Accuracy of non-invasive scoring systems for diagnosing non-alcoholic steatohepatitis-related fibrosis: Multicenter validation study

Aim: Hepatic fibrosis is the most important factor for estimating the prognosis of patients with non-alcoholic fatty liver disease (NAFLD). A novel non-invasive scoring system, the FM-fibro index, showed high accuracy in a pilot study. The purpose of this study was to validate the efficacy of the FM-fibro index in a multicenter cohort. Methods: Among 18 institutions, we analyzed 400 Japanese patients with biopsy-proven NAFLD. We evaluated the accuracies of the FM-fibro index, CA-fibro index, and European Liver Fibrosis (ELF) panel by area under the receiver operator characteristics curves (AUROC). The FM-fibro index includes three formulas for type IV collagen 7S, hyaluronic acid, and vascular cell adhesion molecule-1. Results: Among 400 patients, 205 were women, and the median age was 56 years. The histological distribution of Matteoni types 1, 2, 3, and 4 was 11, 40, 15, and 334, and the distribution of hepatic fibrosis stages 0 to 4 was 67, 183, 55, 63, and 32, respectively. The AUROCs of the FM-fibro index, CA-fibro index, and ELF panel for non-alcoholic steatohepatitis (NASH)-related fibrosis were 0.7178/0.7095/0.7065, 0.7093, and 0.7245, respectively. The sensitivity and specificity of the FM-fibro index for predicting NASH-related fibrosis was 0.5359/0.5210/0.4641 and 0.8333/0.8182/0.8788, respectively. The accuracy of the FM-fibro index was not significantly different from that of the CA-fibro index or the ELF panel. Conclusions: The FM-fibro index can predict NASH-related fibrosis with sufficient accuracy compared with previous scoring systems. Further analyses that verify the accuracy of the FM-fibro index to distinguish significant or advanced fibrosis in patients with NAFLD are awaited. (UMIN-CTR: UMIN000018158).