AI Tool Reveals Cortical Lesions on Standard MRI in Multiple Sclerosis

Clinicians have long recognized that lesions in the brain’s gray matter, or cortex, are central to progression, yet these lesions are largely invisible on conventional magnetic resonance imaging (MRI). This invisibility limits prognostic accuracy and masks disease activity that continues despite treatment targeting white matter. Researchers have now developed an artificial intelligence (AI) approach that reveals cortical lesions on standard, legacy MRI scans.

A University at Buffalo–led team created a suite of AI-based image-processing methods that extract information from relationships among multiple MRI contrasts. The workflow includes a new technique, multimodal cortical lesion enhancement (MMCLE), which synthesizes signals across sequences to highlight subtle cortical abnormalities that are not apparent on single images. The researchers describe the capability as an advance made possible by computational methods that can finally “pull out” previously hidden information in routine scans.

The methods were validated using legacy MRI from the Phase III U.S. Food and Drug Administration (FDA) regulatory ORATORIO clinical trial of ocrelizumab, which included more than 700 participants. While individual scans predominantly showed white matter lesions, the AI pipeline revealed an additional 15 to 20 cortical lesions per patient, totaling more than 11,000 cortical lesions across the dataset. The work demonstrates that generative AI can synthesize multiple images to expose tissue behavior that diverges from healthy cortex, enabling visualization of key indicators of MS progression and cognitive impairment.

The international collaboration involved academic and industry partners, including Genentech, the manufacturer of ocrelizumab. The findings were published in Communications Medicine on July 7, 2026, with the team noting potential value for re-analyzing past clinical trials and informing future study designs that more fully capture gray matter disease activity.

“Detecting previously invisible cortical lesions on conventional legacy MRI scans has major implications for MS research and clinical care. The ability to see for the first time these previously hidden indicators of MS disease progression, including cognitive impairment and disability, is an important advance,” said Robert Zivadinov, MD, Ph.D., SUNY Distinguished Professor of Neurology and director of the Buffalo Neuroimaging Analysis Center at the Jacobs School of Medicine and Biomedical Sciences, University at Buffalo.

“The ability to see for the first time these previously hidden indicators of MS disease progression, including cognitive impairment and disability, is an important advance,” said Zivadinov.

Related Links
University at Buffalo