The KOSMOS AI platform is aimed at making KOSMOS easy to learn, while greatly improving confidence curves in point-of-care ultrasound for providers of all experience levels
EchoNous is proud to announce FDA approval of its KOSMOS Platform, which consists of a distinctive 8oz. ultrasound-based tool combined with deep learning for clinical assessment of the heart, lungs and abdomen. It is the first tool in medicine to apply the mathematical framework of machine learning to mimic the human learning and decision-making process.
“It was 11 years ago that Dr. Rod Hochman, then-CEO of Swedish Hospital Seattle, urged me to one day create a point-of-care ultrasound tool that actually teaches users how to use it. Today, the FDA approval of KOSMOS marks the beginning of our delivery on that dream. The KOSMOS platform is enabled by the arrival of useful deep learning tools and the unique ultrasound miniaturization skills we possess at EchoNous,” said Kevin Goodwin, CEO and co-founder of EchoNous, Inc.
The KOSMOS AI platform is aimed at making KOSMOS easy to learn, while greatly improving confidence curves in point-of-care ultrasound for providers of all experience levels. The device is not only ideal for novice users but still serves super users who demand high-quality imaging to enable clinical decision-making immediately at the point of care.
“KOSMOS, in the hands of a clinician at the bedside, can move their diagnostic needle into the confident zone. I’m a POCUS advocate, daily user, and educator. I know that physicians, including myself, are often not quite confident enough after their POCUS exam to avoid the need to confirm what they are seeing with a formal test such as a CT or echocardiogram. KOSMOS’ image quality and AI-confirmation will give us greater accuracy and confidence at the bedside, reducing the need for additional testing and delays,” said Dr. David Tierney, Program Director of the Internal Medicine Residency and IMBUS Program, Abbott Northwestern Hospital.
The KOSMOS platform enables deep learning by using high-grade image quality, which is essential to train deep learning algorithms, because high-quality data ensures accurate results. KOSMOS also reveals its AI calculations in a transparent manner to clinicians while executing clinical tasks, such as AI-calculated ejection fraction.
“For AI techniques to be useful in POCUS, a data-rich image is required to avoid classic ‘garbage-in, garbage-out’ problems. For deep learning to be of actual value to medicine, we have, since the beginning, carefully listened to numerous leading clinicians as to how they want AI techniques delivered with ultrasound,” said Niko Pagoulatos, Ph.D., and co-founder of EchoNous.
FDA approval for KOSMOS has come at a vital moment as the world fights the COVID-19 pandemic. EchoNous thanks the FDA for its timely approval. “There is a palpable need for additional bedside diagnostic capability for identifying pneumonia and monitoring patient progress when septic or in septic shock with COVID-19. As such, in this great time of need by our country and the world, the approval couldn’t have been more timely for the healthcare community,” said Dr. Michael Blaivas, MD, MBA, FACEP, FAIUM, and Chief Medical Officer at EchoNous.
KOSMOS is particularly suited for the moment because its handheld probe and proprietary tablet can move quickly and easily through busy hospital wards. Its durable design makes it simple to disinfect. Its hospital-grade materials can withstand frequent disinfection with harsh cleaning chemicals, making KOSMOS not only an excellent imager of the torso but also a time-efficient tool in overwhelmed emergency rooms and ICUs that are desperate to shrink the infection footprint quickly.
Following several months of clinical evaluations, KOSMOS has been routinely described by physicians as an “exciting new kind of medical tool.” KOSMOS includes an 8oz ultrasound imaging tool with embedded and synchronized ECG and digital auscultation, and a 23oz tablet, the Kosmos Bridge. The Bridge has been built to withstand hospital-grade disinfection procedures and heavy use in hospital and clinical settings. All AI computation functions run in the Bridge itself, eliminating the need for cloud-connectivity during use and protecting patient and hospital data as well as ensuring there are no drops in service during scanning. The Bridge is also designed to seamlessly connect data to the health system IT environment without the need for internet connectivity.
Currently, the KOSMOS AI platform can calculate systolic heart function measurement (Left Ventricular Ejection Fraction), which has been clinically tested vs. expert measurements, yielding excellent results in precision and consistency.
In a parallel, we will soon file the world’s first-ever trio of algorithms that will enable users to have their image quality graded, their technique guided, and the heart anatomy labeled by machine learning algorithms.
The KOSMOS platform has been rigorously evaluated in acute care clinical settings, specifically on over 300 patients, while also being tested by hundreds of physicians and imaging experts across the United States, Canada, Europe and Japan in live clinical demonstrations.
EchoNous has validated that KOSMOS is the world’s first AI-assisted handheld tool to move the needle into increased diagnostic confidence at the bedside, with capabilities proven to be equivalent to larger cart-based machines – all for under US $10,000.
Near-term steps for KOSMOS include the release of continuous and pulsed wave doppler (Spring 2020). The KOSMOS platform also has a full suite of clinical deep learning applications planned for its 12-24-month roadmap.
Source: Company Press Release