drug conjugate comprising a humanized anti–B7-H3 IgG1 monoclonal antibody (ifinatamab)
isomerase I inhibitor payload, has shown efficacy in triple-negative and hormone covalently linked to a potent topoisomerase I inhibitor payload (DXd; an exatecan derivative)
receptor-positive breast cancers. This study aims to investigate the efficacy and safety by a cleavable linker. In the phase 1/2 DS7300-A-J101 study, I-DXd monotherapy showed
of sacituzumab govitecan in patients with advanced ESCC. Methods: This investigator- promising antitumor activity in participants (pts) with advanced ESCC. KEYMAKER-U06 is an
open-label, phase 1/2, umbrella platform study designed to evaluate investigational agents
initiated, prospective, phase II, single-arm, multi-center trial evaluates the efficacy and
with or without pembrolizumab and/or chemotherapy for advanced gastroesophageal cancer.
safety of sacituzumab govitecan (10 mg/kg IV on days 1 and 8 of a 21-day cycle) in
Substudy 06E (NCT06780111) will be conducted to evaluate I-DXd plus pembrolizumab with
advanced ESCC patients. Eligible patients must have failed prior platinum-based or without chemotherapy as first-line therapy for advanced ESCC. Methods: Eligible pts are
chemotherapy and anti-PD-1/PD-L1 therapy, exhibit measurable disease per RECIST aged $18 years with previously untreated, histologically or cytologically confirmed, locally
1.1, and have an ECOG performance status #1. The primary endpoint is the objective advanced unresectable or metastatic ESCC, measurable disease per RECIST v1.1 by in-
response rate (ORR) by RECIST 1.1. Secondary endpoints include overall survival, vestigator review and verified by blinded independent central review (BICR), and an Eastern
progression-free survival, duration of response, and safety outcomes. Biomarker an- Cooperative Oncology Group performance status of 0 or 1. Pts will be assigned to 1 of 4
alyses will explore Trop-2 expression and other molecular markers associated with treatment arms: arm 1 (reference treatment; pembrolizumab 200 mg IV Q3W for #35 cycles
treatment efficacy and resistance as well as toxicity. A total of 35 patients will be plus chemotherapy [mFOLFOX6: oxaliplatin 85 mg/m2 IV Q2W plus 5-FU 400 mg/m2 (bolus)
enrolled employing Simon’s two-stage design, with a type I error rate of 0.1 and 80% and 2400 mg/m2 (continuous) IV Q2W plus leucovorin 400 mg/m² IV Q2W]); arm 2 (I-DXd 12
power to detect an ORR $25%, considered promising compared to the historical control mg/kg IV Q3W plus pembrolizumab); arm 3 (I-DXd 12 mg/kg plus pembrolizumab plus 5-FU
of #10%. In the first stage, 16 patients will be accrued, with $2 responses required to 400 mg/m2 [bolus] and 2400 mg/m2 [continuous] IV Q2W plus leucovorin 400 mg/m² IV Q2W);
proceed to the second stage of 15 additional patients. Accounting for an anticipated 10% and arm 4 (I-DXd [8 mg/kg or 12 mg/kg] IV plus pembrolizumab plus 5-FU 2400 mg/m2 IV and
dropout rate, the study aims to complete enrollment within 24 months. Enrollment began oxaliplatin 60 mg/m2). Approximately 209 pts will be enrolled. A safety lead-in phase
in August 2024, and as of December 2024, 5 of the planned 35 patients have been with #29 pts will be conducted in arms 2 (n #6), 3 (n #10), and 4 (n #13) using a Bayesian
enrolled. Clinical trial information: NCT06329869. Research Sponsor: Gilead Sciences. optimal interval design to confirm the safety and recommended phase 2 dose (RP2D; arm 4
only) of I-DXd in combination with other agents; this phase will be conducted sequentially,
starting with arm 2, followed by arms 3 and 4. Thereafter, #180 pts will be included in the
randomized phase (#60 in arm 1; #40 each in arms 2-4). Pts will be randomly assigned 1:2 to
arm 1 and the investigational arms. Primary outcomes are safety and tolerability, RP2D of I-
DXd, and objective response rate per RECIST v1.1 by BICR for the selected dose. Secondary
outcomes include DOR and PFS per RECIST v1.1 by BICR, OS, and pharmacokinetics of I-DXd
in combination with other agents. Enrollment is ongoing. Clinical trial information:
NCT06780111. Research Sponsor: Daiichi Sankyo Company, Limited and Merck Sharp &
Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA.

TPS4210 Poster Session TPS4212 Poster Session

A phase II study of perioperative intraperitoneal paclitaxel in patients with Repurposing itraconazole for secondary prevention of metaplasia and pri-
gastric adenocarcinoma and carcinomatosis or positive cytology. First Au- mary prevention of cancer in patients with high-risk Barrett’s esophagus in
thor: Brian D. Badgwell, The University of Texas MD Anderson Cancer Center, Houston, combination with ablation. First Author: Ajay Bansal, Department of Gastroenter-
TX ology and Hepatology, University of Kansas Medical Center, Kansas City, KS
Background: Over the last 2 decades, intraperitoneal chemotherapy has been found to Background: To prevent invasive esophageal adenocarcinoma (EAC), endoscopic eradi-
have activity for select subgroups of patients with carcinomatosis from colon, ovarian, cation therapy (EET) is used to remove its precursor, Barrett’s esophagus (BE) with
appendiceal, and recently, gastric origins. However, there is little data to support an dysplasia. EET combines endoscopic removal of visible lesions with radiofrequency ab-
aggressive surgical approach of cytoreduction (debulking) and intraperitoneal therapy lation (RFA) of surrounding BE to achieve complete remission of intestinal metaplasia
for patients with gastric cancer and positive cytology or carcinomatosis. Recently, the (CRIM) and complete eradication of dysplasia (CED) to halt progression to cancer. Un-
DRAGON-01 randomized trial reported improvement in outcomes for the addition of fortunately, EET has metaplasia recurrence rates of 12.4%/year; thus, adjunctive cancer
intraperitoneal paclitaxel as part of a bidirectional approach with systemic paclitaxel and interception agents are needed to maintain the gains of EET. Data from pre-clinical studies
S-1 for patients with gastric cancer and peritoneal metastases. However, there are few and patient tissues demonstrate that the Hedgehog (Hh) pathway regulates esophageal
studies supporting intraperitoneal paclitaxel in Western populations. As systemic stem cell activity and cell fate determination (squamous versus intestinal). Post-RFA
therapy is improving with concomitant targeted and immunotherapy, intraperitoneal reactivation of Hh signaling is hypothesized to drive BE recurrence; however, current FDA-
therapy may be best utilized in Western populations after standard of care systemic approved Hh inhibitors are expensive and toxic. The antifungal itraconazole inhibits Hh
therapy. Therefore, the purpose of this clinical trial is to determine the efficacy and signaling and has demonstrated antitumor activity in multiple cancers. In addition, it
safety of perioperative intraperitoneal paclitaxel in patients with stage IV gastric cancer inhibits VEGFR and PI3K-AKT pathways, which are critical to BE development and
limited to the peritoneum after treatment with systemic chemotherapy. neoplastic progression. Given its safety and affordability, itraconazole represents a
Methods: Patients with gastric and gastroesophageal adenocarcinoma and positive promising strategy to reduce BE recurrence and EAC risk. Methods: This randomized,
peritoneal cytology or carcinomatosis that have completed treatment with systemic phase 2b, double-blind, placebo-controlled trial will evaluate itraconazole’s efficacy in
chemotherapy are offered participation in the study. Patients with metastatic disease accelerating BE eradication. Participants with high-risk BE, defined as BE $2 cm with low/
not limited to the peritoneum are excluded. Type and duration of systemic chemotherapy high-grade dysplasia or intramucosal/T1 adenocarcinoma, undergoing ablation will be
enrolled. Participants will be randomized 1:1 to receive 300 mg of oral itraconazole or
is left to the discretion of the treating medical oncologist. Immunotherapy or Her2-
placebo for two weeks before and four weeks after their first 2 sessions of EET. The primary
directed therapy may continue during the trial. We have recently completed a Phase I
endpoint is time to CRIM, a surrogate for long-term BE recurrence, measured in days.
clinical trial demonstrating doses of up to 100 mg/m2 were safe (NCT04220827; PMID:
Secondary endpoints include time to CED, 12-month BE recurrence rates, safety, toler-
39287936). Therefore, 100 mg/m2 is administered intraperitoneal every 2 weeks for
ability, and correlations between itraconazole levels and patient-reported outcomes. We
three treatments before and after gastrectomy. We also modified the trial to allow for the will enroll 74 patients (37 per arm). We shall use a two-sided log rank test for right-
inclusion of heated intraperitoneal chemotherapy during gastrectomy. The primary censored time to event analysis to assess differences. The sample size calculation is based
outcome is overall survival from the date of diagnosis of stage IV disease, with sec- on anticipated surviving proportions at specific study times. We assume that the surviving
ondary outcomes of safety. After completion of study–related treatment, subjects will non-CRIM proportions of patients at 3, 6, 9, and 12 months in the control arm to be 0.8, 0.5,
be followed until recurrence and/or death for up to three years. Sixteen of planned 30 0.2, 0.1, respectively, and in the treatment arm to be 0.5, 0.2, 0.1, 0.05, respectively. To
patients have been enrolled (NCT05977998). Clinical trial information: NCT05977998. detect this effect size with 80% power at 5% level of significance, we need 30 evaluable
Research Sponsor: None. participants in each group (with a plan to enroll 37 per arm to account for attrition). A Cox
model will adjust for demographic and clinical variables. If successful, this trial could
establish itraconazole as a novel adjunct to EET, reducing BE recurrence and lowering EAC
risk. Clinical trial information: NCT06732388. Research Sponsor: National Cancer Institute;
UG1-CA242632.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
322s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

TPS4213 Poster Session TPS4214 Poster Session

Total neoadjuvant therapy with induction immunochemotherapy and che- PROSPERO: A phase 3 randomized, placebo (Pbo)-controlled study of
moradiotherapy followed by surgery for locally advanced esophageal squa- amezalpat (TPST-1120), a peroxisome proliferator-activated receptor a
mous cell carcinoma (TNT-ESCC). First Author: Chien-Huai Chuang, National (PPARa) inhibitor, in combination with atezolizumab + bevacizumab (AB)
Taiwan University Cancer Center, Taipei, Taiwan for patients (pts) with unresectable or metastatic hepatocellular carcinoma
Background: Locally advanced esophageal squamous cell carcinoma (ESCC) is indi- (mHCC) not previously treated with systemic therapy. First Author: Mark
cated for multi-modalities treatment strategies, including a neoadjuvant chemo- Yarchoan, Johns Hopkins Medicine, Baltimore, MD
radiotherapy (CRT) followed by surgery. While the CROSS trial established neoadjuvant Background: PPARa is a fatty acid ligand-activated transcription factor that regulates
CRT as a standard of care, distant metastasis remains a significant cause of treatment genes involved in fatty acid oxidation (FAO), angiogenesis, and inflammation. In HCC and
failure. Immune checkpoint inhibitors (ICIs) have demonstrated survival benefits in other tumor types, PPARa signaling promotes tumor growth and also modulates the tumor
advanced or metastatic ESCC, and adjuvant nivolumab has shown efficacy following immune microenvironment to suppress antitumor immunity. Amezalpat (TPST-1120) is an
neoadjuvant CRT in locally advanced ESCC. Integrating ICI earlier in the treatment investigational PPARa antagonist that inhibits FAO, targeting the bioenergetic requirements
sequence through total neoadjuvant therapy may enhance the immune response against of cancer cells and restoring anticancer immune pathways. HCC has the highest PPARa
the primary tumor and the hidden metastases and potentially lead to improved survival expression of any major tumor type. In preclinical studies of HCC, including ß-catenin
outcomes. This phase II study evaluates induction immunochemotherapy followed by activated disease, amezalpat exhibits anti-cancer activity as a single agent and demon-
CRT before surgery in locally advanced ESCC. Methods: This is a single-center, single- strates complementary efficacy in combination with PD-L1 and VEGF inhibitors. In an
arm, phase II study enrolling 50 patients with histologically confirmed ESCC (T3/4aN0M0 ongoing global randomized Phase 1b/2 study in pts with unresectable or mHCC not pre-
or T1-4aN1-3M0 according to the AJCC Cancer Staging System 8th ed). Eligible patients viously treated with systemic therapy, amezalpat in combination with atezolizumab +
must have primary intrathoracic esophageal tumor #10 cm in length and #5 cm in bevacizumab (TPST-AB) was tolerable and was associated with a clinically meaningful
improvement in multiple efficacy endpoints, including overall survival (OS) and confirmed
radial diameter, an ECOG performance status of 0-1, and adequate organ function.
objective response rate (ORR), compared to AB alone. Here we describe a follow-up pivotal
Patients will receive induction immunochemotherapy consisting of tislelizumab (200 mg
Phase 3 study to evaluate the safety and efficacy of TPST-AB vs Pbo plus AB (Pbo-AB) in pts
every 3 weeks), paclitaxel (175 mg/m² every 3 weeks), and cisplatin (75 mg/m² every with unresectable or mHCC (NCT06680258). Methods: This Phase 3, global, randomized,
3 weeks) for two cycles. This is followed by CRT consisting of radiotherapy (45 Gy in 25 double-blind study will enroll ~740 pts with unresectable or mHCC. Key eligibility criteria
fractions, 1.8 Gy/day, 5 days/week) plus chemotherapy with weekly paclitaxel (50 mg/ include no prior systemic therapy (prior locoregional therapy allowed), ECOG PS 0-1, Child-
m²) and cisplatin (30 mg/m²) for 5 weeks. Esophagectomy will be performed 6 to Pugh Class A, and measurable disease by RECIST v1.1; pts with fibrolamellar/sarcomatoid
8 weeks after completing CRT. The primary endpoint is pathologic complete response HCC, mixed cholangiocarcinoma/HCC, and untreated active HBV are not eligible. Pts will be
(pCR) rate, defined as no residual tumor in the resected primary site and lymph nodes. randomized 1:1 to receive oral amezalpat 600 mg or Pbo twice daily along with the approved
We hypothesize that the pCR rate will increase from 35% (the historical control) to 55%. doses of atezolizumab and bevacizumab every 3 weeks, until unacceptable toxicity or loss of
Based on a binomial precision design, the study is of 80% power and a unilateral a error clinical benefit. Randomization will be stratified by geographic region (Asia excluding Japan
of 0.05 to detect a statistically significant difference in pCR rate. Secondary endpoints vs rest of world), macrovascular invasion and/or extrahepatic spread (y/n), baseline
include major pathological response rate, R0 resection rate, disease-free survival, event- a-fetoprotein ( , 400 vs $400 ng/mL), and baseline ECOG PS (0 vs 1). The primary efficacy
free survival, distant metastasis-free survival, overall survival and safety. The study endpoint is OS. Key secondary efficacy endpoints include progression-free survival and ORR
starts patient enrollment in March 2025 (registered at [Link] as (RECIST v1.1). Exploratory analyses will include outcome by PD-L1 expression and ß-catenin
NCT06764355). Clinical trial information: NCT06764355. Research Sponsor: National mutational status. Interim analyses for futility (30% OS events) and efficacy (70% OS events)
Taiwan University Hospital; BeiGene. are planned. Findings of this pivotal study will inform the efficacy and safety profile of
amezalpat added to AB vs AB alone in pts with unresectable or mHCC not previously treated
with systemic therapy. Clinical trial information: NCT06680258. Research Sponsor: Tempest
Therapeutics, Inc.

TPS4215 Poster Session TPS4216 Poster Session

RHEA-1: First-in-human (FIH) study of AZD9793, a first-in-class CD8- Phase II trial of zanzalintinib (XL-092) in combination with durvalumab and
guided T cell-engager (TCE) for glypican-3-positive (GPC3+) advanced or tremelimumab in unresectable hepatocellular carcinoma (ZENOBIA). First
metastatic hepatocellular carcinoma (HCC). First Author: Stephane Champiat, Author: Meghana Singh, Department of Hematology and Oncology, University of
Department of Investigational Cancer Therapeutics, Division of Cancer Medicine, The Pittsburgh Medical Center, Pittsburgh, PA
University of Texas MD Anderson Cancer Center, Houston, TX Background: Despite the evolving novel treatment options, the prognosis for advanced
Background: GPC3 is an oncofetal protein expressed in 70–80% of HCC and is associated hepatocellular carcinoma (HCC) remains poor, with a 4-year survival rate of approximately
with poor prognosis. Preliminary clinical research on GPC3-specific CAR-T treatment has 25%. Immune checkpoint inhibitors (ICIs), including the combination of durvalumab (Durva)
validated GPC3 as a therapeutic target in HCC. Its expression is largely confined to the & tremelimumab (Treme), represents the current standard-of-care for frontline HCC
surface of tumor cells, making it an ideal target for TCEs. AZD9793, a trispecific IgG1 treatment. However, therapy resistance, either primary or secondary, often arises due to the
monoclonal antibody, is a first-in-class CD8-targeted TCE that directly engages tumor- immunosuppressive tumor microenvironment (TME). VEGFR is a well-established thera-
infiltrating T cells and GPC3+ tumors, forming a bridge that activates T cells leading to tumor peutic target in HCC. Cabozantinib (Cabo), a multi-target tyrosine kinase inhibitor (TKI), is
cell lysis and T cell proliferation. AZD9793 promotes potent GPC3+ HCC cell killing through approved for later-line use per the CELESTIAL trial, and it also demonstrated significant TME
preferential engagement of CD8+ T cells, while minimizing CD4+ T cell activation and modulation through antiangiogenic effects. A Phase II study (CHECKMATE 040) reported an
unwanted cytokine release. The novel mechanism of action combines bivalent GPC3 binding, impressive objective response rate (ORR) of 29% with the combination of ipilimumab/
CD8-biased engagement, and low-affinity T cell receptor binding to improve cytotoxicity and nivolumab and Cabo but noted high toxicity rates. Zanzalintinib (Zanza) is a novel TKI
reduce the risk of cytokine release syndrome compared with other TCEs. Methods: RHEA-1 targeting VEGFR, MET, & TAM kinases (TYRO3, AXL, MER), key mediators of angiogenesis,
is the FIH trial of AZD9793 monotherapy. Eligible patients in this modular, Phase I/II, open- tumor growth, metastasis, and TME immunosuppression. With a target profile similar to
label, multicenter study are adults ($18 years old) with prospective centrally determined Cabo but an improved pharmacokinetic profile & a shorter half-life (16–22 hours), Zanza has
GPC3+ advanced or metastatic HCC with $1 measurable lesion by RECIST v1.1, who have demonstrated potential synergy with ICIs in preclinical and early-phase trials, suggesting
received $1 line of prior systemic treatment and have an ECOG performance status of 0 or 1. enhanced sensitivity by fostering an immune-permissive TME. This phase II study evaluates
Patients with hepatitis B are eligible if they receive antiviral treatment to ensure adequate the safety & efficacy of Zanza combined with Durva and Treme in HCC. Methods: This open-
viral suppression before enrollment and for $6 months after the study; and with hepatitis C if label, non-randomized Phase II trial consists of two parallel cohorts. Eligible patients must
they are being managed per local practice. The study includes Module 1 (intravenous have unresectable HCC, be treatment-naı̈ve in the unresectable setting, & have ECOG
AZD9793) and Module 2 (subcutaneous AZD9793), each comprising dose escalation (Part A) performance status of 0-1. Exclusion criteria include Child-Pugh score .7, known auto-
and dose expansion (Part B). Module 1 Part A1 (fixed dose) will start with an accelerated immune diseases, heightened risk of gastrointestinal perforation or fistula formation, and
titration design and will then switch to a modified toxicity probability interval-2 algorithm known gastric or esophageal varices. The study begins with a safety lead-in phase of 9–12
after the first 4 dose cohorts or earlier if dose-limiting toxicities (DLTs) are reported. Part A2 patients to establish the recommended Phase II dose. The two cohorts aim to explore the
(step-up dosing) may open in either Module based on emerging safety data from Part A1. optimal sequential strategy for combining Zanza with Durva & Treme. Cohort A: Zanza is
Part B may be initiated in one or both Modules. Primary endpoints include safety and administered during Cycle 1, followed by Durva + Treme in Cycle 2. Cohort B: Durva + Treme
tolerability in terms of DLTs (only in dose escalation) and adverse events to establish is administered during Cycle 1, followed by Zanza and Durva in Cycle 2. Both cohorts will
maximum tolerated dose, optimal biological dose, and recommended phase II dose; and continue with Zanza and Durva in subsequent cycles. A total of 40 participants (20 per
objective response rate (only in dose expansion) by investigator assessment (IA) per RECIST cohort) will be enrolled. The primary endpoint is the ORR assessed by imRECIST 1.1.
v1.1. Secondary endpoints include preliminary efficacy (only in dose escalation) in terms of Secondary endpoints include the conversion rate to resectable or transplant-eligible dis-
response and progression-free survival by IA, as well as overall survival (only in dose ex- ease, disease control rate, median PFS and OS, and landmark PFS & OS at 6, 12, 24, and
pansion), pharmacokinetics, immunogenicity, and CD8+ T cell infiltration pre- and post- 36 months. Safety and tolerability will also be evaluated. Comprehensive translational
treatment. No formal statistical hypothesis is proposed; all variables will be reported de- analyses include bulk RNA sequencing, spatial transcriptomics of baseline tumor biopsies, &
scriptively. The study (NCT06795022) is currently enrolling in the US and APAC. Clinical trial serial ctDNA monitoring. Trial enrollment commenced in December 2024. Clinical trial
information: NCT06795022. Research Sponsor: AstraZeneca. information: NCT06698250. Research Sponsor: Exelixis.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 323s

TPS4217 Poster Session TPS4218 Poster Session

A randomized phase 2 study of casdozokitug, an IL-27 targeting antibody, in A first-in-human study of MT-303, an innovative in vivo mRNA chimeric
combination with toripalimab plus bevacizumab in patients with unresect- antigen receptor (CAR) therapy targeting GPC3, in adults with hepatocellular
able and/or locally advanced or metastatic hepatocellular carcinoma. First carcinoma. First Author: Timothy Guy Humphries, Sir Charles Gairdner Hospital/
Author: Daneng Li, City of Hope National Comprehensive Cancer Center, Duarte, CA University of Western Australia, Nedlands, Australia
Background: IL-27 is a heterodimerized cytokine, a member of the IL-12/IL-23 cytokine Background: Hepatocellular carcinoma (HCC) remains a leading cause of cancer
family, and an immunoregulatory cytokine expressed by myeloid cells that dampens T and mortality worldwide, with advanced cases posing significant treatment challenges.
NK effector function. IL-27 is highly expressed by tumor-associated macrophages in several GPC3, a cell surface protein highly expressed in HCC, represents a promising therapeutic
cancers, including hepatocellular carcinoma (HCC) and non-small cell lung cancer (NSCLC), target. MT-303 is an in vivo chimeric antigen receptor (CAR) therapy, leveraging mRNA-
and suppresses antitumor immune responses. Casdozokitug (Casdozo) is the only clinical- lipid nanoparticle (LNP) technology to reprogram myeloid cells directly within the body.
stage IL-27 targeting antibody and it increases IFN-g and T and NK cell activation in This novel platform eliminates the logistical and technical barriers of ex vivo CAR
preclinical/clinical studies. A phase 1 study (NCT04374877) demonstrated a favorable therapies while retaining the ability to activate targeted immune responses. MT-303’s
safety profile and antitumor activity of casdozo as monotherapy and in combination with mRNA encodes a GPC3-targeted CAR receptor incorporating a single-chain variable
PD-1 blockade in indications, including HCC, with known high levels of IL-27 activation fragment (scFv) linked to the transmembrane domain and cytoplasmic tail of CD89.
signature (Marron T, et al. Ann Oncol. 2023). A phase 3 study of toripalimab (tori) + Crucially, functional CAR expression is restricted to Fc receptor common gamma chain-
bevacizumab (bev) demonstrated significant improvements in efficacy (overall survival [OS], expressing cells, predominantly myeloid cells, ensuring precise immune activation.
progression-free survival [PFS], and objective response rate [ORR]) compared to sorafenib
Preclinical studies demonstrated that MT-303 effectively infiltrates tumors, triggers
(Yinghong S, et al. CSCO 2024) in HCC. A phase 2 study of casdozo + atezolizumab + bev
tumor cell killing, produces chemokines and cytokines, eliciting an adaptive anti-tumor
showed an acceptable safety profile and antitumor activity (ORR 38%, CR 17.2%, mPFS 8.1
immunity. Rodent models of “cold” tumors and nonhuman primate studies have
mo) (Li D, et al. ASCO GI 2025). CHS-388-202 (NCT06679985) will evaluate the efficacy,
safety, and biomarkers of tori + bev 6 casdozo and optimize the dose for casdozo in
highlighted MT-303’s safety, pharmacodynamic effects, and dose-dependent activity,
combination with tori + bev as first-line treatment for patients (pts) with unresectable and/or supporting its clinical development (Argueta, SITC 2024, #1125). Methods: This first-in-
locally advanced/metastatic HCC. Methods: CHS-388-202 is a Phase 2, open-label, ran- human, multicenter, open-label, Phase 1 dose-escalation trial evaluates the safety,
domized study and will enroll up to 72 pts randomized ([Link]) to 1 of 3 treatment arms (IV tolerability, pharmacokinetics (PK), pharmacodynamics (PD), and preliminary efficacy of
Q3W): Arm A (tori 240 mg + bev 15 mg/kg + casdozo 700 mg), Arm B (tori 240 mg + bev 15 MT-303 in participants with GPC3-expressing tumors, with a primary focus on Hepa-
mg/kg + casdozo 1400 mg), Arm C (tori 240 mg + bev 15 mg/kg). Key eligibility criteria tocellular Carcinoma. MT-303 is administered intravenously every 14 days using a
include treatment-naı̈ve unresectable metastatic HCC with $1 measurable lesion; not Bayesian Optimal Interval design, with backfill cohorts for dose refinement. The primary
suitable for surgical or local therapy; Child-Pugh A; ECOG PS 0 or 1; controlled hepatitis B endpoints are safety, maximum tolerated dose (MTD), and recommended Phase 2 dose
virus or cured hepatitis C virus. Pts will be stratified by geographic region (Asia excluding (RP2D). Secondary endpoints include detailed PK profiling and assessment of immune-
Japan vs the rest of the world) and macrovascular invasion or extrahepatic spread of related adverse events (e.g., ICANs, CRS). Exploratory endpoints encompass efficacy
disease (presence vs absence). Primary endpoints are ORR by investigator review according measures (e.g., objective response rate [ORR], duration of response [DOR]), immune
to RECIST v1.1 and safety. Secondary endpoints are ORR by investigator review according to reprogramming (e.g., peripheral CAR expression, cytokine profiles), and intratumoral
HCC modified RECIST (mRECIST) criteria; duration of response, PFS, and disease control immune changes, including T-cell receptor clonality and GPC3 modulation. Enrollment is
rate by investigator review according to RECIST v1.1 and mRECIST criteria; OS, pharma- ongoing, with safety and preliminary efficacy data expected to inform future devel-
cokinetics. A safety run-in evaluation will be conducted after the first ~6 pts are enrolled in opment. Clinical trial information: NCT06478693. Research Sponsor: None.
Arms A and B, with $3 from each arm completing 1 cycle of treatment. Pts will remain on
study treatment for #2 years or until documented disease progression or unacceptable
toxicity. Enrollment is ongoing. Clinical trial information: NCT06679985. Research Sponsor:
Coherus BioSciences.

