How a U of A spinoff is using artificial intelligence to make medical sci-fi a reality
In the original Star Trek TV series, Dr. McCoy used a tiny device called a tricorder to instantly diagnose his patient’s condition. While it seems improbable, a real-life version of the tricorder may be widely available as soon as 2022, thanks to University of Alberta research.
Medo.ai, a University of Alberta spinoff company, uses artificial intelligence (AI), smartphones and ultrasound technology to diagnose common and critical conditions. The company’s unique technology analyzes ultrasound images and diagnoses disorders using machine learning accumulated from thousands of similar cases.
Imaging technology makes it possible to see inside the human body with astonishing precision. But MRIs and CT scans are very expensive, often subject to long wait times and accessible only in larger urban centres. Traditional ultrasound, while much more accessible, produces images that are blurry, two-dimensional and extremely difficult to interpret. While ultrasound imaging can happen almost anywhere, a diagnosis still has to wait on the availability of very highly skilled clinicians.
Two-thirds of the world does not have access to medical imaging. In developing countries, access is non-existent.
“Even in [countries] such as Canada, it is very limited,” says Dornoosh Zonoobi, Medo.ai’s CEO, who adds she’s seen patients drive for hours in the middle of winter to get an ultrasound scan. “We have patients flying in helicopters from the northern territories.”
Driven by the urgent need for better options, she and radiologists Jacob Jaremko and Jeevesh Kapur decided to create software to augment ultrasound technology. Leading a diverse team, the three launched Medo.ai to commercialize the technology they developed.
The company’s technology makes it possible to put diagnostic ultrasound in the hands of caregivers in assisted living facilities and minimally trained care providers in the most remote communities. Rather than transporting trauma victims to an MRI scanner, emergency room doctors or even ambulance attendants could diagnose injuries using a hand-held scanner. Your family doctor could diagnose liver and kidney disease in her office.
Medo.ai’s first platform was created to diagnose hip dysplasia – a common hip joint abnormality that causes osteoarthritis if it isn’t caught early.
Ultrasound is very portable. It is cheap and safe; some units fit in your pocket. The problem our technology addresses is that the expertise needed to use these devices is rare.
“It’s one of the leading causes of arthritis of the hip, and the leading cause in young women,” explains Dr. Jaremko, Medo.ai’s chief technology officer. “It can be fixed very easily in infancy with a soft harness worn for four to six weeks. If we screened every newborn, we would prevent so much pain and disability for these people in later life. Until now, however, it simply wasn’t possible or practical to do so. The mother of one of my daughter’s friends is in her early 40s; she has already had bilateral hip replacements because of arthritis caused by hip dysplasia.
“It is very hard for her to get through her days with young children. And it could have been prevented.”
Medo.ai’s technology allows Dr. Jaremko to teach almost anyone how to do a scan in about an hour. In fact, he and Zonoobi gave medical students one hour of training and found they were able to produce 3D images as reliable as those produced by senior technicians.
In Alberta, four babies are born with hip dysplasia each day. Two are undiagnosed, leading to painful osteoarthritis and premature hip replacement surgeries later in life, adds Dr. Zonoobi. This technology can change that, she says.
“Ultrasound is very portable. It is cheap and safe; some units fit in your pocket. The problem our technology addresses is that the expertise needed to use these devices is rare.”
Four leading global children’s hospitals are now using the platform in research projects. Once it’s approved, it is expected to prevent up to 17 premature hip replacement surgeries each day in Canada and 151 in the United States, representing savings of $1-billion a year in surgical costs in the U.S. alone. A recent multimillion-dollar seed capital round was oversubscribed, and the platform is now moving toward FDA clearance.
Medo.ai – which employs 24 people in Edmonton, mostly computer programmers, many of them U of A grads – is one of more than 130 U of A spinoffs in operation today. The university ranks third in the country for spinoffs created between 2013 and 2017, with 108 patents received for its various discoveries.
Drs. Jaremko and Zonoobi say Edmonton is well placed for this type of startup, due to the high quality and integration of its health-care system and the U of A’s AI expertise.
The quality-of-life implications are enormous. And with Medo.ai now turning its focus to platforms that address more situations in which cheap, portable and easy-to-use imaging technology could save lives, it’s clear – the future of global health care is about to get much brighter.
Produced by Randall Anthony Communications. The Globe’s editorial department was not involved in its creation.