Ramucirumab monotherapy for previously treated advanced gastric or gastro-oesophageal junction adenocarcinoma (REGARD): an international, randomised, multicentre, placebo-controlled,

Articles

www.thelancet.com Vol 383 January 4, 2014 31

Ramucirumab monotherapy for previously treated advanced gastric or gastro-oesophageal junction adenocarcinoma (REGARD): an international, randomised, multicentre, placebo-controlled, phase 3 trialCharles S Fuchs, Jiri Tomasek, Cho Jae Yong, Filip Dumitru, Rodolfo Passalacqua, Chanchal Goswami, Howard Safran, Lucas Vieira dos Santos, Giuseppe Aprile, David R Ferry, Bohuslav Melichar, Mustapha Tehfe, Eldar Topuzov, John Raymond Zalcberg, Ian Chau, William Campbell, Choondal Sivanandan, Joanna Pikiel, Minori Koshiji, Yanzhi Hsu, Astra M Liepa, Ling Gao, Jonathan D Schwartz, Josep Tabernero, for the REGARD Trial Investigators*

SummaryBackground Vascular endothelial growth factor (VEGF) and VEGF receptor-2 (VEGFR-2)-mediated signalling and angiogenesis can contribute to the pathogenesis and progression of gastric cancer. We aimed to assess whether ramucirumab, a monoclonal antibody VEGFR-2 antagonist, prolonged survival in patients with advanced gastric cancer.

Methods We did an international, randomised, double-blind, placebo-controlled, phase 3 trial between Oct 6, 2009, and Jan 26, 2012, at 119 centres in 29 countries in North America, Central and South America, Europe, Asia, Australia, and Africa. Patients aged 24–87 years with advanced gastric or gastro-oesophageal junction adenocarcinoma and disease progression after fi rst-line platinum-containing or fl uoropyrimidine-containing chemotherapy were randomly assigned (2:1), via a central interactive voice-response system, to receive best supportive care plus either ramucirumab 8 mg/kg or placebo, intravenously once every 2 weeks. The study sponsor, participants, and investigators were masked to treatment assignment. The primary endpoint was overall survival. Analysis was by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT00917384.

Findings 355 patients were assigned to receive ramucirumab (n=238) or placebo (n=117). Median overall survival was 5·2 months (IQR 2·3–9·9) in patients in the ramucirumab group and 3·8 months (1·7–7·1) in those in the placebo group (hazard ratio [HR] 0·776, 95% CI 0·603–0·998; p=0·047). The survival benefi t with ramucirumab remained unchanged after multivariable adjustment for other prognostic factors (multivariable HR 0·774, 0·605–0·991; p=0·042). Rates of hypertension were higher in the ramucirumab group than in the placebo group (38 [16%] vs nine [8%]), whereas rates of other adverse events were mostly similar between groups (223 [94%] vs 101 [88%]). Five (2%) deaths in the ramucirumab group and two (2%) in the placebo group were considered to be related to study drug.

Interpretation Ramucirumab is the fi rst biological treatment given as a single drug that has survival benefi ts in patients with advanced gastric or gastro-oesophageal junction adenocarcinoma progressing after fi rst-line chemotherapy. Our fi ndings validate VEGFR-2 signalling as an important therapeutic target in advanced gastric cancer.

Funding ImClone Systems.

IntroductionWorldwide, gastric cancer is the fourth most common malignant disease and the second leading cause of cancer mortality.1 Standard cytotoxic chemotherapy is typically used as fi rst-line treatment for advanced gastric adenocarcinoma, with median survival ranging from 8 months to 10 months.2 After fi rst-line treatment of gastric cancer, no second-line treatment is yet approved by the US Food and Drug Administration, European Medicines Agency, or most other govern-mental drug regulatory agencies. New targeted systemic drugs are needed to improve the outcome of patients with advanced disease.

Vascular endothelial growth factor (VEGF) and VEGF receptor-2 (VEGFR-2)-mediated signalling and

angiogenesis seem to have an important role in the pathogenesis of gastric cancer. In animal models of gastric adenocarcinoma, VEGFR-2 inhibition has reduced tumour growth and vascularity.3 In patients with gastric cancer, circulating and tumoral con-centrations of VEGF are associated with increased tumour aggressiveness and reduced survival.4–13 Ramucirumab is a fully human IgG1 monoclonal antibody VEGFR-2 antagonist that prevents ligand binding and receptor-mediated pathway activation in endothelial cells.14

We assessed the safety and effi cacy of ramucirumab in patients with advanced gastric or gastro-oesophageal junction adenocarcinoma who had disease progression after fi rst-line chemotherapy.

Lancet 2014; 383: 31–39

Published OnlineOctober 3, 2013http://dx.doi.org/10.1016/S0140-6736(13)61719-5

See Comment page 4

*See appendix for list of principal investigators

Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA (Prof C S Fuchs MD); Department of Complex Oncology Care, Masaryk Memorial Cancer Institute, Brno, Czech Republic (J Tomasek MD); Department of Medical Oncology, Yonsei University College of Medicine, Gangnam Severance Hospital, Seoul, Korea (Prof C J Yong MD); Oncology Department, Emergency County Hospital “Dr Constantin Opris”, Baia Mare, Romania (F Dumitru MD); Medical Oncology Division, Istituti Ospitalieri di Cremona, Cremona, Italy (R Passalacqua MD); Department of Medical Oncology, B P Poddar Hospital and Medical Research, Kolkata, West Bengal (Prof C Goswami MD); Department of Medicine, The Brown University Oncology Group, Brown University, Providence, RI, USA (Prof H Safran MD); Medical Oncology Department, Gastrointestinal Oncology Division, Hospital de Câncer de Barretos and Hemomed Instituto de Oncologia e Hematologia, São Paulo, Brazil (L V dos Santos MD); Department of Oncology, University and General Hospital, Udine, Italy (G Aprile MD); Department of Medical Oncology, New Cross Hospital, West Midlands, UK (Prof D R Ferry MD); Department of Oncology, Palacky University

