Genomics and Targeted Therapies in Gastroesophageal Adenocarcinoma

Published OnlineFirst November 14, 2019; DOI: 10.1158/2159-8290.CD-19-0487

REVIEW

Ankur K. Nagaraja 1 , Osamu Kikuchi 1 , and Adam J. Bass 1 ,2 ,3

ABSTRACT Gastroesophageal adenocarcinomas (GEA) are devastating diseases with stark global presence. Over the past 10 years, there have been minimal improvements in treatment approach despite numerous clinical trials. Here, we review recent progress toward understanding the molecular features of these cancers and the diagnostic and therapeutic challenges posed by their intrinsic genomic instability and heterogeneity. We highlight the potential of genomic heterogeneity to inﬂ uence clinical trial outcomes for targeted therapies and emphasize the need for comprehensive molecular profi ling to guide treatment selection and adapt treatment to resistance and genomic evolution. Revising our clinical approach to GEA by leveraging genomic advances will be integral to the success of current and future treatments, especially as novel targets become therapeu- tically tractable.

Signiﬁ cance: GEAs are deadly cancers with few treatment options. Characterization of the genomic landscape of these cancers has revealed considerable genetic diversity and spatial heterogeneity. Understanding these fundamental properties of GEA will be critical for overcoming barriers to the development of novel, more effective therapeutic strategies.

INTRODUCTION proximal stomach, directly adjacent to the esophagus ( 3 ). This rise in the stomach parallels an alarming increase in adenocar- Stomach and esophageal cancers are the third and sixth cinomas of the lower esophagus and gastroesophageal junc- most common causes of cancer mortality worldwide, respec- tion (GEJ). Key risk factors include gastroesophageal refl ux tively, and together are responsible for more than 1.2 million disease and obesity. Both gastric and esophageal adenocarci- deaths ( 1 ). These two cancers were once considered distinct nomas commonly emerge with intestinal metaplasia, which diseases plainly separated into adenocarcinomas in the stom- can result from chronic infl ammatory stimuli. The shared ach and squamous cell carcinomas in the esophagus. However, epidemiology, pathology, and genomic and molecular features recent decades have witnessed a shift in the epidemiologic of these adenocarcinomas suggest the common pathophysiol- and anatomic patterns of these cancers, contributing to a ogy of esophageal and proximal gastric adenocarcinomas (3, 4). revised and evolving understanding of their classifi cation and Indeed, The Cancer Genome Atlas (TCGA) has revealed defi ni- pathogenesis. The worldwide incidence of gastric cancer has tive genomic overlap between gastric and esophageal adenocar- been declining for at least 40 years (2 ). Nevertheless, there are cinomas, and absolute molecular distinction from squamous more than one million new cases annually, with the majority cell carcinomas of the upper and mid-esophagus (5, 6 ). This occurring in Eastern Asia ( 1 ). In North America and Western review will focus on gastroesophageal adenocarcinomas (GEA); Europe, cancers of the distal stomach, typically associated with additional information on the genomics of esophageal squa- Helicobacter pylori ( H. pylori ) infection, have decreased dramati- mous cell carcinomas (ESCC) can be found in ref. 6 . cally (3 ). In contrast, there is a rising incidence of cancers of the In addition to the rising incidence of esophageal, GEJ, and proximal gastric adenocarcinomas, another epidemiologic trend involves an increase in cancers of the gastric corpus 1 Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard or body (and fundus, to a lesser extent), predominantly in Medical School, Boston, Massachusetts. 2 Division of Molecular and Cel- lular Oncology, Dana-Farber Cancer Institute, Harvard Medical School, non-Hispanic white women younger than 50 years old, and Boston, Massachusetts. 3 Broad Institute of MIT and Harvard, Cambridge, restricted to areas with less than 20% poverty (7–9 ). Whereas Massachusetts. the current gastric cancer male:female incidence rate ratio Corresponding Author: Adam J. Bass, Dana-Farber Cancer Institute, 450 for patients ages 60 to 74 years is 2.5, the ratio is 1.0 for Brookline Avenue, Dana 810B, Boston, MA 02215. Phone: 617-632-2477; patients ages 25 to 29 years (7 ). It has been estimated that if Fax: 617-582-9830; E-mail: [email protected] the upward trend in early-onset disease continues, by 2030 Cancer Discov 2019;9:1656–72 overall gastric cancer incidence will be increasing, and female doi: 10.1158/2159-8290.CD-19-0487 incidence will surpass male incidence ( 7 ). The histologic and ©2019 American Association for Cancer Research. molecular subtypes of these “CYF” (corpus-dominant, young

1656 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in GEA REVIEW

age-dominant, female-dominant) gastric cancers have not workings of a cancer cell, nor clinical insight into rational been reported, and their risk factors are unknown, though systemic therapeutic strategies. In recent years, new genomic they may be rooted in the evolving gastric microbiome in the technologies have enabled groundbreaking studies that have wake of H. pylori decline, and/or linked to autoimmunity and revolutionized our understanding of the genomic composi- reproductive factors (7, 8). tion of GEA, ultimately providing a framework for molecular GEAs have dismal outcomes, with cumulative five-year classification and new leads for therapeutic development. relative survival of 21% to 31% in the United States (10, Pivotal studies have been led by TCGA, the Asian Cancer 11). Five-year relative survival for those with locoregional Research Group (ACRG), and the Oesophageal Cancer Clini- gastric cancer (31%–67%) is far inferior to that for colorectal cal and Molecular Stratification (OCCAMS) Consortium (5, cancer (70%–91%; ref. 10), indicating that later diagnosis 6, 19, 20). Among these, ACRG focused on gastric adenocar- alone does not account for these poor outcomes. GEAs also cinoma at a South Korean medical center, OCCAMS focused have substantial propensity for early spread of disease, and on EAC in a Western European population, and the TCGA systemic therapy for disseminated disease remains woefully studied both adenocarcinoma and squamous cell carcinoma inadequate, with 5-year relative survival of 5% (10). The spanning the stomach and esophagus using a diverse interna- convergence of a new molecularly based classification, recent tionally derived sample collection. genomic insight into drivers of GEA pathogenesis, and an Within TCGA, the first published analysis included 295 imperative clinical need make this an opportune time to gastric adenocarcinomas, inclusive of GEJ tumors, using the address how our emerging understanding of GEA can ulti- tools standard to TCGA-led studies: array-based somatic mately be translated into new therapeutic strategies. copy-number analysis, whole-exome sequencing, array-based DNA methylation profiling, mRNA sequencing, miRNA sequencing, and reverse-phase protein array (5). Four dis- DISEASE CLASSIFICATION IN tinct molecular subtypes of gastric cancer emerged from THE PREGENOMIC ERA integration of these data: Epstein–Barr virus (EBV)–positive Fifty years ago, the Lauren classification subtyped gas- (8.8%), microsatellite-unstable (MSI; 21.7%), genomically tric cancers into intestinal, diffuse, and indeterminate/mixed stable (GS; 19.7%, enriched for Lauren diffuse type), and histologies (12). Intestinal-type tumors are most common, tumors with chromosomal instability (CIN; 49.8%; Fig. 1). consisting of cohesive cells in glandular formations, often In comparison with the TCGA classification, where underly- associated with intestinal metaplasia and H. pylori infection. ing somatic genomic processes were central to development Diffuse-type tumors have noncohesive scattered cells, some- of the classifier, ACRG developed a classification into four times with signet-ring features, that tend toward peritoneal groups based on array-based gene-expression profiling of dissemination due to locally invasive properties. The World 300 gastric adenocarcinomas: MSI (22.7%), microsatellite sta- Health Organization (WHO) further refines the histologic ble with epithelial-to-mesenchymal transition (MSS/EMT; classification of gastric cancer into tubular, papillary, muci- 15.3%), MSS/TP53+ (26.3%), and MSS/TP53− (35.7%; ref. 19). nous, poorly cohesive (including Lauren diffuse type), and Applying the TCGA classifiers to the ACRG tumors showed mixed variants. Most esophageal adenocarcinomas (EAC) the expected concordance in MSI tumors. The MSS/EMT resemble intestinal-type gastric cancer, especially given their class was enriched in diffuse-type tumors classified as GS by evolution from intestinal metaplasia. To date, these histo- TCGA, and MSS/TP53+ was enriched for EBV+ tumors. A key logic classifications have not appreciably affected clinical difference was that all four ACRG subtypes had CIN tumors, care, including the selection of systemic agents. Another with the highest proportion of CIN tumors in MSS/TP53−. classification, the Siewert classification, is focused on refined The ACRG analysis demonstrated inferior survival among anatomic staging of GEJ adenocarcinomas to guide surgical the MSS/EMT+ group. Although no survival associations approaches (13). Beyond histologic and anatomic types, it is were seen in the TCGA study, potentially owing to the thera- also worth noting that there have been long-standing debates peutic heterogeneity of the internationally diverse sample regarding the comparative features of GEA in the Asian and set, subsequent studies applying the TCGA classes to two Western populations. Asian patients, beyond having higher large independent cohorts demonstrated that EBV+ tumors rates of gastric cancer, tend to have more distal gastric tumors have superior survival and GS tumors have inferior survival associated with H. pylori, and less predilection toward GEJ dis- (21). Major genomic features of these molecular subtypes are ease and adenocarcinomas of the tubular esophagus, or its discussed below. antecedent, Barrett’s esophagus (14–16). Although extensive Although the second TCGA study was nominally an analy- discussion of the comparisons of Eastern/Western tumors is sis of esophageal cancer, this study evaluated both ESCC beyond the scope of this review, early data have pointed to and EAC with the entire gastric cancer dataset in an effort the genomic similarity of these tumors (17), but potential to address debates regarding the comparative features of distinctions in their tumor–immune microenvironment (18). EAC, GEJ, and gastric tumors (6). As would be predicted by studies in other cancer types, the TCGA analysis identified systematic differences between squamous cell carcinoma MOLECULAR CLASSIFICATION and adenocarcinoma across all molecular platforms. Indeed, IN THE GENOMICS ERA these data robustly argued that high-level histologic clas- The Lauren and WHO classifications are basic descriptors sifications transcended the organ of origin. Whereas ESCCs for the microscopic organization of gastric tumors. However, clustered more strongly with head and neck SCCs than they offer no immediate scientific insight into the inner with EACs, EACs were more similar to gastric tumors than

DECEMBER 2019 CANCER DISCOVERY | 1657

REVIEW Nagaraja et al.

CIN: Chromosomal instability

• Most common subtype GS: Genomically stable • Increasing incidence • Increasing gradient from distal to proximal stomach • Most are Lauren diffuse type • TP53 mutations • Non-cohesion, invasion, dissemination • Whole-genome doubling • CDH1 and RHOA mutations • RTK/RAS/cell-cycle gene amplifications • CLDN18–ARHGAP26/6 fusions (e.g., ERBB2, EGFR, KRAS, CCNE1, CCND1/2/3, CDK6) LOWER • VEGFA amplification common in EAC ESOPHAGUS/ GEJ A GEJ/CARDIA MSI: Microsatellite instability FUNDUS EBV : Epstein–Barr virus B • Hypermutation • Hypermethylation • Hypermethylation • CDKN2A silencing • MLH1 silencing • PIK3CA, ARID1A, BCOR mutations • ARID1A, RNF43, PIK3CA, KRAS mutations C • PD-L1/2 overexpression • Elevated immune signatures • Prominent immune signatures • Sensitive to immune checkpoint blockade • Sensitive to immune checkpoint blockade

BODY

PYLORUS

ANTRUM

47.9% 47.6% 74.6% 3.5% 100% 9.5% 13.6% 19.3% 21.7% 6.3% 27.3% 9.5% 19.3%

ABCD

Figure 1. Molecular classification of GEA. Key features of the four TCGA subtypes of GEA spanning the lower esophagus to the distal stomach. Frequencies at each anatomic location are from ref. 6. to ESCCs. More subtle analysis refined the comparison of Consortium found that patients with EAC with synchro- EAC with gastric cancer, noting that EAC showed the most nous Barrett’s esophagus had better survival compared with similarity to the CIN class of gastric cancer. Furthermore, a patients without evident Barrett’s esophagus (22), suggesting joint analysis of GEA showed an anatomic gradient, with the that the most aggressive tumors might outstrip the potential proportion of CIN tumors increasing proximally such that 71 for Barrett’s esophagus to take hold in the lower esophagus. of 72 tumors categorized as definitively esophageal based on Although the TCGA study did not define subgroups within anatomic location and presence of Barrett’s esophagus were EAC, whole-genome sequencing by OCCAMS suggests that CIN tumors. Although specific methylation patterns were mutational signatures may inform targeted therapy (23). enriched in EACs relative to distal gastric cancers, multiple Together, these studies directly inform our broader view analyses attempting to differentiate the molecular features of of gastroesophageal cancer. The dramatic and consistent CIN tumors of the stomach and esophagus failed to dichoto- separation of adenocarcinoma and squamous cell carcinoma mize these cancers. Still, several somatic features were differ- clearly challenge prior notions of treating esophageal cancer entially distributed between EACs and gastric cancers. EACs as a single entity. Although still debated, these data also sug- have fewer APC mutations, suggesting less WNT depend- gest the field should transition to viewing GEA more holisti- ence or alternate means of activating WNT signaling, and cally, especially for the CIN class of tumors. These data also higher frequency of RUNX1 deletion and VEGFA and MYC indicate that any previous effort to compare gastric adenocar- amplifications. Intriguingly, new studies from the OCCAMS cinoma and EAC, or to compare Eastern and Western gastric

