FDA Grants Breakthrough Status to Two Generative AI Radiology Devices

What Happened

The U.S. Food and Drug Administration has granted breakthrough device designation to two generative AI products designed to interpret chest X-rays and draft radiology reports. The designations were awarded to devices developed by Cognita and Aidoc, marking a notable regulatory step for generative AI applied to diagnostic imaging.

Breakthrough device designation is a formal FDA program that expedites the development, assessment, and review of medical devices that provide more effective treatment or diagnosis of serious conditions. Receiving the designation does not constitute market approval, but it grants companies access to intensive FDA guidance and prioritised review.

Background

Aidoc is a clinical AI company whose software is already deployed across hundreds of hospital systems in the United States, primarily for flagging urgent findings in radiology workflows. The company has previously received FDA clearances for AI tools focused on detecting conditions such as pulmonary embolism, intracranial hemorrhage, and aortic dissection.

Cognita is a less widely known entity in the radiology AI space. Details on the company's broader product portfolio were not available in the wire reports at the time of publication.

The FDA's breakthrough designation program was established under the 21st Century Cures Act of 2016. It applies to devices intended to treat or diagnose life-threatening or irreversible conditions. Chest X-ray interpretation is one of the most common and time-sensitive tasks in hospital radiology departments, and errors or delays in reading chest X-rays can affect outcomes in conditions including pneumonia, lung cancer, and cardiac abnormalities.

Generative AI in this context refers to models capable of producing natural language text, in this case the narrative radiology report that accompanies an image interpretation, rather than simply flagging a finding or producing a probability score.

What the Designations Cover

Both devices use generative AI to interpret chest X-ray images and produce drafted radiology reports. The reports would then be reviewed by a licensed radiologist before clinical use. The exact architecture of each model, the datasets used for training, and the specific clinical indications covered by each designation were not detailed in available wire reports.

The FDA has been expanding its engagement with AI-enabled medical devices over the past several years. The agency has cleared or authorised several hundred AI and machine learning-based medical devices to date, the large majority in radiology. However, generative AI devices that produce free-text clinical reports represent a newer and more complex category for regulatory evaluation, given concerns around accuracy, hallucination, and clinician over-reliance.

Industry Context

The radiology AI sector has attracted substantial investment over the past decade, with companies including Aidoc, Nuance, Viz.ai, and Rad AI developing tools to address radiologist workload and diagnostic speed. The integration of large language model capabilities into imaging workflows has accelerated since 2023, with several vendors announcing or piloting report-drafting features.

Breakthrough designation for generative AI radiology tools is rare. The FDA has previously granted breakthrough status to AI devices for detecting specific pathologies, but designations explicitly tied to generative AI report drafting represent a more recent development in the agency's regulatory posture toward the technology.

Nuance, a Microsoft subsidiary, has deployed AI-assisted radiology reporting tools widely across U.S. health systems, though its products operate under different regulatory pathways.

What Comes Next

Both Cognita and Aidoc will now enter an expedited FDA review process under the breakthrough designation program, with the timeline for potential market authorisation dependent on the submission and review of premarket approval or 510(k) applications for each device.