In 2017, when artificial intelligence (AI) was not widely used in patient care or research to improve outcomes, Dr. Barry Rubin envisioned its potential.
The Medical Director of UHN's Peter Munk Cardiac Centre recruited a team with the vision of integrating the transformative power of AI with clinical expertise. AI would empower a personalized medicine approach enabling doctors to analyze genetic sequences and imaging studies like CT and MRI scans, and to develop specialized treatments for each patient.
The use of AI would also help health care providers do their jobs more efficiently, with less time spent on administrative tasks so that more time could be spent caring for patients.
"We put heart surgeons, cardiologists and vascular surgeons together with experts in AI, and created an environment where all the medical information on every patient can be analyzed with AI in a protected, secure way," says Dr. Rubin.
The decision to invest big in AI was prescient – the Peter Munk Cardiac Centre AI team has led over 135 transformative projects at the centre, many of which have been published in top medical journals worldwide.
"We had a lot of foresight," says Dr. Rubin. "Algorithms that the Peter Munk Cardiac Centre AI team has developed are being used by health care providers and researchers around the world and are impacting the lives of millions of patients."
Now, the Peter Munk Cardiac Centre is a leader in the AI space, engaged in groundbreaking, AI-driven projects for cardiac and vascular conditions. These projects are already making a big impact on the length of patients' stays in hospital, the accuracy of their diagnoses and, most importantly, on improving patient outcomes.
Maximizing efficiency
"The Peter Munk Cardiac Centre has been really at the forefront of developing and integrating AI into cardiac and the vascular care," says Dr. Bo Wang, who was recruited by Dr. Rubin from Stanford University to lead the Peter Munk Cardiac Centre AI team, and is now UHN's Chief AI Scientist – the first role of its kind in Canada.
He says that one of the biggest achievements has been fine-tuning AI's language model to ensure that it understands medical terms to generate the most accurate clinical notes possible.
The team also developed AI algorithms that can help cardiologists and surgeons to segment – divide images into multiple parts to provide the most precise imaging – when viewing organs on CT scans, MRIs or echocardiograms.
"These tools can help improve the efficiency of segmentation by almost tenfold," says Dr. Wang, improving diagnoses and better informing treatment.
Dr. Rubin says while many other hospitals are just beginning to explore AI, the Peter Munk Cardiac Centre is already assessing and finessing its AI applications.
"We're in the process right now of measuring how well it does," says Dr. Rubin. "Did patients have a better experience with the health care system? Were there fewer complications? And did their survival rates increase?"
A big part of this refinement comes down to maximizing efficiency, says Dr. Rubin.
"Budgets are precious," he says. "If you can use AI to improve the efficiency of health care operations, and if we can do more with the same or more with less, and improve patient outcomes, that's a great use of AI."
Predicting risk and saving lives
Dr. Rubin is particularly proud of a series of AI-enabled projects that Peter Munk Cardiac Centre researchers are working on that will help prevent emergency surgeries, keep more patients out of hospital and save lives. The ability to better predict risk and guide treatment decisions is at the heart of these projects.
Dr. Kathryn Howe, a vascular surgeon, is exploring how the composition of plaques in a person's neck arteries can predict stroke risk.
Some plaques – collections of fat and inflammatory cells in blood vessel walls that build up over time – can remain stable for a lifetime, while others are vulnerable to rupture, causing a stroke. The problem is that doctors don't know which is which.
Dr. Howe's project leverages AI to analyze huge amounts of ultrasound and biological data to identify who will be at risk before a devastating event like a stroke occurs.
"She's looking at how you can predict who's going to benefit from neck artery surgery – and who's not," says Dr. Rubin.
Dr. Howe is also investigating what drives the formation of the plaques in the first place, in order to develop new, preventive treatments.
Dr. Phyllis Billia, a cardiologist and heart failure specialist, is using AI to analyze the thousands of different heart proteins to determine which ones could lead to heart failure – a life-altering condition affecting 64 million people globally and the leading cause of hospitalization in Canada.
She and her team are developing a machine learning model that can analyze the 10,000-plus proteins present in the heart at any given time and identify which proteins can predict rehospitalization or recovery in people with heart failure. Ultimately, this will allow doctors to provide the right care for each patient, according to their individual risk profile.
AI is also being used by cardiologist Dr. Arnon Adler to study enlarged hearts – a condition called hypertrophic cardiomyopathy (HCM) – to see how genes interact to cause this condition. Patients with HCM develop abnormal thickening of the heart muscle, leading to symptoms that can include chest pain, shortness of breath and even sudden cardiac death.
Dr. Adler is developing precision-medicine tools to better understand the root causes of HCM and determine the optimal treatment for each patient.
Another AI-driven project is being helmed by Dr. Maral Ouzounian, the Peter Munk Cardiac Centre's Head of the Division of Cardiovascular Surgery. She is examining which patients are at risk of rupture of an aortic aneurysm – a bulge in the main artery that comes out of the heart – which has a mortality rate of 65 per cent to 90 per cent.
Dr. Ouzounian is using AI to interpret cardiac MRI and CT scans and better predict outcomes for individual patients that have an aortic aneurysm.
"This could mean the difference between going to emergency surgery versus sitting in a doctor's office and using AI to predict what your aneurysm is going to do," says Dr. Rubin.
Enhancing the 'human element'
Dr. Wang is buoyed by the world-leading research happening now at the Peter Munk Cardiac Centre – and what research AI will enable in the future. These developments will enable doctors to do more to help people live longer, healthier lives, he says.
"AI does not replace the human element in medicine – it enhances it," Dr. Wang says. "In the long run, AI will serve as a powerful tool to empower clinicians, enhance patient relationships and improve outcomes without sacrificing compassion."
There's certainly buy-in at the Peter Munk Cardiac Centre, which is now a world leader in the most revolutionary technology of our age.
"The Peter Munk Cardiac Centre and UHN have embraced AI because we see how AI will transform health care," Dr. Rubin says. "The future of medicine is going to be AI-enabled."
This article originally appeared in The Globe and Mail.
