This browser is not actively supported anymore. For the best passle experience, we strongly recommend you upgrade your browser.
Welcome to Reed Smith's viewpoints — timely commentary from our lawyers on topics relevant to your business and wider industry. Browse to see the latest news and subscribe to receive updates on topics that matter to you, directly to your mailbox.
| 1 minute read

FDA receives comments on making information about AI/ML-enabled medical devices more transparent

In response to the U.S. Food and Drug Administration's (FDA's) virtual public workshop on transparency surrounding artificial intelligence/machine learning (AI/ML)-enabled medical devices, the American College of Radiology (ACR) and other stakeholders have provided comments to the FDA. 

As we noted earlier, FDA has begun to gather and publish a list of AI/ML-enabled medical devices that have been authorized by the agency to be marketed. Hundreds of those authorized devices are in the radiology space. The FDA's interest in promoting the dissemination of useful information about these devices was the impetus for the agency's virtual workshop on transparency.

ACR's comments focused on radiology device considerations of interest to providers, with the ACR defining providers as decision-makers responsible for technology acquisition as well as radiologist end-users of the software devices. The ACR's input addressed the type of device data that the FDA could make available that would be helpful to radiology providers. Information about the AI/ML-enabled devices, ACR wrote, should include the following: patient demographics; identification of the imaging acquisition device (manufacturer, model, version, contrast); both performance and findings metrics; "enhanced" product identifications (use case definition and CAD classification instead of, or in addition to, product codes); and intended user qualifications such as "qualified radiologist."  ACR offered ideas for usable presentations of device information to stakeholders and provided a recommendation for an alternative reporting mechanism to openly share software performance issues that do not result in patient harm. 

The ACR believes that the real-world performance of these technologies may differ from that of the training and testing environments. Thus, having access to meaningful information could help radiology providers discover appropriate innovations for their clinical needs and estimate software performance when paired with their respective patient populations and subpopulations, image acquisition/input devices, and imaging protocols. Currently, access to such information is relatively limited and unintuitive. ACR recommends that FDA make available summaries with links to additional details. Of particular interest was ACR's request that FDA's information about AI/ML-enabled devices be publicly available in concise, understandable and "readily discoverable" formats by interested providers who may not be familiar with FDA's regulatory processes.

The Medical Imaging & Technology Alliance (MITA) also provided comments, writing in support of transparent device labeling to provide clinicians with the information they need to confidently use a device in accordance with its intended use.

You can review the ACR, MITA and other stakeholder comments here.


health care & life sciences, fda, artificial intelligence, machine learning, transparency