Published Ahead of Print on February 11, 2013 as 10.1200/JCO.2012.46.2762 The latest version is at http://jco.ascopubs.org/cgi/doi/10.1200/JCO.2012.46.

2762

JOURNAL OF CLINICAL ONCOLOGY

R E V I E W

A R T I C L E

Markers of Response for the Antiangiogenic Agent Bevacizumab

Diether Lambrechts, Heinz-Josef Lenz, Sanne de Haas, Peter Carmeliet, and Stefan J. Scherer
Diether Lambrechts and Peter Carmeliet, Vesalius Research Center, Flanders Institute for Biotechnology (VIB), and University of Leuven, Leuven, Belgium; HeinzJosef Lenz, Keck School of Medicine, University of Southern California, Los Angeles; Stefan J. Scherer, Genentech, South San Francisco, CA; and Sanne de Haas, F. Hoffmann-La Roche, Basel, Switzerland. Published online ahead of print at www.jco.org on February 11, 2013. Supported by the Fund for Scientic Research Flanders (D.L.), the Agency for Innovation by Science and Technology, the Stichting tegen Kanker, Grant No. KULPFV/10/016SymBioSys from the University of Leuven, and by the Seventh Framework Programme of the European Community for Research (AngioPredict). Authors disclosures of potential conicts of interest and author contributions are found at the end of this article. Corresponding author: Diether Lambrechts, PhD, Vesalius Research Center, VIB, Herestraat 49, bus 912, Leuven, Belgium B-3000; e-mail: [email protected]. 2013 by American Society of Clinical Oncology 0732-183X/13/3199-1/$20.00 DOI: 10.1200/JCO.2012.46.2762

Bevacizumab is the rst antiangiogenic therapy proven to slow metastatic disease progression in patients with cancer. Although it has changed clinical practice, some patients do not respond or gradually develop resistance, resulting in rather modest gains in terms of overall survival. A major challenge is to develop robust biomarkers that can guide selection of patients for whom bevacizumab therapy is most benecial. Here, we discuss recent progress in nding such markers, including the rst results from randomized phase III clinical trials evaluating the efcacy of bevacizumab in combination with comprehensive biomarker analyses. In particular, these studies suggest that circulating levels of short vascular endothelial growth factor A (VEGF-A) isoforms, expression of neuropilin-1 and VEGF receptor 1 in tumors or plasma, and genetic variants in VEGFA or its receptors are strong biomarker candidates. The current challenge is to expand this rst set of markers and to validate it and implement it into clinical practice. A rst prospective biomarker study known as MERiDiAN, which will treat patients stratied for circulating levels of short VEGF-A isoforms with bevacizumab and paclitaxel, is planned and will hopefully provide us with new directions on how to treat patients more efciently. J Clin Oncol 31. 2013 by American Society of Clinical Oncology

INTRODUCTION

Bevacizumab is a humanized monoclonal antibody against the vascular endothelial growth factor A (VEGF-A), a key factor inducing the formation of blood vessels (angiogenesis) in tumors.1 Bevacizumab is currently approved in Europe and the United States in combination with standard chemotherapy for the treatment of metastatic colorectal cancer (mCRC)2 and nonsmall-cell lung cancer (NSCLC).3 The drug is also approved in combination with interferon alfa-2a for renal cell carcinoma (RCC),4-6 with standard chemotherapy for advanced ovarian cancer in Europe,7,8 and as a single agent for recurrent glioblastoma in the United States.9 In early 2008, bevacizumab also received the US Food and Drug Administrations accelerated approval for treatment of metastatic breast cancer (mBC). Its efcacy and safety were shown in a multicenter trial (E2100) that randomly assigned women with mBC to conventional chemotherapy alone or in combination with bevacizumab.10 The trial showed that bevacizumab signicantly improved progression-free survival (PFS). Adverse effects were moderate and manageable. Subsequent completion of the E2100 trial and publication of other trials in rst-line mBC, in which bevacizumab

failed to improve overall survival (OS), revealed excess toxicity and less benet in terms of PFS than expected on the basis of E2100.11,12 As a result, the US Food and Drug Administration revoked the license for bevacizumab in the setting of rstline mBC.13 This failure needs to be seen in the context of a more general debate, in which it is increasingly being realized that biomarkers for targeted cancer therapies are necessary, because only a subset of patients respond, and the overall clinical benet is limited.14 Considering the high cost of these therapies, predictive markers are not only a clinical necessity but are also an economic requirement. With this in mind, extensive biomarker programs have been built into numerous clinical studies with bevacizumab. However, a marker that predicts bevacizumab treatment outcome has not yet been validated. What are the reasons for these problems? And what can be done to move forward? Here, we try to provide an answer to these questions. First, we describe the challenges in identifying biomarkers for bevacizumab. Then, we identify a set of markers that we consider most promising, either because they were predictive in placebo-controlled studies involving large numbers of patients or have been replicated in several other studies. Finally, we also speculate on how to further improve this set of markers and discuss how to implement them into clinical practice.
2013 by American Society of Clinical Oncology

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Copyright 2013 by American Society of Clinical Oncology

Lambrechts et al

CHALLENGES IN IDENTIFYING BIOMARKERS FOR BEVACIZUMAB

The search for a biomarker predictive of bevacizumab treatment outcome has proven to be challenging for various reasons. First, angiogenesis is a complex and highly adaptive biologic process. Despite the predominant role of VEGF-A, multiple other factors can play an essential role during angiogenesis, including the placental growth factor (PlGF), broblast growth factors (FGFs) and platelet-derived growth factors (PDGFs), angiopoietins (ANGs), and various cytokines.1 In addition, other factors that promote proteolytic degradation of the matrix or induce maturation of the vasculature by stimulating pericyte coverage (PDGF-BB, ephrin-B2, and NOTCH) also critically contribute to the process. As a result, the activity of bevacizumab may be compensated by at least a dozen alternative angiogenic signals. Second, preclinical studies revealed that VEGF-A blockade has heterogeneous effects on tumors, ranging from inhibition of vessel expansion and regression of pre-existing vessels to inhibition of bone marrow derived cell and/or endothelial progenitor cell recruitment to the vascular wall.15,16 However, in humans, studies are often hindered by the inability to perform serial tumor biopsies, thereby preventing histologic analysis of vascular changes. A seminal study in rectal carcinoma revealed that a single infusion of bevacizumab rapidly decreased tumor perfusion, vascular volume, microvascular density, and interstitial uid pressure (consistent with a reduction in vascular permeability) and increased the fraction of vessels with pericytes.17 The net result was a more functional and normal vasculature, with the potential for improved delivery and efcacy of chemotherapeutic agents.18,19 In patients with malignant glioma, bevacizumab and the pan-VEGF receptor (VEGFR) tyrosine kinase inhibitor (TKI) cediranib also exhibited features of vessel normalization, leading to reduced peritumoral edema and therapeutic effects of both drugs.20,21 Normalization of the tumor vasculature may, however, be transient and context dependent and has yet to be conrmed in many of the common human tumors.22 Insights on how the tumor vasculature differs between cancers, the same cancer at different stages of progression (eg, adjuvant v metastatic setting), or after different treatment regimens and, importantly, how these differences could inuence vessel normalization after bevacizumab, are thus still limited.23,24 Finally, clinical end points fail to accurately identify which patients benet from bevacizumab. First, objective response rates do not predict the magnitude of PFS or OS benet from bevacizumab therapy in some studies of mCRC,25,26 whereas other studies observed that bevacizumab induces clear improvements in PFS without increasing objective response rates.27 Second, because patients may not participate if the experimental drug is not offered as either rst-line or second-line therapy, some studies include a crossover option on disease progression. This crossover rate may be high (for instance, up to 37% in the Avastin and Docetaxel in Metastatic Breast Cancer [AVADO] trial) thereby also compromising OS as an end point for biomarker studies with bevacizumab.11,12 Finally, distinguishing patient subgroups with differential effects of bevacizumab is challenging on the basis of PFS alone, since bevacizumab is effective only in combination with chemotherapy, which by itself may also differentially affect survival. Although several of the issues we raised apply to targeted therapies in general, differences in the tumor vasculature and limited in2
2013 by American Society of Clinical Oncology

sights into how it reacts to VEGF inhibition are particularly relevant for antiangiogenic therapies.28 Because of these numerous challenges, a single biomarker may not sufce to reliably predict bevacizumab treatment outcome across cancers. The following is a discussion of various biomarkers that have been identied as predictors of bevacizumab treatment outcome.

