FDA Designates Two Generative AI Radiology Tools as Breakthrough Devices

What Happened

The U.S. Food and Drug Administration has granted breakthrough device designation to two artificial intelligence-powered devices built to assist in drafting radiology reports, according to a report published by STAT. The designations, part of the FDA's formal breakthrough device program, are intended to accelerate the review and development of medical technologies that may provide more effective treatment or diagnosis of serious conditions.

What the Breakthrough Device Designation Means

The FDA's breakthrough device program allows manufacturers of certain medical devices to receive more intensive guidance from the agency during development and priority review of their submissions. Designation does not constitute approval or clearance of a device for clinical use. It is a procedural status that expedites the path toward potential authorization. The two generative AI devices in question are focused specifically on drafting radiology reports, a task that currently requires significant time from trained radiologists.

Generative AI systems in medical imaging contexts typically analyze imaging data, such as X-rays, CT scans, or MRIs, and produce structured written summaries or preliminary reports that a clinician then reviews and approves. The FDA has been developing its regulatory framework for AI-enabled medical devices in recent years as the number of such products submitted for review has grown substantially.

Background

The FDA has been tracking AI and machine learning-enabled medical devices for several years. As of recent counts, the agency has authorized hundreds of AI-enabled devices, with medical imaging representing the largest single category. Radiology has been among the most active areas for AI development in health care, given the volume of imaging studies performed annually and the documented shortage of radiologists in parts of the United States.

Generative AI, which can produce original text, has emerged as a distinct category within health care AI tools. Unlike earlier AI systems that flagged anomalies or prioritized worklists, generative AI systems can produce draft narrative reports, which places them in a different regulatory classification given their more direct role in clinical documentation.

The STAT report also noted a separate development: the FDA dropped an enforcement action against Whoop, a wearable health technology company, though details of that matter were reported independently of the radiology AI designations.

The Regulatory Pipeline

The volume of generative AI devices moving through the FDA's breakthrough pipeline reflects a broader increase in submissions from companies developing clinical AI tools. The breakthrough designation for the two radiology report drafting devices signals that the FDA has determined these tools have the potential to address an unmet medical need and that existing therapies or technologies may be inadequate.

Manufacturers that receive breakthrough designation work directly with FDA staff through a more collaborative and iterative review process. This can shorten the total time from submission to a regulatory decision, though the agency does not guarantee approval at the end of that process.

Radiology report drafting tools, if eventually authorized, would be used within clinical workflows where a licensed radiologist retains responsibility for reviewing, editing, and signing any report before it is used to guide patient care. The devices are designed to function as assistive tools rather than autonomous decision-makers.

What Comes Next

The two AI radiology device manufacturers will now engage in the FDA's expedited review process, with the timeline for final regulatory decisions subject to the companies' submission schedules and the agency's review procedures.