Atrial anatomy alone significantly influences atrial fibrillation complexity in-silico

Abstract Background and Aims Current therapies for atrial fibrillation (AF) rely on "one-size-fits-all" approaches that do not consider inter-patient differences. The mechanisms driving variability in AF inducibility and the role of the individual anatomy on AF inducibility are poorly understood. This study uses a recently developed patient-specific computer model of AF with realistic 3D bundle architecture to study how anatomical variability alone may affect AF inducibility and complexity. Methods Patient-specific atrial wall anatomies were derived from late gadolinium-enhanced magnetic resonance imaging (LGE-MRI) to create personalized models for 10 patients undergoing catheter ablation for AF. Detailed intra- and inter-atrial structures, such as pectinate muscles, Bachmann's bundle, fossa ovalis, coronary sinus, and fiber orientations, were incorporated through a novel approach by using a highly detailed reference model as a template. All models were assigned identical electrophysiological properties, with optional endomysial fibrosis mapped according to LGE-MRI distributions. Incremental pacing was applied from 20 positions in each model, without and with fibrosis, to quantify the inducibility of AF and of supra-ventricular tachycardias (SVTs, including AF and macro-reentrant tachycardias). AF complexity was assessed by measuring functional reentry formation and termination rates, analyzed using renewal theory to estimate each patient's average number of simultaneous reentries. Results All 10 personalized models were successfully generated (Figure 1, left), with low fibrotic content in all but one patient (Utah stages <2). AF initiation rates showed no significant variation across models, regardless of fibrosis (Chi-squared test, 50.0% [IQR: 36.25%-53.75%] without fibrosis, p=0.45; 45.0% [41.25%-55.0%] with fibrosis, p=0.36). However, SVT initiation rates varied significantly only in models with fibrosis (57.5% [47.5%-60.0%], p=0.04, Figure 1, right). Reentry dynamics differed significantly across patients, revealing an impact of the anatomy on AF complexity irrespective of fibrosis. Reentry formation rates varied significantly among patients both without (ANOVA, p<0.01) and with fibrosis (p<0.03). Reentry termination rates varied only in the presence of fibrosis (p<0.001). These differences combined led to substantial variability in the expected number of simultaneous reentries among patients (ANOVA, p<0.001 without fibrosis; p<0.05 with fibrosis, Figure 2). Conclusion Anatomical differences alone led to variability in AF reentry dynamics, contributing to inter-patient differences in SVT susceptibility and AF complexity irrespective of the presence of fibrosis. This variability highlights the importance of considering anatomical features in AF management and could support the development of more personalized therapeutic approaches.Fig. 1:model examples and initiation Fig. 2:AF complexity per patient model