Articles

32 www.thelancet.com Vol 383 January 4, 2014

Medical School and Teaching Hospital, Olomouc, Czech

Republic (Prof B Melichar MD); Department of Medical

Oncology, Hôpital Notre Dame de CHUM, Montreal, Quebec

(M Tehfe MD); State Budgetary Educational Institution of

Higher Professional Education (SBEIHPE), “Northwest State

Medical University na II Mechnikov”, Ministry of

Healthcare of the Russian Federation, Russia

(Prof E Topuzov MD); Division of Cancer Medicine, Peter

McCallum Cancer Centre, East Melbourne, VIC, Australia

(Prof J R Zalcberg MD); Departments of Medicine and

Oncology, Faculty of Medicine, University of Melbourne,

Australia (Prof J R Zalcberg); Department of Medicine, Royal Marsden Hospital, London and

Surrey, England (I Chau MD); Department of Oncology, Hospital Herrera Llerandi-

Clinicas Médicas, Guatemala (Prof W Campbell, MD); Division

of Cancer Research, Regional Cancer Centre, Kerala, India

(C Sivanandan MD); Wojewódzkie Centrum

Onkologii Gdańsk, Poland

MethodsStudy design and patientsWe did this international, randomised, double-blind, placebo-controlled, phase 3 trial between Oct 6, 2009, and Jan 26, 2012, at 119 centres in 29 countries in North, Central, and South America, and Europe, Asia, Australia, and Africa. Eligible patients aged 24–87 years had metastatic or unresectable, locally recurrent gastric or gastro-oesophageal junction adenocarcinoma; had had disease progression within 4 months of the last dose of fi rst-line platinum-containing or fl uoropyrimidine-con-taining chemo therapy for metastatic disease, or within 6 months of the last dose of platinum-containing or fl uoropyrimidine-containing adjuvant treatment; and had an Eastern Cooperative Oncology Group (ECOG) performance status score of 0 or 1. We included patients with measurable disease (defi ned by Response Evaluation Crieria In Solid Tumours [RECIST] version 1.0)15 or evaluable disease. Major exclusion criteria included any grade 3 or higher gastrointestinal bleeding (defi ned by National Cancer Institute Common Terminology Criteria for Adverse Events [NCI-CTCAE], version 4.02)16 within 3 months before randomisation; any arterial thrombo-embolic event, including myocardial infarction, unstable angina, transient ischaemic event, cerebrovascular acci-dent within 6 months before random isation; or poorly controlled hypertension. The appendix provides full inclusion and exclusion criteria.

Each centre’s institutional review board or independent ethics committee approved the study. The trial followed the guiding principles of the Declaration of Helsinki and the Good Clinical Practice Guidelines of the International Conference on Harmonization. All patients provided written informed consent.

Randomisation and maskingPatients were randomly assigned in a 2:1 ratio, via a centralised interactive voice-response system, to receive best supportive care plus either ramucirumab 8 mg/kg or placebo, intravenously once every 2 weeks. Ran-domisation was stratifi ed by weight loss (≥10% vs <10% of bodyweight in the previous 3 months), geographic region (North America, Europe, Australia, and New Zealand vs South and Central America, India, South Africa, and Middle East vs Asia), and location of the primary tumour (gastric vs gastro-oesophageal junc-tion). Patients, medical and ancillary staff , the study investigators, and the sponsor were masked to group assignment. Unmasking could be done for indiv idual patients in emergencies only; serious adverse events did not necessarily precipitate immediate unmasking.

ProceduresAll patients received best supportive care, excluding other investigational antitumour drugs or antineoplastic chemotherapy, hormonal treatment, or immunotherapy. Patients received study treatment until radiographically confi rmed disease progression (investigator assessed by RECIST15), unacceptable toxic eff ects, or death. No crossover between treatment groups was allowed. We permitted predefi ned dose delays, reductions, and dis-continuations to manage clinically signifi cant treatment-related toxic eff ects (appendix).

The primary endpoint was overall survival, defi ned as the time from randomisation to death from any cause. Secondary endpoints were progression-free survival (defi ned as the time from randomisation to progres sive disease or death from any cause, whichever happened fi rst), 12 week progression-free survival, objective response rate, duration of response, quality of life, safety, and ramucirumab immunogenicity. Analyses of pharma-co dynamic markers will be reported separately else-where. Investigator-assessed objective response rate was measured according to RECIST;15 responses were con-fi rmed at least 28 days after the fi rst documentation of a complete or partial response.

Study investigators did tumour assessments at base-line and every 6 weeks thereafter until disease pro-gression. Quality of life was assessed at baseline and after 6, 12, and 18 weeks of treatment with the Euro-pean Organ isation for Research and Treatment of Cancer quality-of-life questionnaire (EORTC QLQ-C30, version 3.0);17 patients who discontinued treatment before scheduled assessments were not expected to pro-vide quality-of-life data. Local laboratory assessments

Figure 1: Trial profi le*Assessed both radiographically and symptomatically. †As of the July 25, 2012, data cutoff date.

459 patients screened

355 randomised

Intention-to-treatpopulation

Safety population

238 assigned to ramucirumab 117 assigned to placebo

2 did not receive study drug

236 receiving ramucirumab 115 receiving placebo

14 receiving ramucirumab† 1 receiving placebo†

2 did not receive study drug

104 not randomised 93 did not meet inclusion criteria 9 withdrew consent 2 died before randomisation

222 discontinued 167 had progressive disease* 25 had adverse events 20 died 7 withdrew consent 3 other

114 discontinued 89 had progressive disease* 7 had adverse events 13 died 3 withdrew consent 2 other

Articles

www.thelancet.com Vol 383 January 4, 2014 33

(J Pikiel MD); ImClone Systems LLC, Bridgewater, NJ, USA (M Koshiji MD, Y Hsu PhD, L Gao PhD, J D Schwartz MD); Eli Lilly, Indianapolis, IN, USA (A M Liepa PharmD); and Medical Oncology Department

and assessments of performance status were done at baseline, on the day of each treatment, and 30 days after the last dose of study drug. Blood for analysis of anti-ramucirumab antibodies was collected before treatment, before cycles 4 and 7 (6 and 12 weeks after the fi rst dose), and at 30 day follow-up, and patient sera were analysed with a validated ELISA format (appendix). We monitored adverse events continuously and graded them according to NCI-CTCAE scales.16

Statistical analysisAll analyses were done with use of data obtained as of the cutoff date of July 25, 2012, at which point 278 patients had died. The primary effi cacy analysis was a stratifi ed log-rank comparison of overall survival in the intention-to-treat population. Safety analyses included all patients who received at least one dose of study drug.

