Israeli Startup Viz.ai Uses AI, Deep Learning To Detect Source Of Stroke In Seconds

When it comes to dealing with a patient that has suffered a stroke, speed is often more important than healthcare. The idea of being able to detect a blocked artery, clot, or rupture that bars blood flow to the brain in minutes, rather than hours, could mean the difference between patient disability and walking out of the hospital. 

A stroke detection platform developed by Tel Aviv-based medical imaging startup Viz.ai uses artificial intelligence to identify the source of a stroke in just seconds. This can give a stroke patient the chance to receive treatment sooner and save them from long term brain damage, disability, and paralysis. Founded in 2016 by a global team of experts including Dr. David Golan, an Israeli statistics and AI expert, Viz.ai leverages AI and deep learning tech to analyze CT scans. The technology automatically detects early signs of large vessel occlusion (LVO) strokes and alerts physicians. 

 The startup has offices in San Francisco and has raised over $80 million to date, including a $50 million Series B round in October 2019.  

A stroke occurs when blood supply to the brain is suddenly interrupted. When an ischemic stroke occurs, a patient’s arteries leading to the brain are blocked abruptly, whereas a hemorrhagic stroke occurs when a blood vessel bursts, leading to bleeding into the brain tissue. Viz.ai’s technology “provides easy to use synchronized and secure technologies to healthcare providers and it helps to provide patients access to the right doctor at the right time,” says  Andrew Colbert, senior marketing manager at Viz.ai. The company’s commercial products are centered around ischemic and hemorrhagic strokes, he explains to NoCamels. “

Using data to pinpoint certain arteries and what they should look like, Viz’s technology allows stroke teams to consult via a HIPAA-compliant mobile interface to effectively locate blockages. The computer-aided triage software helps patients connect directly to a stroke specialist to fast-track life-saving care. 

“We offer an app on the phone that connects to a CT scanner,” Colbert tells NoCamels. “When a patient comes into a hospital and gets a CT scan for a suspected stroke, those images go up into our cloud. If the algorithm detects that there’s a stroke it alerts the entire care team. So that’s the synchronized part, it substitutes a serial process where essentially a very long game of telephone of 10 different calls used to happen previously into everyone getting alerted at the exact same time.”

The platform automatically detects suspected LVOs, specifically in the carotid arteries, blood vessels that carry oxygen-rich blood to the head, brain, and face, and in the middle cerebral artery which supplies blood to the cerebrum in the brain. Patients with blockages in the carotid and middle cerebral arteries account for about 20 percent of stroke volume and often need surgical intervention. Colbert notes that Viz.ai’s LVO product can effectively detect and locate these blockages for quicker treatment.

As of November 2019, Viz.ai’s platform was available in over 300 US hospitals.

Viz.ai LVOs for health systems 

Using Viz.ai’s technology, health systems can capture potential missed revenue opportunities by detecting and treating more large vessel occlusions; a case study at six hospitals revealed an average annual volume of 31 untreated LVOs, with the average MS-DRG reimbursement totaling $28,000. This means that if these 31 LVOs can be treated, the hospital can make an additional $5.2 million for potential new revenue opportunities. 

“[Viz LVO] really is helpful and beneficial because patients do better when you detect [LVOs] faster and we also detect more of them, all of which is better primarily for the patient, but also for the hospital from a financial standpoint,” Colbert says.

Viz.ai hopes to get treatment rates for mechanical thrombectomies, a minimally-invasive procedure in which an interventional radiologist removes a clot from a patient’s artery to 15 to 20 percent, compared to many hospitals today operating between three to four percent.

Ischemic stroke portfolio

In addition to Viz LVO, Viz CTP is a more commoditized software that uses advanced imaging technology to automatically analyze computed tomography (CT) perfusion images of the brain, produce parametric color maps, and calculate CT perfusion parameters. CT perfusion imaging reveals which areas of the brain are adequately supplied or perfused with blood and noninvasively gives detailed information on blood flow to the brain. Colbert notes that Viz CTP does not rely as heavily on artificial intelligence as LVO yet still uses AI to help minimize motion artifacts. 