1658 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in GEA REVIEW cancers, had an unrecognized confounding by subtype, as phenotype between GEA and colorectal cancer have not been CIN tumors are dominant in the esophagus and GEJ (and established. A specific mutational signature of >A C transver- more prevalent in the West) compared with distal stomach sions at AA dinucleotides has been observed in GEAs (20, 29) tumors (more prevalent in the East) where GS/diffuse, MSI, and is associated with the CIN-F phenotype (25). Although and EBV+ are more common. Although the anatomic gradi- the etiology of this signature remains unresolved, it has been ents of molecular subtypes and molecular features could be posited to be caused by oxidative damage (30) and associated used to argue firmer separation of cancers in the anatomic with greater inflammatory infiltrates in EAC (23). Intrigu- esophagus or stomach, it is notable that such gradients are ingly, this signature has been reported in colorectal cancer present within analogous diseases such as colorectal cancer, associated with inflammatory bowel disease (IBD; ref. 31). where methylation and MSI are more prevalent in the ascend- As IBD-associated colorectal cancers also resemble CIN GEA ing colon and CIN more common in the distal colon and in their relative paucity of APC mutations and greater preva- rectum. For the purpose of this review and the discussion of lence of oncogene activation via amplification, these data biologically driven therapies, we will treat CIN GEA as a sin- raise the hypothesis that inflammation may contribute to the gle category given the consistent overlap in oncogenic drivers patterns of genomic evolution in GEA. and other molecular features. As explored in greater depth below, the pattern and puta- tive mechanisms of the predominantly focal CIN seen in GEA have definite implications for optimal therapy. Beyond CIN the need to evaluate and target amplified rather than mutant CIN tumors are the predominant form of GEA. Half of oncoproteins, numerous studies have documented the gastric antrum/pylorus tumors classify as CIN, and this pro- genomic complexity and heterogeneity of oncogene profiles portion increases proximally according to the aforementioned in GEA (23, 32–35). These features include both transforma- gradient such that nearly all EACs are CIN (6). Although a tion leading to cancer cells with multiple driving mitogenic firm definition of CIN has yet to be established, CIN tumors oncogenes (e.g., cancer cells with coamplification ofERBB2 are characterized by extensive aneuploidy and lack the somatic and MET) as well as heterogeneous tumors where distinct hypermutation seen in MSI tumors. Although the term CIN driving amplified oncogenes are present in distinct subclones has been applied to many cancers to describe the state of of tumors. Either case poses a clear challenge to simple tar- harboring frequent gross structural chromosomal aberrations geted therapy approaches, likely contributing to the failure and aneuploidy, the CIN present within GEAs is overwhelm- of many such therapies in GEA. ingly associated with widespread focal events (i.e., less than half the chromosome arm length), especially recurrent focal amplifications at numerous oncogenic loci (24, 25). GS A meta-analysis of TCGA data spanning GEA and colo- GS tumors are most prevalent in the antrum/pylorus and rectal cancer detailed the differential features of CIN in fundus/body (5). The origin of the GS nomenclature in these diseases (25). In this report, a score was developed TCGA followed their lack of CIN, MSI/hypermutation, and to subclassify CIN tumors based on high-amplitude focal the hypermethylation found in EBV+ tumors (5). Seventy amplifications (CIN-F) or low-amplitude broad amplifica- percent are Lauren diffuse type, a histology that is enriched tions (CIN-B). This analysis revealed a dramatic distinction in the estimated 10% to 15% of patients who develop gastric between GEA and colorectal cancer, as 75% of CIN GEA cancer prior to 45 years of age (36, 37). Diffuse gastric cancers tumors are CIN-F, in contrast to only 25% of CIN colorectal are also enriched in Alaska Natives and Hispanics compared cancer. These data correspond to the discrepant patterns of with non-Hispanic whites (38, 39). The most characteristic oncogene activation in these tumors, with amplifications somatic features of GS tumors are CDH1 (E-cadherin) loss- predominant in GEA, and mutations, chieflyRAS , prevalent of-function mutations in 30% (25); notably, this gene is also in colorectal cancer. responsible for hereditary diffuse gastric cancer when altered Within GEA, CIN-B tumors sustain more frequent somatic in the germline (40). In families without altered CDH1, muta- activating mutations of receptor tyrosine kinases (RTK) and tions have been described in CTNNA1, which functions in RAS oncogenes, and inactivation of tumor suppressors in the same complex as E-cadherin; in DNA-repair genes such cell-cycle, WNT, and TGFβ pathways. CIN-F GEAs frequently as PALB2, BRCA1, BRCA2, RAD51C, and ATM; and in MSR1 harbor amplifications of RTKs, RAS, and cell-cycle genes. and STK11 (40–42). A proteomic analysis of 84 pairs of dif- TP53 mutations are more common in CIN-F tumors (76% as fuse gastric cancer and adjacent normal tissue identified compared with 54% of CIN-B tumors), as is whole-genome three subtypes, in order of worsening overall survival (OS): doubling (68% vs. 42%). These somatic features of the more PX1, characterized by cell-cycle dysregulation; PX2, enriched prevalent CIN-F GEAs suggest distinct pathogenesis whereby in EMT pathways; and PX3, comprised of immune response a TP53-null cell may undergo genome doubling to an unstable proteins (43). It remains to be seen whether this classification tetraploid intermediate, which is then prone to catastrophic translates clinically beyond prognostic stratification. genomic events leading to transformation, as has been pos- Beyond CDH1, the most notable constellation of genomic ited from studies of the evolution of Barrett’s esophagus (26, findings in these cancers relates to the RHO pathway. Highly 27). Indeed, this pattern is consistent with what has been recurrent missense mutations in the RHOA GTPase were observed via whole-genome sequencing of these cancers (28). identified contemporaneously by three groups in 2014 (5, The etiologies for the distinct genomic pathways of can- 44, 45). Although missense mutations cluster in the amino- cer development and the marked differences of the CIN terminal GTPase domain, studies on the activity of these

DECEMBER 2019 CANCER DISCOVERY | 1659

REVIEW Nagaraja et al. mutations are mixed, with data supporting a mechanistic (5, 25). Perhaps most importantly, MSI tumors exhibit high gain-of-function but also attenuation of the GTP loading of gene-expression scores for IFNγ signaling, CD8+ T cells, and RHOA, a proxy for its activation. Intriguingly, the most recur- M1 macrophages, second only to EBV+ tumors, highlighting rent RHOA mutation is in the highly conserved core effector their immunogenicity (25, 53). As in other tumors with MSI, region at codon Y42. This residue is analogous to HRAS immune checkpoint blockade has demonstrated efficacy in Y40, where mutations selectively modulate specific down- this population (54, 55). Beyond MSI, there are also rare stream effects of HRAS (46). Aside from RHOA, recurrent tumors with hypermutation not due to mismatch repair but fusion genes have been identified which include nearly the secondary to POLE mutations. These tumors demonstrate a entire coding sequence of the transmembrane tight junction high frequency of SNVs, yet without the insertion/deletion protein CLDN18 fused by its cytoplasmic tail to the RHO events observed in MSI cancers (25). Unlike MSI tumors, GTPase–activating proteins ARHGAP26 or ARHGAP6 (5, 47). these hypermutated SNV tumors do not display elevated Intriguingly, the CLDN18–RHOGAP fusions are mutually immune signatures (25). exclusive with both CDH1 and RHOA alterations, suggesting this fusion may modulate both cellular adhesion and the + RHO pathway in gastric cancer pathogenesis. Moreover, given EBV the roles of these genes in cell adhesion and motility, their EBV+ tumors localize to the fundus/body and proximal unique recurrence in this disease suggests clear connection stomach and are notable for their marked DNA hypermethyl- with a tumor phenotype characterized by noncohesion, inva- ation. Indeed, these tumors harbor more DNA hypermethyla- sion, and dissemination. tion than any other class of cancer in the TCGA collection (5). From a therapeutic perspective, the GS/diffuse gastric EBV+ gastric tumors epigenetically silence CDKN2A, similar cancer population remains a vexing problem. As suggested to many CIN GEAs, but infrequently harbor TP53 mutation by the more mesenchymal nature of these tumors, standard or CIN (5, 25). EBV+ gastric cancers have long been recognized cytotoxic chemotherapeutic agents may be less active in these to have strong lymphoid infiltration, leading to the term patients (21, 48). Coupled to the paucity of classic targetable “lymphoepithelioma-like carcinoma” on histologic analy- oncogenes and the propensity for invasion and metastasis, sis. Accordingly, focused molecular analysis of these tumors this disease carries a dismal prognosis, as documented by noted prominent immune signatures (5, 53). Genomic analy- the ACRG study (19). Deeper mechanistic inquiry aiming to sis also identified recurrent 9p24.1 amplification at the loci define vulnerabilities of these cancers is a substantial priority of genes encoding PD-L1, PD-L2, and JAK2, consistent with within the GEA field. elevated expression of these genes (5), an event that parallels the presence of this amplification recurrently in EBV-associated lymphoid cancers (56). JAK2 transcriptionally activates MSI both PD-1 ligands in lymphoma models (56). Moreover, EBV MSI tumors arise in the distal stomach in the antrum/ infection alone can induce PD-L1 expression independent pylorus and fundus/body. Within the TCGA project, MSI of 9p24.1 amplification (57). This confluence of immune tumors were not observed in the esophageal lineage despite features imparts sensitivity to immune checkpoint blockade, the enrichment of hypermethylation in EAC (6). Clinically, as observed by the impressive response rates in lymphomas MSI is a favorable prognostic factor in resectable primary gas- with 9p24.1 amplification and in EBV+ lymphomas (58), and tric cancer (49). Post hoc exploratory analyses of the MAGIC more recently, in EBV+ gastric cancers (55). Beyond immuno and CLASSIC trials suggest that in contrast to MSS patients, therapy, somatic analysis has also revealed highly recurrent who benefit from the addition of adjuvant or perioperative PIK3CA-activating mutations and mutations in ARID1A and chemotherapy, MSI patients demonstrate better outcomes BCOR (5, 25). Thus, EBV status provides key guidance for with surgery alone, and the addition of chemotherapy may prioritizing an immunotherapeutic approach, and raises have a detrimental effect in MSI patients (50–52). These data intriguing possibilities about the utility of PI3K pathway suggest that MSI tumors may be intrinsically chemoresistant, inhibitors and DNA-demethylating agents in these cancers. or that chemotherapy blunts the robust immune response stimulated by these hypermutated tumors. However, these observations have yet to be validated through prospective TARGETED THERAPY IN THE clinical trials. PRECISION MEDICINE ERA: As in colorectal cancer, MSI gastric adenocarcinomas typi- ERBB2 AS A CAUTIONARY TALE cally occur following MLH1 silencing in the presence of a Given the paucity of targets in GS/diffuse gastric cancer and DNA hypermethylation phenotype, or less commonly MLH1 the exciting results with immunotherapy in MSI/EBV+ disease, or MSH2 mutations, leading to defective DNA mismatch the bulk of this review will discuss issues revolving around tar- repair resulting in insertion/deletion events and single- geted molecular therapy for the most common subtype, CIN nucleotide variants (SNV). Numerous other genomic fea- GEA, especially in the Western-predominant proximal gastric/ tures of MSI gastric cancers resemble MSI colorectal cancer, GEJ/EAC cancers. The genomic characterization of CIN GEA including ARID1A, RNF43, PIK3CA, and KRAS mutations has unveiled a multitude of candidate oncogenic drivers that (25). Strikingly, however, the classic BRAFV600E mutations that may be potential biomarkers for targeted therapies, most frequently occur in sporadic MSI colorectal cancer with DNA prominently RTKs. The poster child for this paradigm shift hypermethylation are notably absent in MSI gastric cancer to precision medicine in GEA was amplification ofERBB2 , despite otherwise highly similar DNA methylation patterns which occurs in 15% to 20% of GEA (5, 6, 25). The ToGA trial

