Abstract
BACKGROUND AND AIMS: Hepatitis B virus (HBV) causes chronic hepatic infection, leading to various advanced liver diseases. Currently, there is still a lack of effective non-invasive biomarkers for evaluating hepatic inflammation and fibrosis. Circulating microRNAs (miRNAs) serve as key regulators and potential biomarkers in the progression and pathogenesis of HBV-associated progressive liver diseases. Here, we characterized the miRNA profile in chronic hepatitis B (CHB) patients and probed their association with liver inflammation and fibrosis. APPROACHES AND RESULTS: We profiled 34 candidate miRNAs in serum from a well-characterized HBV Disease Continuum Cohort, comprising 165 individuals in the discovery set (42 healthy controls, HC; 40 CHB; 42 cirrhosis; 41 hepatocellular carcinomas, HCC) and 143 in an independent testing set. Serum miRNA levels were quantified by qRT-PCR. Supervised clustering and correlation analyses revealed distinct miRNA expression profiles across disease stages. Statistical analyses included logistic regression, and ROC/AUC evaluation with 5-fold cross-validation and external validation. Unsupervised clustering and correlation with histological G/S staging revealed stage-specific miRNA signatures: miR-224-5p, miR-125a-5p, and miR-15a-5p peaked in cirrhosis and strongly correlated with fibrosis stage (miR-224-5p: r = 0.606, p = 1.5E-12), while miR-200a-3p and miR-939-5p were predominantly upregulated in HCC. Critically, miR-224-5p emerged as a robust non-invasive biomarker for cirrhosis, with exceptional diagnostic accuracy (AUC = 0.973 in discovery; 0.906 in external validation), significantly outperforming APRI (0.803) and FIB-4 (0.809), and remained the sole independent predictor in multivariable analysis (p = 0.009). For HCC detection, the miR-200a-3p/AFP combined model achieved outstanding performance (AUC > 0.9), substantially improving upon AFP alone (0.737). Bioinformatic prediction of targets (297 for miR-224-5p; 616 for miR-200a-3p) highlighted associated in cancer- and senescence-related pathways; however, these associations are in silico and require experimental validation. CONCLUSION: We identify miR-224-5p as a fibroinflammatory activity indicator for early cirrhosis detection and miR-200a-3p as a synergistic enhancer of AFP for non-invasive HCC diagnosis, establishing a dual miRNA signature that spans the HBV disease continuum and addresses critical gaps in current risk stratification. These findings highlight the potential of specific serum miRNAs as non-invasive biomarkers for monitoring disease progression and improving the differential diagnosis during the process of HBV-related liver diseases.