A Clinically-Compatible Workflow for Computer-Aided Assessment of Brain Disease Activity in Multiple Sclerosis...

Area of Science:

• Neurology
• Radiology
• Medical Imaging

Background:

• Multiple Sclerosis (MS) treatment has advanced with more disease-modifying drugs.
• Personalized medicine in MS aims for no clinical or radiological activity.
• Accurate detection of new lesions on longitudinal MRI is crucial for monitoring MS activity.

Purpose of the Study:

• To develop and evaluate a comprehensive workflow for monitoring new FLAIR lesions in MS patients using longitudinal MRI.
• To provide a tool usable by hospital and private practice neurologists and radiologists in France.
• To improve the accuracy and efficiency of lesion detection in MS patients.

Main Methods:

• Developed a workflow with three components: automated data anonymization/transfer, automated lesion segmentation from MRI scans (T1, T2, FLAIR), and a web viewer for lesion visualization.
• Evaluated the workflow on 54 longitudinal MRI scan pairs analyzed by 3 experts (neuroradiologist, radiologist, neurologist).
• Compared expert analysis with and without the workflow to assess accuracy and time efficiency.

Main Results:

• The workflow significantly improved the accuracy of new MS lesion detection (2.3 lesions with workflow vs. 1.8 without, p=5.10^-4).
• It reduced expert analysis time (mean difference 2'45", p=10^-4).
• The workflow increased the sensitivity of lesion detection, particularly for experienced neuroradiologists (0.90 with workflow vs. 0.74 without, p=0.003).

Conclusions:

• The developed workflow is a valuable aid for clinicians in detecting new MS lesions.
• Improved lesion detection has implications for classifying MS activity and guiding therapeutic management.
• This tool supports the paradigm of personalized medicine in Multiple Sclerosis.