Results of surgery alone for pancreatic cancer are disappointing. We retrospectively evaluated the efficacy and tolerability of neoadjuvant chemotherapy (NAC) with gemcitabine and oral S-1 in patients with potentially resectable pancreatic cancer. A total of 34 patients with pancreatic ductal adenocarcinoma, radiologically diagnosed preoperatively as having potentially resectable tumors, underwent pancreatic resection with lymphadenectomy at Kanazawa University Hospital. NAC was administered to 13 patients (NAC group). The remaining 21 patients were surgically treated without preoperative chemotherapy (control group). Surgical results were compared between these two groups, with follow-up for at least 24 months. No statistically significant differences were found in the clinicopathological background data (tumor location, age, gender, lymph node metastases, tumor stage and tumor size) between the NAC and control groups. Following preoperative chemotherapy, no patients were judged to be unable to undergo laparotomy, i.e., neither distant metastasis nor tumor progression was observed. Radiologically, all 13 NAC group patients had stable disease, whereas, histopathologically, all tumor specimens showed evidence of tumor cells. The treatment effect was judged by Evans grading to be grade IIa in 11 patients and grade IIb in 2 patients. Toxicity was evaluated in 11 patients. Grade III side effects were regarded as hematological toxicity, i.e., leucopenia (7.7%) and thrombocytopenia (15.4%). Moreover, the incidence of perioperative complications did not differ significantly between the NAC and control groups. The one- and three-year overall survival rates of the NAC group with pancreatic head cancer were 88.9 and 55.6%, respectively, superior to 88.9 and 29.6% in the control group (p=0.055). Therefore, NAC with gemcitabine and S-1 is well tolerated and potentially effective against pancreatic head cancer. A phase I study of NAC with gemcitabine and S-1 is under way in patients with resectable pancreatic cancer.