Journal of Surgical Oncology 2014;109:533–541

Transplant Versus Resection for the Management of Hepatocellular Carcinoma

Meeting Milan Criteria in the MELD Exception Era at a Single Institution in a
UNOS Region with Short Wait Times

MALCOLM H. SQUIRES III, MD, MS,1 STEVEN I. HANISH, MD,2,3 SARAH B. FISHER, MD,1
CRISTEN GARRETT, BS,2 DAVID A. KOOBY, MD,1 JUAN M. SARMIENTO, MD,4 KENNETH CARDONA, MD,1
ANDREW B. ADAMS, MD, PhD,2 MARIA C. RUSSELL, MD,1 JOSEPH F. MAGLIOCCA, MD,2
STUART J. KNECHTLE, MD, PhD,2 CHARLES A. STALEY III, MD,1 AND SHISHIR K. MAITHEL, MD1*
1
Department of Surgery, Division of Surgical Oncology, Winship Cancer Institute, Emory University, Atlanta, Georgia
2
Department of Surgery, Division of Liver Transplantation, Emory Transplant Center, Emory University, Atlanta, Georgia
3
Department of Surgery, Division of Transplant Surgery, University of Maryland School of Medicine, Baltimore, Maryland
4
Department of Surgery, Division of General and GI Surgery, Emory University, Atlanta, Georgia

Background: Management of hepatocellular carcinoma (HCC) in the Model for End‐Stage Liver Disease (MELD) exception era remains regionally
variable. Outcomes were compared for patients undergoing transplant versus resection at a single institution in a UNOS region with short wait times
for organ availability.
Methods: All patients who underwent resection of HCC from January 2000 to August 2012 and patients who underwent transplant post‐
January 2006, during the Milan Criteria (MC)‐based MELD exception policy for HCC, were identified. Primary outcomes were overall survival (OS)
and recurrence‐free survival (RFS).
Results: Two hundred fifty‐seven patients were analyzed, of whom 131 underwent transplant and 126 underwent resection. All transplant patients
met MC; 45 (36%) resection patients met MC. Median follow‐up time was 30 months. Median wait time to transplant was 55 days; no patients
dropped off the waitlist while awaiting an organ.
Among patients meeting MC, transplant demonstrated significantly greater 5‐year OS (65.7% vs. 43.8%; P ¼ 0.005) and RFS (85.3% vs. 22.7%;
P < 0.001) versus resection. For patients with hepatitis C, transplant (n ¼ 87) demonstrated significantly improved 5‐year outcomes compared to
patients meeting MC who underwent resection (n ¼ 21; OS: 63.5% vs. 23.3%; P ¼ 0.001; RFS: 83.5% vs. 23.7%; P < 0.001).
Conclusion: In a region with short waitlist times for organ availability, liver transplant is associated with improved survival compared to resection for
HCC within MC and should be considered for all patients meeting MC, particularly those with hepatitis C.
J. Surg. Oncol. 2014;109:533–541. ß 2013 Wiley Periodicals, Inc.

KEY WORDS: hepatocellular carcinoma; hepatic resection; liver transplant; Milan Criteria; waitlist time

INTRODUCTION applicable, but is limited by the substantial morbidity and mortality

associated with hepatectomy in patients with significant underlying liver
Hepatocellular carcinoma (HCC) is the third‐leading cause of cancer dysfunction [8]. Given the limited number of organs available for
mortality worldwide, and the annual incidence of both new diagnoses and transplant and the risk of disease progression while awaiting an organ,
deaths attributed to HCC continues to increase [1,2]. Despite advances in
nonsurgical interventional therapies, the best potential curative treatment
options for HCC remain liver transplantation or hepatic resection [3–5]. Abbreviations: HCC, hepatocellular carcinoma; MC, Milan Criteria; MELD,
Optimal surgical management of HCC patients remains a point of debate, Model for End‐Stage Liver Disease; OLT, orthotopic liver transplantation;
OS, overall survival; RFA, radiofrequency ablation; RFS, recurrence‐free
with practices varying between institutions globally and regionally survival; TACE, transarterial chemoembolization; UNOS, United Network
within the United States. In 1996, Mazzaferro et al. [6] first reported their for Organ Sharing.
improved results after transplantation for patients with HCC meeting
their Milan Criteria (MC), defined as a single lesion <5 cm or 3 lesions Grant sponsor: Katz Foundation; Grant sponsor: National Institutes of
Health; Grant number: UL1TR000454.
each <3 cm in size, with no evidence of macrovascular invasion or
extrahepatic disease on imaging. Numerous studies worldwide, many Presented at the 2013 Annual Meeting of the Society of Surgery for the
Alimentary Tract, Digestive Disease Week, May 2013, Orlando, Florida.
included in a comprehensive 2011 meta‐analysis by the Milan group,
have confirmed the favorable outcomes that can be achieved with Conflict of interest disclosure: None.
transplantation for patients meeting these criteria [7]. In light of these Malcolm H. Squires and Steven I. Hanish shared first authorship.
results, United Network for Organ Sharing (UNOS) guidelines, *Correspondence to: Shishir K. Maithel, MD, Division of Surgical
universally implemented since 2006, have allocated Model for End‐ Oncology, Winship Cancer Institute, 1365C Clifton Road NE, 2nd Floor,
Stage Liver Disease (MELD) exception points to patients on the liver Atlanta, GA 30322. Fax: þ1‐404‐778‐4255. E‐mail: [email protected]
transplant waitlist with a diagnosis of HCC meeting MC. Received 03 October 2013; Accepted 22 November 2013
Liver transplantation for HCC provides the additional benefit of DOI 10.1002/jso.23531
removing the underlying diseased liver, however, it is limited by donor Published online 17 December 2013 in Wiley Online Library
organ availability. Hepatic resection is potentially more widely (wileyonlinelibrary.com).

