Particle.news
Download on the App Store

AI Analysis Identifies Two Biological Subtypes of Multiple Sclerosis

The peer-reviewed work links routine scans with a blood marker to sort disease by biology, pointing to tailored treatment paths once confirmed.

Overview

  • University College London and Queen Square Analytics report the findings in Brain after assessing about 600 people with relapsing‑remitting and secondary progressive MS.
  • The SuStaIn machine‑learning model combined MRI features with serum neurofilament light chain levels to define early‑sNfL and late‑sNfL subtypes.
  • Early‑sNfL shows high sNfL early, damage to the corpus callosum, and rapid lesion development, suggesting a more active disease course.
  • Late‑sNfL is marked by early cortical and deep grey matter volume loss with a later rise in sNfL, indicating overt tissue injury occurs later.
  • Researchers and the MS Society say the approach could enable more personalised monitoring and treatment decisions, though clinical adoption awaits further validation.