| Description |
The spinal cord provides the major pathway for the signal transmission between the brain and peripheral nervous system. Any injury on the spinal cord may disrupt the signal transmission partially or completely, and lead to the permanent disability of the patient. Therefore, a technique which can evaluate the spinal cord disease burden and monitor the treatment progress noninvasively is very essential. Magnetic resonance imaging (MRI) has emerged as a powerful tool for imaging of the spinal cord because of its high soft-tissue contrast and specificity to the pathologic cord. However, using the conventional MRI methods such as T1-weighted and T2-weighted imaging, the disease burden and monitoring process cannot always be assessed accurately. An advanced imaging method, the diffusion MRI of spinal cord, has been proven as a more successful imaging method than the conventional MRI methods to detect the lesions in earliest stages; however, diffusion MRI of spinal cord is challenging. The major technical challenges for the high-resolution diffusion MRI of the spinal cord include the low signal-to-noise ratio (SNR) from the small cross-sectional area and deep location of the cord, large field inhomogeneity in the static magnetic field due to the magnetic susceptibility difference between tissue-bone interface, and patient’s involuntary as well as voluntary motions. In addition to the above technical challenges, the signal behavior and outcomes of the diffusion MRI cannot be easily interpreted in the spinal cord because of its complex microscopic structure. This dissertation contributes significantly in three areas to overcome the difficulties currently faced in diffusion MRI of the spinal cord. A Monte Carlo simulation (MCS) of water diffusion in white matter (WM) has been developed and performed. The simulation provides the deeper understanding of the signal measured in diffusion MRI, which facilitates easier interpretation of the outcomes of diffusion MRI. The results of the ultrahigh-b radial diffusion-weighted imaging (UHB-rDWI) of excised pig cervical spinal cord (CSC) agree fairly well with the results of the simulation. An improvement in the SNR of the spinal cord images was achieved by constructing an 8-channel CSC dedicated coil, which does not require a commonly used preamplifier decoupling technique to minimize the interaction between nonadjacent elements. The newly constructed coil provides 1.4âˆ'2 time SNR improvement compared with the manufacturer’s coil (Siemens’ head neck and spine matrix). A new sequence, 2D single-shot diffusion-weighted stimulated echo planar imaging with reduced field of view (2D ss-DWSTEPI-rFOV), has been developed for the UHB-rDWI of the spinal cord. The 2D ss-DWSTEPI-rFOV sequence acquires an image in a single excitation, and thereby reduces motion related artefacts. The reduced phase field of view imaging capability of the new sequence minimizes the off-resonance (field inhomogeneity and chemical shift) related artefacts. The time efficient sequence acquires stimulated echoes (STE) for the high-resolution UHB-rDWI of the spinal cord. |