ß 2013 Wiley Periodicals, Inc.

534 Squires et al.
resection has historically been advocated for most patients with HCC, January 2006. All transplant patients met MC based on preoperative
especially those with no underlying liver disease or with well‐ cross‐sectional imaging. Patients with HCC diagnosed incidentally after
compensated cirrhosis. Given the lack of randomized, prospective transplant on examination of the explant specimen were excluded from
data and the difficulties in comparing retrospective studies with analysis. Waitlist time was calculated as the period of time from the date
heterogeneous cohorts and different inclusion criteria, the question of of active listing with a diagnosis of HCC for which the patient was
whether transplantation or resection provides superior overall survival awarded MELD exception points, to the date of transplantation. For
(OS) and recurrence‐free survival (RFS) for early HCC remains patients who were placed on the organ allocation waitlist prior to the
unclear [4]. MELD exception era implementation in January 2006, wait time was
Wait times for transplant vary widely among reported studies, calculated from January 1, 2006, as this was the date on which they
significantly affecting outcomes and influencing comparisons with received their MELD exception points that facilitated their receipt of a
resection, as a function of patient dropout from the waitlist due to disease liver transplant.
progression or death. In countries and UNOS regions with A comprehensive retrospective chart review was conducted to
comparatively short transplant wait times and greater organ identify patient demographics, clinicopathologic features, and
availability, transplantation may be more feasible and associated with preoperative laboratory values. Preoperative cross‐sectional imaging
better outcomes. While other standards for transplantation such as the was reviewed to determine whether patients who underwent resection
UCSF criteria have been validated, the most widely utilized guidelines met MC based on tumor size and focality. Raw MELD scores and Child‐
for patient selection worldwide are based on the MC. The variance in Pugh scores were calculated retrospectively for each patient, and for
median waitlist times for liver transplantation across UNOS regions is patients in the transplant cohort, these scores were validated with the
considerable, ranging from 27 to 490 days for patients with a MELD pretransplant evaluations from the medical record. Tumor size, presence
score of 19–24 in 2004, prior to the MELD exception era for HCC [9]. of macroscopic or microscopic vascular invasion, tumor differentiation,
Further underscoring this disparity in wait times, the percentage of all and the presence of cirrhosis were confirmed from the surgical pathology
patients underdoing liver transplantation within a wait period of 90 days report.
or less from 1999 to 2008 ranged from 70.9% in Region 3, where the
current study was conducted, to only 42.8% of patients in Region 5 Statistical Analysis
(Fig. 1) [10]. We hypothesized that in a region with short waitlist times
and greater organ availability, transplantation for the treatment of HCC Comparison of perioperative clinicopathologic features between the
in the MELD exception era would be associated with improved OS and transplant and resection cohorts was conducted using chi‐square
RFS as compared to hepatic resection. analysis for categorical variables and the Student’s t‐test for
continuous variables. Kaplan–Meier survival plots and log‐rank
MATERIALS AND METHODS analysis were conducted to compare outcomes for transplantation
versus resection for the primary endpoints of OS and RFS. Survival was
This study was conducted with the approval of the Institutional calculated from the operative date to the date of confirmed death or the
Review Board and all research activities were performed in compliance date of last follow‐up, with those patients lost to follow‐up censored at
with the Health Insurance Portability and Accountability Act of 1996. the date of last contact noted in the medical record or date of confirmed
All patients undergoing curative intent hepatic resection for a diagnosis death. The Social Security Death Index was used to verify survival data.
of HCC at a single academic institution between January 2000 and RFS was calculated from the operative date to the time of confirmed
June 2012 were identified from a prospectively maintained surgical recurrence on cross‐sectional imaging; per Kaplan–Meier methodology,
database. All patients placed on the wait list for orthotopic liver patients without recurrence were censored at their date of death or date of
transplantation (OLT) at our institution who received MELD exception last follow‐up. To account for potential differences in clinicopathologic
points for a known diagnosis of HCC from January 2006 to June 2012 variables between the transplant and resection cohorts, univariate and
were identified, as the allocation policy of awarding MELD exception multivariate Cox regression analyses were performed for OS and RFS.
points for HCC within MC was universally implemented by All variables with a P‐value 0.05 on univariate analysis were included
into the multivariate model for each endpoint. Statistical significance
was defined as a P‐value 0.05. All statistical analysis was performed
with Statistical Package for the Social Sciences 19.0 software (IBM
Corporation, Armonk, NY).