The trial was originally designed for a planned enrolment of 615 patients with 90% power to detect a hazard ratio (HR) for overall survival of 0·71 with a two-sided α level of 0·05. In November, 2010, with all data still concealed to the investigators and the sponsor, and with only 36 patients randomised, the protocol was amended to address slower than expected early enrolment while maintaining capacity to identify an improvement in overall survival that would be of suffi cient clinical interest. As such, total enrolment was changed to enable 80% power to detect an HR of 0·69 for overall survival, with a two-sided α level of 0·05. This change was initially calculated to require 315 patients, but was corrected to a total enrolment of 348 patients with 268 deaths in an amendment made in October, 2011.

A single interim analysis for futility was planned after 94 (35%) of the 268 total deaths had happened, and was done according to a group sequential design with a futility boundary established in accordance with the β-spending function gamma (−1).18

We used a stratifi ed log-rank test to assess overall sur-vival and progression-free survival in the intention-to-treat popu lation, with stratifi cation according to the pre defi ned randomisation strata. Estimation of survival curves was generated with the Kaplan-Meier method. The HR was estimated with stratifi ed Cox proportional hazards models. We did a multivariable analysis with a stepwise Cox regres-sion model of predefi ned baseline charac teristics to examine the eff ect of treatment after adjustment for other statistically signifi cant prognostic factors (appen dix). Add-itionally, we used unstratifi ed analyses and alter native Cox regression models to assess the eff ect of baseline patient and disease characteristics on treatment eff ect.

We estimated 12 week progression-free survival with the Kaplan-Meier method. We compared objective response rates in each treatment group with the exact Cochran-Mantel-Haenzel test, with adjustment for the stratifi cation factors. We calculated the rate of disease control, defi ned as the best overall response for complete or partial response or stable disease, for each treatment group. For each quality-of-life measure, we classifi ed

patients as improved or worsened if change from baseline was 10 points or more (on the basis of a standardised 100 point scale);19 we regarded changes of less than 10 points as stable. Additionally, we used the Kaplan-Meier method to identify time to deterioration of ECOG per-formance status between treatment groups, defi ned as

Ramucirumab (n=238)

Placebo (n=117)

Age (years) 60 (52–67) 60 (51–71)

Sex

Male 169 (71%) 79 (68%)

Female 69 (29%) 38 (32%)

Race*

White 181 (76%) 91 (78%)

Asian 39 (16%) 17 (15%)

Black 4 (2%) 2 (2%)

Other 14 (6%) 7 (6%)

ECOG performance status†

0 67 (28%) 31 (26%)

1 171 (72%) 85 (73%)

2 0 1 (1%)

Weight loss in the previous 3 months*

≥10% 41 (17%) 20 (17%)

<10% 197 (83%) 97 (83%)

Geographic region

North America, Europe, Australia, New Zealand 165 (69%) 80 (68%)

Asia 18 (8%) 8 (7%)

South and Central America, India, South Africa, Middle East 55 (23%) 29 (25%)

Location of primary tumour (%)

Gastric 178 (75%) 87 (74%)

Gastro-oesophageal junction 60 (25%) 30 (26%)

Histological subtype (%)

Intestinal 52 (22%) 35 (30%)

Diff use 96 (40%) 44 (38%)

Unknown or NA 90 (38%) 38 (32%)

Primary tumour present 174 (73%) 86 (74%)

Number of metastatic sites

0–2 163 (68%) 71 (61%)

≥3 75 (32%) 46 (39%)

Peritoneal metastases 64 (27%) 45 (38%)

Measurable disease 218 (92%) 106 (91%)

Progression-free interval after previous treatment‡

<6 months 154 (65%) 83 (71%)

≥6 months 81 (34%) 34 (29%)

Previous anti-cancer treatment

Fluoropyrimidine plus platinum 200 (84%) 88 (75%)

Fluoropyrimidine plus other systemic drug 13 (5%) 17 (15%)

Fluoropyrimidine alone 16 (7%) 7 (6%)

Platinum plus other systemic drug 9 (4%) 5 (4%)

Data are median (IQR) or n (%), unless otherwise indicated. ECOG=Eastern Cooperative Oncology Group. NA=not applicable. *As established by self-report. †Assessed according to ECOG guidelines, with 0 as asymptomatic, 1 as restricted in strenuous activity but ambulatory and able to do light work, or 2 as ambulatory and capable of all self-care but unable to work. ‡Defi ned as the time from start of previous treatment to disease progression.

Table 1: Baseline characteristics

Articles

34 www.thelancet.com Vol 383 January 4, 2014

the time from randomisation to the fi rst documentation of an ECOG performance status score of 2 or worse.

This trial is registered with ClinicalTrials.gov, number NCT00917384.

Role of the funding sourceThe sponsor of the study provided the study drug and collaborated with the investigators on protocol, study design, data collection, data analysis, data interpretation, and preparation of this report. CSF prepared the fi rst draft of the report in collaboration with the sponsor and other coauthors. The corresponding author had full access to all the data in the study and all authors had fi nal responsibility for the decision to submit for publication.

ResultsFigure 1 shows the trial profi le. 355 patients were randomly assigned to either the ramucirumab group (n=238) or the placebo group (n=117). Baseline charac-teristics were mostly similar between groups (table 1). Slightly fewer patients in the ramucirumab group had

peritoneal metastases than did those in the placebo group (table 1).

At the time of data cutoff , 278 patients had died: 179 (75%) patients given ramucirumab and 99 (85%) of those given placebo. Treatment with ramucirumab signifi cantly improved overall survival compared with placebo (fi gure 2). Median overall survival was 5·2 months (IQR 2·3–9·9) in patients receiving ramucirumab and 3·8 months (1·7–7·1) in those receiving placebo. Esti-mated rates of 6 month overall survival were 41·8% (35·4–48·1) in the ramucirumab group and 31·6% (23·2–40·2) in the placebo group; rates at 12 months were 17·6% (11·8–24·3) versus 11·8% (6·0–19·7; fi gure 2).

