Novel Algorithm for Non-Invasive Assessment of Fibrosis in NAFLD

• DOI: 10.1371/journal.pone.0062439
• Publication Date: 2013-05-01T22:57:06Z
• Authors: Jan-Peter Sowa; Dominik Heider; Lars Peter Bechmann; Guido Gerken; Daniel Hoffmann; Ali Canbay
• Keywords: Adult; Liver Cirrhosis; Male; 0303 health sciences; Science; Q; Decision Trees; Medizin; R; Middle Aged; Prognosis; 3. Good health; Fatty Liver; 03 medical and health sciences; Liver; Artificial Intelligence; Non-alcoholic Fatty Liver Disease; Medicine; Humans; Female; Biologie; Algorithms; Research Article

Introduction Various conditions of liver disease and the downsides biopsy call for a non-invasive option to assess fibrosis. A score would be especially useful identify patients with slow advancing fibrotic processes, as in Non-Alcoholic Fatty Liver Disease (NAFLD), which should undergo histological examination Patients/Methods Classic serum parameters, hyaluronic acid (HA) cell death markers 126 undergoing bariatric surgery morbid obesity were analyzed by machine learning techniques (logistic regression, k-nearest neighbors, linear support vector machines, rule-based systems, decision trees random forest (RF)). Specificity, sensitivity accuracy evaluated datasets predict fibrosis assessed. Results None single parameters (ALT, AST, M30, M60, HA) did differ significantly between 1 or 2. However, combining these using RFs reached 79% prediction more than 60% specificity 77%. Moreover, identified M30 M65 important classic parameters. Conclusion On basis generation scoring system seems feasible, even when only marginally tissue is available. Prospective evaluation novel markers, i.e. performed an optimal set predictors.