Radiomic characteristics of epicardial adipose tissue in patients with coronary artery disease
K Krauz, I Poprawa, J Zmuda, K Momot, M Zarebinski, D Gralak, M WojciechowskaAbstract
Background
Epicardial adipose tissue (EAT) has been implicated in the pathogenesis of coronary artery disease (CAD), as it may modulate atherosclerotic plaques formation through changes of local metabolic and inflammatory microenvironment. EAT volume has been previously linked with the presence of CAD. In recent years, there has been increasing interest in radiomics which allows for quantitative representation of EAT heterogeneity, which may improve chances to capture biologically significant tissue alterations.
Purpose
The aim of this study was to compare the radiomic features of EAT between patients with and without CAD.
Methods
A total of 40 individuals undergoing coronary computed tomography angiography (CCTA) were recruited to this study. According to the presence of atherosclerotic plaques in the coronary arteries, the patients were stratified into two groups: CAD (n = 20) and no-CAD (n = 20). Pericardial borders were manually segmented on non-contrast-enhanced scans. The attenuation threshold -200 to -50 HU was applied, and the resulting EAT masks were used for radiomic data extraction including: Shape, First Order, Gray Level Co-occurrence Matrix (GLCM), Gray Level Run Length Matrix (GLRLM) and Gray Level Size Zone Matrix (GLSZM) features.
Results
Patients characteristics and summary of selected laboratory findings were presented in Table 1 and Table 2, respectively. The EAT volume was significantly higher in the CAD group (Figure 1). We found that studied groups differed in terms of 2 out of 88 extracted radiomic features. Energy and Gray Level Variance were significantly higher in no-CAD group (Table 3).
Conclusions
Radiomic characteristics of EAT may help identify potential biomarkers associated with CAD. Decreased First Order Energy in CAD patients may reflect generally less uniform voxel intensities across EAT. However, reduced GLSZM Gray Level Variance may indicate lower heterogeneity between local intensity zones. Consistently with previous data, the presence of CAD was associated with higher EAT volume.Figure