1660 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in GEA REVIEW demonstrated an OS benefit with the addition of the ERBB2 Although these newer agents offer some optimism, it is mAb trastuzumab to chemotherapy for patients with advanced critical to assess the potential reasons why patients progress or metastatic gastric and GEJ cancers (13.8 vs. 11.1 months, P = on first-line trastuzumab-based therapy and to understand 0.0046; ref. 59). The survival benefit climbed to 4.2 months in why many agents that have been successful in breast cancer the subset of patients with higher ERBB2 levels (59, 60), com- fail in GEA. Potential differential features of GEA and breast parable to the effect of trastuzumab in patients with ERBB2+ cancer may include fundamental lineage-related cancer cell metastatic breast cancer (61). These data led to its widespread pathophysiology, as well as more clearly delineated distinc- regulatory approval and adoption in clinical practice. Although tions in somatic genomic patterns such as the extent of it is not FDA-approved in EAC with ERBB2 amplification, these spatial heterogeneity (Fig. 2). In intralesion heterogeneity, patients are also routinely treated per the ToGA results. individual oncogenes may be uniquely amplified in distinct Unfortunately, in stark contrast to breast cancer, the thera- subclones of the same tumor, or different oncogenes may peutic efficacy of ERBB2-directed agents in GEA has not be co-amplified within the same cancer cell. Compared with extended beyond first-line trastuzumab plus chemotherapy. breast cancers, gastroesophageal cancers have more intrale- Several successful FDA-approved agents for ERBB2+ breast sion heterogeneity of ERBB2 expression and amplification as cancer have failed in large phase III trials in GEA (Table 1). The observed by IHC and multiregion next-generation sequencing addition of pertuzumab, a mAb that blocks ERBB2 heterodi- (NGS) of primary tumors (35, 71). This intralesion heteroge- merization with other ERBB family receptors, to trastuzumab neity has direct therapeutic implications. First, the degree of plus chemotherapy in the first-line setting did not signifi- ERBB2 amplification by NGS correlates with the benefit from cantly improve OS for ERBB2+ metastatic gastric or GEJ can- trastuzumab (34), possibly serving as a measure of ERBB2 cer in the JACOB trial, although it trended toward a positive dependence in the tumor. Second, trastuzumab resistance has result (62). Importantly, major pharmacokinetic differences been observed to be commonly mediated by the rise of a non– in pertuzumab between ERBB2+ gastric and breast cancers ERBB2-amplified clone (34, 72). Although it is often referred were identified and adequately addressed through a phase IIa to as “loss” of ERBB2, this situation more likely reflects preex- dose-finding study ( JOSHUA) prior to JACOB and confirmed isting heterogeneity and the selection of ERBB2-negative sub- by pharmacokinetic assessments of JACOB participants (63, clones during trastuzumab therapy. This phenomenon might 64). T-DM1, an antibody–drug conjugate of trastuzumab and have contributed to the failure of second-line T-DM1 in the a microtubule inhibitor, was not superior to taxane in the sec- GATSBY trial, where more than 75% of patients had already ond-line setting in the GATSBY trial (65). Both the TRIO-013/ progressed on trastuzumab and patients were not rebiopsied LOGiC and TyTAN trials also failed to meet their primary to evaluate ERBB2 status prior to randomization (65, 73). endpoint of OS for the dual ERBB2/EGFR kinase inhibitor Genomic analysis has demonstrated that more than half lapatinib combined with chemotherapy as first- or second- of ERBB2-amplified GEA tumors have concomitant baseline line treatment, respectively (66, 67). A preplanned subgroup genomic alterations that may contribute to intrinsic tras- analysis in the LOGiC study did show a significant OS benefit tuzumab resistance, including activating mutations in the in Asian patients and in younger patients, and in the TyTAN PI3K pathway and amplifications ofCCNE1, CDK6, EGFR, study, patients with ERBB2 immunohistochemistry (IHC) 3+ and MET, in line with the focal CIN pattern characteristic of had significantly improved OS, highlighting the importance these tumors (32, 34, 74, 75). As observed by the OCCAMS of refining patient selection, fine-tuning the definition of Consortium, this phenomenon of genomic co-amplification biomarker positivity, and designing more effective targeted in GEA holds not only for ERBB2 but for other RTKs too approaches for GEA with ERBB2 amplification. (23). Moreover, genomic analysis demonstrated that patients Several emerging ERBB2 agents have shown encouraging with GEA with ERBB2 amplification and co-occurring lesions efficacy in early-phase clinical trials. These new drugs include had inferior survival, especially when compared with those antibodies that direct a payload to ERBB2+ cells and those with the highest level of ERBB2 amplification by NGS (34). that are optimized to enhance immunologic attack against Preclinical studies suggest that dual blockade of multiple the cancer cell following binding of the antibody to ERBB2. kinases present in GEA can overcome resistance to single- A phase I study of trastuzumab deruxtecan (DS-8201), a drug therapy, results confirmed clinically (23, 32, 33, 76). For humanized anti-ERBB2 antibody conjugated to a topoisomer- example, a phase II study of afatinib in trastuzumab-resistant ase I inhibitor payload, demonstrated objective responses in ERBB2+ GEA demonstrated an association between afatinib 19 of 44 patients with advanced unresectable or metastatic response and tumors harboring dual amplification ofERBB2 ERBB2+ gastric/GEJ tumors, all of whom were previously and EGFR (76), suggesting that combined targeting of ERBB2 treated with trastuzumab (68). Margetuximab (MGAH22) and specific co-occurring genomic alterations may be essen- is a monoclonal antibody that binds the same epitope of tial for therapeutic efficacy. ERBB2 as trastuzumab, but with an altered Fc domain Perhaps even more striking than evolution of the tumor that enhances binding to the CD16A receptor on natural to an ERBB2-negative state upon trastuzumab therapy, gas- killer cells and macrophages to boost antibody-dependent troesophageal cancers also demonstrate extensive baseline cell-mediated cytotoxicity. A phase I study showed good toler- discordance in ERBB2 status between synchronous primary ability and cytotoxic activity against ERBB2+ advanced solid tumors and metastases prior to systemic therapy, in contrast tumors, including partial responses in 2 of 20 patients with to breast cancers that are ERBB2 concordant in more than GEA (69). A phase I/II study of margetuximab plus pembroli- 90% of cases (35). This interlesion heterogeneity suggests zumab for the second-line treatment of patients with ERBB2+ that the field should revisit the guidelines for assessment of GEA (post-trastuzumab) is ongoing (NCT02689284; ref. 70). ERBB2 status in GEA, which recommend testing of only a

DECEMBER 2019 CANCER DISCOVERY | 1661

REVIEW Nagaraja et al.

Table 1. Key clinical trials for GEA

Trial Median OS Median PFS Target (reference) Phase Line of therapy Treatment ORR (%) (months) (months) HER2 ToGA (59 ) III 1L, HER2+ Cisplatin/fl uoropyrimidine + 47 13.8 6.7 trastuzumab Cisplatin/fl uoropyrimidine 35 11.1 5.5 P = 0.0046 HER2 JACOB (62 ) III 1L, HER2+ Chemotherapy + trastu- 57 17.5 8.5 zumab + pertuzumab Chemotherapy + 48 14.2 7 trastuzumab P = 0.057 NS HER2 TRIO-013/ III 1L, HER2+ CapeOx + lapatinib 53 12.2 6 LOGiC (66 ) CapeOx 39 10.5 5.4 P = 0.3492 NS HER2 TyTAN (67 ) III 2L, HER2+ Paclitaxel + lapatinib 27 11.0 5.4 Paclitaxel 9 8.9 4.4 P = 0.1044 NS HER2 GATSBY (65 ) II/III 2L, HER2+ T-DM1 21 7.9 2.7 Paclitaxel or docetaxel 20 8.6 2.9 P = 0.86 NS EGFR EXPAND (85 ) III 1L Cisplatin/capecitabine + 30 9.4 4.4 cetuximab Cisplatin/capecitabine 29 10.7 5.6 P = 0.32 NS primary endpoint EGFR REAL3 (86 ) III 1L EOC + panitumumab 46 8.8 6 EOC 42 11.3 7.4 P = 0.013 NS EGFR COG (87 ) III 2L Gefi tinib 2.7 3.7 1.6 Placebo 0.4 3.7 1.2 P = 0.29 NS MET METGastric III 1L, MET+, HER2- mFOLFOX6 + onartuzumab 46 11.0 6.7 (94 ) mFOLFOX6 41 11.3 6.8 (early termi- P = 0.24 NS nation) HGF RILOMET-1 III 1L, MET+ ECX + rilotumumab 30 8.8 5.6 (96 ) ECX 45 10.7 6 (early termi- P = 0.032 NS nation) MET AMG 337 (97 ) II ≥2L for 98% AMG 337 18 7.9 3.4 (early termi- of patients, primary nation) MET+ endpoint mTOR GRANITE-1 III ≥2L Everolimus 4.5 5.4 1.7 (99 ) Placebo 2.1 4.3 1.4 P = 0.124 NS FGFR1/2/3 SHINE (103 ) II 2L, FGFR2+ AZD4547 2.6 5.5 1.8 Paclitaxel 23 6.6 3.5 P = 0.9581 NS primary endpoint VEGFA AVAGAST III 1L Cisplatin/fl uoropyrimidine + 46 12.1 6.7 (107 ) bevacizumab Cisplatin/fl uoropyrimidine 37 10.1 5.3 P = 0.1002 NS

(continued)

1662 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in GEA REVIEW

Table 1. Key clinical trials for GEA (Continued)

Trial Median OS Median PFS Target (reference) Phase Line of therapy Treatment ORR (%) (months) (months) VEGFR2 RAINFALL III 1L Cisplatin/ﬂ uoropyrimidine + 41 11.2 5.7 (106 ) ramucirumab Cisplatin/ﬂ uoropyrimidine 36 10.7 5.4 P = 0.0106 (investigator review) P = 0.74 NS (central independent review) primary endpoint VEGFR2 REGARD ( 104 ) III 2L Ramucirumab 3 5.2 2.1 Placebo 3 3.8 1.3 P = 0.047 VEGFR2 RAINBOW III 2L Paclitaxel + ramucirumab 28 9.6 4.4 (105 ) Paclitaxel 16 7.4 2.9 P = 0.017 PD-1 KEYNOTE- III 2L Pembrolizumab CPS≥1: CPS≥1: CPS≥1: 061 (114 ) Paclitaxel 16 9.1 1.5 14 8.3 4.1 P = 0.0421 NS P = Not reported OS and PFS in OS and PFS in CPS ≥ 1 are CPS≥1 are both primary both primary endpoints endpoints PD-1 KEYNOTE- II ≥3L, gastric/GEJ Pembrolizumab (Cohort 1) CPS≥1: CPS≥1: 5.8 CPS≥1: 2.1 059 (112 ) 15.5 CPS<1: 6.4 CPS<1: 4.9 CPS<1: 2.0 Primary endpoint PD-1 KEYNOTE- II ≥3L, ESCC/EAC/ Pembrolizumab ESCC: ESCC: 6.8 ESCC: 2.1 180 (115 ) GEJ 14.3 EAC/GEJ: EAC/GEJ: 3.9 EAC/GEJ: 1.9 5.2 primary endpoint PD-1 ATTRACTION-2 III ≥3L Nivolumab 11 5.3 1.6 ( 113 ) Placebo 0 4.1 1.5 P < 0.0001 PARP GOLD (129 ) III 2L Paclitaxel + olaparib 24 8.8 3.7 Paclitaxel 16 6.9 3.2 P = 0.026 NS

NOTE: Primary endpoint is median OS unless otherwise specifi ed. Abbreviations: CapeOx, capecitabine/oxaliplatin; CPS, PD-L1 combined positive score; ECX, epirubicin/cisplatin/capecitabine; EOC, epirubicin/ oxaliplatin/capecitabine; mFOLFOX6, leucovorin/ﬂ uorouracil/oxaliplatin; NS, not signifi cant; ORR, objective response rate.

single tumor site (ref. 77 ; Fig. 3 ). Although it is diffi cult to lesion. Later, we discuss the potential of using cutting-edge argue at this time that the presence of heterogeneous ERBB2 liquid biopsy technology to confront this heterogeneity. is suffi cient rationale to deprive a patient of trastuzumab Aside from differences encoded in the somatic genome, therapy, research may ultimately show that therapy against a optimal targeting also requires consideration of the under- target clearly present in metastatic disease could be superior lying biology of a lineage. A telling example stems from a to targeting ERBB2 that is detectable only in the primary basket trial evaluating the ERBB2 kinase inhibitor neratinib

DECEMBER 2019 CANCER DISCOVERY | 1663

REVIEW Nagaraja et al.

