New software improves detection of kidney transplant rejection
French researchers have developed software to better diagnose kidney transplant rejection cases. Doctors can better tailor treatment to their patients.
It took four and a half years This new tool should be born This makes the work of doctors easier. This software works like a Artificial intelligence to decideAnalysis of a patient, his biopsy, all his medical data, whether kidney transplant is rejected or not. “The machine does the work we used to do in our heads, it does it by itself but it’s not wrong”Valentin Goutaudier explains.
>> Health: Artificial intelligence “can lighten the doctor’s technical work”, explains Professor Jean-Emmanuel Bibault
He is part of the team of researcher and kidney transplant surgeon Alexandre Lupi, the brains behind this medical breakthrough. He is the one who came up with the idea for this online platform, diagnostic errors are very frequent, around 40%. These errors were mainly due to the more complex classification, the Banff classification, a reference for diagnosing transplant rejection, with millions of possible scenarios to confirm or invalidate rejection. So, a team of 40 people, doctors, data scientists, IT developers, all supported by AP-HP, Paris-Cité and Inserm University, Alexandre Loupy tackled this problem.
“Physicians have long complained that this classification is too complicated, and he speculated that automating the classification would stop making mistakes.”Alexandre Lupi, physician
From the interface, patient data can be entered one by one, “And in less than a second, you’ll have your diagnosis”Alexandre Loupy explains. “We integrated these millions of views into an algorithm, which is like a calculator. The human brain can make mistakes, but not a calculator.”, Daniel Yu, a data scientist who worked on the project. A small revolution technical and medical French, which allows to modify the treatment depending on the response.
The trial was conducted on 3,000 patients
The tool is already in practice at several Parisian hospitals, such as Necker or Saint-Louis, where Dr. Gillian Dewart works. “This kind of tool doesn’t replace a doctor.”He says, “But it really allows us to come up with the correct diagnosis and the fact that we can move on to a treatment, which is very useful in practice.”
“It’s safer to get a patient because the most common type of undiagnosed rejection is untreated rejection. And untreated rejection increases the risk of losing parasites.”Gillian DeWard, a physician at St. Louis Hospital
“So a patient returns to dialysis or a transplant, which has devastating consequences in terms of the patient’s quality of life or life expectancy.”, adds Gillian DeWard. An experiment conducted with 3,000 patients has proven the effectiveness of this computer tool, which has been validated by all international transplant societies. It will soon be used by groups around the world to improve patient care.