It has been reported that the response rate to TS-1 of advanced recurrent gastric cancer was the highest rate (46.5%) of effectiveness among anti-cancer agents, but the incidence of adverse reactions to this drug has been found to be as high as 83.2%, with grade 3 or severer reactions occurring in 20.3% of patients. Taking into consideration the post-marketing survey finding that adverse reactions to the drug first appear 2-3 weeks after the start of oral TS-1 therapy, we attempted a new dosing regimen for this drug, wherein each session of therapy lasted for 2 weeks, with a one-week interval between two consecutive sessions (herein-after called "the 2-week regimen"). This regimen was employed based on the expectation that the adverse reactions to the drug would be minimized and that the consecutive dosing period could be prolonged, while keeping the anti-cancer potency at a level similar to that expected with the 4-week dosing regimen with a 2-week interval between sessions (the 4-week regimen). The subjects were 38 patients with advanced or recurrent stomach cancer who were treated with TS-1 at our center between September 1999 and November 2001. Twenty-four patients treated using the 4-week method until January 2001 were taken as a historical control, and compared with 14 patients treated using the 2-week method from February 2001 and afterwards. The incidence of adverse reactions was 71% in the 2-week regimen group against 92% in the 4-week regimen group. The incidence of grade 3 or severe adverse reactions was 8% in the 2-week group and 21% in the 4-week group. Thus, the incidence of adverse reactions was lower in the 2-week group. The percentage of patients who complied with the dosing instructions completely during a 6-month period, as evaluated by the Kaplan-Meier method, was 86% in the 2-week group and 58% in the 4-week group. The response rate, as calculated in patients whose lesions could be evaluated, was 25% in the 2-week group and 19% in the 4-week group. These results suggest that the 2-week regimen may allow safer outpatient drug therapy using TS-1 and merits a trial when considering the QOL of patients. We propose conducting a phase-II multi-center clinical study of this regimen in the near future.