Of all cancers, pancreatic cancer is one of the deadliest. After five years, on average, only one in ten people manage to survive this disease.
What if artificial intelligence allowed us to fight this serial killer more effectively?
This is the bet made by Doctor David Hénault, hepato-pancreato-biliary surgeon at the CHUM.
He is currently conducting a postdoctoral research fellowship at the Ontario Cancer Research Institute, where he specializes in artificial intelligence.
“We are specifically working to improve the prediction of survival or recurrence in cases of pancreatic cancer,” he explains.
Before getting to the heart of the matter and letting him tell you how he does it, here’s why I interviewed him: I recently learned of the existence of a School of Artificial Intelligence in Health at CHUM. And it is organizing a conference next Tuesday with three Quebec doctors who have obtained scholarships (including Dr.r Hénault), in recent years, to improve in the field of artificial intelligence.
The meeting aims to explain to what extent “the health sector will be transformed” by AI. David Hénault and the establishment’s director of education and academy, Nathalie Beaulieu, agreed to tell me more.
To understand the challenges that David Hénault faces, you must first know that in his field, “in 2024, the best clinical tools we have to predict a patient’s survival are the same as 10 years ago. or 20 years.”
We are talking, specifically, about histopathology, that is to say the microscopic study of tumors. During cancer surgery, the human tissue removed will be sent to a pathologist for analysis, for which microscope slides will be used.
“When we put the tissue on the slide, we have to put dyes according to the types of molecules we are looking for in order to characterize the tissue, then to determine its nature, its aggressiveness,” he explains.
It’s a long and laborious process… which can now be automated thanks to developments in artificial intelligence.
But AI also allows us to go even further. When analyzing the slides “with an algorithm that has seen the same cancer thousands and thousands of times, we are able to generate an automated preliminary diagnosis.” A method that we are starting to use at the CHUM.
The next step is to assess a patient’s chances of survival.
We traditionally use a series of variables such as age, other illnesses, other operations undergone, etc. We will also examine “the DNA of the cancer and try to identify which mutation makes it more aggressive”, and then try to find the most effective chemotherapy. Finally, medical imaging will also be used to form the prognosis.
At this stage too, artificial intelligence can be a game-changer.
Artificial intelligence algorithms make it possible to determine which signals will help us predict the outcome of a cancer patient. After that, we are able to tell ourselves where we can intervene to improve the survival rate and patient care. That’s precisely what I do.
David Hénault, hepatopancreato-biliary surgeon at the CHUM
Clarification: the progress made thanks to AI regarding pancreatic cancer will not be limited to this disease. “It is entirely reasonable to believe that the discoveries made for this cancer and the tools that we are developing will have repercussions and impacts on other cancers,” adds the doctor.
We also suspect that the progress made by AI in health will not be limited to cancer either.
“Accessibility to care and services, equity, coordination, continuity… Artificial intelligence is a new tool that can help us respond to the complexity of all these needs,” explains Nathalie Beaulieu, head of the School of Intelligence. artificial in health.
It offers another concrete example, currently developed by CHUM researchers. “We have a team that is looking at the use of connected objects, connected clothing, etc., to be able to have data that will help patients who have epilepsy, for example how to we can predict crises to better intervene. »
But the advances and progress announced do not fall from the sky. My two interlocutors emphasize the importance of being informed, but also of training in artificial intelligence.
“I think it is essential for the population, for caregivers and for all personnel in the health sector to educate themselves about artificial intelligence to improve the quality of what we discover, with the aim of “improve the quality of care administered to patients,” says David Hénault.
Nathalie Beaulieu, for her part, points out that the CHUM School of Artificial Intelligence offers conferences on a regular basis like the one in which three doctors will participate shortly.
“Everyone is invited without exception. The people of our community, the patients… It’s an invitation to a wide audience. The idea is really to raise awareness,” she says.
Because if the adventure of artificial intelligence in the health sector offers us several reasons for hope, it is only just beginning.
In the case of cancer, for example, “we are at the stage of massive gains in our understanding of the disease,” summarizes David Hénault.
“But we are still in the early stages of artificial intelligence. And we will have to continue to work so that this understanding translates into clear survival benefits for patients. »
Visit the Facebook page of the CHUM School of AI in Health
What do you think ? Participate in the dialogue