Abstract:
This study was conducted to align in-vivo cardiac tagged magnetic resonance (MR) images and ex-vivo diffusion tensor (DT) MR images of a beagle´s heart provided by our collaborators at NIH. Such data will be used in a mathematical modelling framework in order to analyze regional left ventricular (LV) mechanics, which will provide important insight to pathological conditions such as myocardial infarction. DT-MR data provide the myocardial fiber orientations for each pixel of an image. The tagged MR and the DT-MR images were orientated and scaled differently, since the shape of the heart was different when imaged in-vivo versus ex-vivo. Image re-sampling was required to obtain new DTMR images with the heart in the same orientation as the tagged images. A graphical user interface was developed for the segmentation of the surface contours and readily identifiable landmark points on the re-sampled DT-MR images, together with the corresponding landmark target points on the tagged MR images. Based on this geometrical data, a free-form deformation is used to warp the fiber orientation data from the DTMR images to an end-diastolic geometrical model determined from the tagged MR images. Cardiac mechanics simulations will be validated against data obtained from the tagged images throughout the cardiac cycle. We conclude by assessing the effectiveness of using DT fiber orientation data for cardiac mechanics modelling. This type of modelling will be useful for analyzing cardiac mechanics during diseased states and will be extended for biventricular modelling.
