Stroke Risk Assessment in Atrial Fibrillation via Phase-Field Modeling of Thrombus Biomechanics

Atrial Fibrillation (AF), the most common arrhythmia, is linked to one-third of all thromboembolic strokes. AF-related strokes are typically ischemic and cardioembolic, often fatal or leading to disability, with a high risk of recurrence. Despite the well-established correlation between AF and stroke, its underlying mechanisms remain poorly understood. Structural changes and eventual instabilities in a clot can give rise to microthrombi, potentially releasing emboli and significantly impacting stroke risk. To address this, we introduce a computational pipeline that investigates clotting biomechanics and thromboembolism within a unified mathematical framework. Our approach employs a phase-field model to represent the thrombus system as a continuum undergoing deformation and incorporates information from high-resolution, time-resolved medical imaging to track the thrombus behavior over time in a patient-specific manner. Coagulation cascade pathways are also integrated to mimic thrombus initiation. We test our pipeline with ground-truth data simulations in both idealized, fixed-wall geometries and patient-specific, moving-wall left atrial meshes. We demonstrate its clinical relevance using 4D CT acquisitions from clot- and/or stroke-positive AF patients.