BACKGROUNDOutcomes are poor in patients with HER2-negative, advanced gastric or gastro-oesophageal junction adenocarcinomas. In this study, we investigated efficacy and safety of the first-in-class, afucosylated, humanised IgG1 anti-fibroblast growth factor receptor 2 isoform IIb (FGFR2b) monoclonal antibody bemarituzumab with modified 5-fluorouracil, leucovorin, and oxaliplatin (mFOLFOX6) in patients with FGFR2b-selected gastric or gastro-oesophageal junction adenocarcinoma.METHODSIn the randomised, double-blind, placebo-controlled phase 2 trial (FIGHT), patients aged 18 years and older with HER2 non-positive, FGFR2b-selected gastric or gastro-oesophageal junction adenocarcinoma, and an Eastern Cooperative Oncology Group performance status of 0-1 were recruited from 144 clinical sites across 17 countries. Patients with previous treatment with any selective inhibitor of the FGF-FGFR pathway were excluded. Eligible patients were randomly assigned (1:1), using permuted-block randomisation (block size of four) and a central interactive voice-web-based response system, stratified by geographical region, previous treatment with curative intent, and administration of mFOLFOX6 while being screened for FGFR2b status, to either bemarituzumab (15 mg/kg of bodyweight) or matched placebo intravenously every 2 weeks. All patients also received mFOLFOX6 (oxaliplatin 85 mg/m2, leucovorin 400 mg/m2, and 5-fluorouracil as a 400 mg/m2 bolus followed by 2400 mg/m2 over approximately 46 h) intravenously every 2 weeks. Patients were given treatment until disease progression (defined by Response Evaluation Criteria in Solid Tumours [RECIST] version 1.1), unacceptable toxicity, withdrawal of consent, or death. The primary endpoint was progression-free survival in the intention-to-treat population (defined as all patients randomly assigned to treatment). Safety was assessed in all patients who received at least one dose of assigned treatment. This study is registered with ClinicalTrials.gov, NCT03694522, and is now complete.FINDINGSBetween Nov 14, 2017, and May 8, 2020, 910 patients were screened and 155 were randomly assigned to the bemarituzumab (n=77) or placebo group (n=78). Median age was 60·0 years (IQR 51·0-67·0), 44 (28%) participants were women, 111 (72%) were men, 89 (57%) were Asian, and 61 (39%) were White. At the time of the primary analysis and at a median follow-up of 10·9 months (IQR 6·3-14·2), median progression-free survival was 9·5 months (95% CI 7·3-12·9) in the bemarituzumab group and 7·4 months (5·8-8·4) in the placebo group (hazard ratio [HR] 0·68 [95% CI 0·44-1·04; p=0·073). Common grade 3 or worse adverse events were decreased neutrophil count (23 [30%] of 76 in the bemarituzumab group vs 27 [35%] of 77 in the placebo group), cornea disorder (18 [24%] vs none), neutropenia (ten [13%] vs seven [9%]), stomatitis (seven [9%] vs one [1%]), and anaemia (six [8%] vs ten [13%]). Serious treatment-emergent adverse events were reported in 24 (32%) patients in the bemarituzumab group and 28 (36%) in the placebo group. Serious mFOLFOX6 treatment-related adverse events occurred in nine (12%) patients in the bemarituzumab group and in 15 (19%) patients in the placebo group. All-grade corneal events (adverse events of special interest) occurred in 51 (67%) patients in the bemarituzumab group and eight (10%) in the placebo group; grade 3 corneal events were reported only in 18 (24%) patients in the bemarituzumab group. Treatment-related deaths occurred in three patients in the bemarituzumab group (two due to sepsis, one due to pneumonia) and none in the placebo group.INTERPRETATIONIn this exploratory phase 2 study, despite no statistically significant improvement in progression-free survival, treatment with bemarituzumab showed promising clinical efficacy. Confirmatory phase 3 trials of bemarituzumab plus mFOLFOX6 powered to demonstrate statistical significance are being investigated in patients with previously untreated, FGFR2b-overexpressing, advanced gastric or gastro-oesophageal junction adenocarcinoma.FUNDINGFive Prime Therapeutics.