AI Tool Improves Detection of Bone Fractures on X-rays

By Megan Brooks

April 06, 2020

NEW YORK (Reuters Health) - An artificial-intelligence tool outperforms radiologists in flagging fractures on imaging and may help get more patients at risk for osteoporosis on appropriate therapy to prevent another fracture.

The X-Ray Artificial Intelligence Tool (XRAIT) uses natural-language processing to screen radiology reports for fractures and identifies patients who may unknowingly have osteoporosis. In testing, the tool detected a significantly greater number of major fractures on x-ray and CT scans compared with standard reading of images by radiologists.

The study was featured at the Endocrine Society's ENDO 2020 virtual press conference on March 30.

"With XRAIT, limited health care resources can be optimized to manage the patients identified as at risk rather than used on the identification process itself," study co-investigator Dr. Jacqueline Center of the Garvan Institute of Medical Research in Sydney, Australia, said in a press release from the Endocrine Society.

"By improving identification of patients needing osteoporosis treatment or prevention, XRAIT may help reduce the risk of a second fracture and the overall burden of illness and death from osteoporosis," she added.

The research team had XRAIT analyze more than 5,000 digital radiology reports from patients over 50 who presented to a hospital emergency department and had bone imaging over three months. The XRAIT results were compared with manual review of the records of the 224 patients referred to the hospital's fracture liaison service in the same period.

XRAIT detected 349 people with fractures likely due to low bone mass compared with 98 people with fractures identified by manual reads, a 3.5-fold higher detection rate.

The researchers also tested XRAIT on an independent dataset of another population of Australian adults over 60. From 327 reports of confirmed known fractures and nonfractures, XRAIT accurately identified fractures nearly seven of 10 times and correctly ruled out patients without fractures more than nine of 10 times, according to the Endocrine Society press release.

Speaking at the briefing, senior author Dr. Christopher White of Prince of Wales Hospital in Randwick, Australia, noted that many hospitals now have fracture liaison services to identify people with fractures that could be due to undetected osteoporosis. But manually reading the radiology reports of referred patients is time consuming and misses some people at risk for osteoporosis.

The new AI tool can screen high volumes of imaging reports and identify more patients at risk for osteoporosis, "with high analytical and clinical specificity, and in numbers likely to overwhelm osteoporosis-fracture-prevention services," Dr. White said. In some respects, "you have to be careful what you wish for," he added.

Looking ahead, Dr. White said further development includes linking the AI tool with clinical risk factors and treatment data to help "target the right patient to increase patients with osteoporosis on treatment, improve productivity and safety, and reduce the burden of care to save money safely."

South Eastern Sydney Local Health District developed XRAIT, using software licensed to Abbott Diagnostics. Australia's Prince of Wales Hospital Foundation gave seed funding for XRAIT.

SOURCE: ENDO 2020, March 30, 2020.