AI’s applications and implications in health care
Artificial intelligence (AI) and its applications in and impact on health care, from data analysis to social accountability, was the focus of the 2023 symposium hosted by the Institute for Biomedical Engineering, Science and Technology (iBEST).
The annual research symposium, held in person on April 14 at the Li Ka Shing Knowledge Institute at St. Michael’s Hospital, Unity Health Toronto, offered attendees the opportunity to hear from a keynote speaker on synthetic data and tips on commercialization, two panels on different aspects of AI and health care and to explore a trainee poster competition. Trainees were also invited to attend an AI and data analytics workshop held in partnership with IBM to gain practical experience in building machine learning pipelines and deep neural networks.
“It’s iBEST researchers and clinicians who are shaping the future of health care,” said Dr. Steven N. Liss, Toronto Metropolitan University’s (TMU) vice-president, research and innovation. He noted the hope that iBEST researchers from TMU and Unity Health will be at the forefront of TMU’s upcoming School of Medicine, which is set to open in 2025. Both he and iBEST co-director Dr. Tom Schweizer, who shared greetings on behalf of Dr. Ori Rotstein, Unity Health Toronto’s vice-president, research and innovation, highlighted the importance of collaboration. “iBEST has grown into a centre for collaboration and a significant destination for trainees,” said Dr. Schweizer, adding there are now 182 registered iBEST trainees.
Keynote address
The symposium’s keynote address was delivered by Dr. Khaled El Emam, who is the Canada Research Chair in Medical AI and professor at the University of Ottawa, as well as a senior scientist at the Children’s Hospital of Eastern Ontario Research Institute, director of the Electronic Health Information Laboratory and co-founder of Replica Analytics. Dr. El Emam covered two distinct topics as part of his presentation: an overview of synthetic data generation and tips for researchers pursuing commercialization.
For the first part, Dr. El Emam focused on the privacy benefits of synthetic data generation. Synthetic data is created by using machine learning to develop models that reflect the statistical patterns found in real health-care data while preserving the privacy of individuals. He gave an overview for the audience of how synthetic data is generated and evaluated, noting its benefits and successes.
In the second part of his address, he shared 10 lessons he has learned about the commercialization process, ranging from negotiating the academic institution’s equity stake to securing access to data for team management and attracting investment.
Panel – AI & Algorithms: Clinicians Beware?
The two AI-focused panels were moderated by Dr. Venkat Bhat, iBEST’s Healthcare AI & Analytics theme lead, who oversaw the development of the program and brought together the panellists as the symposium chair. In the first panel, Dr. Bhat asked participants to discuss the challenges and opportunities of algorithm development and deployment, how AI can enhance decision-making and how organizations will have to adjust to these innovations.
The panellists included:
- Dr. Gary Bader, Professor, The Donnelly Centre, University of Toronto
- Dr. Timothy Chan, Associate Vice-president and Vice-provost, Strategic Initiatives, Canada Research Chair in Novel Optimization and Analytics in Health and Professor, University of Toronto
- Dr. April Khademi, Principal Investigator of the Image Analysis in Medicine Lab and Assistant Professor, Toronto Metropolitan University
- Dr. Sri Krishnan, Associate Dean, Research, Development and External Partnerships, iBEST Co-director and Professor, Electrical, Computer and Biomedical Engineering, Toronto Metropolitan University
- Dr. Muhammad Mamdani, Vice President, Data Science and Advanced Analytics, Unity Health Toronto and Director, University of Toronto Temerty Centre for Artificial Intelligence Research and Education in Medicine (T-CAIREM)
- Tony Orsi, Partner at Bereskin & Parr LLP
“Should clinicians beware? I think clinicians should be aware,” said Dr. Krishnan, a sentiment echoed by several on the panel, who noted that AI use is already happening in health care, and its adoption will continue to grow. “It’s not a pipedream. This is happening as we speak,” said Dr. Mamdani, while Dr. Khademi predicted that AI has the opportunity to revolutionize medical practice.
Challenges for developing and deploying AI for health care purposes drew examples ranging from the complexity of health data to protecting intellectual property to reconsidering how student and researcher success is measured.
Panel – AI & Society: Time for A Social Contract?
In answer to this panel’s theme, the five panellists invited to the discussion agreed that the need to examine and regulate AI’s societal impact through an ethics, fairness, transparency, privacy and trust lens is a given. The panellists also agreed that interdisciplinary collaborations are crucial to addressing challenges that aren’t anticipated from a single perspective.
The panellists included:
- Dr. Ebrahim Bagheri, Canada Research Chair in Social Information Retrieval, NSERC Industrial Research Chair in Social Media Analytics and Professor, Toronto Metropolitan University
- Dr. Yvonne Bombard, Canada Research Chair in Genomics Health Services Research and Policy and Associate Professor, University of Toronto, Scientist and Director, St. Michael’s Hospital, Unity Health Toronto
- Dr. Ali Mazalek, Canada Research Chair in Digital Media and Innovation, Director, Synaesthetic Media Lab and Professor, Toronto Metropolitan University, Affiliate Scientist, Keenan Research Centre for Biomedical Science
- Tony Orsi, Partner, Bereskin & Parr LLP
- Dr. Reza Samavi, Faculty Affiliate, Vector Institute and Associate Professor, Toronto Metropolitan University
When asked to address the challenges of AI in their domains, the panellists cited AI’s built-in biases as significant. “Search will continue to be one of the biggest areas of AI tech, and there are huge biases and prejudices there,” said Dr. Bagheri. According to Dr. Bombard, “Our tools are only as good as the information we put into them.”
For Dr. Mazalek, challenges also involve designing a Human-Computer Interaction (HCI) interface that engages our bodies and minds. “We need to examine how AI is changing our cognitive capabilities. What are we losing with these technologies?” According to Dr. Samavi, “Security and privacy cannot be an afterthought in AI tools. They need to be part of the design.”
To ensure trainees and PIs develop an understanding of ethical, privacy and societal issues, the panellists noted that the future belongs to those who understand how AI tools work and have immense creativity, so the best skill to develop in AI students is how to think outside the box.
Trainee poster competition
Several iBEST trainees presented their research in a walk-about poster competition. The top three participants were invited to give a 10-minute presentation on their research and findings and to participate in a Q&A session later in the day.
The winners of the poster competition, sponsored by Bereskin & Parr LLP, were:
First Place: Wound healing: Exploring mechanism whereby a new drug Ag373K augments VEGF secretion and keratinocyte migration by Vida Maksimoska.
Maksimoska’s research tested the effects of a new antimicrobial drug on skin cells. She found the drug is nontoxic and affects certain cell proteins that play key roles in cell migration, ultimately increasing wound healing. The next step is to investigate how and why the drug activates the proteins.
Second Place: Ultrasound activation of gold nanoparticle carriers functionalized with anticancer drugs: Modelling and measurements by Tyler Hornsby.
Third Place: Investigating the characterization of airborne droplets with photoacoustic sensing by Omar Nusrat.
