623P - Exploratory Phase Ii Trial Assessing Relevance of in Vitro Chemosensitivity Test As a Predictor of Survival in Gastric Cancer Undergoing Adjuvant Chemotherapy (Jaccro-Gc04)

Abstract

ABSTRACT Aim: Postoperative S-1 (a 5-FU derivative) monotherapy has become the standard for Stage II/III gastric cancer in Japan following a phase III trial designated ACTS-GC, which implies that S-1 is also given to the patients whose tumors may not be sensitive to 5-FU. A nationwide clinical trial, JACCRO-GC04, was conducted to explore whether individualization of the treatment is possible through chemosensitivity testing. Methods: Using surgically resected specimens, the collagen gel droplet-embedded drug sensitivity test was performed at 5-FU concentrations (conc) of 0.2, 0.4, 1.0 and 2.0 µg/ml. Patients who underwent R0 resection for histologically proven stage II ∼ IV gastric cancer and for whom the chemosensitivity test was successful were eligible. All eligible patients were to receive adjuvant treatment with S-1 for 1 year at a dose of 80 mg/m2 bid for 4 weeks followed by 2 weeks rest. Optimal cut-off values at each 5-FU conc to decide on the chemosensitivity were to be sought within the study. Ultimate objective of the study was to prove the hypothesis that a significant difference in 3-year relapse-free survival (RFS), the primary endpoint, is observed between the responder (Res) and non-responder (Non). Results: Between 2005 and 2013, 311 patients were enrolled, of which 206 were eligible. Hazard ratios (HR) for relapse in Res as compared with Non were calculated at various cut-off values of in vitro growth inhibition rates from 10% to 90% with an increment of 10%. The appropriate cut-off values could be defined based on HR Conclusions: The appropriate cut-off values classifying patients as Res or Non were found. The low HR for relapse indicates the in vitro drug sensitivity predicts efficacy of the postoperative S-1 treatment. Disclosure: M. Fujii: Advisory board; Taiho Honarium; Taiho Travel Grant; JACCRO. All other authors have declared no conflicts of interest.