Distributions of epicardial adipose tissue in atrial fibrillation using photon counting CT and 3D atrial segmentation: Insights from the Prospective AFAT Study

Abstract Background Previous studies have suggested EAT to be an important contributor to the pathogenesis of atrial fibrillation (AF) as well as a plausible predictor for AF recurrences after ablation. However, data on EAT using precise volumetric quantification is limited. We investigated EAT distributions across both atria as well as possible EAT level differences between AF subtypes. Methods and Results AFAT is a prospective observational study conducted at our Heart Centre. 210 patients underwent first time radiofrequency ablation for paroxysmal (n=93) or persistent AF (n=107) and were included between 2022-2024. Prior to ablation all patients received photon counting atrial cardiac computed tomography (CT) which was then, analysed using 3D segmentation to precisely quantify left atrial (LA) and right atrial (RA) EAT levels as well as wall thickness levels. Mean age was 67.6 ± 9.5 years with 34.3 % (n=66) being female. Mean BMI was 27.5 ± 4.5. The median total biatrial EAT Volume was 62.7 ml (IQR = [48.0 - 78.7] ml) (Median LA EAT 30.45 ml (IQR: 23.54 - 40.07); RA EAT: 31.18 ml (IQR: 23.80 - 37.98). We found a strong linear relationship between left and right atrial fat volumes (correlation coef = 0.911; r2=0.830; p < 0.001) (Figure 1) demonstrating the highlight the uniform distribution of epicardial fat across all cardiac regions. BMI and total EAT levels were also significantly correlated (correlation coef=0.36; r2 =0.133; p < 0.001). When comparing EAT levels adjusted for BMI between the paroxysmal and persistent AF group, we observed significantly higher LA and RA volumes in the persistent AF group compared to the paroxysmal AF group. (LA EAT levels 28.0 ml, IQR = [21.0 - 36.3] ml vs. 32.1 ml, IQR = [25.7 - 42.6] ml (P-value = 0.004) and RA EAT levels 27.7 ml, IQR = [22.6 - 34.8] ml vs. 33.4 ml, IQR = [26.1 - 44.1] ml (P-value = 0.0001), respectively for the paroxysmal vs. persistent AF group (Figure 2). Conclusion EAT levels seem to be equally distributed across bot atria. We observed significantly higher LA and RA EAT levels in patients suffering from persistent AF compared to paroxysmal AF regardless of the underlying BMI.Correlation between LA and RA EAT levels Comparison of EAT levels by AF typ