Generative Vision Language Model for Chest X-Ray Gets FDA’s Breakthrough Device Designation

The Food and Drug Administration (FDA) has granted a breakthrough device designation for Cognita Chest X-Ray (CXR), a generative AI vision language model that may bolster detection and efficiency for CXR interpretation.

Developed by Cognita, an AI business unit of Mosaic Clinical Technologies, Cognita CXR’s vision language model reportedly provides comprehensive initial assessment of CXRs with a subsequent radiology report that can be reviewed and finalized by radiologists.

Noting that Cognita CXR is the first generative AI model to garner the FDA’s breakthrough device designation, Mosaic Clinical Technologies cites preliminary internal clinical data that revealed an 18 percent increase in interpretation efficiency. The company said the data also demonstrated improved detection for various CXR findings ranging from 16 to 65 percent with the adjunctive use of Cognita CXR.

“We believe this is an important step toward bringing AI into everyday radiology practice in a way that enables physicians to practice at the top of their license and expands access to high-quality imaging care worldwide,” noted Nina Kottler, M.D., the chief medical AI officer at Mosaic Clinical Technologies.