On March 1, 2023, FDA’s Center for Drug Evaluation and Research (CDER) published a discussion paper on artificial intelligence (AI) in drug manufacturing.  The publication acknowledges the many possible applications and benefits AI offers in a manufacturing context and identifies regulatory questions focused on five areas where CDER is seeking feedback for potential policy development.

The discussion paper is part of FDA’s wider efforts to facilitate the implementation of new pharmaceutical product manufacturing technologies to improve drug quality, address shortages of medicines, and reduce time to market (advanced manufacturing).

Key highlights

CDER is starting to grapple with the broader implications of AI-enabled pharmaceutical manufacturing. Currently, the Office of Pharmaceutical Manufacturing Assessment (OPMA), within the Office of Pharmaceutical Quality, is responsible for FDA regulatory oversight of advanced manufacturing technologies, and such technologies are reviewed as part of assessment and inspectional activities connected to new product approvals.

CDER appreciates the emerging role for AI in this space and has been actively analyzing existing regulatory requirements. The newly published discussion paper focuses on the manufacture of drug products that would be marketed under a New Drug Application (NDA), Abbreviated New Drug Application (ANDA), or Biologics License Application (BLA) and identifies five areas of consideration, each presented with a list of associated regulations and guidance, for potential policy development.

The five areas of consideration called out in the discussion paper are as follows:

• Cloud applications may affect oversight of pharmaceutical manufacturing data and records.
• The internet of things may increase the amount of data generated during pharmaceutical manufacturing, affecting existing data management practices.
• Applicants may need clarity about whether and how the application of AI in pharmaceutical manufacturing is subject to regulatory oversight.
• Applicants may need standards for developing and validating AI models used for process control and to support release testing, and;
• Continuously learning AI systems that adapt to real-time data may challenge regulatory assessment and oversight.

We note that many of these areas of consideration bear a striking resemblance to issues that FDA’s Center for Devices and Radiological Health (CDRH) has been tackling in relation to AI-enabled medical devices for the past few years. We anticipate that thinking and findings from the medical device world may be informative in the pharmaceutical manufacturing space. For example, might predetermined change control plans[1] be deployed for continuously learning AI systems in drug manufacturing, in addition to the regulation of AI-enabled medical devices?

The paper also raises familiar and critical issues of data integrity, explainability, and ability to transfer learning between models, applying them to the manufacturing side of the house.  Does this portend the same level of expected validation of AI models that FDA has come to expect of software as a medical device (SaMD) and clinical decision support software (CDS)?

And, while the discussion paper includes a broad swath of manufacturing issues in its scope, including AI-enabled bioinformatics pipelines to “generate and select candidates for precision medicine complex biological products,” also of note is what is not included: AI in drug supply chain management – an issue of particular and recent importance in the pandemic and its global disruption of supply chains, including pharmaceutical ones.  Will separate guidance for AI-enabled supply chain management be forthcoming?

CDER notes that, while there are areas of consideration not covered in the paper, the five areas presented are those where CDER thinks public feedback would be valuable. Arising from these five considerations, CDER is seeking stakeholder input on the following eight questions:

• What types of AI applications do you envision being used in pharmaceutical manufacturing?
• Are there additional aspects of the current regulatory framework (e.g., aspects not listed above) that may affect the implementation of AI in drug manufacturing and should be considered by FDA?
• Would guidance in the area of AI in drug manufacturing be beneficial? If so, what aspects of AI technology should be considered?
• What are the necessary elements for a manufacturer to implement AI-based models in a CGMP environment?
• What are common practices for validating and maintaining self-learning AI models and what steps need to be considered to establish best practices?
• What are the necessary mechanisms for managing the data used to generate AI models in pharmaceutical manufacturing?
• Are there other aspects of implementing models (including AI-based models) for pharmaceutical manufacturing where further guidance would be helpful?
• Are there aspects of the application of AI in pharmaceutical manufacturing not covered in [the discussion paper] that FDA should consider?

The above questions are not meant to be exhaustive. CDER is interested in any pertinent information the public may have to share arising from the discussion paper. The publication of previous discussion papers in the advanced manufacturing space have been followed by FDA workshops, so a public meeting announcement on AI in drug manufacturing may very well be on the horizon.

Advanced manufacturing momentum

The release of the discussion paper on AI in drug manufacturing is just one example of FDA activity in advanced manufacturing policy development. On March 1, 2023, FDA also issued final guidance on continuous manufacturing of drug substances and drug products, implementing the International Council for Harmonisation (ICH) guideline on the subject.[2]

The pace of policy addressing advanced manufacturing technologies is likely to increase, building upon the 2014 CDER Emerging Technology Program and the 2021 National Academies of Sciences, Engineering, and Medicine (NASEM) report on advanced drug manufacturing.

Further activity, anticipated this coming year, as reflected in the recently enacted Food and Drug Omnibus Reform Act,[3] includes:

• A framework for FDA to designate up to five eligible institutions as National Centers of Excellence in Advanced and Continuous Pharmaceutical Manufacturing. Such centers will collaborate with FDA, sharing and publishing research, strategic plans, and best practices for the progression of advanced and continuous manufacturing, and
  
  
• An “Advanced Manufacturing Technologies Designation Pilot Program” to expedite development and review of applications using the designated technology. FDA is required to hold a public meeting to discuss the program and then issue guidance.

Interested in providing comments to FDA?

The comment period for the discussion paper ends May 1, 2023. If you are interested in submitting information to FDA or have any questions about AI in drug manufacturing or related matters, please reach out to the authors of this alert.

[1] FDA has sent draft guidance on Marketing Submission Recommendations for A Change Control Plan for AI/ML-Enabled Device Software Functions to the White House for review and potential clearance, indicating that the document will soon be ready for publication.

[2] FDA has been engaged in developing an ICH guideline to standardize terminology and regulatory approach to continuous manufacturing (ICH guideline Q13) for a number of years. On November 16, 2022, ICH finalized and adopted ICH guideline Q13. FDA implemented ICH guideline Q13 by publishing it as final guidance.

[3] Part of the Consolidated Appropriations Act for 2023, also known as the omnibus appropriations or funding bill.

For more industry insights from DLA Piper’s Life Sciences team, click here to subscribe to the sector’s dedicated blog, Cortex.