PLASMA MARKERS PREDICTIVE OF BEVACIZUMAB TREATMENT OUTCOME

Many studies have measured circulating angiogenic factors (CAFs) to predict outcome of bevacizumab treatment (Table 1). Most studies included VEGF-A as an obvious candidate, because high levels of VEGF-A could indicate VEGF-A dependency of the tumor vasculature. Although increased tumor or plasma VEGF-A levels are well established as indicators of poor prognosis, data related to the predictive effect of pretreatment VEGF-A levels have largely been inconsistent.42 In the E4599 study for NSCLC, response rates in patients with high VEGF-A levels were signicantly higher in the bevacizumab arm than in the placebo arm.30 Single-arm studies involving cancer of the breast, ovary, and endometrium observed similar correlations, but many other studies failed to observe such effect (Data Supplement). A recent meta-analysis of 1,816 patients participating in phase III trials in CRC, NSCLC, and RCC conrmed that pretreatment VEGF-A levels serve as a prognostic rather than predictive marker (Table 1).31 In contrast, pretreatment soluble VEGFR1 (sVEGFR1) levels inversely correlated with outcome of either bevacizumab or anti-VEGFR TKIs in at least ve different trials. In particular, there was an inverse correlation in patients with rectal cancer after bevacizumab and chemoradiotherapy,32 BC after bevacizumab with chemotherapy,33 hepatocellular carcinoma after cediranib,43 and mCRC after vandetanib.44 Conversely, baseline levels of most other CAFs failed to predict benet after bevacizumab (Data Supplement), although potentially interesting correlations were observed for interleukin-8 (IL-8) and ANG-2 (Table 1). Recent data suggest that a novel enzyme linked immunosorbent assay (ELISA) with a preference to detect short VEGF-A isoforms may be more promising as a predictive marker. Through alternative RNA splicing, several VEGF-A isoforms are generated, of which the short VEGF-A121 isoform is freely diffusible, because it lacks basic amino acid residues that bind the extracellular matrix (ECM). The longer isoforms, consisting of 165, 189, or 206 amino acids, bind to heparin and heparan sulfate proteoglycans in the ECM. Because of differential afnities to the ECM, VEGF-A isoforms lay down a spatial VEGF-A gradient, with VEGF-A121 diffusing over long distances, VEGF-A165 reaching distant and nearby target cells, and ECM-bound VEGF-A189 providing guidance cues over short ranges.45 High circulating levels of short VEGF-A isoforms provide a more specic readout of the tumor-secreted VEGF-A. Several phase III randomized trials revealed that patients with mBC (AVADO trial) as well as patients with pancreatic cancer (Avastin and Tarceva in Advanced Pancreatic Cancer [AViTA] trial) who express high baseline levels of VEGF-A, as measured with this novel ELISA, exhibit improved PFS and/or OS after bevacizumab.13,37,38 Likewise, in the randomized Avastin in Advanced Gastric Cancer (AVAGAST) trial, for which plasma was
JOURNAL OF CLINICAL ONCOLOGY

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Biomarkers for Antiangiogenic Therapies

Table 1. Circulating Angiogenic Factors As Predictive Biomarkers for Bevacizumab Treatment Outcome Protein VEGF-A Cancer Type Colorectal Reference Goede et al29 Study Acronym Sample Size 34 Phase Study Details All patients received bevacizumab combined with either FOLFIRI, FOLFOX, XELIRI, or XELOX Three arms, receiving either 0, 7.5, or 15 mg/kg bevacizumab, each combined with cisplatingemcitabine All patients received bevacizumab carboplatinpaclitaxel or cisplatingemcitabine Two arms, receiving either bevacizumab or placebo, each combined with carboplatinpaclitaxel See Jayson et al38 for details on individual studies Two arms, receiving either bevacizumab or placebo, each combined with irinotecan uorouracil leucovorin Three arms, receiving 0, 7.5, or 15 mg/kg bevacizumab, each combined with cisplatin-gemcitabine Three arms, receiving 0, 7.5, or 15 mg/kg bevacizumab, each combined with docetaxel Two arms, receiving either bevacizumab or placebo, each combined with gemcitabine-erlotinib Two arms, receiving either bevacizumab or placebo, each combined with capecitabine-cisplatin Two arms, receiving either bevacizumab or placebo, each combined with IFN-2a Single arm, receiving four cycles of therapy consisting of bevacizumab for each cycle; uorouracil in cycles 2 to 4; external-beam irradiation and surgery after therapy Preoperative trial with a run-in of single-agent bevacizumab followed by ddACT chemotherapy All patients received bevacizumab All patients received bevacizumab FOLFIRI All patients received bevacizumab erlotinib All patients received bevacizumab combined with either FOLFIRI, FOLFOX, XELIRI, or XELOX Correlation With Outcome No (OR/PFS/OS)

Lung

Leighl et al34

AVAiL

358

III

No (PFS), no (OS)

Lung

Mok et al35

BO21015

287

II

Yes (PFS), trend (OR)

Lung

Dowlati et al30

ECOG E4599

160

II/III

Yes (OR), no (OS)

Short VEGF-A isoforms

Colorectal/lung/ renal cell Colorectal

Bernaards et al31 Jayson et al38 AVF2107g

1,816 398

Different phase III clinical studies III

No (PFS/OS) No (PFS/OS)

Lung

Jayson et al38

AVAiL

859

III

Trend (PFS), no (OS)

Breast

Miles et al37

AVADO

396

III

Yes (PFS/OS)

Pancreatic

Van Cutsem et al36

AViTA

225

III

Trend (PFS), yes (OS)

Gastric

Van Cutsem et al39

AVAGAST

712

III

Trend (PFS), yes (OS)

Renal cell

Jayson et al38

AVOREN

404

III

No (PFS/OS)

sVEGFR1

Rectal

Willett et al32

32

I/II

Yes (tumor stage)

Breast

Tolaney et al33

104

Yes (pathologic response)

IL-8

Hepatocellular Colorectal

Boige et al40 Kopetz et al

42

43 43 40 34

II II II

Yes (PFS/OS) Yes (PFS) No (PFS), yes (OS) Yes (OR/PFS/OS)

ANG2

Hepatocellular Colorectal

Kaseb et al41 Goede et al29

NOTE. All peer-reviewed publications available from PubMed and abstracts presented at international meetings were screened, but only a limited number of studies are included in this table. Preference was given to studies assessing a large number of patients, studies including a placebo-controlled arm, and markers for which consistent ndings were reported in several studies. Abbreviations: ANG2, angiopoietin 2; AVADO, Avastin and Docetaxel in Metastatic Breast Cancer; AVAGAST, Avastin in Advanced Gastric Cancer; AVAiL, Avastin in Lung Cancer; AViTA, Avastin and Tarceva in Advanced Pancreatic Cancer; AVOREN, Avastin and Roferon in Renal Cell Carcinoma; ddACT, dose-dense doxorubicin cyclophosphamide paclitaxel; ECOG, Eastern Cooperative Oncology Group; FOLFIRI, folinic acid uorouracil irinotecan; FOLFOX, folinic acid uorouracil oxaliplatin; IFN-2a, interferon alfa-2a; IL-8, interleukin-8; OR, objective response; OS, overall survival; PFS, progression-free survival; sVEGFR1, soluble vascular endothelial growth factor receptor 1; VEGF-A, vascular endothelial growth factor A; XELIRI, capecitabine irinotecan; XELOX, capecitabine oxaliplatin.

www.jco.org

2013 by American Society of Clinical Oncology

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Lambrechts et al

available from 712 patients or 92% of the study population, patients with high baseline plasma VEGF-A levels exhibited improved OS (hazard ratio [HR], 0.72) relative to patients with low VEGF-A levels (HR, 1.01).39 In the randomized studies for CRC, NSCLC, and RCC, no such correlation was observed (Table 1), possibly because citrated plasma rather than EDTA plasma was collected in these trials.38

Treatment-related changes in CAF might also predict adaptive resistance to antiangiogenic therapies. Several studies have shown acute increases in circulating VEGF-A levels on delivery of bevacizumab.32,46,47 The magnitude of these changes was proposed as a predictive marker, but ndings have not been consistent (Table 2). Because most of the circulating VEGF-A is bound by bevacizumab and cannot be differentiated from

Table 2. Changes in Expression of Circulating Angiogenic Factors During Bevacizumab Treatment Protein VEGF-A Cancer Type Colorectal Reference Willett et al32 Study Acronym NCI#5642 Sample Size 32 Phase II Study Details All patients received bevacizumab uorouracil Two arms, docetaxel bevacizumab Change During Bevacizumab Treatment Increased at different time points after start of bevacizumab treatment Increased at weeks 17 to 30 after start of bevacizumab treatment; no increase in chemotherapy-only arm Decreased at day 3 after start of bevacizumab treatment Decreased at 2 weeks after start of bevacizumab treatment Decreased from cycle 2 to 4 of bevacizumab treatment Increased gradually until progression (weeks 2 to 4 PD) Increased at weeks 8 to 24 after start of bevacizumab treatment, then normalized at PD Increased at different time points after start of bevacizumab treatment Increased prior to and at progression Increased at progression

Breast

Baar et al46

49

II

Hepatocellular

Boige et al40

43

II

Melanoma

Fuerstenberger et al49

SAKK 50/07

60

II

Ovarian

Smerdel et al50

38

All patients received bevacizumab as a single agent All patients received bevacizumab temozolomide All patients received bevacizumab All patients received bevacizumab FOLFIRI All patients received bevacizumab FOLFOXIRI All patients received bevacizumab uorouracil All patients received bevacizumab FOLFIRI All patients received bevacizumab FOLFIRI Two arms, receiving either bevacizumab or placebo, each combined with carboplatinpaclitaxel All patients received bevacizumab FOLFIRI

