Circulating MicroRNA Profiling for Phenotypic Stratification in Patients with Metabolic Dysfunction-Associated Fatty Liver Disease: A Candidate-Based Study

Metabolic dysfunction-associated fatty liver disease (MAFLD) comprises phenotypic subgroups, including type-2 diabetes-associated MAFLD (T2D-MAFLD), obesity-associated MAFLD (OB-MAFLD), and lean MAFLD (L-MAFLD). Emerging evidence indicates that dysregulation of miRNAs plays a key role in MAFLD pathogenesis and progression. This study evaluated the diagnostic accuracy of a plasma miRNA-based signature as a non-invasive biomarker for early detection and phenotypic stratification of MAFLD. A total of 393 MAFLD patients and 109 healthy controls were enrolled. Plasma expression of miR-122, miR-103a, miR-222, miR-15a, miR-34a, miR-192, miR-197, and miR-99a was quantified using Reverse transcription polymerase chain reaction. Compared to controls, MAFLD patients exhibited significant upregulation of miR-122, miR-103a, miR-222, miR-15a, and miR-34a, alongside downregulation of miR-197 and miR-99a. Multinomial logistic regression revealed phenotype-specific associations: miR-103a, miR-34a, and miR-197 with T2D-MAFLD; miR-122, miR-222, and miR-99a with OB-MAFLD; and miR-15a with L-MAFLD. Receiver operating characteristic analysis demonstrated highest individual diagnostic accuracy for miR-197 in T2D-MAFLD (AUC = 0.784), miR-99a in OB-MAFLD (AUC = 0.869), and miR-15a in L-MAFLD (AUC = 0.776). Integrating combined miRNA panels with biochemical markers further improved diagnostic performance and clinical utility, achieving high positive and negative predictive values. In conclusion, plasma miRNA signatures enable phenotype-specific discrimination of MAFLD subtypes and may serve as promising non-invasive tools pending multi-center validation.