| ObjectivesTo establish and validate the nomogram to predict probability of patients with hepatitis B cirrhosis after splenectomy developing into PVST by analyzing and screening the independent risk factors of PVST.Evaluate whether the nomogram was generally applicable to clinical practice.MethodsWe recorded information of patients with hepatitis B cirrhosis after splenectomy from January 2014 to June 2019 in the Second Affiliated Hospital of Chongqing Medical University retrospectively.we randomly divided all patients into the development group(n=100)and the validation group(n=33).The two groups were divided into PVST group and control group according to whether PVST occurred within 1 month after splenectomy Univariate and multivariate Logistic regression were used to screen the independent risk factors of PVST,and the nomogram prediction model was established based on the independent risk factors.We also estimated the discrimination,the calibration and clinical efficiency of the prediction model,respectively.ResultsThere were 133 patients in our study,including 100 patients in the development group and 33 in the validation group,the incidence of PVST after splenectomy was nearly 40% in all 133 patients.There was no significant difference in clinical data between the two groups(P>0.05).Univariate logistic regression analysis showed that the risk factors of PVST included history of hemorrhage,diameter of portal vein,diameter of spleen vein,volume of spleen,D-dimer and PLT addition(P<0.05);multivariate logistic regression analysis showed that the diameter of spleen vein,diameter of portal vein,PLT addition were independent risk factors of PVST after splenectomy.The AUROC(Cindex)were 0.847(95%CI:0.772~0.922)and 0.835(95CI%:0.693~0.976)in development group and validation group,that indicated the discrimination of the nomogram model was good.The 80% and 95% CI region of GiViTI calibration belt didn’t cross the 45-degree diagonal bisector line(P=0.586 and 0.589),P-values of the Hosmer-Lemeshow test were 0.568 and 0.540,suggested that the calibration of the prediction model in both groups were perfect.DCA curve show the threshold probability of 32%,the net benefits were 27% in development group and 32% in validation group respectively.Thus,the predict model has the good clinical efficiency.ConclusionsThe diameter of portal vein,diameter of spleen vein,PLT addition were the independent risk factors of PVST in patients with hepatitis B cirrhosis after splenectomy.The accuracy of the model was excellent,can facilitate help to early identification of patients having high-risk of PVST and treat it at the early stages. |