PlGF

Colorectal

Kopetz et al42

43

II

Loupakis et al52

25

II

Willett et al32

NCI #5642

32

II

VEGF-C

Colorectal

Lieu et al51

42

II

VEGF-D

Colorectal

Lieu et al51

42

II

bFGF

Lung

Dowlati et al30

ECOG E4599

160

II/III

Increased after cycle 2 (similar increase in placebo arm)

Colorectal

Kopetz et al43

43

II

PDGF-BB

Colorectal

Kopetz et al42

43

II

SDF1

Colorectal

Kopetz et al42

43

II

All patients received bevacizumab FOLFIRI All patients received bevacizumab FOLFIRI

Unchanged at weeks 2 to 4 after start of bevacizumab treatment, but increased prior to and at progression Unchanged, but increased prior to and at progression Unchanged, but increased prior to progression

NOTE. All peer-reviewed publications available from PubMed and abstracts presented at international meetings were screened, but only a limited number of studies are included in this table. Preference was given to studies assessing a large number of patients, studies including a placebo-controlled arm, and markers for which consistent ndings were reported in several studies. Abbreviations: bFGF, basic broblast growth factor; ECOG, Eastern Cooperative Oncology Group; FOLFIRI, folinic acid uorouracil irinotecan; FOLFOX, folinic acid uorouracil oxaliplatin irinotecan; FOLFOXIRI, irinotecan oxaliplatin uorouracil folinic acid; NCI, National Cancer Institute; PD, progressive disease; PDGF-BB, platelet-derived growth factor BB; PlGF, placental growth factor; SAKK, Schweizerische Arbeitsgemeinschaft fu r Klinische Krebsforschung; SDF1, stromal cell-derived factor 1; VEGF-A, vascular endothelial growth factor A.

2013 by American Society of Clinical Oncology

JOURNAL OF CLINICAL ONCOLOGY

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Biomarkers for Antiangiogenic Therapies

unbound VEGF-A, it is not clear how an increase in VEGF-A expression could contribute to resistance. Conversely, bevacizumab quite consistently led to persistent increases in PlGF levels,32,48 with patients exhibiting a two-fold increase showing improved clinical benet.53 Interestingly, in a cediranib study, levels of PlGF and basic FGF (bFGF) were also associated with radiographic response or survival.53 A landmark study in NSCLC measured 31 CAFs and investigated the relationship between CAF changes and tumor shrinkage during treatment with the anti-VEGFR TKI pazopanib.54 Pazopanib induced signicant changes in eight CAFs, of which changes in plasma sVEGFR2 and IL-4 correlated signicantly with tumor shrinkage. Finally, several studies collecting plasma at disease progression under bevacizumab observed that levels of circulating bFGF,48 PDGF-BB,48 VEGF-C,51 and VEGF-D51 increased on progression. Overall, these data conrm that upregulation of alternative proangiogenic signaling pathways may act as a mechanism of evasive resistance against bevacizumab (Table 2; Data Supplement) and may identify patients that develop adaptive resistance. A pending question is whether this compensatory upregulation can efciently and cost-effectively be used in a clinical setting.

CIRCULATING ENDOTHELIAL CELLS AND PROGENITOR CELLS

Most circulating endothelial cells (CECs) exhibit a mature phenotype and represent apoptotic cells derived from the endothelial wall. A subpopulation of CECs consists of circulating endothelial progenitors (CEPs) that are derived from bone marrow and exhibit a proliferative potential.55 Tumor angiogenesis driven by VEGF-A depends at least partly on the mobilization of CEPs, which integrate into growing tumors and contribute to the formation of a functional vascular bed. Increased concentrations of CEPs may reect active tumor angiogenesis and could serve as predictive markers for antiangiogenic therapies.58 Preliminary data correlating changes in CEC levels with response to bevacizumab are conicting (Data Supplement), probably because of the different cancers analyzed and the various methodologies used. In particular, there is continuing debate about the ideal CEC marker that should be used for ow cytometry.55 In addition, chemotherapy backbones might differentially affect the tumor vasculature and inuence the uctuations in CEC and CEP levels.55 Finally, the number of samples analyzed was small, and data remain to be conrmed in larger studies.

not in other smaller single-arm studies (Data Supplement). Other studies assessed expression of the VEGFRs (including the co-receptor neuropilin-1 [NRP1]), VEGF ligands, or other angiogenic factors (bFGF, IL-8) by immunohistochemistry (Data Supplement). Interestingly, in the mCRC Australian Gastro-Intestinal Trials Group testing Mitomycin, Avastin, Xeloda (AGITG MAX), low baseline VEGFR1 expression correlated with improved OS after bevacizumab (Table 3), whereas in AVAGAST, low VEGFR1 correlated with improved PFS (HR, 0.67 v 0.89) but not OS.39 Furthermore, in more than 700 gastric tumors from AVAGAST, low baseline NRP1 expression correlated with reduced OS in patients receiving placebo but with prolonged OS in patients receiving bevacizumab.60 In particular, patients with low NRP1 showed improved OS (HR, 0.75) versus patients with high NRP1 (HR, 1.07). In mBC and CRC, low NRP1 had similar effects on disease progression (Table 3). Notably, the latter studies involved large tumor sets, indicating that low NRP1 expression represents one of the most consistent and promising markers identied thus far.58,59 Emerging evidence also suggests that stromal cells play an important role in mediating response to antiangiogenic therapies. In particular, the presence of Gr-1CD11b myeloid cells renders murine tumors refractory to antiVEGF-A therapy.61 Myeloid cells provide a rich reserve of angiogenic molecules and possess potent immunosuppressive activity, both of which favor tumor progression.62,63 Tumor-associated broblasts also upregulate PDGF-C expression after delivery of a neutralizing VEGF-A antibody in murine lymphoma models, thereby ensuring the continued formation of tumor vessels.64 In xenograft lung adenocarcinoma models, gene expression changes associated with acquired resistance to bevacizumab occur predominantly in stromal cells.65 In particular, components of the epidermal growth factor receptor (EGFR) and broblast growth factor receptor (FGFR) pathways were upregulated. Few human studies have assessed expression changes in tumor-associated cells during bevacizumab therapy. A seminal study in rectal cancer revealed that bevacizumab upregulates stromal cell derived factor 1 (SDF-1), its receptor CXCR4, and CXCL6 in tumor cells and also upregulates ANG-1 but downregulates NRP1 in tumor-associated macrophages.66 Notably, similar observations were reported for hepatocellular carcinoma treated with sunitinib, an anti-VEGFR TKI.67 Overall, these studies highlight the critical role of stromal cells in regulating resistance to bevacizumab. It is less clear, however, how to evaluate the presence of stromal cells or the markers that they express in a routine clinical setting during treatment.
IMAGING THE RESPONSE TO BEVACIZUMAB

PREDICTIVE MARKERS IN THE TUMOR AND ITS ENVIRONMENT

Several studies assessed expression of VEGF pathway genes in tumor and stromal cells. VEGF-A expression on tumor cells was studied by using immunohistochemistry, real-time polymerase chain reaction, and in situ hybridization, but few positive correlations were observed (Data Supplement). Because microvascular density correlates with VEGF-A expression and serves as a surrogate marker of tumor angiogenesis, it was also investigated. In the NO16966 randomized study, a higher density of CD31 vessels was associated with greater benet from bevacizumab. These ndings were conrmed in one study56 but
www.jco.org

Because VEGF-A blockade is believed to reduce tumor vascular permeability and perfusion, dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), which monitors changes in vascular structure and function, represents an attractive biomarker to assess bevacizumab treatment response. The majority of DCE-MRI studies performed so far involved phase I dose-escalation studies in patients who had already received extensive treatment with chemotherapy or in investigator-led trials that test a single or narrow dose range of bevacizumab.68 Nevertheless, statistically signicant changes in microvascular physiology, involving mainly reductions in the volume
2013 by American Society of Clinical Oncology

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Lambrechts et al

Table 3. In Situ Biomarkers in Tumor or Stroma Predictive of Bevacizumab Treatment Outcome Cancer Type Colorectal Study Acronym NO16966 Sample Size 247 Quantication Method IHC on tumor Association of Biomarker With Clinical Outcome No (PFS/OS)

Marker VEGFR1

Reference Foernzler et al
59

Phase III

Study Details 2 2 factorial design: XELOX v FOLFOX, and bevacizumab v placebo Three arms, receiving either bevacizumab, mitomycin, or placebo, each combined with capecitabine Two arms, receiving either bevacizumab or placebo, each combined with capecitabinecisplatin 2 2 factorial design: XELOX v FOLFOX, and bevacizumab v placebo Two arms, receiving either bevacizumab or placebo, each combined with capecitabinecisplatin Two arms, receiving either bevacizumab or placebo, each combined with capecitabine

Weickhardt et al57

AGITG MAX

268

III

IHC on tumor

No (PFS), yes (OS); low VEGFR1 increases benet from bevacizumab

Gastric

Van Cutsem et al39

AVAGAST

763

III

IHC on tumor

Yes (PFS), no (OS)

NRP1

Colorectal

Foernzler et al59

NO16966

247

III

IHC on tumor

Low NRP1 increases benet from bevacizumab Low NRP1 is negative prognostic and positive predictive for OS Trend toward improved PFS in low NRP1expressing patients

