OBJECTIVE: To identify rates of positive circumferential resection margin (CRM) for colon cancer surgery in the U.S. SUMMARY BACKGROUND DATA: CRM is one of the most important determinants of local control in colorectal cancers. The extent to which CRM involvement exists after colon cancer surgery is unknown. METHODS: Colon cancer cases with resection 2010-2015 were identified from the National Cancer Data Base. Adjusting for patient and tumor characteristics, comparisons were made between cases with CRM >1 mm (negative margin) and those with margin involved with tumor or ≤1 mm (positive margin, CRM+). Hospital-level analysis was performed, examining observed-to-expected CRM+ rates. RESULTS: In total, 170,022 cases were identified: 150,291 CRM- and 19,731 CRM+ (11.6%). Pathologic T-category was the greatest predictor of CRM+, with higher rates in pT4(25.8%), pT4A(24.7%), and pT4B(31.5%) versus pT1(4.5%), pT2(6.3%) and pT3 (10.9%, P < 0.001). Within pT4 patients, predictors of CRM+ included signet-ring histology (38.1% vs. 26.7% non-mucinous, and 26.9% mucinous adenocarcinoma, P < 0.001), removing < 12 lymph nodes (36.5% vs. 26.1% ≥12, P < 0.001), community facilities (32.7%) vs. academic/research (23.6%, P < 0.001), year (30.1% 2010 vs. 22.6% 2015, P < 0.001), and hospital volume (24.5% highest quartile vs. 32.7% lowest, P < 0.001). Across 1,288 hospitals, observed-to-expected ratios for CRM+ ranged from 0 to 7.899; 429 facilities had higher than expected rates. CONCLUSIONS: Overall rate of CRM+ in U.S. colon cancer cases is high. Variation exists across hospitals, with higher than expected rates in many facilities. Although biology is a major influencing factor, CRM+ rates represent an area for multidisciplinary improvement in quality of colon cancer care.
OBJECTIVE: To identify rates of positive circumferential resection margin (CRM) for colon cancer surgery in the U.S. SUMMARY BACKGROUND DATA: CRM is one of the most important determinants of local control in colorectal cancers. The extent to which CRM involvement exists after colon cancer surgery is unknown. METHODS:Colon cancer cases with resection 2010-2015 were identified from the National Cancer Data Base. Adjusting for patient and tumor characteristics, comparisons were made between cases with CRM >1 mm (negative margin) and those with margin involved with tumor or ≤1 mm (positive margin, CRM+). Hospital-level analysis was performed, examining observed-to-expected CRM+ rates. RESULTS: In total, 170,022 cases were identified: 150,291 CRM- and 19,731 CRM+ (11.6%). Pathologic T-category was the greatest predictor of CRM+, with higher rates in pT4(25.8%), pT4A(24.7%), and pT4B(31.5%) versus pT1(4.5%), pT2(6.3%) and pT3 (10.9%, P < 0.001). Within pT4 patients, predictors of CRM+ included signet-ring histology (38.1% vs. 26.7% non-mucinous, and 26.9% mucinous adenocarcinoma, P < 0.001), removing < 12 lymph nodes (36.5% vs. 26.1% ≥12, P < 0.001), community facilities (32.7%) vs. academic/research (23.6%, P < 0.001), year (30.1% 2010 vs. 22.6% 2015, P < 0.001), and hospital volume (24.5% highest quartile vs. 32.7% lowest, P < 0.001). Across 1,288 hospitals, observed-to-expected ratios for CRM+ ranged from 0 to 7.899; 429 facilities had higher than expected rates. CONCLUSIONS: Overall rate of CRM+ in U.S. colon cancer cases is high. Variation exists across hospitals, with higher than expected rates in many facilities. Although biology is a major influencing factor, CRM+ rates represent an area for multidisciplinary improvement in quality of colon cancer care.