Imaging and diagnostic equipment based on advanced technologies like artificial intelligence (AI), machine learning (ML) and deep learning accelerate innovation in healthcare. While there are many cutting-edge, large ultrasound machines, some of which have AI capabilities, housed in healthcare settings, their usage was limited during the COVID-19 pandemic due to the need for high infection security. Hand-carried ultrasound devices saw an uptick in demand to help with virus detection. Additionally, along with devices and 3-signal probes that can hear acoustic sounds and measure electrical impulses from the heart, other AI-driven handheld tools also demonstrate the recent growing number of AI use cases in healthcare diagnostics. EchoNous, which stands for ‘intelligent ultrasound’ in Greek, honors the roots of one of the company’s founders and set out to create tools by leveraging the benefits of deep learning, AI and ML to deliver high-performance, cutting-edge hand-held ultrasound machines.
Kevin Goodwin, the CEO of EchoNous is a veteran of ultrasound imaging and medicine since 1987. After having founded his first-ever POCUS company in the late 90s, when the global ultrasound market was valued at nearly 3 billion USD, Goodwin highlights the sizable proportion that the POC segment has on today’s ultrasound market. “Today, global ultrasound is 8 billion and POCUS is 35 percent or nearly 3 billion. So POCUS is that natural diffusion based on miniaturization and simplification of technology out into the world of clinical medicine where clinicians use it to do their work,” Goodwin says. “What’s happening is not only clinical physicians but also care providers such as physician assistants, nurse practitioners, even RN’s and patients will be taking advantage of ultrasound because of the contribution of AI.”
Goodwin further adds that AI arriving into ultrasound is the field of mathematics on top of the world of physics which has dominated ultrasound for years! With the arrival of mathematical algorithms, improvements can be made to ultrasound devices. “We can improve cycle times for learning, becoming competent and reduce variance,” he adds.
Uscan, one of EchoNous’ intelligent devices for scanning the bladder and measuring its volume, is used by more than 25 percent of healthcare systems. The company’s AI-driven Bladder tool is accessible to 97 percent of the U.S. healthcare market alongside another peripheral IV finding tool. The Bladder tool is based on convolutional neural networks to improve the measurement of residual bladder volume in the bladder, basically fluid or urine. This enables the nursing community to make better decisions about whether to catheter a patient or not. “This is the beginning of us proving the use of AI on a small narrow application such as bladder volume residuals,” Expresses Goodwin. Cleveland Clinic was among the very first that EchoNous signed on.
Kosmos, a handheld AI-driven tool of EchoNous is benchmarked against the cart-based system. To understand the tool’s applicability in the care practice and protocols for clinicians, EchoNous conducted a benchmark study involving a diverse group of 1,000 patients in Greece. “We learned that the imaging, Doppler measurements, and AI algorithm were all equivalent with a large $120K euro system on a broad set of patients. We did not select patients, which is vital to understand because patients can be difficult to image. That’s the real challenge of ultrasound which they call technically difficult patients.” Kosmos was created as a new category of ultrasound, packaged on a silicon chip. This makes Kosmos a piece of hardware with the quality and functionality of a large-middle market console-based system. “Conventional hand carry, or highly mobile ultrasound devices such as the tiny ones out there, have limited image quality, limited functionality and therefore limited to no use of AI. Our product, with its 64/128 channel engine, is the opposite. It creates a new category of being an adjunct to the large cart world and not the compromise found in the small hand-carried world.” Kosmos’ functionality includes Continuous Wave Doppler and also three signals that can come through one ultrasound probe, including ECG and auscultation, along with imaging and Doppler. Together, these would create a stack of four signals: the image, the Doppler measurements, the sounds and the electrical patterns. “In the future, this will be a great platform for applying AI and we’ve been told this by the cardiology community. So as far as the benchmarking is concerned, we did that to understand how we stacked up and we’re very pleased with the outcome,” Goodwin elaborates.
It all begins with a very high-resolution ultrasound device. This is hard to accomplish in a small package and EchoNous’ Kosmos Platform is a result of its game-changing AI and innovative hardware,” Goodwin says.
Design philosophy is central to the usage of imaging devices with zero friction for EchoNous. It means, the tools limit the interactions or delays for users to get to the endpoint. AI has been a predominant factor in moving the user interaction to zero, reducing variance, enabling better care protocols and accuracy. “It all begins with a very high-resolution ultrasound device. This is hard to accomplish in a small package and EchoNous’ Kosmos Platform is a result of its game-changing AI and innovative hardware,” Goodwin says. From a use case standpoint, EchoNous’ AI-based devices are designed to help novice users learn faster. For example, the toolkit from deep learning called Object Detection, where the ultrasound device could find the heart and then label all the elements of the heart, the chambers and the valves etc or the anatomical labeling algorithm set is a unique capability. The company has been expanding it into the abdomen for abdominal labeling for the FAST exam. In both cases, fast exams are designed to find fluid in the abdomen or thorax and ascites is a clinical problem that results in fluid from the kidneys. And so importantly, labeling the body changes if done sonographically with deep learning changes the learning curve for novices. “It gets them up on their feet and we’ve heard this from critical care residents that have to get into the firing line right away with critical care patients and they’re able to use our device more rapidly ’cause they understand what they’re looking at,” Goodwin recalls. “One of the core problems of ultrasound is, ‘how do I get a good image?’ and that’s probably 75 percent of the challenge of using ultrasound. The remainder is ‘Ok, I’ve got a pretty good image, but what am I looking at?’.” AI enables a user to understand the anatomy, sonographically deployed, displayed rather, because of the labeling technology.
In 2022, EchoNous is gearing for its rapid expansion in partnerships across medicine. “The field is just beginning. I’m pleased to say that with our platform and our position in AI, we’re ahead of the pack and doing things that are based on clinical feedback and not just trivial,” Goodwin says. “Problems like heart failure and valve disease are out there in society. With our Kosmos Platform and these AI connections, you can actually go into the community and engage in successful population health management or what we were more previously would call effective accurate health screening,” Goodwin concludes.