RESULTS
A total of 261 patients were identified. Of the 133 patients with HCC
who underwent transplantation during the MELD exception era, 2
patients were excluded from analysis because HCC was only
incidentally diagnosed on explants, leaving 131 patients for analysis.
Hepatic resection was performed in 128 patients for a diagnosis of HCC.
In order to clearly determine the outcomes of each procedure
independently, two patients who initially underwent resection and
subsequently underwent salvage transplantation for recurrence of HCC
were excluded from analysis, leaving 126 patients in the resection cohort
(Fig. 2). Demographics and clinicopathologic features for the qualifying
257 patients, classified by operation type, are presented in Table I.
Median wait time to transplant was 55 days, and no patients listed with a
Fig. 1. Proportion of liver transplant recipients with a waiting time of diagnosis of HCC dropped off the waitlist while awaiting an organ
90 days or less by UNOS region, 1999–2008, per Organ Procurement during the study period. Ninety patients (68.7%) listed for transplant
and Transplantation Network (OPTN)/Scientific Registry of Transplant underwent bridging therapy while on the wait list; 68 patients received
Recipients (SRTR) 2009 Annual Report. transarterial chemoembolization (TACE) and 22 patients underwent

Journal of Surgical Oncology

 Transplant Versus Resection for HCC 535

Fig. 2. Stratification of patients undergoing surgical treatment of HCC by surgical modality, Milan Criteria, and raw MELD score. HCC,
hepatocellular carcinoma; OLT, orthotopic liver transplantation; MELD, Model for End‐Stage Liver Disease.

radiofrequency ablation (RFA). The median wait time for the 41 patients Recurrence and Survival
who did not receive bridging therapy prior to transplant was 34 days.
Median overall follow‐up time was 30 months. Among the entire cohort, 80 patients (30.9%) had recurrence of HCC,
Overall, median age was 59.0 years, and 71% of patients were male. with significantly greater rates of recurrence among patients undergoing
Transplant patients were more likely than patients who underwent resection as compared to transplant (52.4% vs. 10.7%, P < 0.001). The
resection to have hepatitis B (34.4% vs. 16.7%, P ¼ 0.002) or hepatitis C majority of recurrences in both groups occurred within the liver
(66.4% vs. 28.6%, P < 0.001). The mean raw MELD score for the (Table I). Of the 45 patients within MC who underwent resection, 22
transplant cohort was significantly higher than the resection cohort (15 (48.9%) experienced recurrence of their HCC (Table II). Only 4 of these
vs. 8, P < 0.001). Based on preoperative imaging, all 131 of the 22 patients had a recurrence within MC following resection. Of these 22
transplant patients met MC, whereas 45 patients (36%) within the patients, 6 subsequently underwent repeat hepatic resection, 5 were
resection cohort met MC. Of the transplant cohort, only 1 patient who treated with TACE or Yttrium (Y‐90) radio‐embolization, 3 with RFA,
initially presented outside of MC was down‐staged with locoregional and 6 were started on sorafenib therapy. Among the 131 transplant
therapy before being placed on the waitlist and subsequently patients, 14 (10.7%; P < 0.001) experienced HCC recurrence; 2
transplanted. Among the resection patients meeting MC, 30 patients underwent surgical resection, 4 were treated with TACE or Y‐90
had relatively preserved hepatic function, defined as a raw MELD  8, radio‐embolization, 2 with RFA, 1 with systemic chemotherapy, and 3
versus 12 such patients among the transplant cohort (Fig. 2). The were started on sorafenib.
majority of patients within the transplant cohort were Child‐Pugh Class Among all patients, transplant was associated with significantly
B or C (71.8%), as opposed to the resection cohort, where most patients greater 5‐year OS and 5‐year RFS as compared to resection (OS: 65.7%
were Child‐Pugh Class A (85.7%). vs. 33.2%, P < 0.001; RFS: 85.3% vs. 22.9%, P < 0.001). When
The comparison of transplant patients to the 45 patients who comparing all transplant patients to those undergoing resection who met
underwent resection and met MC is detailed in Table II. Unifocal tumors MC (n ¼ 45), transplantation was associated with significantly improved
were present in 73 patients (55.7%) among the transplant cohort, as 5‐year OS (65.7% vs. 43.8%, P ¼ 0.005; Fig. 3) and RFS (85.3% vs.
compared to 39 (86.7%, P < 0.001) patients in the Milan‐meeting 22.7%, P < 0.001; Fig. 4). On subset analysis of patients meeting MC
resection cohort. Patients undergoing resection had significantly larger with hepatitis C, patients undergoing transplantation (n ¼ 87)
maximum tumor size (3.9 vs. 2.5 cm, P < 0.001), and had a greater demonstrated significantly greater 5‐year OS (63.5% vs. 23.3%,
proportion of poorly differentiated tumors (22.2% vs. 6.9%, P ¼ 0.01). P ¼ 0.001; Fig. 5) and RFS (83.5% vs. 23.7%, P < 0.001; Fig. 6) as
Rates of tumor macrovascular and microvascular invasion were not compared to those patients undergoing resection (n ¼ 21).
significantly different for patients undergoing transplant versus resection On survival analysis of patients meeting MC with preserved hepatic
(P ¼ 0.68 and 0.13, respectively). Three patients (6.7%) had function, defined as a raw MELD score  8, transplantation (n ¼ 12) and
microscopically positive margins following resection. Pathologic resection (n ¼ 30) were associated with similar 5‐year OS (62.5% vs.
evidence of cirrhosis or significant fibrosis was present in all 131 48.9%, P ¼ NS; Fig. 7), but transplantation demonstrated a trend
patients who underwent transplantation and 39 patients (86.7%) within towards greater RFS (71.6% vs. 30.8%, P ¼ 0.08; Fig. 8). When
the resection cohort. No patients undergoing transplantation had comparing outcomes for patients meeting MC with well‐compensated
fibrolamellar HCC on pathology, versus 1 such patient in the Child‐Pugh Class A cirrhosis, transplantation (n ¼ 37) demonstrated a
resection cohort. trend towards greater 5‐year OS (56% vs. 35%, P ¼ 0.07; Fig. 9) and was
associated with significantly improved RFS (71% vs. 37%, P ¼ 0.04;
Fig. 10) as compared to resection (n ¼ 16).

