FDA Finalizes Guidance to Streamline Updates for AI-Powered Medical Devices

In December 2024, the U.S. Food and Drug Administration (FDA) finalized its highly anticipated guidance on Marketing Submission Recommendations for a Predetermined Change Control Plan (PCCP) for Artificial Intelligence/Machine Learning (AI/ML)-Enabled Device Software Functions. The guidance introduces a regulatory mechanism allowing manufacturers to proactively define and seek pre-authorization for software updates to AI-enabled medical devices, without needing to file a new premarket submission for every change.

The new framework is a pivotal step in modernizing medical device regulation to keep pace with the dynamic and iterative nature of AI. It offers a formal pathway for adaptive AI/ML-enabled devices to evolve safely post-market while reducing regulatory friction that could otherwise delay critical innovations reaching patients.

What is a PCCP?

A Predetermined Change Control Plan (PCCP) is a section of a marketing submission that details proposed future modifications to an AI/ML-enabled device’s software that are anticipated post-clearance or approval. It includes:

• Description of Modifications: The types of intended changes to the device’s performance, inputs, or algorithm logic (e.g., expanding a training dataset, improving image segmentation accuracy).

• Methodology for Implementing Changes: Technical and procedural information about how the modifications will be validated and integrated (e.g., retraining workflows, version control).

• Performance Evaluation Protocols: Statistical and clinical performance testing methods that will be used to assess whether changes maintain safety and effectiveness.

• Update Procedures: How the changes will be deployed in a controlled, traceable, and secure manner, including cybersecurity considerations.

  
  

Submitting a PCCP to the FDA

PCCPs can be included in the original marketing application, whether a 510(k), De Novo request, or PMA, or in a supplemental application for an already authorized device. Manufacturers should:

1. Clearly Separate PCCP Content: The PCCP must be distinguishable from the current version of the device and any near-term changes that would otherwise be part of the traditional premarket review.

2. Provide Robust Rationale: Each proposed change must include a justification showing it can be verified and validated through the methods described in the PCCP.

3. Labeling Updates: Manufacturers must indicate how changes resulting from the PCCP will be communicated to end users, including revised instructions for use (IFU) and updated product labeling.

4. Postmarket Oversight: While changes implemented under an approved PCCP do not require a new submission, FDA expects ongoing monitoring and documentation to ensure continued compliance.

   
   

Once approved, a PCCP provides a streamlined path for deploying the outlined changes. This can significantly reduce regulatory burden, especially for developers of continuously learning or performance-adaptive algorithms in diagnostics, imaging, and decision support tools.

Industry Implications

The finalized PCCP guidance reflects years of stakeholder engagement and public comment following the FDA’s 2019 discussion paper on AI/ML-based Software as a Medical Device (SaMD). It aligns with the agency’s broader commitment to enabling safe innovation in digital health.

Professional groups like AdvaMed have praised the guidance, noting that it provides “predictability and clarity” for developers navigating regulatory pathways for AI-enabled devices .

This framework also places the FDA at the forefront of global AI regulation. It may serve as a model for regulators in the EU and elsewhere as they consider how best to manage the unique lifecycle of learning health technologies.

A Polaris Biomedical Perspective

At Polaris Biomedical, we see this guidance as a milestone that enables the safe and timely evolution of AI-driven health solutions. Too often, the speed of innovation in software outpaces traditional regulatory processes, especially in high-impact domains like cancer diagnostics, cardiovascular monitoring, and imaging.

The PCCP framework gives companies the flexibility to keep their devices clinically relevant while maintaining transparency and traceability. From a quality systems standpoint, this guidance also emphasizes the importance of integrating rigorous software lifecycle planning into risk management and design control.

For manufacturers exploring AI/ML pathways, now is the time to assess your product roadmaps for features or improvements that could be pre-authorized under a PCCP. Doing so not only accelerates your time to market but also ensures sustained alignment with FDA expectations.

Need help navigating the submission process or integrating PCCPs into your quality system? Polaris Biomedical offers regulatory strategy, risk management, and quality assurance support tailored to high-tech and AI-enabled medical devices. Reach out to learn how we can help your product stay compliant and competitive.

Sources

[1] FDA. (2024). Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence/Machine Learning (AI/ML)-Enabled Device Software Functions. Link to guidance

[2] AdvaMed. (October 2024). Principles for Artificial Intelligence in Medical and Digital Health Technologies. Advanced Medical Technology Association. Link to press release