Machine Learning Assessment for Severity of Liver Fibrosis for Chronic HBV Based on Physical Layer With Serum Markers

Abstract

Noninvasive assessment of severity of liver brosis is crucial for understanding histology and making decisions on antiviral treatment for chronic HBV in view of the associated risks of biopsy. We aimed to develop a computer-assisted assessment system for the evaluation of liver disease severity by using machine leaning classi er based on physical-layer with serum markers. The retrospective data set, including 920 patients, was used to establish Decision Tree Classi er (DTC), Random Forest Classi er (RFC), Logistic Regression Classi er (LRC), and Support Vector Classi er (SVC) for liver brosis severity assessment. Training and testing samples account for 50% of the data set, respectively. The best indicator combinations were selected in random combinations of 24 indicators including 67 108 760 group indicators by four different machine learning classi ers. The resulting classi ers prospectively tested in 50% testing patients, and the sensitivity, speci city, overall accuracy, and receiver operating characteristics (ROC) were used to compare four classi ers to existed 19 models. Results show that the RFC-based classi er system, with 9 indicators, is feasible to assess severity for liver brosis with diagnostic accuracy (greater than 0.83) superior to existing 19 models. Additional studies based on a large data set with full serum markers and imaging information are necessary to enhance diagnostic accuracy and to expand clinical application.