Gastric

Van Cutsem et al39

AVAGAST

763

III

IHC on tumor

Breast

Jubb et al58

AVF2119g

223

III

IHC on tumor

NOTE. All peer-reviewed publications available from PubMed and abstracts presented at international meetings were screened, but only a limited number of studies are included in this table. Preference was given to studies assessing a large number of patients, studies including a placebo-controlled arm, and markers for which consistent ndings were reported in several studies. Abbreviations: AGITG MAX, Australasian Gastro-Intestinal Trials Group Mitomycin, Avastin, Xeloda trial; AVAGAST, Avastin in Advanced Gastric Cancer; FOLFOX, folinic acid uorouracil oxaliplatin; IHC, immunohistochemistry; N/R, not reported; NRP1, neuropilin-1; OS, overall survival; PFS, progression-free survival; VEGFR1, vascular endothelial growth factor receptor 1; XELOX, capecitabine oxaliplatin.

transfer constant Ktrans, were observed after bevacizumab monotherapy in at least nine studies. Similar results have been reported in studies involving sunitinib and pazopanib.69 Although these data are clearly encouraging, additional research is needed. In particular, current challenges involve standardization of DCE-MRI to enable its application in a wider context, such as in multicenter clinical studies. Whether changes in tumor blood ow measured by DCE-MRI predict outcome of bevacizumab in combination with chemotherapy still needs to be assessed.
GENETIC SUSCEPTIBILITY AS A BIOMARKER FOR BEVACIZUMAB OUTCOME

Unlike tumor cells, in which genes are mutated, deleted, or amplied, tumor endothelial cells are genetically stable. The response of the tumor vasculature to bevacizumab could therefore be considered a host-mediated process inuenced by genetic variability in the host DNA. Several studies have meanwhile assessed whether single nucleotide polymorphisms (SNPs) in candidate genes predict bevacizumab treatment outcome (Data Supplement).
6
2013 by American Society of Clinical Oncology

In the mBC E2100 trial, mutant carriers of the rs699947 and rs1570360 SNPs, which correlate with reduced expression of VEGF-A,69 predicted favorable median OS in the bevacizumab arm but not in the control arm.70 Surprisingly, neither SNP predicted superior PFS for either arm. In the AVADO trial, it was found that rs699947, but not rs1570360, correlated with PFS in the placebo arm,71 whereas in patients with mCRC, rs699947 and rs1570360 correlated with OS in the bevacizumab arm.72 Findings for both SNPs are thus inconsistent. Several other SNPs in VEGFA, including rs833061 or rs3025039, have been proposed as predictive markers, but except for rs499946, which is located near rs699947 and which was conrmed in a meta-analysis of ve studies,73 most SNPs have not been conrmed (Table 4). Various SNPs in other angiogenic factors have been assessed, but only SNPs in VEGFR2, IL8, and CXCR2 were replicated in at least one other study. In particular, the rs2305948 SNP in VEGFR2, which encodes a Val273Ile substitution that reduces binding of VEGF-A to VEGFR2 in mutant carriers,75 has been correlated with reduced bevacizumab treatment outcome.76 The mutant A allele of rs4073 in the IL8
JOURNAL OF CLINICAL ONCOLOGY

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Biomarkers for Antiangiogenic Therapies

Table 4. Genetic Markers Evaluated As Predictive Biomarkers for Bevacizumab Treatment Outcome Gene VEGFA Cancer Type Reference Study Acronym Sample Size Phase 218 (111 with bevacizumab) Study Details All patients received FOLFIRI bevacizumab Genetic Variant Association With Clinical Outcome 460C increases PFS/ OS in bevacizumab patients No (OR)

Colorectal Loupakis et al74

Hansen et al83

Nordic ACT

218

Koutras et al72

209

Gerger et al76

119

Lung

Zhang et al81

ECOG 4599

133 (66 with bevacizumab)

Breast

Schneider et al70

ECOG 2100

363 (180 with bevacizumab)

Miles et al71

AVADO

336 (231 with bevacizumab)

2578A/C (rs699947), 460C/T (rs833061), 634G/C (rs2010963), 936C/T (rs3025039) III Two arms receiving either Five SNPs (not further specied) FOLFOX/XELOX or FOLFIRI/XELIRI bevacizumab III All patients received 634G/C (rs2010936), 936C/T bevacizumab (rs3025039), 1154G/A FOLFIRI or XELIRI (rs1570360), 2578A/C (rs699947) All patients received 634G/C (rs2010936), 936C/T FOLFOX/XELOX (rs3025039), 1154G/A bevacizumab (rs1570360), -460C/T (rs833061), 2578A/C (rs699947) II/III Two arms, receiving 634G/C either bevacizumab or placebo, each combined with carboplatin-paclitaxel III Two arms, receiving 634G/C (rs2010936), 936C/T paclitaxel (rs3025039), 1154G/A bevacizumab (rs1570360), -460C/T (rs833061), 2578A/C (rs699947) III Three arms, receiving 634G/C (rs2010936), 936C/T either 0, 7.5, or 15 mg/ (rs3025039), 1154G/A kg bevacizumab, each (rs1570360), -460C/T combined with (rs833061), 2578A/C docetaxel (rs699947) 158 SNPs in VEGF pathway Various regimens (see Escudier et al, 5, Miles 12 et al, and Saltz et al 27 for details) III Two arms, receiving 158 SNPs in VEGF pathway gemcitabine-erlotinib bevacizumab

2578A and 1154A increase OS but not PFS No (response/PFS/ OS)

Various

Lambrechts NO16966, AVAiL, 1,348 (669 with AViTA, bevacizumab) et al73 AVOREN, AVADO 154 (77 with bevacizumab)

VEGFR1 Pancreatic Lambrechts AViTA et al82

Colorectal Hansen et al83

Nordic ACT

218

III

Breast

Miles et al71

AVADO

336 (231 with bevacizumab)

III

VEGFR2 Colorectal Gerger et al76 Breast Schneider et al70 ECOG 2100

119

363 (180 with bevacizumab)

III

Various

Lambrechts NO16966, AVAiL, 1,348 (669 with AViTA, bevacizumab) et al73 AVOREN, AVADO 35

Two arms receiving either FOLFOX/XELOX or FOLFIRI/XELIRI bevacizumab Three arms, receiving either 0, 7.5, or 15 mg/ kg bevacizumab combined with docetaxel All patients received FOLFOX/XELOX bevacizumab Two arms, receiving paclitaxel bevacizumab Various regimens (see Escudier et al, 5, Miles et al,12 and Saltz et al 27 for details) All patients received chemotherapy (not specied) bevacizumab

rs9582036 (intronic)

-634G/C correlates with PFS in bevacizumab patients 2578A and 1154A increase OS but not PFS in bevacizumab patients 2578C increases PFS in placebo arm; 1154A increases PFS in bevacizumab arm (trend) rs699946-A allele correlates with improved PFS in bevacizumab patients rs9582036-A allele increases PFS and OS in bevacizumab arm but not in placebo arm rs9582036-A allele increases RR to bevacizumab No (PFS/OS)

rs9554316, rs9582036 (both intronic)

889G/A 889G/A, 1416A/T

Yes (RR), G allele increases RR No (PFS/OS)

158 SNPs in VEGF pathway

IL8

Colorectal Giudice et al78

251T/A

rs11133360-T allele correlates with improved PFS in bevacizumab patients TT carriers have improved OR

(continued on following page)

www.jco.org

2013 by American Society of Clinical Oncology

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Lambrechts et al

Table 4. Genetic Markers Evaluated As Predictive Biomarkers for Bevacizumab Treatment Outcome Gene Cancer Type Ovarian Reference Schultheis et al79 Zhang et al81 Study Acronym PH-II-45 Sample Size Phase 53 II Study Details 251T/A Genetic Variant Association With Clinical Outcome TT carriers have increased RR Yes (PFS in bevacizumab patients) Yes (PFS), C allele increases PFS Yes (RR), C allele increases RR

CXCR2

Lung

ECOG 4599

133 (66 with bevacizumab)

Ovarian

Schultheis et al79

PH-II-45

53

Colorectal Gerger et al76

119

All patients received cyclophosphamide bevacizumab II/III Two arms, receiving either bevacizumab or placebo, each combined with carboplatin-paclitaxel II All patients received cyclophosphamide bevacizumab All patients received FOLFOX/XELOX bevacizumab

785C/T

785C/T 785C/T

NOTE. All peer-reviewed publications available from PubMed and abstracts presented at international meetings were screened, but only a limited number of studies are included in this table. Preference was given to studies assessing a large number of patients, studies including a placebo-controlled arm, and markers for which consistent ndings were reported in several studies. Abbreviations: ACT, doxorubicin cyclophosphamide paclitaxel; AVADO, Avastin and Docetaxel in Metastatic Breast Cancer; AVAiL, Avastin in Lung Cancer; AViTA, Avastin and Tarceva in Advanced Pancreatic Cancer; AVOREN, Avastin and Roferon in Renal Cell Carcinoma; CXCR2, chemokine receptor 2; ECOG, Eastern Cooperative Oncology Group; FOLFIRI, folinic acid uorouracil irinotecan; FOLFOX, folinic acid uorouracil oxaliplatin; IL8, interleukin-8; OR, objective response; OS, overall survival; PFS, progression-free survival; RR, response rate; SNP, single nucleotide polymorphism; VEGF, vascular endothelial growth factor; VEGFA, vascular endothelial growth factor A; VEGFR1, VEGF receptor 1; VEGFR2, VEGF receptor 2; XELIRI, capecitabine irinotecan; XELOX, capecitabine oxaliplatin.