Journal of Surgical Oncology

536 Squires et al.
TABLE I. Demographics and Clinicopathologic Features of All Patients Undergoing Transplantation versus Resection for HCC

Total (n ¼ 257), OLT pts (n ¼ 131), Resection pts (n ¼ 126),

n (% total) n (% group) n (% group) P‐Value

Demographics
Gender
Male 183 (71.2%) 102 (77.9%) 81 (64.3%) 0.02
Race
White 166 (64.6%) 100 (76.3%) 66 (52.4%) <0.001
Black 47 (18.3) 17 (13.0) 30 (23.8)
Other 44 (17.1) 14 (10.7) 30 (23.8)
Age, median [range], years 59.0 [20.3–89.8] 57.0 [38.9–73.3] 61.5 [20.3–89.8] 0.02
ASA class
2 26 (10.1%) 26 (20.6%) <0.001
3 86 (33.5) 86 (68.2)
4 145 (56.4) 131 (100%) 14 (11.1)
Hepatitis B 66 (25.7) 45 (34.4) 21 (16.7) 0.002
Hepatitis C 123 (47.9) 87 (66.4) 36 (28.6) <0.001
Alcoholic cirrhosis 55 (21.4) 36 (27.5) 19 (15.1) 0.03
Preoperative labs, median [range]
Platelet (103/ml) 107 [11–747] 54 [11–171] 223 [54–747] <0.001
Albumin (gm/dl) 3.2 [1.4–4.8] 2.9 [1.4–4.1] 3.5 [1.6–4.8] <0.001
Total bilirubin (mg/dl) 1.1 [0.1–16.9] 2.8 [0.5–16.9] 0.7 [0.1–8.4] <0.001
INR 1.16 [0.86–5.13] 1.41 [0.93–5.13] 1.03 [0.86–1.50] <0.001
Creatinine (mg/dl) 0.90 [0.40–5.50] 1.00 [0.46–5.50] 0.90 [0.40–4.95] 0.03
AFP 19 [0–38000] 13.5 [0–2400] 34 [1–38000] 0.02
Raw MELD score 10 [6–39] 15 [6–39] 8 [6–23] <0.001
Patients within Milan Criteria 176 (68.5%) 131 (100%) 45 (35.7%) <0.001
Child‐Pugh Class
Class A 145 (56.4%) 37 (28.2%) 108 (85.7%) <0.001
Class B 93 (36.2) 75 (57.3) 18 (14.3)
Class C 19 (7.4) 19 (14.5) 0
Pathology
Number of lesions
Unifocal 176 (68.5%) 73 (55.7%) 103 (81.7%) <0.001
Multifocal 81 (31.5) 58 (44.3) 23 (18.3)
Largest tumor size (cm) 3.7 [0.8–29.0] 2.5 [0.8–6.0] 7.0 [1.5–29.0] <0.001
Differentiation
Well/moderate 227 (87.5%) 122 (93.1%) 103 (81.7%) 0.01
Poor 32 (12.5) 9 (6.9) 23 (18.3)
Macrovascular invasion 20 (7.8) 7 (5.3) 13 (10.3) 0.21
Microvascular invasion 78 (30.4) 23 (17.6) 55 (43.7) <0.001
Cirrhosis/fibrosis present 204 (79.4) 131 (100) 73 (57.9) <0.001
Outcomes
30 day mortality 12 (4.6%) 2 (1.5%) 10 (7.9%) 0.02
Recurrence, any 80 (30.9) 14 (10.7) 66 (52.4) <0.001
Intrahepatic recurrence 62 (23.9) 11 (8.4) 51 (41.8) <0.001
Extrahepatic recurrence 30 (11.6) 6 (4.6) 24 (19.0) <0.001

HCC, hepatocellular carcinoma; OLT, orthotopic liver transplantation; ASA, American Society of Anesthesiology; INR, International Normalized Ratio; AFP, alpha
fetoprotein; MELD, Model for End‐Stage Liver Disease.

Prognostic Factors for OS and RFS in Patients Meeting Milan OS and RFS versus resection. The more clinically relevant comparison
Criteria limited to those patients meeting MC demonstrated significantly greater
5‐year OS (65.7% vs. 43.8%, P ¼ 0.005) and RFS (85.3% vs. 22.7%,
The univariate and multivariate Cox regression analyses of all P < 0.001) in favor of transplantation as compared to resection.
patients meeting MC (n ¼ 176) for OS and RFS are presented in The optimal surgical management of early HCC within MC remains
Tables III and IV, respectively. After accounting for other adverse controversial and regionally dependent. The results of the present study
pathologic features, resection remained independently associated with are consistent with several studies over the past decade that have
decreased OS (HR 2.91; 95% CI: 1.52–5.57; P ¼ 0.001) and decreased suggested that transplantation, despite the potential morbidity of the
RFS (HR 9.98; 95% CI; 2.60–38.39; P ¼ 0.001), as compared to procedure and the burden of life‐long immunosuppression, may offer
transplant. superior survival and substantially less risk of recurrence compared to
resection [11–15]. Other studies have suggested that hepatic resection
DISCUSSION may provide equivalent or superior results to transplantation for select
patient subgroups, particularly those with minimal or well‐compensated
This study represents a single‐institution comparison of outcomes for hepatic dysfunction [16–18]. A recent meta‐analysis of outcomes for
transplantation versus hepatic resection for the treatment of HCC in the patients undergoing hepatic resection of HCC meeting MC concluded
MELD exception era in a region with short waitlist times for organ that resection in patients with preserved liver function produced good
availability. Overall, transplantation was associated with greater 5‐year outcomes, with a 5‐year OS of 67% (range, 27–81%), but was associated