(79, 80). Key lineage-dependent differences in the response to a targeted agent can have far-reaching consequences on Metastases with therapeutic efficacy. As many approaches being evaluated in Gene B alteration ERBB2+ GEA originated with clinical success in breast cancer, Primary site with these results are a reminder of the importance of focused Gene A alteration biological study of GEA. Interlesion heterogeneity Metastasis with Gene A alteration BEYOND ERBB2: PARALLEL LESSONS FROM OTHER TARGETED THERAPIES Intralesion Genomic EGFR heterogeneity coalteration Phase III trials of other targeted therapies in gastro esophageal cancer have fallen short of the mark left by trastuzumab (Table 1). EGFR is amplified in 5% of GEA and Gene A alteration Gene A amplification the protein is overexpressed in 30% to 50% of GEA (81–84). Neither the addition of cetuximab (EXPAND trial) nor pani- Gene B alteration Gene A and B coamplification tumumab (REAL3 trial), both anti-EGFR mAbs, to chemotherapy improved progression-free survival (PFS) or OS (85, 86). In REAL3, patients in the panitumumab plus chemotherapy arm actually had worse OS, with the important caveat that this combination arm also received less chemotherapy Figure 2. Types of genomic heterogeneity in GEA. because of overlapping toxicity with panitumumab (86). In the second-line or higher setting, the EGFR tyrosine kinase inhibitor (TKI) gefitinib had no benefit compared with pla- in patients with ERBB2 somatic mutations (78). Although cebo (COG trial; ref. 87). However, it is critical to note that caveats should be drawn to the differences between mutation these negative EGFR trials were all conducted in unselected versus amplification, and small-molecule kinase inhibitor ver- patient populations, irrespective of EGFR status. Indeed, sus antibody, the results were notable for robust responses in subset analysis of the COG data showed a statistically signifi- ERBB2-mutant breast cancer but remarkable lack of efficacy cant OS benefit inEGFR -amplified patients (88), and a similar in GEA or colorectal cancer (78). These results evoke memo- trend was observed in the EXPAND trial for tumors with high ries of the failure to reproduce the stunning albeit transient EGFR IHC expression (89), again emphasizing the simple effects of BRAF inhibition from V600E-mutant melanoma yet crucial lesson of testing targeted therapies in biomarker- in colorectal cancers also harboring the identical mutation positive patient populations. More recent data highlight the

Next-generation Circulating Response to Gene A– Primary Metastasis Gene A FISH sequencing tumor DNA directed therapy

Endoscopic Gene A Gene A + Sensitive biopsy amplification amplification

Metastasis Gene A Gene A ± B Initial sensitivity but with + amplification amplification shorter time to progression

Gene A Gene B ± A + Resistant Primary amplification amplification

Gene A and B Gene A and B + Resistant amplification amplification

+ in a Gene A and B Gene B ± A Resistant subset of cells amplification amplification Gene A amplification

Gene B amplification Gene B Gene A ± B Initial sensitivity but with − Gene A and B coamplification amplification amplification shorter time to progression

Figure 3. Genomic heterogeneity affects the molecular diagnosis and treatment response of GEA. Potential outcomes with testing of a single tumor site, in this case, endoscopic biopsy of the primary lesion. The inset represents a scenario in which the biopsy captures both cells with Gene A amplification and cells with Gene B amplification.

1664 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in GEA REVIEW potential efficacy of EGFR-directed therapy both with and however, an earlier phase II study identified phosphorylated without chemotherapy in patients with bona fideEGFR - S6, a downstream effector of mTOR, as a potential predictive amplified GEA (81). factor that remains to be validated (100). MET FGFR2 Clinical trials using MET-directed agents in GEA have also FGFR2 amplifications occur in 5% of GEA, with enrich- been disappointing. Support for targeting MET followed ment in the CIN and GS subtypes (5, 25). As with MET, cell marked MET inhibitor sensitivity against kinase inhibitors line studies in GEAs with FGFR2 amplification show robust in MET-amplified GEA cell lines (90). WhereasMET ampli- responses to kinase inhibition (101, 102). Mechanistic stud- fications occur in 6% of GEA, MET overexpression has been ies suggest that excess FGFR2 activates PI3K signaling indi- reported in 25% to 60% of GEA (5, 91–93). The METGastric rectly through transactivation of alternative RTKs such as study examined chemotherapy with onartuzumab, a mAb ERBB3 and IGF1R (102). However, no differences in PFS were against the MET extracellular domain that blocks interac- observed in the randomized phase II SHINE study comparing tion with the MET ligand HGF (94). This trial was stopped the FGFR1/2/3 TKI AZD4547 to paclitaxel for the second- early because no predictive cutoff for MET IHC could be line treatment of gastric cancers with FGFR2 polysomy or identified in a parallel phase II study (95). Even MET IHC amplification as determined by FISH on archival tumor speci- 2/3+ patients in the phase III trial had no OS benefit (11 vs. mens (103). A limitation of this study was the consider- 9.7 months, P = 0.06), although statistical significance may have able intratumor heterogeneity of FGFR2 status and lack of been thwarted by decreased power from premature termination correlation of FGFR2 genomic status with FGFR2 mRNA (94). The RILOMET-1 study was also discontinued early, in this expression, possibly explaining the negative results (103). case because of decreased survival in the experimental arm of Indeed, another translational clinical trial demonstrated that chemotherapy plus rilotumumab, a mAb against HGF (96). As response to AZD4547 can be predicted by high-level homo- compared with METGastric, which required MET IHC ≥1+ in geneous FGFR2 amplification as opposed to subclonal ampli- at least 50% of cells, RILOMET-1 had looser inclusion criteria fication (102), again highlighting the critical importance of (≥1+, ≥25% cells), included fewer IHC 2/3+ patients, and had identifying an effective biomarker threshold and considering no trend toward an interaction between MET expression and heterogeneity. outcomes. Also, there were very few MET-amplified patients in RILOMET-1, and FISH ratios were not reported in METGastric. Antiangiogenic Therapies In contrast to these studies, MET amplification was used Although antiangiogenic therapies are currently stand- as the inclusion criteria for a single-arm phase II study of ard of care and are distinct from the targeted oncogene the MET TKI AMG 337 in heavily pretreated patients with approaches described above, clinical trials of these agents advanced-stage tumors (97). Disappointingly, this trial was in unselected patients with GEA have had mixed efficacy, also stopped early because the primary endpoint of objective suggesting there may be room for improvement through response rate (ORR; 18%) was lower than expected compared investigation of biomarkers and effective combination strate- with phase I data (30%; ref. 98). A key limitation of the AMG gies. The anti-VEGFR2 antibody ramucirumab demonstrated 337 studies was that MET status was derived from archi- significant albeit modest OS benefit in the second-line setting val tumor specimens, generally obtained prior to first-line both as monotherapy (vs. placebo in the REGARD trial) and cytotoxic therapy. In light of the above discussion regarding combined with chemotherapy (vs. chemotherapy alone in the ERBB2 heterogeneity, this raises the hypothesis that some RAINBOW trial; refs. 104, 105). On the other hand, in the patients may not have been truly MET-amplified at enroll- first-line setting, there was no improved OS or PFS (by central ment. Indeed, similar to ERBB2, MET amplification is hetero- independent review) with the addition of ramucirumab to geneous between primary tumors and metastases, and MET chemotherapy (RAINFALL trial; ref. 106). Similarly, the anti- inhibition permits outgrowth of non–MET-amplified clones VEGFA antibody bevacizumab improved overall response (33). Moreover, in one study, 10 of 21 MET-amplified GEAs rate and PFS, but not OS, when added to chemotherapy for harbored coamplification ofERBB2 and/or EGFR with MET the first-line treatment of advanced gastric cancer (AVAGAST in the same tumor cells, conferring de novo resistance to MET trial; ref. 107). Interestingly, biomarker analysis pinpointed inhibition (33). These results again emphasize the inadequacy high baseline plasma VEGFA levels (but not tumor VEGFA) of simple biomarker assessment in CIN GEA, and suggest we and low tumor protein expression of neuropilin 1 (a VEGF have rejected targeted therapies that might be effective if used family coreceptor) as candidate predictive biomarkers in in optimal patients or in optimal combinations. patients from non-Asian regions (108). Future correlative studies will be integral to maximizing the efficacy of antian- mTOR giogenic therapy, especially given that VEGFA amplifications The PI3K–AKT–mTOR pathway is frequently activated occur in 28% of EACs and 7% of gastric adenocarcinomas (5, in GEA, particularly in the EBV+ subtype, through PIK3CA- 6). Furthermore, the clinical success of pairing anti-VEGF activating mutations and amplifications, andPTEN truncat- agents with immunotherapy in renal cell carcinoma (109) ing mutations and deletions (5, 6). The GRANITE-1 study warrants evaluation of these combinations in GEA. enrolled unselected patients with gastric cancer to evaluate second- and third-line treatment with everolimus compared Immune Checkpoint Blockade with placebo, unfortunately without improvement in OS The role of immune checkpoint blockade for GEA is reviewed (99). Biomarker analysis for this trial has not been published; elsewhere (110) and will be addressed only briefly. Recent work

DECEMBER 2019 CANCER DISCOVERY | 1665

REVIEW Nagaraja et al. has demonstrated clear enrichment of responses in patients therapy, underscores the need for comprehensive molecu- with MSI and EBV+ gastric cancer (55), again emphasizing the lar profiling to guide treatment selection, including combi- importance of molecular stratification in the application of nation strategies. Furthermore, the breadth of interlesion immunotherapy for GEA. Pembrolizumab is approved for the tumor heterogeneity within an individual patient means that second-line treatment of MSI-high/mismatch repair–deficient biopsy of a single lesion incompletely represents the clinically gastric cancer. Notably, unlike in colorectal cancer, there are relevant genomic composition of a patient’s cancer burden. also responses to immune checkpoint inhibition in non-MSI To bypass the impracticality of taking multiple tissue biop- and non-EBV+ GEA. The phase III KEYNOTE-062 trial dem- sies, including synchronous primary and metastatic biop- onstrated noninferior OS and a more favorable safety profile sies, emerging data suggest the clear potential of minimally for pembrolizumab compared with chemotherapy as first-line invasive liquid biopsies for the isolation and sequencing of treatment for patients with PD-L1+/ERBB2− advanced gastric/ cell-free circulating tumor DNA (ctDNA) in peripheral blood GEJ adenocarcinoma (NCT02494583; ref. 111). In addition, (reviewed in ref. 117). Although systematic validation of the pembrolizumab prolonged OS in the subset of patients with accuracy, reliability, and reproducibility of ctDNA assays is tumors expressing high levels of PD-L1 (17.4 vs. 10.8 months). a work in progress, clinical utility has been demonstrated in Although the combination of pembrolizumab plus chemo- several studies of GEA with amplifications ofERBB2, EGFR, therapy improved ORR, it did not improve OS or PFS com- MET, and FGFR2 (33, 35, 72, 76, 81, 102). In an umbrella pared with chemotherapy alone. The phase II KEYNOTE-059 trial of targeted therapy in GEA, approximately one third trial led to FDA approval of pembrolizumab for patients of patients had discordant oncogene amplification between with PD-L1+ advanced gastric/GEJ adenocarcinoma who have synchronous primary tumor and metastatic biopsies, leading received two or more lines of therapy (112). In an Asian popu- to treatment reassignment (35, 118). Biomarker concord- lation of relapsed/refractory gastric/GEJ cancers, the phase III ance between metastases and ctDNA surpasses correlation ATTRACTION-2 study showed a modest median OS benefit between metastases and primary tumor, raising the hypoth- for nivolumab compared with placebo (5.3 vs. 4.1 months, esis that ctDNA may complement traditional tissue-based P < 0.0001), and a substantial benefit in 12-month OS rate genomic profiling to guide treatment in the metastatic set- (26.2% vs. 10.9%; ref. 113). ting. Beyond initial treatment selection, ctDNA may be sam- However, other studies have tempered expectations from pled longitudinally during the course of treatment to trace these agents and emphasize the need for additional bio- the evolution of clonal heterogeneity, including identifying markers beyond PD-L1. KEYNOTE-061, a phase III trial of drivers of intrinsic and acquired resistance. Biomarker detec- pembrolizumab compared with paclitaxel for the second-line tion and validation might also improve with ctDNA analysis; treatment of PD-L1+ gastric/GEJ tumors, failed to show a for example, higher EGFR copy number in ctDNA correlated survival benefit for immune checkpoint blockade (114). The with response to anti-EGFR therapy (81), and a similar find- phase II single-arm KEYNOTE-180 trial of pembrolizumab ing was reported for FGFR2 amplification and response to for advanced, metastatic esophageal cancer progressing after an FGFR TKI (102). In addition, ctDNA mutational load two or more lines of therapy demonstrated an ORR of only predicted response to pembrolizumab in MSI-high meta- 5.2% in patients with EAC/GEJ, in contrast to 14.3% in static gastric cancer, and lower ctDNA levels at 6 weeks post- patients with ESCC (115). The phase III KEYNOTE-181 study treatment predicted improved outcomes (55). compared pembrolizumab to chemotherapy in patients with Integrating NGS and ctDNA technologies into clinical advanced or metastatic ESCC or EAC/GEJ after first-line trials will enhance patient and treatment selection, espe- standard therapy (NCT02564263; ref. 116). Despite favora- cially for innovative trial designs like basket and umbrella ble trends and a better safety profile, pembrolizumab did studies. PANGEA (NCT02213289; ref. 118) and VIKTORY not significantly improve OS in the intention-to-treat popu- (NCT02299648; ref. 119) are both biomarker-based umbrella lation. Pembrolizumab did improve OS in a prespecified trials for metastatic GEA or gastric cancer, respectively, that subgroup analysis of patients with tumors expressing high prospectively use tumor genomic profiling to assign patients levels of PD-L1 (9.3 vs. 6.7 months); however, this benefit was to targeted therapy treatment arms spanning key RTK restricted to patients with ESCC (10.3 vs. 6.7 months) and pathway amplifications and mutations. Whereas VIKTORY not observed in patients with EAC/GEJ (6.3 vs. 6.9 months). focuses on primary tumor specimens, PANGEA analyzes both Whereas these trials led to FDA approval of pembrolizumab primary and metastatic lesions at baseline and at progres- in the ESCC population, numerous ongoing studies will sive disease, and both trials perform ctDNA sequencing for address pivotal questions about how to enhance the efficacy correlative studies. Outcomes from PANGEA are pending; of immunotherapy in GEA and identify combinations with however, the biomarker-driven treatment group in VIKTORY other targeted and conventional therapies. demonstrated better OS (9.8 vs. 6.9 months, P < 0.001) and PFS (5.7 vs. 3.8 months, P < 0.0001) compared with patients who received conventional second-line treatment (119). More MOVING FORWARD: MAXIMIZING adaptive trials such as these are needed to address the intri- OPPORTUNITIES FOR THERAPEUTIC cate, heterogeneous genomes of patients with GEA. Although EFFICACY these trials will certainly be more complex, recent experience As a field, we are now poised to leverage the failures of has demonstrated that unless we change our approach, we targeted therapy to revise our clinical approach to GEA. The will continue to conduct negative studies that fail to identify prevalence of concomitant genomic alterations, especially the efficacy of drugs that may be beneficial in a more tailored RTK coamplifications that confer resistance to single-agent patient population. These refined clinical trial approaches