We examined the eff ect of treatment with adjustment for other statistically signifi cant prognostic factors with a stepwise Cox proportional hazards model that identifi ed three signifi cant independent predictors for reduced overall survival: ECOG performance status (≥1 vs 0), location of the primary tumour (gastro-oesophageal junc-tion vs gastric), and presence of peritoneal metastases. After adjustment for these factors, ramucirumab still conferred a signifi cant improvement in overall survival (appendix). When we repeated this multivariable analysis to include the predefi ned stratifi cation factors, the eff ect of ramucirumab on overall survival was unchanged and remained signifi cant (HR for death 0·767, 95% CI 0·598–0·984; p=0·037). The eff ect of ramucirumab on overall survival was consistent across almost all sub-groups, including the prespecifi ed stratifi cation factors (fi gure 3). The eff ect of treatment did not diff er signifi -cantly between men and women (p=0·063).

Treatment with ramucirumab resulted in a 52% reduc-tion in the risk of disease progression or death from any cause (fi gure 2). Median progression-free survival was 2·1 months (IQR 1·3–4·2) in patients receiving ramu-cirumab and 1·3 months (1·1–2·1) in those receiving placebo (fi gure 2). Estimated rates of 12 week progression-free survival were 40·1% (33·6–46·4) in patients in the ramucirumab group and 15·8% (9·7–23·3) in the placebo group (fi gure 2). The eff ect of ramucirumab on pro-gression-free survival was maintained after adjustment for other signifi cant baseline prognostic factors (appendix). Additionally, the progression-free survival benefi t asso-ciated with ramucirumab was consistent across almost all subgroups (fi gure 4). Table 2 shows the numbers of patients with an objective tumour response. The rate of disease control was signifi cantly higher in patients given ramucirumab than in those given placebo (table 2). Add-itionally, the duration of disease control was signifi cantly longer in the ramucirumab group than in the placebo group (median 4·2 months [IQR 2·8–8·1] vs 2·9 months [2·7–4·3]; p=0·036).

At data cutoff , 222 (93%) patients had discontinued ramucirumab and 114 (97%) had discontinued placebo (fi gure 1). Radiological and symptomatic progression was the most common reason for discontinuation followed by adverse events (fi gure 1, appendix). Median duration

Figure 2: Kaplan-Meier estimates of overall survival (A) and progression-free survival (B)HR=hazard ratio.

Number at riskRamucirumab

Placebo

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 26 27 28

238117

15466

9234

4920

177

74

32

01

00

HR (95% CI)=0·776 (0·603–0·998)Log-rank p value (stratified)=0·047

0

20

40

60

80

100

Ove

rall

surv

ival

(%)

A

Ramucirumab (n=238)Placebo (n=117)Censored

Number at riskRamucirumab

Placebo

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

238117

11327

21392

6511

617

302

182

112

52

454

182

41

21

10

10

10

10

00

Time since randomisation (months)

HR (95% CI)=0·483 (0·376–0·620)Log-rank p value (stratified) <0·0001

0

20

40

60

80

100

Prog

ress

ion-

free

surv

ival

(%)

B

Vall d’Hebron University Hospital, UniversitatAutònoma

de Barcelona, Barcelona, Spain (J Tabernero MD)

Correspondence to:Dr Charles S Fuchs,

Gastrointestinal Oncology Division, Dana-Farber Cancer Institute, Boston, MA 02215,

[email protected]

See Online for appendix

Articles

www.thelancet.com Vol 383 January 4, 2014 35

of treatment was 8 weeks (IQR 6–16) in the ramucirumab group and 6 weeks (4–8) in the placebo group. Relative dose intensity of study drug was high in both the ramucirumab and placebo groups (99·6% vs 100%). After discontinuation of study drug, systemic antineoplastic treatment was used more often in the placebo group than in the ramucirumab group (appendix).

223 (94%) patients in the ramucirumab group had adverse events of any grade compared with 101 (88%) in the placebo group (table 3). Hypertension was more

common in the ramucirumab group than in the placebo group (table 3). 134 (57%) patients in the ramucirumab group had grade 3 or higher adverse events compared with 67 (58%) in the placebo group. More patients in the ramucirumab group had grade 3 hypertension than those in the placebo group (table 3); we recorded no events of grade 4 hypertension. Ramucirumab was not associated with increased rates of fatigue, decreased appetite, vomit-ing, anaemia, or other notable toxic eff ects (table 3). Ramu-cirumab was not associated with increased bleeding,

Figure 3: Forest plot for subgroup analyses of overall survivalHRs are based on a stratifi ed proportional-hazards model. The size of the circles is proportional to the size of the subgroup. Horizontal bars represent 95% CIs. HR=hazard ratio. ECOG=Eastern Cooperative Oncology Group. *0 is asymptomatic, 1 is restricted in strenuous activity but ambulatory and able to do light work, 2 is ambulatory and capable of all self-care but unable to work. †1 is North America, Europe, Australia, New Zealand; 2 is South and Central America, India, South Africa, Jordan, Egypt, Saudi Arabia, Lebanon; 3 is Asia.

HR (95% CI)Ramucirumabgroup

Placebo group

Age (years) <65 ≥65 Sex Male FemaleRace White Asian OtherEthnic origin Hispanic Not HispanicECOG score* 0 ≥1Measurable tumour Yes NoPrevious treatment First-line Adjuvant or neo-adjuvantFirst-Line type Doublets TripletsHistological subtype Diffuse Intestinal UnknownNumber of metastatic sites 0–2 ≥3Peritoneal Yes NoProgression-free interval (months) <6 ≥6Weight loss ≥10% <10%Primary tumour Gastric Gastro-oesophageal junctionGeographical region† 1 2 3

Overall

15682

16969

1813918

41197

67171

21820

19939

9992

965290

16375

64174

13265

37201

17959

1655518

238

7146

7938

91179

1998

3186

10611

10314

6336

443538

7146

4572

7428

17100

8532

8029

8

117

0·846 (0·611–1·171)0·722 (0·471–1·106)

0·676 (0·499–0·916)1·431 (0·852–2·405)

0·784 (0·590–1·042)0·636 (0·306–1·321)1·426 (0·448–4·539)

0·563 (0·296–1·072)0·814 (0·617–1·072)

1·075 (0·638–1·810)0·682 (0·508–0·915)

0·808 (0·620–1·054)0·434 (0·143–1·313)

0·793 (0·605–1·041)0·784 (0·372–1·654)

0·756 (0·517–1·105)0·790 (0·510–1·224)

0·560 (0·366–0·857)1·009 (0·583–1·745)0·908 (0·579–1·423)

0·751 (0·541–1·042)0·795 (0·516–1·225)

0·871 (0·556–1·366)0·800 (0·582–1·101)

0·876 (0·631–1·216)0·660 (0·389–1·119)

0·883 (0·425–1·834)0·728 (0·554–0·957)

0·823 (0·608–1·114)0·756 (0·472–1·211)

0·896 (0·667–1·205)0·464 (0·265–0·813)0·694 (0·265–1·818)

0·776 (0·603–0·998)

Favours ramucirumab Favours placebo

10·50·1 1·5 2 3 4 5

Articles

36 www.thelancet.com Vol 383 January 4, 2014

venous thrombosis, proteinuria, gastrointestinal perfor-ation, fi stula formation, or infusion-related reactions (table 3). Grade 3 or higher arterial thromboembolic events were slightly more common in the ramucirumab group than the placebo group (p=0·55; table 3).

