Blood Test Combined with MRI Brain Scans Reveals Two Distinct Multiple Sclerosis Types

Multiple sclerosis (MS) affects more than 2.8 million people worldwide, yet predicting how the disease will progress in individual patients remains difficult. Current MS classifications are based on clinical symptoms, which often do not reflect the underlying biological mechanisms driving nerve damage. New research now shows that combining a simple blood test with standard brain imaging and artificial intelligence (AI) can distinguish biologically distinct forms of MS for the first time.

In a study led by researchers from University College London (London, UK) and Queen Square Analytics (London, UK), the team combined blood levels of serum neurofilament light chain, a marker of nerve cell damage, with MRI brain scans showing disease spread. These data were analyzed using a UCL-developed machine learning model called Subtype and Stage Inference, or SuStaIn.

Researchers analyzed data from 634 participants drawn from two clinical trial cohorts. Serum neurofilament light chain levels were measured from blood samples, while MRI scans assessed structural brain changes and lesion development. The SuStaIn model integrated these inputs to identify distinct disease patterns and stages based on biological features rather than clinical symptoms.

The analysis revealed two distinct biological types of multiple sclerosis. In the early-sNfL type, patients showed high blood levels of neurofilament light chain early in the disease, along with early damage to the corpus callosum and rapid lesion formation, indicating a more aggressive form. In the late-sNfL type, brain shrinkage in regions such as the limbic cortex and deep grey matter occurred before blood biomarker levels rose, suggesting a slower disease course. The findings were published in Brain.

The approach allows clinicians to more accurately predict which patients are at higher risk of developing new brain lesions and worsening disability. By identifying disease biology earlier than clinical deterioration appears, doctors may be able to tailor monitoring and treatment more precisely. Researchers believe these data-driven subtypes could help match patients to therapies that target the underlying mechanisms of their disease.

“MS is not one disease, and current subtypes fail to describe the underlying tissue changes, which we need to know to treat it,” said Arman Eshaghi, MD, PhD, founder of Queen Square Analytics and lead author of the study. “By using an AI model combined with a widely available blood marker and MRI, we have shown two clear biological patterns of MS for the first time.”

Related Links:
University College London
Queen Square Analytics