A combined MRI and histological evaluation of cortical pathology in multiple sclerosis

Multiple sclerosis (MS) is a chronic, demyelinating disease affecting the central nervous system. White matter lesions are the typical hallmark of the disease, however, cortical pathology has been demonstrated to influence MS disease progression. While neuropathological assessments evidenced areas of cortical demyelination, inflammation and neurodegeneration, in-vivo detection of cortical pathology is still challenging. Quantitative magnetic resonance imaging (qMRI), measuring biophysical properties sensitive to the composition/microstructure of the underlying tissue, has the potential to reveal subtle pathological changes. Despite the potentiality of the qMRI approach, a detailed multimodal characterization of cortical pathology is still lacking. The main objective of this thesis was to design and optimize acquisition and processing of a multimodal MRI protocol comprehending the estimations of T1 and T2* relaxation times, Magnetization Transfer Ratio (MTR) and Magnetization Transfer Saturation (MTsat) for the characterization of cortical pathology carried out on brain samples consisting of bi-hemispheric coronal slices donated from human MS donors. More in detail the thesis aimed to i) evaluate the feasibility of a B1 corrected variable flip angle approach for the estimation of T1 relaxation times in formalin-fixed ex-vivo samples at 3T field strength; ii)  evaluate the effect of cortical demyelination, histologically-validated, on each MRI modality estimated at both 3 and 7T and iii) compare the magnitude of the lesion contrast for each MRI modality obtained at the lower with respect to the higher field strength; iv) evaluate the association between each MRI modality acquired at 3 and 7T and activated microglia cell density in lesions and normal appearing tissue.  The MRI protocol was acquired on all the brain samples (8) on a 3T MRI system while a subset of the samples (5) were acquired also at 7T. Five of the samples acquired at 3T were selected to investigate the feasibility of the B1-corrected variable flip angle T1 mapping approach for ex-vivo analysis. Neuropathological assessments of myelin and activated microglia were performed on five samples (three acquired only at 3T and two acquired at both field strengths). Cortical demyelination was delineated on immunohistological stainings and registered to the MRI space.  The analysis of the T1 mapping approaches for ex-vivo investigations at 3T revealed that B1-correction improves, as compared to the uncorrected approach, the reliability of the variable flip angle (VFA) method both in terms of root mean square error and correlation with the gold standard inversion recovery approach. The analyses of histologically validated cortical demyelination revealed a significant decrease in MTsat values as well as an increase in T2* in areas of demyelination as compared to layer-matched normal appearing tissue in all the subjects when examining the global cortical voxel distributions. The same MRI modalities also reported the highest percentage of lesions with marked local contrast with respect to adjacent normal appearing tissue. MTsat also showed the strongest association with microglia activation in the cortex. In conclusion the qMRI approaches undertaken in this combined MRI-histology study demonstrated sensitivity to pathological changes in the cortex of MS patients, hardly visible with conventional MRI, that seemed to be amplified at higher field strength. MTsat have been demonstrated to be sensitive to cortical pathology and might be a candidate for the indirect assessment of demyelination and/or microglia activation in the cortex.