Background: Addition of oral leucovorin (LV) to S-1 significantly improved progression-free survival (PFS) in a previous randomized phase II trial in Japanese patients (pts) with gemcitabine (GEM)-refractory advanced pancreatic cancer (PC). TAS-118 is an oral drug containing S-1 and LV. This phase III trial conducted in Japan and Korea compared overall survival (OS) between GEM-refractory advanced PC pts treated with TAS-118 and S-1. Methods: GEM-refractory PC pts were randomized in a 1:1 ratio to receive TAS-118 (S-1; 40-60 mg and LV; 25 mg bid for 1w, q2w) or S-1 (S-1; 40-60 mg bid for 4w, q6w). The primary endpoint was OS. The secondary endpoints included PFS, overall response rate, disease control rate, duration of response, and safety. Results: Five hundred and eighty-six pts were eligible for efficacy assessment (TAS-118: n=296 and S-1: n=290). Baseline characteristics were well balanced between the treatment arms. TAS-118 did not result in a statistically significant improvement in OS compared with that achieved with S-1 (median OS, 7.6 months vs. 7.9 months; hazard ratio [HR], 0.98; 95% CI, 0.82 to 1.16; P=0.756). However, it significantly improved PFS compared to that achieved with S-1 (median PFS, 3.9 months vs. 2.8 months; HR, 0.80; 95% CI, 0.67 to 0.95; P=0.009). Pre-planned subgroup analysis of OS showed significant interactions between the treatment effects and pancreatic resection (P=0.025), and between the treatment effects and country (P=0.004). Grade 3/4 drug-related adverse events (≥5% incidences) in TAS-118 and S-1 arms included diarrhea (7.0% vs. 7.3%), anorexia (6.7% vs. 5.0%), stomatitis (6.7% vs. 0.7%), and anemia (3.3% vs. 5.0%). Conclusions: The primary endpoint was not met. Further, the interactions between the treatment effects and pancreatic resection, and between the treatment effects and country, might affect the results. Clinical trial information: 132172.