BACKGROUND.
The prognostic factors and natural history of recurrence in patients with colorectal carcinoma who underwent curative resection and no other therapy were analyzed. METHODS. The object of analysis was the potentially curative resection only subgroup in the randomized clinical trial (RCT) that we performed. Cox's proportional hazards model was used mainly to analyze recurrence rates during the first 5 years after surgery. RESULTS. The analysis was performed on a subgroup of the RCT (279 patients with colon carcinoma and 293 patients with rectal carcinoma). Five-year disease free survival rates were 76.3% and 56.5% for colon and rectal carcinomas, respectively. The prognostic factors for recurrence for colon carcinoma patients were different from those with rectal carcinoma. For colon carcinoma, only Dukes stage was significant, whereas for rectal carcinoma, Dukes stage, age, location of the tumor, and serosal and venous invasion by cancer cells were prognostic factors. Logtransformed disease free survival rates were linear in Dukes Stage B and biphasic in Dukes Stage C for both colon and rectal carcinoma. The two phases in Dukes Stage C intersected at 2.85 and 3.04 years, respectively. The annual hazard value was high for the first 3 years in both colon and rectal carcinoma. CONCLUSIONS. We conclude that follow-up of patients with colorectal carcinoma who undergo potentially curative resection is of particular importance in the first 3 years after surgery. Furthermore, the usefulness of adjuvant chemotherapy can be adequately evaluated from data yielded during this postoperative period. Cancer