Evaluating the American College of Surgeons national surgical improvement project risk calculator in extrahepatic biliary malignancy: results from the U.S. extrahepatic biliary malignancy consortium

Abstract

Objective: To assess the accuracy of the ACS NSQIP calculator for estimating risk after operation for gallbladder cancer and extrahepatic cholangiocarcinoma. Methods: Adult patients with gallbladder cancer, distal or hilar cholangiocarcinoma who underwent curative-intent, complete resection January, 2000–December, 2014 were included. The ability of the NSQIP calculator to accurately predict a particular outcome was assessed using the c-statistic (0.7 or above score for reasonable models) and Brier score (zero score for perfect models). Results: 854 adult patients were included. The most commonly performed procedures included right or left hepatic lobectomy (254, 29.7%), pancreaticoduodenectomy (132, 15.5%) and trisegmentectomy (118, 13.8%). 185 patients experienced surgical site infection (SSI) (21.7%, median predicted risk of 9.6%, range: 1.9–36.6%). There was 2.1% mortality (median predicted risk of 0.5%, range: 0–19.4%). There were 57 (6.7%) reoperations (median predicted risk of 3.3%, range: 0.9–13.2%). 169 (19.7%) patients were readmitted (median predicted risk of 9.1%, range: 2.6–32.2%). The median length of stay in the cohort was 8 days (Range: 0 119 days) and the median predicted length of stay was 6.5 days (Range: 2.5–25.5 days). The c-statistics for SSI, mortality, reoperation and readmission were 0.64, 0.74, 0.68 and 0.57 and Brier scores were 0.17, 0.02, 0.06 and 0.16 respectively (FIGURE). Conclusion: The ACS NSQIP Risk Calculator estimates the risk of mortality reasonably well for patients with gallbladder and extrahepatic biliary malignancy but generally underestimates risk of several other postoperative complications. Future modifications of the calculator for some patient populations are needed before widespread use of the tool can be encouraged.