FDA OKs Marketing of AI-enabled Stroke Alert for Providers

Dive Brief:

• The Food and Drug Administration permitted the marketing of Viz.AI’s clinical decision support software, which analyzes computed tomography brain scans and alerts clinicians if a patient is at risk of a stroke.
• The Viz.AI Contact application was cleared for marketing via the De Novo premarket review process, which is reserved for some low- to moderate-risk medical devices that have no legally marketed predicate device on which to base a determination of substantial equivalence.
• The decision “also creates a new regulatory classification, which means that subsequent computer-aided triage software devices with the same medical imaging intended use may go through FDA’s premarket notification 510(k) process, whereby devices can obtain marketing authorization by demonstrating substantial equivalence to a predicate device,” the company stated.

Dive Insight:

To support the review, Viz.AI submitted a retrospective review of CT images that evaluated the performance of the product’s algorithm and notification app against two trained neuro-radiologists in detecting large vessel blockages in the brain. Real world evidence was used to show that the app could alert a neurovascular specialist sooner than a clinician in cases where a blockage was seen.

Tools that leverage machine learning to support diagnosis of disease and detection of abnormalities have generated a lot of investor enthusiasm recently and will continue to do so.

With Viz.AI, images of the brain are analyzed and a neurovascular specialist is notified via text if a large vessel blockage is identified. A specialist still needs to review the images on a clinical workstation, FDA stated.

In October, FDA issued draft guidance clarifying when it will regulate certain types of clinical and patient decision support software as medical devices. The draft exempts certain clinical decision support tools, such as when a physician can independently review the underlying basis for the recommendation being made by the app.

Apps that analyze medical imaging, in vitro diagnostic signals or patterns from an electrocardiogram or other processor would be regulated due to the potential for patient harm, according to the draft.

The proposal has been skewered by industry. In public comments, the Clinical Decision Support Coalition charged the document over-regulates“to the detriment of patient care.” The group — which includes software developers, IT infrastructure manufacturers, providers and medical device and drug makers — urged the agency to revise and re-issue the draft with a more risk-based approach.