Serum samples for detection of anti-ramucirumab anti-bodies were available for 207 (88%) patients receiving ramucirumab and 106 (92%) receiving placebo. Anti-ramucirumab antibodies were detected in six (3%) patients given ramucirumab and one (<1%) patient given

placebo; none of these participants had an infusion-related reaction (data not shown). No patients developed neutralising antibodies to ramucirumab (data not shown).

Five (2%) deaths in the ramucirumab group (caused by myocardial infarction, gastric haemorrhage, intestinal per foration [two events], and pneumonia) two (2%) in the placebo group (intestinal perforation and pulmonary embol ism) were considered to be related to study drug.

230 (97%) patients in the ramucirumab group and 110 (94%) in the placebo group provided baseline

Figure 4: Forest plot for subgroup analyses of progression-free survivalHRs are based on a stratifi ed proportional-hazards model. The size of the circles is proportional to the size of the subgroup. Horizontal bars represent 95% CIs. HR=hazard ratio. ECOG=Eastern Cooperative Oncology Group. *0 is asymptomatic, 1 is restricted in strenuous activity but ambulatory and able to do light work, 2 is ambulatory and capable of all self-care but unable to work. †1 is North America, Europe, Australia, New Zealand; 2 is South and Central America, India, South Africa, Jordan, Egypt, Saudi Arabia, Lebanon; 3 is Asia.

HR (95% CI)Ramucirumabgroup

Placebo group

Age (years) <65 ≥65 Sex Male FemaleRace White Asian OtherEthnic origin Hispanic Not HispanicECOG score* 0 ≥1Measurable tumour Yes NoPrevious treatment First-line Adjuvant or neo-adjuvantFirst-Line type Doublets TripletsHistological subtype Diffuse Intestinal UnknownNumber of metastatic sites 0–2 ≥3Peritoneal Yes NoProgression-free interval (months) <6 ≥6Weight loss ≥10% <10%Primary tumour Gastric Gastro-oesophageal junctionGeographical region† 1 2 3

Overall

15682

16969

1813918

41197

67171

21820

19939

9992

965290

16375

64174

13265

37201

17959

1655518

238

7146

7938

91179

1998

3186

10611

10314

6336

443538

7146

4572

7428

17100

8532

8029

8

117

0·450 (0·327–0·620)0·490 (0·319–0·752)

0·377 (0·277–0·515)0·805 (0·504–1·288)

0·450 (0·338–0·599)0·735 (0·356–1·517)0·651 (0·209–2·030)

0·335 (0·174–0·648)0·502 (0·381–0·661)

0·571 (0·338–0·965)0·423 (0·315–0·569)

0·484 (0·371–0·631)0·277 (0·081–0·948)

0·531 (0·406–0·694)0·431 (0·194–0·958)

0·479 (0·331–0·694)0·531 (0·337–0·837)

0·487 (0·318–0·746)0·456 (0·267–0·778)0·550 (0·347–0·870)

0·460 (0·334–0·635)0·599 (0·386–0·927)

0·695 (0·438–1·103)0·423 (0·308–0·580)

0·535 (0·385–0·743)0·465 (0·282–0·769)

0·421 (0·199–0·893)0·480 (0·367–0·629)

0·513 (0·384–0·685)0·386 (0·229–0·649)

0·474 (0·352–0·639)0·456 (0·270–0·771)0·690 (0·268–1·773)

0·483 (0·376–0·620)

10·50·1 1·5 2 2·5

Favours placeboFavours ramucirumab

Articles

www.thelancet.com Vol 383 January 4, 2014 37

quality-of-life data. At the fi rst 6 week assessment, only 114 (48%) patients receiving ramucirumab and 29 (25%) receiving placebo provided quality-of-life data, mainly because of treatment discontinuation (fi gure 5). Of patients who provided 6 week data, a larger proportion of those in the ramucirumab group reported stable or improved global quality of life than those in the placebo group; however, this diff erence was not signifi cant (p=0·23). Median time to deterioration in ECOG performance status to a score of 2 or worse was 5·1 months (IQR 1·9–16·8) in the ramucirumab group and 2·4 months (1·3 to not reached) in the placebo group (fi gure 6).

DiscussionRamucirumab signifi cantly prolonged overall survival in patients with advanced gastric or gastro-oesophageal junction adenocarcinoma that had progressed after fi rst-line chemotherapy. This drug reduced the risk of death from any cause compared with placebo and reduced risk of disease progression by half. The survival benefi t with ramucirumab was consistent across almost all sub-groups. Although the eff ect on overall survival was attenuated in women, the progression-free survival estimate in women favoured ramucirumab. Moreover, when we addressed potential imbalances in baseline characteristics between treatment groups, the benefi ts in both overall and progression-free survival remained un-changed after adjustment for other prognostic factors.

Ramucirumab was well tolerated in this patient population, with similar rates for most adverse events between the ramucirumab and placebo groups. Hyper-tension—an adverse event associated with most anti-angiogenic treatments20—was more common in the ramucirumab group, although grade 3 hypertension was noted in only a few patients given ramucirumab. Although patient-reported quality-of-life data were scarce, ramucirumab conferred no worse eff ect than placebo; furthermore, patients receiving ramucirumab had a sig-nifi cantly longer time before deterioration of perfor-mance status.

