Summary
“Clinicians need to to trust the (AI) algorithm, and knowing how it works helps them understand where it might fail, or where extra attention might be needed.”
In this episode of the Let's Talk Risk Podcast, Dr. Jay Vaishnav discusses the rapid growth of artificial intelligence (AI) in the medical technology (MedTech) field, particularly in Radiology.
She shares insights from her extensive experience at the FDA and in MedTech, both in roles ranging from scientific to medical affairs and regulatory affairs, emphasizing the importance of understanding the regulatory landscape, navigating benefit-risk assessments, and the challenges of AI model validation.
The discussion also covers innovations in triage and notification systems, the future of AI in clinical care, and best practices for regulatory professionals. Dr. Vaishnav highlights the need for transparency in AI development and the importance of involving clinicians early in the process. She concludes with career insights and the value of mentorship in navigating the evolving landscape of MedTech.
Listen to the full 30-minute podcast or jump to a section of interest listed below.
Chapters
00:00 Introducing Dr. Jay Vaishnav
01:50 The Rise of AI/ML devices in MedTech
04:40 FDA perspective on benefit-risk
06:05 Managing risk of clinician over-reliance on AI
08:25 Case study: De Novo granted for a triage & notification AI device
11:56 Challenges in developing triage & notification AI devices
14:00 Current stat and barriers to adoption of AI technologies in healthcare
15:35 The need for explainability in AI/ML devices
17:16 Future directions for AI/ML in MedTech
18:00 Why clinician involvement in AI/ML development is important
19:12 Best Practices for QA/RA professionals
21:30 Career insights and power of mentoring
27:46 Closing comments
Suggested links:
* DEN 170073 - ContaCT viz.AI, Inc..
* LTR: AI in MedTech.
* LTR: Regulating Generative AI.
Key Takeaways
* AI is growing rapidly in MedTech, especially in radiology applications.
* Regulatory considerations are crucial for AI applications.
* Benefit-risk assessments are complex and vary by device indication.
* Triage applications can significantly improve patient outcomes.
* AI models face challenges with false positives and negatives.
* Cultural mistrust of AI algorithms can hinder adoption.
* Involving clinicians early in AI development is essential.
* Transparency about AI limitations is necessary for trust.
* The future of AI in healthcare is promising but uncertain.
* Career growth opportunities exist for regulatory professionals in AI.
Keywords
AI, MedTech, FDA, regulatory affairs, machine learning, healthcare, medical devices, radiology, clinical applications, patient safety.
About Jay Vaishnav, Ph.D.
Dr. Jay Vaishnav is currently Director of Regulatory Affairs at Canon Medical Informatics, where she leads US regulatory strategy and FDA submissions for Canon's portfolio of Healthcare IT solutions. She holds a Ph.D. from Harvard in theoretical physics, and after some time in academia moved to the FDA, spending over seven years there in positions of increasing responsibility. She eventually joined the medical device industry in Medical Affairs before moving into a Regulatory role.She is a Fellow of the Regulatory Affairs Professional Society, an occasional instructor at UCSC Silicon Valley Extension, and co-editor of the book "From X Rays to AI: Navigating US Regulations in Radiological Health."
Disclaimer
Information and insights presented in this podcast are for educational purposes only. Views expressed by all speakers are their own and do not reflect those of their respective organizations.
Let's Talk Risk! is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit naveenagarwalphd.substack.com/subscribe