TPS4219 Poster Session TPS4220 Poster Session

A phase 1b/2, safety lead-in and dose-expansion trial of ivosidenib plus A phase 2, randomized, multicenter study of adjuvant adebrelimab plus
durvalumab and gemcitabine/cisplatin as first-line therapy in patients with capecitabine in resected cholangiocarcinoma with high-risk factors:
locally advanced, unresectable or metastatic cholangiocarcinoma with an ACHIEVE. First Author: Xiangcheng Li, Hepatobiliary Center, the First Affiliated
IDH1 mutation. First Author: James J. Harding, Memorial Sloan Kettering Cancer Hospital of Nanjing Medical University, Nanjing, China
Center, New York, NY Background: Cholangiocarcinoma is a rare and aggressive group of gastrointestinal
Background: Cholangiocarcinomas (CCAs) are often advanced and incurable at the time of cancers. For early-stage disease following curative resection, capecitabine is a category
diagnosis. The phase 3 TOPAZ-1 trial showed improved OS and ORR with gemcitabine/ 1 recommendation for adjuvant therapy for biliary tract cancer (BTC) according to the
cisplatin (GEM/CIS) and durvalumab (DURVA) vs GEM/CIS in unresectable advanced or NCCN guidelines. However, recurrence rates remain high. For example, the BILCAP study
metastatic biliary tract cancers. The phase 3 ClarIDHY trial demonstrated that the mIDH1 reported a 5-year recurrence-free survival (RFS) rate of 34% with adjuvant capecitabine.
inhibitor ivosidenib (IVO) improved progression-free survival in CCA patients who have Based on cholangiocarcinoma-specific cohort data from the Hepatobiliary Center, The
progressed from first or second-line chemotherapy and who have activating mutations in First Affiliated Hospital of Nanjing Medical University, the 1-year RFS rate for intra-
isocitrate dehydrogenase-1 (mIDH1). Additionally, mIDH1 suppresses key immune-related hepatic cholangiocarcinoma (ICC) and hilar cholangiocarcinoma (HCCA) with high-risk
genes, with reversal of this effect when mIDH1 inhibitors are administered in preclinical CCA factors is approximately 50% (unpublished data), highlighting a substantial unmet
models. Finally, encouraging activity has been observed in treatment-naive mIDH1 patients medical need. Immunotherapy has shown efficacy as adjuvant therapy in other cancer
administered with the mIDH1 inhibitor LY3410738 in combination with GEM/CIS. Given the types. Results from the TOPAZ-1 and KEYNOTE-966 studies support the combination of
ability of ivosidenib to stabilize advanced CCA, ability of IDH1 inhibition to restore immune immunotherapy and chemotherapy for advanced BTC, including locally advanced non-
activity, promising clinical activity of an mIDH1 inhibitor in combination with GEM/CIS, and
metastatic disease. Adebrelimab, a PD-L1 inhibitor, has shown promising results in
the limited overlapping toxicities of these treatments, this study seeks to explore safety and
several cancers. In China, it is approved for first-line treatment of extensive-stage small
preliminary activity of the quadruplet combination. Methods: This is a phase 1b/2, mul-
cell lung cancer with chemotherapy. The ACHIEVE study will assess the efficacy of
ticenter, safety lead-in and dose expansion, open-label study of IVO in combination with
DURVA/GEM/CIS in first-line therapy of locally advanced, unresectable, or metastatic CCA
adebrelimab plus standard adjuvant chemotherapy in ICC/HCCA patients after curative
with mIDH1. Treatment with up to one cycle of DURVA/GEM/CIS is permitted before ini- resection with high-risk factors. Methods: ACHIEVE is a Phase 2, randomized, open-
tiation of study treatment. Key eligibility criteria include: a histopathological diagnosis; label, multicenter study designed to assess the efficacy and tolerability of adebrelimab
tumor mIDH1 based on local or centralized tissue testing (local testing by plasma ctDNA administered intravenously every three weeks for one year in combination with
may be used); at least 1 measurable lesion as defined by RECIST v1.1; and adequate bone capecitabine (8 cycles) as adjuvant therapy for ICC or HCCA after curative resection. The
marrow, hepatic, and renal function. The study has a safety lead-in phase where IVO will be study will enroll about 120 adult patients with histologically confirmed ICC or HCCA who
administered orally to the first 6 patients at a starting dose of 500 mg QD on every day of the have undergone complete resection (R0). Eligible participants must have an ECOG
21-day cycle, plus DURVA 1500 mg IV infusion every 3 weeks for up to 8 cycles, plus GEM performance status of 0–1, confirmed complete response (CR) on imaging 4–8 weeks
1000 mg/m2 IV and CIS 25 mg/m2 IV on days 1 and 8 of each 21-day cycle, followed by IVO post-surgery, and at least one high-risk factor. High-risk factors are defined as follows:
500 mg QD and DURVA 1500 mg every 4 weeks of a 28-day cycle. Dose-limiting toxicities ICC: Single tumor . 5 cm, multiple tumors, liver capsule breach, vascular invasion,
(DLTs) will be evaluated during the first cycle of quadruplet treatment. 6 additional patients regional lymph node metastasis; HCCA: Tumor invasion into surrounding tissues,
may be enrolled to evaluate an alternative reduced dose of IVO 250 mg QD. The primary vascular invasion, regional lymph node metastasis. Key exclusion criteria include locally
objective is to evaluate the safety and tolerability of the quadruplet combination, and to advanced, unresectable, or metastatic disease at diagnosis and prior anti-cancer therapy
determine the recommended combination dose (RCD). The expansion phase will enroll before surgery. The primary endpoint is the 1-year recurrence-free survival rate (RFSR).
approximately 40 patients who will be treated with the RCD, with the primary objective being Key secondary endpoints include overall survival (OS) and RFS, minimal residual disease
to assess the clinical activity of the combination, as determined by a primary endpoint of (MRD), and patient-reported tolerability, and safety. Enrollment has begun, with six sites
confirmed complete or partial response using RECIST v1.1 criteria. Clinical trial information: in mainland China participating. Clinical trial information: NCT06607276. Research
NCT06501625. Research Sponsor: Servier. Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
324s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

TPS4221 Poster Session TPS4222 Poster Session

EMERALD-Y90: A phase 2 study to evaluate transarterial radioembolization Trial in progress: A phase Ib/II study to evaluate the safety and efficacy of
(TARE) followed by durvalumab (D) and bevacizumab (B) for the treatment atezolizumab plus bevacizumab as adjuvant therapy following carbon ion
of participants (pts) with unresectable hepatocellular carcinoma (uHCC) radiotherapy in hepatocellular carcinoma (VANGUARD trial). First Author:
eligible for embolization. First Author: Riad Salem, Northwestern Feinberg School of Keisuke Koroki, Department of Gastroenterology, Graduate School of Medicine, Chiba
Medicine, St. Clair, Chicago, IL University, Chiba, Japan
Background: Locoregional therapy, such as transarterial chemoembolization (TACE) or Background: While hepatic resection remains the standard curative treatment for
TARE, is commonly used to treat uHCC eligible for embolization. Despite advances in TACE hepatocellular carcinoma (HCC), many cases are ineligible due to impaired liver function
and TARE delivery, median progression-free survival (PFS) following treatment is , 1 year, or patient-related factors. Percutaneous ablation is an option for small HCC but is limited
highlighting a need for new treatment options. EMERALD-1 (NCT03778957), a global phase by tumor size. Carbon ion radiotherapy (C-ion RT) has emerged as a promising modality,
3 study, demonstrated a statistically significant improvement in PFS with TACE + D + B characterized by superior biological efficacy and dose distribution compared to con-
versus TACE + placebos in pts with uHCC eligible for embolization (hazard ratio, 0.77 [95% ventional radiotherapy. Although HCC is relatively radiosensitive, conventional radio-
confidence interval, 0.61–0.98]; two-sided p-value = 0.032). With the increased use of TARE therapy has limited efficacy due to low radiation tolerance of surrounding liver tissue. C-
for patients with uHCC eligible for embolization in the US, a need exists for evidence to ion RT achieves effective treatment while minimizing radiation exposure through the
support additional treatment options in settings where TARE is preferred. The EMERALD- superior dose localization of the carbon ion beam’s Bragg peak. The establishment of
Y90 study will evaluate the efficacy and safety of TARE with D monotherapy, followed by D + adjuvant systemic therapy to prevent recurrence remains an urgent unmet need in HCC
B in pts with uHCC eligible for embolization. Methods: EMERALD-Y90 (NCT06040099) is a management. The IMbrave050 trial demonstrated the efficacy of combined atezoli-
phase 2, single-arm study that will enroll ~100 pts aged $18 years with uHCC amenable to zumab and bevacizumab (Atezo+Bev) after resection or ablation, but recent analyses
embolization who are ineligible for or have declined treatment with resection and/or ablation
suggest diminishing long-term benefits. The combination of immune checkpoint in-
or liver transplant (transplant candidates are those listed for transplant or eligible to be
hibitors (ICIs) and radiotherapy has shown promise in several malignancies, with
listed and within Milan criteria). Eligible pts also must have Child-Pugh class A liver function
and an Eastern Cooperative Oncology Group performance status of 0–1. Pts are allowed to
preclinical studies demonstrating synergistic enhancement of ICI efficacy through
receive a single TACE or TARE $6 months before the study or .1 TACE or radiation-induced immunogenic cell death. Additionally, carbon ion radiation induces
TARE $12 months before the study. Prior TACE or TARE must have been administered for a stronger immune responses compared to proton therapy. Based on these findings, we
different primary intrahepatic lesion unrelated to the current lesion, and pts should have a designed a phase Ib/II study to evaluate sequential C-ion RT followed by ICI as a novel
functional liver remnant . 30%. Exclusion criteria include prior systemic therapy, evidence therapeutic approach. Methods: This multicenter, open-label, single-arm phase Ib/II
of extrahepatic spread, or major portal vein invasion. Pts will receive partition-based dosing study evaluates the safety and efficacy of Atezo+Bev as adjuvant therapy following C-ion
of TARE using Y-90 glass microspheres. Following TARE, pts will receive D 1500 mg (one RT for HCC. Key inclusion criteria include treatment-naı̈ve, Child-Pugh class A, maximum
dose) followed by D 1120 mg + B 15 mg/kg every 3 weeks until study completion or intrahepatic tumor diameter $4 cm and #3 intrahepatic tumors. After initial enrollment,
discontinuation criteria are met. The primary endpoint is PFS (time from start of TARE until patients undergo C-ion RT followed by a two-week observation period with eligibility
date of disease progression [investigator (INV)-assessed per modified Response Evaluation screening for second enrollment. Eligible patients receive atezolizumab (1200 mg) and
Criteria in Solid Tumors (mRECIST)] or death due to any cause). Secondary endpoints bevacizumab (15 mg/kg) every 3 weeks for up to 48 weeks, with radiological as-
include safety and tolerability, 6-, 12-, and 24-month PFS, objective response rate (per- sessments every 3 months. The Phase Ib part will enroll six patients to evaluate dose-
centage of pts with confirmed complete or partial response [INV-assessed per mRECIST]), limiting toxicities. Secondary endpoints include adverse events (AEs) and serious AEs
overall survival (time from start of TARE until death due to any cause), and duration of for safety and 1-year RFS, overall survival (OS) and 6-month OS rates for efficacy. If
response (time from date of first documented response until date of progression or death tolerability is confirmed, the trial will proceed to Phase II. Clinical trial information:
due to any cause). An early safety review is planned when approximately 15 pts have jRCT2031240284. Research Sponsor: Chugai Pharmaceutical Co., Ltd.
completed their first cycle of D + B dosing. Study enrollment is ongoing at 22 US sites.
Clinical trial information: NCT06040099. Research Sponsor: AstraZeneca.

TPS4223 Poster Session TPS4224 Poster Session

Donafenib combined with capecitabine for postoperative adjuvant therapy Australasian Gastro-Intestinal Trials Group (AGITG) STOPNET: A random-
of biliary malignant tumors with high risk of recurrence: A multi-center, ized study of cessation of somatostatin analogues (SSA) after peptide
randomized controlled, phase II study. First Author: Jianhua Rao, Hepatobiliary receptor radionuclide therapy (PRRT) in mid, hind-gut, and pancreatic
center, Jiangsu Province Hospital - The First Affiliated Hospital with Nanjing Medical neuroendocrine tumours. First Author: Matthew E. Burge, Royal Brisbane and
University, Nanjing, China Women’s Hospital, Herston, QLD, Australia
Background: Biliary Tract Cancer (BTC) is an aggressive malignancy with rising incidence. Background: PRRT is a standard therapeutic option for patients with metastatic well-
Surgery is the only curative option, but only 10% of patients are eligible at diagnosis, and differentiated somatostatin receptor-positive gastroenteropancreatic neuroendocrine tu-
recurrence rates post-surgery can reach 67% within a year. The 5-year survival rate is only 5- mours (GEP-NETs) following progression on SSA. It is uncertain whether current practice of
15%. Emerging therapies, such as targeted and immunotherapies, show promise. A study continuing SSA after commencing PRRT is beneficial, especially in non-functioning NETs.
combining GEMOX, tislelizumab, and donafenib (a tyrosine kinase inhibitor) in advanced BTC Studies by Yordanova et al. (2018) and Sygula et al. (2022) have included heterogenous study
showed an 87.5% disease control rate (DCR) with strong safety and efficacy. The BILCAP populations and yielded conflicting results. Methods: STOPNET is a prospective, randomized,
study found that adjuvant capecitabine improved overall survival (OS) in resected BTC non-comparative, open-label, multi-center phase II trial led by the AGITG in collaboration with
patients (median OS: 49.6 vs. 36.1 months; HR = 0.84). However, clinical research on the Canadian Cancer Trials Group (CCTG) under the Commonwealth Neuroendocrine Tumor
adjuvant treatments for high-risk postoperative BTC remains limited, with no consensus on research collaborative (CommNETS). The trial aims to evaluate the outcomes of SSA cessation
high-risk factors. This study evaluates the efficacy and safety of donafenib combined with or continuation in patients with GEP-NETs undergoing PRRT after progression on SSA. The co-
capecitabine as adjuvant therapy for postoperative BTC with high recurrence. Methods: The primary endpoints are 20-month PFS and feasibility for a phase III trial, assessed by re-
study selected BTC patients prior to radical resection without any anti-tumor systemic cruitment over a 24-month period & patient acceptance of SSA cessation. Secondary end-
therapy (including radiotherapy, chemotherapy, targeted therapy, immunotherapy) with at points include OS, rate of SSA recommencement, time to subsequent therapy, quality of life,
least one high-risk postoperative recurrence factors including specific stages according to cost-effectiveness and psycho-oncological impacts of SSA cessation. Eligible participants
the UICC/AJCC TNM 8th edition staging system: T2-4, N0, M0 or T1-4, N1, M0 (applicable to must have advanced or unresectable WHO grade 1/2 non-functioning GEP-NETs (excluding the
extrahepatic cholangiocarcinoma); T1b-4, N0-1, M0 or T1a, N1, M0 (applicable to intrahepatic foregut), and disease progression after receiving $3 months of SSA at standard growth-
cholangiocarcinoma), vascular invasion or neurophilic invasion as research subjects. Pa- control doses. SSA must have been primarily commenced for growth control, as opposed to
functional symptoms and for mid/hindgut NET’s 24-hr urine 5-hydroxyindoleacetic acid must
tients will be randomly divided into 1:1 groups. The experimental group consisted of
be , 1.5 times upper limit normal at screening. Participants will be randomized 2:1 to SSA
donafenib (200mg, bid for 6 months) combined with capecitabine (1250mg/m2, bid, treated
cessation or continuation. SSA cessation arm will receive their last SSA $28 days prior to first
for 2 weeks and stopped for 1 week, with 3 weeks as a treatment cycle, 8 cycles). The control
PRRT, and the SSA continuation arm will continue SSA during and after PRRT. Following PRRT,
group was capecitabine (same as experimental group). Treatment will start at least 4 weeks
participants will be assessed every 12 weeks (minimum 20 months) until disease progression
after radical resection and stop until patients experience disease recurrence or intolerable
or study closure, whichever occurs first. The sample size was calculated using Fleming’s single
toxic side effects. The primary endpoint of the study was the 1-year recurrence free survival stage design, assuming uninteresting and interesting 20-month PFS rates of 60% and 77%
(RFS) rate. Secondary endpoints consisted of 2-year RFS, OS and safety assessment in- respectively. Novel translational research will be conducted to define and validate NET tissue
cluding incidence, severity, and relationship to study drugs of all adverse events (AEs), and circulating biomarkers, with a particular focus on analysis of microRNA. Formalin-fixed
treatment-related adverse events (TRAEs), and serious adverse events (SAEs). Based on the paraffin-embedded (FFPE) tumor tissue will be retrieved (if available), with the collection of
data analysis of BTC cohort at our center, the 1y-RFS rate for the control group is set at 30%, bloods at 3 time-points during study. The trial implemented the Australian Teletrial Program
while that for the experimental group is set at 60%. With a two-sided alpha of 0.05, a power (ATP) to enhance equity of access for participants in regional, rural or remote locations. The
of 0.80, and a randomization ratio of 1:1, the required number of RFS events is 64. trial will enroll 78 participants across 13 sites. Enrolment is open at 1 site in Australia & 4 sites
Considering a 10% dropout rate, it is planned to enroll 35 participants per group, with a total in Canada, with 3 participants randomized as of Jan 2025. Clinical trial information:
planned enrollment of 70 participants. Dated by 20 January 2025, 8 of planned 70 patients NCT06345079. Research Sponsor: Medical Research Future Fund (MRFF); Tour de Cure; AGITG
have been enrolled. Clinical trial information: NCT06685289. Research Sponsor: None. philanthropic funding; Canadian Neuroendocrine Tumour Society (CNETS); Canadian Institutes
of Health Research (CIHR).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 325s

TPS4225 Poster Session TPS4226 Poster Session

A multi-centre, stratified, open, randomized, comparator-controlled, parallel NCI 10479: A phase I dose escalation-expansion trial of sunitinib malate
group phase II trial comparing adjuvant treatment with 177Lu-DOTATATE to plus lutetium (Lu-177) dotatate in somatostatin receptor positive pancreatic
standard of care in patients after resection of neuroendocrine liver metas- neuroendocrine tumors. First Author: Nikolaos Trikalinos, Washington University
tases (NELMAS). First Author: Andreja Frilling, Imperial College London, London, School of Medicine in St. Louis, Siteman Cancer Center, St. Louis, MO
United Kingdom Background: Patients with metastatic or unresectable pancreatic neuroendocrine tumors
Background: Gastro-entero-pancreatic (GEP) neuroendocrine tumours (NET) are (PanNETs) have a poor prognosis, even with currently available treatments, with a 5-year
steadily increasing in incidence and prevalence. About 65%-95% of GEP NET show overall survival (OS) of less than 20%. Lutetium Dotatate (Lu-177) is a radiopharmaceutical
hepatic metastases. Surgery is the mainstay of treatment for NE LM. While macro- that consists of the somatostatin analogue DOTA-Tyr3-Octreotate, coupled to the metal-
scopically complete resection for NE LM is associated with favourable overall survival ion chelating moiety, DOTA, and radiolabeled with lutetium-177. Lu-177 was approved by
(OS), recurrence rates of up to 70% at 3 years and up to 95% at 5 years are reported. the FDA in 2018 for treatment of somatostatin receptor (SSTR)-positive gastro-
These results call for adjuvant treatment concepts which have not yet been established. enterohepatic NETs, but it is limited in its efficacy to achieve cytoreduction and provide
Methods: A prospective open-label, multicentre randomised parallel-group trial was durable responses. Sunitinib malate is an oral small-molecule tyrosine kinase inhibitor
conducted in patients with resected GEP NE LM. Adjuvant treatment with 177Lu-DOTA0- targeting VEGFRs, PDGFRs, and KIT and is also FDA approved as a monotherapy for the
Ty3-ocreotate (177Lu-DOTATATE) (total administered activity 14.8 GBq) is compared treatment of metastatic unresectable PanNETs. There is preclinical, as well as clinical
with standard of care (SOC). The frequency of administration is 2 cycles (861weeks evidence of sunitinib being used as a radiosensitizer with classic radiation, but it has never
between each cycle). The first cycle is applied 862 weeks after liver resection. The been combined with a radiolabeled analogue in patients with PanNETs. Methods: This is a
control arm consists of SOC. Main inclusion criteria are well differentiated grade 1 or Phase I dose escalation/expansion study aiming to enroll up to 24 patients across several
grade 2 (Ki67 , 20%) GEP NET, R0 or R1 resection of NE LM, primary tumour already sites. Eligible patients will be offered fixed dose Lu-177 at 200 mCi for 4 fractions with
resected or resected synchronously with LM, 68Ga DOTATATE PET/CT prior to surgery concurrent oral sunitinib administration initiating on C1D1 and concluding 28 days after
confirming LM and no extrahepatic disease (except resectable perihilar lymph node the last Lu-177 infusion. Dose escalation applies to sunitinib and will be guided by a 3+3
involvement and/or primary tumour, if still in place). Main exclusion criteria are high design to determine the maximum tolerated dose (MTD) and recommended phase 2 dose
grade NET, neuroendocrine carcinoma, R2 resection of LM, peptide receptor radionuclide (RP2D). Once the RP2D has been established, up to 12 more patients will be offered
participation in the expansion phase in an attempt to further record antitumor activity and
therapy at any time prior to randomisation in the study, and any type of liver directed
correlation with imaging, tumor markers, as well as Lu-177 dosimetry. Treatment will
therapy within 12 weeks prior to randomisation in the study. Primary endpoint are
continue until disease recurrence/progression, unacceptable toxicity, or completion of
disease-free survival (DFS) at 3 years after liver resection. The sample size of 106
planned protocol. Key eligibility criteria include age . = 18 years, ECOG performance
patients in total is powered to detect an HR of 0.27, reflecting a 44% DFS probability at 3
status , = 2, histologic diagnosis of metastatic, unresectable well- or moderately-
years post-surgery in the 177Lu-DOTATATE arm compared with a 25% in the SOC arm. differentiated SSTR-positive PanNETs of any grade, up to 1 prior treatment except for
Secondary endpoints OS, time to tumour recurrence, time to administration of sub- somatostatin analogues and appropriate baseline hematological parameters. Key ex-
sequent antineoplastic therapy, safety and tolerability of 177Lu-DOTATATE, health- clusion criteria are prior use of sunitinib, Lu-177 or other radiopharmaceuticals, myocardial
related quality of life, patient reported outcomes, and cost effectiveness. Ancillary or cerebrovascular accident within the prior 12 months and left ventricular ejection fraction
objectives explore the clinical utility of novel molecular based biomarkers in identifi- of , = 50%. The study uses an 8-week safety window to determine its primary endpoint,
cation of residual microscopic disease and early detection of recurrent disease. which is DLTs during administration of the combination. Secondary endpoints are ob-
Enrolment has begun. Follow-up data will be collected for 5 years overall from the date of jective response (ORR), duration of response (DOR), progression-free survival (PFS) and
randomisation of the last patient. Discussion: The NELMAS trial aims to investigate the overall survival (OS), intensity of tumor uptake on pre-treatment SSTR PET and post Lu-
efficacy of adjuvant therapy with 177Lu-DOTATATE (2 cycles) compared to standard of 177, chromogranin A level response as well as optional dosimetry imaging. Enrollment is
care in preventing tumour recurrence in patients following R0/R1 resection of LM of well ongoing. Clinical trial information: NCT05687123. Research Sponsor: National Cancer
differentiated GEP NET. Clinical trial information: NCT05987176. Research Sponsor: Institute.
Novartis/AAA; The Taylor Family 2010 Charitable Trust.

