Fluid-dynamics modelling of the human left ventricle with dynamic mesh for normal and myocardial infarction: Preliminary study S.S. Khalafvand a , E.Y.K. Ng a,n , L. Zhong b , T.K. Hung c a School of Mechanical and Aerospace Engineering, College of Engineering, 50 Nanyang Avenue, Nanyang Technological University, Singapore 639798, Singapore b Department of Cardiology, National Heart Centre, Mistri Wing 17, 3rd Hospital Avenue, Singapore 168752, Singapore c Department of Bioengineering, University of Pittsburgh, Pittsburgh, PA 15261, USA article info Article history: Received 25 February 2011 Accepted 23 June 2012 Keywords: Left ventricle CFD MRI Dynamic mesh Vortices Pressure difference abstract Pulsating blood flow patterns in the left ventricular (LV) were computed for three normal subjects and three patients after myocardial infarction (MI). Cardiac magnetic resonance (MR) images were obtained, segmented and transformed into 25 frames of LV for a computational fluid dynamics (CFD) study. Multi-block structure meshes were generated for 25 frames and 75 intermediate grids. The complete LV cycle was modelled by using ANSYS-CFX 12. The flow patterns and pressure drops in the LV chamber of this study provided some useful information on intra-LV flow patterns with heart diseases. & 2012 Elsevier Ltd. All rights reserved. 1. Introduction Heart failure is one of the greatest killers in the world. Analysis of intra-ventricular blood flow patterns could provide under- standing to the complex relationships for treatments of LV dysfunction. In recent years, researchers reported various models for studying flow patterns in the left ventricle. Hunter et al. [1] proposed to establish a computational framework for all heart functions including structure and cardiac cells behaviour. How- ever, the lack of knowledge about linking the different functions in heart made it difficult to achieve this final goal of LV modelling. Cheng et al. [2] reviewed and classified numerical modelling into three types: geometry-prescribed CFD methods [3–8], fictitious and realistic Fluid–Structure Interaction (FSI) methods [2,9–13]. Geometry-prescribed methods employed MRI data of LV motion as boundary conditions for CFD with dynamic mesh for the movement of boundary structure. Both fictitious and realistic FSI methods were aimed at modelling the interaction of fluid and structure [14]. The ambitious task of modelling the cardiac and valvular motions is much more complicated and appears to be impractical for patient study. Saber et al. [4] studied a patient-specific model of LV and used CMR tools [15] for reconstruction of realistic geometries of LV. Baccani et al. [5] simulated diastolic blood flow in an axisymmetric model of LV. Long et al. [6] used prescribing boundary conditions and concluded that the CFD simulation was highly sensitive to the boundary conditions imposed during diastolic filling. Domenichini et al. [16] reported a numerical study of the three-dimensional diastolic flow in a prolate spheroid with moving wall. A patient-specific modelling with MRI was reported by Long et al. [7] for LV flow patterns in normal subjects. Schenkel et al. [8] presented a three-dimensional CFD method to simulate LV blood flow. Although 3D modelling is more realistic than 2D analyses, the input data are enormous if the twisting components of cardiac motion are included in the analysis. As 3D pulsating blood flow processes are complicated to quantify and present graphically, 2D models are useful and practical to capture the main characteristics of the flow phenomena. In the present study, patient specific 2D geometries were used to compute velocity and pressure fields. The combination of non- invasive magnetic resonance imaging (MRI) and CFD has been well recognised as an effective method to study the complex dynamics of the cardiovascular system, such as diastolic and systolic blood flows in a left ventricle. 2. Methodology 2.1. MRI scans and data processing This study involved three normal subjects and three patients after myocardial infarction. All subjects underwent diagnostic Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/cbm Computers in Biology and Medicine 0010-4825/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compbiomed.2012.06.010 n Corresponding author. E-mail address: mykng@ntu.edu.sg (E.Y.K. Ng). Computers in Biology and Medicine 42 (2012) 863–870