Viz.ai, Hyperfine team up to add image-reading AI to bedside MRI scanners

In what seems to be a match made in high-tech heaven, Viz.ai and Hyperfine will bring together their respective technologies designed to speed up the process of detecting signs of stroke and other neurological conditions.

Viz.ai, as the name suggests, is developing deep learning artificial intelligence software that analyzes imaging data collected in emergency situations to help triage cases of aortic disease, pulmonary embolisms, stroke and more.

Hyperfine, meanwhile, is about two years into the commercial rollout of its Swoop portable MRI system, a compact, wheeled machine that’s designed to bring emergency brain scans directly to a patient’s bedside in the ER or ICU.

So, through their newly announced collaboration, the companies are aiming to significantly accelerate the entire imaging, analysis and triage process for patients suspected of serious neurological conditions.

“The partnership between Hyperfine and Viz.ai will help us integrate two breakthrough technologies that already enable us to better assess patients in a time-critical fashion and make timely decisions that will ultimately save more lives and improve outcomes,” said Shahid Nimjee, M.D., Ph.D., surgical director of the Ohio State Medical Center’s Comprehensive Stroke Center.

“By combining these two proven technologies, we hope to further expedite care, when every second matters and transportation presents additional risk,” Nimjee said.

Like its new partner, Hyperfine, too, has dabbled in AI. The BrainInsight software embedded in the Swoop system uses AI algorithms to automatically assess each of the brain scans collected by the portable MRI machine—measuring ventricular volume, brain extraction, brain alignment and midline shift to help assess brain injuries—while additional deep learning capabilities help the system churn out high-quality images similar to those from a full-size MRI scanner.

That deep learning software earned FDA clearance in November. The image reconstruction technology helps clear up T1, T2 and FLAIR scans, three of the most commonly used MRI sequences.

And just last month, the FDA cleared the deep learning software to run another pair of sequences: a T1 sequence that can capture images of the inside of the brain, rather than just the surface, and an accelerated T2 sequence that images the brain’s ventricles in half the time of standard T2 processes.

Viz.ai, meanwhile, is expanding its own AI capabilities. In April, the San Francisco-based startup collected $100 million in venture capital, which it said would be used to build out its brain scan-reading software to spot more conditions.

To date, Viz.ai has racked up regulatory clearances for its AI tools spanning large vessel occlusion, cerebral perfusion, intracranial hemorrhage, pulmonary embolism, aortic disease and cerebral aneurysm. Next up: AI-powered detection of subdural hematoma, which is already under FDA review.