| Cardiovascular disease is a major cause of death in the civilized world. Magnetic resonance imaging (MRI) is a powerful non-invasive method for the quantification of heart motion and function, which provide valuable information about the health of the heart. Harmonic phase (HARP) is a promising technique using MRI for automatic and fast quantification of cardiac motion that has facilitated clinical cardiac function MRI studies. This dissertation addresses both computational and imaging improvements related to the HARP methodology.; One of the limitations of HARP is its inadequate computational speed even though a real-time HARP MR imaging protocol exists. In this dissertation, we developed an integrated 2D imaging system that dramatically improves the computational engine of HARP, making it 50 times faster than before, now capable of keeping up with the real-time imaging protocol. The system was validated and demonstrated on a human subject experiment using a commercial clinical MRI scanner.; Another limitation of the HARP approach is its inability to capture 3D motion and strain from single image planes. In the second contribution of this dissertation, we developed, and validated a new imaging method and algorithm, collectively called zHARP, which enables rapid 3D motion imaging and quantification from a single image slice and 3D strain measurements from adjacent pairs of slices.; For further improvements in the HARP methodology, we developed a nonlinear smoothing algorithm for reducing artifacts in strain measurement due HARP processing, a new strain calculation formula using zHARP, and a novel single breath-hold imaging sequence that suppresses HARP-related interference and thus allows HARP to use a larger amount of data.; The imaging techniques and algorithms presented in this dissertation were evaluated and validated in moving phantom and human studies and support improved heart motion measurements and allow faster, more comprehensive, and more accurate whole heart myocardial strain measurements and modeling. |
