Background S-1 plus weekly split-dose cisplatin demonstrated promising results in previous phase I and II studies for advanced gastric cancer (AGC) patients. Methods In this randomized phase II study, the efficacy and safety of S-1 plus weekly split-dose cisplatin (SWP, S-1 daily oral dose of 80-120 mg according to body surface area on days 1-14, and cisplatin 20 mg/m 2 i.v. on days 1 and 8 every 3 weeks) were compared with those of S-1 plus standard-dose cisplatin (SP) as first-line chemotherapy for AGC patients. The primary endpoint was 1-year survival rate.Results Patients were randomized into two groups: 18 in the SWP arm and 19 in the SP arm. This trial was terminated early because of low patient enrollment. The 1-year survival rate was 61 % [95 % confidence interval (CI), 36-86 %] and 53 % (95 % CI, 30-75 %) in the SWP and SP arms, respectively. However, the median survival time was 12.3 months (9.9-14.6 months) and 15.7 months (4.0-27.4 months), respectively (P = 0.064). Progressionfree survival was significantly shorter in the SWP arm than in the SP arm (P = 0.047). Toxicity tended to be milder in the SWP arm than in the SP arm. For approximately 40 % of patients in the SWP arm, cisplatin was omitted on day 8 and treatment delayed because of prolonged myelosuppression. Conclusions No clear benefits of adding cisplatin to S-1 in the SWP arm were demonstrated in this study. At this point, split-dose cisplatin combined with S-1 cannot be recommended for use in clinical practice.