Multiple sclerosis: Researchers have identified two distinct biological forms of multiple sclerosis using artificial intelligence, a development that could transform how doctors treat the debilitating neurological condition.
The disease affects millions worldwide, but current treatment approaches rely heavily on managing symptoms rather than addressing the specific biological mechanisms at work in individual patients. That often means medications don’t work as well as they should.
A team from University College London and Queen Square Analytics used AI technology, standard blood tests and brain imaging to pinpoint two separate MS variants in a study of 600 patients. Medical experts called the discovery a potential game-changer for care.
Also Read | Sleep apnoea linked to 40% higher odds of mental health conditions, shows study
The investigation centred on measuring blood levels of serum neurofilament light chain, or sNfL, a protein that reveals how much nerve damage is occurring and how aggressively the disease is progressing.
Researchers fed the protein measurements and brain scan images into a machine learning program called SuStaIn. The analysis, detailed in the medical journal Brain, uncovered two categories: early sNfL and late sNfL multiple sclerosis.
Patients with the first type showed elevated sNfL levels from the disease’s onset, along with damage to the corpus callosum, a brain region that connects the two hemispheres. These individuals also developed brain lesions rapidly. Scientists described this variant as more aggressive.
The second group experienced brain tissue loss in the limbic cortex and deep grey matter regions before their sNfL numbers climbed. This form progresses more slowly, with obvious damage appearing later in the disease course.
The findings should help physicians better predict which patients face higher risks of specific complications, opening the door to more tailored treatment plans.
Study lead Dr Arman Eshaghi from UCL said MS isn’t a single condition, and existing classifications don’t capture the tissue-level changes doctors need to understand for effective treatment. “By using an AI model combined with a highly available blood marker with MRI, we have been able to show two clear biological patterns of MS for the first time,” Eshaghi said. “This will help clinicians understand where a person sits on the disease pathway and who may need closer monitoring or earlier, targeted treatment.”
Patients identified with early sNfL MS might qualify for more potent medications and receive closer medical supervision, according to Eshaghi. Those with late sNfL could receive different therapies, possibly including customised treatments designed to shield brain cells and neurons from damage. The work represents both a modernisation of centuries-old clinical examination methods through AI and a path toward treatments matched to each patient’s disease profile, he added.
