Patients with DCM/NDLVC have significantly higher epicardial adipose tissue radiodensity (HU -84.6 vs -96.9, p=0.009) and distinct radiomic phenotype (AUROC 0.86) vs controls.
Does the CT-derived radiomic phenotype of epicardial adipose tissue differ between patients with DCM/NDLVC and healthy controls?
Patients with DCM/NDLVC exhibit a distinct epicardial adipose tissue radiomic phenotype with higher radiodensity, suggesting altered EAT composition potentially linked to the underlying cardiomyopathy.
Absolute Event Rate: 0% vs 0%
Abstract Introduction Epicardial adipose tissue (EAT) biology changes in the presence of myocardial disease. Data on EAT in inherited cardiomyopathies are scarce. We explored whether EAT imaging phenotype is different is patients with dilated cardiomyopathy (DCM) and non-dilated LV cardiomyopathy (NDLVC). Aim To evaluate computed tomography (CT)-derived radiomic features of EAT in patients with DCM/ACM and explore potential differences vs. healthy controls. Methods This study included n=23 patients diagnosed with DCM/NDLVC from the Cardiogenetics Unit, 1st Cardiology Department, Hippokration General Hospital, Athens, Greece. All patients underwent coronary computed tomography angiography (CCTA) to exclude significant coronary artery disease. A control group (n=19) comprising individuals without underlying cardiomyopathy who also underwent CCTA was included for comparison. EAT segmentation was performed on CCTA scans. The mean CT attenuation and volume of EAT was calculated. In addition, a total number of 843 radiomic features, including first-order statistics, texture-based parameters, and shape descriptors, were extracted. Principal component analysis (PCA) was applied to reduce dimensionality and identify principal components that capture the most significant variance in EAT radiomic features. The selected principal components were then compared between cardiomyopathy and control groups to explore potential differences. Statistical analysis was conducted to assess the discriminative power of these features. Results The median segmented EAT volume showed no significant difference between patients with DCM/NDLVC and control patients (90.5 58.6, 163.5 vs. 78.7 55.9, 121-6, p=0.5). The median HU values of DCM/NDLVC patients were found to be significantly higher that the corresponding values of control patients, indicating that EAT has higher radiodensity in DCM patients (-84.6 -94.4, -79.9 vs. -96.9 -101.3, -88.5, p=0.009) compared to age- and sex-matched control patients. A logistic regression model incorporating PC1–PC5 of EAT radiomics was developed to classify patients with DCM/NDLVC versus controls (AUROC 0.86). Conclusions The radiomic phenotype of EAT is distinct in patients DCM/NDLVC. Such differences in radiomic textures may indicate altered EAT composition in DCM/NDLVC patients, potentially as a result of underlying cardiomyopathy and inside-to-outside signalling from the heart.Figure 1
Tsampras et al. (Sat,) reported a other. Patients with DCM/NDLVC have significantly higher epicardial adipose tissue radiodensity (HU -84.6 vs -96.9, p=0.009) and distinct radiomic phenotype (AUROC 0.86) vs controls.