Abstract: Objective To establish a mathematical model for predicting significant liver histopathological changes in chronic HBV infection, and evaluate the diagnostic value of the model. Methods A retrospective analysis was performed on 457 patients with chronic HBV infection who were hospitalized and underwent liver biopsy without antiviral therapy in the Department of Hepatobiliary Medicine of Shanghai Public Health Clinical Center from December 2011 to December 2017. The pathological results of liver biopsy and routine laboratory indexes were collected, and anti-HBc quantitative detection was performed. According to Scheuer method, the inflammatory grade (G) and fibrosis stage (S) were divided into significant and non-significant hepatic necrotizing inflammation, hepatic fibrosis and hepatic injury groups. A mathematical model for predicting significant liver necrotizing inflammation, significant liver fibrosis and significant liver injury was constructed based on univariate analysis and multivariate Logistic regression analysis.Compared with FIB-4, GPR, APRI and RPR, the predictive performance of the model was evaluated by ROC curve analysis, and the diagnostic value was compared according to the area under ROC curve (AUC). Results Of the 457 patients, 178 had significant hepatic necrotizing inflammation (G≥2) and 279 had non-significant hepatic necrotizing inflammation (G<2), 248 had significant liver fibrosis (S≥2) and 209 had non-significant liver fibrosis (S<2), 264 had significant liver injury (G≥2 or/and S≥2) and 193 had non-significant liver necrotizing inflammation (G< 2 and S<2). According to the results of univariate analysis and multivariate Logistic regression analysis, the mathematical model M-SHN composed of anti-HBc, AST, PLT and TTR was established to predict significant liver necrotizing inflammation, and the mathematical model M-SHF composed of anti-HBc, PLT, ChE, TTR and gender to predict significant liver fibrosis, and M-SHI, a mathematical model composed of anti-HBc, PLT, TTR and sex, predicted significant liver injury, respectively. The predictive value of each model was analyzed by ROC curve. The AUC of M-SHN was 0.826 (95%CI: 0.788~0.860), and M-SHF was 0.776 (95%CI: 0.735~0.814), and M-SHI was 0.789 (95%CI: 0.748~0.825). Conclusion Based on routine laboratory indicators and serum anti-HBC quantification, M-SHN, M-SHF and M-SHI models were established, which have reliable predictive value for significant liver necrotizing inflammation, significant liver fibrosis and significant liver injury, and can help clinical evaluation of whether patients need antiviral therapy.

Key words: Chronic HBV infection, Anti-HBc quantification, Inflammation, Fibrosis, Diagnosis