The Sooner the Better

How AI-assisted scanning can solve many (big) problems


Adam McArthur doesn’t recall much about the first three years of his life, other than the facts – he needed 15 surgical operations to correct his hip dysplasia; he didn’t begin walking until he was three years old, and his mom left work so she could take him to the hospital four hours away in Edinburgh (he grew up in Scotland) for his surgeries, follow-up procedures and “a lot” of physiotherapy.

“Even though I don’t remember, I think it’s something that’s ingrained in my parents’ memory forever,” says McArthur, 22, a PhD student at the University of Alberta.

He’s still not out of the woods. “The problem now is how long my hips will last. Best case scenario is 40 years from now, when I’m in my 60s. Worst case, 20 years.”

McArthur is a key player in the research of Dr. Jacob Jaremko, a radiologist at the University of Alberta Hospital, Canada CIFAR AI Chair at the Alberta Machine Intelligence Institute (AMII) and a practitioner of teleradiology. Teleradiology refers to the distribution of medical imaging for diagnoses. This ensures patients receive timely diagnoses, whether they live within major cities or rural communities.

“If you’re on a First Nations reserve or you live in the far northeast corner of Alberta, you can get an ultrasound scan with one of these portable ultrasounds, but what’s missing is the expertise. What’s really needed is a way for people who are less expert, but still know a little bit of anatomy and are able to scan, to have some guidance and support to interpret the images, and to acquire good images. And so, that’s where AI comes in,” says Jaremko.

“When it comes to improving care in Indigenous communities, we’re interested,” says Dr. Jodi Abbott, President and CEO of the University Hospital Foundation. “Dr. Jaremko’s work is a perfect example of what transformational patient care looks like.”

As a way to test drive the concept of AI-assisted ultrasounds, Jaremko and his team have embarked upon a multi-year pilot project to scan infants for hip dysplasia.

“We’ve done over 1,300 scans all over Alberta, and the images are assessed by AI, which says whether it’s normal or not. If the AI says the hip is abnormal, then you proceed to get it checked or you get sent to orthopedics to have a more thorough scan done,” says Jaremko.

If diagnosed within the first three months, hip dysplasia can be corrected relatively easily and without any invasive procedures.

Other uses of teleradiology that Jaremko plans to pursue in the future include AI-assisted scanning in cardiology and ultrasounds during pregnancy.

“In Canada, it’s standard practice to do ultrasounds, but in many other countries around the world, it’s not, because of the lack of resources and expertise to understand the images. AI can solve that problem.”

Share this article