Improving healthcare for all
radiology

AZmed partners with GE Healthcare to better diagnose fractures

Rayvolve is the first French CE-marked algorithmic AI tool designed for use in standard radiology.  It saves time and improves accuracy for radiologists.

7 Dec 2021

AZmed‘s Rayvolve software is now available on GE Healthcare’s PACS (Picture Archiving and Communication Systems) via the Edison™ Software Marketplace platform.  This cutting-edge deep learning algorithm optimizes the workflow of radiology professionals, reducing image interpretation time by 36% on average, improving fracture detection and reducing diagnostic errors by 20%.  

3 million X-rays were collected, labeled and then used to train the AI algorithm.  Result: 96% sensitivity in X-ray analysis, with a specificity of over 86%.

AZmed has entered into this partnership with GE Healthcare, a world leader in medical image diagnostics, in order to “ingeniously integrate artificial intelligence into General Electrics’ tools in a manner that is accessible and appropriate, without any additional clicks for healthcare professionals,” said Julien Vidal, CEO and co-founder. 

When AZmed was created in 2018, the co-founders – including an external doctor from the AP-HP – took as their starting point the fact that medical imaging needs were increasing dramatically, but the number of specialists was not, and that therefore the lack of time and availability of professionals was clearly a problem.

“By partnering with GE Healthcare, AZmed is enabling healthcare professionals to become familiar with AI, with a consequent saving of time when examining X-rays, due to all the information being in the right place at the right time and the diagnosis being precise, which thus provides an additional safety net, more often than not in an emergency context,” Julien Vidal, CEO and co-founder of AZmed.

Once configured, Rayvolve receives the X-rays sent from the PACS, downloads and analyzes them before sending back the originals with a duplicate, all in less than one second.  This duplicate is an AI-annotated radiographic image that provides information to radiologists to help detect fractures.

Professionals do not therefore need to change their display console.  The AI prediction appears seamlessly on the same image, without obstructing the radiologist’s view.  

AZmed‘s first fracture detection solution received its CE mark in June 2019, and the second solution for thoracic pathologies before the summer of 2021.  Rayvolve is already used by more than 50 medical imaging centers in France.  The software is expected to receive its U.S. marketing approval during 2022.

What are you looking for?