Accelerate Updates to AI Algorithms with PCCP: Webinar

We in MedTech are learning how to more rapidly evolve the software in our connected medical devices and Software as a Medical Devices (SaMD) while maintaining our quality standards for device safety and effectiveness.

Historically, it has been a challenge to minimize the regulatory burden of a full resubmission each time we’re ready to release a new software version. However, the ability of AI/ML to rapidly evolve far outpaces software since it does not require human intervention. This takes the opportunity and challenge of releasing frequent updates for medical devices to an entirely new level. Regulators and manufacturers need to reconcile the instantaneous speed of AI learning with our methodical approach to updating devices.

As a key step in this process, the FDA has created Predetermined Change Control Plans (PCCP), giving manufacturers more latitude to make post-market algorithm changes without resubmission as long as they fall within specific guardrails and scope parameters established during the original device submission. How can the industry work with regulators to leverage this new regulatory framework so we can “move faster and break nothing” with AI/ML algorithms and medical device software in general?

On June 28th, 2024, Orthogonal held a webinar looking at the potential impact of PCCP on software-enabled medical devices and SaMD. Expert speakers discussed the challenges organizations face when embracing PCCP, emerging best practices and the possibilities it opens up for medical device manufacturers going forward.

This webinar built on a recent white paper developed by a group of seasoned industry professionals co-convened by Orthogonal titled “Making the Significant Insignificant: Implementing a Predetermined Change Control Plan (PCCP) for Class II SaMD Products Beyond AI/ML.”

Video Recording

Key Takeaways Points

1. As a preface, the topic discussed in this webinar is based on emerging information gleaned from conversations with current and ex-FDA staff, others who have completed PCCP for medical devices and a collective understanding of the industry.
2. In a nutshell, a Predetermined Change Control Plan (PCCP) is a regulatory tool that allows manufacturers to refine and tune AI/ML algorithms through ongoing updates within the constraints of safety and effectiveness, without needing to refile with the FDA before release. PCCP has the potential to reduce the regulatory submission burdens for both manufacturers and regulators, but it has a steep learning curve for both parties.
3. AI/ML algorithms are currently the FDA’s top priority with PCCP, so our speakers focused on it for this discussion, but the legal language around PCCP put forth in the 2023 Ominus Bill does not limit it to AI/ML. In the future, PCCP may be used for other software and even hardware changes.
4. The three key elements of a PCCP are a list of modifications, predetermined protocols for developing, verifying, validating and implementing those modifications, and a process for assessing their impact. The level of detail that the FDA expects on these subjects  is significantly higher than in a typical pre-market submission. Yu Zhao noted that if a PCCP is less than ten pages, it’s probably not detailed enough. For a good starting point, we recommend this white paper and PCCP template from Ketryx.
5. PCCPs give manufacturers more latitude to make changes while staying within their device’s intended use and indications for use. For example, if your device was tested on patients up to 83 years of age, the FDA may limit your indication for use to that age group. If older patients use your device off-label, and provide real-world data on their usage, you may want to expand your indication to patients up to 95 years old. However, the PCCP as currently implemented by the FDA will likley not accommodate such a change.
6. A PCCP is not a pass to ignore existing processes when making updates to your device. For changes outlined in a PCCP, manufacturers will still need to follow their quality management system and collect clinical data when making changes that impact the device’s claims. They will also need to verify the data they use to test their algorithms, and ensure that modified software meets performance standards laid out in the PCCP.
7. To date, the FDA has not cleared an “unlocked” AI/ML algorithm; an algorithm that continuously updates itself in production. The PCCP draft guidance lays out a potential pathway for such algorithms, including outlining the modifications the algorithm will autonomously make and the modifications that humans will review. Yu Zhao believes that the FDA will likely favor clearing an unlocked algorithm under a PCCP rather than through a De Novo filing.
8. Assessing the risk/benefit of proposed changes to AI/ML algorithms in advance can be daunting, especially when these changes may not be predictable. Yu Zhao gave an example from a PCCP he worked on earlier in 2024. Instead of describing individual changes in detail, this PCCP described specific types of changes, which gives the manufacturers more flexibility in accommodating different modifications under the PCCP.
9. As PCCP is still in its draft state, the FDA may be less inclined to allow certain changes to be made without their oversight, particularly if no other PCCP have been authorized in that medical specialty. Our speakers hope and expect that over time, as regulators gain more confidence in reviewing and approving PCCP, they will be more accommodating and flexible with manufacturers.

