Non-invasive diagnosis of non-alcoholic fatty liver disease using an algorithm combining clinical indexes and ultrasonographic measures

Abstract

Introduction and objectiveNon-Alcoholic Fatty Liver Disease (NAFLD) is a metabolic liver disease related to insulin resistance, which requires invasive methods for diagnosis. The aim of this study was to analyze whether the use of an algorithm involving both clinical indices and hepatic ultrasound measurements improves the accuracy for the non-invasive diagnosis of NAFLD. Patients and methodsCross-sectional study with patients undergoing elective cholecystectomy. We collected anthropometric, metabolic, liver biopsy, and liver ultrasonography data. We calculated unpaired t-test and Pearson’s coefficient, and areas under the receiver-operating characteristic curves (AUROC) for the Fatty Liver Index (FLI), Lipid Accumulation Product (LAP) indexes, right liver index diameter, and for predictive models constructed with discriminant analysis. ResultsOne hundred patients in groups with and without NAFLD. FLI, LAP, right and caudate liver lobe diameters, and congestion index were higher in NAFLD group (p = 0.011, p = 0.011, p = 0.001, p = 0.027, p = 0.009). The right liver lobe diameter had the highest AUROC. Predictive models that combined sensitivity and specificity for the clinical indexes and liver ultrasound had an AUROC over 0.7. ConclusionThe ultrasonography measure of right liver lobe diameter by itself can reliably identify patients with NAFLD with a good sensitivity and specificity, however, this can be improved by adding the LAP mathematical index in our population.