About 29% of AI-enabled radiology devices approved by the FDA between 1995 and 2024 had undergone clinical testing, according to a study published Nov. 7 in JAMA Network.
Researchers analyzed premarket submission materials of 950 AI and machine learning-enabled devices from November 1995 to June 2024 to assess the levels of prospective, “human-in-the-loop” and clinical testing undergone prior to FDA authorization.
Here are five things to know from the study:
- Of the 950 devices, 723, or 76%, were designed for use in radiology.
- Also, of the 950 devices, 924 were cleared through the FDA’s 510(k) pathway.
- Of the 723 radiology devices, 717 had submission documentation available to review. According to the documentation, 5% of the devices underwent prospective testing, 8% included a human in the loop and 29% underwent clinical testing.
- Only six of the 717 devices employed all three testing mechanisms and 15 employed both prospective and clinical testing.
- “Institutional governance is needed to determine whether a device has undergone adequate testing and confirm that it functions as intended,” the study authors wrote. “Clinical studies and frameworks such as Health Technology Assessment adopted in Europe could also help address this evidence gap by evaluating clinical and economic value postapproval.”
Read the full study here.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