1666 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in GEA REVIEW will pave the way for exploring and ultimately translating For CCNE1-amplified GEA, preclinical data from gastric new candidate genetic dependencies and therapeutic vulner- cancer patient-derived xenografts and from high-grade serous abilities of GEA, as described below. ovarian cancer (HGSOC) demonstrate that targeting CDK2, the partner kinase of CCNE1, may be effective (123, 124). Notably, CCNE1 amplifications are mutually exclusive and UNCHARTED TERRITORY: NOVEL TARGETS synthetic lethal with BRCA1 loss in HGSOC (125), suggesting AND TREATMENTS ON THE HORIZON the necessity of homologous recombination (HR) DNA repair The diversity of oncogenic amplifications in the CIN sub- in these cancers secondary to CCNE1-induced replication type of GEA have made it challenging to identify a unifying stress. Thus, targeting the DNA-damage response may be a vulnerability of these tumors. The common thread is their key therapeutic approach that is further supported by enrich- underlying genomic instability and aneuploidy. Long-term ment of an HR deficiency (HRD) mutational signature (also therapeutic advancements may require a deeper understand- known as COSMIC signature 3, or the BRCA signature) in the ing of the causes and consequences of CIN, and the vulnera- CIN subtype, specifically associated with focal somatic copy- bilities stemming from CIN. This not only includes targeting number alteration events (25, 126). The HRD signature cor- oncogenes that are activated in the presence of genomic insta- relates with response to platinum- and PARP inhibitor–based bility, but also defining vulnerabilities inherent to the under- therapies in breast and ovarian cancers (127), suggesting lying state of a cell with such marked genomic disruption. there may be a subset of CIN GEA harboring similar suscep- Meanwhile, the focus of targeted therapy efforts for CIN tibility. Indeed, a retrospective analysis of patients with South GEA has been RTK amplifications. However, there are a host of Korean gastric cancer showed that those with CIN tumors other targets worthy of evaluation. One example followed the yielded the greatest benefit from adjuvant chemotherapy recognition that CIN GEAs frequently activate the KRAS onco- (21), perhaps hinting at a connection between CIN, HRD, and gene not via canonical mutations seen in many other cancers platinum sensitivity in gastric cancer. but through focal high-level genomic amplification compris- PARP inhibitors have been clinically evaluated in GEA. ing 14% of CIN GEA and causing as much as several hundred- Despite promising phase II data (128), the phase III GOLD fold overexpression of wild-type KRAS protein (25, 120). Labo- trial failed to show benefit from adding the PARP inhibitor ratory-based evaluation of the intrinsic wiring of these tumors olaparib to paclitaxel in the second-line setting (129). These demonstrated that MEK inhibitors, which form the backbone negative results in the overall patient population might be of many current approaches for developing combination ther- understandable because sensitivity to PARP inhibition typi- apy for RAS-driven cancers, led to marked adaptive KRAS acti- cally depends on HRD. Unexpectedly, however, the subgroup vation owing to the massive KRAS expression. This adaptive of patients with loss of ATM expression by IHC also did not response could be overcome by dual inhibition of MEK and the benefit from olaparib. Potential explanations include imma- protein tyrosine phosphatase SHP2, also known as PTPN11, ture follow-up of the ATM-negative subgroup, and confound- leading to tumor regression in xenograft models (120). Phase ing of treatment by favorable prognostic factors enriched in I trials are under way to evaluate the safety and tolerability of ATM-low tumors, such as PD-L1 expression (129, 130). The the SHP2 inhibitors TNO155 (NCT03114319) and RMC-4630 data also suggest that additional genomic and functional (NCT03634982) in advanced solid tumors. measures of HRD and replication stress should be explored Beyond RTK/MAPK pathways, molecular dissection of CIN as predictive biomarkers of response to PARP inhibition. GEA also revealed frequent alterations in cell-cycle regulators Platinum sensitivity may itself be a predictive biomarker, a (5, 6, 25, 121). Epigenetic silencing or deletion of CDKN2A is concept that has led to a phase III trial of PARP inhibition common, particularly in EAC and EBV+ gastric cancer, as are versus placebo as maintenance therapy in inoperable, locally amplifications of cyclin–CDK proteins that coordinate the advanced, or metastatic gastric cancer that responded to

G1–S transition, specificallyCCNE1, CCND1, and CDK6. On platinum-based first-line chemotherapy (NCT03427814; ref. the surface, these alterations seemingly converge on a final 131). More broadly, there is a widening spectrum of therapeu- common pathway to phosphorylate and disable the RB1 tic agents in the DNA-damage sphere, for example, inhibitors tumor suppressor. Interestingly, however, primary RB1 dele- of CHEK1, ATR, and WEE1, which may be combined with tion rarely occurs in CIN GEA (6, 25), raising the possibility PARP inhibition to attack cancer cells with high basal replica- that cyclin–CDK amplifications may also promote distinct tion stress, even when HR remains intact (132). Evaluation of oncogenic functions and confer specific genetic dependen- these agents in GEA, both preclinically and clinically, may be cies independent of RB1 inhibition. In parallel to investi- a critical path for the future. gating these fundamental questions at the bench, there are The development of targeted therapies for diffuse gastric already tangible opportunities for translation at the bedside. cancer is a pressing unmet need. Although these tumors are CDK4/6 inhibition is a candidate therapeutic strategy for generally characterized by altered cell adhesion and motil- GEA with amplifications ofCCND1 or CDK6, or silencing/ ity, up to 75% retain strong protein expression of the tight deletion of CDKN2A. Although the response to palbociclib junction molecule claudin 18 isoform 2 (CLDN18.2), a lin was not encouraging in a small phase II trial of patients with eage marker highly specific to differentiated gastric mucosal gastroesophageal cancer, inclusion was based on IHC for RB epithelial cells (133). CLDN18.2 is also expressed at lower nuclear expression, which may not have been an optimal bio- frequency in intestinal-type gastric cancers and ectopically marker (NCT01037790; ref. 122). Furthermore, efficacy may activated in EAC (133). The phase II FAST trial randomized require combinations, as seen by effects of CDK4/6 inhibitors patients with advanced/recurrent gastric and GEJ cancers, in concert with hormonal therapy in breast cancer. including 45% with diffuse type histology, to first-line

DECEMBER 2019 CANCER DISCOVERY | 1667

REVIEW Nagaraja et al.

