Abstract: Objective To explore the noninvasive diagnostic model based on Logistic regression and ROC curve in liver fibrosis and cirrhosis. Methods 58 patients with chronic hepatitis B（CHB）,84 with compensated cirrhosis and 35 with decompensated cirrhosis were included in this study. Fibroscan was conducted to obtain liver stiffness measurement（LSM） in all the patients,and serological examination were carried out. Logistic regression,plotted ROC curve and area under ROC curve（AUC） were applied to assess the diagnostic value of all the diagnostic model. Results The LSM were 5.2（4.2~7.3） kPa,7.3（5.8~15.5） kPa and 12.2（8.9~20.0） kPa in patients with CHB,compensated and decompensated cirrhosis,respectively（P<0.05）;Frons index was strongly positively correlated to fibrosis index based on the 4 factors（Fib-4）,red blood cell volume distribution width （RDW） to platelet ratio（RPR）,and PLT counts（r=0.93,r=0.89,r=-0.91,P<0.01）;aspartate aminotransferase-to-platelet ratio index（APRI） was positively correlated to FIB-4 or PLT（r=0.83,r=-0.80,P<0.01）;FIB-4 was positively correlated to RPR or PLT （r=0.86,r=-0.87,P<0.01）;RPR was negatively correlated to PLT （r=-0.98,P<0.01）;ROC curve analysis showed that the diagnostic accuracy of Logistic prediction model（PRE） was 80%,and the area under the curve（AUC） was 0.86 with the sensitivity and specificity of 86% and 78%,respectively. Conclusion Fibroscan detection has a promising value in the diagnosis of hepatic fibrosis in patients with chronic hepatitis B and the prediction models are noninvasive procedure in clinical practice.

Key words: Hepatitis B, Cirrhosis, Fibroscan, Serological index, Hepatic fibrosis, Noninvasive diagnosis models