promoter, which correlates with increased IL-8 production after stimulation with lipopolysaccharide,77 is associated with poor response to bevacizumab and pazopanib.78-80 Finally, rs2230054 in CXCR2, which could affect splicing of CXCR2,76 has been correlated with reduced PFS in at least three independent studies.76,79,81 In nearly all studies, limited numbers of SNPs were selected on the basis of candidate gene approaches. As a result, selected SNPs differ between studies, leading to heterogeneous data sets and few possibilities for assessing data consistency. Recently, the rst approaches to systematically cover SNPs in all genes of the VEGF pathway were reported.82 In particular, up to 158 SNPs located in 14 genes were assessed in the AViTA randomized pancreatic cancer trial. Four SNPs in VEGFR1 correlated highly signicantly with both PFS and OS in the bevacizumab arm (per allele HR, 2.1). No effect was observed in the placebo arm. Fine mapping of this locus identied rs7993418, a synonymous SNP affecting tyrosine 1213 in the tyrosine kinase domain of VEGFR1, as the functional variant underlying the association.82 Mutant carriers of the rs7993418 allele increased VEGFR1 expression by almost 20% and exhibited worse outcome after bevacizumab, thereby conrming previous observations for VEGFR1 expression in plasma or tumors. Intriguingly, this association was also replicated in patients with RCC, NSCLC, and CRC but not in those with BC.71,72 Overall, although interesting SNPs have been identied, several questions remain. First, interesting SNPs still need to be replicated in as many trials as possible. Second, with the exception of the VEGFR1 locus, predictive effects of most individual SNPs have been rather modest, raising the question of whether SNPs are sufciently informative to assist with patient selection. One possible way to increase their predictive effect would be to consolidate effects of individual SNPs into a combined predictive score.
8
2013 by American Society of Clinical Oncology

CONCLUDING REMARKS AND FUTURE DIRECTIONS

The most promising markers for bevacizumab treatment outcome consist of circulating short VEGF-A isoforms and modied expression of VEGFRs (VEGFR1 or NRP1), either in plasma or tumors. Although there is abundant functional evidence that these markers could indeed determine bevacizumab outcome (Note 1 and Fig 1), none of them has consistently been replicated across different studies involving various cancer types. Perhaps this is not surprising, since biomarkers for bevacizumab may differ between cancer types. The question is whether additional biomarker discovery to more consistently predict bevacizumab treatment outcome across cancer types is still needed or whether prospective translation of existing markers into clinical practice should be considered a priority. Ideally, all biomarkers that most consistently replicate in a particular cancer type should be considered for prospective validation. Currently, this strategy is not feasible because there are too few large-scale biomarker studies in the same cancer that assess the same panel of markers. Therefore, future studies should continue to focus on the discovery of novel biomarkers in different tissue types (plasma, DNA, tumor, and so on) and should not be limited to testing one or two markers in plasma or genotyping only a few genetic variants. Instead, they should include homogeneous sets of candidate markers to allow for comparison between studies. Furthermore, since it is becoming obvious that single biomarkers may not be sufcient to predict the complex phenotype of response to bevacizumab, studies should integrate the individual effects of these markers by using advanced statistical analyses and combine them into a general predictive score. Possibly, such integrated analyses will generate more robust predictions that are valid across cancers. Markers in plasma, DNA, or tumor tissue might even have to be combined to obtain accurate
JOURNAL OF CLINICAL ONCOLOGY

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Biomarkers for Antiangiogenic Therapies

A
Short VEGFisoform driven chaotic tumor vasculature

Response to Bevacizumab
Leaky tumor vasculature

Diffusion of short VEGF-isoforms

Low VEGFR1 and sVEGFR1 expression

Low NRP1 expression

B
Long VEGFisoform driven tumor vasculature

Resistance to Bevacizumab
Macrophage and myeloid cell infiltration Well-perfused tumor vasculature Pericyte coverage

High VEGFR1 and sVEGFR1 expression Limited diffusion of long VEGF isoforms

Recruitment of bone marrowderived myeloid cells

Compensation by proangiogenic factors

High NRP1 expression

Fig 1. (A) Characteristics of a tumor responsive to bevacizumab based on current biomarker data. High expression of the short vascular endothelial growth factor A (VEGF-A) isoform VEGF121 leads to a chaotic vessel structure (highly irregular diameter and very leaky), naked tumor vessels (no pericyte coverage), and a low perfusion index. Expression of VEGFR1 on tumors, vessels, and macrophages is low. Soluble VEGFR1 (sVEGFR1) expression in plasma is low. Neuropilin-1 (NRP1) expression is low on tumor cells. There are few stromal cells, such as macrophages and broblasts, that secrete VEGF-A present in the tumor. (B) Characteristics of a tumor resistant to bevacizumab based on current biomarker data. High expression of long VEGF-A isoforms (VEGF165 and VEGF189) leads to a less chaotic vessel structure (normal diameter and not leaky) and to tumor vessels that are covered with some pericytes and have a high perfusion index. Expression of VEGFR1 on tumors, vessels, and macrophages is high. sVEGFR1 expression in plasma is high. Expression of NRP1 on tumor cells is also high. There are many stromal cells in the tumor, including macrophages and broblasts. These will secrete various other angiogenic molecules, including interleukin-8 (IL-8), basic broblast growth factor (bFGF), platelet-derived growth factor BB (PDGF-BB), PDGF-C, VEGF-C, and VEGF-D.

VEGFR2/NRP1 complex Short VEGF-isoform (VEGF121) VEGF-isoform 165 (VEGF165) Long VEGF-isoform (VEGF189)

Circulating angiogenic factors (IL-8, PDGF-C, VEGF-C, VEGF-D, bFGF, etc.) Transmembrane VEGFR1 Soluble VEGFR1 (sVEGFR1) Macrophage, myeloid cell

Tumor cell Endothelial cell Pericyte

predictive models at the risk of compromising their clinical applicability. Obviously, it will remain challenging to discover which set of individual markers has the highest sensitivity and specicity to predict outcome of treatment with bevacizumab. A potential solution might be to explore biomarkers in relation to vascular responses determined by DCE-MRI as an alternative end point of treatment response. Another possibility is to focus on cancers in which chemotherapy does not signicantly contribute to the outcome of bevacizumab (eg, RCC, which is highly dependent on VEGF for disease progression) or in which bevacizumab is given as monotherapy. One could also focus on identifying markers for anti-VEGFR TKIs that are delivered as monotherapies and for
www.jco.org

which objective responses can be directly attributed to the TKI. Most TKIs are not limited to inhibiting VEGFRs, and predictive markers may therefore only partly overlap with those predictive for bevacizumab outcome. Finally, one could consider screening for markers in studies that assess bevacizumab for the treatment of age-related macular degeneration.84 In the latter, VEGF-A inhibition counteracts the excessive growth of leaky blood vessels leading to prevention and reversal of vision loss. Because neoangiogenesis in age-related macular degeneration differs from tumor angiogenesis, only a limited number of markers might overlap. International collaborative efforts will soon release the mutation and methylation proles of thousands of tumors allowing the identication of novel molecular subtypes.
2013 by American Society of Clinical Oncology

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Lambrechts et al

Because these novel subtypes can indirectly modulate tumor angiogenesis, it will be necessary to assess them as markers of treatment outcome. Clinical trials in BC are among the rst to take molecular subtyping into accountBEATRICE is restricted to triplenegative BCwhereas the AVADO and AVEREL trials are limited to human epidermal growth factor receptor (HER) negative and HER-positive patients, respectively. In other cancers, some initial subtyping has retrospectively been performed, for instance, by genotyping KRAS and BRAF mutations in CRC, but it has failed to reveal any differential effects.85 Many other subtypes, such as microsatellite instability or hypermethylator phenotypes, have largely been neglected and will need to be assessed in the future. Conversely, there might already be sufcient evidence to prospectively validate existing biomarkers. In particular, markers that have been replicated in several different cancers are eligible. The short VEGF-A isoforms are an example of such biomarkers, because they have been correlated with outcome in three large randomized studies, but they failed to replicate in a few other studies (possibly because plasma was not appropriately collected). The phase III MERiDiAN trial in mBC (opening in 2012) will evaluate the impact of bevacizumab in patients stratied for plasma short VEGF-A isoforms. This trial represents the rst prospective validation of a biomarker for bevacizumab and, if positive, MERiDiAN will result in the clinical application of short VEGF-A isoforms as a biomarker for bevacizumab in mBC. In addition, MERiDiAN will indicate whether prospective validation of other markers replicated in several studies is meaningful (eg, VEGFR1 or NRP1 expression in mCRC or gastric cancer). Meanwhile, in a commitment to discover and validate other biomarker candidates, extensive biomarker programs should continue in several indications. As long as these markers have not prospectively been validated, the available clinical trial data provide the most compelling evidence for prescribing bevacizumab.

