The ballet of Sunday (and Saturday) athletes is part of the routine of emergency services throughout France at weekends. Simple sprain or nasty fracture? To find out, the injured all have the same reflex: push open the doors of the emergency rooms closest to them to take an x-ray. The problem is that they are never the only ones and come to crowd into already overwhelmed services, generating crazy waiting times sometimes exceeding six hours. To try to improve the care of its patients, the Rennes University Hospital has used artificial intelligence. By developing algorithms, the hospital was able to significantly reduce the time taken to treat fractures but also diagnoses of prostate cancer. Explanations.
The emergency rooms of Rennes saw 66,000 patients last year, including a good number of trauma cases. “In 2019, we realized that those who came for isolated limb trauma, such as a sprained ankle or a broken wrist, stayed on average 4 hours 45 minutes in the department,” says Dr Ulysse Donval. Thanks to Boneview software, this time is now an hour less, or even an hour and twenty minutes. Concretely, the x-rays are displayed on the screen, with a yellow frame on the lesion that the algorithm thinks it has detected if applicable, and the mention “yes” or “no” in the “fracture”, “dislocation” section. and “effusion”. In case of doubt, linked for example to a splint interfering with the image, the software also indicates this. In the very rare cases where an anomaly has escaped the AI, the patient is contacted again for treatment.
In addition to this use in emergencies, the Rennes University Hospital has also developed an algorithm to aid in the diagnosis of prostate cancer. Described as “difficult to interpret” for non-specialist doctors, these MRI images sometimes require “hundreds or even thousands of images”, according to radiologist Luc Beuzit. With the collaboration of the French start-up Incepto, the University Hospital has built a database of 6,000 prostate MRIs. Thus fed, the software is now capable of detecting a suspicious lesion, which a biopsy can confirm. Soon certified, this software will soon be marketed worldwide.