BackgroundWe conducted a multicenter randomized study of adjuvant S-1 administration schedules for surgically treated pathological stage IB-IIIA non-small cell lung cancer patients.MethodsPatients receiving curative surgical resection were centrally randomized to arm A (4 weeks of oral S-1 and a 2-week rest over 12 months) or arm B (2 weeks of S-1 and a 1-week rest over 12 months). The primary endpoints were completion of the scheduled adjuvant chemotherapy over 12 months, and the secondary endpoints were relative total administration dose, toxicity, and 3-year disease-free survival.ResultsFrom April 2005 to January 2012, 80 patients were enrolled, of whom 78 patients were eligible and assessable. The planned S-1 administration over 12 months was accomplished to 28 patients in 38 arm A patients (73.7%) and to 18 patients in 40 arm B patients (45.0%, p = 0.01). The average relative dose intensity was 77.2% for arm A and 58.4% for arm B (p = 0.01). Drug-related grade 3 adverse events were recorded for 11% of arm A and 5% of arm B (p = 0.43). Grade 1–3 elevation of bilirubin, alkaline phosphatase, aspartate aminotransferase, and alanine transaminase were more frequently recorded in arm A than in arm B. The 3-year disease-free survival rate was 79.0% for arm A and 79.3% for arm B (p = 0.94).ConclusionsThe superiority of feasibility of the shorter schedule was not recognized in the present study. The conventional schedule showed higher completion rates over 12 months (p = 0.01) and relative dose intensity of S-1 (p = 0.01). Toxicity showed no significant difference among the shorter schedule and the conventional schedule, except for grade 1–3 elevation of bilirubin.Trial registrationThis randomized multicenter study was retrospectively registered with the UMIN-CTR (UMIN000016086, registration date December 30, 2014).