AUTHORS DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Although all authors completed the disclosure declaration, the following author(s) and/or an authors immediate family member(s) indicated a nancial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a U are those for which no compensation was received; those relationships marked with a C were compensated. For a detailed description of the disclosure categories, or for more information about ASCOs conict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conicts of Interest section in Information for Contributors. Employment or Leadership Position: Sanne de Haas, F. Hoffmann-La Roche (C); Stefan J. Scherer, F. Hoffmann-La Roche (C) Consultant or Advisory Role: Heinz-Josef Lenz, Bayer (C), Genentech/Roche (C), Merck (C), sano-aventis (C); Peter Carmeliet, F. Hoffmann-La Roche (C) Stock Ownership: None Honoraria: Diether Lambrechts, F. Hoffman-La Roche; Heinz-Josef Lenz, Bayer, Genentech/Roche, Merck, sano-aventis; Peter Carmeliet, Roche Research Funding: Diether Lambrechts, F. Hoffman-La Roche; Heinz-Josef Lenz, Bayer, Genentech/Roche, F. Hoffmann-La Roche; Peter Carmeliet, F. Hoffmann-La Roche Expert Testimony: None Other Remuneration: None Other: Diether Lambrechts, patent EP 10744513.2 (publication No. EP 2462242) (U), patent US 13/388,840 (publication No. US 2012-0195858) (U); Sanne de Haas, patent EP 10744513.2 (publication No. EP 2462242) (U), patent US 13/388,840 (publication No. US 2012-0195858) (U); Peter Carmeliet, patent EP 10744513.2 (publication No. EP 2462242) (U), patent US 13/388,840 (publication No. US 2012-0195858) (U); Stefan J. Scherer, patent EP 10744513.2 (publication No. EP 2462242) (U), patent US 13/388,840 (publication No. US 2012-0195858) (U)

AUTHOR CONTRIBUTIONS
Conception and design: Diether Lambrechts, Sanne de Haas, Peter Carmeliet, Stefan J. Scherer

PRECLINICAL EVIDENCE SUPPORTING PROMISING BIOMARKERS FOR BEVACIZUMAB

High Plasma Levels of Short VEGF-A Isoforms 86 Because total body production of VEGF-A eclipses VEGF-A production from tumors, plasma VEGF-A is unlikely to provide a sensitive index of tumor-secreted VEGF-A. A modied ELISA that favors detection of short VEGF-A isoforms might, because these isoforms diffuse over long distances, be more sensitive in detecting vascular dependence of the tumor. Reduced VEGFR1 Expression in Plasma or Tumors VEGFR1 triggers angiogenesis either directly by transmitting intracellular signals or indirectly by inducing trans-phosphorylation of VEGFR2.87 VEGFR1 expression is upregulated in tumors in which it contributes to tumor survival. 88 A neutralizing PlGF antibody suppresses angiogenesis and recruitment of inammatory cells in tumor models. 89 High sVEGFR1 levels sequester tumor-derived VEGF-A, thereby limiting the benets of VEGF-A neutralization through bevacizumab. Aibercept, a fusion protein composed of VEGFR1 and VEGFR2 ligand-binding components fused to the fragment crystallizable portion of human immunoglobulin G1 (IgG1), inhibits all VEGFR1 ligands (VEGF-A, PlGF, and VEGF-B) and prolongs PFS in second-line mCRC.90 Importantly, approximately 30% of patients received prior bevacizumab, thereby illustrating the clinical relevance of VEGFR1 signaling in the context of VEGF inhibition. Low NRP1 Expression in Tumors 91 Antibodies blocking the binding between VEGF-A and NRP1 slow tumor growth in mice. 92 A combination of anti-NRP1 and antiVEGF-A antibodies enhances tumor growth and vascular density reduction in mice. 93 VEGF-A stimulates cancer stemness and renewal through NRP1.

10

2013 by American Society of Clinical Oncology

JOURNAL OF CLINICAL ONCOLOGY

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Biomarkers for Antiangiogenic Therapies

Collection and assembly of data: Diether Lambrechts, Heinz-Josef Lenz, Sanne de Haas, Stefan J. Scherer Data analysis and interpretation: All authors

Manuscript writing: All authors Final approval of manuscript: All authors

REFERENCES
1. Carmeliet P, Jain RK: Molecular mechanisms and clinical applications of angiogenesis. Nature 473:298-307, 2011 2. Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, uorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335-2342, 2004 3. Sandler A, Gray R, Perry MC, et al: Paclitaxel-carboplatin alone or with bevacizumab for non-small-cell lung cancer. N Engl J Med 355: 2542-2550, 2006 4. Escudier B, Bellmunt J, Ne grier S, et al: Phase III trial of bevacizumab plus interferon alfa-2a in patients with metastatic renal cell carcinoma (AVOREN): Final analysis of overall survival. J Clin Oncol 28:2144-2150, 2010 5. Escudier B, Pluzanska A, Koralewski P, et al: Bevacizumab plus interferon alfa-2a for treatment of metastatic renal cell carcinoma: A randomised, doubleblind phase III trial. Lancet 370:2103-2111, 2007 6. Yang JC, Haworth L, Sherry RM, et al: A randomized trial of bevacizumab, an anti-vascular endothelial growth factor antibody, for metastatic renal cancer. N Engl J Med 349:427-434, 2003 7. Burger RA, Brady MF, Bookman MA, et al: Incorporation of bevacizumab in the primary treatment of ovarian cancer. N Engl J Med 365:2473-2483, 2011 8. Perren TJ, Swart AM, Psterer J, et al: A phase 3 trial of bevacizumab in ovarian cancer. N Engl J Med 365:2484-2496, 2011 9. Kreisl TN, Kim L, Moore K, et al: Phase II trial of single-agent bevacizumab followed by bevacizumab plus irinotecan at tumor progression in recurrent glioblastoma. J Clin Oncol 27:740-745, 2009 10. Miller K, Wang M, Gralow J, et al: Paclitaxel plus bevacizumab versus paclitaxel alone for metastatic breast cancer. N Engl J Med 357:2666-2676, 2007 11. Robert NJ, Die ras V, Glaspy J, et al: RIBBON-1: Randomized, double-blind, placebo-controlled, phase III trial of chemotherapy with or without bevacizumab for rst-line treatment of human epidermal growth factor receptor 2-negative, locally recurrent or metastatic breast cancer. J Clin Oncol 29:1252-1260, 2011 12. Miles DW, Chan A, Dirix LY, et al: Phase III study of bevacizumab plus docetaxel compared with placebo plus docetaxel for the rst-line treatment of human epidermal growth factor receptor 2-negative metastatic breast cancer. J Clin Oncol 28:3239-3247, 2010 13. Rugo HS: Inhibiting angiogenesis in breast cancer: The beginning of the end or the end of the beginning? J Clin Oncol 30:898-901, 2012 14. Potti A, Schilsky RL, Nevins JR: Refocusing the war on cancer: The critical role of personalized treatment. Sci Transl Med 2:28cm13, 2010 15. Potente M, Gerhardt H, Carmeliet P: Basic and therapeutic aspects of angiogenesis. Cell 146: 873-887, 2011 16. Ellis LM, Hicklin DJ: VEGF-targeted therapy: Mechanisms of anti-tumour activity. Nat Rev Cancer 8:579-591, 2008 17. Willett CG, Boucher Y, di Tomaso E, et al: Direct evidence that the VEGF-specic antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med 10:145-147, 2004 www.jco.org

