Outcome after curative resection for locally recurrent rectal cancer

PURPOSE: Few biologic markers have been studied as prognostic factors in recurrent rectal carcinoma patients. We sought to determine the influence of clinical, pathologic, and biologic (p53, bcl- 2, and ki- 67) variables on survival after curative resection of locally recurrent rectal cancer. METHODS: Retrospective review of patients with locally recurrent rectal cancer who received surgery with curative intent. RESULTS: From 1988 to 1998, 134 patients with locally recurrent rectal cancer underwent operative exploration. Curative resection was performed in 85 patients. Median follow- up was 43 (range, 1.3 - 149) months. On multivariate analysis, negative predictors of overall survival included an elevated carcinoembryonic antigen level (P = 0.02; hazard ratio 2.41; 95 percent confidence interval, 1.19 - 4.89) and an R1 resection margin (P = 0.01; hazard ratio, 2.81; 95 percent confidence interval, 1.27 - 6.21). In 26 patients for whom biologic variables were available, p53, bcl- 2, and ki- 67 did not significantly impact disease-specific survival or overall survival. Five-year disease-specific survival, overall survival, and pelvic control rates were 46, 36, and 51 percent respectively. Of the 50 patients who relapsed, time to second local recurrence was longer than time to development of metastasis (median, 16.5 vs. 9 months). Median survival for patients with metastatic recurrence was 26.1 vs. 41.5 months for those with a subsequent local recurrence alone. CONCLUSIONS: Approximately two- thirds of patients with locally recurrent rectal cancer can be resected for cure. Preoperative carcinoembryonic antigen and an R0 resection margin were the only significant predictors of overall survival. p53, bcl-2, and ki-67 did not impact survival outcomes.