Transforming multiple sclerosis care using AI: prospects and challenges

This talk considered where artificial intelligence might change multiple sclerosis care, using disease classification as a practical example. It also examined the distance between promising research and routine clinical use.

The question

Can AI move MS classification closer to disease biology while producing evidence that is useful to patients and clinicians?

Deck outline

• A working distinction between artificial intelligence, machine learning and deep learning
• The clinical and regulatory landscape for AI-based medical software
• Why conventional clinical MS types do not fully reflect disease biology
• Data-driven MRI subtypes and single-sequence image analysis
• Lessons from randomised evaluations of medical AI
• Multimodal systems, bias and the path to clinical translation

Take-away

AI can reveal patterns that conventional labels miss, but classification alone is not clinical impact. Translation depends on representative data, prospective evaluation and evidence that an AI-supported decision improves care.