18. Muggia F, Safra T, Borgato L, et al: Pharmacokinetics (PK) of pegylated liposomal doxorubicin (PLD) given alone and with bevacizumab (B) in patients with recurrent epithelial ovarian cancer (rEOC). J Clin Oncol 28:406s, 2010 (suppl; abstr 5064) 19. Jain RK: Normalizing tumor vasculature with anti-angiogenic therapy: A new paradigm for combination therapy. Nat Med 7:987-989, 2001 20. Batchelor TT, Sorensen AG, di Tomaso E, et al: AZD2171, a pan-VEGF receptor tyrosine kinase inhibitor, normalizes tumor vasculature and alleviates edema in glioblastoma patients. Cancer Cell 11:83-95, 2007 21. Fischer I, Cunliffe CH, Bollo RJ, et al: Highgrade glioma before and after treatment with radiation and Avastin: Initial observations. Neuro Oncol 10:700-708, 2008 22. Goel S, Duda DG, Xu L, et al: Normalization of the vasculature for treatment of cancer and other diseases. Physiol Rev 91:1071-1121, 2011 23. Van Cutsem E, Lambrechts D, Prenen H, et al: Lessons from the adjuvant bevacizumab trial on colon cancer: What next? J Clin Oncol 29:1-4, 2011 24. Carmeliet P, Jain RK: Principles and mechanisms of vessel normalization for cancer and other angiogenic diseases. Nat Rev Drug Discov 10:417-427, 2011 25. Grothey A, Hedrick EE, Mass RD, et al: Response-independent survival benet in metastatic colorectal cancer: A comparative analysis of N9741 and AVF2107. J Clin Oncol 26:183-189, 2008 26. Chun YS, Vauthey JN, Boonsirikamchai P, et al: Association of computed tomography morphologic criteria with pathologic response and survival in patients treated with bevacizumab for colorectal liver metastases. JAMA 302:2338-2344, 2009 27. Saltz LB, Clarke S, Daz-Rubio E, et al: Bevacizumab in combination with oxaliplatin-based chemotherapy as rst-line therapy in metastatic colorectal cancer: A randomized phase III study. J Clin Oncol 26:2013-2019, 2008 28. Bottsford-Miller JN, Coleman RL, Sood AK: Resistance and escape from antiangiogenesis therapy: Clinical implications and future strategies. J Clin Oncol 30:4026-4034, 2012 29. Goede V, Coutelle O, Neuneier J, et al: Identication of serum angiopoietin-2 as a biomarker for clinical outcome of colorectal cancer patients treated with bevacizumab-containing therapy. Br J Cancer 103:1407-1414, 2010 30. Dowlati A, Gray R, Sandler AB, et al: Cell adhesion molecules, vascular endothelial growth factor, and basic broblast growth factor in patients with non-small cell lung cancer treated with chemotherapy with or without bevacizumab: An Eastern Cooperative Oncology Group Study. Clin Cancer Res 14:1407-1412, 2008 31. Hegde PS, Jubb AM, Chen D, et al: Predictive impact of circulating vascular endothelial growth factor in 4 phase III trials evaluating bevacizumab. Clin Cancer Res [epub ahead of print on November 20, 2012] 32. Willett CG, Duda DG, di Tomaso E, et al: Efcacy, safety, and biomarkers of neoadjuvant bevacizumab, radiation therapy, and uorouracil in rectal cancer: A multidisciplinary phase II study. J Clin Oncol 27:3020-3026, 2009 33. Tolaney SM, Duda DG, Boucher Y, et al: A phase II study of preoperative (preop) bevacizumab (bev) followed by dose-dense (dd) doxorubicin (A)/ cyclophosphamide (C)/paclitaxel (T) in combination with bev in HER2-negative operable breast cancer (BC). J Clin Oncol 30:55s, 2012 (suppl; abstr 1026)

34. Leighl N, Reck M, de Haas S, et al: Analysis of biomarkers (BMs) in the AVAiL phase III randomised study of rst-line Bevacizumab (Bv) with cisplatingemcitabine (CG) in patients (pts) with non-small cell lung cancer (NSCLC). Eur J Cancer 7:558, 2009 (abstr 9172) 35. Mok T, Gorbunova V, Juhasz E, et al: Biomarker analysis in BO21015, a phase II randomised study of rst-line bevacizumab (BEV) combined with carboplatin-gemcitabine (CG) or carboplatin-paclitaxel (CP) in patients (pts) with advanced or recurrent nonsquamous non-small cell lung cancer (NSCLC). Eur J Cancer 47:S592, 2011 (suppl 1; abstr 9003) 36. Van Cutsem E, Jayson G, Dive C, et al: Analysis of blood plasma factors in the AVITA phase III randomized study of bevacizumab (bev) with gemcitabine-erlotinib (GE) in patients (pts) with metastatic pancreatic cancer (mPC). Eur J Cancer 47: S95-S96, 2011 (suppl 1; abstr 803) 37. Miles DW, de Haas SL, Dirix L, et al: Plasma biomarker analyses in the AVADO phase III randomized study of rst-line bevacizumab docetaxel in patients with human epidermal growth factor receptor (HER) 2-negative metastatic breast cancer. Cancer Res 70, 2010 (suppl 2; abstr P2-16-04) 38. Jayson GC, de Haas S, Delmar P, et al: Evaluation of plasma VEGF-A as a potential predictive pan-tumour biomarker for bevacizumab. Proceedings of the 2011 European Multidisciplinary Cancer Congress, Stockholm, Sweden, September 23-27, 2011 (abstr 804) 39. Van Cutsem E, de Haas S, Kang YK, et al: Bevacizumab in combination with chemotherapy as rst-line therapy in advanced gastric cancer: A biomarker evaluation from the AVAGAST randomized phase III trial. J Clin Oncol 30:2119-2127, 2012 40. Boige V, Malka D, Bourredjem A, et al: Efcacy, safety, and biomarkers of single-agent bevacizumab therapy in patients with advanced hepatocellular carcinoma. Oncologist 17:1063-1072, 2012 41. Kaseb AO, Garrett-Mayer E, Morris JS, et al: Efcacy of bevacizumab plus erlotinib for advanced hepatocellular carcinoma and predictors of outcome: Final results of a phase II trial. Oncology 82:67-74, 2012 42. Kopetz S, Hoff PM, Morris JS, et al: Phase II trial of infusional uorouracil, irinotecan, and bevacizumab for metastatic colorectal cancer: Efcacy and circulating angiogenic biomarkers associated with therapeutic resistance. J Clin Oncol 28:453-459, 2010 43. Zhu AX, Ancukiewicz M, Supko JG, et al: Clinical, pharmacodynamic (PD), and pharmacokinetic (PK) evaluation of cediranib in advanced hepatocellular carcinoma (HCC): A phase II study (CTEP 7147). J Clin Oncol 30:266s, 2012 (suppl; abstr 4112) 44. Meyerhardt JA, Ancukiewicz M, Abrams TA, et al: Phase I study of cetuximab, irinotecan, and vandetanib (ZD6474) as therapy for patients with previously treated metastastic colorectal cancer. PLoS One 7:e38231, 2012 45. Ruiz de Almodovar C, Lambrechts D, Mazzone M, et al: Role and therapeutic potential of VEGF in the nervous system. Physiol Rev 89:607-648, 2009 46. Baar J, Silverman P, Lyons J, et al: A vasculature-targeting regimen of preoperative docetaxel with or without bevacizumab for locally advanced breast cancer: Impact on angiogenic biomarkers. Clin Cancer Res 15:3583-3590, 2009 47. Horn L, Dahlberg SE, Sandler AB, et al: Phase II study of cisplatin plus etoposide and bevacizumab for previously untreated, extensive-stage small-cell lung cancer: Eastern Cooperative Oncology Group Study E3501. J Clin Oncol 27:6006-6011, 2009 11

2013 by American Society of Clinical Oncology

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Lambrechts et al

48. Buysschaert I, Schmidt T, Roncal C, et al: Genetics, epigenetics and pharmaco-(epi)genomics in angiogenesis. J Cell Mol Med 12:2533-2551, 2008 49. Fuerstenberger G, Boneberg E, Simcock M, et al: Predictive and prognostic potential of angiogenic serum factors and circulating endothelial cells in metastatic melanoma patients receiving temozolamide plus bevacizumab (SAKK 50/07). J Clin Oncol 28:632s, 2010 (suppl; abstr 8585) 50. Smerdel MP, Steffensen KD, Waldstrm M, et al: The predictive value of serum VEGF in multiresistant ovarian cancer patients treated with bevacizumab. Gynecol Oncol 118:167-171, 2010 51. Lieu CH, Tran HT, Jiang Z, et al: The association of alternate VEGF ligands with resistance to anti-VEGF therapy in metastatic colorectal cancer. J Clin Oncol 29:229s, 2011 (suppl; abstr 3533) 52. Loupakis F, Cremolini C, Fioravanti A, et al: Pharmacodynamic and pharmacogenetic angiogenesisrelated markers of rst-line FOLFOXIRI plus bevacizumab schedule in metastatic colorectal cancer. Br J Cancer 104:1262-1269, 2011 53. Batchelor TT, Duda DG, di Tomaso E, et al: Phase II study of cediranib, an oral pan-vascular endothelial growth factor receptor tyrosine kinase inhibitor, in patients with recurrent glioblastoma. J Clin Oncol 28:2817-2823, 2010 54. Nikolinakos PG, Altorki N, Yankelevitz D, et al: Plasma cytokine and angiogenic factor proling identies markers associated with tumor shrinkage in early-stage non-small cell lung cancer patients treated with pazopanib. Cancer Res 70:2171-2179, 2010 55. Bertolini F, Shaked Y, Mancuso P, et al: The multifaceted circulating endothelial cell in cancer: Towards marker and target identication. Nat Rev Cancer 6:835-845, 2006 56. Pohl A, El-Khoueiry A, Yang D, et al: Pharmacogenetic proling of CD133 is associated with response rate (RR) and progression-free survival (PFS) in patients with metastatic colorectal cancer (mCRC), treated with bevacizumab-based chemotherapy. Pharmacogenomics J 10.1038/tpj.2011.61 [epub ahead of print on January 10, 2012] 57. Weickhardt AJ, Williams D, Lee C, et al: Vascular endothelial growth factors (VEGF) and VEGF receptor expression as predictive biomarkers for benet with bevacizumab in metastatic colorectal cancer (mCRC): Analysis of the phase III MAX study. J Clin Oncol 29: 228s, 2011 (suppl; abstr 3531) 58. Jubb AM, Miller KD, Rugo HS, et al: Impact of exploratory biomarkers on the treatment effect of bevacizumab in metastatic breast cancer. Clin Cancer Res 17:372-381, 2011 59. Foernzler D, Delmar P, Kockx M, et al: Tumor tissue based biomarker analysis in NO16966: A randomized phase III study of rst-line bevacizumab in combination with oxaliplatin-based chemotherapy in patients with mCRC. 2010 Gastrointestinal Cancers Symposium, January 22-24, 2010 (abstr 374) 60. Shah M, Kang Y, Ohtsu A, et al: Tumor and blood plasma biomarker analyses in the AVAGAST phase III randomized study of rst-line bevacizumab capecitabine/cisplatin in patients with advanced gastric cancer. Ann Oncol 21:S176, 2010 (suppl 8) 61. Shojaei F, Wu X, Malik AK, et al: Tumor refractoriness to anti-VEGF treatment is mediated by CD11b Gr1 myeloid cells. Nat Biotechnol 25:911-920, 2007 62. Carmeliet P: Angiogenesis in life, disease and medicine. Nature 438:932-936, 2005

