Novel Clinical Predictive Model for Stent Dysfunction after TIPS in Patients with Cirrhosis-Related Portal Hypertension

Objectives Transjugular intrahepatic portosystemic shunt (TIPS) is frequently compromised by stent dysfunction (SD). This multi-center retrospective study aimed to develop and validate a predictive tool for post-TIPS SD. Methods Between January 2019 and June 2024, 290 TIPS patients from three Chinese hospitals were collected, with 202 eligible cases randomly divided into a training set (n = 141) and an internal validation set (n = 61) at a 7:3 ratio using R software (version 4.2.3). Variable selection was performed via Least Absolute Shrinkage and Selection Operator (LASSO) regression, followed by multivariate logistic regression modeling. Model performance was assessed using the Hosmer-Lemeshow (H-L) test, calibration curves, receiver operating characteristic (ROC) curves, net reclassification improvement (NRI), integrated discrimination improvement (IDI), and decision curve analysis (DCA). Results Seven independent predictors of SD were identified: cirrhosis etiology, serum albumin, prothrombin time, blood ammonia, initial stent position, portal vein thrombosis, and maximum transverse diameter of the right hepatic lobe (MTDORHL). The area under the ROC curve (AUC) values of Model C (clinical features), Model R (radiological features), and Model CR (combined features) were 0.823, 0.790, and 0.901 in the training set, and 0.785, 0.736, and 0.866 in the validation set, respectively. Conclusion Three SD-specific models were constructed with multi-center data, and Model CR emerged as the optimal one with excellent predictive efficiency and application value, supported by a corresponding nomogram. The identified seven factors serve as independent risk predictors for post-TIPS SD.