This text is part of the special section Artificial Intelligence and Cybersecurity
“From a research point of view, the future is exciting,” says Arnaud Droit, researcher at the CHU de Québec-Université Laval Research Center. He, who has just obtained the Inria International Chair in Artificial Intelligence linked to health, wants to extract as much meaning as possible from health data.
In his laboratory, the associate professor in the Department of Molecular Medicine of the Faculty of Medicine of Laval University and his team of around thirty people work on themes related to human health. They are particularly interested in hormone-dependent cancers (breast, prostate), cardiovascular problems, and degenerative diseases. “We are in a world where technologies generate more and more data. What we want to do is make the data speak, and understand how and why a particular patient reacts to a particular treatment,” he explains.
Major advances
In recent years, major advances in deep learning and machine learning have hinted at several promising future applications in the field of health. “We have a gold mine of data available that could be used for research to improve the prediction and treatment of diseases,” underlines Mr. Droit. This data puts into context the person’s family history, their lifestyle habits, their personal data, to bring out a model that would allow treatment to be better adapted. For example, machine learning can extract data about the patient’s trajectory, which could help predict the risk of recurrence for a breast cancer patient.
The analysis of images has also made dazzling gains in recent years. “There will always be humans, but in certain cases, AI is capable of detecting signs further upstream,” assures Mr. Droit. AI can also stratify patients, that is, categorize them, to determine whether or not some are more likely to develop a particular disease, so that doctors can then ensure closer monitoring. These uses are not yet implemented in the field, but are under development.
For precision medicine
The new Chair, awarded by the National Institute for Research in Digital Sciences and Technologies (Inria), establishes collaboration between the Laval University team and the Université Côte d’Azur, which has an important research center in AI. The chair’s work will focus, among other things, on the development of algorithms to detect patients at risk of aortic valve stenosis (a narrowing of the opening of the aortic valve that blocks blood flow) and who should undergo a surgery to replace it. By combining the analysis of data from blood samples and medical imaging, researchers also wish to determine a treatment that would slow the progression of the disease.
These analysis tools can then be used in other projects, for example the identification of biomarkers to detect diseases such as cancer. “The goal is to identify the molecules and characteristics of the disease, to monitor it longitudinally, and to determine when to intervene,” observes Mr. Droit.
“We are so overwhelmed by information, and humans have limited analytical capacity. The advantage of AI is that it makes it possible to agglomerate astronomical quantities of data that go far beyond the analytical capacity of a human. Once well trained and configured, it is also very quick to analyze data that it has already observed,” believes Mr. Droit.
Indeed, advances in artificial intelligence suggest increasingly precise medicine. In the field, therefore, doctors will ultimately be able to offer more targeted treatments and ensure better patient monitoring. “But to apply these advances made by research, it is another job, which will have to involve confidence in these technologies by the clinical environment,” nuance Mr. Droit.
AI to support pharmacists
This content was produced by the Special Publications team at Duty, relating to marketing. The writing of the Duty did not take part.