Effect of the circumferential resection margin on survival following rectal cancer surgery.

BACKGROUND The aim was to determine the effect of the circumferential resection margin (CRM) on overall survival following surgical excision of rectal cancer. METHODS The effect of CRM on survival was examined by case mix-adjusted analysis of patients undergoing potentially curative excision of a rectal cancer between 1998 and 2002. RESULTS Of 1896 patients, 1561 (82.3 per cent) had recorded data on the CRM. In 232 patients (14.9 per cent) tumour was found 1 mm or less from the CRM. In 370 patients (23.7 per cent) it was over 1 mm but no more than 5 mm from the CRM, and in 288 (18.4 per cent) it was over 5 mm but no more than 10 mm from the CRM. The remaining 671 patients (43.0 per cent) had a CRM exceeding 10 mm. Overall 5-year survival rates for these groups were 43.2, 51.7, 66.6 and 66.0 per cent respectively. Compared with patients with a CRM exceeding 10 mm, the adjusted risk of death was significantly increased for patients with a CRM of 1 mm or less (hazard ratio (HR) 1.61, P < 0.001) and those with a margin greater than 1 mm but no larger than 5 mm (HR 1.35, P = 0.005). There was no added risk for patients with tumour more than 5 mm but 10 mm or less from the CRM (HR 1.02, P = 0.873). The adverse effect of a CRM greater than 1 mm but no larger than 5 mm was found particularly in mid-rectal cancers. CONCLUSION A predicted CRM of 5 mm or less on preoperative staging should be considered for neoadjuvant treatment.