Abstract: Objective To develop a noninvasive dignositic model comprising of transabdominal ultrasonography and serum markers for assessing portal vein thrombosis（PVT）in liver cirrhosis patients. Methods During Jan，2007 to Jan，2010，280 cirrhotic patients were randomly allocated to either an estimation group or a validation group. Abdominal spiral CT was performed and whether portal vein thrombosis existed or not was assessed in every patient. Seventeen common ultrasonographic and serum markers were analyzed initially in the estimation group to derive a predictive model. The model created was then assessed with ROC analysis. It was also applied to the validation group to derive its accuracy. Results Among severteen variables associated with portal vein thrombosis selected by univariate analysis，the D-dimer，main portal vein width and platelet were indentified by multivariate Logistic regression analysis as independent risk factors of portal vein thrombosis. A portal vein thrombosis index constructed from the above three markers was established. In ROC analysis，the AUC was 0.864（0.753，0.946）for predicting existence of portal vein thrombosis using the optimal cutoff score. The diagonstic sensitivity，specificity， positive predictive value，negative predictive value and accuracy were 82.1%，86.7%，93.53%，64.58%，81.77% respectively. For the validation group，the AUC and diagnostic accuracy were 0.886（0.785，0.962）and 82.16% respectively. Conclusion A noninvasive model consisting of the D-dimer，main portal vein width and platelet is helpful in predicting existence of portal vein thrombosis in patients with cirrhosis.

Key words: Liver cirrhosis, Portal vein thrombosis, Predictive model