Abstract:?Objective?To find the independent risk factors of microvascular invasion(MVI)in hepatitis B(HBV)-related hepatocellular carcinoma(HCC), build a logistic regression model based on the data of clinical laboratory tests and analyze the prediction value of the model.?Methods?A cohort of 1 628 patients with HBV-related HCC was recruited in Eastern Hepatobiliary Surgery Hospital from January 2015 to December 2015. Their clinical data was collected. The independent risk factors of MVI in HBV-related HCC were anayazed by means of univariate and multivariate logistic regression. Then the logistic regression model was constructed, the prediction value of which was analyzed.?Results?According to the results of multivariate logistic regression, the independent risk factors of MVI in HBV-related HCC contained HBeAg(P<0.01), AFP(P<0.01), AFP-L3(P<0.01), PT(P<0.01), and PLT(P<0.01). Incorporating these 5 factors, the sensitivity, specificity, accuracy and AUCROC(95%CI) of the logistic regression model named Logistic(P)in differentiating M0 and M2(M0 vs M2) were 73.7%, 61.6%, 66.6% and 0.722(0.692,0.751), respectively.?Conclusions?HBeAg, AFP, AFP-L3, PT and PLT were the independent risk factors of MVI in HBV-related HCC. The logistic regression model based on the parameters of clinical laboratory tests was clinically valuable in predicting the existence of high-risk MVI(M2).