artificial intelligence to overcome the “limitations of the human eye” and measure the effectiveness of treatment against liver cancer

On Professor Julien Calderaro’s computer, there is a sample of a liver tumor, a pink spot made up of bean-shaped cells. “Here we have a histological image where we see the tumor cells. This is the section that we use to make our diagnosis”, he describes. Pathologists examine this type of digital slide every day, but to say precisely whether the treatment against liver cancer will work on the patient, the human eye has its limits, hence the importance of the discovery which has just been made. be made by a team of researchers from AP-HP and Inserm and which was published on November 8 in the journal The Lancet Oncology.

It turns out that artificial intelligence is capable of identifying liver cancer patients who will respond to treatment, knowing that not everyone will respond. “The limits of the human eye are that there is already a part of subjectivity, explains Professor Julien Calderaro, pathologist at Mondor hospital in Paris. The other limit, the main one, are the limits of the information that we can obtain. That is to say, this image is very complex information. There is a lot of information that we cannot analyze exhaustively.”. So this is where artificial intelligence comes in.

Professor Calderaro’s team has developed an algorithm capable of identifying in these images the presence of genes that respond to treatment. This is an essential issue since liver cancer is one of the most aggressive and patients are often diagnosed at an advanced stage. For them, there is an injectable treatment but it is only effective on a quarter of patients.

“The problem at the moment is that we give this same treatment to everyone, knowing that some patients will benefit from it and others will not benefit from their cancer disease.”

Professor Julien Calderaro, pathologist

at franceinfo

Some patients even go “develop side effects. They can trigger autoimmune diseases which can be very serious and lead, fortunately in a minority of cases, to death. There is also something to take into account, and that is the cost , these are very expensive drugs for Social Security”indicates Professor Calderaro.

He specifies that his device has the advantage of being fast and inexpensive. It can be used to treat other cancers and even autoimmune diseases. Researchers from the Henri Mondor hospital in Paris, Inserm and the University of Paris-Est Créteil will now test it on a large scale before being able to generalize it in all hospitals.