This trial is one of the largest phase 3 trials of second-line treatment of gastric or gastro-oesophageal junction cancer, and the survival benefi t for ramucirumab versus placebo is similar to that reported in two phase 3 trials that compared second-line cytotoxic chemotherapy with

Ramucirumab (n=238)

Placebo (n=117)

p value

Best overall response

Complete response 1 (<1%) 0 ··

Partial response 7 (3%) 3 (3%) ··

Stable disease 108 (45%) 24 (21%) ··

Progressive disease 78 (33%) 63 (54%) ··

Not evaluable 44 (18%) 27 (23%) ··

Objective response 8 (3%) 3 (3%) 0·76

Disease control rate* 116 (49%) 27 (23%) <0·0001

Data are n (%), unless otherwise indicated. *Denotes best response for complete response, partial response, or stable disease.

Table 2: Objective tumour response

Ramucirumab (n=236) Placebo (n=115)

Any event Grade ≥3 Any event Grade ≥3

Fatigue* 84 (36%) 15 (6%) 46 (40%) 11 (10%)

Abdominal pain† 68 (29%) 14 (6%) 32 (28%) 3 (3%)

Decreased appetite 57 (24%) 8 (3%) 26 (23%) 4 (3%)

Vomiting 47 (20%) 6 (3%) 29 (25%) 5 (4%)

Constipation 36 (15%) 1 (<1%) 26 (23%) 3 (3%)

Anaemia‡ 35 (15%) 15 (6%) 17 (15%) 9 (8%)

Dysphagia 25 (11%) 5 (2%) 12 (10%) 5 (4%)

Dyspnoea 22 (9%) 4 (2%) 15 (13%) 7 (6%)

Adverse events of special interest

Hypertension§ 38 (16%) 18 (8%) 9 (8%) 3 (3%)

Bleeding or haemorrhage¶ 30 (13%) 8 (3%) 13 (11%) 3 (3%)

Arterial thromboembolism|| 4 (2%) 3 (1%) 0 0

Venous thromboembolism** 9 (4%) 3 (1%) 8 (7%) 5 (4%)

Proteinuria 7 (3%) 1 (<1%) 3 (3%) 0

Gastrointestinal perforation 2 (<1%) 2 (<1%) 1 (<1%) 1 (<1%)

Fistula formation 1 (<1%) 1 (<1%) 1 (<1%) 1 (<1%)

Infusion-related reaction 1 (<1%) 0 2 (2%) 0

Cardiac failure 1 (<1%) 0 0 0

Data are n (%), unless otherwise indicated. *Includes asthenia. †Includes upper or lower abdominal pain and hepatic pain. ‡Includes decreased haematocrit and red blood-cell count. §Includes increased blood pressure. ¶Includes epistaxis, gastric haemorrhage, gastrointestinal haemorrhage, gingival bleeding, haematemesis, haematoma, haematuria, haemoptysis, haemorrhage, haemorrhoidal haemorrhage, melaena, nail-bed bleeding, petechiae, rectal haemorrhage, and upper gastrointestinal haemorrhage. ||Includes angina pectoris, cardiac arrest, cerebral ischaemia, cerebrovascular accident, myocardial infarction, and myocardial ischemia. **Includes pulmonary embolism, deep vein thrombosis, thrombosis, and venous thrombosis in a limb.

Table 3: Adverse events, according to grade

Figure 5: Patient-reported global quality of life 6 weeks after start of treatment initiationBaseline quality-of-life data were provided by 230 (97%) patients in the ramucirumab group and 110 (94%) in the placebo group. Of the patients who did not provide data either at baseline or before the 6 week assessment, 113 (88%) of 128 in the ramucirumab group and 82 (90%) of 91 in the placebo group either discontinued treatment before the 6 week assessment (111 [98%] vs 80 [98%]) or were never treated (two [2%] in each group). The main reason for treatment discontinuation was disease progression (81 [73%] vs 61 [76%]). Patients were classifi ed as improved or worsened if change from baseline was 10 points or more (on the basis of a standardised 100-point scale);19 we considered changes of less than 10 points as stable.

Ramucirumab (n=238) Placebo (n=117)

128 (54%)

29 (12%)

10 (9%)

5 (4%)

58 (24%)

23 (10%) 11 (9%)

91 (78%)

No dataWorsenedStableImproved

Articles

38 www.thelancet.com Vol 383 January 4, 2014

best supportive care (median 5·3 months vs 3·8 months21 and 5·2 months vs 3·6 months22). Rates of grade 3 or higher adverse events associated with ramucirumab in our trial compare favourably to those associated with second-line chemotherapy reported in other phase 3 trials of this malignant disease.21–23 A previous placebo-controlled phase 3 trial24 assessed the addition of bevaci-zumab, a monoclonal antibody targeting VEGF-A, with fi rst-line chemotherapy in patients with previously untreated gastric adenocarcinoma and showed a signifi -cant improvement in progression-free survival. Although a trend in overall survival favoured bevacizumab, this diff erence was not signifi cant. In subset analyses, a survival benefi t for bevacizumab was limited to non-Asian patients. In our trial, the survival benefi t associated with ramucirumab was similar between Asian patients and those from America, Europe, and Australia, although relatively few Asian patients were enrolled.

In view of our fi ndings, ramucirumab could be an important treatment option in patients with advanced gastric or gastro-oesophageal junction adenocarcinoma pro gressing after fi rst-line chemotherapy. Additionally, this novel drug has a mechanism of action and toxic-eff ect profi le that is distinctly diff erent and non-overlapping with standard chemotherapy for oesophago gastric adeno-carcinoma (panel). Our results validate the role of VEGFR-2 signalling as an important therapeutic target in advanced gastric and gastro-oesophageal junction adenocarcinoma. Future analyses are needed to identify potential predictive biomarkers for ramucirumab. A recently completed randomised clinical trial (ClinicalTrials.gov, number NCT01170663) has reported a survival benefi t when ramucirumab is added to second-line paclitaxel in patients with advanced oesophagogastric adenocarcinoma, while an ongoing randomised clinical trial (NCT01170663) is assessing a combination of ramucirumab with fi rst-line chemo therapy in this patient population.

ContributorsJDS, JRZ, JTa, and CSF conceived and designed the study. CSF, JTa, JYC,

FD, RP, CG, HS, LVdS, GA, DRF, BM, MT, ET, IC, WC, CS, JP, and JTo

collected the data. JDS, MK, AML, LG, YH, and CSF analysed and

interpreted the data. CSF wrote the fi rst draft. All authors drafted,

reviewed, and completed and approved the manuscript and the decision

to submit for publication.

