In its most recent effort to keep pace with advancing technology, the US Food and Drug Administration (FDA) recently issued two draft guidances on the use of artificial intelligence (AI) in the context of drugs, biologics, and medical devices.

Medical Device Guidance

The first draft guidance is entitled “Artificial Intelligence-Enabled Device Software Functions: Lifecycle Management and Marketing Submission Recommendations.” The guidance provides an overview of the type of documentation and information that companies will need to submit to the FDA to obtain medical device regulatory approval — part of the so-called “marketing submission” process. Among other things, the FDA advises that such documentation and information should include:

• A description of the device inputs and outputs, including whether inputs are entered manually or automatically.
• A description of how AI is used to achieve the device’s intended purpose.
• A description of the device’s intended users, their characteristics, and the level and type of training they are expected to have and/or receive.
• A description of the intended use environment (e.g., clinical setting, home setting).
• A description of the degree of automation that the device provides in comparison to the workflow for the current standard of care.
• A comprehensive risk assessment and risk management plan.
• Data management information, including how data were collected, limitations of the dataset, and an explanation of how the data are representative of the intended use population.
• A cybersecurity assessment, particularly focusing on those risks that may be unique to AI.

Guidance for Drugs and Biological Products

The second draft guidance issued this month is entitled “Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products.” The guidance addresses considerations for the use of AI to support regulatory decision-making for drugs and biologics. Specifically, the draft guidance discusses the use of AI models to produce information or data to support regulatory decision-making regarding safety, effectiveness, or quality for these products. To that end, the FDA recommends utilizing the following seven-step process to establish and assess the credibility of an AI model output for a specific context of use (COU) based on model risk:

• Step 1: Define the question of interest that will be addressed by the AI model.
• Step 2: Define the COU for the AI model.
• Step 3: Assess the AI model risk.
• Step 4: Develop a plan to establish the credibility of AI model output within the COU.
• Step 5: Execute the plan.
• Step 6: Document the results of the credibility assessment plan and discuss deviations from the plan.
• Step 7: Determine the adequacy of the AI model for the COU.

Each of these steps is discussed in detail in the draft guidance, with examples provided. The FDA states that this will provide a “risk-based credibility assessment framework” that will help manufacturers and other interested parties plan, gather, organize, and document information to establish the credibility of AI model outputs when the model is used to produce information or data intended to support regulatory decision-making.

Conclusion

The new guidances from the FDA provide further indication that the agency is closely scrutinizing the use of AI in and in connection with FDA-regulated medical devices, drugs, and biological products. To reduce the risk of unnecessary regulatory delays, companies seeking approval of FDA-regulated products should carefully review their regulatory submissions to ensure they align with the new AI guidance documents.