In oncology, now, algorithms allow researchers to be more precise and faster.
Guest of the Sans Rendez-Vous show on Wednesday on Europe1, Professor Eric Deutsch explained how artificial intelligence can, for example, make it possible to predict a patient's response to cancer treatment and improve radiotherapy.
INTERVIEW
Innovations in the fight against cancer are increasing, especially with the rise of new therapies, but not only: artificial intelligence (that is to say computer programs that learn, reason, classify, here in particular images) can also help manage patients, better targeting areas and treatments.
Professor Eric Deutsch, head of the radiotherapy department at the Gustave-Roussy Institute in Villejuif, heads a research team made up of doctors, mathematicians, physicists and engineers.
All of them pool their skills to improve cancer care, in particular thanks to artificial intelligence.
He presented their progress on Wednesday in the program
Sans Rendez-vous
, the health program of Europe 1.
Why use artificial intelligence in oncology?
Immunotherapy is a process that works on a patient's immune system to help them fight their disease.
In the case of cancer, instead of directly attacking cancer cells, immunotherapy will help the cells of the immune system to recognize and destroy them.
A "revolution in cancerology", underlines the doctor.
The "main drugs which are today routinely used in oncology and immunotherapy awaken lymphocytes which will induce anti-tumor immunity", he explains.
But for these treatments to be effective, "there must be lymphocytes in the tumor and the immunotherapy can wake them up. And unfortunately, patients for whom there are no lymphocytes in the tumor do not. not benefit from immunotherapy ".
>> Find all of Sans rendez-vous in replay and podcast here
His team therefore conducted a study, partly funded by the Foundation for Medical Research (FRM) to see if it was possible, thanks to a computer program, to be able to determine the presence or absence of these famous lymphocytes in tumors.
What results?
This study showed that it was possible to predict the response to cancer treatment.
"We set up a database which consisted of samples of tumors from patients for whom we knew whether or not there were lymphocytes. And at the same time we had the scans of these patients," explains Eric Deutsch.
The team then "trained and developed an algorithm to, from an image, be able to find the presence or absence of lymphocytes in a tumor".
Result: "We were able to show that the algorithm we had developed did as well as the microscope with a specialist doctor to see whether or not there were lymphocytes in the tumor, and that it was able to correlate ( these data) with the response to treatment of patients treated with immunotherapy ".
Since then, adds the professor, this work has been confirmed by other international studies.
>> Also listen
: A new technique for detecting colon cancer using artificial intelligence
How can algorithms improve radiation therapy?
That's not all.
While artificial intelligence can help target cancer treatments better, the lab is also using AI to improve radiation therapy.
"Radiotherapy is a treatment that is 100 years old, constantly evolving and which benefits from advances in imaging, information technology and advances in biology," recalls Eric Deutsch.
"We want to make it more effective, less toxic, better target the areas we want to treat and better define the doses we want to give. And for that, artificial intelligence can help us enormously."
Indeed, radiotherapy is the "second contributor to the healing of cancers after surgery", used for "almost all tumors in current practice".
Artificial intelligence can therefore be used for many applications, which is "already a reality since today we use artificial intelligence to define the target volumes that we want to treat, something which, normally, is done. This saves us both a lot of time and precision. The more precise we are, the more efficient we are and the less toxic we are, "concludes Eric Deutsch.