London medical startup attracts international attention with AI ultrasound tech | CBC News

When a soldier is injured on a battlefield, decisions have to be made quickly.

In the case of a collapsed lung, a common treatment in the heat of the moment is to insert a tube between a soldier’s ribs and into their chest. In such a situation, whether that painful procedure is the right course of action is often a guessing game.

A London startup has developed a solution — and potentially many others — and is receiving international recognition, along with $100,000 from the U.S. Army.

Deep Breathe is a medical technology company based in Western University’s research park. They’ve developed a program that can be installed onto a phone and uses artificial intelligence (AI) to accurately interpret the results of lung ultrasounds in the field and at a moment’s notice.

“Whether it’s the common cold or chest pain, or you’re having a procedure at the hospital, the lungs are a common concern. They drive a lot of attention inside and outside the medical system, so we really feel strongly about using portable ultrasound and AI to give this technology and give these insights to everybody,” said Dr. Rob Arntfield, the founder and CEO of Deep Breathe.

The company started in 2020 and has grown to a team of nine people. Deep Breathe recently placed at the top of a 150-company competition called xTechInternational, which is hosted by the U.S. Army and invites companies to offer up their best pitches that can provide value to the military.

“We beat out companies that do extraordinary things. Ultimately our solution to help save soldiers’s lives by addressing a lethal condition called pneumothorax [collapsed lung] very quickly really caught their attention,” said Arntfield. “The endorsement of the U.S. Army is extraordinarily meaningful to us and our team.”

But it’s not just about saving lives on the battlefield, according to Arntfield. The technology could be used to diagnose a number of issues in almost any setting you could think of, especially because of the technology’s portability.

In Deep Breathe’s office, a portable ultrasound scanner that fits in the palm of a technician’s hand is connected to a phone. The company’s app is loaded onto the phone, a scan is taken of a staff member’s chest, and within seconds a determination is made.

That’s the vision Arntfield and his team want to see playing out wherever possible, whether it be on a battlefield, in a hospital, or in space, and it’s one that’s driven by the AI they’ve carefully sculpted and taught over four years.

“The AI is really just putting a trained doctor’s brain inside a device,” he said.

Demonstration shows how London tech company uses AI to screen for lung problems

Dr. Rob Arntfield explains how his company Deep Breathe’s AI-powered software interacts with a portable ultrasound device to accurately assess if someone has a lung injury.

AI is the future of medicine

Blake VanBerlo is tasked with overseeing the effort to teach the AI what it needs to know to effectively and accurately diagnose a problem. The process of teaching the program is called machine learning.

“Machine learning lets us tweak software until it comes to a solution that does the thing that we wanted to do, essentially,” said VanBerlo. “Where it really excels is coming up with solutions to really complex pattern recognition problems, like image interpretation.”

The app, which is now 93 percent accurate in ruling out the presence of a collapsed lung, isn’t used to diagnose exactly what’s wrong, but rather to rule out issues, determining when they aren’t present. It’s trained by being asked to assess images and being told when it’s wrong, or when it’s right.

The introduction of AI into medicine is a natural progression given its effectiveness, specifically when it comes to pattern recognition, VanBerlo said.

“I believe the future of AI in healthcare is very bright. AI capabilities are expanding rapidly. Our abilities to to assess AI systems is getting much, much better. Our ability to get AI systems to explain their outputs is much better,” he said.

“I think we’re at the start of what’s going to be an eventual like large-scale adoption of AI in healthcare.”

Putting London on the world stage

Delaney Smith is Deep Breathe’s manager of data science research and regulatory affairs, and joined Arntfield for the xTechInternational competition in Spain.

She said the experience of pitching the product and showing off local innovation was a pride-inspiring experience.

“It was a surreal experience to get to be a part of something like that, to get feedback from people that are at the top of all of their different sectors within such a large organization [the U.S. Department of Defence],” she said. 

“That recognition amplifies what we already knew to be true, which is that the technology is going to save lives and other people want it just as much as we do.”

While Deep Breathe’s product isn’t available on the market yet, Smith hopes to have approval from the U.S. Food and Drug Administration within the next year, with eyes on Health Canada approval down the line.

While his company works toward approval, Arntfield said he continues to take pride in what he calls his “incredible” team, and the city they operate in.

“Tapping into the expertise that we have locally, both in the medical sector and the tech sector, has really allowed our dreams to come true in terms of getting this far. And we’re only getting started.”