The influence of circumferential resection margins on long-term outcomes following rectal cancer surgery.

PURPOSE: Circumferential resection margin involvement after rectal cancer surgery is associated with local recurrence and decreased survival, but definitions of "safe" margins vary. This study assessed the influence of various circumferential margins on long-term outcome from rectal cancer surgery.

METHODS: Data were extracted from a rectal cancer database of patients undergoing rectal resection at a tertiary referral center between 1971 and 1996. The influence of circumferential margins on five-year local recurrence and cancer-specific survival were assessed using Cox regression.

RESULTS: Circumferential margin measurements were available from 435 patients (median follow-up, 70.4 months). Cancer-specific survival at five years was 80.8%, 69.2%, 59.2%, and 34.1% for tumors with a circumferential resection margin of >10 mm, 3-10 mm, 2 mm, and < or =1mm, respectively (P < 0.001). Local recurrence at five years was 9.0%, 14.7%, and 25.8% for margins >10 mm, 2-10 mm, and < or =1 mm, respectively (P = 0.001). Independent predictors of cancer-specific mortality were circumferential margins of < or =1 mm vs. >10 mm (odds ratio = 3.38, P = 0.014) or 2 mm (odds ratio = 2.24, P = 0.029), Dukes Stage (C2 vs. A: odds ratio = 15.18, P < 0.001), and vascular invasion (present vs. absent: odds ratio = 1.51, P = 0.033). Local recurrence was predicted by a margin of < or =1 mm (odds ratio = 2.29, P = 0.041), gender (female vs. male: odds ratio = 0.25, P = 0.002), Dukes Stage (C2 vs. A: odds ratio = 28.89, P = 0.003), and vascular invasion (extramural vs. none: odds ratio = 2.04, P = 0.024).

CONCLUSION: Circumferential margins < or =2 mm are associated with significantly reduced cancer-specific survival, and margins < or =1 mm with increased local recurrence, when other factors are accounted for, challenging the assumption that a circumferential resection margin of < or =1 mm is safe.