Imagine an operating room where an artificial intelligence-powered machine monitors a patient for unnoticeable signs of trouble and alerts doctors and nurses to ill effects of surgery before they become critical.
Or imagine that AI is trained to read diagnostic images so accurately that it can detect illness and injury earlier and more accurately than ever before.
Like antibiotics, X-rays and vaccines before them, AI and machine learning (ML) are poised to transform medicine, and the results are already starting to show in our healthcare system.
“AI and machine learning represent the most promising technology that can transform current medical practice and therapy designs,” said Bo Wang, president of the Canadian Institute for Advanced Research (CIFAR) at the Vector Institute for Artificial Intelligence.
AI will ease physicians’ workload by streamlining mundane tasks such as image analysis, medication matching, notes, and so on, said Dr. Wang, who is also the chief scientist on the AI team at University Health Network’s (UHN) Peter Munk Cardiac Center. Abundant clinical data is already available and AI algorithms can be trained to make fast and accurate clinical predictions that benefit doctors and patients, making diagnosis easier. And last but not least, AI enables fast drug discovery platforms.
No medical field will be left untouched by AI
The Royal College of Physicians and Surgeons of Canada predicts these sweeping changes will happen soon.
“Many predict that specialty medicine could fundamentally change, and beyond that, no area of medicine will be untouched by the profound transformation these technologies are expected to bring,” says a report from the College’s Task Force Report on Artificial Intelligence and Emerging. Digital Technologies published last year. “Applications of AI in healthcare are not a future consideration; they are here today,” it reads.
dr. Wang and his UHN lab are developing AI algorithms that integrate data from a variety of sources, such as electrocardiograms (ECGs), magnetic resonance imaging (MRI), clinical outcomes, doctor’s statements, and demographic and genomic information to predict heart-related outcomes for patients.
“Early and accurate diagnosis of cardiac events can save people from catastrophic life-changing events,” says Dr. Cheek. And as AI’s role in drug discovery continues to grow, he adds that it’s important to establish guidelines on how groundbreaking algorithms are communicated to the general public.
“Not as a magic panacea, but with a nuanced, coherent explanation based on science,” says Dr. Cheek.
Monitoring these new applications is an emerging consideration. The College task force report includes 12 recommendations, including the introduction of a new discipline of clinical informatics and measures to ensure AI democratizes healthcare rather than widening the “digital divide” for marginalized populations.
“The ML algorithm learns patterns in the data even when such patterns come from bias, noise, quirks or other sources,” the report states. “Therefore, the data collection process requires significant care and a certain level of expertise and knowledge.”
Better use AI
While a wealth of health data is generated today, it’s not widely used, said Frank Rudzicz, a faculty member and CIFAR chair in AI at the Vector Institute and one of its 14 task force members.
Still, there are a wide range of applications for AI and ML in healthcare, adds Dr. Rudzicz, including genetics, patient-centric consumer apps that interact with patients such as chatbots, clinician workflow apps, monitoring, and data science.
He says the technology could save time and help prevent errors, for example by using AI to monitor operations for side effects.
“So if there’s bleeding during surgery that’s more intense than we’d expect…if we can identify those cases in the first place and what caused them, we can structure surgeries much more effectively,” says Dr. Rudzicz, who is also an associate professor in the computer science department at the University of Toronto.
AI and ML can provide insights that improve medicine and the healthcare system, he says. Patients may not even notice much of a difference in their interactions, but the system as a whole will be made much more effective, he says.
Meeting the challenges of AI
Still, there are challenges to advancing the technology, notes Dr. Rudzicz op. One is the long, expensive process of clinical trials and regulatory approval, and another is the practicality of introducing new technology into the real world. Even the billing process will have to be revised to recognize these emerging technologies.
There is also a natural reluctance to change. “Sometimes you talk to clinicians and they’ve been practicing the same way for decades,” notes Dr. Rudzicz op. “When it comes to a new technology that they don’t know about, that can be a big concern.”
AI will not replace doctors, he says, but will improve patient outcomes and quality of life for doctors and nurses. “It won’t be a robot doctor coming up to you and asking you how you feel. More doctors will gain insights from data that will improve things,” he says. “It’s a revolution, but it’s a revolution that won’t happen overnight.”
Regenerative Medicine: Where Will Stem Cells Take Us? This Globe and Mail webinar, presented by Bayer, looks at how researchers are working on stem cell developments that could change the future of medicine. Join the free virtual event on November 30, 1:30 PM to 2:30 PM ET. Register here: http://globeandmailevents.com/regenerativemedicine