Capecitabine versus S-1 as adjuvant chemotherapy for patients with stage III colorectal cancer (JCOG0910): an open-label, non-inferiority, randomised, phase 3, multicentre trial.

BACKGROUND: Adjuvant chemotherapy with oral fluoropyrimidine alone after D3/D2 lymph node dissection improves disease-free survival and overall survival in patients with stage III colon cancer. Adjuvant S-1 has been shown to be non-inferior to uracil and tegafur plus leucovorin in terms of disease-free survival. This study aims to confirm the non-inferiority of S-1 compared with capecitabine as adjuvant treatment in patients with stage III colorectal cancer.

METHODS: This study was an open-label, non-inferiority, randomised, phase 3, multicentre trial done in 56 Japanese centres to assess the non-inferiority of S-1 to capecitabine as adjuvant chemotherapy. Eligible patients were aged 20-80 years with stage III colorectal adenocarcinoma, as defined by the presence of an inferior margin of the primary tumour above the peritoneal reflection; R0 resection; and colectomy with D3 or D2 lymph node dissection. Patients were randomly assigned (1:1) to receive eight courses of capecitabine (1250 mg/m2 orally twice daily, days 1-14, every 21 days) or four courses of S-1 (40 mg/m2 orally twice daily, days 1-28, every 42 days). Randomisation was done via phone call, fax, or web-based systems to the Japan Clinical Oncology Group Data Center and used a minimisation method with a random component adjusted by institution, tumour location (colon vs rectosigmoid and upper rectum), number of positive lymph node metastases (≤3 vs ≥4), and surgical technique (conventional vs non-touch isolation). The primary endpoint was disease-free survival with a non-inferiority margin for the hazard ratio (HR) set at 1·24, analysed by intention to treat. This trial was registered with UMIN Clinical Trial Registry, number UMIN000003272.

FINDINGS: Between March 1, 2010, and Aug 23, 2013, 1564 patients were randomly assigned to capecitabine (n=782) or S-1 (n=782), all of whom were included in the efficacy analysis; 777 patients in the capecitabine group and 768 in the S-1 group were included in the safety analysis. At the prespecified second interim analysis after final accrual, 258 (48%) of 535 required events were reported, and the Data and Safety Monitoring Committee recommended early publication because S-1 could not show non-inferiority compared with capecitabine for disease-free survival. With a median follow-up of 23·7 months (IQR 14·1-35·2), 3-year disease-free survival was 82·0% (95% CI 78·5-85·0) for the capecitabine group and 77·9% (74·1-81·1) for the S-1 group (HR 1·23, 99·05% CI 0·89-1·70; one-sided pnon-inferiority =0·46). The most frequent grade 3 or higher adverse events in the capecitabine group were hand-foot skin reactions (123 [16%] of 777 patients), and in the S-1 group were diarrhoea (64 [8%] of 768 patients) and neutropenia (61 [8%]). There was one (<1%) treatment-related death in each group.

INTERPRETATION: Adjuvant capecitabine remains one of the standard treatments for stage III colorectal cancer in Japan; S-1 is not recommended.

FUNDING: National Cancer Center and Ministry of Health, Labour and Welfare of Japan.