A new study has found that artificial intelligence (AI) can not only read electronic radiology reports including X-rays, it can also accurately identify and separate patients at risk of osteoporosis. The study has been accepted for presentation at ENDO 2020, the Endocrine Society’s annual meeting, and will be published in a special supplemental section of the Journal of the Endocrine Society.
In their abstract, Chris White, PhD, and Jacqueline Center, PhD, conclude that "automated methods of patient identification may assist fracture liaison services to identify fractures that still remain largely untreated." #osteoporosis @POWHFoundation #AI #artificialintelligence
— Endocrine Society (@TheEndoSociety) March 30, 2020
Osteoporosis affects 200 million women worldwide, most of whom are above the age of 60.
Dr. Christopher White, F.R.A.C.P., of Prince of Wales Hospital in Randwick, and his colleagues conducted this study at Garvan Institute of Medical Research in Sydney, Australia, to determine the efficiency of artificial intelligence in reading the X-ray reports in comparison with manual methods.
A tool, called X-ray Artificial Intelligence Tool (XRAIT), found five times higher number of major fractures in X-ray computed tomography report when compared with manual methods. Fracture determining services have been implemented to figure out the patients whose fractures may be due to osteoporosis.
A language processing software is also installed in XRAIT to understand the human language. 5,089 radiology reports were studied. The reports were collected from patients with age greater than 50 who visited hospital emergency and had bone imaging in three months. The XRAIT results were then compared with records handled and checked by human.
The results revealed that XRAIT found 349 people with fractures due to low bone mass as compared to 98 people who were found manually.
Professor Jacqueline, co-author and head of the Clinical Studies and Epidemiology Lab at Garvan Institute of Medical Research in Sydney, Australia, says:
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. 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.
Later, the scientist checked XRAIT on digital reports of the population of Australian adults, older than 60, who were part of the Dubbo Osteoporosis Epidemiology Study.
The results showed that XRAIT rightly identified fractures 7 out of 10 times and completely separated patients without fractures more than 9 out of 10 times. It was done on 327 reports of fractures and no fractures.
This means XRAIT can be used in hospitals and osteoporosis centers. The tool has been developed by South Eastern Sydney Local Health District, using software licensed to Abbott Diagnostics.