Radiology societies weigh in on FDA’s push to regulate generative AI
Two leading American radiology societies are weighing in on the U.S. Food and Drug Administration’s push to regulate generative artificial intelligence.
Back in August, the federal agency asked for feedback from the public on how to govern this emerging technology, which can generate content such as images and audio. The FDA recently closed the comment period, receiving over 300 responses, radiologist and Digital Health Center of Excellence Director Rick Abramson, MD, reported.
Both the American College of Radiology and the Radiological Society of North America recently shared their comments online. The two trade groups are emphasizing the need to prioritize patient safety while working closely with the radiology community to develop guidance and related policies.
“Patient safety should remain the central organizing principle, and regulatory flexibility should be paired with clear, enforceable safeguards that promote transparency, accountability, clinical validation, and continuous real-world monitoring,” RSNA wrote in its comments, submitted Oct. 19.
The society notes that generative AI models will have the potential to evolve after deployment, creating risks that “cannot be addressed” through a one-time premarket review. Instead, RSNA wants regulators to develop a plan for continuously monitoring new gen AI products, including model performance, human-AI interaction, workflow effects and appropriate use.
The radiology group believes that AI manufacturers—rather than radiologists—must remain accountable for device safety and effectiveness.
“Healthcare organizations, clinicians, professional societies, and standards bodies have complementary roles based on what each can observe and control, but those roles should not substitute for manufacturer obligations,” the society wrote.
RSNA believes the regulatory process should require timely and detailed communication of any changes to gen AI models and their system configurations. It also wants established channels for end users to report performance deviations and request reviews of these changes, based on their real-world experience. Manufacturers also should provide transparency to help support safe customization of AI models to local patient populations and clinical workflows.
Meanwhile, ACR also emphasized the need for continuous post-market performance monitoring to ensure the responsible use of AI in clinical practice. The college said it welcomes continued partnership between the FDA and the radiology community to ensure the agency is able to take on this challenge.
ACR offered a list of key priorities in governing gen AI. It’s urging the FDA to regulate the complete clinical function of an AI tool and any associated risk. ACR also wants the agency to allow these technologies to evolve, balance premarket flexibility with real-world performance monitoring, and “measure the failures that matter clinically.”
“Evaluation should focus on clinically consequential performance rather than generic hallucination rates, nondeterminism or textual similarity,” the college wrote.
ACR said it can help the FDA to define meaningful categories such as false assertions, omissions, incorrect comparisons or unsupported recommendations. It’s urging the agency to work closely with the imaging community in these processes. To do so, the FDA should convene a radiology-focused public workshop, digital health advisory committee meeting or solicit imaging-specific feedback prior to publishing draft guidance.
“As always, ACR welcomes continued dialogue with the FDA on these and other radiology-relevant topics,” the letter closed.
You can read the American College of Radiology’s full comments here and find the Radiological Society of North America’s here.
Marty Stempniak has covered healthcare since 2012, with his byline appearing in the American Hospital Association's member magazine, Modern Healthcare and McKnight's. Prior to that, he wrote about village government and local business for his hometown newspaper in Oak Park, Illinois. He won a Peter Lisagor and Gold EXCEL awards in 2017 for his coverage of the opioid epidemic.