Investigating biomarkers of clinical progression in multiple sclerosis using magnetic resonance imaging, blood biomarkers, and machine learning

Multiple sclerosis (MS) is a chronic, inflammatory, and demyelinating disease of the central nervous system characterized by heterogeneous clinical presentations and variable progression. Sensitive biomarkers capable of detecting subtle tissue damage and predicting disease progression are needed to improve clinical management and therapeutic development. This thesis examined advanced MRI and blood biomarkers to better characterize tissue integrity, clinical progression, and predictors of disability in MS across spinal cord and brain imaging studies and multimodal prediction algorithms. The first research chapter investigated spinal cord myelin integrity using myelin water imaging (MWI) in individuals with relapsing-remitting and progressive MS. Lower mean myelin water fraction and greater perilesional myelin heterogeneity were associated with worsening disability over five years, suggesting that tissue surrounding spinal cord lesions is particularly vulnerable to ongoing degeneration. The second research chapter evaluated two definitions of clinical progression in the CanProCo cohort and found that a composite definition incorporating the Expanded Disability Status Scale (EDSS), processing speed, hand dexterity, and ambulation tests was more sensitive than EDSS alone in detecting early progression. The third research chapter demonstrated that diffusion tensor imaging (DTI) and MWI metrics can differentiate people who clinically progress from those who remain clinically stable over two years, with baseline and longitudinal changes in axonal integrity and myelin content corresponding to clinical worsening or stability. In the final research chapter, MRI, blood protein, and clinical data were integrated using an ensemble logistic regression model to predict 2-year progression, showing that multimodal integration significantly improved predictive performance over models that only integrated MRI or blood biomarker data with clinical data. The findings highlight the value of advanced MRI and molecular biomarkers to capture subtle disease processes and enhance early detection of clinical progression in MS, contributing to the development of personalized prognostic tools and improved monitoring of therapeutic efficacy.