AI-assisted identification of disability patterns within identical EDSS grades

Area of Science:

• Neurology
• Biostatistics
• Artificial Intelligence in Medicine

Background:

• The Expanded Disability Status Scale (EDSS) is standard for multiple sclerosis (MS) trials.
• EDSS scores ≥4.5 primarily assess ambulation, potentially missing other functional impairments.

Purpose of the Study:

• To investigate if identical EDSS scores can represent diverse disability patterns.
• To refine disability assessment in progressive MS.

Main Methods:

• Analysis of 13,103 assessments from the EXPAND trial in secondary progressive MS patients.
• Utilized Functional System scores, subscores, and ambulation data.
• Applied machine learning clustering to identify distinct subscore patterns within EDSS categories.

Main Results:

• Identified numerous clusters representing similar disability patterns.
• Differentiated four distinct subscore patterns within the same EDSS score for patients with EDSS ≥4.0.

Conclusions:

• Artificial intelligence can discern varied disability profiles among patients with the same EDSS score.
• This approach enhances the granularity of disability assessment in MS clinical trials.