Electronic Thesis and Dissertation Repository


Master of Science


Medical Biophysics

Collaborative Specialization

Molecular Imaging


Dr. Donna Goldhawk


Lawson Research Institute

2nd Supervisor

Dr. Neil Gelman


Lawson Research Institute



Myocardial hemorrhage (MH) may occur following myocardial infarction (MI). Myocardial hemorrhage can be assessed by image-based measurements of R2* (1/T2*) which are sensitive to iron-based byproducts of hemorrhage. However, automated detection of MH throughout the entire left ventricle (LV) is challenging due to R2* contributions from magnetic field distortions (susceptibility artifact). To provide insight for improved MH detection, we analyzed canine cardiac magnetic resonance images that had previously been acquired at baseline and several time points (3 to 41 days) following experimentally induced MI. The images were processed to generate R2* maps and cardiac polar plots. A baseline composite polar plot was generated to identify a large LV region (~ 75% of LV) with relatively low and uniform R2* (suggesting low susceptibility artifact). The probability distribution of R2* values on the cumulative and individual polar plots was analyzed to provide insight for future development of automated analysis of MH. Also, the boundaries of this uniform region were used to define the region over which post-MI polar plots were analyzed. A threshold (0.08 ms-1) to differentiate abnormally high R2* regions from normal R2* regions was determined from baseline polar plots. The mean R2* values of normal regions in the post-MI images remained approximately constant as a function of time post-MI and close to the mean R2* from baseline images. The mean R2* values of abnormal regions appeared to peak at day 3 post-MI but remained elevated compared to the remote region (0.04 ms-1): day 3 (0.11 ms-1, p < 0.001), day 7, 15, 21 and 41 (0.08 ms-1, p < 0.05). At all post-MI times investigated, these data suggest that hemorrhage and its byproducts remained for several weeks.