TPS4227 Poster Session TPS4228 Poster Session

An open-label, dose-finding, phase Ib study to assess the safety, tolerability ALTER-PA001: A multicenter, randomized study of anlotinib and benmel-
of nesuparib (JPI-547), a dual inhibitor of PARP/TNKS, in combination with stobart in combination with AG chemotherapy vs. AG as first-line treatment
modified FOLFIRINOX (mFOLFIRINOX) and gemcitabine-nab-paclitaxel for metastatic pancreatic cancer. First Author: Jiujie Cui, Department of Medical
(GemAbraxane) in patients with locally advanced and metastatic pancreatic Oncology and State Key Laboratory of Oncogene and Related Genes, Shanghai Cancer
cancer. First Author: Do-Youn Oh, Department of Internal Medicine, Seoul National Institute, Renji Hospital, School of Medicine, Shanghai, China
University Hospital Cancer Research Institute, Seoul National University College of Background: Metastatic pancreatic cancer (mPC) remains one of the most challenging
Medicine, Seoul, South Korea malignancies to treat, with limited effective therapeutic options. While AG chemotherapy
Background: Pancreatic ductal adenocarcinoma (PDAC) is a leading cause of cancer deaths, (nab-paclitaxel and gemcitabine) is a current standard first-line regimen, new combi-
with most cases diagnosed at advanced stages. Current maintenance therapy with the PARP nations are needed to improve outcomes. Preclinical data suggest potential synergistic
inhibitor, olaparib, benefits only patients with germline BRCA1/2 mutations (Approximately 5- effects of anlotinib, a multi-target tyrosine kinase inhibitor, and benmelstobart, a novel
7 % of PDAC cases). However, homologous recombination deficiency (HRD)-related muta- anti-PD-L1 antibody, with chemotherapy. This study aims to evaluate the efficacy and
tions occur in ~15% of PDACs, potentially expanding the utility of PARP inhibitors. safety of this combination in mPC. Methods: ALTER-PA-001 is a multicenter, open-
Nesuparib, a next-generation PARP inhibitor, also targets tankyrase, disrupting WNT and label, randomized, controlled phase 2 trial that compared anlotinib plus benmelstobart
Hippo signaling pathways which are critical for homologous recombination repair. This dual and AG with AG in patients with treatment-naı̈ve mPC. Eligible patients are aged 18-75,
mechanism mimics BRCA loss ("BRCAness"), sensitizing HRD-positive tumors without BRCA ECOG 0-1, with histologically or cytologically confirmed PC. A total of 104 patients will be
mutations to PARP inhibition, broadening therapeutic options. Combining PARP inhibitors
randomly assigned in a 2:1 ratio to receive anlotinib (8 mg orally, QD, d1–14), ben-
with chemotherapy (e.g., irinotecan or platinum-based drugs) enhances sensitivity to DNA
melstobart (1200 mg IV, d1), nab-paclitaxel (125 mg/m² IV, d1, d8), and gemcitabine
damage. Preclinical studies showed nesuparib inhibited tumor growth as monotherapy and
achieved higher efficacy in combination with standard treatments. In a prior phase I trial,
(1000 mg/m² IV, d1, d8) every 21 days or AG regimen with nab-paclitaxel and gem-
nesuparib showed promising antitumor activity, with overall response rate of 28.2% and citabine at the same doses and schedule. The randomisation is done centrally and
disease control rate of 64.1%. This Phase Ib study aims to evaluate the efficacy of nesuparib stratified by the presence of liver metastasis. Patients achieving CR, PR, or SD after 8
in combination with standard chemotherapy for advanced PDAC using a 3+3 dose-escalation cycles will enter a maintenance phase with continued treatment based on their assigned
design. Methods: This multicenter, open-label, Phase Ib, dose-finding study will enroll 24–48 arm. Tumor assessment is performed every 6 weeks for induction treatment, and every
patients with locally advanced or metastatic PDAC. Two arms are included: Arm A 9 weeks for maintenance phase. The primary endpoint is objective response rate (ORR),
(mFOLFIRINOX combination) and Arm B (GemAbraxane combination), each with 12–24 with secondary endpoints including progression-free survival (PFS), disease control rate
subjects across four dose groups (3–6 patients per group). Nesuparib is administered orally (DCR), duration of response (DoR), overall survival (OS), and safety. Exploratory bio-
under fasting conditions ranging from Dose Levels -2 (12.5 mg qd) to 4 (100 mg qd), starting marker analyses will assess correlations between baseline tumor characteristics and
at Dose Level 1 (25 mg qd) with a 5 days on/2 days off schedule. Based on the occurrence of therapeutic outcomes. This trial is actively recruiting in November 2024. Clinical trial
dose-limiting toxicities (DLTs), the dose may be escalated to higher levels (Dose Levels 2, 3, or information: NCT06621095. Research Sponsor: None.
4) or reduced to lower levels (Dose Levels -1 or -2) with a 5 days on/2 days off or 3 days on/4
days off schedule. Arm A includes mFOLFIRINOX chemotherapy with biweekly oxaliplatin
(65 mg/m²), leucovorin (400 mg/m²), irinotecan (135 mg/m²), and 5-FU (2,400 mg/m²). Arm B
involves gemcitabine (1,000 mg/m²) and nab-paclitaxel (125 mg/m²) on Days 1, 8, and 15 of a
28-day cycle. Primary objectives are to determine the maximum tolerable dose (MTD) and
recommended Phase II dose (RP2D) and to identify the optimal combination regimen based
on safety. Secondary objectives include evaluating safety and antitumor activity. Enrollment
began in Q1 2022. [Link] ID: NCT05257993. Clinical trial information:
NCT05257993. Research Sponsor: Onconic Therapeutics.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
326s GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY

TPS4229 Poster Session TPS4230 Poster Session

A phase 1b/2 trial of onvansertib in combination with NALIRIFOX for first line Trial in progress: RASolute 302—A phase 3, multicenter, global, open-label,
treatment of advanced pancreatic cancer (PANCONVA trial). First Author: Anup randomized study of daraxonrasib (RMC-6236), a RAS(ON) multi-selective
Kasi, University of Kansas Cancer Center, Fairway, KS inhibitor, versus standard of care chemotherapy in patients with previously
Background: Pancreatic cancer is a highly lethal disease. Despite research and drug treated metastatic pancreatic ductal adenocarcinoma (PDAC). First Author:
development efforts focused on KRAS, no effective RAS inhibitors have been approved Brian M. Wolpin, Dana-Farber Cancer Institute, Gastrointestinal Cancer Center, Boston,
for the treatment of pancreatic cancer with KRAS mutation. PLK1 inhibition is a potential MA
target in KRAS-mutated pancreatic cancer and may provide a new first-line treatment Background: Patients with previously treated metastatic PDAC have significant need for
option. Onvansertib (also known as PCM-075 and NMS-1286937) is the first PLK1- treatments with improved efficacy and tolerability. RAS is an oncogenic driver in . 90% of
specific adenosine triphosphate competitive inhibitor administered by oral route to enter patients with PDAC. Daraxonrasib (RMC-6236) is an oral, RAS(ON) multi-selective, tri-
clinical trials with proven antitumor activity in different preclinical models. complex inhibitor of GTP-bound mutant and wild-type RAS. In the ongoing Phase 1
Methods: This is a phase 1b/II, non-randomized, open label single arm study being monotherapy trial (NCT05379985), daraxonrasib exhibited a manageable safety profile with
conducted at the University of Kansas Cancer Center and its affiliated sites. The study is primarily low-grade rash and GI toxicities, and encouraging ORR, PFS and OS in a broad
open for enrollment. Eligibility: Key inclusion includes pts with locally advanced, population of previously treated patients with RAS mutant metastatic PDAC (J Clin Oncol 43,
unresectable, or metastatic pancreatic adenocarcinoma who are treatment naı̈ve, have 2025 [suppl 4; abstr 722]). The significant unmet need for alternative treatment options,
adequate archival tissue for biomarker evaluation or are willing to undergo a biopsy, and along with the preliminary clinical data of daraxonrasib monotherapy, support its evaluation
have an ECOG of 0-1. Key Exclusion: Planned concomitant use of medications known to in the ongoing Phase 3 clinical trial, RASolute 302, in patients with previously treated
prolong the QT/QTc interval, use of strong CYP3A4 or CYP2C19 inhibitors or strong metastatic PDAC. Methods: RASolute 302 is a global, multicenter, open-label, randomized
CYP3A4 inducers. Treatment Plan: The phase 1b (safety lead-in) will follow a dose de- study (NCT06625320) designed to evaluate daraxonrasib outcomes compared to investi-
gator’s choice of standard of care chemotherapy as a 2L treatment in patients with
escalation phase in which up to 2 different Onvansertib dose levels will be tested in
metastatic PDAC. Eligibility includes patients $18 years old, ECOG performance status 0 or
combination with standard NALIRIFOX. Onvansertib starting dose level is 30mg orally
1, disease progression on 1 prior line of either a 5-fluorouracil or gemcitabine-based regimen
once daily. The Phase II portion of the study will be a single-arm open-label enrollment in the metastatic setting, and documented RAS mutation status (mutant or wild-type).
with dosing based on the starting dose determination in the Phase Ib portion of the study Eligible RAS mutations are defined as nonsynonymous mutations in KRAS, NRAS, or HRAS
(30mg or 20mg). NALIRIFOX (Nano-liposomal Irinotecan 50 mg/m2, Oxaliplatin 60 mg/ at codons 12, 13, or 61 (G12, G13, or Q61). Patients with tumors that are RAS wild-type and
m2, Leucovorin 400 mg/m2, 5-FU 2400 mg/m2) will be administered intravenously on D1 received appropriate approved targeted therapy for actionable mutations are also eligible. A
of the 14-day cycle. Onvansertib will be dosed orally on D1-5 of each 14-day cycle. 1:1 randomization of approximately 460 patients will receive daraxonrasib 300 mg daily or
Imaging will be performed at baseline and after every 4 cycles. Objectives: The primary investigator’s choice of chemotherapy (gemcitabine/nab-paclitaxel, mFOLFIRINOX, nal-IRI/
objective of this study is to determine anti-tumor activity by measuring Overall Response 5-FU/LV, or FOLFOX) until unacceptable toxicity or disease progression. For patients
Rate (ORR). The secondary objectives are to determine treatment safety based on randomized to daraxonrasib, recommended prophylactic measures for rash include oral
toxicities in participants who have received at least one dose of onvansertib, to de- antibiotics and topical corticosteroids. Response Evaluation Criteria in Solid Tumors version
termine anti-tumor activity by Progression Free Survival (PFS), to determine anti-tumor 1.1 (RECIST v1.1) tumor assessments will be performed every 8 weeks until disease
activity by Disease Control Rate (DCR), to determine Overall Survival (OS). Statistical progression, withdrawal of consent, lost to follow up, or death, whichever occurs first. Dual
Plan: Simon’s two-stage Optimum design will be used. The null hypothesis that the true primary endpoints are progression-free survival (PFS) as assessed by blinded independent
response rate is 41% will be tested against a one-sided alternative that the true response central review and overall survival (OS) in the RAS G12X-mutant population. Key secondary
rate is 65%. In the first stage, 10 evaluable pts will be enrolled. If there are 4 or fewer endpoints include PFS, OS, objective response and quality of life measures in the all-patient
responses in these 10 pts, the study will be stopped. Otherwise, 11 additional evaluable population with tumors carrying RAS mutations (G12X, G13X or Q61X) or RAS wild-type.
pts will be accrued for a total of 21 evaluable pts. The null hypothesis will be rejected if Enrollment for the trial commenced in October 2024. Clinical trial information:
12 or more responses are observed in 21 evaluable pts. Clinical trial information: NCT06625320. Research Sponsor: Revolution Medicines, Inc.
NCT06736717. Research Sponsor: Cardiff Oncology.

TPS4231 Poster Session TPS4232 Poster Session

A phase 2 study of botensilimab and AGEN1423, an anti-CD73-TGFb-trap Adaptive organoid-based precision therapy study in pancreatic cancer
bifunctional antibody, with or without chemotherapy in subjects with ad- (ADOPT): A phase II single-arm study to evaluate the efficacy of patient-
vanced pancreatic cancer. First Author: Bruno Bockorny, Beth Israel Deaconess derived organoid (PDO)-directed therapy in advanced pancreatic ductal
Medical Center, Boston, MA adenocarcinoma (PDAC). First Author: Ronan Andrew McLaughlin, Division of
Background: Traditional immune checkpoint inhibitors have shown limited benefit in Medical Oncology and Hematology, Princess Margaret Cancer Centre – University Health
pancreatic ductal adenocarcinoma (PDAC) owing to non-redundant immune resistance Network, University of Toronto, Toronto, ON, Canada
mechanisms dominating the tumor microenvironment (TME). Transforming growth factor Background: PDAC is a devastating malignancy. High-throughput genomic technologies have
(TGF)-b and cluster of differentiation (CD)73-adenosine represent two major immuno- yielded insights regarding the molecular underpinnings and heterogeneity of PDAC. Systemic
regulatory and pro-tumorigenic pathways responsible for therapeutic resistance and treatment options are limited to cytotoxic chemotherapies, except for approx. 10%, who receive
progressive disease in PDAC. AGEN1423 (also known as dalutrafusp alfa and GS-1423) is a targeted treatment based on genomic profiling. PDO’s are three-dimensional ex vivo experimental
bifunctional, humanized, aglycosylated immunoglobulin G1 kappa antibody that selectively models grown directly from tumor tissue and can provide a direct assessment of drug response. By
directly exposing cancer cells to potential drug therapies, functional profiling provides a dynamic
inhibits CD73-adenosine production and neutralizes active TGF-b signaling. Botensilimab
measurement of response that is more informative than static gene panels. PDOs can theoretically
(BOT) is an Fc-enhanced multifunctional anti-cytotoxic T-lymphocyte-associated protein 4 be used to direct therapeutic decisions, offering an opportunity to expand the reach of precision
(CTLA-4) antibody. We hypothesize that the combination of AGEN1423 with BOT can therapies for PDAC beyond genomics. To date, PDO testing has been limited by small sample sizes,
rescue T-cell functional activity leading to responses in advanced PDAC. Methods: An few drugs included in the screens, and retrospective studies. To expand the impact of precision
investigator-initiated open label Phase 2 study to evaluate the safety, tolerability, and therapy, we developed a rapid high-throughput screening (HTS) platform where over 3,000 drugs can
initial efficacy of BOT + AGEN1423 +/- chemotherapy in patients with metastatic PDAC be tested in PDOs within 8-10 weeks of diagnosis. In ADOPT, we aim to formally investigate the
(NCT05632328). In cohort 1, 12 patients with metastatic PDAC with disease progression to efficacy of PDO-directed therapy in a prospective phase II study, leveraging our existing platforms
at least one line of treatment will receive AGEN1423 30mg/kg IV Q2W for 4 doses + BOT using real-time HTS of PDOs. This study represents one of the first formal trials of PDO-directed
150mg IV Q6W ongoing for up to 2 years. If the combination is considered safe and therapy in solid tumors. Our novel approach will enroll pts with advanced PDAC who do not have
tolerable, and objective response is achieved in at least 1 subject, the study will proceed to alternative treatment options. Methods: This is an actively recruiting prospective, single-arm phase
Cohort 2. In Cohort 2, 12 additional patients with disease progression on first-line II trial. Patients (pts) with advanced epithelial PDAC are eligible if they either: 1) progressed on, were
intolerant to, or refused first-line or subsequent therapies (Cohort A), or 2) have stable disease
fluorouracil-based chemotherapy will be enrolled to receive second-line gemcitabine
after $8 cycles of FOLFIRINOX (“Maintenance” Cohort B) and have a PDO showing sensitivity to an
and nab-paclitaxel in combination with AGEN1423 30mg/kg IV Q2W for 4 doses + BOT approved HC drug. Pts will be recruited, from multiple ongoing studies including PROSPER-PANC
150mg IV Q6W. Key eligibility criteria include histologically or cytologically confirmed where we have successfully generated and tested a PDO. PDO-directed treatment will be selected
metastatic pancreatic adenocarcinoma, age $18 years, Eastern Cooperative Oncology based on drug sensitivity as tested through our validated HTS platform. Each case will be discussed
Group (ECOG) performance status #1, adequate organ function, and measurable disease at our PDO dedicated tumor board. All pts must meet the inclusion/exclusion and drug-specific
by RECISTv1.1. A pre-treatment and on-treatment tumor biopsy will be obtained for eligibility criteria. The primary endpoint is disease control rate. A Simon’s two-stage optimal design
translational studies. The primary endpoint is to estimate the objective response rate will be used to test the hypothesis: H0: P # 0.05 versus H1: P $ 0.25. In the first stage, 9 pts will be
(ORR) according to RECISTv1.1 criteria. Secondary endpoints include safety and toler- evaluated. The trial will be discontinued if no disease control response is observed in this stage. If at
ability as defined by the incidence of AEs as assessed according to CTCAE v5, disease least one response is observed, then the trial will continue to the second stage and an additional 17
control rate (DCR), progression-free survival (PFS), and overall survival (OS). Translational pts will be evaluated for a total of 26 evaluable. This design has a one-sided alpha of 0.05 and power
endpoints include the characterization of the transcriptional signatures in paired biopsies of 80%. We will reject the null hypothesis after 26 if 3 or more responses are observed. Clinical trial
information: awaited. Research Sponsor: Ontario Institute for Cancer Research and Princess
obtained before and on-treatment with BOT + AGEN1423, as well as the changes in cell Margaret Cancer Foundation; Terry Fox Research Institute - Marathon of Hope Cancer Centres
composition of the TME following treatment using multiplexed immunofluorescence Network Funding; Sinai Health Foundation; The MNitz Pancreatic Cancer Research Fund (Michelle
spatial technology. Enrolment has started and accrual is anticipated to complete in Q4 Reisman); The MNitz Pancreatic Cancer Research Fund (Veroli Cultural Society); The MNitz Pan-
2025. Clinical trial information: NCT05632328. Research Sponsor: Agenus, Inc. creatic Cancer Research Fund (Elyse Graff).

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GASTROINTESTINAL CANCER—GASTROESOPHAGEAL, PANCREATIC, AND HEPATOBILIARY 327s

TPS4233 Poster Session TPS4234 Poster Session

Phase I trial of MK2 inhibitor in combination with mFOLFIRINOX for un- An open-label, phase 1 trial with expansion cohort of botensilimab
treated metastatic pancreatic ductal adenocarcinoma. First Author: Moh’d M. (AGEN1181) + balstilimab (AGEN2034) + nab-paclitaxel + gemcitabine +
Khushman, Washington University School of Medicine, St. Louis, MO cisplatin + chloroquine + celecoxib in adult patients with previously un-
Background: Zunsemetinib (also known as ATI-450) is an investigational small molecule treated metastatic pancreatic cancer. First Author: Erkut H. Borazanci,
inhibitor targeting MAPK-Activated Protein Kinase (MAPKAPK2, or MK2). Preclinical work HonorHealth Research Institute, Scottsdale, AZ
conducted by the Lim Lab at Washington University in St. Louis demonstrated that Background: The unfolded protein response (UPR) is an adaptive endoplasmic retic-
FOLFIRINOX activates heat shock protein 27 (Hsp27), a molecule with pleiotropic pro- ulum (ER) stress pathway that and can prevent cellular death under moderate stress
survival properties, and beclin1, a key mediator of autophagy, in pancreatic ductal ade- conditions or promote apoptosis under severe stress. The UPR is often upregulated in
nocarcinoma (PDAC) models. In an autochthonous PDAC (KPC) mouse model, zunsemetinib pancreatic cancer cells and has been identified as a promising target for therapeutic
synergized with FOLFIRINOX, resulting in near-complete ablation of all PDAC foci and intervention. We hypothesize that inducing severe ER stress by combining multiple
significantly improved mouse survival. Additionally, mice treated with zunsemetinib ex- chemo and immunotherapy agents will cause pancreatic cancer cells to enter the
perienced significantly less intestinal damage and weight loss—common concerns asso- apoptotic UPR pathway, destroying these cells and improving patient survival. Pro-
ciated with FOLFIRINOX. These preclinical data support the rationale for combining longed ER stress is achieved in this study by using chemotherapy (nab-paclitaxel +
zunsemetinib with FOLFIRINOX in PDAC patients. Methods: We are conducting a phase I, gemcitabine + cisplatin) in combination with 2 immunotherapy agents: botensilimab
single-arm, open-label study of zunsemetinib in combination with mFOLFIRINOX in patients (AGEN1181), an Fc-engineered anti-CTLA-4 monoclonal antibody, and balstilimab
with untreated metastatic PDAC. The study consists of two phases: a dose escalation phase
(AGEN2034), a human monoclonal immunoglobulin (Ig) G4 (IgG4) antibody designed to
and an expansion phase. During the dose escalation phase, zunsemetinib dosing will
block programmed cell death protein (PD-1) binding by PD-L1 and PD-L2. Additionally, 2
proceed according to the BOIN design with a cohort size of 3. A total of 6–21 patients will be
enrolled at Washington University. Patients will receive zunsemetinib starting at Dose Level
agents are included to help block apoptosis escape routes: chloroquine to inhibit
1 (40 mg twice daily), with dose escalation continuing until the recommended phase 2 dose autophagy and celecoxib to reduce tumor microenvironment inflammation.
(RP2D) is determined. Patients will remain in the study until disease progression or Methods: This single-center, open-label, phase 1 study evaluates the safety, tolerability,
treatment intolerance. In the expansion phase, up to 30 additional patients will be enrolled to and preliminary efficacy of two botensilimab doses in combination with fixed doses of
further assess the toxicity profile of zunsemetinib in combination with mFOLFIRINOX. These balstilimab (240 mg) + nab-paclitaxel (125 mg/m2) + gemcitabine (1000 mg/m2) +
patients will begin at the RP2D and continue on study treatment until disease progression or cisplatin (25 mg/m2) + chloroquine (500 mg) + celecoxib (200 mg) in adult patients with
treatment intolerance. Eligible patients must be treatment-naı̈ve, newly diagnosed, and have previously untreated metastatic pancreatic cancer (NCT06076837). The study design
histologically or cytologically confirmed PDAC for which mFOLFIRINOX is deemed a suitable consists of 6 patients in a dose 1 cohort at 50 mg botensilimab + combination regimen
treatment option. The primary objective is to determine the dose-limiting toxicities (DLTs) and an escalated dose 2 cohort of 6 patients at 75 mg botensilimab + combination
and RP2D of zunsemetinib in combination with mFOLFIRINOX. Secondary objectives include regimen (pending dose 1 cohort safety signals), with an additional 6 patients in an
assessing toxicity profiles, progression-free survival (PFS) at six months and overall PFS, expansion cohort treated at the determined maximum tolerated dose (MTD) of
disease control rate, overall response rate, overall survival, CA 19-9 response at the RP2D, botensilimab (Total N = 18). Adverse events (AEs) are evaluated according to NCI CTCAE
and pharmacokinetics of zunsemetinib in PDAC treated with mFOLFIRINOX. Exploratory v5.0 and tumor response is assessed by RECIST v1.1. Key eligibility criteria include 1)
objectives include evaluating pharmacodynamic markers via immunohistochemistry (e.g., histologically confirmed diagnosis of metastatic pancreatic ductal adenocarcinoma with
phospho-Hsp27 to assess pharmacodynamics of zunsemetinib, LC3B to assess autophagy, measurable disease on baseline imaging, 2) life expectancy of at least 3 months, 3) no
and TUNEL staining to evaluate DNA damage) and analyzing pathway suppression through previous radiotherapy, surgery, chemotherapy, or investigational therapy for the
RNA sequencing. Pre- and post-treatment serum samples will also be collected for the treatment of metastatic disease, and 4) no prior immune checkpoint inhibitor therapy.
analysis of inflammatory cytokines. Clinical Trial Registration: NCT06648434. Clinical trial Enrollment began in January 2025 at the HonorHealth Research Institute. Clinical trial
information: NCT06648434. Research Sponsor: Pancreatic SPORE (P50 CA272213); Aclaris. information: NCT06076837. Research Sponsor: TGen Foundation; Purple Pansies
Foundation.

