2621-P: Associations between MRI-Derived Visceral Adipose Tissue and Incident Cardiometabolic Outcomes in a Large Population Cohort

Introduction and Objective: Visceral adipose tissue (VAT) is a key driver of cardiometabolic risk, including insulin resistance and type 2 diabetes (T2D), with growing relevance for risk stratification and treatment response in obesity and cardiometabolic disease trials. This study used large-scale UK Biobank imaging data to evaluate the association between MRI-derived VAT, measured using both single-slice and volumetric approaches, with incident cardiometabolic clinical outcomes. Methods: VAT was segmented automatically from 3D Dixon MRI using U-Net models to derive whole-volume VAT, while cross-sectional VAT area was extracted from a single axial slice at the L3 vertebral level. Cox proportional hazards models (adjusted for age, sex, BMI) were used to estimate associations with incident type 2 diabetes (T2D), liver-related hospitalisation, major adverse cardiovascular events (MACE), and chronic kidney disease (CKD). Results: A total of 75,331 UK Biobank participants were included (median age 66 years, 52% female, median BMI 26 kg/m2, 31% had preclinical obesity, and 38% had clinical obesity). Single-slice VAT showed strong associations with incident T2D (HR 1.82, 95% CI 1.57-2.11) and liver-related hospitalisation (HR 1.56, CI 1.37-1.78), with more modest associations for MACE (HR 1.17, CI 1.10-1.24) and CKD (HR 1.16, CI 1.02-1.31). Volumetric VAT showed comparable associations across outcomes. Conclusion: Both single-slice and volumetric MRI-derived VAT measurements were independently associated with major cardiometabolic outcomes. Volumetric VAT provides a comprehensive reference-standard assessment, while single-slice VAT demonstrated comparable prognostic performance, supporting its use as a complementary, scalable approach that can be incorporated within existing MRI protocols. These findings support MRI-based VAT as a clinically meaningful imaging biomarker for large population studies and obesity trials to improve risk stratification beyond BMI. Disclosure M. Nowak: Employee; Current; Perspectum Ltd. Y. Deng: Employee; Ended; Perspectum Ltd. L.F. Cardiel Castro: None. A. Mermekli: Employee; Current; Perspectum. H.B. Thomaides Brears: Employee; Current; Perspectum Ltd. Stock/Shareholder; Current; Perspectum Ltd.