Artificial Intelligence and Multiple Sclerosis

Summary

This narrative review synthesises advances in artificial intelligence and machine learning applications across multiple sclerosis research and clinical practice, spanning disease pathogenesis, diagnosis, treatment selection, and prognostic prediction. The authors analyse how ML models leverage multimodal data sources—MRI, genetic, and clinical information—to distinguish MS from mimicking conditions, predict disease progression, and enable personalised treatment strategies. The review identifies model interpretability and clinical transparency as critical barriers to adoption, and proposes future directions including open data initiatives, federated learning, and generative AI to enhance explainability and facilitate clinical integration.

UK applicability

The review's findings on AI applications in MS diagnosis and prognostication are broadly applicable to UK clinical practice and the NHS, where such tools could support earlier diagnosis and individualised treatment decisions. However, successful implementation will depend on addressing data governance, NHS integration pathways, and ensuring that model transparency meets regulatory and clinical governance requirements.

Key measures

Application domains of AI/ML in MS (diagnosis, prognosis, lesion segmentation, biomarker identification); model interpretability and transparency metrics; potential future implementations (federated learning, generative AI approaches)

Outcomes reported

The review analysed advances in AI approaches across MS disease pathogenesis investigation, diagnosis, treatment, and prognosis prediction. It examined how machine learning models utilise MRI, genetic, and clinical data for MS distinction from other conditions, lesion segmentation, biomarker identification, and personalised treatment strategy development.

Topic tags