Tumor cell DNA (ploidy) content was measured prospectively in samples from 320 patients resected for colorectal cancer with a minimum follow-up time of 2 years. All patients were followed and those with recurrence were investigated carefully. There was no correlation between tumors with an abnormal cellular DNA content (aneuploid or tetraploid) and patient age, sex, tumor site, pathologic stage, or histologic grade. In 236 patients who underwent potentially curative operations, 75 (32%) had local and/or distant recurrence. The recurrence rate was significantly higher (test statistic, 4.3; P = 0.04) for those patients with aneuploid tumors (52 of 142, 37%) compared with those with diploid tumors (23 of 94, 24%). The subgroups of patients where ploidy exerted an effect were in patients with Stage B tumors or mobile tumors and in patients over 65 years of age. Further analysis showed that there was a twofold increase in local recurrence and a threefold increase in distant recurrence in patients with aneuploid tumors, but no excess of patients who had both local and distant recurrence. Measurement of DNA ploidy can identify a group of patients undergoing curative surgery for colorectal cancer at high risk for recurrence. In combination with clinicopathologic factors, DNA ploidy may be useful in analyzing the results of trials and in planning adjuvant therapy.