The Artificial Intelligence (AI) Program in the FDA’s Center for Devices and Radiological Health (CDRH) conducts regulatory science research to ensure patient access to safe and effective medical devices using AI/ML. This is one of 20 research programs in CDRH’s Office of Science and Engineering Laboratories (OSEL).
Artificial Intelligence, Machine Learning, and Medical Devices
AI technologies are transforming healthcare by producing diagnostic, therapeutic and prognostic medical recommendations, or decisions, in some cases autonomously, informed by the vast amount of data generated during the delivery of healthcare. In medical devices, application areas include:
- Image acquisition and processing
- Early disease detection
- More accurate diagnosis, prognosis, and risk assessment
- Identification of new patterns in human physiology and disease progression
- Development of personalized diagnostics
- Therapeutics treatment response monitoring
The breadth of applications has continually and rapidly increased in the last few years and is not predicted to decelerate in the near future. The rigorous and least burdensome evaluation of the safety and effectiveness of these products is the focus of this program.
The application of AI-based technology into many different clinical areas combined with the unique nature of clinical medical data (e.g., low prevalence of disease and lack of or difficulty in obtaining truth data), produces many challenges in developing robust evaluation methods and understanding the effects of AI-based technology in real-world settings. Moreover, AI-enabled medical devices can be designed to continuously learn, update, and adapt based on the availability of more data or to respond to changes in the data. This ability presents unique regulatory challenges for CDRH with a need to develop appropriate regulatory controls and test methods that balance the potential benefits and risks of AI adoption in the clinic.
Regulatory Science Gaps and Challenges
Major regulatory science gaps and challenges that drive the Artificial Intelligence Program are:
- Lack of methods that can enhance AI algorithm training for limited labeled training and test data
- Lack of methods to analyze training and test methods to understand, measure, and minimize bias of AI-enabled devices
- Lack of metrics for performance estimation, reference standards, and uncertainty of AI devices
- Lack of methods to evaluate the safety and effectiveness of continuously learning AI algorithms
- Lack of methods to evaluate the safety and effectiveness of emerging clinical applications of AI-enabled medical devices
- Lack of methods for post-market monitoring of AI devices
The Artificial Intelligence Program is intended to fill these knowledge gaps by developing robust AI test methods and evaluation methodologies for assessing AI performance both in premarket and real-world settings to reasonably ensure the safety and effectiveness of novel AI algorithms.
Artificial Intelligence Program Activities
The Artificial Intelligence Program focuses on regulatory science research in these areas:
- Data augmentation and synthetic data to enhance algorithmic training and testing and to address the limitations of small and fragmented patient datasets for the training and testing of medical AI models
- Methods to measure and quantify algorithmic bias, reduce performance difference among subpopulations, and ensure generalizability
- Study designs and analysis methods for AI/ML-based computer-aided triage (CADt).
- New evaluation frameworks for continual learning and multi-class systems
- Training for robustness and metrics for assessment of machine learning algorithms
- Assessment techniques for evaluating the reliability of adaptive AI/ML algorithms to support non-clinical test method development.
- Assessment approaches to estimate and report the robustness of AI/ML to variation in data acquisition factors.
For more information, email OSEL_AI@fda.hhs.gov