ARTIFICIAL INTELLIGENCE is increasingly being explored as a tool to improve diagnosis, monitoring, and prognostic assessment in central nervous system (CNS) autoimmune diseases, according to a new review of current evidence.
The review examined applications across multiple sclerosis, neuromyelitis optica spectrum disorder, autoimmune encephalitis, and myelin oligodendrocyte glycoprotein antibody associated disease.
These conditions often present diagnostic and therapeutic challenges because of overlapping clinical features, imperfect biomarkers, and the need for long term follow up.
Researchers reported that artificial intelligence has been evaluated across several clinical tasks, including automated neuroimaging analysis, differential diagnosis, disease classification, relapse prediction, prognostic stratification, and risk informed care planning.
The technology is also being applied to laboratory data, electronic health records, molecular datasets, and multimodal sources of patient information.
Artificial Intelligence Evidence Strongest in Multiple Sclerosis
Evidence was found to be unevenly distributed across diseases. Multiple sclerosis accounted for the largest body of published research, particularly in neuroimaging applications such as lesion segmentation, disease activity monitoring, volumetric assessment, and disability prediction.
Some imaging tools have progressed to multicentre and prospective workflow evaluation.
By contrast, studies in neuromyelitis optica spectrum disorder, autoimmune encephalitis, and myelin oligodendrocyte glycoprotein antibody associated disease were generally smaller and more heterogeneous.
Although several studies demonstrated encouraging diagnostic and prognostic performance, many lacked external validation and relied on retrospective datasets.
The review highlighted the limited representation of diagnostically uncertain, seronegative, and overlap populations in current research.
The researchers also noted growing interest in advanced applications, including federated learning, digital biomarkers, multi omics integration, foundation models, and large language models.
However, these approaches currently provide hypothesis generating or exploratory evidence rather than proven clinical benefit.
Key Challenges for Clinical Translation
Despite increasing progress, important translational gaps remain. No prospectively validated artificial intelligence model was identified for comparative treatment effect estimation or individual treatment selection across the diseases reviewed.
Most available models focus on diagnosis, monitoring, and risk prediction rather than guiding therapeutic decisions.
The review emphasised the need for larger multicentre datasets, stronger external validation, prospective clinical evaluation, and assessment of real-world impact on clinical decision making and patient outcomes.
According to the researchers, near term progress is most likely to come from multimodal data integration, standardised imaging analysis, and improved risk stratification, while autonomous diagnosis and treatment selection remain unsupported by the current evidence base.
Reference
Cao Z et al. Artificial intelligence in central nervous system autoimmune diseases: from clinical decision support to mechanistic discovery and translation. EMBO Mol Med. 2026;DOI:10.1038/s44321-026-00527-0.
