MS Speaks

In 2025, substantial breakthroughs were made with respect to earlier diagnosis, improved mechanistic understanding, and new hope for delaying progression of multiple sclerosis. Over successive refinements, the McDonald criteria have enabled earlier diagnosis of multiple sclerosis, with the 2017 update allowing diagnosis at the initial clinical event. The newly revised 2024 criteria allow for identification of asymptomatic disease with biological and radiological biomarkers. Because optic neuritis is the initial clinical event in about 25% of people with multiple sclerosis, inclusion of the optic nerve as a fifth anatomical location for demonstrating dissemination in space, incorporation of optical coherence tomography, visual evoked potentials, and orbital MRI as supportive paraclinical tests are welcome additions to the diagnostic criteria.

The central vein sign, if present on MRI findings, can now substitute for the presence of dissemination in time in specific situations. Diagnosis is possible even without dissemination in space or dissemination in time if supported by specific MRI biomarkers (eg, central vein sign or paramagnetic rim lesions) or CSF findings. The inclusion of the kappa free light chain index as an alternative to oligoclonal bands simplifies CSF analysis, reducing reliance on isoelectric focusing and subjective band interpretation.

Although the goal is earlier diagnosis and more timely intervention, these newer criteria increase reliance on MRI pattern recognition, in particular the newly introduced high-specificity imaging biomarkers central vein sign and paramagnetic rim lesions require specialised neuroimaging expertise. In some cases, diagnosis can now be made with MRI findings solely, without symptoms, CSF analysis, or evidence of dissemination in time. Early diagnosis allows initiation of disease modifying agents (eg, B cell depleting therapy) at a very early stage in disease progression, potentially transforming multiple sclerosis into a non-disabling chronic disease.

In parallel, the 2024 MAGNIMS–CMSC–NAIMS consensus outlines the importance of standardised imaging of the brain, optic nerves, and spinal cord, and recommends fat-suppressed orbital sequences and susceptibility-based techniques to detect central vein sign and paramagnetic rim lesions. The authors highlight that high T2 lesion load, brainstem or spinal cord lesions, paramagnetic rim lesions, and brain atrophy predict increased likelihood of long-term disability. From a practice perspective, a particularly important recommendation concerning spinal cord imaging is to perform “axial slices at least through the cervical region with moderately T2-weighted spin-echo or T2 gradient echo”. A 2025 population-based analysis showed that half of spinal cord lesions are missed when axial imaging is not performed.

Given that misdiagnosis generally stems from MRI misinterpretation, the 2024 criteria must be applied judiciously and in conjunction with neurologist judgment, in order to avoid overdiagnosis. To reduce misdiagnosis, the authors of the newly revised 2024 criteria emphasised the importance of considering lesion morphology and location to avoid mislabeling non-specific white matter lesions and recognising mimics.

The phase 3 HERCULES trial of the Bruton's tyrosine kinase inhibitor tolebrutinib provided evidence that progression can be slowed in non-relapsing secondary progressive multiple sclerosis. Among more than 1100 participants, tolebrutinib reduced the risk of confirmed disability progression at 6 months by 31% compared with placebo, with concordant benefits in walking speed and MRI lesion burden. For a population that has long awaited therapeutic success, these findings provide hope for patients with progressive multiple sclerosis.

However, reasons for caution include the continuous progression observed in 25% of patients treated with tolebrutinib and the complication of elevated liver enzymes in 4% of patients.

However, tolebrutinib in relapsing multiple sclerosis trials (GEMINI 1 and 2) did not outperform teriflunomide in reducing relapses or MRI activity. This divergence is mechanistically intriguing. The modest anti-inflammatory effect of tolebrutinib in relapsing disease, contrasted with its effect on disability accrual in non-relapsing progressive disease, suggests that Bruton's tyrosine kinase inhibition might modulate microglial-driven, smouldering inflammation rather than acute, lymphocyte-mediated activity. Whether Bruton's tyrosine kinase inhibition signals a new class of progression-directed therapies, or merely opens a transient therapeutic window, remains to be determined.

Two immunopathological studies highlight the importance of novel mechanistic insights into identification of potential therapeutic targets to halt progression. In the first study, using post-mortem tissue from the Netherlands Brain Bank, a distinct lesion type characterised by a broad rim of macrophages and microglia at least 1 mm wide—termed the broad rim lesion—exhibited an immunological and transcriptomic signature that was associated with rapid disease progression. The authors observed strong translocator protein (TSPO) 18-kDA immunoreactivity, considered a marker of myeloid activation, in the rim region. Since TSPO is a validated PET imaging target, they conducted an independent TSPO PET study in 114 patients with multiple sclerosis to search for radiological broad rim lesions. They found radiological broad rim lesions in about 34% of patients. These broad rimmed lesions were strongly associated with disease progression, higher lesion load, and diffuse white matter injury on MRI. The absence of broad rim lesions in a third of rapid progressors could be explained by pathogenic heterogeneity and suggests alternative, neurodegenerative or diffuse inflammatory mechanisms. Different pathological mechanisms will probably require personalised diagnostics and therapeutics in the future.

In the second study, a high-resolution spatial transcriptomic atlas of chronic active lesions revealed how the immune system and cellular metabolism interact to sustain CNS inflammation. Within the rims of chronic active lesions, the authors identified CD8⁺ tissue-resident T-cell niches intertwined with IFNγ-responsive, lipid-laden microglia. These foamy microglia, impaired in cholesterol efflux, accumulate toxic myelin-derived lipids that perpetuate their inflammatory state. Disabling microglial cholesterol transporters in animal models amplified demyelination and T-cell recruitment, whereas pharmacological stimulation of lipid efflux reversed inflammation and improved pathological findings and clinical severity scores in a mouse model of disease. Correcting this microglial cholesterol efflux defect might provide an alternative path to treating people who have progressive multiple sclerosis.


In summary, 2025 has seen the multiple sclerosis scientific community continuing to redefine the condition as a unified, biologically continuous, and increasingly modifiable disease.