The miRFIB-Score

Background: The current diagnosis of early-stage liver fibrosis often relies on a serological or imaging-based evaluation of the stage of fibrosis, sometimes followed by an invasive liver biopsy procedure. Novel non-invasive experimental diagnostic tools are often based on markers of hepatocyte damage, or changes in liver stiffness and architecture, which are late-stage characteristics of fibrosis progression, making them unsuitable for the diagnosis of early-stage liver fibrosis. miRNAs control hepatic stellate cell (HSC) activation and are proposed as relevant diagnostic markers. Methods: We investigated the possibility of circulating miRNAs, which we found to be dysregulated upon HSC activation, to mark the presence of significant liver fibrosis (F ≥ 2) in patients with chronic alcohol abuse, chronic viral infection (HBV/HCV), and non-alcoholic fatty liver disease (NAFLD). Results: miRNA-profiling identified miRNA-451a, miRNA-142-5p, Let-7f-5p, and miRNA-378a-3p to be significantly dysregulated upon in vitro HSC activation, and to be highly enriched in their extracellular vesicles, suggesting their potential use as biomarkers. Analysis of the plasma of patients with significant liver fibrosis (F ≥ 2) and no or mild fibrosis (F = 0-1), using miRNA-122-5p and miRNA-29a-3p as positive control, found miRNA-451a, miRNA-142-5p, and Let-7f-5p, but not miRNA-378a-3p, able to distinguish between the two patient populations. Using logistic regression analysis, combining all five dysregulated circulating miRNAs, we created the miRFIB-score with a predictive value superior to the clinical scores Fibrosis-4 (Fib-4), aspartate aminotransferase/alanine aminotransferase (AST/ALT) ratio, and AST to platelet ratio index (APRI). The combination of the miRFIB-score with circulating PDGFRβ-levels further increased the predictive capacity for the diagnosis of significant liver fibrosis. Conclusions: The miRFIB- and miRFIBp-scores are accurate tools for the diagnosis of significant liver fibrosis in a heterogeneous patient population.