Abstract
Background: Sarcopenia is a clinically important complication of chronic liver disease (CLD), but its underlying mechanisms may differ according to disease etiology. Quantitative MRI biomarkers, including proton density fat fraction (PDFF) and magnetic resonance elastography (MRE), may help characterize etiology-specific patterns of muscle loss. This study aimed to explore etiology-specific associations between MRI-derived biomarkers and sarcopenia, with a particular focus on metabolic dysfunction-associated steatohepatitis (MASH) and viral hepatitis. Methods: This retrospective single-center study included 131 CLD patients (77 with MASH, 54 with viral hepatitis) who underwent MRI, including PDFF and MRE. Sarcopenia was defined by L2 skeletal muscle index thresholds (<42 cm(2)/m(2) for men, <38 cm(2)/m(2) for women). Muscle identification was performed by automatic threshold-based segmentation by a single observer. Multivariable logistic regression analyses incorporating interaction terms were performed to evaluate whether associations between MRI biomarkers and sarcopenia differed by etiology. Results: Sarcopenia was present in 56% of patients. In the overall cohort, older age (OR = 1.05, p = 0.01), lower PDFF (OR = 0.93, p = 0.03), and lower liver stiffness (OR = 0.51, p = 0.006) were independently associated with sarcopenia. A significant interaction between BMI and disease etiology was observed (p = 0.02). Subgroup analyses suggested that in MASH, sarcopenia was associated with aging, hepatic fat depletion, and lower stiffness. In contrast, in viral hepatitis, it tended to be associated with higher stiffness and lower BMI. Conclusions: MRI-derived hepatic fat and stiffness reflect distinct etiologic patterns of sarcopenia in CLD-metabolically depleted in MASH and fibrosis-related in viral hepatitis. These findings suggest that sarcopenia in MASH and viral hepatitis may reflect different underlying phenotypic patterns, highlighting the importance of considering disease etiology in imaging-based sarcopenia assessment. The results should be interpreted as hypothesis-generating and warrant validation in prospective studies.