Contributions of pathologic margins and Dukes' stage to local recurrence in colorectal carcinoma.

Two hundred fourteen patients with colorectal carcinoma who underwent curative resection for biopsy-proved or autopsy-proved local recurrences with a minimum of 2 years follow-up were evaluated. The only predictive variables for anastomotic recurrence were Dukes' stage and tumor margins. There were 49 Dukes' A lesions with no observed recurrences. There were also 83 Dukes' B lesions and 84 Dukes' C lesions with a total of 18 local recurrences in 214 cases or 8.4 percent (of Duke's B lesions or 6 percent and 14 of Dukes' C lesions or 17 percent). When proximal or distal margins were less than 5 cm there were seven total recurrences (three Dukes' B lesions and four Dukes' C lesions). However, when margins were greater than 5 cm, 11 local recurrences were observed (1 Dukes' B and 10 Dukes' C lesions). It appears that margins are not as important in preventing local recurrences of Dukes' A lesions as they are of both Dukes' B and C lesions. Although the numbers are small in this study, it appears that Dukes' B lesions can be satisfactorily resected with a very low incidence of local recurrence if their margins are 5 cm or greater, whereas if the resected margins are less than 5 cm, the incidence of local recurrences increases from 9 percent (1 of 11 lesions) to 43 percent (3 of 7 lesions), or almost a fivefold increase. Therefore, it appears that good surgical technique and adequate margins of greater than 5 cm are very important in reducing local recurrences of Dukes' B lesions. However, when margins are greater than 5 cm, this does not guarantee freedom from local recurrence of Dukes' C lesions. This may merely reflect the difference in the biologic characteristics among Dukes' A, B, and C lesions and the fact that limited resection, particularly in the rectosigmoid region, cannot possibly remove all intralymphatic disease, which is the presumed culprit in locally recurrent Dukes' C lesions.