Abstract Body (Do not enter title and authors here): Introduction Since pulmonary vein isolation (PVI) is often insufficient for persistent atrial fibrillation (PsAF) treatment, rotor ablation targeting AF drivers/substrates has been conducted as potential effective strategy. Recently, substrate location capable of attracting rotor (LR) has been reported to be associated with local fibrosis features. Meanwhile, epicardial adipose tissue (EAT) has been shown to be linked to developing fibrosis and arrhythmogenic substrates. Currently it is not well understood how EAT is related to the LR, and local fibrosis features. Personalized LGE-MRI-based digital twins (DTs) and cardiac CCT can be employed to investigate this.
Hypothesis EAT detected in CCT is correlated with LRs having specific local fibrosis features.
Aim To investigate the correlation between EAT, LRs, and the corresponding local fibrosis features using the merging of LGE-MRI-based DTs and CCT.
Methods Both LGE-MRI and CCT were acquired in 19 consecutive PsAF patients. Bi-atrial DTs incorporating the fibrosis distribution were created, wherein, following virtual PVI, LRs were identified by rapid pacing from 40 bi-atrial sites near fibrosis. In silico ablation was repeated until non-inducibility of rotors (Fig). EAT segmented from CCT and DT were compared after merging them using the bi-atria and 4 PVs as reference points. The distance from each tissue type in DT to the closest EAT (DCE) was calculated. The correlation between local fibrosis density feature (FD) and the median DCE at LRs was analyzed. For each LR, both volumetric FD (vFD) and endo-epi surface FD (sFD), subtracting epicardial surface FD from the endocardial surface FD to measure the dissociation of Endo-Epi surface FDs, were calculated.
Results Among 17 DTs excluding 2 without qualified CCT images, 15 DTs had 95 extra-PVI LRs in total. The DCE of fibrotic tissues (16%) was significantly shorter than that of normal tissues (84%) (p<0.001). The DCE of tissues within LRs (2%) was also significantly shorter than outside LRs (98%) (p<0.001). The mean DCE among 95 LRs was 4.1 mm. The vFD of the LRs with DCE <4.1 mm (56 LRs) was comparable to that of LRs with DCE ≥4.1 mm (39 LRs) (p=0.543). However, the sFD was significantly higher for LRs with DCE ≤4.1 mm (p=0.010).
Conclusions EAT is associated with fibrotic tissue and LR. Higher sFD suggests shorter DCE is linked to higher FD on the endocardial surface at LRs. Detecting EAT with CCT could be useful for evaluation and ablation of PsAF substrate.