Journal of Surgical Oncology

 Transplant Versus Resection for HCC 537
TABLE II. Clinicopathologic Features of Patients Undergoing Transplant (n ¼ 131) vs. Resection (n ¼ 45) for HCC Within Milan Criteria

OLT patients (n ¼ 131), n (% group) Resection patients within MC (n ¼ 45), n (% group) P‐Value

Clinical features
Hepatitis B 45 (34.4) 8 (17.8) 0.04
Hepatitis C 87 (66.4) 21 (46.7) 0.02
Alcoholic cirrhosis 36 (27.5) 9 (20.0) 0.43
Raw MELD, median [range] 15 [6–39] 8 [6–23] <0.001
Child‐Pugh Class
Class A 37 (28.2%) 39 (86.7%) <0.001
Class B 75 (57.3) 6 (13.3)
Class C 19 (14.5) 0
Pathologic features
Number of lesions
Unifocal 73 (55.7%) 39 (86.7%) <0.001
Multifocal 58 (44.3) 6 (13.3)
Largest tumor size (cm) 2.5 [0.8–6.0] 3.9 [1.5–7.0] <0.001
Differentiation
Well/moderate 122 (93.1%) 35 (77.8%) 0.01
Poor 9 (6.9) 10 (22.2)
Macroscopic vascular invasion 7 (5.3) 1 (2.2) 0.68
Microscopic vascular invasion 23 (17.6) 13 (28.9) 0.13
Margin positivity 0 3 (6.7) 0.02
Cirrhosis/fibrosis present 131 (100) 39 (86.7) <0.001
Outcomes
30 day mortality 2 (1.5%) 3 (6.7%) 0.11
Recurrence, any 14 (10.7) 22 (48.9) <0.001
Intrahepatic recurrence 11 (8.4) 19 (42.2) <0.001
Extrahepatic recurrence 6 (4.6) 4 (8.9) 0.28

MC, Milan Criteria; HCC, hepatocellular carcinoma; OLT, orthotopic liver transplantation; MELD, Model for End‐Stage Liver Disease.

with a substantial risk of disease recurrence (5‐year RFS: 37%; range, demonstrated a trend towards improved 5‐year OS with
21–57%) [19]. transplantation (56% vs. 35%, P ¼ 0.07) and significantly lower rates
When analyzing patients in the current study with preserved hepatic of recurrence (5‐year RFS: 71% vs. 37%, P ¼ 0.04). These results
function and a raw MELD score  8, similar 5‐year OS was observed for suggest that while patients with minimal liver dysfunction undergoing
transplantation and resection (62.5% vs. 48.9%, P ¼ NS), although resection for HCC within MC may achieve comparable survival in some
transplantation was associated with a trend towards improved RFS cases, the majority of such patients are living with recurrent disease.
(71.6% vs. 30.8%, P ¼ 0.08). When stratified by Child‐Pugh score, A separate meta‐analysis by Dhir et al. [20] comparing outcomes for
patients with well‐compensated Child‐Pugh Class A cirrhosis transplantation versus resection in patients with early HCC found no

Fig. 3. Overall survival for all patients meeting Milan Criteria based on Fig. 4. Recurrence‐free survival for all patients meeting Milan Criteria
preoperative imaging undergoing transplant (n ¼ 131) versus resection based on preoperative imaging undergoing transplant (n ¼ 131) versus
(n ¼ 45). resection (n ¼ 45).

Journal of Surgical Oncology

538 Squires et al.

Fig. 5. Overall survival for patients meeting Milan Criteria with

hepatitis c undergoing transplant (n ¼ 87) versus resection (n ¼ 21). Fig. 6. Recurrence‐free survival for patients meeting Milan Criteria
with hepatic C undergoing transplant (n ¼ 87) versus resection (n ¼ 21).

significant difference in survival, based on an intention‐to‐treat strategy. resection, only 4 of the 22 patients analyzed in this study with recurrent
When the analysis was limited to only patients with HCC and well‐ disease following primary resection of HCC within MC had a recurrence
compensated cirrhosis, transplantation was associated with significantly within MC that would have afforded them the opportunity for
improved 5‐year OS, although the authors remarked that only three small transplantation. Given the substantial number of patients expected to
studies qualified for this subset meta‐analysis [20]. A 2009 study by recur following resection and the low likelihood of qualifying for
Cherqui et al. [21] of 67 patients with Child‐Pugh Class A cirrhosis and salvage transplantation, primary transplantation appears to confer an
HCC meeting MC reported excellent 5‐year OS of 72% following oncologic advantage over resection and should be considered for most
resection, but significant risk of recurrence. Some authors have patients with HCC within MC in regions with short waitlist times.
advocated a strategy of initial resection for patients with HCC within Another consideration in evaluating the role of these two surgical
MC with preserved liver function, followed by “salvage transplantation” modalities for patients with HCC is the presence of underlying hepatitis
for recurrent disease [21–24]. Unfortunately a significant number of C. Patient with hepatitis C tend to have high rates of recurrence and poor
these patients never reach salvage transplantation due to recurrence outcomes following resection for HCC [25,26]; even in the setting of
outside of MC or extra‐hepatic recurrence, questioning the role of such a relatively preserved hepatic function, some have argued that these
strategy [21–23]. While the present study was not designed to assess the patients may derive significant benefit from transplantation [27]. Of
utility of salvage transplantation for recurrent HCC following initial particular interest in the present study was the significantly improved OS

Fig. 8. Recurrence‐free survival for patients meeting Milan Criteria

Fig. 7. Overall survival for patients meeting Milan Criteria with raw with raw MELD  8 undergoing transplant (n ¼ 12) versus resection
MELD  8 undergoing transplant (n ¼ 12) versus resection (n ¼ 30). (n ¼ 30).

