Abstract / Summary
Abstract Metabolic dysfunction-associated fatty liver disease (MAFLD) shows significant inter-patient heterogeneity, likely due to differences in the hepatic immune microenvironment. We aimed to identify clinically meaningful immunophenotypic subtypes based on quantitative expression of CD68, CD163, IL-6, and ZNF580. Liver biopsies from 19 patients with biopsy-proven MAFLD were analyzed. Immunohistochemical positive rates and counts for CD68, CD163, IL-6, and ZNF580 were quantified using digital image analysis. Hierarchical clustering identified patient subtypes. Clinical parameters, NAS, fibrosis stage, and ALT levels were compared across subtypes. Results : Three stable immunophenotypic subtypes were identified: Type M (macrophage-rich, high CD68/CD163, n = 7, 36.8%), Type I (inflammation-rich, high IL-6/ZNF580, n = 6, 31.6%), and Type D (immune-desert, low all markers, n = 6, 31.6%). NASs differed significantly ( p = 0.006, Kruskal-Wallis). Type D patients had the highest NAS (5.67 ± 0.82), significantly higher than Type I (3.83 ± 0.98, adj. p = 0.010) and Type M (4.14 ± 0.69, adj. p = 0.023). Fibrosis stage ( p = 0.542) and ALT levels ( p = 0.798) did not differ significantly. Intra-patient dominance analysis revealed macrophage-dominant (73.7%) and ZNF580-dominant (26.3%) cases, with no IL-6-dominant cases. Extreme case analysis showed that isolated high IL-6 (Patient 4) did not cause significant liver injury (NAS = 2, F0), whereas the immune-desert phenotype (Patient 23) presented with advanced bridging fibrosis (F3, NAS = 5) despite low marker expression. In this pilot cohort, our findings suggest that MAFLD may comprise three reproducible immunophenotypic subtypes with distinct histological activity but similar fibrosis burden. The immune-desert subtype is associated with high NAS despite low immune marker expression, suggesting immune-independent fibrogenic pathways. These findings provide a framework for precision immunophenotyping and subtype-targeted therapeutic strategies in MAFLD. However, due to the exploratory nature of this study and the small exploratory pilot cohort ( n = 19), our findings should be considered hypothesis-generating and require validation in larger cohorts.