A Computational Framework for Personalized Blood Flow Analysis in the Human Left Atrium TOMOHIRO OTANI, 1,3 ABDULLAH AL-ISSA, 1 AMIR POURMORTEZA, 2,4 ELLIOT R. MCVEIGH, 2,5 SHIGEO WADA, 3 and HIROSHI ASHIKAGA 1,2 1 Cardiac Arrhythmia Service, Division of Cardiology, Johns Hopkins University School of Medicine, 600 N. Wolfe Street, Carnegie 568, Baltimore, MD 21287, USA; 2 Department of Biomedical Engineering, Johns Hopkins University School of Medicine, Baltimore, MD, USA; 3 Department of Mechanical Science and Bioengineering, Osaka University Graduate School of Engineering Science, Osaka, Japan; 4 Department of Radiology and Imaging Sciences, National Institutes of Health Clinical Center, Bethesda, MD 20814, USA; and 5 Departments of Bioengineering, Medicine, Radiology, University of California, San Diego, La Jolla, CA, USA (Received 13 October 2015; accepted 7 March 2016) Associate Editor Peter E. McHugh oversaw the review of this article. Abstract—Atrial fibrillation (AF), the most common human arrhythmia, is a marker of an increased risk of embolic stroke. However, recent studies suggest that AF may not be mechanistically responsible for the stroke events. An alter- native explanation for the mechanism of intracardiac throm- bosis and stroke in patients with AF is structural remodeling of the left atrium (LA). Nevertheless, a mechanistic link between LA structural remodeling and intracardiac throm- bosis is unclear, because there is no clinically feasible methodology to evaluate the complex relationship between these two phenomena in individual patients. Computational fluid dynamics (CFD) is a powerful tool that could poten- tially link LA structural remodeling and intracardiac throm- bosis in individual patients by evaluating the patient-specific LA blood flow characteristics. However, the lack of knowl- edge of the material and mechanical properties of the heart wall in specific patients makes it challenging to solve the complexity of fluid–structure interaction. In this study, our aim was to develop a clinically feasible methodology to perform personalized blood flow analysis within the heart. We propose an alternative computational approach to perform personalized blood flow analysis by providing the three-dimensional LA endocardial surface motion estimated from patient-specific cardiac CT images. In two patients (case 1 and 2), a four-dimensional displacement vector field was estimated using nonrigid registration. The LA blood outflow across the mitral valve (MV) was calculated from the LV volume, and the flow field within the LA was derived from the incompressible Navier–Stokes equation. The CFD results successfully captured characteristic features of LA blood flow observed clinically by transesophageal echocardiogram. The LA global flow characteristics and vortex structures also agreed well with previous reports. The time course of LAA emptying was similar in both cases, despite the substantial difference in the LA structure and function. We conclude that our CT-based, personalized LA blood flow analysis is a clinically feasible methodology that can be used to improve our understanding of the mechanism of intracardiac throm- bosis and stroke in individual patients with LA structural remodeling. Keywords—Image-based simulation, Computed tomogra- phy, Computational fluid dynamics, Cardiac mechanics, Left atrium. ABBREVIATIONS AF Atrial fibrillation CFD Computational fluid dynamics CT Computed tomography LA Left atrium LAA Left atrial appendage LV Left ventricle MV Mitral valve PV Pulmonary vein TEE Transesophageal echocardiogram 3D Three-dimensional 4D Four-dimensional INTRODUCTION Atrial fibrillation (AF) is associated with an increased risk of embolic stroke. 32 However, recent studies using extended electrocardiographic monitor- Address correspondence to Hiroshi Ashikaga, Cardiac Arrhyth- mia Service, Division of Cardiology, Johns Hopkins University School of Medicine, 600 N. Wolfe Street, Carnegie 568, Baltimore, MD 21287, USA. Electronic mail: hashika1@jhmi.edu Annals of Biomedical Engineering (Ó 2016) DOI: 10.1007/s10439-016-1590-x Ó 2016 Biomedical Engineering Society