Purpose A randomized controlled trial was conducted to determine whether pathologic necrosis in response to preoperative treatment with uracil-tegafur(UFT) could be used to identify patients with colorectal cancer most likely to beneWt from postoperative adjuvant therapy with the drug. Patients and methods The 152 patients with colorectal cancer who received preoperative UFT at a dose of 600 mg/ day for at least 10 days were classiWed into two groups according to the pathologic necrosis in resected tumor specimens: 90% or more necrosis (sensitive) versus less than 90% necrosis (insensitive). After excluding 13 ineligible patients, the remaining 139 were then randomly assigned to receive postoperative adjuvant UFT (400 mg/day) for 12 months or no treatment. Results Preoperative and postoperative UFT produced no serious toxicity in any of the patients. Among the 22 patients with sensitive tumors, overall survival was signiWcantly better in the UFT group (n = 12) than in the control (n = 10) (100 vs. 70.0%; P = 0.023). Among the 117 patients with insensitive tumors, there was no signiWcant diVerence between the two groups (n = 60, 68.1% vs. n = 57, 76.6%; P = 0.373). Conclusion Our method involving neoadjuvant UFT can identify patients most likely to beneWt from postoperative UFT, as well as those unlikely to beneWt from such treatment.