In February 2018 the US Food and Drug Administration granted Viz.ai a De Novo clearance for its Viz LVO software, the first clearance for an AI-based computer-aided triage and notification system. Where the contemporaneous IDx-DR clearance covered autonomous diagnosis, this one addressed a different role: speeding the right patient to the right specialist.
The software uses deep learning to analyze CT angiography scans for large vessel occlusion, the kind of blockage that causes severe strokes. When it detects a suspected case, it automatically alerts an on-call stroke specialist and pushes the images to their smartphone, rather than waiting for the scan to work its way through the normal radiology queue. In stroke care, where brain tissue dies by the minute, shaving time off that handoff can change outcomes. The company reported a study with roughly 90 percent sensitivity and specificity and median notification within minutes.
Because it was a De Novo authorization, the decision also created a new regulatory category, computer-aided triage (CADt), giving later triage products a defined pathway through the FDA’s clearance process.
For a business or general reader, this milestone marks the arrival of a distinct flavor of clinical AI: not a system that diagnoses on its own, but workflow software that uses pattern recognition to reorder the queue and get urgent cases in front of humans faster.