Nomogram Predictive Model Analysis And Validation Of Hepatic Cirrhosis Complicated With Portal Venous Thrombosis

Objective: To investigate the nomogram predictive model of PVT formingin cirrhosis patients and validation.Method: The general data,serological indicators,ultrasound,imaging data of patients with cirrhosis combined with PVT and non-PVT admitted to our hospital from January 2014 to October 2021 were retrospectively analysis.The independent risk factors of cirrhosis PVT were screened by Lasso regression and multivariate Logistic regression model in the modeling group(before December 2020)and then established the nomogram prediction model.verified the model with modeling group date internally and validation group data(after December 2020)externally(time verification).Result: The independent risk factors for cirrhosis combined with PVT are history of splenectomy,diameter of portal vein,diameter of splenic vein and monocyte-to-lymphocyte ratio(P<0.05).The internal validation of nomogram prediction model in modeling group suggested that the AUC of the nomogram was 0.723(95%CI:0.684～0.761),the discrimination of nomogram is well.The P-values of the Hosmer-Lemeshow test were 0.443,suggesting a high reliability of the predicted probability by the model.The external validation of nomogram prediction model in validation group suggested that the AUC of the nomogram was 0.782(95%CI:0.698～0.857),the discrimination of nomogram is well.The P-values of the Hosmer-Lemeshow test were 0.065,suggesting a high reliability of the predicted probability by the model.Conclusion: We developed an easy-to-use nomogram to predict the PVT forming in cirrhosis patients.The nomogram can help clinicians screen patients at high risk of PVT,make timely and individualized clinical decisions for PVT,improve patients outcomes ultimately.