FDA to Establish Oversight Rules for AI in Medicine

The FDA review may include examination of the underlying performance of a product’s algorithms, a manufacturer’s plan to make modifications, and the manufacturer’s ability to manage the risks associated with any modifications.

The FDA has already approved medical devices that rely on “locked algorithms,” which do not change each time they are used, but instead are changed by a manufacturer at intervals, using specific training data and a validation process to ensure proper functioning of the system. Among such devices approved during 2018 were a device that is used to detect degenerative diabetic retinopathy, and another one designed to alert providers of a potential stroke in patients.

According to the FDA, the proper performance of those locked algorithms, and others like them, is crucial to ensuring that doctors base life-and-death treatment decisions on accurate information. That task is harder for products that learn and evolve on their own, in ways that are difficult even for the manufacturers of such systems to understand. An example of such one system uses AI algorithms to identify breast cancer lesions on mammograms and learns to improve its confidence and identify new subgroups of cancer, based on its exposure to additional real world data.

“A new approach to these technologies would address the need for the algorithms to learn and adapt when used in the real world. It would be a more tailored fit than our existing regulatory paradigm for software as a medical device,” explained FDA outgoing commissioner Scott Gottlieb, MD. “I can envision a world where, one day, artificial intelligence can help detect and treat challenging health problems, for example by recognizing the signs of disease well in advance of what we can do today.”

The FDA recently launched a fellowship program with Harvard University (Boston, MA, USA) on AI and machine learning, which is focused on designing, developing, and implementing algorithms for regulatory science applications. One such example is innovative clinical decision support software that uses AI algorithms to help alert neurovascular specialists of brain deterioration.

Related Links:
US Food and Drug Administration