Application and validation of a surgical risk calculator in a liver surgery cohort of patients in Veracruz, Mexico.

Objective: Multiple models have tried to predict the morbidity and mortality of liver resections (HR). This study aims to determine the efficacy and validity of the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) surgical risk calculator in a cohort of patients undergoing HR in Veracruz, Mexico.

Material And Methods: Retrospective analysis of patients undergoing HR between 2005 and 2019. To estimate the performance of the calculator, the observed results were compared with the average risk predicted by the calculator, using ROC curve, Brier score and Z test.

Results: 67 patients were evaluated, mean age 51.9 years of age, 50.7% female. The majority of resections (56.7%) were for malignancy and 62.1% were partial hepatectomies (up to 3 liver segments). The morbidity was 25.4% and the mortality 4.5%. There was a good prediction in the complications (serious complication: C = 0.725 statistic, Brier score 0.26, p = 0.006 and any statistical complication C = 0.731, Brier score 0.33, p = 0.005) and mortality (C = 0.922 statistic, Brier score 0.005, p = 0.014).

Conclusions: The application of the ACS-NSQIP calculator in patients undergoing HR has good discrimination power and good predictive ability. Prediction of postoperative risks achieves a preoperative planning of the appropriate procedure.