Published on
Updated
Reading 2 mins.
Cock-a-doodle Doo ! French researchers have developed software to quickly diagnose cases of kidney transplant rejection. A valuable aid for doctors who will be able to better adapt the treatment for their patients and thus reduce the risk of loss of grafts.
This invention, we owe it to Pr Alexandre Loupy, professor of nephrology and epidemiology at the Necker-Enfants Malady hospital and director of the Research team in organ transplantation at the Cardiovascular Research Center of Paris (Parcc). This software works like an artificial intelligence that is able to determine whether or not a kidney transplant has been rejected in a patient based on several data (blood tests, biopsies and other medical data).
Once on the software interface, the doctor enters the patient’s data one by one and the diagnosis falls in less than a second. “The machine does the work we did before in our head, it does it itself but it is not mistaken”explained Valentin Goutaudier, nephrologist, to our colleagues from Franceinfo.
The solution to avoid diagnostic errors?
Prof. Loupy, who worked for more than four years on this project, aimed to reduce too frequent diagnostic errors (40% of kidney transplant rejections are not diagnosed). Mistakes that must be avoided as much as possible given the shortage of organs. As a reminder, a rejection must be treated as soon as possible at the risk of losing the graft. These errors are partly explained by an overly complex method for diagnosing rejection, the Banff classification. This integrates millions of possible scenarios to confirm or invalidate a rejection.
Thanks to this new software, these millions of possible scenarios are analyzed by an artificial intelligence. “We put these millions of scenarios into the algorithm, it’s kind of like a calculator, the human brain can make mistakes, but the calculator can’t.”said Daniel Yoo, datascientist, at France info.
A tool already tested on 3000 patients
This software was developed by 40 people. Among them, doctors, data scientists, IT developers. It was then tested with 3,000 patients. An experiment that has shown its effectiveness and has it validated by all international transplant companies. Alexandre Loupy’s team has announced that this tool will soon be used by doctors around the world to improve patient care. It is already used in several Parisian hospitals (Necker, Saint-Louis).