Diffusion Magnetic Resonance Imaging of the Human Spinal Cord in Vivo: Feasibility and Application of Advanced Diffusion Models

Abstract

Multiple sclerosis (MS) is an autoimmune disease of the central nervous system (CNS) that is marked by inflammation, demyelination, gliosis and axonal loss. The damage to the CNS from these mechanisms can result in an accumulation of sensorimotor impairment. Diffusion magnetic resonance imaging (MRI) offers the potential to reveal the microstructural integrity of the cervical spinal cord resulting from these pathological mechanisms, which would be useful in the diagnosis and management of MS. This dissertation investigates the application of a spectrum of diffusion models. Starting from the conventional signal model diffusion tensor imaging (DTI) and working towards biophysically based models (i.e., NODDI, SMT and DBSI), these methods are assessed based on their reproducibility in healthy controls and sensitivity to distinguish disparity in MS patients. In comparison to healthy controls, decreased axonal volume fractions were estimated in MS patients using NODDI and SMT. Furthermore, these techniques were robust when optimized for shorter acquisition times and increased coverage. Taken together, the work presented here describes the feasibility and potential of novel diffusion MRI methods for the cervical spinal cord, serving as a vital stepping stone towards the clinical implementation of characterizing spinal cord microstructure in vivo.