7518 Background: We conducted a randomized phase II trial for patients with resected stage II-IIIA NSCLC comparing postoperative oral S-1 (80 mg/m2/day for consecutive 2 weeks q3w for 1 year) (S) (N=100) or cisplatin (CDDP) (60 mg/m2 day1) plus oral S-1, (80 mg/m2/day for 2 weeks) q3w for 4 cycles (PS)(N=100). We reported that the disease-free survival rate at 2 years (DFS@2) (95% confidence interval: CI), a primary endpoint, was 66 (55-74) % for S and 58 (48-67)% for PS. Here, we report the preliminary results of preplanned biomarker analysis, a co-primary endpoint, to identify molecules whose expression is significantly associated with patient outcome. Methods: cDNA extracted from macro-dissected formalin-fixed paraffin-embedded specimens were available for 197/200 patients. Thirty-one genes including those whose expressions have been potentially associated withCDDP (e.g. ERCC1, XRCC1, BRCA1, GSTpi, HMG1, TBP) or fluorouracil (FU) sensitivity (TS, DHFR, DPD, UMPS, UPP1) were measured by QGE analysis (MassArray, Sequenom, CA). Expression of each gene was dichotomized according to its median value of expression. Results: The only gene with interaction P<0.05 after adjustment with sex, histology, age, stage was UMPS (uridine monophosphate synthase) (P=0.0348). UMPS catalyzes the conversion of fluoro UMP from 5FU and thus high UMPS expression is implicated to enhance 5FU effect. DFS@2 % (95% CI) for UMPS high/S, UMPS high/CS, UMPS low/S, UMPS low/CS were 69 (54-80), 53 (37-66), 64 (49-76), 61 (46-73)%, respectively. However, molecules such as ERCC1 and GSTpi whose expression have been previously associated with CDDP sensitivity did not emerge as predictive markers (P=0.7908, 0.6406, respectively). Conclusions: UMPS expression may define patient subset that would benefit from postoperative S-1 therapy. Clinical trial information: 000001658.