Resources

• Making the Significant Insignificant: Implementing a Predetermined Change Control Plan (PCCP) for Class II SaMD Products Beyond AI/ML – White Paper from Orthogonal and MedSec
• Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence/Machine Learning (AI/ML)-Enabled Device Software Functions – Draft guidance from FDA
• FDA Predetermined Change Control Plan (PCCP): Compliance Guide and PCCP Template from Ketryx

Speakers

Yu Zhao, Founder and President, Bridging Consulting LLC

Yu Zhao leads a cutting-edge regulatory, quality, and clinical consulting firm dedicated to servicing AI startups and medical device companies. With a dedicated team of over 20 expert consultants, he provides invaluable services to clients across North America, Europe, and Asia Pacific. With over two decades of leadership experience in the medical device industry, Mr. Zhao’s career also includes a 16-year tenure at Medtronic. During this time, he held successive leadership roles, including that of director and interim vice president for several multi-billion-dollar business units.

Throughout his career, Mr. Zhao and his teams have secured more than 150 U.S. FDA approvals and clearances. Their success extends across diverse device classifications and submission categories, encompassing PMAs, PMA supplements, IDEs, 510(k)s and De Novos, including submissions for many AI/ML-enabled SaMDs and recent ones with Predetermined Change Control Plans.

Mr. Zhao’s academic background includes a BS in Electrical Engineering from Zhejiang University in China, complemented by an MBA and an MSc from Washington University in St. Louis, U.S.

Ashley Miller, Senior Director of Regulatory Affairs, Digital Diagnostics

Ashley Miller is the Senior Director of Regulatory Affairs at Digital Diagnostics, a medical device company that designs AI systems that can diagnose disease by analyzing high-quality images and received the first-ever FDA clearance to market a fully autonomous AI diagnostic platform. Ashley has over a decade of experience in regulatory affairs in the medical device industry and developing regulatory strategies for a variety of device classifications across the globe.

Ashley has extensive knowledge supporting software as a medical device, including AI/ML-enabled SaMDs, and imaging systems, and has led cross-functional teams throughout the product life cycle, from design and development through regulatory review, to ultimately gain market access.

Ashley’s academic background includes a BS in Nuclear, Plasma, and Radiological Engineering from the University of Illinois at Urbana-Champaign.

Bernhard Kappe, CEO and Founder, Orthogonal

Bernhard Kappe is the Founder and CEO of Orthogonal. For over a decade, Bernhard has provided thought leadership and innovation in the fields of Software as a Medical Device (SaMD), Digital Therapeutics (DTx) and connected medical device systems. As a leader in the MedTech industry, Bernhard has a passion for launching successful medical device software that makes a difference for providers and patients, as well as helping companies deliver more from their innovation pipelines. He’s the author of the eBook Agile in an FDA Regulated Environment and a co-author of the AAMI Consensus Report on cloud computing for medical devices. Bernhard was the founder of the Chicago Product Management Association (ChiPMA) and the Chicago Lean Startup Challenge. He earned a Bachelor’s and Masters in Mathematics from the University of Pennsylvania, and a Bachelor’s of Science and Economics from the Wharton School of Business.

Moderator

Randy Horton, Chief Solutions Officer, Orthogonal

Randy Horton is Chief Solutions Officer at Orthogonal, a software consulting firm that improves patient outcomes faster by helping MedTech firms accelerate their development pipelines for Software as a Medical Device (SaMD), digital therapeutics (DTx) and connected medical device systems. Orthogonal makes that acceleration happen by fusing modern software engineering and product management tools and techniques (e.g., Agile, Lean Startup, User-Centered Design and Systems Thinking) with the regulated focus on device safety and effectiveness that is at the heart of MedTech.

Horton serves as Co-Chair for AAMI’s Cloud Computing Working Group, as well as AAMI CR:510(2021) and the in-process Technical Information Report #115, all of which address how to safely move medical device computing functions into the cloud. He is a frequent speaker at conferences and webinars, including events hosted by AdvaMed, AAMI, HLTH, RAPS and the Human Factors and Ergonomics Society (HFES).