Journal of Surgical Oncology

 Transplant Versus Resection for HCC 539
recurrence; all patients who remain recurrence‐free at the time of death
or time of last follow‐up are censored. While transplant patients
experienced mortality that related to their liver transplant (sepsis, graft
failure, etc.), only 8 of the 32 deceased transplant patients had recurrent
HCC at the time of their death. The remaining 24 deceased transplant
patients had no evidence of HCC recurrence and thus their RFS duration
was appropriately censored at their time of death. This finding is what
accounts for the fact that RFS exceeded OS in the transplant cohort.
These results contrasted starkly with the resection cohort. The
recurrence rate for HCC was not only significantly greater for patients
within MC undergoing resection versus transplantation (49% vs. 11%,
P < 0.001), but increased recurrence directly translated into significantly
worse survival for patients treated with resection. Of the 20 deceased
resection patients, 14 had recurrent HCC at the time of their death. These
findings further highlight the significantly greater risk of recurrence and
the detrimental effect of recurrence on survival for patients undergoing
resection for HCC within MC, compared to transplantation in a UNOS
region where organs are readily available and wait times are routinely
less than 60 days.
The present study, similarly to most previously published series of
Fig. 9. Overall survival for patients meeting Milan Criteria with Child‐ HCC patients, is limited by its retrospective design and nonrandomized
Pugh Class A cirrhosis undergoing transplant (n ¼ 37) versus resection patient selection. As expected, the clinicopathologic features of the
(n ¼ 16). transplantation and resection groups in the present study were
significantly different, with patients in the resection cohort having
larger and more poorly differentiated tumors than those in the transplant
cohort. By limiting the analysis to patients within MC and stratifying
and RFS associated with transplantation as compared to resection for patients by MELD score and Child‐Pugh score, we sought to minimize
those patients meeting MC with hepatitis C. Given similar recent confounding discrepancies and optimize the comparisons between the
findings of poor outcomes for hepatitis C‐positive patients undergoing two cohorts. Additionally, Cox regression analyses were performed in
resection for HCC within MC [25–27], these current results suggest that order to account for potential differences between the transplant and
transplantation should be favored over resection for patients with resection cohorts. Even after accounting for adverse pathologic features
underlying hepatitis C. such as tumor size, differentiation, and vascular invasion, resection
The 5‐year RFS for patients who underwent transplantation was in remained independently associated with worse OS and RFS as compared
fact greater than the 5‐year OS; this was not the case for resection to transplant for patients meeting MC.
patients. The explanation for this finding lies in the fact that, unlike the Another limitation is the potential for bias in terms of patient follow‐
resection cohort, very few patients within the transplant cohort died due up; transplant patients are routinely seen in clinic every 4–6 months
to recurrent HCC (n ¼ 8). Per Kaplan–Meier methodology, an analysis posttransplant, while more patients who underwent resection were lost
of RFS measures the event of recurrence and the duration of survival to follow‐up over the study period. The perioperative mortality rate of
until that event. By definition, RFS is terminated at the date of the patients undergoing resection was higher than expected, suggesting
that some of these patients may have been borderline resection
candidates, and perhaps would have been better served with
transplantation or primary therapy with interventional techniques
such as RFA, TACE, or Y‐90. All patients at our institution placed on
the transplant wait list with a diagnosis of HCC are considered for
bridging therapy. Ninety patients (68.7%) received pretransplant TACE
or RFA, and the median time on the wait list for those patients who did
not receive bridging therapy was only 34 days, compared to 55 days for
the entire cohort. Studies have not demonstrated a correlation between
pretransplant locoregional bridging therapy and improved survival, and
given the short waitlist times in the current study, it is unlikely that the
improved outcomes associated with transplant are attributable to
bridging therapy [28,29].
The results of the current study need to be interpreted in the context
of short waitlist times for transplant. They may not be applicable or
reproducible in regions with longer waitlist times or lesser organ
availability. A study by Shah et al. [30] determined that transplantation
was associated with superior survival to resection for patients with
HCC within MC and Child‐Pugh Class A or B cirrhosis only when wait
time to transplant was less than 4 months. A median waitlist time of
55 days in the present study contrasts starkly with median wait times of
6–12 months in many published reports from other regions of the
United States and globally [31,32]. In this series, no patients were
Fig. 10. Recurrence‐free survival for patients meeting Milan Criteria removed from the waitlist due to disease progression or death while
with Child‐Pugh Class A cirrhosis undergoing transplant (n ¼ 37) versus awaiting transplantation; this is certainly not the case in other
resection (n ¼ 16). regions [17,31].

