Hidden Damage in the Brain’s Outer Layers Drives Multiple Sclerosis Progression

For a quarter of a century, cortical lesions have been quietly rewriting the textbook understanding of multiple sclerosis. A sweeping new review published in Nature Reviews Neurology traces this transformation, showing how structures once dismissed as pathological footnotes have become central to explaining why some patients worsen relentlessly even when their relapses are fully suppressed. Written by a team led by Massimiliano Calabrese, Roberta Magliozzi and Damiano Marastoni at the University of Verona, the review synthesizes 25 years of neuropathological, imaging and clinical evidence into a unified account of how grey matter damage shapes the disease.

The pathological story began with post-mortem studies in the late 1990s, notably Kidd and colleagues’ landmark 1999 paper in Brain, which established that demyelination in the cerebral cortex is widespread in multiple sclerosis. Subsequent work revealed a striking heterogeneity: leukocortical lesions straddling the white and grey matter boundary, purely intracortical lesions, and subpial lesions that extend along the outer surface of the brain. Subpial lesions proved particularly devastating, often spanning entire gyri and representing the dominant form of cortical pathology in progressive disease. Crucially, these lesions are linked to inflammatory clusters, including B-cell follicle-like structures, in the meninges — the membranes bathing the brain in cerebrospinal fluid.

That link between meningeal inflammation and cortical damage was cemented by Magliozzi and colleagues in 2010, who described a gradient of neuronal loss radiating inward from the pial surface. Neurons closest to the inflamed meninges die first, creating an outside-in pattern of destruction that has since been confirmed in thalamic and cortical compartments, even in children with multiple sclerosis. The review argues that this compartmentalized inflammation — largely segregated from the blood and resistant to many peripherally acting therapies — is a key biological driver of progression independent of relapse activity, a form of worsening that defies the classic inflammatory relapse model.

Seeing this pathology in living patients has been the field’s great technical challenge. Conventional magnetic resonance imaging dramatically underestimates cortical lesion burden, a fact established when Geurts and colleagues combined post-mortem MRI with histopathology in 2005. The breakthrough came with double inversion recovery sequences, which suppress cerebrospinal fluid and white matter signals to isolate the cortex, and with consensus scoring standards published in 2011. Ultrahigh-field 7 Tesla MRI then pushed detection further, revealing cortical lesions as dynamic entities that expand, stabilize and accrue tissue damage over time, and enabling longitudinal characterization of lesion evolution that earlier studies could only infer.

The clinical payoff has been substantial. Cortical lesions are highly specific to multiple sclerosis, appearing in radiologically isolated syndromes, clinically isolated syndromes and early relapsing disease, while remaining essentially absent in migraine and neuromyelitis optica spectrum disorders. Their formal incorporation into the 2017 revision of the McDonald criteria as a valid site for dissemination in space marked a turning point, and more recent work shows that combining cortical lesions with the central vein sign and paramagnetic rim lesions sharpens diagnostic accuracy further.

Prognostically, cortical lesion burden has emerged as one of the most powerful imaging predictors available. Studies spanning decades of follow-up show that cortical lesions detected at diagnosis independently forecast physical disability progression, conversion to secondary progressive disease, cognitive impairment and even seizure susceptibility, given the established association between temporal lobe cortical pathology and epilepsy in multiple sclerosis. A 20-year study by Ziccardi and colleagues demonstrated that lesions seen at diagnosis predict long-term cognitive decline, while large cohort analyses have linked cortical damage to earlier transition to the progressive phase. Newer research from 2024 to 2026 extends this picture, showing that intrathecal inflammatory profiles combined with grey matter damage predict early progression independent of relapse activity, and that choroid plexus enlargement on 7 Tesla MRI correlates with cortical and paramagnetic rim lesions to flag the same risk.

Treatment effects on the cortical compartment are now being systematically mapped. Disease-modifying therapies show differential impacts: glatiramer acetate, dimethyl fumarate, natalizumab, fingolimod, ocrelizumab and cladribine have all been associated with reductions or stabilization in cortical lesion accumulation or grey matter atrophy in various studies, while sphingosine-1-phosphate modulators such as siponimod address progression in secondary progressive disease. Yet the review emphasizes that compartmentalized meningeal inflammation remains a therapeutic blind spot, and emerging data on Bruton’s tyrosine kinase inhibitors and lymphotoxin-driven meningeal injury pathways suggest that targeting this compartment could finally offer true neuroprotection.

The next frontier, the authors argue, lies in integrating single-cell RNA sequencing, spatial transcriptomics and quantitative MRI to define biologically meaningful cortical endotypes. Techniques such as myelin water imaging, magnetization transfer ratio, neurite orientation dispersion and density imaging, and quantitative susceptibility mapping are converting the cortex from a radiological blind spot into a measurable substrate of disease. Fluid biomarkers, including cerebrospinal fluid profiles of chemokines and osteopontin, complement these imaging advances by capturing the intrathecal inflammatory milieu that drives grey matter injury.

Artificial intelligence stands poised to multiply these gains. Deep learning algorithms trained on ultrahigh-field and synthetic double inversion recovery images are already improving cortical lesion detection across centers, while machine learning models that fuse imaging with high-dimensional blood signatures promise individualized prediction of disease trajectory and treatment response. If these tools mature, the review concludes, clinicians could move beyond population-level risk scores toward precision medicine — predicting each patient’s demyelination and repair capacity and tailoring therapies to the specific cortical endotype driving their disease. What was invisible for most of multiple sclerosis history may soon become its most informative biomarker.

Subject of Research: Cortical lesion pathology, imaging and clinical relevance in multiple sclerosis

Article Title: Cortical lesions in multiple sclerosis: 25 years of progress, remaining challenges and future directions

Article References: Calabrese, M., Magliozzi, R., Tamanti, A., Ziccardi, S., Eccher, C., Biasi, P., Pizzini, F. B., & Marastoni, D. (2026). Cortical lesions in multiple sclerosis: 25 years of progress, remaining challenges and future directions. Nature Reviews Neurology. https://doi.org/10.1038/s41582-026-01263-2

Image Credits: AI Generated

DOI: 10.1038/s41582-026-01263-2

Keywords: multiple sclerosis, cortical lesions, meningeal inflammation, MRI, 7 Tesla imaging, grey matter pathology, progression independent of relapse activity, cognitive impairment, disease-modifying therapy, artificial intelligence, fluid biomarkers, quantitative MRI