Using Artificial Intelligence to Identify Fractures on X-rays  

According to a study by Boston University School of Medicine researchers, artificial intelligence can assist clinicians in detecting fractures on x-rays. 

“Our AI algorithm can quickly and automatically detect x-rays that are positive for fractures and flag those studies in the system so that radiologists can prioritize reading x-rays with positive fractures,” chief of radiology at VA Boston Healthcare System and Professor of Radiology & Medicine at BUSM, Ali Guermazi, MD, PhD said in a press release. 

“The system also highlights regions of interest with bounding boxes around areas where fractures are suspected. This can potentially contribute to less waiting time at the time of hospital or clinic visit before patients can get a positive diagnosis of fracture.” 

According to researchers, fracture interpretation errors represent around 24 percent of harmful diagnostic errors seen in the emergency department. Additionally, inconsistencies in the radiographic diagnosis of fractures are more common during the evening and overnight hours. 

The AI algorithm was trained on a large x-ray data set from multiple institutions to identify limb, pelvis, torse, lumbar spine, and rib cage fractures. Expert human readers defined the gold standard in the study and compared the performance of human readers with and without AI assistance.  

A variety of readers, including radiologists, orthopedic surgeons, emergency physicians, physician assistants, rheumatologists, and family physicians, were used to simulate real-life scenarios.  

“Each reader’s diagnostic accuracy of fractures, with and without AI assistance, were compared against the gold standard. They also assessed the diagnostic performance of AI alone against the gold standard, the press release stated.  

“AI assistance helped reduce missed fractures by 29 percent and increased readers’ sensitivity by 16 percent, and by 30 percent for exams with more than one fracture, while improving specificity by 5 percent.” 

According to Guermazi, AI can be a powerful tool to assist radiologists and other physicians in improving diagnostic performance and increasing efficiency. Additionally, the technology has the potential to improve patient experience at the hospital or clinic.  

“Our study was focused on fracture diagnosis, but similar concept can be applied to other diseases and disorders. Our ongoing research interest is to how best to utilize AI to help human healthcare providers to improve patient care, rather than making AI replace human healthcare providers. Our study showed one such example,” Guermazi added. 

The study was funded by GLEAMER Inc and the findings appeared online in Radiology.