Atrial fibrillation (AF) currently afflicts more than 3 million people in the USA and this number is expected to rise to 7.56 million by 2050. AF is the most common sustained arrhythmia, especially among elderly patients, costing the health care system ~$6 Billion each year. AF patients have 3-5 folds higher risk of stroke. All thromboprophylaxis decisions require physician assessment of the patients stroke risk, which is currently determined by clinical data included in the CHA2DS2-VASc score. However, we remain very limited in predicting who will have a stroke in the setting of AF. The left atrial appendage (LAA) has been shown to be a culprit location for formation of thrombi in non-valvular AF (91%) and in valvular AF (50%). This suggests a possible correlation between the LAA shape and stroke risk. We propose two novel, patient-specific indices to improve the stroke prediction in AF patients: a) A hemodynamics-based calculation of the scalar residence index (SRI) in LAA; b) A generalized approach to quantify the LAA appearance/shape. To achieve these goals, we will pursue the following specific aims: 1) To further develop the SRI, including the evaluation of potentially confounding variables (waveform shape and fluid viscosity). 2) To further develop indices of LAA appearance/shape. 3) To improve CHA2DS2-VASc-based stroke risk stratification using LAA SRI and LAA appearance indices.We will use cardiac computed tomography images from 157 patients with AF (57 of these patients had prior stroke). Images will be segmented and prepared for computational fluid dynamics (CFD) analysis. SRI for each patient will be calculated using patient-specific left atrium/LAA geometry and a CFD-based hemodynamic model. In addition, features of the LAA appearance/shape will be quantified using two novel indices, LAA complexity and balloonability.The calculated SRI and appearance/shape indices, together with the CHA2DS2-VASc score, will be used as the inputs of classifiers to predict stroke. The overall hypothesis is that adding LAA SRI and LAA appearance/shape indices will significantly improve stroke risk stratification in AF patients as compared to the stratification based solely on the CHA2DS2-VASc score, thus enhancing clinical decision making.
|Program type||Predoctoral Fellowship|
|Effective start/end date||01/01/2020 → 12/31/2021|