How Canada is leading the charge in AI-assisted surgery
Dr. Amin Madani, a general surgeon and director of the Surgical AI Research Academy at University Health Network (UHN), has spent most of his career obsessed with making surgery better—both for patients and surgeons.
While completing his PhD in surgical education, Dr. Madani aimed to understand how errors occur in the operating room and what characteristics and behaviours define elite surgeons. “For the longest time, I’ve been focused on mapping that out,” he says. “It’s not so much about having good, steady hands but instead it’s about how they use their brain and make decisions.”
As technological capabilities and accessibility advanced, Dr. Madani realized there were exciting opportunities to use computer brains—or artificial intelligence (AI)— to supercharge surgeons’ skills. “I met people doing computer vision AI research who started looking at my work and that’s how my research took a completely different turn,” he says. “The rest is history.”
Visionary new skills
While UHN is known for handling some of the most complex surgeries in the country, complications from even the most routine surgeries are actually quite common. “On average, in approximately five per cent of procedures performed worldwide—no matter which subspecialty or part of the world—people suffer complications,” says Dr. Madani. “When you look at what led to those complications, it’s typically a preventable lapse in judgment, or a decision that wasn’t made properly or the surgeon wasn’t aware they weren’t cutting in the right place.”
All too often, these problems aren’t caused by lack of skill or expertise—they’re simply the result of human limitations, such as scope of vision and the fact that every patient’s body is unique, though human anatomy tends to be the same. “AI can be like decision support,” he says. “For instance, it can act like a GPS for surgeons. It gives an extra set of eyes to watch you and make sure you don’t cut in the wrong area.”
At UHN, which is home to the largest surgical program in Canada, the Sprott Department of Surgery, Dr. Madani is working with a team of interdisciplinary experts, including computer and software engineers, to take AI-enhanced surgeries from potential to reality. Having launched the first prototype that uses AI to provide real-time information about where it’s safe to dissect, he’s now working on preliminary clinical trials to confirm the benefits.
“Right now, we are starting to use AI to assist with breast cancer surgery, gallbladder surgery and colorectal surgery,” Dr. Madani says. “We have to make sure it’s done properly, improves patient outcomes and presents evidence that it works. There are a lot of research questions to answer—everything from feasibility and workflow questions like "Does this technology improve this specific type of operation? But from a proof-of concept point of view, we’ve seen AI really does what we’ve trained it to do.”
“My dream is one day we can have a global data infrastructure where we can create the algorithms and disseminate them anywhere in the world, anytime” —Dr. Amin Madani
Improving surgeries of tomorrow
It may still be a while before the world can experience the benefits of Dr. Madani’s augmented procedures, but step inside his operating room today and it’s possible to catch a glimpse of the future.
“I know people might think of a robot sitting in the operating room, but it’s really just a piece of software,” says Dr. Madani. “It’s an AI algorithm that receives an input of data and then it spits out an interpretation or prediction on what you’ve trained it to do.”
While that sounds fairly simple, Dr. Madani says there’s a learning curve. “Bringing AI into a real-life clinical setting is challenging,” he says. “It’s similar to an experience I had recently buying a new car for the first time in many years—I wasn’t used to all the different alerts and warning signals that suddenly come into play.”
Having shown that AI surgery is possible, for Dr. Madani it’s now about making the technology fit with how surgeons work. “We’re trying to figure out how to give the right kind of guidance, warning signals and information at the right time without interfering with the surgery,” he says.
As his team learns more, this may take on different forms. “The current prototypes our engineers created are plug-and-play solutions where you can insert the software into your OR and it will support you during the surgery,” Dr. Madani says. “In this case, we’re training the AI to see the surgical field and make interpretations on the anatomy—things like ‘This is where critical anatomy is—be careful.’”
“We are talking about minimally invasive keyhole surgeries where we put a camera inside the body to do the surgery. The video feed is the input to the AI and then it gives us suggestions on a screen while we are operating.”
Local innovation—global potential
While augmenting surgeries with AI has the potential to improve the care of Canadian patients, Dr. Madani’s hope is that, in time, the benefits will be even further reaching. Specifically, he hopes that with AI guidance, surgeons in remote areas will be able to expand the types of treatments they offer. “Working at UHN, if I need to get a second opinion, I have access to world experts,” Dr. Madani says. “I can just pick up the phone, if they’re not already across the room.” That’s expertise he wants to democratize.
“My dream is one day we can have a global data infrastructure where we can create the algorithms and disseminate them anywhere in the world, anytime,” he says. For example, a remote surgeon operating in the middle of the night who needs a second opinion would be able to tap into guidance or support to make sure they don’t cause a complication.
If this all sounds cool, it’s also worth noting that Dr. Madani himself is blown away by the potential, too. “My wife took a video of me back when I saw AI in action for the first time when we developed the prototype a few years ago. My mind was blown and I almost fell off my chair!”
