Comparison of Data-Driven Phenotyping Systems for Metabolic Dysfunction-Associated Steatotic Liver Disease: Implications for Treatment Response.
文献摘要
Background Although emerging classification systems address the pathophysiological heterogeneity of metabolic dysfunction-associated steatotic liver disease (MASLD), the implications of these subtypes for treatment responses remain unclear. We systematically evaluated three phenotyping frameworks for stratifying responses to lifestyle management. Methods This prospective observational cohort study included 618 patients with MASLD who received 48 weeks of guideline-based lifestyle management. The classification systems assessed were metabolic clusters-mild obesity/dyslipidemia-associated MASLD (MODM), age-related MASLD (AM), severe insulin resistance-associated MASLD, high lipoprotein(a)-associated MASLD (HLPAM), and severe mixed hyperlipidemia-associated MASLD (SMHLM); hepatic and cardiometabolic traits (control, liver-specific, and cardiometabolic); and biochemical patterns (hepatocellular, mixed, and cholestatic). The primary endpoints were responses in magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF) (≥30% relative liver fat reduction), alanine aminotransferase (ALT) (≥17 IU/L decrease), and liver stiffness measurements (LSM) (≥1-stage improvement in the LSM-defined fibrosis stage). Results Metabolic clusters demonstrated significantly greater temporal stability than the hepatic and cardiometabolic trait and biochemical pattern systems (62.8% vs. 56.6% vs. 53.4%, P =0.003). The HLPAM cluster demonstrated an attenuated MRI-PDFF response (odds ratio [OR], 0.35; 95% confidence interval [CI], 0.15 to 0.79), while the AM cluster showed a lower ALT response (OR, 0.33; 95% CI, 0.18 to 0.62), and the MODM cluster displayed a superior LSM response compared to the others (all P Conclusion Metabolic cluster-based MASLD subtypes exhibited differential responses to lifestyle management, supporting the integration of metabolic subtyping into clinical algorithms to optimize management.