MR phase and susceptibility-weighted imaging of iron deposition in multiple sclerosis and radiation-treated brain tumors.

摘要

Magnetic Resonance Imaging is a non-invasive imaging technique that is widely used in medicine. Conventionally, MRI contrasts rely on differences in longitudinal or transverse relaxation times of different tissues and the resulting images display primarily anatomical information of the tissues. With progress in MRI techniques, image contrasts that provide functional or physiological information are now available. Tissue susceptibility is one of this kind and recently has been gaining interest in MRI, especially at high MR field strengths. To generate susceptibility contrast, MR phase and susceptibility-weighted imaging (SWI) are currently two imaging modalities that are used most often. In this work, we focused on using these imaging methods to study abnormal iron accumulation in multiple sclerosis (MS) and radiation-treated brain tumors. We first report in this dissertation a serial phase imaging study of chronic MS lesions in which the phase contrast, presumed due to iron deposition in the lesions, was investigated longitudinally. The observations from the study contribute to a better knowledge of the mechanism of the phase contrast in MS lesions and their evolution. Then we present a comparison study of SWI of iron-containing cerebral microbleeds (CMBs) between 3 Tesla and 7 Tesla MR scanners for patients who had brain tumors and received radiation therapy. This study was aimed at knowing how much sensitivity gain can be achieved when choosing 7T over 3T for detection of CMBs. Followed by the study, a gradient-echo sequence with multiple echoes is introduced, which is able to acquire MR angiography and susceptibility-weighted images simultaneously. This sequence provides a way to characterize CMBs together with veins and arteries in the brain. In addition, by integrating data from several echoes, the SWI can be made more flexible and its imaging quality of CMBs can be improved compared to sing-echo SWI. Finally, an automated CMB detection algorithm is developed, which is able to identify CMBs on images from SWI with a high sensitivity. Its high accuracy and fast speed significantly reduces radiological time burden in identifying CMBs, which will speed up the exploring of the clinical relevance of CMBs.