An AI phenotyping study of 14,707 patients with metabolic dysfunction-associated steatohepatitis (MASH) identified cardiovascular, kidney, and hematologic features associated with rapid fibrosis progression, long-term complications, and high healthcare costs. Using diagnoses, procedures, medications, laboratory results, and demographic data from the year before MASH diagnosis, investigators found that anemia and thrombocytopenia showed the largest absolute differences between patients with rapid versus non-rapid fibrosis progression. Cardiovascular conditions, including atrial fibrillation, coronary artery disease, and congestive heart failure, showed some of the strongest relative associations, while acute kidney failure, chronic kidney disease, prior hospitalization, and heart failure–related procedures also distinguished patients with more rapid progression. Cardiovascular and kidney-related signals remained prominent in analyses of subsequent liver and cardiovascular complications and in patients in the highest 10% of healthcare costs.
The findings suggest that MASH risk assessment may benefit from looking beyond liver-specific measures to the broader cardiovascular-kidney-metabolic picture. Patients with kidney disease, cardiovascular comorbidities, diabetes, hypertension, hematologic abnormalities, or recurrent hospitalization may warrant closer attention and more frequent fibrosis assessment, although the study does not establish these factors as causal predictors or validated screening tools. Important limitations include reliance on retrospective EHR and claims data, potentially incomplete records, and diagnosis codes rather than consistent biopsy or imaging confirmation. Further prospective research is needed to determine whether these AI-derived phenotypic signals can reliably improve early identification, monitoring, and clinical decision-making in MASH.
Reference: McCormick B. AI phenotyping links cardiovascular, kidney disease to rapid MASH progression, higher costs. AJMC. 2026 Jul 18.
