Objective: To determine whether surgical subspecialty training in colorectal surgery or frequency of rectal cancer resection by the surgeon are independent prognostic factors for local recurrence (LR) and survival.
Summary background data: Variation in patient outcome in rectal cancer has been shown among centers and among individual surgeons. However, the prognostic importance of surgeon-related factors is largely unknown.
Methods: All patients undergoing potentially curative low anterior resection or abdominoperineal resection for primary adenocarcinoma of the rectum between 1983 and 1990 at the five Edmonton general hospitals were reviewed in a historic-prospective study design. Preoperative, intraoperative, pathologic, adjuvant therapy, and outcome variables were obtained. Outcomes of interest included LR and disease-specific survival (DSS). To determine survival rates and to control both confounding and interaction, multivariate analysis was performed using Cox proportional hazards regression.
Results: The study included 683 patients involving 52 surgeons, with > 5-year follow-up obtained on 663 (97%) patients. There were five colorectal-trained surgeons who performed 109 (16%) of the operations. Independent of surgeon training, 323 operations (47%) were done by surgeons performing < 21 rectal cancer resections over the study period. Multivariate analysis showed that the risk of LR was increased in patients of both noncolorectal trained surgeons (hazard ratio (HR) = 2.5, p = 0.001) and those of surgeons performing < 21 resections (HR = 1.8, p < 0.001). Stage (p < 0.001), use of adjuvant therapy (p = 0.002), rectal perforation or tumor spill (p < 0.001), and vascular/neural invasion (p = 0.002) also were significant prognostic factors for LR. Similarly, decreased disease-specific survival was found to be independently associated with noncolorectal-trained surgeons (HR = 1.5, p = 0.03) and surgeons performing < 21 resections (HR = 1.4, p = 0.005). Stage (p < 0.001), grade (p = 0.02), age (p = 0.02), rectal perforation or tumor spill (p < 0.001), and vascular or neural invasion (p < 0.001) were other significant prognostic factors for DSS.
Conclusion: Outcome is improved with both colorectal surgical subspecialty training and a higher frequency of rectal cancer surgery. Therefore, the surgical treatment of rectal cancer patients should rely exclusively on surgeons with such training or surgeons with more experience.