For decades, multiple sclerosis (MS) has been defined primarily by its symptoms, rather than its underlying biology. Now, a new study aims to challenge that approach, presenting evidence that MS may actually follow two distinct biological pathways. It’s a shift that could reshape how clinicians think about diagnoses, disease progression and treatment strategy.
In everyday clinical care, multiple sclerosis is treated as a single disease with a broad variety of symptoms. Doctors rely on patterns of inflammation, imaging changes, and neurological symptoms to guide treatment decisions. But for many patients, managing the disease is far from straightforward. Therapies that work well for some fall flat for others, and progression often unfolds in ways that resist standard treatment strategies.
This kind of patient-to-patient variation is a common part of MS complexity. Now, a new study published in the journal Brain suggests that complexity may be pointing to something more fundamental. Rather than reflecting endless variation within one disorder, the study suggests that different disease experiences may stem from distinct biological patterns.
To explore that possibility, researchers at University College London and Queen Square Analytics set out to look beyond symptoms and clinical labels and focus instead on the biological signals of how MS damages the brain.
The goal was not just to track disease activity, but to see whether hidden patterns in degeneration might reveal different underlying pathways.
To do that, the team paired two complementary sources of information. One came from blood measurements of serum neurofilament light chain (sNfL), a protein released when nerve cells are damaged and widely used as a marker of disease activity. The other came from MRI scans that captured how structural degeneration spread through the brain over time.
Rather than examining each dataset in isolation, the researchers analyzed them together using a machine-learning system developed at UCL called SuStaIn (Subtype and Stage Inference). The model is designed to detect subtle disease patterns and map how they evolve, allowing the team to test whether MS follows a single biological trajectory or something more complex.
When the team examined the combined imaging and biomarker data from 634 people with multiple sclerosis, an unexpected pattern began to take shape.
Instead of detecting a smooth disease spectrum, two distinct structural patterns emerged. The team found that patients clustered into separate groups that reflected different underlying pathways of neurodegeneration.
One subtype was marked by early damage concentrated in the brain’s cortex, while the other was dominated by degeneration in white matter regions. Although both patterns ultimately produced the symptoms associated with multiple sclerosis, the location of tissue damage and the path it followed through the brain differed substantially between the two groups.
Beyond anatomy, the two subtypes followed different timelines of disease progression. One group showed a slower, more gradual pattern of structural decline, while the other experienced more rapid neurodegeneration, reinforcing the idea that MS may not unfold along a single biological clock.
That biological split helps explain one of the most persistent frustrations in multiple sclerosis care: why patients with similar diagnoses often experience very different outcomes.
If MS unfolds along more than one disease pathway, prognosis may depend not only on symptom severity, but on which underlying biological pattern a patient follows. In practical terms, this could change how clinicians interpret early disease signals and assess long-term risk.
That same pattern shows up in treatment response. Therapies that slow progression in one subtype may prove far less effective in another, helping explain why some patients respond well to certain medications while others see little benefit.
Because changes in brain structure and blood biomarkers often appear before obvious clinical deterioration, the researchers suggest this type of data-driven subtyping could eventually help clinicians anticipate disease worsening earlier than symptom-based methods alone.
For researchers, the framework also opens new directions for studying disease mechanisms that have long been difficult to disentangle under a single diagnostic label. If those efforts continue to hold up, they could help move MS care closer to a more personalized, biology-driven model of treatment.
For now, the work remains in the research phase, with the approach not yet intended for clinical diagnosis or treatment guidance. The team’s next step will be to expand the study across larger and more diverse patient populations to confirm whether the same biological patterns hold up in real-world settings.
This study was published in the journal Brain.
Source: University College London via Medicalxpress
