Real-world brain volumetry in multiple sclerosis: importance of methodological consistency and clinical relevance of gray matter atrophy

Joke Temmerman^1,2,3Ann Vanremoortel⁴Maria Bjerke^2,5,6,7Marie Beatrice D'Hooghe^2,4Guy Nagels^2,3,8Ayla Pauwels^2,3,4Diana Sima^2,8Dirk Smeets^2,8Jeroen Van Schependom^2,9Mandy Melissa Jane Wittens^2,3,5,8Sebastiaan Engelborghs^2,3,5Miguel D'haeseleer^2,3,4*

• ¹Department Biomedische Wetenschappen, University of Antwerp, Antwerp, Belgium
• ²Vrije Universiteit Brussel, Center for Neurosciences, NEUR and/or AIMS Research Groups, Brussels, Belgium
• ³Universitair Ziekenhuis Brussel, Department of Neurology, Brussels, Belgium
• ⁴Nationaal Multipele Sclerose Centrum, Melsbroek, Belgium
• ⁵Universiteit Antwerpen Departement Biomedische Wetenschappen, Antwerp, Belgium
• ⁶Laboratory of Clinical Neurochemistry, Universitair Ziekenhuis Brussel Klinische biologie, Brussels, Belgium
• ⁷Department of Neurobiology, Center for Alzheimer Research, Division of Clinical Geriatrics, Karolinska Institutet, Stockholm, Sweden
• ⁸Icometrix NV, Leuven, Belgium
• ⁹Department of Electronics and Informatics, Vrije Universiteit Brussel, Brussels, Belgium

Background: Brain volume loss (BVL) is a marker of neurodegeneration associated with clinical disability in multiple sclerosis (MS). However, its application in routine clinical practice is limited due to measurement errors introduced by the use of different magnetic resonance imaging (MRI) scanners across and within centers. Objective: To confirm the existence and clinical relevance of longitudinal BVL in a real-world MS cohort with scanner variability, employing a dedicated quantification pipeline combined with post-acquisition harmonization. Methods: We analyzed MRI data from 72 MS patients scanned across multiple Belgian centers over 48-60 months. Clinical disability was assessed using the Expanded Disability Status Scale, Timed 25-Foot Walk Test, 9-Hole Peg Test (9HPT), and Symbol Digit Modalities Test. Percentage volume change (PVC) in whole brain (WB), total gray matter (TGM), cortical gray matter (CGM), and deep gray matter was quantified using the icobrain ms pipeline. A similarity index was applied to account for scanner differences. Twenty-seven healthy volunteers served as controls. Results: No significant differences in annualized PVC were observed between MS patients and controls. Within the MS group, 9HPT performance correlated with TGM (ρ = -0.30, P = 0.017) and CGM (ρ = -0.31, P = 0.015) volume loss. Modified MS Functional Composite scores correlated with WB (R = 0.28, P = 0.03), TGM (ρ = 0.31, P = 0.014), and CGM (ρ = 0.31, P = 0.013) volume loss and could be independently predicted by these measures. Conclusion: Using automated brain volumetry with post-acquisition harmonization to address scanner variability, we did not detect accelerated BVL in this real-world MS cohort compared to healthy individuals. Nonetheless, GM volume loss was found to be clinically relevant in MS.