Image-based models of cardiac structure with applications in arrhythmia and defibrillation studies

Abstract

The objective of this article is to present a set of methods for constructing realistic computational models of cardiac structure from high-resolution structural and diffusion tensor magnetic resonance images and to demonstrate the applicability of the models in simulation studies. The structural image is segmented to identify various regions such as normal myocardium, ventricles, and infarct. A finite element mesh is generated from the processed structural data, and fiber orientations are assigned to the elements. The Purkinje system, when visible, is modeled using linear elements that interconnect a set of manually identified points. The methods were applied to construct 2 different models; and 2 simulation studies, which demonstrate the applicability of the models in the analysis of arrhythmia and defibrillation, were performed. The models represent cardiac structure with unprecedented detail for simulation studies.

Fig. 1

The application of the segmentation pipeline to an example image slice. The sequence of rectangular blocks illustrates the results as an example image slice is processed through the steps 1 to 4 in the pipeline.

Fig. 2

Mesh generation: (A) mesh corresponding to the segmented image slice shown in Fig. 1; (B) enlarged view of the small region enclosed by the magenta box in A.

Fig. 3

Assignment of fiber orientations: (A) 2D projection of orientations assigned to the mesh shown in Fig. 2A; (B) enlarged view of the small region enclosed by the magenta box in A.

Fig. 4

The models of normal and infarcted rabbit hearts. In each row, the first column shows the anterior view of the entire model, and the second and third columns show the model split in half along a horizontal long-axis view plane.

Fig. 5

Demonstration of the structural detail that can be obtained using our processing pipeline: (A) the endocardial trabeculations in the basal region of the RV in the normal rabbit model; (B) a papillary muscle that is attached to the mitral valve of the infarcted rabbit model; (C) blood vessels and interlaminar clefts in the normal rabbit model.

Fig. 6

Visualization of fiber tracks, indicative of regionally prevailing cell orientation, in the infarcted rabbit model. (A) Epicardial view of tracks in the entire model; (B) tracks in a slab of tissue between 2 short-axis planes.

Fig. 7

Reconstructing the PS: (A) a slab of the mesh that lies between 2 short-axis planes; (B) the PMJs, PPJs, and tags on a short-axis slice of the image. The inset displays the enlarged view of the small region enclosed in the red box; (C) the reconstructed PS; (D) the PS overlaid on the mesh.

Fig. 8

Simulations of the infarcted rabbit model showing the effect of the infarct core on shock-end virtual electrode polarizations. A, The whole heart model with infarct core (blue), border zone (green), and the plate electrodes. B, The heart cut in half at a coronal plane. C, Transmembrane potential at shock-end of a 1 V/cm left ventricular shock. D, Transmembrane potential at shock-end of a 5 V/cm left ventricular shock. E, Transmembrane potential at shock-end of a 1 V/cm right ventricular shock. F, Transmembrane potential at shock-end of a 5 V/cm right ventricular shock. The color bar for C through F is shown in the bottom panel. Transmembrane potentials of −90 mV or less are colored as −90 mV, and potentials of 20 mV or more are colored as 20 mV.

Fig. 9

Initiation of an apical wave reentry in the rabbit right ventricular free wall. A, Surface of rabbit ventricles with right ventricular wall boxed in. B, Voltage map of second reentrant beat initiated by premature line stimulus in the wake of a propagating wave. C and D, Activation and voltage maps, respectively, of third and final pacing stimulus and subsequent wave propagation. E and F, Activation and voltage maps, respectively, of a reentrant beat after initiation as shown in C and D. The color bar in B applies to D and F also.