Po-01-210 stroke risk is identified by slow blood flow and stagnant blood particles in the left atrium

Atrial fibrillation (AF) is the most common type of arrhythmia, causing irregular contraction patterns in the atrial chambers. This rhythm disorder may result in stagnant flow, leading to the formation of blood clots, especially in the left atrial appendage (LAA), and causing a thrombotic stroke. Current stroke risk assessment in AF patients is done using CHADS2 or CHA2 DS2–VASc scores, which are based on selected clinical characteristics, without accounting for potential risk factors such as LAA thrombi formation. Thus, there is an urgent need to develop personalized approaches to stroke risk prediction that are based on understanding the mechanistic hemodynamic consequences of AF. To quantify thrombotic stroke risk using hemodynamic indicators for disturbed blood flow in the left atria by means of personalized fluid dynamics modeling. Our patient cohort consists of subjects who experienced a thrombotic stroke and those who did not. In a novel approach, we merge static 3D LGE MRI images with CINE MRI data to accurately construct the LA endocardial geometry over time. This allowed us to construct a framework to simulate, quantify and compare left atrial hemodynamics in these patients in terms of multiple indicators such as flow stasis, time-averaged wall shear stress, relative residence time, and mean age of blood particles (see Figure). Patients who experienced stroke exhibited regions of the left atrium with stagnant blood flow, especially in the LAA. The comparison of mean age of blood particles and flow stasis revealed a distinct division between the no-stroke and stroke groups; the stroke group exhibited larger values for both indicators, promoting optimal conditions for clot formation. Furthermore, we investigated the importance of including the atrial contraction in the model in terms of stroke risk assessment. We found that the usage of CINE MRI data in our personalized fluid dynamics model is crucial since it significantly improves our results when compared with a simplified model with a rigid atrial endocardium. Fluid dynamics modeling showed that stroke risk is strongly correlated with an abnormal rinsing of the left atrium, promoting the stagnation of particles, especially in the LAA. The development of personalized frameworks, with the inclusion of functional data as the endocardial displacement, can distinctively quantify atrial hemodynamic anomalies, allowing for accurate prediction of stroke risk.