Hyperfine has received FDA clearance for its deep learning, image reconstructive technology

Hyperfine gets FDA nod for deep learning, image reconstructive tech

November 30, 2021
by John R. Fischer, Senior Reporter
Hyperfine, known for creating the first portable MR system, Swoop, has now scored FDA clearance for its deep learning, advanced image reconstruction technology.

The company made the announcement this week at the 2021 RSNA annual meeting in Chicago, and says the solution will help Swoop produce crisp, clear T1, T2 and FLAIR images that are on par in quality with those produced by fixed 1.5T systems.

Access to the reconstructive technology will enable providers to make more accurate and faster diagnoses, which opens up the possibility of shorter hospitals stays, earlier diagnoses and an improved overall healthcare experience, according to Dr. Edmond Knopp, senior medical director at Hyperfine. “If you look at the old images, you can see the same things. But it’s a question of confidence and diagnostic ability. As a neuroradiologist, I'm now looking at an image that I’m more accustomed to, so I feel more comfortable making the diagnosis,” he told HCB News.

An alert on the new deep-learning technology will soon be sent to the system for users to download it if they wish. The reconstructive technology functions through a two-prong approach. The first utilizes a gridding approach, in which small grids made around the scan are broken down by neural networks to analyze and ensure errors do not show up in the image. The image is then sent to the imaging space, where it undergoes denoising. This brings it into the realm of high-field, which applies to T1, T2 and FLAIR images.

The company previously acquired FDA clearance for another advanced AI application back in January that enables providers to measure brain structure and pathology in images acquired by Swoop. It does so by providing tools for automatic measuring of ventricular volume, brain extraction, brain alignment and midline shift. With it, clinicians can diagnose and measure acute neurological conditions at a patient’s bedside.

Almost a year prior to this, Hyperfine got the nod from the FDA for its actual Swoop system. The solution is equipped with a .064 Tesla magnet — significantly lower than that of a standard scanner — requires no shielding, was designed with the nearly 90% of the world lacking access to MR in mind, and can be used to assess patients that are too fragile to move from the bedside, as well as ER patients who complain of headaches or dizziness.

Users can control it with a wireless tablet, and Hyperfine says the system is 20 times cheaper, 10 times lighter and consumes 35 times less power than fixed conventional MR solutions.

Back in September, Ohio State University Wexner Medical Center became the first in its home state to deploy Swoop for the diagnosis of brain injuries. Two are now operating in the provider’s emergency room and Comprehensive Stroke Center operating room. It also is used in Yale New Haven Hospital, UCI Medical Center and the University of Illinois and Chicago's Surgical Innovation Training Lab (SITL).

“There is no question this device can help save lives in resource-limited settings, such as rural hospitals or developing countries,” said Kevin Sheth, professor of neurology and neurosurgery at Yale School of Medicine and co-corresponding author of the research, back in September.

The company also announced in July that it would go public by merging with Liminal Sciences and HealthCor Catalio Acquisition Corp. into one company with a total enterprise value of approximately $580 million. All three plan to develop solutions together across the care continuum that will expand the existing $23 billion global imaging market, particularly for MR and brain sensing technologies.

Hyperfine is currently working on getting Health Canada approval, U.K. CA mark, European CE mark, and regulatory approval in Australia and New Zealand for its advanced image reconstructive technology.