“The CTP will tell you how much of the brain is likely dead now and how much of the brain is at risk,” Colbert said. “And if there is, for example, no difference between how much of the brain is currently dead, or what looks to be dead and at risk, then obviously the patient won’t benefit from an intervention. If there’s a lot of brain to save that you can see on the CTP, then that’s when the patient will benefit.”

In order to coordinate patient care and treatment decisions, medical professionals from hospitals in a hub and spoke network can use Viz HUB, a secure, HIPAA-compliant text messaging and calling platform. Viz HUB rounds out Viz.ai’s ischemic stroke portfolio along with Viz VIEW, a mobile, non-diagnostic image viewer that enables the dynamic viewing of Digital Imaging and Communications in Medicine (DICOM) images on Android or Apple phones.

For better-synchronized care, Viz.ai has also developed Viz ASPECTS for simplified aspects scoring, a quantitative topographic CT scan score used in patients with middle cerebral artery stroke to determine the degree of severity.

Viz ICH for intracerebral hemorrhage strokes

In addition to ischemic strokes, Viz also offers Viz ICH for intracerebral hemorrhage (ICH) strokes, using artificial intelligence to automatically identify suspected ICH strokes on non-contrast CT imaging. Viz ICH automatically alerts on-call stroke teams in less than 30 seconds, saving an average of 37 minutes from picture to alert relative to standard care. 

Viz’s hemorrhage stroke portfolio consists of Viz ICH and the aforementioned Viz HUB and Viz VIEW for automated ICH workflow and triage. Clinical benefits of Viz ICH range from significant increases in cost savings to decreased length of hospital stay and ventilator use to heightened neurological case volume. In 2018, Viz ICH automatically detected intracerebral hemorrhage with a sensitivity of 93 percent and a specificity of 95 percent in 191 NCCT images.

SEE ALSO: Medical Imaging Startup Aidoc Gets FDA Clearance For AI Solution To Spot Stroke

“We’re saving between 30 minutes and an hour at every hospital, that’s typically what we expect to be able to cut off, which is a massive time for the brain,” Colbert said. “Neurons die in the millions per minute, and each minute translates into days of improvement of disability-free life for the patient.”

According to Colbert, Viz’s technology is used in many major comprehensive stroke centers and has participated in conferences run by the Society of NeuroInterventional Surgery and the Society of Vascular and Interventional Neurology. Currently, Viz’s technology is limited solely to the United States, but Viz hopes to expand this technology to other telehealth-related platforms in the coming years. 

“We want to help and be a resource as much as we can to hospitals, especially through this COVID-19 time, and we’re looking at ways to do that, whether it’s using leveraging our communication platform or image viewing for enhanced telehealth across other disease states essentially outside the brain through more kind of traditional telemedicine but implementing some of the algorithms that we have,” Colbert said.

Helping patients and hospitals during COVID-19

Viz recently unveiled its Viz COVID-19 telehealth platform to hospitals at no charge and with no limitations on the number of providers and scanners that can be connected to the platform. After performing a chest X-ray on a patient, hospitals and providers can send real-time COVID-19 alerts, view radiographic images on a mobile phone, and engage in secure team communication. This system allows for remote consultations, effective resource allocation management, and ultimately a method of flattening the curve. 

In May, the company announced that it launched the COVID-19 patient triage software at Boca Regional Regional Hospital in Florida “ to improve patient management and allow for a safer hospital workplace during the pandemic,” according to a statement. 

“Our goal is to put patients first and help leaders in the medical community like Boca Raton Regional Hospital do the same. If our software can help in some way during this time of national emergency, we want to ensure its availability and utility for all health care professionals,” said Chris Mansi, CEO of Viz.ai.

According to Colbert, offering telehealth-type platforms that allow providers to remotely triage patients and do it quickly in a synchronized fashion will allow patients to receive the care that they need during COVID-19 times, where there are delays aggravated by health risks and slower communication. 

“We will select chest X-rays, and we will share those with the care team for suspected COVID patients, and the care team can then communicate directly about the needs for that patient, and all of this is remote because it’s texting on the app using HIPAA compliant chat,” Colbert says. The tech also allows “for resource planning, resource management, bed management, [and] ventilator use, so we can start giving providers the ability to communicate more effectively remotely,” he adds.

Subscribe to NoCamels weekly newsletter and get our top stories