Improvements on Whole-Heart Coronary Magnetic Resonance Angiography.

摘要

CAD is the leading cause of death in the United States. The current gold standard for diagnosing CAD, x-ray angiography, is a costly and invasive procedure that involves ionizing radiation. Whole-heart coronary MRA is a promising non-invasive technique for detecting significant (>50%) coronary stenosis. However, its clinical application remains limited due to unresolved challenges including low resolution, long scan time, complex scanning procedure and variable performance. This dissertation aims to address some of these challenges.;First, a 3DPR based, self-navigated retrospective respiratory motion correction framework was developed to achieve 100% scanning efficiency without the need of setting up a diaphragm navigator and prospective respiratory gating. The proposed method offered excellent image quality, comparable to those from both conventional navigator gating and data binning using navigator, at (1.0 mm)3 spatial resolution and around 7 minute of scan time.;Second, a non-Cartesian sensitivity encoding framework with self-calibration and motion correction was developed to suppress the streaking artifacts and further shorten the scan time. Excellent image quality was observed with as few as 10,000 projections, corresponding to a scan time of around 5 minutes. No significant difference in image quality was found between the 10,000 projection images and those reconstructed from 20,000 projections.;Third, the 3DPR-based motion correction and image reconstruction framework was applied to contrast-enhanced coronary MRA at 3T with inversion-recovery prepared spoiled GRE. Healthy volunteer studies (N=10) demonstrated the feasibility of a 5- minute scan time and its potential SNR and CNR advantages over longer scan times. Initial patient results (N=1) showed excellent agreement with x-ray angiography and areas of myocardial enhancement in accordance with 2D LGE images.;Last, an ECG and navigator-free 4D coronary MRA technique was developed for simultaneous cardiac anatomy and function visualization. A contrast-enhanced, ungated spoiled GRE sequence with SG and 3DPR was used for image acquisition. Data was binned into different cardiac and respiratory phases based on PCA of the multi-channel SG data. Respiratory motion was corrected separately for each cardiac phase. Cardiac and respiratory SG performed reliably for all subjects. The reconstructed 4D images offered favorable performance in LV function assessment and coronary artery visualization.