The Impact of Anatomical Variability on Atrial Fibrillation Dynamics: a Novel Personalized In-Silico Approach

Aims:We aimed to assess the effects of anatomical variability on atrial fibrillation (AF) initiation and complexity through a novel patient-specific model strategy incorporating detailed atrial structures. Methods: Personalized models from 10 patients undergoing AF ablation were created by combining atrial anatomies extracted from magnetic resonance imaging with detailed structures mapped from a reference model. Pectinate muscles, Bachmann’s bundle, coronary sinus, and fiber orientations were mapped from the reference to the patient models using an extended implementation of the universal atrial coordinates algorithm. All models had identical electrophysiological properties. We assessed the inducibility of AF and macroreentrant tachycardia in each model after incremental pacing (20 sites). For AF, complexity was analyzed via the functional reentry formation/termination rates and the estimated number of simultaneous reentries using the renewal theory. Results: AF initiation rates did not vary significantly across models (50.0% [IQR: 36.2%-53.8%], p=0.43). While reentry termination rates were similar across models, formation rates varied significantly among patients (1.28 [0.66; 1.68] %/ms, p<0.01), leading to variability in the number of co-existing reentries and in the resulting AF complexity. Conclusion: Anatomical differences alone caused inter-patient variability in the AF susceptibility and complexity.