63. Loges S, Schmidt T, Carmeliet P: Mechanisms of resistance to anti-angiogenic therapy and development of third-generation anti-angiogenic drug candidates. Genes Cancer 1:12-25, 2010 64. Crawford Y, Kasman I, Yu L, et al: PDGF-C mediates the angiogenic and tumorigenic properties of broblasts associated with tumors refractory to anti-VEGF treatment. Cancer Cell 15:21-34, 2009 65. Cascone T, Herynk MH, Xu L, et al: Upregulated stromal EGFR and vascular remodeling in mouse xenograft models of angiogenesis inhibitorresistant human lung adenocarcinoma. J Clin Invest 121:1313-1328, 2011 66. Xu L, Duda DG, di Tomaso E, et al: Direct evidence that bevacizumab, an anti-VEGF antibody, up-regulates SDF1alpha, CXCR4, CXCL6, and neuropilin 1 in tumors from patients with rectal cancer. Cancer Res 69:7905-7910, 2009 67. Zhu AX, Sahani DV, Duda DG, et al: Efcacy, safety, and potential biomarkers of sunitinib monotherapy in advanced hepatocellular carcinoma: A phase II study. J Clin Oncol 27:3027-3035, 2009 68. OConnor JP, Jackson A, Parker GJ, et al: Dynamic contrast-enhanced MRI in clinical trials of antivascular therapies. Nat Rev Clin Oncol 9:167-177, 2012 69. Lambrechts D, Storkebaum E, Morimoto M, et al: VEGF is a modier of amyotrophic lateral sclerosis in mice and humans and protects motoneurons against ischemic death. Nat Genet 34:383-394, 2003 70. Schneider BP, Wang M, Radovich M, et al: Association of vascular endothelial growth factor and vascular endothelial growth factor receptor-2 genetic polymorphisms with outcome in a trial of paclitaxel compared with paclitaxel plus bevacizumab in advanced breast cancer: ECOG 2100. J Clin Oncol 26:4672-4678, 2008 71. Miles DW, De Haas SL, Romieu G, et al: Polymorphism analysis in the AVADO randomised phase III trial of rst-line bevacizumab (BEV) combined with docetaxel in HER2-negative metastatic breast cancer (mBC). Eur J Cancer 47:S176, 2011 (suppl 1; abstr 1423) 72. Koutras AK, Antonacopoulou AG, Eleftheraki AG, et al: Vascular endothelial growth factor polymorphisms and clinical outcome in colorectal cancer patients treated with irinotecan-based chemotherapy and bevacizumab. Pharmacogenomics J 12:468-475, 2012 73. Lambrechts D, Delmar P, Miles DW, et al: Single nucleotide polymorphism analysis and outcome in advanced-stage cancer patients treated with bevacizumab. Eur J Cancer 47:S173, 2011 (suppl 1; abstr 1414) 74. Loupakis F, Ruzzo A, Salvatore L, et al: Retrospective exploratory analysis of VEGF polymorphisms in the prediction of benet from rst-line FOLFIRI plus bevacizumab in metastatic colorectal cancer. BMC Cancer 11:247, 2011 75. Wang Y, Zheng Y, Zhang W, et al: Polymorphisms of KDR gene are associated with coronary heart disease. J Am Coll Cardiol 50:760-767, 2007 76. Gerger A, El-Khoueiry A, Zhang W, et al: Pharmacogenetic angiogenesis proling for rst-line bevacizumab plus oxaliplatin-based chemotherapy in patients with metastatic colorectal cancer. Clin Cancer Res 17:5783-5792, 2011 77. Hull J, Thomson A, Kwiatkowski D: Association of respiratory syncytial virus bronchiolitis with the interleukin 8 gene region in UK families. Thorax 55:1023-1027, 2000 78. lo Giudice L, Di Salvatore M, Astone A, et al: Polymorphisms in VEGF, eNOS, COX-2, and IL-8 as

predictive markers of response to bevacizumab. J Clin Oncol 28:97e, 2010 (suppl; abstr e13502) 79. Schultheis AM, Lurje G, Rhodes KE, et al: Polymorphisms and clinical outcome in recurrent ovarian cancer treated with cyclophosphamide and bevacizumab. Clin Cancer Res 14:7554-7563, 2008 80. Xu CF, Bing NX, Ball HA, et al: Pazopanib efcacy in renal cell carcinoma: Evidence for predictive genetic markers in angiogenesis-related and exposure-related genes. J Clin Oncol 29:2557-2564, 2011 81. Zhang W, Dahlberg SE, Yang D, et al: Genetic variants in angiogenesis pathway associated with clinical outcome in NSCLC patients (pts) treated with bevacizumab in combination with carboplatin and paclitaxel: Subset pharmacogenetic analysis of ECOG 4599. J Clin Oncol 27:414s, 2009 (suppl; abstr 8032) 82. Lambrechts D, Claes B, Delmar P, et al: VEGF pathway genetic variants as biomarkers of treatment outcome with bevacizumab: An analysis of data from the AViTA and AVOREN randomised trials. Lancet Oncol 13:724-733, 2012 83. Hansen TF, Christensen RD, Andersen RF, et al: The predictive value of single nucleotide polymorphisms in the VEGF system to the efcacy of rst-line treatment with bevacizumab plus chemotherapy in patients with metastatic colorectal cancer: Results from the Nordic ACT trial. Int J Colorectal Dis 27:715-720, 2012 84. Rosenfeld PJ, Brown DM, Heier JS, et al: Ranibizumab for neovascular age-related macular degeneration. N Engl J Med 355:1419-1431, 2006 85. Price TJ, Hardingham JE, Lee CK, et al: Impact of KRAS and BRAF gene mutation status on outcomes from the phase III AGITG MAX trial of capecitabine alone or in combination with bevacizumab and mitomycin in advanced colorectal cancer. J Clin Oncol 29:2675-2682, 2011 86. Rudge JS, Holash J, Hylton D, et al: VEGF trap complex formation measures production rates of VEGF, providing a biomarker for predicting efcacious angiogenic blockade. Proc Natl Acad Sci U S A 104:18363-18370, 2007 87. Autiero M, Waltenberger J, Communi D, et al: Role of PlGF in the intra- and intermolecular cross talk between the VEGF receptors Flt1 and Flk1. Nat Med 9:936-943, 2003 88. Fischer C, Mazzone M, Jonckx B, et al: FLT1 and its ligands VEGFB and PlGF: Drug targets for antiangiogenic therapy? Nat Rev Cancer 8:942-956, 2008 89. Mazzone M, Dettori D, Leite de Oliveira R, et al: Heterozygous deciency of PHD2 restores tumor oxygenation and inhibits metastasis via endothelial normalization. Cell 136:839-851, 2009 90. Van Cutsem E, Tabernero J, Lakomy R, et al: Addition of aibercept to uorouracil, leucovorin, and irinotecan improves survival in a phase III randomized trial in patients with metastatic colorectal cancer previously treated with an oxaliplatin-based regimen. J Clin Oncol 30:3499-3506, 2012 91. Bagri A, Tessier-Lavigne M, Watts RJ: Neuropilins in tumor biology. Clin Cancer Res 15:1860-1864, 2009 92. Pan Q, Chanthery Y, Liang WC, et al: Blocking neuropilin-1 function has an additive effect with antiVEGF to inhibit tumor growth. Cancer Cell 11:53-67, 2007 93. Beck B, Driessens G, Goossens S, et al: A vascular niche and a VEGF-Nrp1 loop regulate the initiation and stemness of skin tumours. Nature 478:399-403, 2011

12

2013 by American Society of Clinical Oncology

JOURNAL OF CLINICAL ONCOLOGY

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48

Biomarkers for Antiangiogenic Therapies

Acknowledgment We thank Bart Claes for his helpful and critical comments.

www.jco.org

2013 by American Society of Clinical Oncology

13

Information downloaded from jco.ascopubs.org and provided by at Penn State Hershey Medical Center on February 18, Copyright 2013 American Society of Clinical Oncology. All rights reserved. 2013 from 128.118.88.48