Pooled analysis of multiple sclerosis findings on multisite 7 Tesla MRI: Protocol and initial observations.

Area of Science:

• Neuroimaging
• Radiology
• Neurology

Background:

• 7 Tesla MRI research has advanced multiple sclerosis (MS) pathology understanding.
• Prior studies were limited by small sample sizes and single-center methodologies.
• Multicenter studies are essential for validating 7T MRI findings in MS.

Purpose of the Study:

• To describe methods for pooling a large, retrospective multicenter 7T MRI dataset in MS.
• To establish unified data processing and harmonization techniques.
• To provide a methodological reference for future multicenter MS imaging studies.

Main Methods:

• A collaborative effort by the North American Imaging in MS (NAIMS) Cooperative.
• Pooling 528 7T MRI scans from 350 individuals with MS across five institutions.
• Developing unified image processing, data harmonization, and lesion segmentation methods.

Main Results:

• Initial observations on intersite acquisition differences (brain coverage, image quality).
• Developed and trained image evaluators achieving median Dice Similarity Coefficients of 0.73–0.82 for lesion annotation.
• Established a common data repository with plans for future expansion.

Conclusions:

• The developed methods enable the creation of large, harmonized multicenter 7T MRI datasets for MS research.
• This work serves as a resource for investigators combining multicenter 7T MRI data in MS.
• Standardized methods are key for reliable analysis and understanding of MS pathology.