Confl icts of interestCSF has been a consultant for Amgen, Pfi zer, Sanofi Aventis, Roche,

Genentech, Metamark Genetics, Momenta Pharm, Infi nity Pharm, Bayer,

and Bristol-Myers Squibb, and is supported by the US National Cancer

Institute (grant number P50CA127003). JTo has received payment support

for travel to meetings for the study. Istituti Ospitalieri di Cremona

(Cremona, Italy) has received funding for this clinical study done by RP

from ImClone and Eli Lilly. RP has been a board member with Novartis,

Roche, Pfi zer, consultancy with Bayer, and Astellas; and has a pending

research grant with Amgen. B P Poddar Hospital and Medical Research

(Kolkata, India) has received funding for this clinical study done by CG

from ImClone LLC and Eli Lilly. DRF has received honorarium, support

for travel to meetings for this study, has been a consultant for, and

received payments for lectures including speakers bureaus from ImClone

and New Cross Hospital (West Midlands, UK); and has received grants or

has a grant pending from ImClone. BM has been a consultant for Roche,

Astellas, and GlaxoSmithKline; received payment for lectures including

Figure 6: Time to deterioration in ECOG performance status to a score of 2 or worseHR=hazard ratio. ECOG=Eastern Cooperative Oncology Group.

Number at riskRamucirumab

Placebo

0 3 6 9Time to ECOG performance status ≥2 (months)

12 15 18

238117

8513

324

132

21

10

00

HR (95% CI)=0·586 (0·414–0·829)Log-rank p value (stratified)=0·002

0

20

40

60

80

100

No

dete

riora

tion

in E

COG

perfo

rman

ce

stat

us (%

)

Ramucirumab (n=238)Placebo (n=117)Censored

l ll llll lllll lllll l lllllllll lll llll

ll ll l lll lllll llll ll ll ll llll l lllllllllll ll llll llllll

lllllll

lll

ll l

lllll lll l ll llll

lll

lllll

llllllll

Panel: Research in context

Systematic reviewWe searched PubMed and the abstracts of major oncology congresses (American Society of Clinical Oncology [ASCO] and ASCO Gastrointestinal Symposium, European Society for Medical Oncology [ESMO], World Congress on Gastrointestinal Cancer, International Society of Gastrointestinal Oncology Conference, American Association for Cancer Research-National Cancer Institute-European Organisation for Research and Treatment of Cancer Congress, ESMO annual meeting, and European Multidisciplinary Cancer Congress) from Jan 1, 2009, to Dec 31, 2012. We used MeSH and full-text search terms for gastric and gastro-oesophageal junction cancer and molecular-targeted treatments, limiting our results to articles in English published in the past 4 years. For PubMed, the search was (“molecular targeted therapy” OR (“molecular” AND “targeted”) AND (“therapy” OR “therapies”) AND (“gastric neoplasms” OR “gastric cancer”) OR (“gastric” AND “cancer”) (“gastroesophageal neoplasms” OR “gastroesophageal cancer”) OR (“gastroesophageal” AND “cancer”) AND (“2009/01/01”: “2012/12/31”). For conferences, the search was “metastatic gastric cancer” or “advanced gastric cancer”, manually limited to abstracts of targeted treatments or phase 3 studies. We identifi ed several potential targeted drugs (either monoclonal antibodies or small-molecule tyrosine-kinase inhibitors) that are being investigated either in synergy with, or in place of, established treatments, including inhibitors of growth factors and their receptors (ie, vascular endothelial growth factor [VEGF], epidermal growth factor receptor, insulin-like growth factor, human epidermal growth factor receptor-2, c-MET), MEK inhibitors, and drugs targeting the Hedgehog pathway.25–31 We used information from the abstracts and ClinicalTrials.gov to identify the latest stage of clinical developments of these drugs in gastric and gastro-oesophageal junction cancer. We have limited our discussion to the drugs we believe are most promising and relevant to the patient population, on the basis of fi ndings for effi cacy from clinical trials.

InterpretationVarious signalling processes have been implicated in the development and progression of gastric and gastro-oesophageal junction adenocarcinoma, and experience with the monoclonal antibodies trastuzumab and bevacizumab show that these pathways might be valid treatment targets. Our fi ndings show that in patients with progressive gastric or gastro-oesophageal adenocarcinoma after standard fi rst-line treatment, ramucirumab can signifi cantly prolong survival compared with placebo, providing evidence for the role of a VEGF-receptor-targeted treatment and off ering an important new treatment option in this patient population.

Articles

www.thelancet.com Vol 383 January 4, 2014 39

speakers bureau from Roche, Novartis, Pfi zer, and GlaxoSmithKline; and

for travel accommodations from Roche, Amgen, Novartis, and Merck. JZR

has received honorarium and support for travel to meetings for study

purposes from ImClone; has been a member of advisory boards for Bayer

and Roche; has received research grants from Bayer and Bristol-Myers

Squibb; and has received payment for lectures including service on

speakers bureaus from Roche. Royal Marsden Hospital (London and

Surrey, England) has received research support for IC undertaking this

study from Eli Lilly, and an honorarium from ImClone and Eli Lilly; and

IC has been a consultant for Roche, Merck-Serono, Novartis,

Sanofi -Aventis, and Bristol-Myers Squibb; has received payment for

lectures including speakers bureau from Roche and Sanofi -Aventis;

payment for educational presentations from Roche; and payment for

travel accommodations and meeting expenses unrelated to this study

from Roche and Sanofi -Aventis. IC acknowledges National Health Service

funding to the National Institute for Health Research Biomedical

Research Centre. Regional Cancer Centre (Kerala, India) has received

grant support to CS from Quintiles and ImClone. JDS, MK, AML, LG,

and YH are employees of and have stock ownership of ImClone and

Eli Lilly. JTa has been a consultant for Amgen, Boehringer, Bristol-Myers

Squibb, Genentech, ImClone, Eli Lilly, Merck KGaA, Millennium,

Novartis, Onyx, Pfi zer, Roche, Sanofi -Aventis, and Celgene.

AcknowledgmentsWe thank the participating patients and staff at each of the study centres,

and members of the independent data monitoring committee—

Paul Catalano (Chair), John S Macdonald, Franco Muggia,

Daniel G Haller, and M Wasif Saif.

References1 Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global

cancer statistics. CA Cancer J Clin 2011; 61: 69–90.

2 Wagner AD, Unverzagt S, Grothe W, et al. Chemotherapy for advanced gastric cancer. Cochrane Database Syst Rev 2010; 3: CD004064.

