A nomogram model to predict the portal vein thrombosis risk after surgery in patients with pancreatic cancer

Abstract

BackgroundPortal vein thrombosis (PVT) is a common postoperative complication in patients with pancreatic cancer (PC), significantly affecting their quality of life and long-term prognosis. Our aim is to establish a new nomogram to predict the risk of PVT after PC surgery.MethodWe collected data from 416 patients who underwent PC surgery at our hospital between January 2011 and June 2022. This includes 87 patients with PVT and 329 patients without PVT. The patients were randomly divided into a training group and a validation group at a ratio of 7:3. We constructed a nomogram model using the outcomes from both univariate and multivariate logistic regression analyses conducted on the training group. The nomogram’s predictive capacity was assessed using calibration curve, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA).ResultsIn the study, the prevalence of PVT was 20.9%. Age, albumin, vein reconstruction and preoperative D-dimer were independent related factors. The model achieved a C-index of 0.810 (95% confidence interval: 0.752–0.867), demonstrating excellent discrimination and calibration performance. The area under the ROC curve of the nomogram was 0.829 (95% CI: 0.750–0.909) in the validation group. DCA confirmed that the nomogram model was clinically useful when the incidence of PVT in patients was 5%–60%.ConclusionWe have established a high-performance nomogram for predicting the risk of PVT in patients undergoing PC surgery. This will assist clinical doctors in identifying individuals at high risk of PVT and taking appropriate preventive measures.