Methods. CPI-11, 25mg/m2, determined by dividing the MTD dose per month by 12, was given on days 1, 2, and 3 of every week, to 21 consecutive patients; 12 with metastatic colon cancer and 9 with metastatic gastric cancers. Results. The total delivered dose of CPI-11 per patient was more than 1000mg in 17 (80.1%) of the 21 patients. Grade 3 marrow depression developed in 3 (14.3%) patients, and although nausea, vomiting, alopecia, and diarrhea developed in some patients, these side effects were all categorized as grade 2 or milder. The antitumor effect was evaluated in 18 patients with measurable lesions, who had received CPI-11 according to our regimen for at least 3 weeks. Of these 18 patients, 10, 7, and 1, respectively, had a found to have partial response, no change, or progression of disease, demonstrating a 55.6% efficacy rate [colon 6/10 (60.0%) and stomach 4/8 (50.0%)]. Moreover, time to progression (TTP) was greater than 90 days in 12 (75.0%) of these 18 patients. Conclusion. These results show that our low-dose, divided MTD of CPI-11 regimen is a promising method of reducing toxicity and strengthening the antitumor effect, justifying further large-scale comparative clinical studies to verify this potential. We devised a new treatment regimen, delivering a frequent low dose of CPT-11, calculated by dividing the maximum tolerated dose (MTD) to reduce its toxicity without impairing its efficacy.