| MRI has become a gold standard for quantification of cardiac left ventricular (LV) function. The LV is imaged in the short axis using a 2D breath-hold tine balanced steady state free precession (b-SSFP) sequence. The resultant images are then subject to post-processing in order to assess ejection fraction, stroke volume, and mass. However, several challenges remain to be addressed. Post processing is often done via manual tracing of the systolic and diastolic cardiac phases, which is time consuming and subject to variability. During data acquisition, patients must perform a breath-hold for each short axis slice imaged, which is problematic for patients who are too ill to do perform the task. This work proposes a suite of image acquisition, image reconstruction, and image processing technologies that attempt to address these challenges. They include a novel left ventricular segmentation algorithm that accounts for the effect of partial voxels, a novel radial GRAPPA image reconstruction algorithm, and a self-navigated 3D free-breathing radial pulse sequence. |