TPS4235 Poster Session TPS4236 Poster Session

5
A supervised prehabilitation program for patients with pancreatic cancer. IMMUNORARE : A national platform of 5 academic phase II trials coordi-
First Author: Philip Chang, Cedars-Sinai Medical Center, Los Angeles, CA nated by Lyon University Hospital to assess the safety and the efficacy of
Background: Individuals who develop pancreatic cancer tend to be older, with 70% of the immunotherapy with domvanalimab + zimberelimab combination in
pancreatic diagnoses occurring in those $ 65 years (2). Older patients are at increased patients with advanced rare cancers—The Peritoneal Mesotheliomas
risk for sarcopenia which is the progressive loss of skeletal muscle mass, tone, quality Cohort. First Author: Julien Peron Sr., Centre Hospitalier Lyon Sud - Hospices Civils
and strength and has been reported to affect 65% of pancreatic cancer patients (4). de Lyon, Pierre-Bénite, France
PREHAB is the process of improving the functional capability and psychological health Background: In patients with rare cancers, there is an unmet medical need for in-
of the individual to reduce the incidence and/or severity of future impairments (6). The vestigating innovative therapeutics beyond standard first-line treatment. Indeed, these
foundation of PREHAB is functional exercise although components of nutrition and diseases are rarely assessed in clinical trials. The standard 1st-line treatment of
stress reduction may be included (9). PREHAB sessions are typically delivered through peritoneal mesothelioma relies on platinum and pemetrexed, with no validated 2nd-line
structured programs and have been shown to have a number of benefits such as treatment so far. Several studies suggested that immunotherapy, such as ipilimumab +
improvements in functional activity and decreased postoperative complications (8). In a nivolumab or atezolizumab + bevacizumab, is active in this disease. There is a strong
study by Ngo-Huang et al, 50 pancreatic cancer participants participated in a home- biological rationale for concurrent blockage of TIGIT and PD1 pathways in mesothe-
based multimodal program resulting in improved physical function and health related lioma. Methods: IMMUNORARE5 (NCT06790706) is a platform of 5 single arm phase II
quality of life (15). Given the numerous benefits, the purpose of this study is to trials testing the safety and the efficacy of DOMVANALIMAB (anti-TIGIT) and ZIM-
demonstrate the feasibility of a multimodal supervised PREHAB program in pancreatic BERELIMAB (anti PD-1) in 5 independent cohorts of rare cancers. The trial, sponsored by
cancer patients which we believe could have greater benefits than unsupervised Lyon University Hospital, is conducted in 15 French centers, in partnership with the
programs. Methods: This is a single arm pilot study assessing the feasibility of a corresponding national networks of reference centers. The PERITONEAL MESOTHE-
supervised prehabilitation program for patients with pancreatic cancer. Inclusion criteria LIOMA cohort, led in collaboration with the RENAPE network ([Link]
include a diagnosis of any stage pancreatic cancer, independence with ambulation, and a [Link]/), will enroll 27 patients in progression after at least 1 line of platinum +
lower level of physical activity as assessed by the Godin-Shepard Leisure-Time Physical pemetrexed based-chemotherapy regimen, with evaluable lesions at the baseline
Activity Questionnaire. To our knowledge, this is the first study in which all exercises (modified RECIST criteria). Patients previously treated with immunotherapy will not be
sessions are in-person and supervised by exercise technicians in pancreatic cancer. eligible. Patients will receive intra-venous DOMVANALIMAB and ZIMBERELIMAB, every
Additionally, while prehabilitation typically takes place during the neoadjuvant therapy three weeks, until disease progression. The primary endpoint will be the progression-free
period, this study will also include patients with metastatic disease on continuous survival rate at 6 months. The disease progression (clinical or radiological) will be
chemotherapy. Participants will undergo baseline evaluations testing strength, en- confirmed by the RENAPE experts. The secondary endpoints are tolerance, overall
durance, balance, subjective measures and sarcopenia measures. This will be imme- response rate and duration of response, progression-free and overall survival. A two-
diately followed by one-hour long supervised exercise sessions 3x per week for 6 weeks stage Simon design was used, with early termination rules for futility (5% one-sided
in which participants will engage in aerobic training and resistance training targeting alpha level, 80% power). The treatment would be considered interesting if the percentage
major muscle groups. Following the intevention, measures will be collected immediately of patients free from disease progression at 6-months is statistically higher than 35%;
afterwards and at 3 month follow-up. The primary analysis will test the hypothesis of 60% is expected. Translational research projects will be developed aiming at deciphering
feasibility using an one-sided exact Binomial test at 25% significance level. If 10 or more cellular and molecular mechanisms involved in response to treatment. Moreover, data
patients attend a minimum of 60% of exercise sessions during the initial 6-week period, from the prospectively-maintained RENAPE database will be investigated to build a
then the study will be declared feasible. 11 of 16 patients have been enrolled to date. synthetic historical arm representative of the efficacy of the standard treatments in a
Clinical trial information: NCT05692323. Research Sponsor: Cedars Sinai Medical Center similar population of patients. Clinical trial information: NCT06790706. Research
(Internal Funding). Sponsor: GILEAD.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
328s GENITOURINARY CANCER—KIDNEY AND BLADDER

4500 Oral Abstract Session 4501 Oral Abstract Session

Nivolumab plus ipilimumab (NIVO+IPI) vs gemcitabine-carboplatin (gem- Avelumab + sacituzumab govitecan (SG) vs avelumab monotherapy as first-
carbo) chemotherapy for previously untreated unresectable or metastatic line (1L) maintenance treatment in patients (pts) with advanced urothelial
urothelial carcinoma (mUC): Final results for cisplatin-ineligible patients carcinoma (aUC): Interim analysis from the JAVELIN Bladder Medley phase
from the CheckMate 901 trial. First Author: Michiel Simon Van Der Heijden, 2 trial. First Author: Jeannie Hoffman-Censits, The Sidney Kimmel Comprehensive
Netherlands Cancer Institute, Amsterdam, Netherlands Cancer Center at Johns Hopkins, Johns Hopkins Greenberg Bladder Cancer Institute,
Background: Platinum-based chemotherapy is a standard of care (SOC) for unresectable or mUC; Baltimore, MD
patients (pts) ineligible for cisplatin (cis) have worse outcomes. The phase 3, global, open-label, Background: In the JAVELIN Bladder 100 phase 3 trial, avelumab 1L maintenance + best
randomized CheckMate 901 trial (NCT03036098) compared NIVO+IPI vs gem-carbo in cis-ineligible supportive care (BSC) significantly prolonged overall survival (OS) and progression-free survival
pts with previously untreated unresectable or mUC. Here, we report final results. Methods: Pts with (PFS) vs BSC alone in pts with aUC without progression following 1L platinum-based chemo-
previously untreated, histologically confirmed, unresectable or mUC who were cis-ineligible (glo- therapy (PBC). The JAVELIN Bladder Medley phase 2 trial is investigating the combination of
merular filtration rate $ 30 to , 60 mL/min) were randomized 1:1 to NIVO 1 mg/kg + IPI 3 mg/kg Q3W avelumab with other antitumor agents in this pt population to assess efficacy and safety vs
up to 4 cycles, then NIVO 480 mg Q4W until disease progression/unacceptable toxicity or up to 2 years, avelumab maintenance monotherapy. SG is a Trop-2–directed antibody and topoisomerase
or to gem-carbo Q3W for up to 6 cycles. Pts were stratified by tumor PD-L1 expression and liver inhibitor drug conjugate being investigated in solid tumors. Here we report an interim analysis of
metastasis. The primary endpoint was overall survival (OS). Progression-free survival (PFS) by blinded
avelumab + SG vs avelumab monotherapy. Methods: Eligible pts had unresectable locally ad-
independent central review (BICR) was a secondary endpoint. Objective response rate (ORR) per BICR,
vanced or metastatic UC, ECOG performance status (PS) 0-1, and no disease progression after 4-6
duration of response (DOR) per BICR, and safety were exploratory. Results: 445 pts were randomized
cycles of 1L PBC. Pts were randomized 2:1 to avelumab + SG or avelumab monotherapy, stratified
(NIVO+IPI, n = 221; gem-carbo, n = 224). Median time to treatment discontinuation (95% CI) was 2.2
(2.1–3.5) mo with NIVO+IPI vs 3.8 (3.5–3.9) mo with gem-carbo. After minimum follow-up (58.3 mo), by presence of visceral metastases at start of 1L PBC. Primary endpoints were investigator-
the primary endpoint of OS did not meet the threshold for significance (median, 19.1 mo with NIVO+IPI assessed PFS and safety; OS was a secondary endpoint. For PFS and OS analyses, data in the
vs 13.2 mo with gem-carbo; HR 0.79 [98.27% CI, 0.61–1.01]; P = 0.0245; Table). PFS, ORR, and DOR are avelumab monotherapy arm were extended per protocol using propensity score-weighted
shown in the Table. Any-grade treatment-related adverse events (TRAEs) occurred in 89.0% (grade JAVELIN Bladder 100 data. Results: At data cutoff (Sep 16, 2024), 38 of 74 pts (51.4%) in
3–4, 47.2%) of NIVO+IPI-treated and 92.9% (grade 3–4, 76.3%) of gem-carbo-treated pts; any-grade the avelumab + SG arm and 10 of 37 pts (27.0%) in the avelumab monotherapy arm were still
TRAEs leading to discontinuation occurred in 31.2% and 14.2% of pts, respectively. There were 8 receiving study treatment. In the avelumab + SG and avelumab monotherapy arms, respectively,
deaths related to toxicity (NIVO+IPI, 7; gem-carbo, 1). Conclusions: NIVO+IPI did not meet the median age was 70 and 67 years, 50.0% and 51.4% had visceral metastases at start of 1L PBC,
threshold of statistical significance for improved OS vs gem-carbo in cis-ineligible pts with untreated and a lower proportion of pts in the avelumab + SG arm had ECOG PS of 1 (31.1% vs 54.1%).
unresectable or mUC. Durable response and favorable landmark OS with NIVO+IPI show meaningful Median PFS was 11.17 months (95% CI, 7.43-not estimable [NE]) with avelumab + SG vs
activity from a chemotherapy-free regimen of finite duration. No new safety signals were identified. 3.75 months (95% CI, 3.32-6.77) with avelumab monotherapy (hazard ratio [HR], 0.49 [95% CI,
Clinical trial information: NCT03036098. Research Sponsor: Bristol Myers Squibb. 0.31-0.76]). OS data were immature at cutoff; median OS was not reached (95% CI, 15.51-NE) in
the avelumab + SG arm vs 23.75 months (95% CI, 18.79-30.82) in the avelumab monotherapy arm
NIVO+IPI; Gem-carbo; (HR, 0.79 [95% CI, 0.42-1.50]). In the avelumab + SG and avelumab monotherapy arms, re-
Efficacy (95% CI) n = 221 n = 224 HR
spectively, treatment-related adverse events (TRAEs) of any grade occurred in 71 (97.3%) vs 23
mOS 19.1 (13.5–22.6) 13.2 (11.6–15.2) 0.79 (98.27% CI, 0.61–1.01)a; pts (63.9%), and were grade $3 in 51 (69.9%) vs 0 pts. TRAEs led to discontinuation of both
P = 0.0245b avelumab + SG in 4.1% (SG only in 12.3%) vs avelumab monotherapy in 2.8%. One pt in the
12-mo OS rate 59.7 (52.8–65.9) 54.3 (47.4–60.7) –
36-mo OS rate 29.6 (23.5–35.9) 19.3 (14.3–24.9) – avelumab + SG arm had a SG-related AE that led to death (acute subarachnoid hemorrhage in the
mPFS 5.3 (3.8–6.0) 5.9 (5.6–7.6) 0.90 (95% CI, 0.72–1.12) setting of sepsis and pancytopenia). Conclusions: In pts with aUC without progression after 1L
12-mo PFS rate 31.5 (25.0–38.2) 17.2 (11.8–23.4) – PBC, avelumab + SG as maintenance treatment improved PFS vs avelumab monotherapy. TRAEs
36-mo PFS rate 20.0 (14.3–26.5) 4.9 (2.1–9.6) – were more frequent in the combination arm and were consistent with the known safety profile of
ORR 35.3 (29.0–42.0) 38.8 (32.4–45.6) - SG and avelumab. Further investigation of avelumab in combination with anti–Trop-2 antibody-
mDOR 25.0 (14.8–61.8); n = 78 7.4 (5.8–8.5); n = 87 -
drug conjugates in aUC is warranted. Clinical trial information: NCT05327530. Research Sponsor:
a
Statistical inference based on adjusted CI: 95% CI, 0.64–0.97. the healthcare business of Merck KGaA, Darmstadt, Germany (CrossRef Funder ID: 10.13039/
b
mOS did not reach the prespecified threshold of statistical significance of P = 0.0173. 100009945).
m, median.

4502 Oral Abstract Session 4503 Oral Abstract Session

Exploratory analysis of responders from the phase 3 EV-302 trial of enfor- Circulating tumor DNA (ctDNA) in patients with muscle-invasive bladder
tumab vedotin plus pembrolizumab (EV+P) vs chemotherapy (chemo) in cancer (MIBC) who received perioperative durvalumab (D) in NIAGARA. First
previously untreated locally advanced or metastatic urothelial carcinoma Author: Thomas Powles, Barts Cancer Institute, Experimental Cancer Medicine Centre,
(la/mUC). First Author: Shilpa Gupta, Cleveland Clinic Taussig Cancer Institute, Queen Mary University of London, St Bartholomew’s Hospital, London, United Kingdom
Cleveland, OH Background: In the phase 3 NIAGARA trial (NCT03732677) of patients (pts) with cisplatin-eligible MIBC,
Background: EV-302/KEYNOTE-A39 (NCT04223856) showed significant PFS and OS benefits for pts with addition of perioperative D to neoadjuvant chemotherapy (NAC) demonstrated a statistically significant
previously untreated la/mUC treated with EV+P vs chemo, which established EV+P as the SOC in this and clinically meaningful improvement in event-free survival (EFS) and overall survival compared with
population. This exploratory analysis presents efficacy and safety results for responders, focusing on pts with NAC alone, and a 10% higher pathological complete response (pCR) rate, with a manageable safety profile
confirmed complete response (cCR). Methods: Pts were randomized 1:1 to receive EV (1.25 mg/kg; Days 1 and and no impact on the feasibility of surgery. Here, we report a planned exploratory analysis of ctDNA and
8; IV) + P (200 mg; Day 1; IV) or chemo (gemcitabine + cisplatin/carboplatin); all Q3W. Primary endpoints were association with clinical outcomes from NIAGARA. Methods: NIAGARA enrolled cisplatin-eligible pts
PFS by BICR and OS. Secondary endpoints included ORR, DOR, and safety. An exploratory analysis evaluated with MIBC (cT2-T4aN0/1M0) planned for radical cystectomy (RC). Pts were randomized 1:1 to receive
outcomes in pts with cCR. A genAI tool (01/09/25; Pfizer; GPT-4o) developed the 1st draft; authors assume either neoadjuvant D (1500 mg IV Q3W) and NAC (cisplatin + gemcitabine IV Q3W) for 4 cycles followed by
content responsibility. Results: Median follow-up (data cutoff: Aug 8, 2024) was 29.1 mo (95% CI, 28.5-29.9). RC, then adjuvant D monotherapy (1500 mg IV Q4W) for 8 cycles (D arm), or NAC followed by RC alone
886 pts were randomized to EV+P (n = 442) vs chemo (n = 444). Confirmed ORR (CR+PR) was 67.5% with EV+P (comparator [C] arm). Dual primary endpoints were pCR and EFS. Disease-free survival (DFS) was a
and 44.2% with chemo; cCR was 30.4% and 14.5%, respectively. Baseline characteristics of responders were secondary endpoint. Plasma ctDNA was assessed using the Signatera personalized, tumor-informed
generally consistent with the ITT population. Among pts with cCR in the EV+P arm, 38 (28.6%) had upper tract molecular residual disease (MRD) assay (Natera, Inc, Austin, TX, USA). ctDNA was assessed at baseline
disease and 20 (15%) had liver metastases. For pts with cCR, mPFS by BICR was not reached (NR) with EV+P (screening or neoadjuvant C1D1, n = 460), after neoadjuvant treatment prior to RC (pre-RC, n = 422), and at
and 26.9 mo with chemo (HR, 0.36; 95% CI, 0.21-0.61); mOS was NR with both EV+P and chemo (HR, 0.37; 95% C1D1 of the adjuvant phase (post-RC, n = 345). Results: Of 1063 randomized pts, 462 comprised the
CI, 0.17-0.80). Median duration of CR was NR for EV+P and 15.2 mo for chemo. Efficacy data are in the Table.
biomarker-evaluable population (237 D arm; 225 C arm). Patient characteristics were similar to the ITT
Among pts with cCR in the EV+P arm, the median number of cycles was 13 (range, 1-50) for EV and 27 (range, 1-
population. Overall, the ctDNA+ rate at baseline was 57% (260/460) and decreased to 22% (94/422) after
35) for P; median treatment (tx) duration was 22.0 mo (range, 0.7-35.4). Safety among responders was
neoadjuvant treatment at pre-RC. ctDNA clearance rates from baseline to pre-RC were 41% in the D arm
generally consistent with previous reports. TRAEs leading to dose modification in pts with cCR are in the Table.
Grade $3 TRAEs occurred in 61.7% and 71.9% of pts with cCR in the EV+P and chemo arms, respectively. EV
and 31% in the C arm. The non-pCR rate was 97% (86/89) among pts with pre-RC ctDNA+ status. Overall
tx-related AESIs and P tx-emergent AEOSI profiles were generally consistent with previous reports. There were ctDNA+ rate post-RC was 9% (31/345). EFS benefit in the D arm vs the C arm was observed in both the
no tx-related deaths among pts with cCR. Conclusions: In the EV+P arm, the proportion of pts achieving cCR baseline ctDNA+ and ctDNA2 groups (Table). DFS benefit with perioperative D was observed in post-RC
was twice that in the chemo arm. Consistent with the ITT data, EV+P reduced the risk of progression or death vs ctDNA+ and ctDNA2 groups (Table). Conclusions: In this exploratory analysis, ctDNA+ status at pre-RC
chemo in pts achieving cCR, with appropriate dose modifications. These data reinforce EV+P as the SOC for 1L was associated with non-pCR. Higher ctDNA clearance from baseline to pre-RC in the D arm indicated the
tx of pts with la/mUC. Clinical trial information: NCT04223856. Research Sponsor: The EV-302 study was additional benefit of D plus NAC vs NAC alone. Perioperative D provided an EFS benefit to both pts with
funded by Astellas Pharma Inc., Northbrook, IL, USA; Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., ctDNA+ and ctDNA2 status at baseline; a similar trend was observed with DFS based on ctDNA status
Inc., Rahway, NJ, USA; and Seagen Inc., Bothell, WA, USA, which was acquired by Pfizer in December 2023. post-RC. These results further support the perioperative D regimen for pts with MIBC. Funding:
AstraZeneca. Clinical trial information: NCT03732677. Research Sponsor: AstraZeneca.
EV+P Chemo
cCR: n=133 cCR: n=64 EFS DFS

Median PFS, mo (95% CI) NR (NE-NE) 26.9 (16.6-NE) Baseline Baseline Post-RC Post-RC
24-mo PFS rate, % (95% CI) 78.2 (69.8-84.6) 53.7 (40.0-65.5) ctDNA+ ctDNA- ctDNA+ ctDNA-
Median OS, mo (95% CI) NR (39.3-NE) NR (32.1-NE) D C D C D C D C
24-mo OS rate, % (95% CI) 95.4 (90.0-97.9) 85.8 (74.6-92.4)
Median DOCR, mo (95% CI) NR (NE-NE) 15.2 (10.3-NE) n 137 123 99 101 9 8 129 126
24-mo cCR rate, % (95% CI) 74.3 (65.1-81.4) 43.2 (28.7-56.9) Median (95% CI), NR 32.3 NR NR 9.5 6.2 NR NR
months (NR–NR) (24.3–NR) (NR–NR) (NR–NR) (2.8–NR) (2.9–NR) (NR–NR) (NR–NR)
TRAEs leading to dose modification (n, %)
Hazard ratio 0.73 0.45 NC* 0.49
Dose interruption of EV 94 (70.7) - (95% CI) (0.51–1.06) (0.25–0.84) (0.28–0.84)
Dose reduction of EV 86 (64.7) -
Dose interruption of P 86 (64.7) - CI, confidence interval; NC, not calculable; NR, not reached.
*NC due to ,20 events between arms.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GENITOURINARY CANCER—KIDNEY AND BLADDER 329s

LBA4504 Oral Abstract Session 4505 Oral Abstract Session

Mitomycin plus BCG as adjuvant intravesical therapy for high-risk, non– Nivolumab plus ipilimumab vs sunitinib for first-line treatment of advanced
muscle-invasive bladder cancer: A randomized phase 3 trial (ANZUP 1301). renal cell carcinoma: Final analysis from the phase 3 CheckMate 214 trial.
First Author: Dickon Hayne, UWA Medical School, University of Western Australia, Perth, First Author: Toni K. Choueiri, Lank Center for Genitourinary Oncology, Dana-Farber
Western Australia, Australia Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston,
MA
Background: First-line nivolumab plus ipilimumab (NIVO+IPI) provided substantial long-term survival benefits
over sunitinib (SUN) in patients (pts) with advanced renal cell carcinoma (aRCC) in the CheckMate 214 trial. We
now report final efficacy and safety data in the intent-to-treat (ITT) population and by International Metastatic
Renal Cell Carcinoma Database Consortium (IMDC) risk. Methods: Pts with clear cell aRCC were randomized 1:
1 to NIVO 3 mg/kg + IPI 1 mg/kg Q3W34 then NIVO (3 mg/kg or 240 mg Q2W or 480 mg Q4W); or SUN 50 mg
once daily for 4 weeks on, 2 weeks off. Efficacy endpoints included overall survival (OS), and independent
radiology review committee (IRRC)-assessed progression-free survival (PFS) and objective response rate
(ORR) in intermediate/poor-risk (I/P; primary), ITT (secondary), and favorable-risk (FAV; exploratory) pts.
Response was assessed using RECIST v1.1. Results: With 9 years median follow-up, OS was improved with
NIVO+IPI vs SUN in ITT (HR 0.71) and I/P (HR 0.69) pts. The probability of OS at 108 months was 31% vs 20% in
ITT pts and 30% vs 19% in I/P pts, respectively. In pts with FAV risk, the HR for OS improved from 1.45 at first
report (Motzer NEJM 2018) to 0.80 at 9 years, showing a delayed benefit with NIVO+IPI vs SUN. OS probabilities
at 108 months were 35% vs 22% in FAV pts, respectively (Table). The probability of PFS at 96 months with
NIVO+IPI vs SUN was 23% vs 9% in ITT pts, 25% vs 9% in I/P pts, and 13% vs 11% in FAV pts. The probability of
The full, final text of this abstract will be available at remaining in response through 96 months with NIVO+IPI vs SUN was 48% vs 19% in ITT pts, 50% vs 23% in I/P
[Link] on the day of presentation and in the pts, and 36% vs not available (NA) in FAV pts. No new treatment-related deaths occurred in either arm.
Additional subgroup analyses will be presented. Conclusions: In the longest and final phase 3 follow-up (9
online supplement to the June 10, 2025, issue of the Journal years) of a first-line checkpoint inhibitor combination in aRCC, milestone rates of OS and PFS and durable
of Clinical Oncology. response remained higher with NIVO+IPI vs SUN. No new safety signals emerged. NIVO+IPI remains a standard
first-line option in aRCC. Clinical trial information: NCT02231749. Research Sponsor: Bristol Myers Squibb.

ITT I/P FAV

Arm; n NIVO+IPI; 550 SUN; 546 NIVO+IPI; 425 SUN; 422 NIVO+IPI; 125 SUN; 124
mOS (95% CI), mo 53 (46–64) 38 (32–44) 47 (35–56) 26 (22–33) 78 (65–92) 67 (56–80)
108-mo OS probabilities 31 (27–35) 20 (16–23) 30 (26–35) 19 (15–23) 35 (27–44) 22 (15–30)
(95% CI), %
mPFS (95% CI), mo 12 (10–16) 12 (10–15) 12 (9–17) 9 (7–11) 13 (10–18) 29 (23–43)
96-moa PFS 23 (18–27) 9 (5–15) 25 (20–31) 9 (4–15) 13 (6–22) 11 (3–27)
probabilities (95% CI),
%
ORR per IRRC 39 (35–44); 12 33 (29–37); 3 42 (38–47); 12 27 (23–32); 3 30 (22–38); 13 52 (43–61); 6
(95% CI); CR, %
mDOR (95% CI), mo 76 (59–NE) 25 (20–33) 83 (54–NE) 20 (16–26) 61 (23–NE) 33 (25–51)
96-moa DOR 48 (39–55) 19 (10–31) 50 (41–58) 23 (13–36) 36 (17–56) NAb
probabilities
(95% CI), %
a
96-mo probabilities reported due to small numbers of pts at risk at 108 mo.
b
No pts remain at risk.
CR, complete response; DOR, duration of response; m, median; NE, not estimable.