Journal of Surgical Oncology

540 Squires et al.
TABLE III. Univariate and Multivariate Analysis of Risk Factors Associated With Overall Survival for Patients With HCC Within Milan Criteria (n ¼ 176)

Univariate Multivariate

Variable HR (95% CI) P‐Value HR (95% CI) P‐Value

Age 1.02 (0.98–2.05) 0.36

Gender, male 1.40 (0.70–2.80) 0.34
Race
White Ref Ref
Black 1.97 (1.05–3.69) 0.03 1.57 (0.82–3.00) 0.18
Other 0.71 (0.28–1.82) 0.47 0.56 (0.21–1.47) 0.24
ASA class
2 Ref
3 0.55 (0.12–2.39) 0.42
4 0.30 (0.07–1.24) 0.11
Hepatitis B 1.68 (0.96–2.92) 0.07
Hepatitis C 1.54 (0.85–2.77) 0.15
Alcohol abuse 1.64 (0.90–2.99) 0.11
Tumor size 1.16 (0.96–1.41) 0.15
Macrovascular invasion 3.40 (1.35–8.61) 0.01 3.56 (1.32–9.58) 0.01
Microvascular invasion 1.51 (0.80–2.83) 0.20
Tumor grade, poor 2.30 (1.12–4.72) 0.02 1.68 (0.78–3.62) 0.19
Resection 2.20 (1.26–3.85) 0.006 2.23 (1.20–4.14) 0.01

HCC, hepatocellular carcinoma; ASA, American Society of Anesthesiology; HR, hazard ratio; CI, confidence interval.
The bolded text appearing in Table designates statistical significance (P  0.05).

TABLE IV. Univariate and Multivariate Analysis of Risk Factors Associated With Recurrence‐Free Survival for Patients With HCC Within Milan Criteria
(n ¼ 176)

Univariate Multivariate

Variable HR (95% CI) P‐Value HR (95% CI) P‐Value

Age 1.03 (0.99–1.08) 0.17

Gender, male 1.21 (0.55–2.65) 0.64
Race
White Ref Ref
Black 2.47 (1.16–5.25) 0.02 1.38 (0.60–3.20) 0.45
Other 1.13 (0.42–3.00) 0.81 0.54 (0.18–1.61) 0.27
ASA class 2.29 (1.18–3.46) <0.001 1.54 (0.47–5.02) 0.48
Hepatitis B 0.92 (0.44–1.91) 0.82
Hepatitis C 1.05 (0.54–2.05) 0.90
Alcohol abuse 1.16 (0.53–2.56) 0.71
Tumor size 1.55 (1.25–1.94) <0.001 1.24 (0.95–1.62) 0.12
Macrovascular invasion 4.14 (1.45–11.78) 0.01 8.51 (1.93–37.57) 0.005
Microvascular invasion 2.39 (1.19–4.78) 0.01 1.17 (0.42–3.29) 0.76
Tumor grade, poor 4.18 (1.96–8.92) <0.001 2.37 (0.93–6.03) 0.07
Resection 7.40 (3.77–14.52) <0.001 9.98 (2.60–38.39) 0.001

HCC, hepatocellular carcinoma; ASA, American Society of Anesthesiology; HR, hazard ratio; CI, confidence interval.
The bolded text appearing in Table designates statistical significance (P  0.05).

CONCLUSION ACKNOWLEDGMENTS
In regions with short wait times, transplantation for HCC within MC This study was supported in part by the Katz Foundation. S.B.F. is
may provide improved overall and RFS compared to resection. Given supported by the National Center for Advancing Translational Sciences
the substantial risk of recurrence following primary resection for patients of the National Institutes of Health under Award Number
with HCC and underlying hepatitis C and the associated poor outcomes, UL1TR000454. The content is solely the responsibility of the authors
all patients with hepatitis C should be considered for transplantation. In and does not necessarily represent the official views of the National
patients with preserved liver function, whether defined as Child‐Pugh Institutes of Health.
Class A or a MELD score  8, transplantation and resection appear to
provide similar OS. The significant recurrence rate associated with
resection, however, suggests that transplantation may provide an REFERENCES
oncologic advantage in these patients as well. Transplantation should be 1. Jemal A, Bray F, Center MM, et al.: Global cancer statistics. CA
considered for all patients with HCC meeting MC, including those with Cancer J Clin 2011;61:69–90.
minimal liver dysfunction and particularly those with hepatitis C, when 2. Edwards BK, Ward E, Kohler BA, et al.: Annual report to the nation
being managed in a region with short wait times for organ availability. on the status of cancer, 1975–2006, featuring colorectal cancer

