MACHINE LEARNING PREDICTION OF HEPATIC FIBROSIS IN HEPATITIS B EGYPTIAN PATIENTS BASED ON CLINICAL LABORATORY PARAMETERS

Abstract

Liver fibrosis stage prediction in chronic hepatitis B virus (HBV) infected patients is vital. Liver biopsy is the reference style and gold standard to evaluate fibrosis stage but with many drawbacks. Therefore, using noninvasive methods are better alternatives. In this study, seven clinical laboratory parameters of 235 chronic HBV Egyptian patients with Hepatitis B virus were collected from HBV clinic at National Liver Institute that belongs to Menoufia University in Egypt. The aim of this study is to experiment multiple machine-learning methods based on clinical parameters to build efficient classification models that predict two liver related issues: the fibrosis stage and cirrhosis of liver in chronic HBV Egyptian patients. Also, attribute selection methods were applied to reduce the dimensionality and find the most relevant parameters. For fibrosis stage prediction, a classification model based on Logistic Regression achieved AUROC of 0.991 and accuracy of 93.61%. Besides, using only four parameters selected as the most relevant, AUROC of 0.971 and accuracy of 95.74% were achieved. For cirrhosis of liver prediction, a classification model based on Logistic Regression and cost sensitive with penalty of 2 achieved AUROC of 0.936 and accuracy of 91.49%. Besides, using only three parameters selected as the most relevant, AUROC of 0.92 and accuracy of 85.11% were achieved. The classification models outperformed noninvasive index-based method, FIB-4 that depends on four clinical parameters, in both fibrosis stage and liver cirrhosis prediction in chronic HBV Egyptian patients.