4506 Oral Abstract Session 4507 Oral Abstract Session

Combination casdatifan plus cabozantinib expansion cohort of phase 1 Hypoxia-inducible factor-2a (HIF-2a) inhibitor belzutifan in von Hippel-
ARC-20 study in previously treated patients with clear cell renal cell Lindau (VHL) disease–associated neoplasms: 5-year follow-up of the phase
carcinoma. First Author: Toni K. Choueiri, Department of Medical Oncology, Dana- 2 LITESPARK-004 study. First Author: Vivek Narayan, Hospital of the University of
Farber Cancer Institute, Boston, MA Pennsylvania, Philadelphia, PA
Background: Hypoxia-inducible factor 2-alpha (HIF-2a) is highly dysregulated in clear Background: The HIF-2a inhibitor belzutifan is approved for the treatment of patients with VHL
cell renal cell carcinoma (ccRCC), resulting in increased expression of proteins involved disease–associated renal cell carcinoma (RCC), CNS hemangioblastomas (HB), or pancreatic neuro-
endocrine tumors (pNETs), not requiring immediate surgery based on previously reported results from
with angiogenesis, proliferation, and cancer cell survival. Casdatifan is an orally bio-
the ongoing open-label phase 2 LITESPARK-004 study (NCT03401788). Updated results are presented
available small-molecule HIF-2a inhibitor. We investigated the safety and efficacy of after a minimum of 5 years of follow-up. Methods: Adults with germline VHL alterations, $1 mea-
casdatifan in combination with the anti–vascular endothelial growth factor receptor surable nonmetastatic RCC tumor, no tumor . 3 cm that required immediate surgery, no metastatic
tyrosine kinase inhibitor (VEGFR-TKI) cabozantinib in previously treated patients with disease, no prior anticancer systemic treatment, and an ECOG PS of 0 or 1 received oral belzutifan
ccRCC in an expansion cohort (casdatifan + cabozantinib) of the phase 1, open-label 120 mg once daily until disease progression, unacceptable toxicity, or participant (pt) withdrawal. The
ARC-20 (NCT05536141) trial. Methods: Patients enrolled in the casdatifan + cabo- primary end point was objective response rate (ORR) in VHL disease–associated RCC per RECIST v1.1
zantinib expansion cohort were previously treated with immunotherapy (IO) alone or with by independent review committee (IRC). Secondary end points included safety, ORR in non-RCC
neoplasms, duration of response (DOR), and progression-free survival (PFS) per RECIST v1.1 by IRC.
anti-VEGF therapies. Casdatifan 100 mg and cabozantinib 60 mg were given orally once
Results: Overall, 61 pts received $1 dose of belzutifan. Median study follow-up was 61.8 mo (range,
daily. Endpoints included the incidence of treatment-emergent adverse events (AEs) and 60.2-70.1). As of the April 1, 2024 data cutoff date, 35 pts (57%) remained on treatment. ORR was 70%
objective response rate (ORR) by RECIST v1.1. This study is ongoing; data as of January for RCC, 50% for CNS HB, and 90% for pNETs. Additional efficacy results are in the Table. Among 14 pts
3, 2025, are reported. Results: Overall, 27 patients with a median (range) follow-up of (n = 18 eyes) with retinal HB, 100% (95% CI, 82-100) of eyes showed improvement per ophthalmologic
2.9 (0.1–6.8) months were enrolled. At data cut off, prior treatment settings included assessment; median DOR for retinal HBs was not reached (NR; range, 8.5-61.0+ mo). At baseline, 59 of
adjuvant only (n = 5/26) and metastatic (1L n = 17/26; 2L n = 4/26). Prior therapies 61 pts (97%) had $1 prior VHL-related surgery. Within the 5 years before starting belzutifan, 46 of 61
included IO only (n = 15/26) or IO plus VEGFR-TKI (n = 11/26). All grade AEs occurred in pts (75%) had $1 surgery. Since starting belzutifan, 19 of 61 pts (31%) underwent VHL-related
surgeries; 4 underwent surgery while on treatment and subsequently discontinued treatment, 8 un-
89% of patients with the most common being anemia (n = 16 [59%]) and fatigue (n = 15 derwent surgery after discontinuing treatment, and 7 are continuing treatment as of the data cutoff
[56%]). Most common ( . 10%) grade $3 AEs were anemia (n = 7 [26%]) and hypoxia (n = date. Grade 3 treatment-related adverse events (TRAEs) (most commonly anemia [n = 7; 11%]) were
3 [11%]). No cardiac events were reported. AEs leading to casdatifan-only, cabozantinib- reported in 11 pts (18%). No grade 4 or 5 TRAEs occurred. Belzutifan was discontinued in 2 pts (3%) due
only, or both casdatifan + cabozantinib dose reductions occurred in 3 (11%), 7 (26%), and to TRAEs (grade 1 dizziness and grade 2 intracranial hemorrhage). Conclusions: After 5 years of follow-
2 (7%) patients, respectively. Only one (4%) pt discontinued due to an AE, hypoxia related up, belzutifan continues to demonstrate durable antitumor activity and a manageable safety profile,
to casdatifan. Responses continue to be observed among the efficacy evaluable consistent with prior reports. Most pts remain on treatment after this period. Results continue to
population (n = 22; as of January 27, 2025) with ORR of 41% (n = 1 complete response; n support the use of belzutifan in pts with VHL disease–related RCC, CNS HB, and pNETs who do not
require immediate surgery. Clinical trial information: NCT03401788. Research Sponsor: Merck Sharp &
= 8 partial response). Activity was seen across all IMDC risk groups. Conclusions: In Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA; The Intramural Research Program of the
previously treated patients with ccRCC, casdatifan 100 mg in combination with National Institutes of Health, National Cancer Institute Center for Cancer Research, and a grant (UO1
cabozantinib 60 mg had a manageable AE profile with promising clinical activity. These CA236489) from the National Cancer Institute.
data support continued evaluation of this combination in the phase 3 PEAK-1 clinical
RCC CNS HB pNETs
trial. Clinical trial information: NCT05536141. Research Sponsor: Arcus Biosciences, Inc. n = 61 n = 50 n = 20
ORR, % (95% CI) 70 (57-82);7 CRs, 36 PRs 50 (36-64); 6 CRs; 19 PRs 90 (68-99); 13 CRs, 5 PRs
DOR, median (range), mo NR (5.8+ to 60.8+) 60.3 (0.0+ to 60.3) NR (11.0+ to 59.6+)
48-mo DOR rate 76% 82% 94%
PFS, median (95% CI), mo NR (NR-NR) 63.5 (63.5-NR) NR (NR-NR)
48-mo PFS rate 81% 79% 96%

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
330s GENITOURINARY CANCER—KIDNEY AND BLADDER

4508 Oral Abstract Session 4509 Clinical Science Symposium

ALLO-316 in advanced clear cell renal cell carcinoma (ccRCC): Updated Exploratory analysis from NEOAVAX, a neoadjuvant trial of avelumab/
results from the phase 1 TRAVERSE study. First Author: Samer Ali Srour, axitinib in patients (pts) with localized renal cell carcinoma (RCC) who
Department of Stem Cell Transplantation and Cellular Therapy, The University of Texas are at high risk of relapse after nephrectomy. First Author: Axel Bex,
MD Anderson Cancer Center, Houston, TX Netherlands Cancer Institute, Amsterdam, The Netherlands, Netherlands
Background: Treatment options are limited for ccRCC after disease progression on immune checkpoint Background: NEOAVAX (NCT03341845) is an open label, single arm, phase II trial, investigating
inhibitors (ICIs) and vascular endothelial growth factor (VEGF) inhibitors. CD70 is highly expressed on ccRCC. 12 weeks of neoadjuvant avelumab/axitinib prior to nephrectomy in 40 pts with high-risk non-
ALLO-316 is an investigational, healthy donor–derived allogeneic CD70 CAR T-cell product designed to metastatic clear-cell (cc) RCC (cT1b-4 cN0-1 M0, Grades 3-4). Dynamic on-treatment increase of CD8+
recognize and kill both CD70 positive tumor cells and CD70 positive host T cells that drive allorejection. Initial tumor infiltrating lymphocytes (TILs) in the tumour microenvironment (TME) and the primary endpoint
data from the multicenter, phase 1a/b TRAVERSE study (NCT04696731) showed that ALLO-316 had (radiographic partial response rate (RECIST 1.1) in the primary tumour (PT) in $25%) were reported
manageable safety and promising antitumor activity. Updated results are presented. Methods: Patients previously together with safety and tolerability [1]. Methods: Exploratory analysis included pathologic
were aged $18 years, had advanced ccRCC, ECOG PS of 0 or 1, and disease progression after ICI and VEGF- response in the PT according to the International Neoadjuvant Melanoma Consortium (INMC) and
targeted therapy. After lymphodepletion (LD) with fludarabine and cyclophosphamide 6 ALLO-647 (anti-
multiplex immune histochemistry (IHC) of the TME including CD8+, CD8-granzyme-B+, CD8+CD39+,
CD52), patients received a single infusion of ALLO-316 following a 3+3 design (40-240 3 106 allogeneic CAR+
Foxp3+ cells and MHC-I in paired samples (pre-treatment biopsy and nephrectomy) from 40 pts;. IHC
T cells). Primary end points were incidence of dose-limiting toxicities and adverse events. Objective response
rate (ORR) was a secondary end point. Results: Of 44 patients who underwent LD, 39 received ALLO-316, and
data were compared to RECIST 1.1 and pathologic response in the PT, and recurrence; Visium spatial
38 were evaluable for disease outcome. Median age was 60 years, median of 3 prior therapies (range, 1-8), transcriptomics was performed on 18 PT from pts with diverging clinical outcome. Results: The
and 36 (82%) had CD70 positive ccRCC. As of January 2, 2025, median follow-up was 6.8 months (range, 0.8- majority of pts (n=25 (62.5%) had no pathologic response (pNR) by INMC criteria. Twelve patients
39.5). Dose-limiting toxicities occurred in 2 patients (autoimmune hepatitis and cardiogenic shock in the (30%) had a partial (pPR) and 3 (7.5%) a major pathological response (MPR). There was no association
setting of multiorgan failure). Treatment-emergent adverse events occurred in 42 patients (96%; grade $3, between pathological and radiographic response of the PT. Recurrence occurred in 1 of 3 pts (33%)
37 patients [84%]). Grade $3 CRS occurred in 1 patient (2%; any grade, 25 patients [57%]), grade $3 ICANS in with MPR at 36 mo, in 7 of 12 (58%) with a pPR at a median of 12 mo and in 14 of 25 (56%) with pNR
0 patients (any grade, 4 patients [9%]), and grade $3 IEC-HS in 1 patient (2%; any grade, 8 patients [18%]). No at a median of 3 mo. Of 25 pts with pNR 7 died of disease (DoD; 28%). On IHC, intratumoural
GvHD occurred. As previously reported, 3 grade 5 adverse events were related to ALLO-316 (cardiogenic CD8+CD39+ on post-treatment PT samples was significantly associated with recurrence (p,0.0001).
shock, failure to thrive, and sepsis). ORR for all LD regimens was 20% (6/30) overall for patients with CD70 MPR associated with spatial co-localisation of tumour cells with tissue-resident macrophages, CD8+
positive tumors (Table). Confirmed ORR was 33% (3/9) for patients with CD70 $50% treated with the phase cytotoxic T-cells, memory T-cells and B-cells. Gene Set Enrichment Analysis (GSEA) results for
1b regimen; all confirmed responses were ongoing (2.1, 6.7, and 8.4 months at the data cut-off). Reactome pathways in each Visium tumor spot cluster demonstrated intratumoural heterogeneity in
Conclusions: After a median follow-up of 6.8 months, a single infusion of ALLO-316 had manageable post-treatment PT in select patients. Conclusions: Pathologic response and IHC post-treatment influx
safety and encouraging antitumor activity in heavily pretreated patients. Further evaluation of ALLO-316 in of CD8+CD39+ TILs associates with prolonged disease-free survival following neoadjuvant avelumab/
CD70 positive ccRCC is warranted. Clinical trial information: NCT04696731. Research Sponsor: The anti- axitinib. Particularly, pts with MPR had distinct spatial co-localisation gene signatures of tumour and
CD70 AlloCAR T program utilizes Cellectis technology and is licensed exclusively from Cellectis. Allogene immune cells in the TME. Despite 3 months of treatment, 62.5% of pts had no pathologic responses
holds global development and commercial rights. This research was in part made possible by an award from (defined by INMC as .50% vital tumour remaining in the tumour bed). [1] Bex A et al. Efficacy, safety,
the CIRM (CLIN2-15343).; California Institute for Regenerative Medicine; Award Number CLIN2-15343.
and biomarker analysis of neoadjuvant avelumab/axitinib in patients (pts) with localized renal cell
CD70 positive carcinoma (RCC) who are at high risk of relapse after nephrectomy (NeoAvAx), in 2022 ASCO
All CD70 receiving phase CD70 negative Genitourinary Cancers Symposium. Journal of Clinical Oncology, Volume 40, Number 6_suppl, https://
positive 1b regimena or unknown [Link]/10.1200/JCO.2022.40.6_suppl.2. Clinical trial information: NCT03341845. Research Sponsor:
n = 30 n = 12 n=8
Pfizer, CellCarta, The Netherlands Cancer Institute and Queen Mary University London.
Best overall response (CR or PR at any 8/30 (27) 4/12 (33) 0/8 (0)
visit), n/N (%)
CD70 ‡50b 8/24 (33) 4/9 (44) –
CD70 <50b 0/6 (0) 0/3 (0) –
ORR (confirmed CR or PR), n/N (%) 6/30 (20) 3/12 (25) 0/8 (0)
CD70 ‡50b 6/24 (25) 3/9 (33) –
CD70 <50b 0/6 (0) 0/3 (0) –
a
ALLO-316 80 3 106 CAR+ T cells and LD with fludarabine 30 mg/m2 + cyclophosphamide 500 mg/m2.
b
IHC-based tumor proportion score.

4510 Clinical Science Symposium 4511 Clinical Science Symposium

Genomic characterization of baseline and post-progression tumors in An integrative analysis of circulating and tumor microenvironment (TME)
IMmotion010, a randomized, phase 3 study of adjuvant (adj) atezolizumab determinants of patient response in the Checkmate 9ER (CM 9ER) trial of
(atezo) vs placebo (pbo) in patients (pts) with high-risk localized renal cell nivolumab and cabozantinib (NIVO+CABO) in advanced renal cell carcinoma
carcinoma (RCC). First Author: Sumanta Kumar Pal, Department of Medical Oncology (aRCC). First Author: David A. Braun, Center of Molecular and Cellular Oncology, Yale
& Therapeutics Research, City of Hope Comprehensive Cancer Center, Duarte, CA Cancer Center, Yale School of Medicine, New Haven, CT
Background: Although adj atezo did not prolong disease-free survival (DFS) in the IMmotion010 trial, accompanying Background: The CM 9ER trial demonstrated increased objective response rate (ORR),
studies identified post-nephrectomy serum KIM-1 levels to be potentially predictive of benefit with atezo. We inves-
tigated whether tissue genomics could complement these findings. Methods: Tumors were obtained from patients pre- progression-free and overall survival in patients with aRCC treated with NIVO+CABO
treatment and (if additional consent obtained) at disease recurrence. Whole transcriptome profiles were generated using compared to sunitinib (SUN). For vascular modulating therapies (e.g. CABO and SUN) and
TruSeq RNA Access Technology (Illumina). Previously, non-negative matrix factorization (NMF) was performed in a immunotherapies (like NIVO), the state of the TME and activity of stromal cells can
separate phase 3 study in advanced RCC (IMmotion151), establishing 7 molecular subgroups (NMF 1-7) (Motzer et al
Cancer Cell 2020). For each IMmotion010 tumor, we derived signature scores dichotomized at the median as well as
modulate tumor response to therapy. Methods: We investigated how the TME and
NMF1-7 subtype using random forest. Clinical outcome was assessed within these groups alone and with the addition of circulating factors were associated with response to NIVO+CABO using pre-treatment
baseline serum KIM-1 levels, dichotomized into KIM-1 high (KIM-1H) and KIM-1 low (KIM-1L) using the previously tumor PD-L1 staining, human interpretable features (HIF) derived from H&E tissue
established cutoff of 86 pg/mL. Where possible, we sought to characterize the evolution of molecular profiles at disease sections, circulating immune cell populations quantified by flow cytometry, and circu-
recurrence. Results: Baseline tissue was obtained from 754 pts, reflecting 97% of the intention-to-treat population.
Tumors from KIM-1H patients were enriched in myeloid, granulocyte and proliferation gene signatures at baseline. Among lating extracellular matrix (ECM) markers quantified by competitive ELISAs from 150
722 pts for whom both serum KIM-1 and NMF subtype could be characterized, pts in cluster 6 (stromal/proliferative) patients (23% of ITT) enrolled in CM 9ER. We employed principal component analysis,
appeared to derive benefit from atezo (n=50) (Table). Across all patients, no difference in outcome was observed among varimax rotation, and Feature Set Enrichment Analysis (FSEA) to identify a subset of
baseline TeffH and TeffL subsets. Within the KIM-1H population, TeffH tumors were associated with longer DFS with atezo
vs pbo. Paired baseline/recurrence tissue was obtained from 80 pts (atezo: 49; pbo: 31). At recurrence, tumors exhibited
biological measurements capturing 85% of the data variability. We constructed a logistic
increased stromal and proliferation gene signatures, regardless of treatment, reflected in an increased proportion in regression model to associate the most variable features with patient response (ORR per
NMF6 (baseline: 6%; progression: 22%). Exploratory analyses also revealed a decreased MHC-I signature after treatment BICR) to NIVO+CABO or SUN therapy. Results: An unbiased clustering and feature
with atezo. Conclusions: This is the first report of tissue genomic profiling in a phase 3 adjuvant immune checkpoint extraction approach was used to identify measurements contributing to multi-modal
inhibitor study in RCC. While certain molecularly defined subsets may carry predictive value, serum KIM-1 remains the
most robust predictor of outcome with atezo. Analyses of progression biopsies highlight an evolution in genomic profile variability in the TME across 150 patients (~4000 biological measurements reduced to 16
and offer insights into mechanisms of relapse. Clinical trial information: NCT03024996. Research Sponsor: F. Hoffmann- highly informative measurements): PD-L1 staining, 4 ECM markers, 4 PBMC markers, and
La Roche Ltd. 7 H&E HIF features. A final binary logistic regression model was built employing lasso
Subgroup n DFS HR regularization to associate these 16 features to short-term response to NIVO+CABO or
Serum KIM-1
SUN while minimizing spurious associations. Based on these regression models, the ECM
KIM-1H 290 0.7* marker VICM, a citrullinated fragment of vimentin released by matrix metalloproteases,
KIM-1L 443 1.13
T-effector signature and which measures macrophage activity and immune status, was prognostic across both
TeffH 367 0.87 therapies. Logistic regression models that integrated highly informatic features had AUCs
TeffL 366 0.97
NMF subtype of 0.76 for NIVO+CABO model and 0.72 for the SUN model. Within the NIVO+CABO arm,
NMF1 82 1.14 this integrative model uncovered high VICM (and therefore anti-tumor macrophage
NMF2 271 1.04
NMF3 131 0.96 polarization), high PD-L1 staining, high plasma cell numbers and high cancer cell numbers
NMF4 62 0.88 at the epithelial stromal interface, low levels of circulating fragment of C-terminal type
NMF5 60 0.77
NMF6 50 0.25* VIa3 collagen (Pro-C6), and low percentages of circulating regulatory (Foxp3+ CD4+)
NMF7 66 1.62 T cells as determinants of therapeutic response. Conclusions: Taken together, these
T-effector signature in KIM-1H pts
KIM-1H TeffH 147 0.59* findings indicate that the state of the tumor microenvironment and circulating factors
KIM-1H TeffL 143 0.93 together have an effect on patient responsiveness to NIVO+CABO in aRCC and provides a
*P,0.05. framework for integrative analysis for biomarker discovery. Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GENITOURINARY CANCER—KIDNEY AND BLADDER 331s

4512 Clinical Science Symposium LBA4513 Rapid Oral Abstract Session

Gut-associated checkpoint as a prognostic biomarker in metastatic renal ENLIGHTED phase 3 study: Interim results of efficacy and safety of pade-
cell carcinoma (mRCC): Results from a randomized first-line clinical trial. liporfin vascular targeted photodynamic therapy (VTP) in the treatment of
First Author: Renee Maria Saliby, Yale Cancer Center, New Haven, CT low-grade upper tract urothelial cancer (LG UTUC). First Author: Vitaly Margulis,
Background: The gut microbiota modulates anti-cancer immune response and therefore UT Southwestern Medical Center, Dallas, TX
benefit to immune checkpoint inhibitors (ICIs). Gut dysbiosis impacts the MadCAM-1/a4b7
axis leading to recirculation of immunosuppressive a4b7+Tr17 cells into tumors. From
mechanistic insight to biomarker development, soluble MAdCMA-1 (sMAdCAM-1) is a cir-
culating surrogate marker of gut dysbiosis. We aim to develop sMAdCAM-1 as a prognostic
biomarker to ICI-based therapy in patients (pts) with mRCC. Methods: Using a Luminex assay,
sMAdCAM-1 levels were measured in available plasma samples at baseline from 612 pts (69%
of the intent-to-treat population) from the phase III JAVELIN Renal 101 trial (NCT02684006),
which compared avelumab + axitinib with sunitinib in previously untreated mRCC pts.
sMAdCAM-1 was examined on the original, per 10^4-scaled, and log-transformed scales. Linear
assumption was visually checked by deviance residual and restricted cubic splines (RCS) plots.
Optimal cut-off value was established based on the maximum log-rank statistic. Cox re-
gression models were used to assess associations with progression-free survival (PFS) and
overall survival (OS). The discrimination of the fitted model was assessed by time-dependent
AUC index. Results: Higher sMAdCAM-1 levels were associated with improved PFS (median: The full, final text of this abstract will be available at
13.9 [11.3 - 6.6] vs 8.4 [6.0 - 9.9] months) and OS rate (at 18 months: 84.2% [80.2 - 87.4] vs
68.1% [59.2 - 75.5]). These associations remained after adjusting for IMDC risk groups
[Link] on the day of presentation and in the
(Table 1). The optimal cutoff was 180 ng/ml (25% percentile) based on the OS outcome in the online supplement to the June 10, 2025, issue of the Journal
whole population. Residual and RCS plots further confirmed a non-linear relationship of of Clinical Oncology.
sMAdCAM-1 levels with OS and PFS. Median follow up was 18.9 months. The prognostic model
incorporating IMDC + sMAdCAM-1 demonstrated a significant improvement in the AUC at
18 months compared to IMDC alone (0.72 vs 0.68; p=0.01). These associations were inde-
pendent of study arm. Conclusions: Higher sMAdCAM-1 is associated with improved out-
comes to 1st line regimens in mRCC. sMAdCAM-1 may have an added prognostic value to
IMDC. As a diagnostic test of gut dysbiosis, it might guide the selection of pts eligible to
microbiota-modulating strategies. The validation in two independent datasets and multi-omics
correlation (i.e. fecal metagenomics) are ongoing under an international collaboration network.
Research Sponsor: None.
Association of MAdCam-1 with clinical outcomes using Cox regression model.
Parameters PFS HR (95% CI) OS HR (95% CI)
Favorable Ref Ref
Intermediate 1.77 (1.31- 2.39) ,0.001 3.01 (1.60 - 5.66) ,0.001
Poor 2.88 (1.99 - 4.17) ,0.001 7.91 (4.02 - 15.56) ,0.001
sMAdCAM-1(high vs low) 0.75 (0.59 - 0.96) 0.021 0.59 (0.41 - 0.85) 0.004

