In silico estimation of thrombogenic risk after left atrial appendage excision: Towards digital twins in atrial fibrillation

Background & aim: The left atrial appendage (LAA) is a highly variable, pouch-like structure in the left atrium prone to thrombus formation, especially in atrial fibrillation (AF) patients. In silico cardiac models can help characterize the LAA's complex morphology and hemodynamics, aiding in identifying pro-thrombotic areas. This study assessed atrial hemodynamics and thrombus formation risk after LAA excision and compared with optimal synthetic excisions and occluder placements in high thrombogenic-risk cases. Methods: We included 33 patients from the MARK-AF study who had persistent AF and underwent excision of the LAA. We quantified the morphological characteristics of the post-excision LAA remnant. With patient-specific atrial geometries and boundary conditions, in silico blood flow simulations were performed. For each patient, we quantified multiple in silico indices to characterize blood flow patterns and identify thrombogenic regions. We performed an in silico comparison of different LAA treatment approaches. Results: In our cohort, 25/33 (76 %) of patients had a post-excision, protruding LAA remnant (LAA depth >10 mm). In silico simulations indicated that patients with a protruding remnant more frequently showed unfavorable values for in silico indices associated with high thrombogenic risk at the excision site. However, a prominent LAA remnant was not the only factor associated with a high thrombogenic risk. An optimal excision or optimal occluder device placement reduced thrombus formation risk. Conclusion: The combination of LAA remnant morphology and hemodynamics contributed to thrombus formation risk. Advanced in silico simulations uniquely enabled the comparison of different therapies, until now only centered on device occluders, contributing to digital twins in AF.