The American College of Radiology and the Radiological Society of North America told the FDA that generative AI-enabled medical devices need strong postmarket monitoring alongside any regulatory flexibility, according to comments both societies submitted Oct. 19 on the agency’s generative AI discussion paper.
RSNA supported a risk-based, total product lifecycle approach but warned that these devices’ variable outputs and ability to change after deployment create risks that cannot be addressed through “one-time premarket review alone.” It urged the FDA to widen its proposed risk framework to account for how easily a user can review an output, how reversible and time-sensitive resulting actions are and whether downstream safeguards exist. It said an output used in a stroke, trauma or oncology workflow may carry more risk because there is little chance to catch an error.
RSNA wants monitoring to evaluate the whole clinical AI system, including human-AI interaction, workflow effects and use beyond validated indications, not just the model. It said manufacturers must stay accountable for safety and effectiveness, while healthcare organizations, clinicians and professional societies have complementary roles that should not “substitute” for that responsibility.
The ACR said the safety of generative AI devices depends on the clinicians, workflows, data sources, governance and monitoring around them, not just the technology, and noted that image-informed draft reporting is among the most active areas of development in radiology. It urged the FDA to regulate a tool’s complete clinical function and risk, since the same foundation model could support low-risk documentation or high-risk autonomous diagnosis. The college asked the FDA to credit qualified end-user oversight only when sponsors show it works in realistic workflows, to let devices evolve within validated performance limits and to pair that flexibility with monitoring for clinically meaningful failures, performance drift and subgroup effects.
The ACR also urged the agency to “measure the failures that matter clinically,” such as false assertions, omissions and laterality errors, instead of generic hallucination rates. It asked the FDA to clarify how any new framework applies to already-authorized devices, such as those that reconstruct or denoise images, to include imaging examples in guidance and to hold a radiology-focused workshop or advisory committee meeting or seek imaging-specific comments before drafting guidance.
The comments echo those of the AHA, which told the FDA its framework should place primary monitoring responsibility on developers and has asked the agency to clarify which AI tools fall outside device oversight.
Becker’s 3rd Annual Fall Chief Pharmacy Officer Summit is the top tier event that the people leading the future of pharmacy convene. This 3-day event is built specifically for Chief Pharmacy Officers and senior pharmacy leaders. The sessions examine the workforce, financial and clinical priorities shaping pharmacy today. Becker’s is where candid discussions flourish and new ideas thrive. Becker's 3rd Annual Fall Chief Pharmacy Officer Summit