4514 Rapid Oral Abstract Session 4515 Rapid Oral Abstract Session
Five-year follow-up results from the phase 3 KEYNOTE-564 study of adju- Zanzalintinib (zanza) + nivolumab (nivo) 6 relatlimab (rela) in patients (pts)
vant pembrolizumab (pembro) for the treatment of clear cell renal cell with previously untreated clear cell renal cell carcinoma (ccRCC): Results
carcinoma (ccRCC). First Author: Naomi Balzer Haas, Abramson Cancer Center, from an expansion cohort of the phase 1b STELLAR-002 study. First Author:
University of Pennsylvania, Philadelphia, PA Jad Chahoud, Moffitt Cancer Center, Tampa Bay, FL
Background: KEYNOTE-564 (NCT03142334) established pembro monotherapy as the first adjuvant Background: VEGFR-targeted tyrosine kinase inhibitors (TKIs) in combination with immune
regimen to significantly improve both disease-free survival (DFS) and overall survival (OS) vs placebo checkpoint inhibitors (ICIs) are standard of care for previously untreated metastatic ccRCC.
(pbo) after surgery for participants (pts) with ccRCC at increased risk of recurrence. We present results Zanza (XL092) is a novel, oral, multi-targeted TKI of VEGFR, MET, and TAM kinases (TYRO3,
from the fourth prespecified interim analysis with a minimum follow-up of 5 yrs. Methods: KEYNOTE- AXL, MER), with a short half-life that may have an improved therapeutic index. In the phase 1
564 is a randomized, double-blind, pbo-controlled, phase 3 study, which enrolled adults with ccRCC
STELLAR-001 study, the tolerability profile of single-agent zanza was manageable and
with intermediate-high (pT2 Gr 4 or sarcomatoid, or pT3 any Gr, N0 M0) or high (pT4 any Gr, N0 M0, or
any pT and Gr, N+ M0) risk of recurrence or M1 with no evidence of disease (NED) who had ne-
antitumor activity was observed in patients with previously treated advanced ccRCC (Pal
phrectomy and/or metastasectomy #12 wks prior to 1:1 randomization to pembro 200 mg or pbo IV et al, IKCS NA 2023). STELLAR-002 (NCT05176483) is a phase 1b, open-label study
Q3W. Treatment continued for ~1 yr (17 cycles) or until recurrence, intolerable AEs, or physician evaluating the tolerability and activity of zanza alone and in combination with ICIs in pts with
decision to discontinue treatment. The primary endpoint was DFS by investigator; the key secondary advanced solid tumors. Here, data from the expansion cohort of pts with previously un-
endpoint was OS. The study met its DFS and OS objectives at earlier analyses; thus, no subsequent treated ccRCC receiving zanza + nivo 6 rela are presented. Methods: Adult patients with
formal statistical testing occurred. AEs were collected for 30-90 days after treatment cessation unresectable advanced or metastatic (adv/met) ccRCC of any IMDC risk, and no prior
depending on severity, with serious treatment-related AEs collected regardless of timing. Results: A systemic anticancer therapy for adv/met ccRCC were enrolled into one of two non-
total of 994 pts were randomized to pembro (n = 496) or pbo (n = 498). The median follow-up time to randomized arms. Patients received zanza 100 mg orally with either nivo 480 mg IV ev-
data cutoff date of 25 Sept 2024 was 69.5 mo (range, 60.2286.9). All pts completed or discontinued ery 4 weeks (q4w) or nivo/rela 480/480 mg IV q4w (fixed-dose combination). Primary
study treatment $3 years earlier. 188 DFS events in the pembro group and 241 in the pbo group had endpoints were investigator-assessed ORR per RECIST 1.1 and safety. Results: In the zanza
occurred. Median DFS was not reached (NR) vs 68.3 mo, respectively (HR 0.71, 95% CI 0.5920.86);
+ nivo arm (n = 40), 75% had intermediate or poor IMDC risk disease. After median follow-up
estimated DFS rate at 5 yrs was 60.9% vs 52.2%. 68 OS events in the pembro group and 99 in the pbo
group had occurred. Median OS was NR in both arms (HR 0.66, 95% CI 0.4820.90); estimated OS rate of 16.1 months, the ORR was 63% (4 complete responses [CRs], 21 partial responses [PRs]),
at 5 yrs was 87.7% vs 82.3%, respectively. DFS and OS outcomes were consistent across key and disease control rate (DCR: CR+PR+SD) was 90%. The 6- and 12-month PFS rates were
subgroups, including by prespecified risk and sarcomatoid features (Table). No new serious treatment- 83.2% and 64.2%, respectively. The most common any grade (G) treatment-emergent
related AEs have been reported for $3 years. Conclusions: With $5 yrs of follow-up, the benefits adverse events (TEAEs) were diarrhea (78%), hypertension (58%), and nausea (58%). The
observed with adjuvant pembro vs pbo are consistent with prior analyses, including in all subgroups. most common G3/4 AEs related to zanza were hypertension (30%) and diarrhea (15%).
No new serious treatment-related safety signals occurred. Adjuvant pembro remains a standard-of- Treatment-related palmar-plantar erythrodysesthesia (PPE) was reported in 28% (8% G3, 0%
care option for patients with ccRCC at increased risk of recurrence. Clinical trial information: G4). Two (5%) pts discontinued both study treatments due to treatment-related AEs
NCT03142334. Research Sponsor: This study was supported by Merck Sharp & Dohme Corp., a (TRAEs). Median average daily zanza dose was 49.5 mg (range: 26-100). In the zanza + nivo/
subsidiary of Merck & Co., Inc., Rahway, NJ, USA. rela arm (n = 40), 70% had intermediate or poor IMDC risk disease. After a median follow-up
Prespecified risk subgroups Sarcomatoid features of 11.9 months, the ORR was 33% (1 CR, 12 PRs) and DCR was 90%. The 6- and 12-month
PFS rates were 80.2% and 58.8%, respectively. The most common any G TEAEs were
All pts Intermediate-high High M1 NED Present Absent
diarrhea (60%) and nausea (50%). The most common G3/4 AE related to zanza was hy-
N 994 855 77 57 111 829 pertension (13%). Treatment-related PPE occurred in 5% (0% G3/4). Seven (18%) pts
DFS events 429 342 49 37 59 339
DFS HR 0.71 0.75 0.61 0.48 0.56 0.75 discontinued all study treatment due to TRAEs. Median average daily zanza dose was
(95% CI) (0.5920.86) (0.6120.93) (0.3521.08) (0.2520.92) (0.3320.96) (0.6020.92) 54.9 mg (range: 31-100). No G5 TRAEs occurred in either arm. Conclusions: First-line zanza
OS events 167 131 22 13 23 130 had acceptable tolerability in combination with nivo or nivo/rela with a low rate of PPE;
OS HR 0.66 0.65 0.86 0.36 0.67 0.64 zanza+nivo showed promising preliminary activity in pts with adv/met ccRCC. Clinical trial
(95% CI) (0.4820.90) (0.4620.92) (0.3721.98) (0.1121.18) (0.2921.56) (0.4520.91)
information: NCT05176483. Research Sponsor: Exelixis, Inc.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
332s GENITOURINARY CANCER—KIDNEY AND BLADDER

4516 Rapid Oral Abstract Session 4517 Rapid Oral Abstract Session
Ipilimumab and nivolumab in patients with metastatic clear cell renal cell Sasanlimab in combination with bacillus Calmette-Guérin (BCG) in BCG-
carcinoma (mccRCC) treated on the phase 3 PDIGREE (Alliance A031704) naive, high-risk non–muscle-invasive bladder cancer (NMIBC): Event-free
trial: Results from Step 1 analysis. First Author: Tian Zhang, Department of Internal survival (EFS) subgroup analyses based on disease stage from the CREST
Medicine, Division of Hematology and Oncology, UT Southwestern, Dallas, TX study. First Author: Thomas Powles, Barts Cancer Centre, Cancer Research UK Ex-
Background: Ipilimumab/nivolumab (ipi/nivo) is a standard of care first-line treatment (tx) perimental Cancer Medicine Centre, Queen Mary University of London, London, United
for patients (pts) with mccRCC; however, ideal timing of subsequent immunotherapy txs are Kingdom
not well defined. We performed an analysis of patients initially treated with ipi/nivo and Background: Sasanlimab in combination with BCG (induction [IND] and maintenance [MNT]) sig-
subsequent cohort assignments on PDIGREE (A031704). Methods: The PDIGREE trial nificantly improved investigator (INV)-assessed EFS vs BCG (IND and MNT) and had a manageable
treated pts with IMDC intermediate/poor risk mccRCC with first-line ipi/nivo (Step 1) at safety profile in patients (pts) with BCG-naive, high-risk NMIBC, according to the primary analysis
NCTN sites (categorized as academic (A), academic regional (R), and community (C)). results from the phase 3 CREST study. Here, we report exploratory EFS subgroup analyses not
Subsequent management was based on iRECIST responses at 12 weeks: pts with complete previously presented based on disease stage at randomization from Arms A and C. Methods: Eligible
pts were randomized [Link] to receive sasanlimab in combination with BCG (IND and MNT; Arm A),
response (CR) received 1-year nivo maintenance; pts with progressive disease (PD) received
sasanlimab in combination with BCG (IND; Arm B), or BCG (IND and MNT; Arm C). To assess the impact
cabozantinib (cabo) monotherapy, pts with non-CR/non-PD were randomized to nivo with or on efficacy of carcinoma in situ (CIS) and T1 tumors at baseline, post hoc INV-assessed EFS analyses
without cabo (for primary endpoint of overall survival [OS], target sample size was 1175). were conducted for the comparison of Arm A vs Arm C. EFS was defined as time from randomization to
Enrollment was held when the randomized sample target was reached. Pts with unresolved recurrence of high-grade disease, progression of disease, persistence of CIS (for patients with CIS at
toxicity at week 22 were managed off protocol. Step 1 pt demographics, assignments into randomization), or death due to any cause, whichever occurred first. Results: At the data cutoff date
Step 2, and adverse event (AE) data are presented. Descriptive statistics were used, and (Dec 02, 2024), the median duration of follow-up for EFS was 36.4 and 36.7 months for Arm A and Arm
cohorts were compared using a chi-square test. These data were released with DSMB C, respectively. A total of 176 pts with CIS (with or without papillary tumors) were in Arms A and C, 102
approval and do not inform the primary OS endpoint (1EP). Results: From May 2019 to May of whom had CIS without papillary tumors. A total of 398 pts with T1 tumor were in Arms A and C, 342
2024, 1111 pts were enrolled and treated with ipi/nivo. Pt characteristics included median of whom had T1 tumor without CIS (Table). Three-year landmark EFS subgroup analyses not previously
age 64.0 years (range 29.0-86.0); 819 (73.7%) males; White (85.1%), Hispanic (10.4%), Black presented are reported in the table. For patients with CIS, with or without concomitant papillary
(4.2%), and American Indian/Native Hawaiian (1.3%); 849 (76.8%) intermediate risk/257 tumors, the 3-year EFS rate was 83.0% in Arm A and 71.8% in Arm C. For patients with T1 tumors, with
or without CIS, the 3-year EFS rate was 81.3% in Arm A and 72.2% in Arm C. Conclusions: Sasanlimab
(23.2%) poor risk; 458 (41.2%) at Academic, 113 (10.2%) at Regional, and 540 (48.6%) at
in combination with BCG (IND and MNT) improved EFS outcomes in the overall population and in the
Community centers; and 603 (54.3%) with de novo metastases. 364 pts (33%) stopped tx in subgroups of pts with BCG-naive, high-risk NMIBC who had CIS or T1 tumors at randomization. This
Step 1: 160 (44%) for AEs, 46 (13%) for PD/clinical PD/suspected PD, 42 (12%) withdrawals, post hoc analysis further supports the potential of sasanlimab in combination with BCG as a practice-
39 (11%) alternative txs, 37 (10%) deaths on the study, 12 (3%) other complicating disease, 8 changing treatment in pts with aggressive disease. Clinical trial information: NCT04165317. Research
(2%) MD decision and 19 (5%) other reasons. Of the 37 deaths, 15 (1.4%) grade 5 SAEs were Sponsor: Pfizer Inc.
reported, 6 of which were due to PD. Of 747 (67%) pts registered to Step 2 at 3 months, 9
Arm A(N=352) Arm C(N=351) HR (95% CI)
(1.2%) achieved CR, and 141 (18.9%) pts were assigned to the PD cohort. 597 pts (80%) were Arm A vs Arm C
randomized for the 1EP, notable with fewer pts with poor risk [(21 vs 27%, p = 0.01] and bone 36-mo EFS rate, 36-mo EFS rate,
metastases [24.5% vs 34.2%, p = 0.0007] compared to Step 1 pts who discontinued tx. Gr 3/4 n (%) % (95% CI) n (%) % (95% CI)
tx-related AEs in 314/1093 (29%) evaluable pts included diarrhea/colitis (8%), transaminase All pts 82.1 (77.4-85.9) 74.8 (69.7, 79.2) 0.68 (0.489-0.941)a
elevation (3%), rash (2%), adrenal insufficiency (2%), fatigue (2%), and hypophysitis (1%). Tumor type at randomization
CIS with and without papillary 88(25.0) 83.0 (72.9-89.6) 88(25.1) 71.8 (60.4-80.5) 0.53(0.285-0.982)
Conclusions: The PDIGREE trial enrolled a representative US-based population with tumors
mccRCC. Step 1 pt characteristics and associated outcomes with induction ipi/nivo reveal CIS without papillary tumors 52(14.8) 81.0 (66.6-89.7) 50(14.2) 75.4 (59.9-85.6) 0.52 (0.233-1.165)
T1 with and without CIS 204(58.0) 81.3 (74.7-86.4) 194(55.3) 72.2 (65.0-78.2) 0.63 (0.406-0.963)
new insights into the tolerability and tx response of 1L ipi/nivo. Clinical trial information: T1 without CIS 178 (50.6) 79.6 (72.1-85.2) 164(46.7) 72.5 (64.6-78.9) 0.70 (0.446-1.109)
NCT03793166. Research Sponsor: Alliance Group: U10CA180821, U10CA180882; SWOG: a
U10CA180888; [Link] Stratified HR; all other HR unstratified.

4518 Rapid Oral Abstract Session 4519 Rapid Oral Abstract Session
First results of SURE-02: A phase 2 study of neoadjuvant sacituzumab 9MW2821, a novel Nectin-4 antibody-drug conjugate (ADC), combined with
govitecan (SG) plus pembrolizumab (Pembro), followed by response- toripalimab in treatment-naı̈ve patients with locally advanced or metastatic
adapted bladder sparing and adjuvant pembro, in patients with muscle- urothelial carcinoma (la/mUC): Results from a phase 1b/2 study. First Author:
invasive bladder cancer (MIBC). First Author: Andrea Necchi, IRCCS San Raffaele Shusuan Jiang, Hunan Cancer Hospital, Changsha, China
Hospital, Vita-Salute San Raffaele University, Milan, Italy Background: Nectin-4 is an adhesion molecule that is highly expressed in variety of
Background: Standard of care for MIBC is radical cystectomy (RC) with neoadjuvant che- solid tumors. Previous study of 9MW2821 has shown promising efficacy and tolerable
motherapy (CT), but ~50% of patients (pts) are ineligible for/refuse CT and survival for RC alone toxicity in different advanced cancers, especially in urothelial cancer, cervical cancer,
is dismal. Neoadjuvant Pembro and SG monotherapies showed activity in MIBC within PURE- esophageal cancer and breast cancer. Here we report preliminary results of 9MW2821
01 and SURE-01 studies. SURE-02 (NCT05535218) is a phase 2 study of neoadjuvant combined with Toripalimab in treatment-naı̈ve patients with la/mUC. Methods: This is
SG+Pembro and adjuvant Pembro, including a bladder-sparing approach depending on clinical an open-label, multicenter, phase 1b/2 study to evaluate the safety and efficacy of
response. We report results of a prespecified interim analysis. Methods: Pts age $18 y, ECOG 9MW2821 combined with Toripalimab in la/mUC. Patients received 9MW2821 on D1/D8
PS 0-1, with histologically confirmed cT2-T4N0M0 MIBC, ineligible/refusing CT, and scheduled
and Toripalimab on D1, 21 days per cycle. Primary objective was safety, and secondary
for RC received 4 cycles of Pembro 200 mg on D1 and SG 7.5 mg/Kg on D1 and D8, Q3W,
objectives were efficacy, pharmacokinetics and immunogenicity. Results: 40 treat-
followed by postsurgical Pembro x 13 cycles, Q3W. A reTURBT is allowed instead of RC,
followed by Pembro x 13 cycles, for pts achieving a clinical complete response (cCR), ment-naı̈ve patients with la/mUC were enrolled and received the combination therapy of
stringently defined as a negative magnetic resonance imaging (MRI) and no residual viable 9MW2821(1.25mg/kg) and Toripalimab(240mg). Median age was 66.5 years [36-78], and
tumor at reTURBT (ypT0). Primary outcome measure is cCR rate: H0#30%, H1$45%. Other 73% patients were ECOG 1. 55% primary tumor sites were upper tract urothelial car-
outcomes: ypT#1N0-x rate including pts undergoing RC, safety (CTCAE v.5.0), survival. The cinoma. As of Dec 19, 2024, ORR was 87.5% [35/40, 95%CI 73.2-95.8], including 7.5% CR
total sample size is 48 pts in a 2-stage design, with 23 pts enrolled at first stage ($7 cCR rate (comfirmed ORR was 80%). DCR was 92.5% [37/40, 95%CI 79.6-98.4]. Median PFS
needed). Decipher Bladder (Veracyte, San Diego, CA) was used for transcriptome-wide an- and DoR were not reached, 6-month PFS rate and 3-month DoR rate were 79.1% and
alyses of primary TURBT tissue. Results: From 10/23 to 01/25, 40 pts were treated and 31 100%. Furthermore, ORR of subgroups in liver metastasis, bladder cancer and tumor with
were efficacy evaluable. 20 (64.5%) had a cT2 stage, 12 (38.7%) had a centrally confirmed negative expression of Nectin-4 were 88.2%, 94.4%, 100%, respectively. These showed
variant histology. The cCR-rate was 38.7% (N = 12; 95%CI: 21.8-57.8); all these pts underwent a that different subgroups of treatment-naı̈ve patients could benefit from the combination
reTURBT; ypT#1N0-x rate was 51.6% (N = 16). Grade $3 treatment-related adverse-events therapy of 9MW2821 and Toripalimab. The most common treatment-related AEs(TRAEs)
occurred in 4 pts (12.9%), 2 dose omissions of SG and one dose delay (1W) were recorded. No were grade 1 or 2, 23.8% patients experienced TRAEs of grade 3 or above, including
SG dose-reduction was needed. cCR varied by molecular subtype. Transcriptome results were neutrophil count decreased (7.1%), rash (4.8%), ALT increased (4.8%), etc. No TRAEs led
available for 23 pts: complete pathologic (ypT0) responses varied by Genomic Subtyping to death occurred. No new safety signals of 9MW2821 or Toripalimab were observed in
Classifier (GSC) groups with luminal tumors showing higher ypT0 rates vs non-luminal (73 vs this study. Conclusions: 9MW2821 combined with Toripalimab in treatment-naı̈ve
25%, p = 0.04). Similarly, based on Lund subtypes, genomically unstable (GU) tumors had ypT0
patients with la/mUC demonstrated remarkable efficacy and well-tolerated safety
response in 67%, vs 57% for urothelial-like, 20% for basal/squamous and 0% for
profile. A pivotal phase 3 study is ongoing currently. Clinical trial information:
neuroendocrine-like. Higher stromal signature (. median) was associated with non-ypT0
response (p = 0.004), while neither Trop2 (p = 0.15) nor TOP1 (p = 0.79) gene expression were
NCT06079112. Research Sponsor: Mabwell (Shanghai) Bioscience Co., Ltd.
associated with ypT0 response. Conclusions: Perioperative SG+Pembro revealed a com-
pelling cCR rate, with a manageable safety profile, allowing a bladder preservation in ~40% of
pts. Pre-treatment molecular biomarker analyses suggest a unique tumor profile associated
with cCR. Overall, SURE-02 interim results support the completion of study accrual and further
investigation of SG+Pembro in pts with MIBC. Clinical trial information: NCT05535218.
Research Sponsor: Gilead; Merck; Associazione Italiana per la Ricerca sul Cancro (AIRC); IG
27746.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GENITOURINARY CANCER—KIDNEY AND BLADDER 333s

4520 Rapid Oral Abstract Session 4521 Rapid Oral Abstract Session
CLONEVO: Preoperative abemaciclib for cisplatin-ineligible muscle-invasive AREN1721, a randomized phase 2 trial of axitinib+nivolumab combination
bladder cancer (MIBC) with molecular response assessment. First Author: therapy vs. single agent nivolumab for the treatment of TFE/translocation
Bishoy Morris Faltas, Weill Cornell Medicine, New York, NY renal cell carcinoma (tRCC) across all age groups, an NCI National Clinical
Background: Up to 40% of MIBC patients are ineligible to receive standard neoadjuvant Trials Network (NCTN) phase 2 study. First Author: Nicholas Cost, Denver Chil-
cisplatin-based chemotherapy creating a significant unmet need. Based on our prior drens, Denver, CO
findings of frequent cell cycle alterations, we conducted the first window-of-opportunity, Background: tRCC accounts for approximately 50% of pediatric RCC and 1-5% of RCC cases overall. tRCC,
investigator-initiated trial of the CDK4/6 inhibitor abemaciclib (abema) followed by radical driven by TFE3 or TFEb fusions or amplifications (TFEb), are often aggressive with no existing standard for
systemic therapy. Methods: AREN1721 was a prospective randomized COG-led NCTN phase 2 trial of
cystectomy (RC) in MIBC (NCT03837821). Methods: Eligibility was MIBC appropriate for
nivolumab/axitinib combination therapy vs. axitinib alone (closed early for feasibility) vs. nivolumab alone in
RC and cisplatin-ineligibility or refusal. Planned treatment was abema (200mg BID PO) for children and adults with advanced unresectable or metastatic tRCC. Prior exposure to anti-PD1/PDL1
4-8 weeks prior to RC. We planned to enroll 20 patients (accounting for 20% attrition). 16 therapies or axitinib was prohibited. The primary endpoint was progression-free survival (PFS), defined as the
evaluable patients provided 80% power to detect 0.75 effect size (a = 0.05, r = 0.5 between time from randomization to the earliest of disease progression based on immune-modified RECIST criteria or
pairs). Whole-exome (WES) and RNA sequencing of pre- and post-abema tissues and serial death. The final protocol version targeted enrollment of 28 eligible patients to detect a hazard ratio (HR) of
evaluation of ctDNA WES were performed on Caris Life Sciences’ platform. Results: 20 0.40 for the comparison of nivolumab/axitinib vs. nivolumab alone using a one-sided log-rank test with alpha
= 0.15. Results: Despite aggressive approaches for trial recruitment, AREN1721 was closed after enrolling
patients received abema for a median of 36 days. Median age was 73, 16/20 were males,
15 patients (13 eligible) from 2019 to 2023 secondary to poor accrual. Median age 16 years (range 7-42) with
and 5/20 had cT4. 3 didn’t undergo RC, and 1 withdrew consent. Abema resulted in 9/13 age , 18 years; 9/13 were male. Six patients were randomized to nivolumab+axitinib, 2 to axitinib alone,
pathologic complete response in 18.8% (3/16) and downstaging in 31.3% (5/16). No and 5 to nivolumab alone. There were no unexpected toxicities. Thirty-three percent of patients randomized
unexpected safety signals were detected. Grade 3 abema-related adverse events included to nivolumab+axitinib experienced partial response, compared to 0% in the other arms, and 0% of patients on
anemia (4/20), abdominal pain (1/20) and diarrhea (1/20). Imaging Mass Cytometry of pre- the combination arm experienced primary disease progression. Addition of axitinib to nivolumab significantly
and post-abema tissues showed a significant reduction in RB1 phosphorylation after improved PFS (p = 0.0004), extending median PFS from 1.8 to 10.5 months. Overall survival also improved (p
= 0.003) with the addition of axitinib. Conclusions: Nivolumab+axitinib combination therapy was statis-
abema confirming on-target activity. Variant allele frequency of somatic mutations sig-
tically more active than nivolumab single agent therapy, which itself was inactive. Whether anti-PD1 pathway
nificantly decreased after abema by 20.5% (p = 0.04), confirming its role in decreasing inhibitors add benefit to anti-VEGF therapy for tRCC remains to be determined. Optimizing trial recruitment is
tumor burden. Serial ctDNA showed a significant reduction in tumor fraction (TF) following critical for this rare but aggressive cancer. Clinical trial information: NCT03595124. Research Sponsor: U.S.
abema by 28.6%. Post-TURBT pre-abema TF increased but rapidly decreased within National Institutes of Health; U10CA098543; :U10CA180886; U10CA098413, U10CA180899.
2 weeks of abema (19.36%), confirming TF reduction was driven by abema not TURBT. Descriptive statistics by arm.
Patients with CCND1 amplification had the most significant decrease in TF (63.8%) Arm A:
highlighting CCND1 as a potential response biomarker. Abema significantly downregulated Axitinib/ Arm B: Arm C:
MKI67, CCNA2, and PCNA proliferation markers with log-fold changes of -1.2, -0.7, and Nivolumab Axitinib Nivolumab Overall p-
Characteristic N=6 N=2 N=5 N = 13 value1
-0.6. Gene set enrichment analysis showed significant downregulation of E2F targets and
G1/S transition pathways. Patients who achieved pathologic downstaging had significant Age (Years) 0.715
Mean (SD) 18 (9) 19 (6) 18 (14) 18 (10)
decrease in E2F pathway activity (-1.6 vs. -0.4, p = 0.01) confirming that abema suppressed Median (Q1, Q3) 16 (10, 21) 19 (15, 23) 15 (12, 16) 16 (12, 21)
E2F-dependent cell proliferation. Interestingly, abema significantly inhibited homologous Min, Max 9, 32 15, 23 7, 42 7, 42
Age Category .0.999
recombination repair of double-strand DNA break (DSBs) (FDR = 0.001), particularly Age < 18 4 (67%) 1 (50%) 4 (80%) 9 (69%)
TOPBP1 and RAD51. Conclusions: This first trial of short-term preoperative abema in Age 18+ 2 (33%) 1 (50%) 1 (20%) 4 (31%)
MIBC demonstrated promising efficacy and tolerability while modulating cell cycle- Prior Anti-VEGF therapy 1 (17%) 0 (0%) 1 (20%) .0.999
No prior systemic therapy 5 (83%) 2 (100%) 4 (80%) 11 (85%)
dependent pathways. Our findings support future trials investigating sequential abema Best Overall Response 0.019
with antibody-drug conjugates such as enfortumab vedotin, where abema’s effects on Partial Response 2 (33%) 0 (0%) 0 (0%) 2 (15%)
Stable Disease 4 (67%) 2 (100%) 1 (20%) 7 (54%)
DSBs repair augment treatment response. Clinical trial information: NCT03837821. Progressive Disease 0 (0%) 0 (0%) 4 (80%) 4 (31%)
Research Sponsor: Eli Lilly. 1
Kruskal-Wallis rank sum test; Fisher’s exact test.

