Dr. Amber Simpson isn’t interested in replacing humans with artificial intelligence. For her, it’s all about, “What can AI do that a human absolutely can’t do?”
Like, how about developing new precision-targeted drugs without animal testing?
Simpson — one of the University of Alberta’s newly minted Canada CIFAR AI chairs — is an expert in improving human health using innovative computational strategies.
She says it’s a goal in many countries to reduce the use of animals in biomedicine — not just as a way to improve animal welfare, but also to save money and time.
“Seventy per cent of the drugs that have shown utility and are tested in animals have limited to no utility in humans,” Simpson says. “We’re doing what’s essentially computational drug modelling so we can tell a manufacturer how to make a drug that will work in humans, based on analysis of high-dimensional data that comes from humans.”
Simpson says large data banks of human health information — containing everything from medical images to doctors’ notes and genetic tests — can be analyzed using artificial intelligence to identify new molecular targets involved with disease, then develop drugs that effectively inhibit or promote those molecules, and predict how patients will respond.
Of course, new treatments would still have to be tested for safety and efficacy in animals before moving on to the ultimate test of a human clinical trial. But the process would be quicker, and fewer animals would be needed than with the current process, which starts with animals.
Most important, the chances of success would be much higher.
“Our hope is that it reduces the number of failures in clinical trials because you’ve started from actual human data,” Simpson says.
“A lot of really clever AI people”
Simpson is one of 25 newly appointed faculty members focused on advancing AI research, announced last month at the Upper Bound conference hosted by the Alberta Machine Intelligence Institute (Amii). Thanks to a $30-million investment from the Canadian Institute for Advanced Research through Amii, the goal is to bring together some of the world’s best minds in artificial intelligence from a wide range of fields including engineering, environmental science, education and health.
Appointed in January as a professor Department of Radiology and Diagnostic Imaging in the U of A’s Faculty of Medicine & Dentistry, Simpson was previously a Canada Research Chair in Biomedical Computing and Informatics based at Queen’s University and Affiliate Faculty at Vector Institute.
After completing undergrad studies at Trent University, a PhD at Queen’s University and a postdoctoral fellowship at Vanderbilt University, Simpson went on to work at the Memorial Sloan Kettering Cancer Center and the Weill Cornell Medical College. When she returned to Canada and Queen’s University, she joined the Vector Institute for AI and the Canadian Cancer Trials Group.
Simpson says she was attracted to the U of A for a number of reasons, including the fact that it is home to Dr. Rich Sutton, winner of the 2024 A.M. Turing Award, also known as the Nobel of computing science.
“For a computer scientist like me, that is a really big deal,” she says, noting that her CIFAR chair enables her to dedicate a significant proportion of her time to her own boundary-pushing research.
She is excited to use Alberta’s highly integrated health data, which has spawned many data-based research innovations through the U of A’s AI + Health Hub.
She’s also attracted to what she sees as the prairie work ethic. “Alberta is a place that’s really trying to be innovative and drive forward change. And I think it’s a really exciting place to be.”
Finally, it’s the prospect of collaboration.
“Amii has a lot of really clever AI people who have been put together really thoughtfully, who are trying to solve the hardest problems in artificial intelligence,” she says.
Simpson was recently also named the inaugural director of the Dianne and Irving Kipnes Health Research Institute. Her position and that of Dr. Mohamed Abdalla, assistant professor of medicine, are being funded in part through philanthropic support from The Dianne and Irving Kipnes Foundation.
Using technology for good
Simpson expects the first AI-developed drug breakthroughs to come in cardiovascular disease, because it progresses quickly and affects so many people.
Simpson also sees huge potential in her own field of specialty — applying AI tools to diagnostic cancer imaging. Her lab creates prediction algorithms that can look at a cancer screening image and foresee what’s likely to happen next for the patient, thus directing oncologists towards the best course of treatment.
This is done by comparing a patient’s CT, MRI or X-ray scans with thousands of others, looking at specific features such as tumour measurements, but also other patterns that aren’t visible to the human eye.
AI becomes even more powerful when it is given more information to analyze, such as the molecular makeup of the tumour or the patient’s smoking history. The more data the model includes, the better it works, says Simpson.
Simpson notes that AI tools such as these are getting approval from regulatory agencies such as the FDA and Health Canada, and are being adopted in clinical settings. She cites a recent international randomized trial that found breast cancers can be detected earlier using AI-powered screening tools.
“There are just so many images that get acquired in the process of breast screening,” she explains. “You can train a computer to look at all of them and to track these very minute changes, which then triggers the radiologist to go back and take another look at that one patient’s images.”
This is particularly helpful in parts of the world where there aren’t enough radiologists, so it’s critical to quickly screen out those who don’t have cancer and focus on those who might.
One project Simpson is working on with leading U of A AI scientist Dr. Russ Greiner is to develop models that can track multiple points of data over time.
“Patients get an image, get some treatment, maybe four weeks or eight weeks go by, and they get another image. So there are continuous decisions made on behalf of that patient as their blood tests change and their images change,” she explains. “We will make better predictions and we can be more accurate and more relevant to their journey by looking at all of that data.”
Simpson is actively recruiting master’s, PhD and postdoctoral students to work in her lab, seeking a broad range of backgrounds from ethics to epidemiology to biostatistics.
Simpson is convinced that patients will appreciate the relevance and value of AI in health care within the next decade. In fact, she’s so excited about the advances she expects to see from the new Amii fellows, she boldly predicts “another Nobel Prize for the U of A” in AI and health someday.
“There’s a lot of fear-mongering about AI, but once we’re changing how patients are treated and changing how they proceed through the medical system, that’s really going to make people understand that these technologies can be used for good.”