MICROSTRUCTURAL WHITE MATTER PROPERTIES IN MULTIPLE SCLEROSIS: ANATOMICAL SPATIAL MAPPING VIA NODDI MODELLING TO BETTER UNDERSTAND THE MECHANISM OF INJURY

Introduction. Plenty of literature focused on the topography of multiple sclerosis (MS) injury localization within brain. Various theories were proposed to explain this pattern, calling into question two main hypothesis: 1- preexistent tissue-intrinsic microstructural susceptibility factors 2- cerebrospinal-fluid (CSF)-borne soluble inflammatory factors diffusing through the brain surfaces (inner-ependymal-ventricular and outer-pial-cortical), thence affecting the parenchyma with a decreasing distribution along a distance-from-CSF gradient. However, despite more than one hundred years of efforts, the etio-physio-pathological basis of the onset and development of the MS plaque has not been completely unveiled yet. Recent technological advances in magnetic resonance imaging (MRI) and data post-processing, enabling an in-vivo definition of the local microstructure of the white matter (WM), give new perspectives for untangling the secrets of this complex disease. Materials & Methods. With a novel algorithm to analyze diffusion weighted MRI images, “NODDI”, we created detailed atlases of the microstructural characteristics of the normal WM in a healthy population; parallelly, we defined the topography of the lesions for a MS-affected population. By superposition of the patients lesion maps onto the healthy atlas, we could then test if any of the microstructural NODDI parameters is predictive of development of a T1-visible lesion. We then tested at which deepness of the gradient of distance from CSF, the a-priori microstructural susceptibility factor was more responsible of the T1-visible lesion development. Finally, we computed the mean distance from the CSF of the T1-lesioned tissue, compared to the T1-spared one. Results. In the corresponding areas where the patients developed T1-visible lesions, we found significant higher values of Neurite Density (ND) on the healthy population atlas, if compared to the areas where no lesion was visible on T1 imaging. The a-priori tissue property of high ND was found to have its greatest influence on T1-visible lesion formation especially in the deep WM layer (the furthest from the pial and ventricular surfaces). The NODDI microstructural parameter Orientation Dispersion Index (ODI) showed to have no influence at any level on the tissue proneness to develop a T1-visible lesion. The average distance of the T1-lesioned tissue from the ventricles was higher than the one of the T1-spared tissue. Conclusion. Our results suggest that an higher density of neurites seem to play a role on the probability of development of a T1-visible WM lesion in MS, while the orientation dispersion of the axons does not appear to have any impact on these pathological events. An higher coherence and compactness of structure in the myelin-rich WM areas could constitute a facilitating factor for the auto-inflammatory immune process against myelin antigens. It is interesting to see that this effect, which appears already significant when considered the whole brain, looks to be even more prominent in the “deep WM layer”, which is the furthest-from-CSF part of WM. Conversely, the lesion-promoting effect of high ND seems to be attenuated or neutralized in the WM layers neighboring the CSF: this fact brings to speculate the existence of some underlying interaction between inflammatory soluble factors and tissue structure, at different WM deepness levels.