Investigating the association between grey matter hypoperfusion and atrophy in multiple sclerosis using arterial spin labelling.

The study of grey matter (GM) perfusion abnormalities in multiple sclerosis (MS) is gaining increasing interest, thanks to the recent developments in MRI techniques such as arterial spin labelling (ASL). Previous studies in small mixed cohorts have demonstrated reduced perfusion in different GM regions in patients with MS compared with controls, especially in the frontal/parietal lobes and deep GM. In this multi-centre study, we aimed to investigate the relationship between ASL-perfusion patterns and GM atrophy in a large cohort of MS patients. We hypothesized that GM hypoperfusion is a phenomenon not wholly associated with the presence of structural atrophy. Therefore, we have evaluated whether GM perfusion changes are directly explained by underlying atrophy or occur in the absence of detectable GM tissue loss. Moreover, the relationship between clinical (physical and cognitive) measures and global GM perfusion level was assessed. Conventional structural and pseudo-continuous ASL data were acquired using 3T MRI scanners and a standardized protocol in four European MS centres. We included in the analysis 170 subjects [96 relapsing-remitting MS (RRMS), 41 primary-progressive MS (PPMS) and 33 healthy controls (HC)] that exhibited acceptable data quality and compatible acquisition schemes/ASL maps. Global and regional cerebral blood flow (CBF) patterns were explored to identify group differences, statistically controlling for 'morphometric covariates' (i.e. total brain volume, cortical thickness or subcortical volume) along with age, sex, disease duration and total T2-lesion load in the analysis of covariance ('basic covariates'). Similarly, partial correlation analyses were performed to assess the association of global CBF with physical (Expanded Disability Status Scale) and cognitive (Symbol Digit Modalities Test) scores. Global GM CBF values were lower in RRMS (64.1 ± 20.9 mL/100 g/min) and PPMS (52.9 ± 16.1 mL/100 g/min) compared with HC (78.5 ± 21.1 mL/100 g/min), with significant differences for both MS types compared with controls using both sets of covariates (pBonf≤0.05). Moreover, regional perfusion reduction encompassing several cortical and subcortical GM regions were found in patients versus controls, even after the inclusion of the morphometric covariates, while there were no differences in GM perfusion levels between MS phenotypes. In all patients considered together, no significant associations between global GM CBF and physical disability/cognitive tests scores were found when controlling for the confounding factors. These findings suggest that GM hypoperfusion is at least partially dissociated with atrophy. In this scenario, ASL-based perfusion may be a promising functional paraclinical tool to quantify GM neurovascular impairment in MS, aiding the understanding of the pathological mechanisms.