Dr. Amin Madani, a surgeon-scientist in UHN’s Sprott Department of Surgery and head of the Surgical AI Research Academy, worked with experts in computer vision to develop technology that applies AI to the surgical field.

UHN 

A team of researchers at University Health Network (UHN) in Toronto is developing a revolutionary program that harnesses the skills of surgeons by using artificial intelligence to improve the accuracy of surgery in real time.

The technology, part of the world-leading research at UHN that is transforming medicine and countless patients’ lives, is set to be used in clinical trials beginning this fall. It promises to improve surgical outcomes and find uses around the world, particularly in rural communities, during remote surgeries and in areas that lack surgical expertise.

“It’s a game-changer to be able to apply AI to the surgical field,” says Dr. Amin Madani, a surgeon-scientist in UHN’s Sprott Department of Surgery who specializes in endocrine surgery, cancers of the thyroid, parathyroid and adrenal glands. He is also head of an academic lab at UHN called the Surgical AI Research Academy.

Dr. Madani’s work is among the trailblazing research that has led to unprecedented advances in surgical procedures at UHN, achievements in health care that have made it the No. 1 hospital in Canada and – according to Newsweek magazine’s 2024 World’s Best Hospitals ranking – No. 3 in the world.

Comprising Toronto General and Toronto Western hospitals, Princess Margaret Cancer Centre, Toronto Rehab, West Park Healthcare Centre and the Michener Institute of Education, UHN has the largest hospital-based research program in the country and is a national referral centre for patient care dealing with some of the most complex medical cases.

❝

“We now have a technology that can replicate the advanced cognitive skills of expert surgeons, which would help bridge a lot of the inequities in the world.”

Dr. Amin Madani, 
surgeon-scientist at UHN

❞

“We have world experts at UHN that come together to make [my work] possible,” says Dr. Madani, who got the idea for the technology as a trainee watching senior surgeons operating using judgment that was “at another level” from his. He developed an educational program where these high-level tactical decisions could be drawn on the surgical field.

In 2020 he began working with experts in computer vision, an area of AI that trains computers to interpret and understand images. They used thousands of examples of operations captured on video in “keyhole” procedures, minimally invasive operations where a camera is inserted into the patient’s body, to train AI algorithms to help make decisions about where to dissect and cut tissues.

“It was mind-blowing when I first saw it. Here was my mental model, the way I see the surgical field, represented by a piece of software,” Dr. Madani recalls. “It’s not like an anatomy textbook when you do surgery. You have to really know where you are, and the pattern recognition is something that takes a long time to learn.”

He says the technology can help to avoid adverse events in surgery, mistakes that are estimated to cause serious complications in about 5 per cent of cases that can lead to sepsis, pneumonia and even death.

In collaboration with engineers and computer scientists at UHN – the first hospital to appoint a Chief AI Scientist – his team has produced a prototype that makes recommendations to surgeons in real time about “go” and “no go” zones. The technology projects coloured areas onto the video monitor used during an operation; green means an area of an organ is safe to cut, red means it’s not. Another variation uses a heatmap-style projection that changes colour based on the model’s confidence as to where the safe area is.

Dr. Madani says that by using such techniques it may also be possible to assist surgeons who require more expertise, such as those working in remote areas. “We now have a technology that can replicate the advanced cognitive skills of expert surgeons, which would help bridge a lot of the inequities in the world.”

He says surgeons “are open to the concept” of the technology, especially if they can “summon it whenever they want” and select the level of probability desired. “The human is still really involved. It’s not automated surgery, it’s augmentation of surgery,” he notes, adding that it’s critical to understand how surgeons will use the technology in procedures.

“We want a system that’s responsive and meets real-time surgical requirements,” says Jimmy Qiu, senior engineering manager of Techna, a research core at UHN focused on enabling technology to push the boundaries of what’s possible in medical science and patient care. “We take an idea from the clinical team, work very rapidly to bring it to a proof of concept, validate its value and then work with our collaborators to translate it into practice,” he explains.

The DATA team led by Dr. Michael Brudno and UHN’s AI Hub – a centre designed to augment human intelligence through AI health care innovation, the first of its kind in Canada – have worked closely on the surgical AI application, Qiu says. “UHN gives us the opportunity to look at issues in a multifaceted way. There was a need to build solutions in the local environment, so you can have more real-time feedback and guidance.”

One of the biggest issues with surgical AI is how to make it usable in practical cases, he points out, and to figure out the best platform for the final system.

Nurses Zareen Tasneem (left) and Brooke Benatar prepare for surgery at UHN.

UHN FOUNDATION

“We have to understand how to integrate this technology into our workflow so it doesn’t distract and truly augments our abilities,” says Dr. Madani. He expects the images will be displayed on something much like an Xbox, a plug-and-play device “that’s universal, so you can use whatever system you want.”

He notes the clinical trials will be used only for gallbladder surgery and the treatment of colorectal cancer, procedures that can be complex and that have been captured the most on video, which provided a large amount of data to train the AI algorithms.

The multidisciplinary team at UHN is working on other surgical applications, including breast, brain, lung and colon surgery, cancer operations and Dr. Madani’s own area of endocrine surgery.

Such advances are possible through the top expertise UHN is known for, says Dr. Madani, whose team will continue to push the boundaries of medical technology through the support of the hospital’s UHNited campaign, which focuses on fundraising for research, beyond the government funding that covers the organization’s operational costs.

Donations made through the UHN Foundation are invested in bold ideas that set the stage for more breakthroughs that move medicine forward, allowing patients access to some of the best care and care models in the world.

Dr. Madani is grateful to have support for the work of his academic lab, “where we do the raw research around these kinds of technologies,” from developing and designing AI surgery to interviewing end users and doing tests in the operating room.

He looks forward to the day when the technology will be in everyday use and expects more developments in AI surgery ahead. “Who knows what it’s going to look like in 10 years? I don’t think any of us can predict.”