The Transform Technology Summits start October 13th with Low-Code/No Code: Enabling Enterprise Agility. Register now!
Huiying Medical, a medical devices company based in Huizhou, China, claims to have developed an AI imaging solution that uses CT chest scans to detect the presence of COVID-19. The company asserts that it might be useful in regions of the world without access to reverse transcription polymerase chain reaction (RT-PCR), the standard testing method for COVID-19.
According to Intel, which highlighted Huiying’s work in a blog post today, Huiying — a member of Intel’s AI Builders program, a market enablement program for OEMs, system integrators, and software vendors — developed the solution’s underpinning algorithms based on CT data from over 4,000 coronavirus cases. It analyzes what’s known as ground-glass opacity (GGO) in the lungs, which indicates a partial filling of air spaces, as well as other indicators that inform a probability of suspected COVID-19 infection.
Running on Intel processors optimized for Google’s TensorFlow framework and deployed with Intel’s OpenVINO framework for machine learning development, Huiying’s solution takes just a day to install and only 2-3 seconds to process CT studies with 500 images. Moreover, it has a claimed 96% novel coronavirus pneumonia (NCP) classification rate, and it’s designed to work either in the cloud or on-premises.
Huiying says that it has partnered with Huawei to reach out to health professionals and institutions to share its solution, which it claims is being used in over 20 hospitals. In the Philippines, Huiying plans to install the system in Baguio General Hospital, and it recently brought it to a hospital in Ecuador, where an additional component is tracking patient progress to determine treatment efficacy. (This made Ecuador the first Latin American country with an AI-enabled diagnostic system in two local hospitals, according to Huawei.)
Beyond Huiying, companies including Alibaba, RadLogics, Lunit, DarwinAI, Infervision, and Qure.ai claim they’ve created systems capable of isolating COVID-19 in X-ray or CT scans highly accurately. Early work from Chinese medical researchers and a system published in the journal Radiology demonstrated similar results.
But currently, the U.S. Centers for Disease Control and Prevention recommends against the use of CT scans or X-rays for COVID-19 diagnosis, as does the American College of Radiology (ACR) and radiological organizations in Canada, New Zealand, and Australia. That’s because even the best AI systems sometimes can’t tell the difference between COVID-19 and common lung infections like bacterial or viral pneumonia.
For this reason, Intel cautions that CT scans results aren’t by themselves “conclusive” in determining coronavirus infection. Rather, explains Intel data platforms marketing group business development manager Xu Cheng, they might play a role in the diagnostic process by indicating that further testing is required.
“[This] complement[s] standard lab testing,” added Cheng. “Simply put, it’s challenging for health care professionals and government officials to allocate resources and stop the spread of the virus without knowing who is infected, where they are located, and how they are affected.”
To this end, the ACR itself released an open project call for AI that can detect COVID-19 from CT scans. It hopes this will foster the development of tools in health care settings where there aren’t many human radiologists available; in the U.S. alone, the projected shortage of radiologists by 2025 is in the tens of thousands.
VentureBeatVentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact. Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:
- up-to-date information on the subjects of interest to you
- our newsletters
- gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
- networking features, and more