The US Food and Drug Administration (FDA) has released a white paper as the first step to a new regulatory framework for developing safe and effective medical devices using advanced artificial intelligence (AI).
The FDA has already approved AI devices to detect diabetic retinopathy and potential stroke in patients. However, it said, these devices were “locked” algorithms, modified by the manufacturer at intervals (including “training” of the algorithm using new data), followed by manual verification and validation of the updated algorithm.
“But there’s a great deal of promise beyond locked algorithms,” FDA Commissioner Dr Scott Gottlieb said in a statement. “Machine learning algorithms that continually evolve, often called “adaptive” or “continuously learning” algorithms, don’t need manual modification to incorporate learning or updates.”
Currently for software modifications to a medical device, a sponsor must make a submission to the FDA demonstrating its safety and effectiveness. However, the FDA said, it was now working to develop an appropriate framework allowing the software to evolve to improve its performance while ensuring changes met its “gold standard for safety and effectiveness throughout the product’s lifecycle, from premarket design throughout the device’s use on the market.”
Informed by the feedback it received on the discussion paper, the FDA said it would take further steps, including issuing draft guidance. “Our approach will focus on the continually-evolving nature of these promising technologies. We plan to apply our current authorities in new ways to keep up with the rapid pace of innovation and ensure the safety of these devices.”
