When AI becomes a healthcare tool to detect fractures

BONEVIEW software from French company Gleamer is capable of detecting fractures and reducing the rate of diagnostic errors, thanks to the use of artificial intelligence. Already in use in a dozen countries, including France, the program has just received approval from the Food and Drug Administration to be marketed in the United States.

Tech innovations that can improve accuracy and speed of diagnosis are always welcome in the field. Gleamer, a medtech startup founded in 2017 and creator of the BoneView software incorporating artificial intelligence, has just received market approval in the United States.

In practical terms, this means that US healthcare specialists will be able to use the software to diagnose fractures and other traumatic injuries on X-rays.

The software will first detect the fracture on available X-rays before submitting its result to radiologists who will make the final decision. In order to attain this authorization, key for its international growth, Gleamer launched a clinical study between July 2020 and January 2021 in several medical schools and with manufacturers of radiological instruments in order to analyze a maximum of images.

Misdiagnosis due to fatigue

The results showed that BoneView software not only reduced radiograph reading time for radiologists and non-radiologists, but also reduced the rate of undetected fractures by 29% through the machine-physician combination. In emergency departments, traumatic skeletal injuries are one of the main reasons a consultation is sought.

According to the study, misinterpretation of fractures could account for as much as 24% of harmful diagnostic errors. These failures are more common in the evening or at night due to fatigue or lack of expert radiology staff on duty.

"Radiologists' workload has doubled in the past two decades, and despite technological progress, they must analyze hundreds more images every day, requiring the readings to be highly reliable," explained Ali Guermazi, lead author of the study, Chief of Radiology at VA Boston Healthcare System and Professor of Radiology and medicine at Boston University School of Medicine.

"The assistance of AI should allow us to improve the specificity of the complementary exams prescribed after the radiography, to avoid delays in care, and to direct patients into the right therapeutic pathway. Our study was focused on fracture diagnosis, and a similar concept can be applied to other diseases and disorders."

Gleamer is already working with some hospitals in France and assists more than 3,500 radiologists and emergency physicians worldwide.