Curriculum
- 5 Sections
- 20 Lessons
- 1 Day
Expand all sectionsCollapse all sections
- AI/ML in Medical Devices5
- 1.1Regulatory definitions: AI/ML-based SaMD vs traditional software
- 1.2Locked algorithms, adaptive algorithms, Clinical Evaluation, GMLP (Good Machine Learning Practice)
- 1.3US & EU focus: FDA’s AI/ML Action Plan vs EU MDR/IVDR interpretations
- 1.4Explainability, Bias & Transparency & Cybersecurity by Design
- 1.5Adaptive Learning Control Plans & Real-World Evidence (RWE)
- US & EU Regulatory Pathways4
- 2.1US FDA: 510(k), De Novo, PMA routes for AI/ML; Predetermined Change Control Plan (PCCP)
- 2.2EU MDR/IVDR: Device classification, conformity assessments, notified body oversight.
- 2.3Comparative lens: FDA flexibility vs EU rigidity
- 2.4Common challenges: AI bias, validation datasets, lack of harmonised guidance, notified body capacity issues
- Lessons from Market Approvals4
- Interactive Exercises & Simulation3
- 4.1Hands-On Activity: Draft a mini Predetermined Change Control Plan (PCCP) for an adaptive AI model
- 4.2Notified body reviewer challenges an AI algorithm’s bias risk → participants defend compliance strategy
- 4.3Classify hypothetical AI devices (locked vs adaptive) and propose regulatory submission routes (FDA vs EU MDR)
- Wrap-Up & Key Takeaways4
- 5.1AI/ML devices face unique regulatory scrutiny due to adaptive learning
- 5.2US FDA is leading with proactive PCCP frameworks, while EU MDR remains rigid but evolving
- 5.3Early regulatory engagement, robust clinical evaluation, and cybersecurity evidence
- 5.4Adaptive AI needs a lifecycle regulatory strategy, not a one-time approval mindset