chemotherapy with or without IMAB362 (also known as clau- ions, interpretations, conclusions, and recommendations expressed diximab or zolbetuximab), a chimeric mAb against CLDN18.2 in this material are those of the authors. that functions not via inhibition of CLDN18.2 but instead by using CLDN18.2 as a specific marker for antibody binding Received April 26, 2019; revised August 9, 2019; accepted September to elicit immune activation (NCT01630083; refs. 134–136). 6, 2019; published first November 14, 2019. Inclusion criteria required CLDN18.2 IHC ≥2+ in at least 40% of tumor cells. The addition of IMAB362 improved ORR (39% vs. 25%, P = 0.022), the primary endpoint of PFS (7.5 vs. REFERENCES 5.3 months, P < 0.0005), and OS (13.0 vs. 8.4 months, P = . 1 International Agency for Research on Cancer. 2018 GLOBOCAN: 0.0008), with even greater benefit in the subgroup with Stomach. Available from: http://gco.iarc.fr/today/data/factsheets/ cancers/7-Stomach-fact-sheet.pdf. CLDN18.2 IHC 2 in at least 70% of cells (135). These ≥ + 2. Luo G, Zhang Y, Guo P, Wang L, Huang Y, Li K. Global patterns and encouraging results have spurred two phase III trials for fur- trends in stomach cancer incidence: age, period and birth cohort ther evaluation of IMAB362 (NCT03504397; NCT03653507). analysis. Int J Cancer 2017;141:1333–44. Other approaches to target CLDN18.2+ cells using antibody– 3. Van Cutsem E, Sagaert X, Topal B, Haustermans K, Prenen H. drug conjugates for delivery of a cytotoxic payload, CD3 Gastric cancer. Lancet 2016;388:2654–64. bispecific antibodies to induce T cell–mediated killing of 4. Rustgi AK, El-Serag HB. Esophageal carcinoma. N Engl J Med tumor cells, and CLDN18.2-specific chimeric antigen recep- 2014;371:2499–509. 5. The Cancer Genome Atlas Research Network. Comprehensive molec- tor T cells are under preclinical investigation (137, 138). ular characterization of gastric adenocarcinoma. Nature 2014;513: 202–9. 6. The Cancer Genome Atlas Research Network. Integrated genomic CONCLUSION characterization of oesophageal carcinoma. Nature 2017;541:169–75. Our understanding of gastroesophageal cancer has rapidly 7. Anderson WF, Rabkin CS, Turner N, Fraumeni JF Jr, Rosenberg evolved over the past five years. Delving into the genomic PS, Camargo MC. The changing face of noncardia gastric cancer incidence among US non-Hispanic whites. J Natl Cancer Inst landscape has revealed strong similarities between adenocar- 2018;110:608–15. cinomas of the stomach and esophagus, and their definitive 8. Blaser MJ, Chen Y. A new gastric cancer among us. J Natl Cancer Inst distinction from squamous cell carcinomas of the esopha- 2018;110:549–50. gus. Moreover, these genomic blueprints have enabled clas- 9. Anderson WF, Camargo MC, Fraumeni JF Jr, Correa P, Rosenberg sification into molecular subtypes that elucidate the diverse, PS, Rabkin CS. Age-specific trends in incidence of noncardia gastric complex oncogenic pathways underlying GEA pathogenesis. cancer in US adults. JAMA 2010;303:1723–8. 10. American Cancer Society. Cancer Facts & Figures 2018. Atlanta, GA: These achievements are essential stepping stones on the American Cancer Society; 2018. path toward developing more effective therapeutic strategies. 11. Noone AM, Cronin KA, Altekruse SF, Howlader N, Lewis DR, Unfortunately, however, the Occam’s razor “single target, sin- Petkov VI, et al. Cancer incidence and survival trends by subtype using gle drug” rationale has proved too simplistic for GEA because data from the Surveillance Epidemiology and End Results Program, of the intrinsic genomic instability and heterogeneity of this 1992–2013. Cancer Epidemiol Biomarkers Prev 2017;26:632–41. disease. To address these formidable obstacles, comprehensive 12. Lauren P. The two histological main types of gastric carcinoma: dif- and repeated molecular profiling will be integral to patient fuse and so-called intestinal-type carcinoma. An attempt at a histo- clinical classification. Acta Pathol Microbiol Scand 1965;64:31–49. and treatment selection, and to the success of future trans- 13. Siewert JR, Stein HJ. Carcinoma of the gastroesophageal junction - lational clinical trials in GEA. These changes in our clinical classification, pathology and extent of resection. Dis Esophagus approach to GEA will be complemented and fueled by ongo- 1996;9:173–82. ing laboratory investigation into novel genetic dependencies 14. Russo A, Li P, Strong VE. Differences in the multimodal treatment and vulnerabilities of GEA. Ultimately, although consider- of gastric cancer: East versus west. J Surg Oncol 2017;115:603–14. able challenges remain, the coming years hold the promise 15. Tachimori Y, Ozawa S, Numasaki H, Ishihara R, Matsubara H, Muro K, et al. Comprehensive registry of esophageal cancer in Japan, 2011. of a turning point in the battle against these deadly cancers. Esophagus 2018;15:127–52. 16. Nishi T, Makuuchi H, Ozawa S, Shimada H, Chino O. The present Disclosure of Potential Conflicts of Interest status and future of Barrett’s esophageal adenocarcinoma in Japan. A.J. Bass is a consultant at Lilly and Glenmark, reports receiving Digestion 2018:1–6. commercial research grants from Novartis, Merck, Sanofi, and Bayer, 17. Schumacher SE, Shim BY, Corso G, Ryu MH, Kang YK, Roviello F, and has ownership interest (including patents) in Earli, Helix Nano, et al. Somatic copy number alterations in gastric adenocarcinomas and Signet Therapeutics. No potential conflicts of interest were dis- among Asian and Western patients. PLoS One 2017;12:e0176045. closed by the other authors. 18. Lin SJ, Gagnon-Bartsch JA, Tan IB, Earle S, Ruff L, Pettinger K, et al. Signatures of tumour immunity distinguish Asian and non-Asian Acknowledgments gastric adenocarcinomas. Gut 2015;64:1721–31. 19. Cristescu R, Lee J, Nebozhyn M, Kim KM, Ting JC, Wong SS, et al. We are grateful to Michael Cooper, Medical Illustrator, for col- Molecular analysis of gastric cancer identifies subtypes associated laborative work on figures. A.K. Nagaraja is supported by the Depart- with distinct clinical outcomes. Nat Med 2015;21:449–56. ment of Defense Peer Reviewed Cancer Research Program Horizon 20. Weaver JMJ, Ross-Innes CS, Shannon N, Lynch AG, Forshew T, Award (W81XWH-17-1-0222), the Conquer Cancer Foundation of Barbera M, et al. Ordering of mutations in preinvasive disease stages ASCO/Boehringer Ingelheim Endowed Young Investigator Award of esophageal carcinogenesis. Nat Genet 2014;46:837–43. in Gastrointestinal Cancers, and the Debbie’s Dream Foundation- 21. Sohn BH, Hwang J-E, Jang H-J, Lee H-S, Oh SC, Shim J-J, et al. Stupid Strong-AACR Gastric Cancer Research Fellowship, in memory Clinical significance of four molecular subtypes of gastric cancer of Candace Netzer (19-40-41-NAGA). A.J. Bass is supported by the identified by The Cancer Genome Atlas Project. Clin Cancer Res NCI (P50CA127003, P01CA224428, and R01CA098101). Any opin- 2017;23:4441–9.

1668 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in GEA REVIEW

22. Sawas T, Killcoyne S, Iyer PG, Wang KK, Smyrk TC, Kisiel JB, et al. regulate DNA recombination repair, in patients with gastric cancer. Identification of Prognostic Phenotypes of Esophageal Adenocar- Gastroenterology 2017;152:983–6. cinoma in 2 Independent Cohorts. Gastroenterology 2018;155: 42. Hansford S, Kaurah P, Li-Chang H, Woo M, Senz J, Pinheiro H, et al. 1720–8. Hereditary diffuse gastric cancer syndrome: CDH1 mutations and 23. Secrier M, Li X, de Silva N, Eldridge MD, Contino G, Bornschein J, beyond. JAMA Oncol 2015;1:23–32. et al. Mutational signatures in esophageal adenocarcinoma define 43. Ge S, Xia X, Ding C, Zhen B, Zhou Q, Feng J, et al. A proteomic etiologically distinct subgroups with therapeutic relevance. Nat landscape of diffuse-type gastric cancer. Nat Commun 2018;9:1012. Genet 2016;48:1131–41. 44. Wang K, Yuen ST, Xu J, Lee SP, Yan HH, Shi ST, et al. Whole-genome 24. Dulak AM, Schumacher SE, van Lieshout J, Imamura Y, Fox C, sequencing and comprehensive molecular profiling identify new Shim B, et al. Gastrointestinal adenocarcinomas of the esophagus, driver mutations in gastric cancer. Nat Genet 2014;46:573–82. stomach, and colon exhibit distinct patterns of genome instability 45. Kakiuchi M, Nishizawa T, Ueda H, Gotoh K, Tanaka A, Hayashi A, and oncogenesis. Cancer Res 2012;72:4383–93. et al. Recurrent gain-of-function mutations of RHOA in diffuse- 25. Liu Y, Sethi NS, Hinoue T, Schneider BG, Cherniack AD, Sanchez- type gastric carcinoma. Nat Genet 2014;46:583–7. Vega F, et al. Comparative molecular analysis of gastrointestinal 46. Joneson T, White MA, Wigler MH, Bar-Sagi D. Stimulation of mem- adenocarcinomas. Cancer Cell 2018;33:721–35. brane ruffling and MAP kinase activation by distinct effectors of 26. Li X, Galipeau PC, Paulson TG, Sanchez CA, Arnaudo J, Liu K, et al. RAS. Science 1996;271:810–2. Temporal and spatial evolution of somatic chromosomal altera- 47. Yao F, Kausalya JP, Sia YY, Teo AS, Lee WH, Ong AG, et al. Recurrent tions: a case-cohort study of Barrett’s esophagus. Cancer Prev Res fusion genes in gastric cancer: CLDN18-ARHGAP26 induces loss of 2014;7:114–27. epithelial integrity. Cell Rep 2015;12:272–85. 27. Stachler MD, Taylor-Weiner A, Peng S, McKenna A, Agoston AT, 48. Pattison S, Mitchell C, Lade S, Leong T, Busuttil RA, Boussioutas A. Odze RD, et al. Paired exome analysis of Barrett’s esophagus and Early relapses after adjuvant chemotherapy suggests primary chem- adenocarcinoma. Nat Genet 2015;47:1047–55. oresistance in diffuse gastric cancer. PLoS One 2017;12:e0183891. 28. Nones K, Waddell N, Wayte N, Patch AM, Bailey P, Newell F, et al. 49. Choi YY, Bae JM, An JY, Kwon IG, Cho I, Shin HB, et al. Is microsat- Genomic catastrophes frequently arise in esophageal adenocarci- ellite instability a prognostic marker in gastric cancer? A systematic noma and drive tumorigenesis. Nat Commun 2014;5:5224. review with meta-analysis. J Surg Oncol 2014;110:129–35. 29. Dulak AM, Stojanov P, Peng S, Lawrence MS, Fox C, Stewart C, et al. 50. Pietrantonio F, Raimondi A, Choi YY, Kang W, Langley RE, Kim YW, Exome and whole-genome sequencing of esophageal adenocarci- et al. MSI-GC-01: Individual patient data (IPD) meta-analysis of micro- noma identifies recurrent driver events and mutational complexity. satellite instability (MSI) and gastric cancer (GC) from four rand- Nat Genet 2013;45:478–86. omized clinical trials (RCTs). J Clin Oncol 37:4s, 2019 (suppl; abstr 66). 30. Supek F, Lehner B. Clustered mutation signatures reveal that error- 51. Smyth EC, Wotherspoon A, Peckitt C, Gonzalez D, Hulkki-Wilson prone DNA repair targets mutations to active genes. Cell 2017; S, Eltahir Z, et al. Mismatch repair deficiency, microsatellite insta- 170:534–47. bility, and survival: an exploratory analysis of the medical research 31. Robles AI, Traverso G, Zhang M, Roberts NJ, Khan MA, Joseph C, et al. council adjuvant gastric infusional chemotherapy (MAGIC) trial. Whole-exome sequencing analyses of inflammatory bowel disease- JAMA Oncol 2017;3:1197–203. associated colorectal cancers. Gastroenterology 2016;150:931–43. 52. Choi YY, Kim H, Yang H-K, Kim WH, Kim YW, Kook M-C, et al. 32. Kim J, Fox C, Peng S, Pusung M, Pectasides E, Matthee E, et al. Clinical impact of microsatellite instability in patients with stage II Preexisting oncogenic events impact trastuzumab sensitivity in and III gastric cancer: results from the CLASSIC trial. J Clin Oncol ERBB2-amplified gastroesophageal adenocarcinoma. J Clin Invest 35:15s, 2017 (suppl; abstr 4022). 2014;124:5145–58. 53. Derks S, Liao X, Chiaravalli AM, Xu X, Camargo MC, Solcia E, et al. 33. Kwak EL, Ahronian LG, Siravegna G, Mussolin B, Borger DR, God- Abundant PD-L1 expression in Epstein-Barr Virus-infected gastric frey JT, et al. Molecular heterogeneity and receptor coamplification cancers. Oncotarget 2016;7:32925–32. drive resistance to targeted therapy in MET-amplified esophago- 54. Le DT, Durham JN, Smith KN, Wang H, Bartlett BR, Aulakh LK, gastric cancer. Cancer Discov 2015;5:1271–81. et al. Mismatch repair deficiency predicts response of solid tumors 34. Janjigian YY, Sanchez-Vega F, Jonsson P, Chatila WK, Hechtman JF, to PD-1 blockade. Science 2017;357:409–13. Ku GY, et al. Genetic predictors of response to systemic therapy in 55. Kim ST, Cristescu R, Bass AJ, Kim KM, Odegaard JI, Kim K, et al. esophagogastric cancer. Cancer Discov 2018;8:49–58. Comprehensive molecular characterization of clinical responses 35. Pectasides E, Stachler MD, Derks S, Liu Y, Maron S, Islam M, et al. to PD-1 inhibition in metastatic gastric cancer. Nat Med 2018;24: Genomic heterogeneity as a barrier to precision medicine in gastroe- 1449–58. sophageal adenocarcinoma. Cancer Discov 2018;8:37–48. 56. Green MR, Monti S, Rodig SJ, Juszczynski P, Currie T, O’Donnell 36. Bautista MC, Jiang SF, Armstrong MA, Postlethwaite D, Li D. E, et al. Integrative analysis reveals selective 9p24.1 amplification, Impact of age on clinicopathological features and survival of increased PD-1 ligand expression, and further induction via JAK2 patients with noncardia gastric adenocarcinoma. J Gastric Cancer in nodular sclerosing Hodgkin lymphoma and primary mediastinal 2014;14:238–45. large B-cell lymphoma. Blood 2010;116:3268–77. 37. Chung HW, Noh SH, Lim JB. Analysis of demographic character- 57. Green MR, Rodig S, Juszczynski P, Ouyang J, Sinha P, O’Donnell istics in 3242 young age gastric cancer patients in Korea. World J E, et al. Constitutive AP-1 activity and EBV infection induce Gastroenterol 2010;16:256–63. PD-L1 in Hodgkin lymphomas and posttransplant lymphoprolif- 38. Martinson HA, Shelby NJ, Alberts SR, Olnes MJ. Gastric cancer in erative disorders: implications for targeted therapy. Clin Cancer Res Alaska Native people: a cancer health disparity. World J Gastroen- 2012;18:1611–8. terol 2018;24:2722–32. 58. Merryman RW, Armand P, Wright KT, Rodig SJ. Checkpoint 39. Sanjeevaiah A, Cheedella N, Hester C, Porembka MR. Gastric can- blockade in Hodgkin and non-Hodgkin lymphoma. Blood Adv cer: recent molecular classification advances, racial disparity, and 2017;1:2643–54. management implications. J Oncol Pract 2018;14:217–24. 59. Bang YJ, Van Cutsem E, Feyereislova A, Chung HC, Shen L, Sawaki 40. Fewings E, Larionov A, Redman J, Goldgraben MA, Scarth J, A, et al. Trastuzumab in combination with chemotherapy versus Richardson S, et al. Germline pathogenic variants in PALB2 and chemotherapy alone for treatment of HER2-positive advanced gas- other cancer-predisposing genes in families with hereditary diffuse tric or gastro-oesophageal junction cancer (ToGA): a phase 3, open- gastric cancer without CDH1 mutation: a whole-exome sequencing label, randomised controlled trial. Lancet 2010;376:687–97. study. Lancet Gastroenterol Hepatol 2018;3:489–98. 60. Van Cutsem E, Bang YJ, Feng-Yi F, Xu JM, Lee KW, Jiao SC, et al. 41. Sahasrabudhe R, Lott P, Bohorquez M, Toal T, Estrada AP, Suarez JJ, HER2 screening data from ToGA: targeting HER2 in gastric and et al. Germline mutations in PALB2, BRCA1, and RAD51C, which gastroesophageal junction cancer. Gastric Cancer 2015;18:476–84.

