A randomized, controlled trial has begun to compare neoadjuvant chemotherapy using gemcitabine and S-1 with upfront surgery for patients planned resection of pancreatic cancer. Patients were enrolled after the diagnosis of resectable or borderline resectable by portal vein involvement pancreatic cancer with histological confirmation. They were randomly assigned to either neoadjuvant chemotherapy or upfront surgery. Adjuvant chemotherapy using S-1 was administered for 6 months to patients with curative resection who fully recovered within 10 weeks after surgery in both arms. The primary endpoint is overall survival; secondary endpoints include adverse events, resection rate, recurrence-free survival, residual tumor status, nodal metastases and tumor marker kinetics. The target sample size was required to be at least 163 (alpha-error 0.05; power 0.8) in both arms. A total of 360 patients were required after considering ineligible cases. This trial began in January 2013 and was registered with the UMIN Clinical Trials Registry (UMIN000009634).