French radiology artificial intelligence (AI) software developer Gleamer is highlighting a study in Radiology that shows its BoneView software aided in lowering the false-negative rate of undetected fractures by 30%.
The AI algorithm increased sensitivity by 12% and specificity by 5% per patient without reducing reading speed on exams that were chosen due to their difficulty, according to the firm.
BoneView was trained on approximately 60,000 x-rays from trauma patients prior to the study. Researchers obtained x-rays from 600 adults who had trauma with or without one or more fractures of the shoulder, arm, hand, pelvis, leg, or foot between 2016 and 2018 for this study. Then, six radiologists and six emergency room physicians viewed the images with and without BoneView to detect and localize fractures.