Journal of Surgical Oncology

 Transplant Versus Resection for HCC 541
trends and impact of interventions (risk factors, screening, and Pugh class A patients with cirrhosis who are eligible for liver
treatment) to reduce future rates. Cancer 2010;116:544–573. transplantation. Liver Transpl 2005;11:1242–1251.
3. Llovet JM, Burroughs A, Bruix J: Hepatocellular carcinoma. Lancet 19. Lim KC, Chow PK, Allen JC, et al.: Systematic review of outcomes
2003;362:1907–1917. of liver resection for early hepatocellular carcinoma within the
4. Jarnagin WR: Management of small hepatocellular carcinoma: A Milan criteria. Br J Surg 2012;99:1622–1629.
review of transplantation, resection, and ablation. Ann Surg Oncol 20. Dhir M, Lyden ER, Smith LM, et al.: Comparison of outcomes of
2010;17:1226–1233. transplantation and resection in patients with early hepatocellular
5. Rahbari NN, Mehrabi A, Mollberg NM, et al.: Hepatocellular carcinoma: A meta‐analysis. HPB (Oxford) 2012;14:635–645.
carcinoma: Current management and perspectives for the future. 21. Cherqui D, Laurent A, Mocellin N, et al.: Liver resection for
Ann Surg 2011;253:453–469. transplantable hepatocellular carcinoma: Long‐term survival and
6. Mazzaferro V, Regalia E, Doci R, et al.: Liver transplantation for the role of secondary liver transplantation. Ann Surg 2009;250:738–
treatment of small hepatocellular carcinomas in patients with 746.
cirrhosis. N Engl J Med 1996;334:693–699. 22. Fuks D, Dokmak S, Paradis V, et al.: Benefit of initial resection of
7. Mazzaferro V, Bhoori S, Sposito C, et al.: Milan criteria in liver hepatocellular carcinoma followed by transplantation in case of
transplantation for hepatocellular carcinoma: An evidence‐based recurrence: An intention‐to‐treat analysis. Hepatology 2012;55:
analysis of 15 years of experience. Liver Transpl 2011;17:S44–S57. 132–140.
8. Teh SH, Christein J, Donohue J, et al.: Hepatic resection of 23. Cucchetti A, Vitale A, Gaudio MD, et al.: Harm and benefits of
hepatocellular carcinoma in patients with cirrhosis: Model of End‐ primary liver resection and salvage transplantation for hepatocellu-
Stage Liver Disease (MELD) score predicts perioperative mortality. lar carcinoma. Am J Transplant 2010;10:619–627.
J Gastrointest Surg 2005;9:1207–1215,discussion 1215. 24. Poon RT, Fan ST, Lo CM, et al.: Long‐term survival and pattern of
9. Organ Procurement and Transplantation Network, Liver Kaplan– recurrence after resection of small hepatocellular carcinoma in
Meier Median Waiting Times for Registrations Listed: 1999–2004. patients with preserved liver function: Implications for a strategy of
As of May 24, 2013. Available at: http://optn.transplant.hrsa.gov salvage transplantation. Ann Surg 2002;235:373–382.
10. 2009 OPTN/SRTR Annual Report, Chapter IV: Liver transplanta- 25. Kao WY, Su CW, Chau GY, et al.: A comparison of prognosis
tion in the United States 1999–2008 [Available through U.S. between patients with hepatitis B and C virus‐related hepatocellular
Department of Health and Human Services at: http://www. carcinoma undergoing resection surgery. World J Surg 2011;35:
ustransplant.org/annual_reports/current/chapter_iv_AR_cd.htm]. 858–867.
11. Lee KK, Kim DG, Moon IS, et al.: Liver transplantation versus liver 26. Huo TI, Wu JC, Hsia CY, et al.: Hepatitis C virus infection is a risk
resection for the treatment of hepatocellular carcinoma. J Surg factor for tumor recurrence after resection of small hepatocellular
Oncol 2010;101:47–53. carcinomas. World J Surg 2004;28:787–791.
12. Baccarani U, Isola M, Adani GL, et al.: Superiority of transplanta- 27. Chirica M, Tranchart H, Tan V, et al.: Infection with hepatitis C
tion versus resection for the treatment of small hepatocellular virus is an adverse prognostic factor after liver resection for early‐
carcinoma. Transpl Int 2008;21:247–254. stage hepatocellular carcinoma: Implications for the management of
13. Adam R, Bhangui P, Vibert E, et al.: Resection or transplantation for hepatocellular carcinoma eligible for liver transplantation. Ann
early hepatocellular carcinoma in a cirrhotic liver: Does size define Surg Oncol 2013;20:2405–2412.
the best oncological strategy? Ann Surg 2012;256:883–891. 28. Porrett PM, Peterman H, Rosen M, et al.: Lack of benefit of pre‐
14. Bellavance EC, Lumpkins KM, Mentha G, et al.: Surgical transplant locoregional hepatic therapy for hepatocellular cancer in
management of early‐stage hepatocellular carcinoma: Resection the current MELD era. Liver Transpl 2006;12:665–673.
or transplantation? J Gastrointest Surg 2008;12:1699–1708. 29. Heckman JT, Devera MB, Marsh JW, et al.: Bridging locoregional
15. Kooby DA, Egnatashvili V, Graiser M, et al.: Changing therapy for hepatocellular carcinoma prior to liver transplantation.
management and outcome of hepatocellular carcinoma: Evaluation Ann Surg Oncol 2008;15:3169–3177.
of 501 patients treated at a single comprehensive center. J Surg 30. Shah SA, Cleary SP, Tan JC, et al.: An analysis of resection vs
Oncol 2008;98:81–88. transplantation for early hepatocellular carcinoma: Defining the
16. Koniaris LG, Levi DM, Pedroso FE, et al.: Is surgical resection optimal therapy at a single institution. Ann Surg Oncol 2007;
superior to transplantation in the treatment of hepatocellular 14:2608–2614.
carcinoma? Ann Surg 2011;254:527–537,discussion 537–528. 31. Freeman RB, Edwards EB, Harper AM: Waiting list removal rates
17. Facciuto ME, Rochon C, Pandey M, et al.: Surgical dilemma: Liver among patients with chronic and malignant liver diseases. Am J
resection or liver transplantation for hepatocellular carcinoma and Transplant 2006;6:1416–1421.
cirrhosis. Intention‐to‐treat analysis in patients within and outwith 32. Pelletier SJ, Fu S, Thyagarajan V, et al.: An intention‐to‐treat
Milan criteria. HPB (Oxford) 2009;11:398–404. analysis of liver transplantation for hepatocellular carcinoma using
18. Margarit C, Escartin A, Castells L, et al.: Resection for organ procurement transplant network data. Liver Transpl 2009;
hepatocellular carcinoma is a good option in Child–Turcotte– 15:859–868.

Journal of Surgical Oncology