RESUMO
Objective:To construct a predictive nomogram on postoperative overall survival (OS) in patients with pancreatic ductal adenocarcinoma (PDAC) after R 0 radical pancreaticoduodenectomy, and to evaluate its performance. Methods:The clinicopathological data of patients who underwent radical pancreaticoduodenectomy at the Faculty of Hepato-Pancreato-Biliary Surgery, Chinese PLA General Hospital from January 2014 to December 2019 for pathologically diagnosed PDAC were retrospectively collected and analyzed. There were 119 patients, with 85 males and 34 females, aged (58±11) years. Using multivariate Cox regression analysis (stepwise regression), a prediction nomogram was constructed. Concordance index (C-index), calibration curve, and time-dependent receiver operating characteristic (ROC) curve were applied to evaluate the predictive performance.Results:The 1-, 2-, and 3-year cumulative survival rates of these 119 patients were 67.2%, 35.0%, and 24.8%, respectively. High-grade tumors (poorly differentiated and undifferentiated), vascular carcinoma embolus, systemic immune inflammatory index <279.4×10 9/L, prognostic nutritional index <40.5, alanine aminotransferase-to-aspartate aminotransferase ratio>1.1, total bilirubin>258.5 μmol/L and plasma fibrinogen>3.43 g/L were independent risk factors for poor OS for PDAC patients after radical pancreaticoduodenectomy (all P<0.05). These indicators, together with age >63 years, constituted the regression formula for prediction with a C-index=0.74. The areas under the curve of ROC for the nomogram on predicting survival were 0.795, 0.803, and 0.836 at 1, 2, and 3-year respectively, and only slight deviations were observed on the calibration curves from the standard 45° line, suggesting that the survival prediction of the model in this dataset fitted well with the actual survival status. Conclusion:The predictive nomogram on OS in patients after R 0 radical pancreaticoduodenectomy based on the clinicopathological characteristics of PDAC was internally validated to have a good predictive performance on OS. The nomogram can help to optimize prognostic risk stratification and treatment decisions for this subgroup of patients. This prediction model needs to be further verified and improved by using large-scale cohort studies.