| Coronary artery disease (CAD) affects myocardial function through reduced blood flow, ischemia, and infarction. Various imaging techniques exist for the study of the coronary vasculature and cardiac function, however no single method provides extensive information on both. In particular, computed tomographic angiography (CTA) provides high resolution images of the coronaries and cardiac anatomy, while magnetic resonance (MR) provides myocardial anatomical and functional information. The ability to combine and visualize both sets of information can lead to a better understanding of the CAD process.;This work introduces a method for the coregistration of cardiac CTA and MR data sets based on the myocardial epicardial surface. The data sets were registered temporally based on the electrocardiogram recorded during the CTA acquisition and the timing characteristics of the MR acquisition. A 3D semi-automated segmentation algorithm was developed and evaluated on clinical data. The results of the segmentation (epicardial volume) revealed a good correlation (CTA: R = 0.994, MR: R = 0.985) with manual outlines, however the limits of agreement were large for the MR volumes (+/-44.9 mL).;The coregistration of the CTA and MR data sets was achieved by implementing the iterative closest point (ICP) algorithm. The objective of the ICP approach is to minimize the distance between two surfaces defined by two sets of points (the epicardial surfaces). The resulting spatial transformation was used to resample the CTA data based on the MR sampling scheme. The technique was applied to data obtained on eleven patients. Two experts compared and evaluated the CTA reformats to the MR images. The majority of the cases were acceptable (reader 1: 45% perfect, 55% acceptable, 0% fail; reader 2: 55% perfect, 36% acceptable, 9% fail). There were no statistically significant differences in the results obtained from the manual and semi-automated outlines, although reader 2 revealed a preference for the manual results.;Some potential applications of this technique were investigated. The coronary vasculature was segmented in the CTA data and combined with the MR data in a 3D rendered scene. This approach was also tested in a dynamic framework over the systolic phase of the cardiac cycle. |
