Connecting white and grey matter damage in multiple sclerosis using multimodal MRI

The aim of this thesis was to study the temporal and spatial relationships between white matter (WM) damage and grey matter (GM) atrophy in early stages of multiple sclerosis (MS) using multimodal magnetic resonance imaging (MRI) techniques, in order to better understand how neurodegeneration is related to neuroinflammation and other pathological changes in MS. Our main research question was to investigate whether GM atrophy progressed faster when there is more damage in the connected WM, and how this related to disability, in subjects with early relapsing-remitting (RR)MS. In conclusion, this thesis has given some new insights in the relation between WM and GM damage, both temporal and spatial, in early RRMS brain and spinal cord by using multimodal MRI techniques. In summary, the research presented in this thesis has led to the following conclusions: Chapter 1 - Introduction • In non-progressive MS, more WM lesions (size and/or number) correlated to more GM atrophy, both globally and regionally. In progressive MS, these correlations were less consistent. • Harmonization of methodology in MRI studies of MS is of importance to enable between-study comparisons of data. Chapter 2 – Challenges related to lesions in MS • Lesion segmentation methods vary largely in their volumetric and spatial accuracy compared to manual segmentation. • Using only one manual segmentation image as input for deep learning software (such as nicMSlesions) can already outperform commonly used (semi-)automated methods. • LESIM is a new lesion simulation tool that enables the user to introduce realistic looking MS lesions with correct shapes, locations, intensities and signal-to-noise ratios in healthy control images. This allows optimization and validation of new segmentation software in MS. Chapter 3 – Atrophy and lesions in the MS spinal cord • Upper cervical cord area (UCCA) cannot be robustly reproduced between scanners or between vendors and comparisons in multi-center studies should therefore be made with caution. • UCCA can be robustly measured when cervical cord lesions are present. • There was no temporal nor spatial relation between UCCA and cervical cord lesions in our cohort of early RRMS subjects. • Increasing cord atrophy, and not cord lesions, were predictive of increasing disability over time. Chapter 4 – Grey and white matter damage in the MS brain • Thalamic atrophy progresses faster when more damage is present in the connected WM in early RRMS and CIS. • Cortical GM atrophy progresses faster when more damage is present in the connected WM in early RRMS subjects with increasing disability. • WM integrity measures such as NODDI and MWF had additional value to standard DTI parameters in determining this WM-GM relationship, and including these acquisitions is recommended for future MRI research studies in MS.