MicroRNAs in aflatoxin B1-associated hepatocellular carcinogenesis: a scoping review of molecular networks, multiomics evidence, machine learning and translational potentials

Aflatoxin B1 (AFB 1 ) is a foodborne hepatocarcinogen, but it remains unclear whether microRNA dysregulation reflects AFB 1 -associated hepatocarcinogenesis or general hepatocellular carcinoma (HCC) biology. This scoping review mapped evidence linking AFB 1 exposure, microRNA regulation, molecular mechanisms, multiomics, computational modelling, and machine learning approaches to HCC diagnostics, prognostics, and therapeutics. Peer-reviewed studies published from 2016 to 2026 were identified through Scopus, Embase, PubMed, and Web of Science, supplemented by citation searching, and screened using PRISMA ScR-aligned criteria. The final corpus comprised 104 unique studies classified by AFB 1 attribution strength and evidence type across five strata spanning direct human AFB 1 HCC microRNA evidence to general HCC contextual evidence. Experimental mechanistic studies predominated, whereas only one study directly combined measured AFB 1 exposure, HCC, and microRNA assessment in humans. The study enrolled 366 HCC cases and 662 controls; circulating miR-4651 provided the strongest exposure-anchored clinical signal, discriminating AFB 1 -positive HCC with an area under the curve of 0.85, 78.1% sensitivity, and 92.1% specificity. Other studies evaluated exposure-responsive hepatotoxicity, experimental malignant transformation, or rat hepatocarcinogenesis, which cannot be considered interchangeable endpoints. Functionally investigated injury candidates included miR-1679 and miR-124a-3p. Multiomics findings implicated DNA damage, methylation, oxidative stress, apoptosis, inflammation, and metabolic dysregulation, although pathway overlap was heterogeneous and independent functional confirmation was limited. Computational and machine learning studies have expanded network and feature prioritisation, but they have often relied on inferred exposure and limited external validation. Direct human evidence remains scarce, and candidate microRNAs occupy different validation stages. Biological plausibility does not establish clinical readiness; multicentre exposure-anchored validation, harmonised assays and prospective evaluation are required.