4522 Poster Session 4523 Poster Session

Ipilimumab/nivolumab versus standard of care in non-clear cell renal can- Efficacy and safety of second-line cabozantinib 6 atezolizumab for patients
cer: Results of the SUNNIFORECAST trial and potential role of the CPS score with advanced renal cell carcinoma after progression on immuno-oncology
and tumor nephrectomy. First Author: Lothar Bergmann, University Hospital combinations: Subgroup analysis of CONTACT-03. First Author: Cristina Suárez,
Frankfurt, Medical Clinic II (Hematology/Oncology), Frankfurt, Germany Medical Oncology, Vall d´Hebron Institute of Oncology (VHIO), Hospital Universitari Vall
Background: Non-clear cell renal cancers (nccRCC) are a rare and heterogeneous group d´Hebron, Vall d´Hebron Barcelona Hospital Campus, Barcelona, Spain
of .20 histological and molecular defined entities. Due to the rarity of these entities, the Background: Cabozantinib (cabo), a vascular endothelial growth factor receptor-associated tyrosine
clinical data are limited and large randomized trials are missing resulting in uncertainties for kinase inhibitor (TKI), is a preferred treatment option for second-line (2L) treatment of advanced renal cell
optimal treatment recommendations. So far, TKI therapy with or without immune checkpoint carcinoma (RCC) based on its superior efficacy vs everolimus (Choueiri TK, N Engl J Med 2015); however,
inhibitors (ICI) are considered standard of care (SOC) options in these diseases. Here we its activity after contemporary first-line (1L) immuno-oncology (IO) combinations is not well charac-
terized. CONTACT-03 was a large phase 3 study evaluating the efficacy and safety of cabo 6 atezo-
report the results of the academic prospective randomised European trial in therapy-naı̈ve
lizumab (atezo) after progression on previous IO treatment (Pal SK, Lancet 2023). We report the results
patients with advanced nccRCC entities, which compared ipilimumab/nivolumab (Ipi/Nivo) vs of a subgroup analysis of the safety and efficacy of 2L cabo 6 atezo in patients from CONTACT-03 who
SOC. Methods: We randomly assigned patients (pts) with nccRCC in a 1:1 ratio to receive received standard of care (SOC) 1L IO-IO or IO-TKI combinations. Methods: In CONTACT-03, adults with
either nivolumab 3 mg/kg IV combined with ipilimumab 1 mg/kg IV every 3 weeks for 4 doses metastatic RCC whose disease had progressed on IO-based regimens were randomized to cabo (60 mg PO
followed by a flat dose of 240 mg IV every 2 weeks or 480 mg every 4 weeks versus SOC by QD) alone or with atezo (1200 mg IV Q3W). This subgroup analysis included patients who had received 1L
investigators choice until disease progression or intolerance occurred. Pts were stratified in IO-IO or IO-TKI SOC combinations prior to enrolling in CONTACT-3. Outcomes for 2L treatment included
papillary vs. non-papillary nccRCC and according to IMDC risk score. Central pathology was PFS by blinded independent central review (BICR), OS, ORR, duration of response, and safety. Results: Of
mandatory to confirm the correct diagnosis of the nccRCC subtype according to the WHO 522 patients, 107 in the cabo arm and 129 in the cabo + atezo arm had received prior treatment with IO-IO
classification [Link] primary endpoint was the overall survival (OS) rate at 12 months (ipilimumab-nivolumab) or IO-TKI (axitinib-avelumab, axitinib-pembrolizumab, or lenvatinib-
pembrolizumab). Efficacy outcomes were comparable between treatments (Table). For cabo and
(mos), secondary endpoints were the OS rate at 6 mos and 18 mos, OS, progression-free
cabo + atezo, respectively, median PFS by BICR was 10.3 and 10.2 months, and ORR was 36% and 37%.
survival (PFS) and response rate (RR). Results: 309 pts ( 70.9% male, 29.1% female) out of Grade 3/4 treatment-related adverse events (AEs) were reported in 48% and 58% of patients treated with
316 pts were randomized to receive either Ipi/Nivo or SOC. 173 (56.0%) pts were of papillary cabo and cabo + atezo, respectively, treatment-related serious AEs were reported in 13% and 25% of
subtype (pRCC) and 143 (44.0%) pts of non-papillary subtypes, whereas 59 pts had chro- patients, and AEs led to dose modification in 87% and 92% of patients and discontinuation in 5% and 17%
mophobe (ccRCC), 20 sarcomatoid/rhabdoid, 10 collecting duct, 11 TFE3-rearranged or TFEB- of patients, respectively. Conclusions: Results from this post-hoc subgroup analysis of CONTACT-03
altered RCC and 37 other histological features. According to the IMDC score, 23.9% were of suggest 2L cabo is effective in patients with advanced RCC previously treated with 1L IO-IO or IO-TKI
favorable, 51.8% of intermediate and 24.3% of poor risk. The 12 mos OS rate for Ipi/Nivo of regimens. Safety was consistent with the overall study. These results can inform clinicians making 2L
78.3% (95%-CI 70.9%-83.9%) vs 68.3% (95%-CI 60.0%-75.3%) in the SOC arm was statistically treatment decisions for patients who have progressed on contemporary 1L IO-containing combinations.
significant (p=0.026). Median OS was 33.2 mos for the Ipi/Nivo arm and 25.2 mos for the SOC Clinical trial information: NCT04338269. Research Sponsor: Exelixis, Inc.
arm. The ORR of 32.8% vs. 19.4% and the median PFS of 5.4 mos vs 5.7 mos was not Efficacy outcomes with cabo 6 atezo after 1L IO combinations.
statistically significant different between both arms. The explorative endpoint CPS score Cabo Cabo + atezo
differed between the various subentities and was associated with an advantage in OS. Pts (n=107) (n=129)
with a CPS$1 had an OS-rate at 12 months of 79.3% in the Ipi/Nivo arm vs. 58.3% and a Median PFS by BICR, months (95% CI) 10.3 (7.95, 12.45) 10.2 (8.34, 10.64)
median OS of 38.6 mos vs. 18.8 mos (p=0.007). Furthermore, pts who did not underwent a Median OS, months (95% CI) NE (18.30, NE) 24.2 (20.24, NE)
Best overall response by BICR, % 36 37
tumor nephrectomy (possibly due to high risk) had a survival benefit with 26.3 mos in the Ipi/ Complete response, % 0 0
Nivo arm vs 16.5 mos in the SOC arm (p=0.065) in contrast to nephrectomized pts with 38.9 Partial response, % 36 37
mos vs 34.0 mos. Conclusions: The OS-Rate at 12 mos was significantly superior for Ipi/Nivo Stable disease, % 50 52
Progressive disease, % 10 5
in comparison to SOC and the primary endpoint was met. Additionally, pts in the Ipi/Nivo arm Not evaluable/missing, % 5 6
had a longer median OS, especially those with a CPS$1. Clinical trial information: Duration of response by BICR, months (95% CI) 15.0 (10.28, NE) 10.5 (7.95, NE)
NCT03075423. Research Sponsor: None. NE, not estimable.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
334s GENITOURINARY CANCER—KIDNEY AND BLADDER

4524 Poster Session 4525 Poster Session

68 18
Comparison of Ga-NY104 PET/CT with F-FDG PET/CT in patients with Anlotinib plus everolimus as first-line treatment for advanced non–clear cell
metastatic clear cell renal cell carcinoma (NYCRM): A prospective, com- renal cell carcinoma: 1 year updated results from UC-001, a single-center,
parative phase II study. First Author: Wenjia Zhu, Department of Nuclear Medicine, single-arm, phase II trial. First Author: Wen-Hao Xu, Fudan University Shanghai
Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Cancer Center, Shanghai, China
Peking Union Medical College, Beijing, China Background: For patients (pts) with recurrent or stage IV non-clear cell renal cell
Background: 68Ga-NY104 is a small-molecule PET agent selectively targeting carbonic carcinoma (nccRCC), the current guidelines recommend participation in clinical trials, or
anhydrase IX (CAIX), which is highly expressed on clear cell renal cell carcinoma (ccRCC). This the use of tyrosine kinase inhibitors (TKIs), such as sunitinib, or mTOR inhibitors, such as
phase II study aims to evaluate the diagnostic efficacy of 68Ga-NY104 PET/CT in patients with everolimus. Anlotinib, a novel multi-target TKI, inhibits vascular endothelial growth
metastatic clear cell renal cell carcinoma and compare it with 18F-FDG PET/CT. factor receptors, fibroblast growth factor receptors, platelet-derived growth factor
Methods: Patients with metastatic ccRCC were prospectively recruited in this study (Clin- receptors, and c-kit. The ALTER-UC-001 study (NCT05124431) is a single-center, single-
[Link]: NCT05879471). All participants underwent 68Ga-NY104 and 18F-FDG PET/CT arm, phase II trial evaluating the efficacy and safety of anlotinib plus everolimus as first-
within one week. Tyrosine kinase inhibitors were stopped at least one week before the study. line therapy in pts with advanced nccRCC. Data for 24 pts from Jan 2022 through Dec
Any lesion that can be detected on either PET or CT is included for further analysis. The number
2023 have been published in 2024 ASCO. Here, we present 1-year updated results.
and uptake of lesions were recorded. The diagnostic efficacy was determined at lesion level
and region level based on a comprehensive reference standard protocol. Results: Forty-four Methods: Eligible pts were those with advanced nccRCC and no prior systemic therapy
patients (mean age, 59.6 6 10.7) were recruited, including 40 men and 4 women. A total of 677 for advanced disease, with an Eastern Cooperative Oncology Group (ECOG) performance
lesions in 172 regions were identified, of which 568 lesions and 128 regions were considered status of 0-1. Pts received anlotinib (12 mg orally once daily on days 1-14 of each 3-week
positive for ccRCC based on reference standard. The lesion-level sensitivity and specificity of cycle) and everolimus (5 mg orally once daily). The primary endpoint was objective
68
Ga-NY104 PET/CT to detect ccRCC lesion are 96.6% (95% CI, 95.2% - 98.1%) and 99.1% (95% response rate (ORR). Secondary endpoints included disease control rate (DCR),
CI, 97.3% - 100%), which is significantly higher than that of 18F-FDG PET/CT (sensitivity, 77.8%, progression-free survival (PFS), and overall survival (OS). Adverse events (AEs) were
95% CI, 74.4% - 81.2%, P＜0.001; specificity, 5.5%, 95% CI, 1.2% - 9.8%, P＜0.001). The region- graded according to CTCAE v5.0. Results: Between January 2022 and December 2024,
level sensitivity and specificity of 68Ga-NY104 PET are 98.4% (95% CI, 96.3% - 100%) and 97.7% 32 pts were enrolled and received treatment. The median age was 56 years old (range:
(95% CI, 93.3% - 100%), which is also significantly higher than that of 18F-FDG PET/CT 20-79 years), and 46.9% of pts had papillary renal cell carcinoma (pRCC). Most pts
(sensitivity, 82.0%, 95% CI, 75.4% - 88.7%, P＜0.001; specificity, 11.4%, 95% CI, 2.0% - 20.7%, P (81.3%; 26/32) had a ECOG PS score of 1. At the data cutoff in December 2024, with a
＜0.001).The SUVmax of ccRcc lesions (n = 568) were 12.6 6 11.7 for 68Ga-NY104 versus 7.5 median follow-up of 11.9 months (95% CI 9.3-14.5), the ORR was 54.5% (95% CI 32.2-
610.5 for 18F-FDG (P＜0.001). The TBR is also higher (15.7 6 14.6 v.s. 4.8 65.5, P＜0.001). A 75.6), and the DCR was 100% (95% CI 84.6-100.0). The median PFS was 20.8 months
significant SUVmax difference between ccRCC and non-ccRCC lesion was noted for 68Ga- (95% CI 12.0-29.6). Adverse events (AEs) of any grade occurred in 90.6% of pts, with the
NY104 (12.6 6 11.7 v.s. 1.2 61.0, P＜0.001), which is not true for 18F-FDG (7.5 610.5 v.s. 6.5 most common being proteinuria (40.6%), mucositis and hypertension (37.5%), and
63.5, P = 0.061). Conclusions: 68Ga-NY104 PET/CT is a promising tool with high diagnostic anemia, increased creatinine, elevated transaminases (18.8%), glutamic-pyruvic
efficacy in patients with metastatic ccRCC. It is better than 18F-FDG PET/CT in both sensitivity
transaminase increased and hematuria (15.6%), and hypercholesterolemia (12.5%).
and specificity. Clinical trial information: NCT05879471. Research Sponsor: National Natural
Science Foundation of China; No. 82202218. Grade 3 treatment-related AEs (TRAEs) occurred in 15.6% of pts, with no treatment-
68 18
related deaths. Treatment was suspended in 18.8% (6/32) and 12.5% (4/32) of pts due to
Diagnostic efficacy of Ga-NY104 and F-FDG PET/CT. TRAEs associated with everolimus and anlotinib, respectively. Conclusions: This study
Sensitivity 95% CI Specificity 95% CI demonstrates that anlotinib combined with everolimus is an effective and tolerable first-
Lesion-level (n=677) 68Ga-NY104 96.6% 95.2%-98.1% 99.1% 97.3%-100% line therapy for advanced nccRCC, achieving a high ORR and prolonged PFS, with
18F-FDG 77.8% 74.4%-81.2% 5.5% 1.2%-9.8% manageable toxicity. These findings provide critical evidence supporting the use of this
Region-level (n=172) 68Ga-NY104 98.4% 96.3%-100% 97.7% 93.3%-100%
18F-FDG 82.0% 75.4%-88.7% 11.4% 2.0%-20.7% novel combination in nccRCC. Survival follow-up is ongoing, and further validation in
larger, multi-center randomized trials is warranted. Clinical trial information:
NCT05124431. Research Sponsor: None.

4526 Poster Session 4527 Poster Session

Anlotinib combined with sintilimab as first-line treatment in patients with Investigation of tumor-associated macrophages (TAMs) and therapeutic
advanced non-clear cell renal cell carcinoma (nccRR): Preliminary results resistance to immune checkpoint inhibitors (ICI) through single-cell anal-
from an exploratory prospective multicentre clinical study. First Author: Pei ysis of renal cell carcinoma (RCC). First Author: Soki Kashima, Center of Molecular
Dong, Department of Urology, State Key Laboratory of Oncology in South China, and Cellular Oncology, Yale Cancer Center, Yale School of Medicine, New Haven, CT
Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-Sen University Background: RCC is characterized by a tumor microenvironment (TME) enriched in TAMs,
Cancer Center, Guangzhou, China which suppress antitumor immunity in RCC. We conducted a comprehensive dissection of
Background: Non-clear cell renal cell carcinoma (nccRCC) accounts for approximately the TME using pre- and post-ICI treatment samples to identify specific TAM populations
25% of all kidney cancers, however, the effect of systemic chemotherapy is limited. We associated with ICI treatment resistance in RCC. Methods: A total of 70 tumor samples (58
report the first results of a single-arm, phase 2 study (NCT05220267) evaluating the clear cell and12 non-clear cell) were collected from 63 patients with advanced RCC, in-
efficacy and safety of anlotinib (a multi-target tyrosine kinase inhibitor) combined with cluding 9 patients who were untreated, 10 patients who received non-ICI-based systemic
sintilimab (a monoclonal antibody against programmed cell death protein 1) as first-line therapies, and 44 patients who received ICI-based therapies. We excluded 17 patients with
treatment in patients with advanced nccRCC. Methods: Patients with histologically stable disease from the 44 patients and analyzed 29 samples prior to (n = 15) or after
confirmed advanced nccRCC and measurable disease per RECIST v1.1 who had not exposure to (n = 14) ICI-based therapies (mono-ICI, n = 11; ICI + ICI, n = 11; ICI + VEGFi, n = 6;
previously received systemic therapy were received anlotinib (12 mg qd, d1-14, repeated other, n = 1) from 27 patients. We performed single-cell RNA-sequencing (scRNA-seq; 10x
every 21 days) plus sintilimab (200 mg IV Q3W) till disease progression or intolerant Genomics) on all 70 samples and established a comprehensive transcriptomics atlas of the
toxicity. The primary endpoint is progression-free survival (PFS); secondary endpoints RCC TME. We utilized non-negative matrix factorization (NMF) to identify interpretable gene
programs for TAMs, comparing responders (R) (n = 18; complete or partial response) with
include objective response rate (ORR), disease control rate (DCR), overall survival (OS)
non-responders (NR) (n = 11; progressive disease) to ICI-based therapies according to the
and safety. Results: From April 2022 to January 2024, 44 patients were enrolled with a
best response based on RECIST. P-values from Wilcoxon signed rank test are reported.
median age of 46 years (range: 18-79), 13 (29.5%) had Fumarate deficient RCC, 10 Results: 443,337 high-quality viable cells were annotated to lymphoid, myeloid, tumor,
(22.7%) had Papillary RCC, 9 (20.5%) had TFE3 rearranged RCC and 12 (27.3%) were endothelial, or fibroblast compartments, capturing the RCC TME landscape. Among TAMs,
unclassified. Among these participants, 44 patients were evaluable. 95.5% were IMDC we discovered underlying gene programs through NMF analysis, including “antigen pre-
intermediate- or poor-risk, 72.7% had prior nephrectomy and 97.7% had synchronous sentation”, “S100A8/9 inflammatory”, “stress response”, “C1Q/APOE/TREM2”, “CD163/
metastatic disease. ORR and DCR were 56.8%（95%CI 41.6-72.1）and 86.4%（95%CI MRC1”, “hypoxia”, “interferon-stimulated genes”, and “LILRB/SIGLEC10” programs. We
75.8-96.9）, respectively.$1 and ＜1 Combined Positive Score of PD-L1 expression identified that the LILRB/SIGLEC10-TAM subcluster was significantly increased in frequency
were observed in 50% (22/44) and 38.6% (17/44) patients respectively, and the ORR was in NR compared to R (p = 0.005). Notably, the significant increase in this program in NR was
72.7% (95%CI:52.2-92.9) and 41.2% (95%CI: 15.1-67.3) in the two groups. As of No- also observed in pre-treatment only samples (p = 0.014), suggesting a primary mechanism
vember 13, 2024, median follow-up time was 17.5m (95%CI 14.9-20.1). The median PFS of resistance. This population was characterized by the expression of immune suppressive
was 13.6m (95%CI 8.6-18.6). Treatment-related grade 3/4 adverse events were observed LILRB1/2/3 genes, together with significant upregulation of the macrophage checkpoints
in 22.7% (10/44) of the patients, encompassed proteinuria (3 patients, 7%), hypona- SIGLEC10 (a recently discovered “don’t eat me signal” receptor) and VISTA (an immune
tremia (2 patients, 4%), hypertension (1 patients, 2%), hepatic insufficiency (1 patients, checkpoint) compared to other TAMs (p , 2.22E-16, for each). Conclusions: Our com-
2%), fatigue(1 patients, 2%), rash (1 patients, 2%), decreased lymphocyte count (1 prehensive dissection of the RCC TME reveals an association between TAM population with
patients, 2%). Neither unexpected safety signals nor treatment-related death occurred. an immunosuppressive gene program and ICI resistance through analysis of a large scRNA-
Conclusions: Our results showed promising efficacy and acceptable toxicity of anlotinib seq dataset. This study provides immunobiological insights into potential therapeutic
plus sintilimab for patients with advanced nccRCC. Clinical trial information: targets for next-generation combination therapy with ICIs, offering a foundation for un-
NCT05220267. Research Sponsor: None. derstanding treatment evolution in RCC. Research Sponsor: Kohlberg Chair at Harvard
Medical School and the Trust Family, Michael Brigham, Pan Mass Challenge and Loker
Pinard Funds for Kidney Cancer Research.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GENITOURINARY CANCER—KIDNEY AND BLADDER 335s

4528 Poster Session 4529 Poster Session

Assessment of time-to-treatment-failure (TTF) as a surrogate endpoint for Phase 1 results of Oncobax-AK in combination with ipilimumab/nivolumab
overall survival (OS) to immune checkpoint inhibitor (ICI) regimens in in advanced clear cell renal cell carcinoma (ccRCC; NCT05865730). First
metastatic renal cell carcinoma (mRCC): Findings from an IMDC analysis. Author: Lisa Derosa, Gustave Roussy Cancer Campus (GRCC), ClinicObiome, Villejuif,
First Author: Zachary Yochum, Yale Cancer Center, New Haven, CT France
Background: In Phase III trials for mRCC, OS is a gold standard primary endpoint. Background: Patients with advanced solid tumors who respond to immunotherapy are
However, for ICI-based regimens, this requires extended follow-up times, resulting in more likely to have Akkermansia spp (Akk) in their stools, whereas its absence (45%) is
higher costs and delayed drug approvals. Identification of surrogate or intermediate associated with worse outcome, regardless of other prognostic factors. Here, we report the
endpoints for OS would be beneficial in addressing these challenges. In the current safety and efficacy findings from a Phase I trial evaluating Oncobax-AK live biotherapeutics
study, we investigated 6-month TTF as a potential intermediate endpoint (IE) for OS in in combination with nivolumab/ipilimumab in intermediate- and poor-risk IMDC ccRCC
mRCC. Methods: We included all patients from the International mRCC Database patients who tested negative for stool Akk. We report first results of cohort 1.
Consortium (IMDC) who received ICI-based regimens from 2013 to 2023. TTF was Methods: Intermediate- and poor-risk IMDC advanced ccRCC patients were screened
defined from ICI start until drug cessation or death or censored at date of last follow-up. across 4 centers in France for Akk using a qPCR stool test designed to detect specific Akk
The cohort was divided into 10 equal sub-cohorts based on the decile disease risk strains (SGB9226/SGB9228). Patients who tested negative for stool Akk received either 1X
scores, calculated using multivariable Cox regression for OS, considering all relevant (cohort 1) or 6X oral capsules (cohort 2) of Oncobax-AK (a specific strain of SGB9228,
covariates (IMDC risk groups, presence of bone, brain, or liver metastases, histology, age, p2261) for one week as monotherapy, followed by a combination treatment with nivolumab/
ipilimumab administered continuously until progression in cohorts 1 and 2, respectively. The
prior nephrectomy, ICI type, and year of ICI initiation). For these sub-cohorts, we used
primary endpoints were safety, pharmacodynamics of Oncobax-AK, and overall response
Kaplan-Meier methods to determine 18-month OS and event-free rates for 6, 9, and 12-
rate (ORR) according to RECIST 1.1. Blood and stool samples were collected at multiple time
month TTF. We then performed linear regression of stratum-specific 18-month OS
points for multi-omics analyses. Results: Among the 29 patients screened for Akk, 11 (38%)
against stratum-specific 6-month (and 9- and 12-month) TTF. In the landmark analysis, did not harbor gut Akk and were eligible. Two patients declined to continue. Ultimately, 9
OS was calculated starting at 6 months after therapy initiation, excluding patients who patients (8 intermediate and 1 poor IMDC risk group) were enrolled in cohort 1, with a median
died or had follow-up of less than 6 months. Results: The IMDC cohort consisted of age of 57 years (n = 9: 1 female and 8 males). No serious adverse events related to Oncobax-
1667 patients with a median age of 63 years and 83% had clear cell histology. Median AK was observed. Only one patient experienced a serious immunotherapy-related adverse
follow-up was 15.4 months (IQR: 7.1-28.6). Across the 10 sub-cohorts, 6-month TTF event that led to treatment withdrawal. After a median follow-up of 14.9 months, all patients
accounted for 76% of the variance in 18-month OS (R2= 0.76, 95% CI: 0.26-0.87). Similar are still alive, with 4 showing a partial response (PR) for more than 6 months (including 2
patterns were seen for 9 (R2= 0.65, 95% CI: 0.11-0.81) and 12-month TTF (R2= 0.64, 95% patients with ongoing PR for over 15 months, continuing the intervention). One patient had
CI: 0.09-0.80). In the 6-month landmark analysis (evaluable n = 1255), mRCC patients stable disease (SD), 3 had progressive disease (PD), and 1 was not evaluable (NE). Among
experiencing treatment failure at 6 months had an 18-month OS (i.e. 12 months after the evaluable patients, the ORR was 50%. Multi-omics analyses revealed that Oncobax-AK
landmark time) rate of 68% (95% CI: 63%-72%), compared to 92% (95% CI: 90%-94%) for remains detectable over time and induced changes in the microbiome, leaky gut markers,
those without treatment failure (adjusted HR: 2.74, 95% CI 2.15-3.49). Conclusions: 6- and immune and metabolite profiles. These changes suggest potential pharmacodynamic
month TTF was predictive of 18-month OS in mRCC patients. These findings suggest biomarkers, reflecting the biological effects of Oncobax-AK. Conclusions: This is the first
that 6-month TTF may be a promising intermediate endpoint for OS and provides trial targeting patients lacking Akkermansia spp. in their stools. Oncobax-AK, combined with
supportive evidence of future validation in prospective studies. Research Sponsor: None. Ipilimuab/Nivolumab in intermediate- and poor-risk IMDC ccRCC, appears safe and may
modulate gut microbiota, inflammation, metabolites, and immune profiles. This potentially
mimics a responder’s profile and leads to clinical and radiological benefits in some patients.
Based on these promising results, enrollment in cohort 2 (6X) is ongoing and has now been
expanded to include non-small cell lung cancer patients. Clinical trial information:
NCT05865730. Research Sponsor: None.

