A discussion paper has been published by the US FDA that aims at complementing and informing future guidance on artificial intelligence (AI) and machine learning (ML) usage in drug development. The purpose of this paper is to initiate an open and constructive dialogue with stakeholders, including industry representatives and academia, to foster mutual learning and discussion on the subject.
Section B of the discussion paper contains a series of questions that are carefully formulated to stimulate engagement and solicit valuable feedback from stakeholders. These questions focus on three key areas related to the application of AI/ML in drug development-
- Human-led governance, accountability, and transparency
- Ensuring the quality, reliability, and representativeness of data
- Addressing model development, performance, monitoring, and validation
The FDA has observed a remarkable surge in the number of submissions referencing AI/ML in recent years, prompting the need for comprehensive guidance and regulation in this rapidly evolving field.
The discussion paper acknowledges the significant potential of AI/ML in the pharmaceutical manufacturing industry, particularly in the context of advanced manufacturing. It highlights that the integration of AI/ML with other advanced manufacturing technologies, such as process analytical technology and continuous manufacturing, can yield numerous benefits. These benefits include enhanced process control, prevention of batch losses, and the facilitation of Industry 4.0 implementation. The FDA emphasizes that the use of AI/ML in manufacturing has the potential to improve efficiency, reduce waste, support informed decision-making, and enhance quality control. Furthermore, the paper highlights the role of AI/ML in optimizing the manufacturing supply chain, thereby streamlining operations and ensuring product quality.
Another crucial area addressed in the discussion paper is the application of AI/ML in clinical research. The FDA acknowledges that AI/ML holds immense promise for streamlining and advancing clinical research efforts. One of the primary advantages of AI/ML identified in the paper is its capability to analyze vast amounts of data. For instance, digital health technologies (DHTs), such as wireless and smartphone-connected products, wearables, implantables, and ingestible, are increasingly being incorporated into clinical trials. These technologies generate large and diverse datasets, which can be effectively analyzed using AI/ML algorithms. The paper also signifies the AI and ML potential so as to inform the design and efficiency of non-traditional trials, including decentralized clinical trials and those integrating real-world data (RWD).
In line with its commitment to promoting innovation in AI/ML applications, the FDA recently published draft guidance on decentralized clinical trials in May 2023. This draft guidance reflects the agency’s intention to develop and adopt a flexible, risk-based regulatory framework that facilitates the integration of AI/ML in drug development while safeguarding patient safety and data integrity.
The FDA encourages stakeholders to actively participate in shaping future guidance and regulations by providing their feedback and comments on the discussion paper. By actively seeking input from a diverse range of perspectives and expertise, the FDA aims to ensure that any future guidance adequately addresses the opportunities, challenges, and potential risks associated with AI/ML in drug development.