DECEMBER 2019 CANCER DISCOVERY | 1669

REVIEW Nagaraja et al.

. 61 Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde tion in ERBB2-amplified esophagogastric cancer. Cancer Discov A, et al. Use of chemotherapy plus a monoclonal antibody against 2019;9:199–209. HER2 for metastatic breast cancer that overexpresses HER2. N Engl 77. Bartley AN, Washington MK, Colasacco C, Ventura CB, Ismaila N, J Med 2001;344:783–92. Benson AB III, et al. HER2 Testing and clinical decision making 62. Tabernero J, Hoff PM, Shen L, Ohtsu A, Shah MA, Cheng K, in gastroesophageal adenocarcinoma: guideline from the College et al. Pertuzumab plus trastuzumab and chemotherapy for HER2- of American Pathologists, American Society for Clinical Pathol- positive metastatic gastric or gastro-oesophageal junction cancer ogy, and the American Society of Clinical Oncology. J Clin Oncol (JACOB): final analysis of a double-blind, randomised, placebo- 2017;35:446–64. controlled phase 3 study. Lancet Oncol 2018;19:1372–84. 78. Hyman DM, Piha-Paul SA, Won H, Rodon J, Saura C, Shapiro GI, 63. Kang YK, Rha SY, Tassone P, Barriuso J, Yu R, Szado T, et al. A et al. HER kinase inhibition in patients with HER2- and HER3- phase IIa dose-finding and safety study of first-line pertuzumab mutant cancers. Nature 2018;554:189–94. in combination with trastuzumab, capecitabine and cisplatin in 79. Flaherty KT, Puzanov I, Kim KB, Ribas A, McArthur GA, Sosman patients with HER2-positive advanced gastric cancer. Br J Cancer JA, et al. Inhibition of mutated, activated BRAF in metastatic mela- 2014;111:660–6. noma. N Engl J Med 2010;363:809–19. 64. Kirschbrown WP, Wang B, Nijem I, Ohtsu A, Hoff PM, Shah 80. Kopetz S, Desai J, Chan E, Hecht JR, O’Dwyer PJ, Maru D, et al. MA, et al. Pharmacokinetic and exposure–response analysis of Phase II pilot study of vemurafenib in patients with metastatic pertuzumab in patients with HER2-positive metastatic gastric or BRAF-mutated colorectal cancer. J Clin Oncol 2015;33:4032–8. gastroesophageal junction cancer. Cancer Chemother Pharmacol 81. Maron SB, Alpert L, Kwak HA, Lomnicki S, Chase L, Xu D, et al. Tar- 2019;84:539–50. geted therapies for targeted populations: anti-EGFR treatment for 65. Thuss-Patience PC, Shah MA, Ohtsu A, Van Cutsem E, Ajani JA, EGFR-amplified gastroesophageal adenocarcinoma. Cancer Discov Castro H, et al. Trastuzumab emtansine versus taxane use for 2018;8:696–713. previously treated HER2-positive locally advanced or metastatic 82. Kim MA, Lee HS, Lee HE, Jeon YK, Yang HK, Kim WH. EGFR in gastric or gastro-oesophageal junction adenocarcinoma (GATSBY): gastric carcinomas: prognostic significance of protein overexpres- an international randomised, open-label, adaptive, phase 2/3 study. sion and high gene copy number. Histopathology 2008;52:738–46. Lancet Oncol 2017;18:640–53. 83. Wang KL, Wu TT, Choi IS, Wang H, Resetkova E, Correa AM, et al. 66. Hecht JR, Bang YJ, Qin SK, Chung HC, Xu JM, Park JO, et al. Lapa- Expression of epidermal growth factor receptor in esophageal and tinib in combination with capecitabine plus oxaliplatin in human esophagogastric junction adenocarcinomas: association with poor epidermal growth factor receptor 2-positive advanced or metastatic outcome. Cancer 2007;109:658–67. gastric, esophageal, or gastroesophageal adenocarcinoma: TRIO-013/ 84. Langer R, Von Rahden BH, Nahrig J, Von Weyhern C, Reiter R, Feith LOGiC–a randomized phase III trial. J Clin Oncol 2016;34:443–51. M, et al. Prognostic significance of expression patterns of c-erbB-2, 67. Satoh T, Xu RH, Chung HC, Sun GP, Doi T, Xu JM, et al. Lapa- p53, p16INK4A, p27KIP1, cyclin D1 and epidermal growth factor tinib plus paclitaxel versus paclitaxel alone in the second-line receptor in oesophageal adenocarcinoma: a tissue microarray study. treatment of HER2-amplified advanced gastric cancer in Asian J Clin Pathol 2006;59:631–4. populations: TyTAN–a randomized, phase III study. J Clin Oncol 85. Lordick F, Kang YK, Chung HC, Salman P, Oh SC, Bodoky G, et al. 2014;32:2039–49. Capecitabine and cisplatin with or without cetuximab for patients 68. Shitara K, Iwata H, Takahashi S, Tamura K, Park H, Modi S, et al. with previously untreated advanced gastric cancer (EXPAND): a Trastuzumab deruxtecan (DS-8201a) in patients with advanced randomised, open-label phase 3 trial. Lancet Oncol 2013;14:490–9. HER2-positive gastric cancer: a dose-expansion, phase 1 study. Lan- 86. Waddell T, Chau I, Cunningham D, Gonzalez D, Okines AF, Okines cet Oncol 2019;20:827–36. C, et al. Epirubicin, oxaliplatin, and capecitabine with or without 69. Bang YJ, Giaccone G, Im SA, Oh DY, Bauer TM, Nordstrom JL, et al. panitumumab for patients with previously untreated advanced First-in-human phase 1 study of margetuximab (MGAH22), an oesophagogastric cancer (REAL3): a randomised, open-label phase Fc-modified chimeric monoclonal antibody, in patients with HER2- 3 trial. Lancet Oncol 2013;14:481–9. positive advanced solid tumors. Ann Oncol 2017;28:855–61. 87. Dutton SJ, Ferry DR, Blazeby JM, Abbas H, Dahle-Smith A, Man- 70. Catenacci DVT, Park H, Uronis HE, Kang YK, Ng M, Gold P, soor W, et al. Gefitinib for oesophageal cancer progressing after et al. Biomarker-guided enrichment of the antitumor activity chemotherapy (COG): a phase 3, multicentre, double-blind, placebo- of margetuximab (M) plus pembrolizumab (P) in patients with controlled randomised trial. Lancet Oncol 2014;15:894–904. advanced HER2+ gastric adenocarcinoma (GEA). Ann Oncol 2018; 88. Petty RD, Dahle-Smith A, Stevenson DAJ, Osborne A, Massie D, 29(suppl_8):mdy282.046. Clark C, et al. Gefitinib and EGFR gene copy number aberrations in 71. Hofmann M, Stoss O, Shi D, Buttner R, van de Vijver M, Kim W, esophageal cancer. J Clin Oncol 2017;35:2279–87. et al. Assessment of a HER2 scoring system for gastric cancer: results 89. Lordick F, Kang Y-K, Salman P, Oh SC, Bodoky G, Kurteva GP, from a validation study. Histopathology 2008;52:797–805. et al. Clinical outcome according to tumor HER2 status and EGFR 72. Wang DS, Liu ZX, Lu YX, Bao H, Wu X, Zeng ZL, et al. Liquid biop- expression in advanced gastric cancer patients from the EXPAND sies to track trastuzumab resistance in metastatic HER2-positive study. J Clin Oncol 31:15s, 2013 (suppl; abstr 4021). gastric cancer. Gut 2019;68:1152–61. 90. Smolen GA, Sordella R, Muir B, Mohapatra G, Barmettler A, 73. Shah MA, Kang YK, Thuss-Patience PC, Ohtsu A, Ajani JA, Van Archibald H, et al. Amplification of MET may identify a subset Cutsem E, et al. Biomarker analysis of the GATSBY study of tras- of cancers with extreme sensitivity to the selective tyrosine kinase tuzumab emtansine versus a taxane in previously treated HER2- inhibitor PHA-665752. Proc Natl Acad Sci U S A 2006;103:2316–21. positive advanced gastric/gastroesophageal junction cancer. Gastric 91. Lee HE, Kim MA, Lee HS, Jung EJ, Yang HK, Lee BL, et al. MET in Cancer 2019;22:803–16. gastric carcinomas: comparison between protein expression and 74. Kim ST, Banks KC, Pectasides E, Kim SY, Kim K, Lanman RB, et al. gene copy number and impact on clinical outcome. Br J Cancer Impact of genomic alterations on lapatinib treatment outcome and 2012;107:325–33. cell-free genomic landscape during HER2 therapy in HER2+ gastric 92. Janjigian YY, Tang LH, Coit DG, Kelsen DP, Francone TD, Weiser cancer patients. Ann Oncol 2018;29:1037–48. MR, et al. MET expression and amplification in patients with 75. Lee JY, Hong M, Kim ST, Park SH, Kang WK, Kim KM, et al. The localized gastric cancer. Cancer Epidemiol Biomarkers Prev impact of concomitant genomic alterations on treatment outcome 2011;20:1021–7. for trastuzumab therapy in HER2-positive gastric cancer. Sci Rep 93. Nakajima M, Sawada H, Yamada Y, Watanabe A, Tatsumi M, 2015;5:9289. Yamashita J, et al. The prognostic significance of amplification and 76. Sanchez-Vega F, Hechtman JF, Castel P, Ku GY, Tuvy Y, Won H, et al. overexpression of c-met and c-erb B-2 in human gastric carcinomas. EGFR and MET amplifications determine response to HER2 inhibi- Cancer 1999;85:1894–902.

