Related Experiment Video
Updated: Sep 10, 2025
11:35
The Multiple Sclerosis Performance Test MSPT: An iPad-Based Disability Assessment Tool
Published on: June 30, 2014
58.1K
AI-driven reclassification of multiple sclerosis progression
Habib Ganjgahi1,2, Dieter A Häring3, Piet Aarden3
1Department of Statistics, University of Oxford, Oxford, UK.
Nature Medicine
|August 20, 2025
Summary
Multiple sclerosis (MS) is reclassified using machine learning into a disease continuum, not distinct subtypes. This new model, based on disability, brain damage, and activity, improves understanding and treatment for millions affected by MS.
Area of Science:
- Neurology
- Data Science
- Medical Classification
Background:
- Traditional multiple sclerosis (MS) subtypes lack prognostic value and hinder drug discovery.
- Current classification fails to capture MS pathobiology and disease evolution.
- A new approach is needed to better understand and manage MS.
Purpose of the Study:
- To develop a data-driven classification of MS disease evolution.
- To identify key dimensions defining MS disease states.
- To propose a refined classification for improved patient management and drug discovery.
Main Methods:
- Analysis of a large clinical trial database (approx. 8,000 patients) using probabilistic machine learning.
- Inclusion of clinical data, MRI scans, and patient visits.
- Validation in an independent clinical trial database and real-world cohort (over 4,000 patients).
Main Results:
- Identified four dimensions of MS: physical disability, brain damage, relapse, and subclinical disease activity.
- Defined two poles of a disease spectrum: Early/Mild/Evolving (EME) MS and advanced MS.
- Demonstrated transitions to advanced MS via brain damage accumulation, with progression independent of relapses.
Conclusions:
- MS should be viewed as a disease continuum rather than distinct subtypes.
- The proposed classification offers a unifying understanding of MS.
- This approach can enhance patient management and accelerate drug discovery for MS.
