The pharmacokinetics of irinotecan vary markedly between individuals. This study sought to compare tailored irinotecan and S-1 therapy with S-1 monotherapy for the treatment of patients with advanced/recurrent gastric cancer. Patients with advanced/recurrent gastric cancer were randomized to receive tailored irinotecan and S-1 (arm A) therapy or S-1 therapy alone (arm B). Arm A received S-1 (80-120 mg/m(2)/day) for 14 days, with irinotecan on days 1 and 15. The initial irinotecan dose of 75 mg/m(2) (level 0) was adjusted for toxicity during an earlier course. In arm B, S-1 (80-120 mg/day) was administered alone for 28 days, followed by 14 days without therapy. Ninety-five patients were randomized (48 patients to arm A and 47 patients to arm B). The response rate of the primary tumor (Japanese criteria) was 25.0% in arm A (12 of 48 patients) and 14.9% in arm B (seven of 47 patients), whereas the response rates according to Response Evaluation Criteria In Solid Tumors were 27.8% (10 of 36) versus 21.9% (seven of 32). Hematological toxicity, anorexia, and diarrhea were significantly more common in arm A, but both arms had similar grades 3-4 toxicities. These findings suggest the usefulness of tailored irinotecan and S-1 therapy for gastric cancer.