1670 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in GEA REVIEW

94. Shah MA, Bang YJ, Lordick F, Alsina M, Chen M, Hack SP, et al. ment of metastatic renal-cell carcinoma. Clin Genitourin Cancer Effect of fluorouracil, leucovorin, and oxaliplatin with or without 2019;17:e513–e21. onartuzumab in HER2-negative, MET-positive gastroesophageal 110. Cohen NA, Strong VE, Janjigian YY. Checkpoint blockade in esoph- adenocarcinoma: the METGastric Randomized Clinical Trial. JAMA agogastric cancer. J Surg Oncol 2018;118:77–85. Oncol 2017;3:620–7. 111. Tabernero J, Cutsem EV, Bang Y-J, Fuchs CS, Wyrwicz L, Lee KW, 95. Shah MA, Cho JY, Tan IB, Tebbutt NC, Yen CJ, Kang A, et al. A rand- et al. Pembrolizumab with or without chemotherapy versus chemo- omized phase II study of FOLFOX with or without the MET inhibi- therapy for advanced gastric or gastroesophageal junction (G/GEJ) tor onartuzumab in advanced adenocarcinoma of the stomach and adenocarcinoma: The phase III KEYNOTE-062 study. J Clin Oncol gastroesophageal junction. Oncologist 2016;21:1085–90. 37:18s, 2019 (suppl; abstr LBA4007). 96. Catenacci DVT, Tebbutt NC, Davidenko I, Murad AM, Al-Batran 112. Fuchs CS, Doi T, Jang RW, Muro K, Satoh T, Machado M, et al. SE, Ilson DH, et al. Rilotumumab plus epirubicin, cisplatin, and Safety and efficacy of pembrolizumab monotherapy in patients capecitabine as first-line therapy in advanced MET-positive gas- with previously treated advanced gastric and gastroesophageal junc- tric or gastro-oesophageal junction cancer (RILOMET-1): a ran- tion cancer: phase 2 clinical KEYNOTE-059 trial. JAMA Oncol domised, double-blind, placebo-controlled, phase 3 trial. Lancet 2018;4:e180013. Oncol 2017;18:1467–82. 113. Kang YK, Boku N, Satoh T, Ryu MH, Chao Y, Kato K, et al. 97. Van Cutsem E, Karaszewska B, Kang YK, Chung HC, Shankaran V, Nivolumab in patients with advanced gastric or gastro-oesophageal Siena S, et al. A multicenter phase II study of AMG 337 in patients junction cancer refractory to, or intolerant of, at least two previ- with MET-amplified gastric/gastroesophageal junction/esophageal ous chemotherapy regimens (ONO-4538-12, ATTRACTION-2): a adenocarcinoma and other MET-amplified solid tumors. Clin Can- randomised, double-blind, placebo-controlled, phase 3 trial. Lancet cer Res 2019;25:2414–23. 2017;390:2461–71. 98. Hong DS, LoRusso P, Hamid O, Janku F, Kittaneh M, Catenacci 114. Shitara K, Ozguroglu M, Bang YJ, Di Bartolomeo M, Mandala M, DVT, et al. Phase I study of AMG 337, a highly selective small- Ryu MH, et al. Pembrolizumab versus paclitaxel for previously molecule MET inhibitor, in patients with advanced solid tumors. treated, advanced gastric or gastro-oesophageal junction cancer Clin Cancer Res 2019;25:2403–13. (KEYNOTE-061): a randomised, open-label, controlled, phase 3 99. Ohtsu A, Ajani JA, Bai YX, Bang YJ, Chung HC, Pan HM, et al. trial. Lancet 2018;392:123–33. Everolimus for previously treated advanced gastric cancer: results of 115. Shah MA, Kojima T, Hochhauser D, Enzinger P, Raimbourg J, the randomized, double-blind, phase III GRANITE-1 study. J Clin Hollebecque A, et al. Efficacy and safety of pembrolizumab for heav- Oncol 2013;31:3935–43. ily pretreated patients with advanced, metastatic adenocarcinoma or 100. Yoon DH, Ryu MH, Park YS, Lee HJ, Lee C, Ryoo BY, et al. Phase II squamous cell carcinoma of the esophagus: the phase 2 KEYNOTE- study of everolimus with biomarker exploration in patients with 180 study. JAMA Oncol 2019;5:546–50. advanced gastric cancer refractory to chemotherapy including fluo- 116. Shah MA, Adenis A, Enzinger PC, Kojima T, Muro K, Bennouna J, ropyrimidine and platinum. Br J Cancer 2012;106:1039–44. et al. Pembrolizumab versus chemotherapy as second-line therapy 101. Xie L, Su X, Zhang L, Yin X, Tang L, Zhang X, et al. FGFR2 gene for advanced esophageal cancer: Phase 3 KEYNOTE-181 study. amplification in gastric cancer predicts sensitivity to the selective J Clin Oncol 37:15s, 2019 (suppl; abstr 4010). FGFR inhibitor AZD4547. Clin Cancer Res 2013;19:2572–83. 117. Heitzer E, Haque IS, Roberts CES, Speicher MR. Current and future 102. Pearson A, Smyth E, Babina IS, Herrera-Abreu MT, Tarazona N, perspectives of liquid biopsies in genomics-driven oncology. Nat Peckitt C, et al. High-level clonal FGFR amplification and response Rev Genet 2019;20:71–88. to FGFR inhibition in a translational clinical trial. Cancer Discov 118. Joshi SS, Maron SB, Lomnicki S, Polite BN, Sharma M, Ibe J, 2016;6:838–51. et al. Personalized antibodies for gastroesophageal adenocarcinoma 103. Van Cutsem E, Bang YJ, Mansoor W, Petty RD, Chao Y, Cunning- (PANGEA): a phase II precision medicine trial (NCT02213289). J ham D, et al. A randomized, open-label study of the efficacy and Clin Oncol 36:4s, 2018 (suppl; abstr TPS198). safety of AZD4547 monotherapy versus paclitaxel for the treatment 119. Lee J, Kim ST, Kim K, Lee H, Kozarewa I, Mortimer PGS, et al. of advanced gastric adenocarcinoma with FGFR2 polysomy or gene Tumor genomic profiling guides patients with metastatic gastric amplification. Ann Oncol 2017;28:1316–24. cancer to targeted treatment: The VIKTORY Umbrella Trial. Cancer 104. Fuchs CS, Tomasek J, Yong CJ, Dumitru F, Passalacqua R, Discov 2019;9:1388–405. Goswami C, et al. Ramucirumab monotherapy for previously 120. Wong GS, Zhou J, Liu JB, Wu Z, Xu X, Li T, et al. Targeting wild-type treated advanced gastric or gastro-oesophageal junction adeno- KRAS-amplified gastroesophageal cancer through combined MEK carcinoma (REGARD): an international, randomised, multicentre, and SHP2 inhibition. Nat Med 2018;24:968–77. placebo-controlled, phase 3 trial. Lancet 2014;383:31–9. 121. Frankell AM, Jammula S, Li X, Contino G, Killcoyne S, Abbas S, 105. Wilke H, Muro K, Van Cutsem E, Oh SC, Bodoky G, Shimada Y, et al. The landscape of selection in 551 esophageal adenocarcinomas et al. Ramucirumab plus paclitaxel versus placebo plus paclitaxel defines genomic biomarkers for the clinic. Nat Genet 2019;51:506–16. in patients with previously treated advanced gastric or gastro- 122. Karasic TB, O’Hara MH, Teitelbaum UR, Damjanov N, Giantonio oesophageal junction adenocarcinoma (RAINBOW): a double- BJ, d’Entremont TS, et al. Phase II trial of palbociclib in patients blind, randomised phase 3 trial. Lancet Oncol 2014;15:1224–35. with advanced esophageal or gastric cancer. J Clin Oncol;36:4s, 2018 106. Fuchs CS, Shitara K, Di Bartolomeo M, Lonardi S, Al-Batran SE, (suppl; abstr 68). Van Cutsem E, et al. Ramucirumab with cisplatin and fluoropyrimi- 123. Chen Z, Huang W, Tian T, Zang W, Wang J, Liu Z, et al. Characteri- dine as first-line therapy in patients with metastatic gastric or junc- zation and validation of potential therapeutic targets based on the tional adenocarcinoma (RAINFALL): a double-blind, randomised, molecular signature of patient-derived xenografts in gastric cancer. placebo-controlled, phase 3 trial. Lancet Oncol 2019;20:420–35. J Hematol Oncol 2018;11:20. 107. Ohtsu A, Shah MA, Van Cutsem E, Rha SY, Sawaki A, Park SR, 124. Au-Yeung G, Lang F, Azar WJ, Mitchell C, Jarman KE, Lackovic K, et al. Bevacizumab in combination with chemotherapy as first-line et al. Selective targeting of cyclin E1-amplified high-grade serous therapy in advanced gastric cancer: a randomized, double-blind, ovarian cancer by cyclin-dependent kinase 2 and AKT inhibition. placebo-controlled phase III study. J Clin Oncol 2011;29:3968–76. Clin Cancer Res 2017;23:1862–74. 108. Van Cutsem E, de Haas S, Kang YK, Ohtsu A, Tebbutt NC, Ming Xu 125. Etemadmoghadam D, Weir BA, Au-Yeung G, Alsop K, Mitchell G, J, et al. Bevacizumab in combination with chemotherapy as first-line George J, et al. Synthetic lethality between CCNE1 amplification therapy in advanced gastric cancer: a biomarker evaluation from the and loss of BRCA1. Proc Natl Acad Sci U S A 2013;110:19489–94. AVAGAST randomized phase III trial. J Clin Oncol 2012;30:2119–27. 126. Alexandrov LB, Nik-Zainal S, Siu HC, Leung SY, Stratton MR. A 109. Labriola MK, Batich KA, Zhu J, McNamara MA, Harrison MR, mutational signature in gastric cancer suggests therapeutic strate- Armstrong AJ, et al. Immunotherapy is changing first-line treat- gies. Nat Commun 2015;6:8683.

DECEMBER 2019 CANCER DISCOVERY | 1671

REVIEW Nagaraja et al.

.127 Gulhan DC, Lee JJ, Melloni GEM, Cortes-Ciriano I, Park PJ. Detect- for therapeutic antibody development. Clin Cancer Res 2008;14: ing the mutational signature of homologous recombination defi- 7624–34. ciency in clinical samples. Nat Genet 2019;51:912–9. 134. Schuler M, Al-Batran SE, Zvirbule Z, Manikhas G, Lordick F, 128. Bang YJ, Im SA, Lee KW, Cho JY, Song EK, Lee KH, et al. Rand- Rusyn A, et al. Final results of the FAST study, an international, omized, double-blind phase II trial with prospective classification by multicenter, randomized, phase II trial of epirubicin, oxaliplatin, ATM protein level to evaluate the efficacy and tolerability of olaparib and capecitabine (EOX) with or without the anti-CLDN18.2 anti- plus paclitaxel in patients with recurrent or metastatic gastric cancer. body IMAB362 as first-line therapy in patients with advanced J Clin Oncol 2015;33:3858–65. CLDN18.2+ gastric and gastroesophageal junction (GEJ) adenocar- 129. Bang YJ, Xu RH, Chin K, Lee KW, Park SH, Rha SY, et al. Olaparib cinoma. Ann Oncol 2016;27:614O. in combination with paclitaxel in patients with advanced gastric 135. Sahin U, Tureci Ö, Manikhas GM, Lordick F, Rusyn A, Vynnych- cancer who have progressed following first-line therapy (GOLD): a enko I, et al. Zolbetuximab combined with EOX as first-line therapy double-blind, randomised, placebo-controlled, phase 3 trial. Lancet in advanced CLDN18.2+ gastric (G) and gastroesophageal junction Oncol 2017;18:1637–51. (GEJ) adenocarcinoma: updated results from the FAST trial. J Clin 130. Smyth E. Missing a GOLDen opportunity in gastric cancer. Lancet Oncol 37:4s, 2019 (suppl; abstr 16). Oncol 2017;18:1561–3. 136. Al-Batran S-E, Schuler MH, Zvirbule Z, Manikhas G, Lordick F, 131. Ciardiello F, Bang Y-J, Bendell JC, Cervantes A, Brachmann RK, Rusyn A, et al. FAST: an international, multicenter, randomized, Zhang Y, et al. A phase III, double-blind, randomized study of phase II trial of epirubicin, oxaliplatin, and capecitabine (EOX) pamiparib versus placebo as maintenance therapy in patients with with or without IMAB362, a first-in-class anti-CLDN18.2 antibody, inoperable, locally advanced, or metastatic gastric cancer (GC) that as first-line therapy in patients with advanced CLDN18.2+ gastric responded to platinum-based first-line chemotherapy. J Clin Oncol and gastroesophageal junction (GEJ) adenocarcinoma. J Clin Oncol 37:4s, 2019 (suppl; abstr TPS173). 34:18s, 2016; (suppl; abstr LBA4001). 132. Fang Y, McGrail DJ, Sun C, Labrie M, Chen X, Zhang D, et al. 137. Zhu G, Foletti D, Liu X, Ding S, Melton Witt J, Hasa-Moreno A, et al. Sequential therapy with PARP and WEE1 inhibitors mini- Targeting CLDN18.2 by CD3 bispecific and ADC modalities for the mizes toxicity while maintaining efficacy. Cancer Cell 2019;35: treatments of gastric and pancreatic cancer. Sci Rep 2019;9:8420. 851–67. 138. Jiang H, Shi Z, Wang P, Wang C, Yang L, Du G, et al. Claudin18.2- 133. Sahin U, Koslowski M, Dhaene K, Usener D, Brandenburg G, Seitz specific chimeric antigen receptor engineered T cells for the treat- G, et al. Claudin-18 splice variant 2 is a pan-cancer target suitable ment of gastric cancer. J Natl Cancer Inst 2019;111:409–18.

1672 | CANCER DISCOVERY DECEMBER 2019 www.aacrjournals.org

Genomics and Targeted Therapies in Gastroesophageal Adenocarcinoma

Ankur K. Nagaraja, Osamu Kikuchi and Adam J. Bass

Cancer Discov 2019;9:1656-1672. Published OnlineFirst November 14, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-19-0487

Cited articles This article cites 126 articles, 34 of which you can access for free at: http://cancerdiscovery.aacrjournals.org/content/9/12/1656.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://cancerdiscovery.aacrjournals.org/content/9/12/1656.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/9/12/1656. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.