BRAIN VOLUMES AND COGNITION IN MOG ANTIBODY-ASSOCIATED DISEASE: AN ITALIAN MULTICENTER LONGITUDINAL STUDY

Background. Myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) is a recently recognized demyelinating disorder whose clinical features partly overlap with Multiple Sclerosis (MS). MRI is, to date, one of the best tools to differentiate these diseases; however, little is known about brain atrophy, disease progression on brain MRI and neuropsychological (NPS) profile of MOGAD patients compared to MS. Objectives. This multicenter longitudinal study compares global, white matter, gray matter and regional brain MRI volumes and T2 lesion volume between MOGAD, re-lapsing-remitting MS (RRMS) patients and a healthy control (HC) group with brain MRI scans available from an online repository. Methods. 16 adult MOGAD patients (9 F) with a clinical follow-up ≥6 months have been selected from an ongoing multicenter observational study which started the recruitment on 31/01/2017. 44 RRMS patients (26 F) fulfilling 2017 McDonald Crite-ria, homogenous for age and sex to the MOGAD cases, and a clinical follow-up ≥6 months have been recruited in Verona MS center through the consultation of an electronic database. For both MOGAD and MS patients, clinical and NPS assess-ments (BICAMS battery), as well as brain MRI scan, are performed at T0 and after 18±6 months. T1-3D and FLAIR-3D scans are used for volumetric analysis, with the same MRI protocol and scanner at both timepoints for each patient. Annualized percentage brain volume change (PBVC/y) between the two MRI timepoints, base-line global brain, white matter (WM), grey matter (GM) and regional brain volumes are compared between groups using SIENA, SIENAX and VolBrain softwares. T2 le-sion volume is assessed using ITK-snap. A group of 14 age- and sex-matched HC has been also added as comparator. Results. MOGAD patients show no lesions on brain MRI in 50% of cases, and they have significantly less cortical, periventricular, juxtacortical, callosal and brain-stem lesions compared with MS group. PBVC/y is lower in MOGAD than in RRMS (p=0.014), and lower in HC than in MS patients (p=0.005). Overall, MOGAD shows higher mean global brain (p=0.012) and WM volume (p=0.024), but lower median T2-lesion volume at timepoint 1 (p<0.001) compared to MS; T2-lesion volume increases over time in RRMS (p<0.001) but not in MOGAD cohort (p=0.262). NPS performanc-es are comparable between MOGAD and MS patients, and only one MS and one MOGAD patient fails one of the BICAMS tests. No significant correlations have been established between NPS tests scores and volumetric variables at T0. Conclusions. Structural brain MRI features of MOGAD are characterized by higher global brain and WM volumes as well as less brain volume loss over time compared to RRMS. Lesion distribution has different topography in the two diseases. Moreo-ver, MS shows an increase in T2 volume which is not detected in MOGAD, suggest-ing different underlining pathogenetic mechanisms.