Association of deep learning-derived epicardial fat volume with target organ damage in subjects with nonobstructive coronary artery disease.
Authors
Affiliations (4)
Affiliations (4)
- Department of Radiology, Seoul Metropolitan Government Seoul National University Boramae Medical Center, Seoul National University College of Medicine, Seoul, Korea.
- Department of Internal Medicine, Seoul Metropolitan Government Seoul National University Boramae Medical Center, Seoul National University College of Medicine, Seoul, Korea.
- Department of Radiology, Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seongnam, Korea.
- Department of Radiology, Seoul Metropolitan Government Seoul National University Boramae Medical Center, Seoul National University College of Medicine, Seoul, Korea. [email protected].
Abstract
Epicardial fat exerts both protective and deleterious effects on organs through diverse cytokine-mediated pathways. This study aimed to investigate computed tomography (CT)-based indexed epicardial fat volume (EFVi) in association with target organ damage parameters. The prospectively enrolled cohort of 75 patients with nonobstructive coronary artery disease underwent electrocardiogram-gated CT and was evaluated for target organ damage parameters: estimated glomerular filtration rate, proteinuria, echocardiographic septal e' velocity, E/e' and tricuspid regurgitation velocity, brachial-ankle pulse wave velocity, and ankle-brachial index. EFVi was measured from semiautomated 3D segmentation of electrocardiogram-gated CT. Partial correlation, multiple linear regression, and receiver operating characteristic (ROC) analyses were conducted. Age and EFVi showed moderate positive linear correlation (r = 0.567, P < 0.001). After adjusting for age, EFVi was significantly correlated with the septal e' velocity (r = - 0.489, P < 0.001) and E/e' (r = 0.256, P = 0.034), but not with other target organ damage parameters (P > 0.05). Multiple linear regression analysis showed that the correlations of the EFVi with the septal e' velocity (β = -0.0003, P = 0.007) and E/e' (β = 0.0606, P = 0.024) remained significant after adjusting for potential confounders. ROC analysis identified optimal EFVi thresholds: 95.78 cm<sup>3</sup>/m<sup>2</sup> for reduced septal e' velocity (area under the ROC curve [AUC], 0.750; sensitivity, 88.2%; specificity, 56.8%) and 91.68 cm<sup>3</sup>/m<sup>2</sup> for elevated E/e' (AUC, 0.692; sensitivity, 71.4%; specificity, 64.8%). EFVi was related to left ventricular diastolic function more than other target organ damage parameters, including renal function and arterial stiffness, which suggests that the epicardial fat may have a role in the pathogenesis of left ventricular diastolic dysfunction.