Artificial intelligence has identified two distinct biological types of multiple sclerosis (MS) for the first time, using a combination of a routine blood test and standard brain MRI scans. The groundbreaking study, led by researchers at University College London (UCL) and Queen Square Analytics (QSA), a UCL spinout company, was published in the journal Brain. The research focused on serum neurofilament light chain (sNfL), a protein found in blood that serves as a biomarker for nerve cell damage and reflects how actively MS is progressing. By integrating sNfL levels with MRI data showing the extent of brain lesion spread, the team applied a machine learning model developed at UCL called SuStaIn (Subtype and Stage Inference) to uncover two clearly defined biological subtypes of MS. Analyzing data from 634 participants across two clinical trial groups, the AI model revealed distinct patterns of disease progression that were not previously identifiable through clinical classification alone. These findings suggest that MS is not a single disease but comprises different biological pathways, each with unique underlying tissue changes. Lead author Dr. Arman Eshaghi, from the UCL Queen Square Institute of Neurology and the UCL Hawkes Institute in UCL Computer Science, emphasized the significance of the discovery: “MS is not one disease, and current subtypes fail to capture the underlying tissue changes we need to understand to treat it effectively. By combining a widely available blood test with MRI scans and AI, we’ve identified two clear biological patterns for the first time. This allows clinicians to better determine where a patient is on the disease journey and who might benefit from closer monitoring or earlier, targeted treatment.” Queen Square Analytics, founded in 2020 by Dr. Eshaghi with support from UCL Professors Frederik Barkhof, Geoff Parker, and Daniel Alexander, and UCL Business, the university’s commercialization arm, specializes in data-driven solutions for neurological clinical trials, particularly in MS. The company leverages large datasets and advanced AI to detect subtle patterns in disease progression that were previously invisible. Because changes in sNfL levels and MRI scans often appear before clinical symptoms worsen, this approach enables earlier prediction of disability progression. This opens the door to proactive interventions, potentially slowing or preventing irreversible damage. Multiple sclerosis affects more than 2.8 million people worldwide, often striking young adults and leading to significant disability early in life. Current classification relies on clinical course—such as relapsing or progressive forms—but these categories don’t align with the actual biological mechanisms driving the disease. As a result, treatments chosen based on symptoms may not address the root causes, limiting their effectiveness. By identifying measurable biomarkers from blood and imaging data, this research brings clinicians closer to a precision medicine approach. These data-driven subtypes can now be matched with therapies designed to target specific biological pathways, paving the way for more effective, personalized treatment strategies.