A Simple Multi-Linear Regression Model for Predicting Fibrosis Scores in Chronic Egyptian Hepatitis C Virus Patients

Hepatitis C is considered as a common infection in Egypt, especially genotype 4. The prognosis of hepatitis C and the risk of developing cirrhosis are related to the stage of fibrosis. Liver biopsy is the best indicator for identifying the extent of liver fibrosis, but it has many draw-backs. Furthermore, it is costly and susceptible to sampling error. Non-invasive methods for the assessment of liver fibrosis are alternative in staging chronic liver diseases. The aim of this paper is to develop a simple multi-linear model to predict the levels of risk for liver fibrosis based on standard laboratory tests. In this proposed model, liver fibrosis was assessed via Metavir score; patients were categorized as mild (F0-F1), moderate (F2), or advanced (F3-F4) fibrosis stages. Statistical analysis was performed using Med Calc software. The relationship between serum markers and the presence of significant fibrosis was assessed. The P-value and the correlation coefficients revealed that, age, AST, AFP, Albumin, platelet count, Glucose, Postprandial Glucose test and BMI, were significantly associated with fibrosis. Multi-linear regression analysis is performed to develop a model for prediction of liver fibrosis scores based on serum markers. Sensitivity and Receiver Operating Characteristic (ROC) curve analysis were performed to evaluate the proposed model. In training set, the area under the receiver operating curve (AUROC) for differentiating mild fibrosis from others is 0.78; with sensitivity 68.8 and specificity 75.2 at cutoff point ?1.5, and for differentiating advanced from others is 0.82; with sensitivity 82.48 and specificity 78.3 at cutoff point >1.7. It has been concluded that, multi-linear regression model can predict fibrosis stages in chronic hepatitis C with accepted accuracy that could be used to reduce the need to assess the liver biopsy.