Medical Physics Seminar – Monday, November 12, 2012
Flow in Radially Acquired Magnetic Resonance Imaging: Methods and Applications in Atherosclerotic Disease
Andrew Wentland (student of Drs. Oliver Wieben and Thomas Grist)
Research Assistant, Department of Medical Physics, UW-School of Medicine & Public Health, Madison, WI - USA -
A number of flow-related atherosclerotic biomarkers have been derived from phase contrast (PC) magnetic resonance imaging (MRI) techniques. Pulse wave velocity (PWV) is one such biomarker of vascular stiffening. As arteries are stiffened by the growth of atherosclerotic plaques, the pulse wave of blood down the aorta is accelerated because of the loss of elastic recoil in the vessel, leading to an elevated PWV. Wall shear stress (WSS) is another biomarker of atherosclerosis that represents the drag forces on vessel walls. It is hypothesized that abnormal blood flow patterns, and in turn insufficient stress, lead to dysfunction of the endothelium. Endothelial dysfunction leads to vascular remodeling, such as adaptive intimal thickening, and may progress to advanced atherosclerotic lesions. However, measurements of PWV require high temporal resolution and measurements of WSS require high spatial resolution. Such resolution requirements are difficult with PC MRI due to the lengthy scan times needed for velocity and reference encoding. Scan times are often reduced for PC MR imaging through the sacrifice of spatial and/or temporal resolution, which would compromise the utility and accuracy of PWV and WSS measurements.
The work in this thesis aims to accelerate phase contrast imaging for the sake of acquiring and measuring PWV and WSS. Two-dimensional PC imaging with radial undersampling is investigated to determine the degree of undersampling that permits accurate flow measurements. Experiments in both flow phantoms and healthy volunteers demonstrate that accurate flow measurements can be obtained even with high undersampling factors. To measure WSS and PWV, a four-dimensional PC imaging technique was employed. A PWV tool was developed and a WSS tool was utilized to handle the unique 4D data sets. A McKinnon-Bates sparsification algorithm was developed and implemented to further accelerate the acquisition of PC data. It is demonstrated that this algorithm reduces artifacts in PC images and allows for greater undersampling factors.