All talks & lectures
Conference ·
MS prognosis and monitoring with AI
How artificial intelligence may support prognosis and monitoring while keeping outcomes, validation and bias in view.
ECTRIMS 2023
- Delivered
- Event
- ECTRIMS 2023
- Format
- Conference
- Source deck
- 20 slides
From the slides
Deck notes
This ECTRIMS presentation reviewed artificial intelligence for prognosis and monitoring in multiple sclerosis. It connected disease activity and MRI change with lessons from newer multimodal models.
The question
How can AI support useful monitoring without losing sight of clinical outcomes, validation and bias?
Deck outline
- Disease activity and baseline risk
- Clinical outcomes for prognosis
- MRI change as a monitoring signal
- Lessons from AI in other medical fields
- Multimodal and generative medical models
- Bias, evaluation and clinical use
Take-away
AI can broaden what is measured, but a model is only useful when its outcome is clinically meaningful and its performance has been tested for bias.