Dual-Sensitivity Multiple Sclerosis Lesion and CSF Segmentation for Multichannel 3T Brain MRI

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Radiology

Background:

• Multiple sclerosis (MS) diagnosis and progression monitoring rely on detailed brain MRI analysis.
• Accurate segmentation of brain tissues and lesions is crucial for assessing disease impact.
• Existing methods may face challenges with image quality variations, particularly in specific brain regions.

Purpose of the Study:

• To present a fully automated pipeline for segmenting 3T brain MRI scans in MS patients.
• To develop a tool (3T morphometry or 3TM) for quantifying indicators of MS disease progression.
• To assess brain atrophy and white matter lesion (WML) volume using multichannel MRI data.

Main Methods:

• The 3TM pipeline utilizes high-resolution 3D T1-weighted, T2-weighted, and FLAIR MRI sequences.
• A dual sensitivity approach was implemented to address nonuniform noise and contrast, especially in the posterior fossa.
• Segmentation accuracy was validated against manual delineations, and precision was assessed through scan-rescan analyses.

Main Results:

• High intraclass correlation coefficients (0.86-0.95) were achieved for WML, CSF segmentation, and brain parenchymal fraction (BPF).
• Dual sensitivity effectively reduced false-positive WMLs in the infratentorial region, enhancing global sensitivity without compromising specificity.
• Scan-rescan analyses demonstrated good precision with low coefficients of variation for WMLs, BPF, and tissue volumes (GM, WM, CSF).

Conclusions:

• The 3TM pipeline provides accurate and precise segmentation of brain structures and lesions in MS.
• The automated nature and probability map outputs facilitate efficient expert review and high-throughput analysis of large MS datasets.
• This tool holds potential for robust characterization of MS progression in clinical research and practice.