FDA Proposing New Rules as AI Software Is Evolving into Medical Devices
An AI and machine learning “action plan” released this week by the U.S. Food and Drug Administration (FDA) would advance the agency’s monitoring of AI-based medical software which is increasing being seen as virtual medical devices.
The proposed regulatory framework focuses on software used in medical devices, including guidance on how to handle software upgrades as algorithms learn over time. To that end, the FDA said it would promote best practices for machine learning development as a means of evaluating and improving algorithms incorporated into medical devices.
The initiative also would seek to promote “device transparency for users,” the agency added, as a way of fostering a “patient-centered approach” to healthcare AI and machine learning.
The framework also includes pilot projects in clinical settings as a way of evaluating and improving machine learning algorithms used in future medical devices.
“One of the greatest benefits of AI/ML in software resides in its ability to learn from real-world use and experience, and its capability to improve its performance,” the agency noted in proposed AI/ML rulemaking released in April 2019.
“The ability for AI/ML software to learn from real-world feedback (training) and improve its performance (adaptation) makes these technologies uniquely situated among software as a medical device,” or SaMD. That medical device label is roughly equivalent to the enterprise category, software-as-a-service.
The agency’s AI regulatory effort moved a step further last fall with the launch of the Digital Health Center for Excellence within the FDA’s Center for Devices and Radiological Health.
“This action plan outlines the FDA’s next steps towards furthering oversight for AI/ML-based SaMD,” said Bakul Patel, the center’s director. “The plan outlines a holistic approach based on total product lifecycle oversight to further the enormous potential that these technologies have to improve patient care while delivering safe and effective software functionality that improves the quality of care….”
The FDA expects its AI/ML plan “will continue to evolve over time,” Patel added.
The need for a regulatory framework that would oversee deployment of software based on AI algorithms is growing as healthcare startups attract investors and accelerate development. For instance, healthcare AI platform developer Lumiata Inc. announced a $14 million funding round this week. The startup’s platform transforms disparate medical data sets into “machine learning-ready” software.
The machine learning tools are “purpose-built for healthcare data scientists to build and deploy predictive models,” the startup said.
Meanwhile, AI leaders such as IBM and Nvidia also have established healthcare and medical device units that incorporate early machine learning algorithms and AI toolkits. Nvidia has forged partnerships with groups like the American College of Radiology to leverage AI for diagnostic imaging based on clinical data.
"We appreciate the FDA’s focus on transparency, which will be crucial for AI adoption in healthcare," said Dr. Mona Flores, Nvidia's global head of medical AI.
--Editor's note: This story has been updated with comment from Nvidia.