Prognostic factors in colorectal carcinoma treated by preoperative radiotherapy and immediate surgery

The clinicopathologic staging of colorectal cancer is the subject of recent debate. We studied morphologic variables in a series of tumors resected from 284 patients. Half had been prospectively, randomly allocated to receive a 4-day schedule of preoperative pelvic radiotherapy followed by immediate surgery. There was a significant (P less than 0.01) difference in the distribution of tumors of various histopathologic grades between irradiated (XS) and unirradiated (S) patients and borderline differences in the predictive values of venous spread, tumor grading, and local spread. However, these differences were less marked in 180 tumors examined by one review pathologist. They were thought to be due to misinterpretation of changes induced by radiotherapy. No differences were detected in the distribution of tumors of various sizes and Dukes' stage in the XS and S groups. The review pathologist recorded a borderline (P = 0.049) difference in the distribution of tumors of various CEA staining patterns between the XS and S groups. In a Cox regression model. Dukes' staging remained the most important predictive variable for survival and pelvic recurrence in the XS and S groups. Dukes' staging was apparently unchanged by this schedule of preoperative radiotherapy, but Broders' grading may be unreliable. Any new clinicopathologic staging system for colorectal cancer should record when preoperative radiotherapy is delivered. More studies of radiotherapy effects are required.