In the most recent episode of our podcast, The Life Science Rundown, The FDA Group's Nick Capman sat down with Adam Foresman, the Head of Quality & Regulatory Affairs at VideaHealth, to explore the intersection of AI, machine learning, and diagnostics within the FDA's regulatory environment. Adam brings over 15 years of experience in medical device design, quality management systems, and regulatory affairs, with a recent focus on AI in the dental diagnostic market.
They discuss regulatory challenges and opportunities, particularly focusing on navigating FDA approvals for software as a medical device (SaMD). Adam shares insights on ensuring AI transparency, managing cybersecurity, and the importance of robust quality management systems. The conversation also explores the future of AI in diagnostics and the critical role of post-market surveillance in maintaining patient safety and trust.
Do you need to access specialized SaMD regulatory expertise to navigate the complexities of AI and machine learning? Contact us. We’re helping innovative medtech companies find and partner with these hard-to-find regulatory experts to bring their products to market and keep them there.
Apple | Spotify | YouTube | Web + Others
Summary
Adam begins by explaining the growing role of AI and ML in healthcare, particularly diagnostic aids. He emphasizes how these technologies help identify previously missed detections, enhance patient education, and streamline administrative tasks. Early adopters in healthcare are reporting significant benefits, which is encouraging broader acceptance of AI technologies. Adam notes that AI is particularly valuable in improving diagnostic accuracy and patient engagement, as AI tools can provide detailed explanations and visualizations that help patients understand their conditions and the proposed treatments.
Much of the discussion focuses on VideaHealth's recent achievement of obtaining 510(k) clearance for a dental AI application for pediatric patients. Adam provides practical tips for successful software as a medical device (SaMD) submissions.
He stresses the importance of engaging with the FDA early through Q-submissions to discuss AI design, indications for use, and testing methodologies. Adam explains that these early interactions with the FDA can help clarify expectations and align on critical aspects of the submission, reducing the risk of delays or rejections.
Adam emphasizes that comprehensive documentation is also crucial, especially regarding cybersecurity and post-market surveillance. He explains that the FDA has been emphasizing these areas more, given the potential risks associated with AI and ML technologies. Manufacturers should assume that the FDA will ask for detailed documentation on cybersecurity measures, even if these aspects have not changed significantly since previous submissions.
Adam also highlights the importance of thorough documentation on post-market surveillance processes, as the FDA wants to ensure that manufacturers can quickly identify and address any issues that arise after commercialization.
Adam shares some specific considerations for pediatric applications, noting that the FDA has a guidance document on premarket pediatric considerations for medical device design. He advises manufacturers to comply with this guidance, which includes detailed risk management techniques and design control processes tailored to pediatric patients. This ensures that the devices are safe and effective for younger populations.
Another key topic is the concept of Predetermined Change Control Plans (PCCPs). Adam explains that PCCPs allow manufacturers to include future design changes in pre-market submissions, potentially reducing the need for re-submissions. However, he acknowledges the challenges AI manufacturers face in predicting future design changes accurately. Unlike traditional electromechanical devices, where future changes might be more predictable, AI models often evolve in ways that are harder to foresee. Adam recommends considering PCCPs for well-defined future changes, particularly when manufacturers can anticipate specific updates or improvements.
Adam also compares the regulatory approaches of the US and the EU. Both regions aim to ensure that clinical benefits outweigh risks, focusing on safety and cybersecurity. The FDA emphasizes detailed testing and adherence to good machine learning practices (GMLP), while the EU focuses more on risk management processes and post-market surveillance. Adam praises the EU’s emphasis on postmarket surveillance, suggesting that the FDA could benefit from a similar approach. He believes that robust postmarket surveillance is crucial for quickly identifying and addressing any issues that arise once a product is on the market.
AI transparency and bias considerations are becoming increasingly important. Adam explains that the FDA emphasizes AI transparency, generalizability, and bias considerations. Manufacturers must submit performance data for different subgroups, such as age and gender, to ensure that the AI performs consistently across diverse populations. Adam anticipates that the FDA will develop more detailed guidelines for AI transparency, possibly requiring specific data presentation formats that make it easier for users to understand AI performance metrics.
Regulatory affairs professionals are crucial in ensuring AI transparency, ethics, and patient-centric design within project teams. Adam advises these professionals to include AI transparency requirements in design inputs and maintain robust quality management system (QMS) processes for cybersecurity and post-market surveillance. This will become increasingly important in regulatory reviews. He also stresses the importance of putting oneself in the shoes of clinicians and patients to ensure that AI outputs are understandable and useful.
Looking to the future, Adam predicts an increase in AI submissions bundling multiple models. He emphasizes the importance of ensuring each model is independent to simplify testing and regulatory approval. Early engagement with the FDA to discuss AI designs, outputs, and testing methodologies is also crucial. Maintaining strong QMS processes, particularly for cybersecurity and post-market surveillance, will be vital as these areas gain more attention in regulatory reviews.
Adam encourages participation in industry committees developing AI-specific standards, such as those led by the American Dental Association for dental AI. Engaging with the FDA's Digital Health Steering Committee, which includes cross-functional groups, can help shape AI regulations and standards. These collaborative efforts are essential for developing practical and effective regulatory frameworks that keep pace with rapid technological advancements.
Not a paid subscriber yet?
Here’s what you’re missing:
Upgrade here to unlock all of these issues and new ones »
The FDA Group helps life science organizations rapidly access the industry's best consultants, contractors, and candidates.
Need expert help planning and executing compliance projects or connecting with a quality and compliance professional? Get in touch with us. If we haven’t worked together yet, watch our explainer video below or head to our company introduction page to see if we could help you execute your projects now or in the future.
Our service areas:
Quality Assurance | Regulatory Affairs | Clinical Operations | Commissioning, Qualification, and Validation | Chemistry, Manufacturing, and Controls (CMC) | Expert Witness
Our engagement models:
Consulting Projects | Staff Augmentation | FTE Recruitment | Functional Service Program
Our podcast:
Apple | Spotify | YouTube | Web + Others
Navigating FDA Regulations for SaMD AI and Machine Learning with VideaHealth's Adam Foresman