3 Jung YD, Mansfi eld PF, Akagi M, et al. Eff ects of combination anti-vascular endothelial growth factor receptor and anti-epidermal growth factor receptor therapies on the growth of gastric cancer in a nude mouse model. Eur J Cancer 2002; 38: 1133–40.

4 Kim SE, Shim KN, Jung SA, Yoo K, Lee JH. The clinicopathological signifi cance of tissue levels of hypoxia-inducible factor-1alpha and vascular endothelial growth factor in gastric cancer. Gut Liver 2009; 3: 88–94.

5 Lieto E, Ferraraccio F, Orditura M, et al. Expression of vascular endothelial growth factor (VEGF) and epidermal growth factor receptor (EGFR) is an independent prognostic indicator of worse outcome in gastric cancer patients. Ann Surg Oncol 2008; 15: 69–79.

6 Song ZJ, Gong P, Wu YE. Relationship between the expression of iNOS, VEGF, tumor angiogenesis and gastric cancer. World J Gastroenterol 2002; 8: 591–95.

7 Cabuk D, Basaran G, Celikel C, et al. Vascular endothelial growth factor, hypoxia-inducible factor 1 alpha and CD34 expressions in early-stage gastric tumors: relationship with pathological factors and prognostic impact on survival. Oncology 2007; 72: 111–17.

8 Juttner S, Wissmann C, Jons T, et al. Vascular endothelial growth factor-D and its receptor VEGFR-3: two novel independent prognostic markers in gastric adenocarcinoma. J Clin Oncol 2006; 24: 228–40.

9 Murukesh N, Dive C, Jayson GC. Biomarkers of angiogenesis and their role in the development of VEGF inhibitors. Br J Cancer 2010; 102: 8–18.

10 Oh SY, Kwon HC, Kim SH, et al. Clinicopathologic signifi cance of HIF-1alpha, p53, and VEGF expression and preoperative serum VEGF level in gastric cancer. BMC Cancer 2008; 8: 123.

11 Okines AF, Reynolds AR, Cunningham D. Targeting angiogenesis in esophagogastric adenocarcinoma. Oncologist 2011; 16: 844–58.

12 Suzuki S, Dobashi Y, Hatakeyama Y, et al. Clinicopathological signifi cance of platelet-derived growth factor (PDGF)-B and vascular endothelial growth factor-A expression, PDGF receptor-β phosphorylation, and microvessel density in gastric cancer. BMC Cancer 2010; 10: 659.

13 Zhao HC, Qin R, Chen XX, et al. Microvessel density is a prognostic marker of human gastric cancer. World J Gastroenterol 2006; 12: 7598–603.

14 Spratlin JL, Cohen RB, Eadens M, et al. Phase I pharmacologic and biologic study of ramucirumab (IMC-1121B), a fully human immunoglobulin G1 monoclonal antibody targeting the vascular endothelial growth factor receptor-2. J Clin Oncol 2010; 28: 780–87.

15 Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000; 92: 205–16.

16 Cancer Therapy Evaluation Program. Common terminology criteria for adverse events v4.02. 2009. http://ctep.cancer.gov/protocolDevelopment/electronic_applications/ctc.htm#ctc_40 (accessed Sept 15, 2009).

17 Aaronson NK, Ahmedzai S, Bergman B, et al. The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 1993; 85: 365–76.

18 DeMets DL, Ware JH. Asymmetric group sequential boundaries for monitoring clinical trials. Biometrika 1982; 69: 661–63.

19 Osoba D, Rodrigues G, Myles J, Zee B, Pater J. Interpreting the signifi cance of changes in health-related quality-of-life scores. J Clin Oncol 1998; 16: 139–44.

20 Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000; 407: 249–57.

21 Kang JH, Lee SI, Lim H, et al. Salvage chemotherapy for pretreated gastric cancer: a randomized phase III trial comparing chemotherapy plus best supportive care with best supportive care alone. J Clin Oncol 2012; 30: 1513–18.

22 Ford H, Marshall A, Wadsley J, et al. Cougar-02: a randomized phase III study of docetaxel versus active symptom control in advanced esophagogastric adenocarcinoma. 10th Annual Gastrointestinal Cancer Symposium; San Francisco, CA; Jan 24–26, 2013. LBA4.

23 Thuss-Patience PC, Kretzschmar A, Bichev D, et al. Survival advantage for irinotecan versus best supportive care as second-line chemotherapy in gastric cancer—a randomised phase III study of the Arbeitsgemeinschaft Internistische Onkologie (AIO). Eur J Cancer 2011; 47: 2306–14.

24 Ohtsu A, Shah MA, Van Cutsem E, et al. Bevacizumab in combination with chemotherapy as fi rst-line therapy in advanced gastric cancer: a randomized, double-blind, placebo-controlled phase III study. J Clin Oncol 2011; 29: 3968–76.

25 Bang YJ, Van Cutsem E, Feyereislova A, et al. Trastuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2-positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet 2010; 376: 687–97.

26 Blumenschein GR Jr, Mills GB, Gonzalez-Angulo AM. Targeting the hepatocyte growth factor-cMET axis in cancer therapy. J Clin Oncol 2012; 30: 3287–96.

27 Li H, Adachi Y, Yamamoto H, et al. Insulin-like growth factor-I receptor blockade reduces tumor angiogenesis and enhances the eff ects of bevacizumab for a human gastric cancer cell line, MKN45. Cancer 2011; 117: 3135–47.

28 Paterson AL, Shannon NB, Lao-Sirieix P, et al. A systematic approach to therapeutic target selection in oesophago-gastric cancer. Gut 2013; 62: 1415–24.

29 Seto M, Ohta M, Asaoka Y, et al. Regulation of the hedgehog signaling by the mitogen-activated protein kinase cascade in gastric cancer. Mol Carcinog 2009; 48: 703–12.

30 Song Z, Yue W, Wei B, et al. Sonic hedgehog pathway is essential for maintenance of cancer stem-like cells in human gastric cancer. PloS One 2011; 6: e17687.

31 Waddell TS, Chau I, Barbachano Y, et al. A randomized, multicenter trial of epirubicin, oxaliplatin, and capecitabine (EOC) plus panitumumab in advanced esophagogastric cancer (REAL3). J Clin Oncol 2012; 30 (suppl): LBA4000.