4530 Poster Session 4531 Poster Session

Co-expression network-based analysis of gene programs contributing to Frailty assessment in patients with advanced/metastatic renal cell carci-
immune checkpoint inhibitor (ICI) resistance in renal cell carcinoma (RCC). noma using routine data collected during treatment with tyrosine kinase
First Author: Ro Malik, Center of Molecular and Cellular Oncology, Yale Cancer Center, inhibitors in the STAR trial. First Author: Sophie Trotter, Weston Park Cancer Centre,
Yale School of Medicine, New Haven, CT Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
Background: There is an unmet need to characterize molecular drivers of resistance to ICIs in Background: The STAR trial (ISRCTN06473203) was a phase 2/3 randomised controlled non-
the RCC tumor microenvironment, especially in cell-type-specific contexts. To address this, inferiority trial (N=920) which assessed whether treatment breaks from tyrosine kinase inhibitors
we identified coherent gene programs in each cell type using scRNA-seq data to uncover an (TKIs, oral sunitinib/pazopanib) could be safely used in locally advanced/metastatic renal cell
immune cell phenotype associated with ICI resistance. Methods: We performed per-cell-type carcinoma (RCC). We aimed to determine if it was possible to obtain a frailty index (FI) which could
weighted coexpression network analysis on scRNA-seq data of 443,337 cells from 70 RCC be linked to toxicity and clinical outcomes from routinely collected large trial data, using the STAR
tumor samples, mostly from patients (pts) who received ICI (S. Kashima, ASCO, 2024), to trial as an example. Methods: We developed the FI using Rockwood’s accumulation of deficits
construct mutually co-expressed gene sets (modules) and their cell-level expression. We methodology. STAR data was initially searched for variables measuring baseline health problems
compared module expression scores between responders (R, complete or partial response) and functional limitations. Variables were excluded if there was a significant proportion of missing
values (.10%); if too rare or too common (,1%, .80%); or if significantly correlated with another
and non-responders (NR, progressive disease) with the Wilcoxon rank-sum test plus FDR
variable (r.0.95). Variables were coded from 0 (no deficit) to 1 (full deficit); these were summed
correction, and ran tests with permuted labels to rule out statistical artifacts. For validation,
then divided by the number of variables assessed to give the FI. Frailty thresholds were adopted in
we performed this analysis on another scRNA-seq cohort from a phase II trial (HCRN GU16- line with the literature: Not Frail (FI#0.08); Pre-Frail (FI 0.08-0.24); Frail (FI$0.25). Results: Of 57
260; NCT03117309), and further, we used CoxPH survival analysis on module signature variables screened, 35 variables remained for inclusion in the FI. 50 participants missing .20% of
scores computed by GSVA to assess the prognostic role of module expression in bulk RNA- FI variables were excluded, leaving n=870. FI scores ranged from 0 to 0.43 (median 0.15, IQR 0.10-
seq samples from the IMmotion 150, CheckMate 009/010/025, and Javelin 101 trials, after 0.22). Kaplan-Meier survival analysis showed a statistically significant difference by FI for overall
stratifying them into immune-high (top 50%) and immune-low (bottom 50%) groups according survival (OS) (FI$0.25: HR 2.48, p,0.001, 95% CI 1.89-3.26). In multivariate Cox proportional
to CIBERSORTx-inferred total immune infiltration. Results: Analysis of scRNA-seq (two hazards regression, FI remained a statistically significant risk factor for OS (HR 1.41, P=0.047, 95%
independent datasets) from ICI-treated RCC tumors revealed a module of robustly co- CI 1.00-1.99); other statistically significant variables included age group, sites of metastatic
expressed ribosomal and translation-associated genes that was significantly upregulated in disease, IMDC and MOTZER scores. Toxicity was assessed over the first 6 months, while all
NR in immune cells (but not tumor or stromal cells), including macrophages, CD4+ and CD8+ participants were treated with continuous TKIs. The most common severe toxicities (G3+) were
T, NK, and B cells (p-values ranging from 1e-20 to 1e-200). Analysis of an independent scRNA- hypertension (200/870), hepatobiliary disorders (97/870) and fatigue (63/870). Time free of severe
seq cohort (HCRN GU16-260) also yielded a module of mutually co-expressed ribosomal toxicity was statistically significantly shorter for frail participants. Conclusions: Amongst STAR
proteins only in immune cells that was upregulated in NR (immune p , 1e-20, non-immune trial participants with advanced/metastatic RCC treated with TKIs, frailty was a statistically
p . .05). Finally, validation in large-scale bulk RNA-seq data from clinical trials shows that for significant risk factor for poorer survival and for shorter time to severe toxicity. It is noted that the
patients receiving any ICI-based therapy (nivolumab, avelumab + axitinib, or atezolizumab +/- eligibility criteria for STAR included a baseline PS of ECOG 0-1 which will have excluded many
bevacizumab), module signature scores were significantly associated with worse PFS only in frailer patients. This data shows that it is feasible to use routinely collected trial data to create a
the CoxPH analysis of immune-high samples (HR = 1.65, 95% CI, 1.18–2.3, p , .005), while clinically meaningful FI and this approach could be applied to other trial datasets. Clinical trial
immune-low samples showed no effect (HR = 0.93, 95% CI 0.71–1.25, p . .5), concordant information: ISRCTN06473203. Research Sponsor: UK National Institute for Health and Care
Research (NIHR).
with findings from the scRNA-seq analysis. Those receiving a TKI only (sunitinib) exhibited no
association, regardless of immune infiltration. Conclusions: Though individual ribosomal Toxicity-free survival, by frailty group.
proteins have been found to be prognostic for RCC, a cell-type-specific module-level analysis Toxicity FI group HR p-value 95% CI
elucidated a link between a coherent translation program in immune cells and resistance and Severe toxicity (G3+) Intermediate 1.00 0.960 0.83-1.21
poor PFS for pts receiving ICI specifically, contributing to a model of ICI-resistant molecular Frail 1.28 0.025 1.03-1.59
phenotypes in RCC. Research Sponsor: National Cancer Institute/U.S. National Institutes of All toxicity (G1+) Intermediate 1.11 0.031 1.01-1.22
Health; 1R37CA279822. Frail 1.08 0.201 0.96-1.20

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
336s GENITOURINARY CANCER—KIDNEY AND BLADDER

4532 Poster Session 4534 Poster Session

International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) Early detection of renal cell carcinoma: A novel cfDNA fragmentomics-based
classification and regression tree analysis to characterize objective re- liquid biopsy assay. First Author: Yulu Peng, Sun Yat-sen University Cancer Center;
sponse rates (ORR) in metastatic renal cell carcinoma (mRCC). First Au- State Key Laboratory of Oncology in South China, Collaborative Innovation Center of
thor: Martin Zarba, Arthur J.E. Child Comprehensive Cancer Centre, Calgary, AB, Canada Cancer Medicine, Guangzhou, China
Background: Therapies for mRCC have evolved significantly, making treatment decisions Background: Renal cell carcinoma (RCC) is a leading cause of cancer-related mortality,
more complex. We used machine learning (ML) to identify whether this could help identify with a rapidly rising global incidence. Early detection greatly improves the outcomes of
subgroups of patients who have a high probability of response. Methods: Patients from IMDC RCC, yet current diagnostic methods have limitations in sensitivity, specificity, and
were identified and a ML classification and regression tree analysis was conducted, in which accessibility. This study develops and evaluates a cfDNA fragmentomics-based liquid
we grew a complex tree up to a depth of 30 with a minimum node split size of 2 with no biopsy integrated with machine learning as a non-invasive and scalable tool for RCC
constraints on the cost-complexity parameter. The resulting tree was pruned according to the early detection. Methods: This case-control cohort study recruited 442 participants
cost-complexity parameter that minimized the leave one out cross-validated error rate and
(223 RCC patients and 219 non-cancer controls, including healthy individuals and those
had a minimum bucket size of 25 patients. Results: 2,549 patients were included, 73.2% male,
with benign renal conditions) at a single cancer referral center from December 2021 to
13.5% non-clear cell histology, 70.3% nephrectomy, and 19.4%, 54.2%, and 26.4% had fa-
vorable, intermediate and poor IMDC risk respectively. 1L treatment regimens consisted of December 2023. Plasma-derived cfDNA underwent low-pass WGS (5X coverage), and
VEGF inhibitors (51.5%), IO-IO combinations (32.3%), and IO-TKI combinations (16.2%). The three fragmentomics features—copy number variation (CNV), fragment size ratio (FSR),
ORR was 36.0% overall, with 29.6% for VEGF inhibitors, 39.1% for IO-IO, and 50.2% for IO-TKI and nucleosome footprint (NFP)—were extracted. A stacked ensemble machine learning
combinations. ML identified 5 hierarchal variables —therapy type, prior nephrectomy (PN), lung model was trained on 280 participants and validated on 162 independent participants.
metastasis (LM), other metastases, and age— that divided patients into 7 different categories Performance was assessed using area under the curve (AUC), sensitivity, and specificity.
with different response probabilities (see Table). VEGF therapy showed the poorest response, Results: The ensemble model achieved an AUC of 0.9656 in the validation cohort, with
with no additional variables able to predict response. The best ORR was observed in patients sensitivity and specificity of 90.5% and 93.8%, respectively. Stratified analyses dem-
treated with IO-TKI and PN; and in those treated with IO-IO, PN, and only lung metastasis. onstrated consistent performance across RCC stages, histological subtypes, and
Factors associated with poorer responses included non-clear cell histology, older age, bone Fuhrman grades, with sensitivities of 87.8% for Stage I RCC and 100% for Stage IV RCC.
and liver metastases, poor performance status, elevated neutrophils, and poor IMDC risk score. Additionally, the model effectively differentiated malignant RCC from benign renal
Conclusions: This large-scale ML analysis identified five key clinical variables that predict conditions, further validating its clinical utility. Stability evaluations confirmed the
treatment response in mRCC, with treatment type emerging as the primary determinant. These model’s robustness across diverse sample types, storage conditions, and processing
results suggest that treatment selection for mRCC could potentially be optimized by con- scenarios, underscoring its potential applicability in routine clinical practice.
sidering these hierarchical variables, though further validation is needed. Research Sponsor: Conclusions: This cfDNA fragmentomics-based liquid biopsy represents a highly
None.
sensitive, specific, and non-invasive approach for the early detection of RCC. Its robust
ML analysis results: Groups of patients and associated outcomes. performance across diverse clinical scenarios highlights its potential to enhance current
Risk Groups N (%) ORR (%) Odd Ratio TTNT 18-month survival RCC diagnostic workflows, facilitate timely interventions, and improve patient out-
1) VEGF 1313 (51.5) 29.6 Ref. 9.4 (8.6-10.3) 0.62 (0.59-0.65) comes. Future integration of this method into clinical practice could address critical
2) IO-IO or IO-TKI and no PN 443 (17.4) 35.0 1.28 (1.02-1.60) 10.2 (8.8-11.3) 0.59 (0.55-0.65) gaps in RCC management, providing substantial benefits in early detection and per-
3) IO-IO and PN sonalized care. Research Sponsor: National Natural Science Foundation of China;
a) No LM 137 (5.4) 29.2 0.98 (0.66-1.43) 17.2 (10.6-30.1) 0.85 (0.78-0.92)
b) LM and other met 267 (10.5) 43.8 1.87 (1.42-2.44) 13.0 (10.1-20.5) 0.78 (0.72-0.83) Science and Technology Projects in Guangzhou; The China National Postdoctoral
c) Only LM 85 (3.3) 60.0 3.56 (2.28-5.63) 39.2 (14.4-NA) 0.93 (0.87-0.99) Program for Innovative Talents.
4)IO-TKI and PN
a) Age 70+ 78 (3.2) 43.6 1.84 (1.15-2.91) 35.7 (19.8-NA) 0.80 (0.71-0.91)
b) Age < 70 226 (8.9) 58.4 3.34 (2.50-4.47) 24.7 (22.4-36.4) 0.88 (0.84-0.93)
Overall 2549 36.0 11.5 (10.7-12.2) 0.68 (0.67-0.70)

4535 Poster Session 4536 Poster Session

Efficacy and safety of neoadjuvant toripalimab plus axitinib in renal cell First-line benmelstobart plus anlotinib versus sunitinib in advanced renal
carcinoma with tumor thrombus: A combined analysis of two phase II trials. cell carcinoma: Subgroup analysis from the phase 3 ETER100 trial. First
First Author: Liangyou Gu, Chinese PLA General Hospital, Beijing, China Author: Xinan Sheng, Key Laboratory of Carcinogenesis and Translational Research
Background: Tumor thrombectomy for renal cell carcinoma (RCC) with tumor thrombus (Ministry of Education/Beijing), Department of Genitourinary Oncology, Peking Uni-
(TT) is associated with high morbidity and mortality. Two trials have demonstrated the versity Cancer Hospital & Institute, Beijing, China
efficacy and safety of neoadjuvant toripalimab plus axitinib in patients with TT. Here, we Background: Dual immune checkpoint inhibitors (ICIs) or ICIs plus VEGF-directed
present the pooled analysis results. Methods: These two phase II trials (NEOTAX and therapies, have been approved as first-line treatment in patients (pts) with advanced
NCT04118855) shared similar target populations. Patients with clear cell RCC (ccRCC) and renal cell carcinoma (RCC). The phase 3 ETER100 trial showed that benmelstobart (PD-L1
TT received up to 12 weeks of toripalimab plus axitinib before surgery. The primary blockade) plus anlotinib improved the progression-free survival (PFS) (19.0 months
endpoint was the downstaging rate of TT based on the Mayo classification. For level 0 9.8 months) and objective response rate (ORR) (71.6% vs 25.1%) of advanced clear cell
thrombus, the downstaging defined as: for right RCC, the proximal end of TT retracted from RCC (ccRCC) pts significantly. Pts with factors, such as intermediate-poor International
the main renal vein to branches of the renal vein; for left RCC, the superior mesenteric Metastatic Renal Cell Carcinoma Database Consortium (IMDC) risk, liver metastasis, or
artery (SMA) was an anatomical landmark, the TT retracted from lateral to the abdominal bone metastasis were considered to have a poor prognosis. Here we report PFS and ORR
aorta to the level of SMA or from the main renal vein lateral to SMA or the branches of the in clinically relevant subgroups. Methods: ETER100 (NCT04523272) was a multicentre,
renal vein. Secondary endpoints included response rate, change in thrombus length, randomised, open-label, controlled phase 3 trial conducted at 37 sites in China. Eligible
surgical morbidity, progression-free survival (PFS), overall survival (OS), and safety. patients were randomly assigned in a 1:1 ratio using stratified block randomisation to
Results: A total of 40 patients were enrolled, 4 (10%), 4 (10%), 14 (35%), 7 (17.5%), 11 receive benmelstobart plus anlotinib or sunitinib. Randomisation was stratified according
(27.5%) patients had level 0, I, II, III, IV tumor thrombus, respectively. After 12 weeks to the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) risk
treatment, 45.0% (18/40) patients experienced a reduction in TT level, one patient (2.5%) (favourable [score of 0], intermediate [score of 1-2], or poor risk [score of 3-6]). PFS
had an increase in Mayo level. Thirty-six (90%) patients exhibited shrinkage in TT length, analyses of clinically relevant subgroups were assessed using Kaplan-Meier method and
with a median reduction of 1.9 cm (IQR: 0.9 to 3.7 cm). According to the RECIST criteria, the 95% CIs of response rate were calculated with the Clopper-Pearson method.
the objective response rate and disease control rate of overall tumor was 37.5% and 97.5%. Results: Overall, 527 pts received the trial treatments (264 in the benmelstobart-anlotinib
In total, 35 patients underwent radical nephrectomy with IVC thrombectomy, including 16 group and 263 in the sunitinib group) and were evaluated for efficacy. A total of 454 (86%)
open, 2 laparoscopic and 17 robotic. No patient had a surgery delay due to treatment- pts had intermediate-poor IMDC risk, 62 (12%) pts had liver metastasis and 111(21%) had
related adverse events (TRAEs). 54.3%(19/35) patients experienced changes in surgical bone metastasis. Data cutoff for the interim analysis occurred on January 31, 2024. The
strategy compared with planned surgery. Median operation time was 300 min (IQR: 180- median follow-up was 22.8 months. Benmelstobart plus anlotinib significantly improved
420 min). Median estimated blood loss was 700 ml (IQR: 300-1800 ml). The postoperative PFS of subgroups with intermediate-poor IMDC risk (17.0 months [95% CI 14.0-20.1] vs
complication rate was 60% (21/35), including three (8.6%) major complications and one 9.7 months [8.0-11.3], HR 0.55, 95% CI 0.43-0.72; p , 0.0001), liver metastasis
(2.9%) postoperative death. The most common TRAEs included hypertension (35.0%), (11.9 months [95% CI 5.8-NE] vs 5.4 months [1.5-6.7], HR 0.44, 95% CI 0.23-0.85; p ,
proteinuria (35.0%), fatigue (27.5%), and diarrhea (22.5%). Grade $3 adverse events were 0.0121), or bone metastasis (19.5 months [95% CI 16.5-27.2] vs 8.3 months [4.2-19.8], HR
reported in 27.5% (11/40) of patients, no patients experienced Grade 4 or 5 TRAEs. With a 0.52, 95% CI 0.30-0.89; p , 0.0154). ORR of benmelstobart-anlotinib group was sig-
median follow-up of 32.2 (IQR: 27.0-35.5) months, the median PFS and OS were not nificantly higher in the subgroups with intermediate-poor IMDC risk (70.0% [95%CI, 63.6-
reached. The estimated PFS rate at 1 year was 87.5% (95% CI, 72.4% to 95.3%). The PFS (P 75.9] vs 21.6% [16.4-27.5]), liver metastasis (60.0% [95%CI, 42.1-76.1] vs 7.4% [0.9-24.3])
= 0.20) and OS (P = 0.44) were similar between responders (partial response) and non- and bone metastasis (63.2% [95%CI, 49.3-75.6] vs 16.7% [7.9-29.3]). Conclusions: The
responders (stable disease or progressive disease). Conclusions: Toripalimab in com- RCC pts with a poor prognosis such as intermediate-poor IMDC risk, liver metastasis and
bination with axitinib downstages TT level in a significant proportion of patients leading to bone metastasis could significantly benefit from benmelstobart plus anlotinib. Clinical
simplification in the procedure of surgery. Clinical trial information: NCT04118855. trial information: NCT04523272. Research Sponsor: None.
Research Sponsor: None.

Visit [Link] and search by abstract for the full list of abstract authors and their disclosure information.
 GENITOURINARY CANCER—KIDNEY AND BLADDER 337s

4537 Poster Session 4538 Poster Session

Ongoing phase 1/2 trial of the hematopoietic progenitor kinase 1 (HPK1) Genomic and proteomic predictors of sites of metastases in renal cell
inhibitor NDI-101150 as monotherapy or in combination with pembrolizu- carcinoma. First Author: Clara Steiner, Dana-Farber Cancer Institute, Boston, MA
mab: Clinical safety and efficacy update in clear cell renal cell carcinoma Background: Among patients with renal cell carcinoma (RCC), the most common sites
(ccRCC). First Author: David A. Braun, Center of Molecular and Cellular Oncology, Yale of metastasis are lung, lymph nodes, and bone. While some sites of metastases are
University, New Haven, CT associated with better cancer-specific outcomes than others, the underlying biology of
Background: NDI-101150 is a potent and selective oral inhibitor of HPK1, a serine/threonine metastatic organ tropism is not well understood. We performed genomic and proteomic
kinase that acts as a negative regulator of immune cell function. Pre-clinically, NDI-101150 analyses to investigate the biological underpinnings of different metastatic sites in RCC.
can enhance immune cell function, leading to potent anti-tumor immunity. Methods: NDI- Methods: Institutional cohorts of patients with metastatic RCC from the Dana-Farber
101150 is currently being investigated in a first-in-human, multi-center, open-label, phase 1/2 Cancer Institute (DFCI) were analyzed using a next-generation tumor somatic mutation
trial (NCT05128487) in patients with advanced solid tumors, as a monotherapy (50–200 mg assay (n = 633) and with a highly multiplexed plasma proteomics assay (n = 258). Data
once daily [QD]) or in combination with pembrolizumab (50–100 mg QD NDI-101150 + 200 mg were clinically annotated for sites of RCC metastasis. Genomic analyses were performed
Q3W pembrolizumab). Results: As of 20 November 2024, 106 patients were dosed [NDI- using a two-sided Fisher’s exact test on the cBioPortal platform at DFCI with pairwise
101150 monotherapy (n = 94) or NDI-101150 + pembrolizumab (n = 12)]. The tumor types comparison of patients with versus without metastases to lung, liver, brain, bone,
were RCC (n = 38), NSCLC (n = 17), gastric/GEJ (n = 12), and other solid tumors (n = 39). We
adrenal, and lymph nodes. The Benjamini-Hochberg method was applied for FDR-
report here updated safety data in all patients from monotherapy and combination arms (n =
adjusted q-values. Exploratory proteomic analyses were performed using logistic re-
106), and efficacy data in patients with ccRCC (n = 29) receiving NDI-101150 monotherapy.
NDI-101150 monotherapy was generally well tolerated, with 150 mg identified as the
gression for each metastatic site with multivariate adjustment for other sites of me-
maximum tolerated dose. The most common treatment-related adverse events (TRAEs) of tastasis. For each metastatic site, the top five associated proteins were selected to
any grade were nausea (39%), diarrhea (35%), vomiting (29%), fatigue (27%), and anemia build a multivariate model to predict the presence of each metastatic site. Bootstrapping
(11%). Grade $3 TRAEs occurred in 13 (14%) patients, of which only 1 (1%) patient with R = 1,000 was employed for the assessment of model performance.
experienced a grade 4 TRAE. The safety profile was comparable in the combination cohort, Results: Tumor genomic alterations in SETD2 (q-value = 0.004) and CDKN2A (q-value =
with 2 (17%) patients experiencing Grade $3 TRAE. 20 of the 29 ccRCC patients who received 0.04) were associated with lung and lymph node metastases, respectively. Logistic
50, 100, 140, or 150 mg of NDI-101150 monotherapy were response-evaluable. The objective regression analyses of proteomic data were used to identify circulating proteins with the
response rate was 15.0% [CR, n = 1 and PR, n = 2]. Clinical benefit rate (CR + PR + strongest associations with the sites of metastases. For instance, circulating collagen
SD $6 months) was 25%, which includes 2 patients who experienced durable SD for alpha-1(IX) chain (CO91A) and relaxin receptor 1 (RXFP1) were the top circulating
~9 months and ~25 months. The disease control rate (CR+PR+SD) was 60%. The ccRCC proteins associated with bone metastases, while GGT2 and tenascin were associated
patients had received a median of 2 (1-9) lines of prior therapy. A nearly dose-proportional with liver metastases and matrilysin (MMP-7) was associated with lymph node me-
increase in NDI-101150 exposure was observed at day 1, with steady state achieved by day tastases. Multivariate models using the top five proteins to predict the presence of each
15. At all doses tested, steady state exposures inhibited the pharmacodynamic biomarker metastatic site demonstrated bootstrapped C-statistics from 0.72 to 0.80 for lymph
pSLP76 by . 50% and for a period consistent with preclinical efficacy predictions. To nodes, lung, adrenal, brain, liver, and bone, respectively. Conclusions: We identified
demonstrate proof of biology, a custom 12-plex immunofluorescence panel and the GeoMx genomic and proteomic predictors of organ-tropic metastases in RCC. Next, we will
whole transcriptomic assay were utilized. By day 28, on-treatment tumor biopsy samples validate these findings in independent external cohorts. Research Sponsor: None.
showed immune activation when compared to pre-treatment samples, including an increased
infiltration of activated CD8+ T-cells and dendritic cells. Conclusions: NDI-101150 continues
to demonstrate an acceptable safety profile and encouraging antitumor activity in patients
with ccRCC, supporting continued clinical development of NDI-101150 as monotherapy and in
combination with other agents as a promising next-generation immunotherapy oral small
molecule. Clinical trial information: NCT05128487. Research Sponsor: Nimbus Therapeutics.

4539 Poster Session 4540 Poster Session

Analysis of phase II study of cabozantinib (Cabo) with nivolumab (Nivo) and Efficacy of second line (2L) treatment with tivozanib (Tivo) as monotherapy
ipilimumab (Ipi) in advanced renal cell carcinoma with divergent histologies or with nivolumab (Nivo) in patients (pts) with metastatic renal cell carci-
(RCCdh). First Author: Bradley Alexander McGregor, Dana-Farber Cancer Institute, noma (mRCC) previously treated with an immune checkpoint inhibitor (ICI)
Boston, MA combination of ipilimumab (Ipi)/Nivo or vascular endothelial growth factor
Background: We previously reported on treatment intensification with the combination of Cabo/ receptor-tyrosine kinase inhibitor (VEGFR-TKI)/ICI in the phase 3 TiNivo-2
Nivo/Ipi in 39 patients (pts) with metastatic RCCdh in a multi-center single arm phase II trial with a study. First Author: Alex Chehrazi-Raffle, City of Hope Comprehensive Cancer Center,
starting cabozantinib dose of 40 mg/day (d). Clinical utility was limited with an objective response Duarte, CA
rate (ORR) of 21% and significant treatment related adverse events (TrAEs) (77% $Grade 3 TrAEs). Background: In TiNivo-2, the addition of Nivo to Tivo did not prolong progression-free survival (PFS)
Therefore, we explored the safety and potential efficacy by using a lower starting dose of relative to Tivo alone (Choueiri, Lancet 2024). To assess study outcomes in the context of con-
cabozantinib of 20 mg/d (NCT04413123). Methods: Eligible pts had metastatic RCCdh with ECOG temporary treatment sequencing, a subset of pts treated in the 2L who failed 1L Ipi/Nivo or VEGFR-
performance status of 0-1 and may have received one line of prior therapy excluding immuno- TKI/ICI therapy was evaluated. Methods: Pts were randomized 1:1 to receive Tivo once daily for 21/28
therapy or Cabo. Pts underwent a baseline biopsy and re