Algorithm assesses CT scans to find patients at risk for bone fractures

An algorithm is being paired with existing CT data to identify patients with bone fragility who would be at high risk for bone fractures.


An algorithm is being paired with existing CT data to identify patients with bone fragility who would be at high risk for bone fractures.

A medical study to determine the viability of the approach makes the case that the results can be used to find patients who could benefit from preventive treatment, researchers say.

The study was conducted at the Clalit Research Institute and Ben-Gurion University, using an osteoporosis algorithm developed by Zebra Medical Vision, an Israel-based artificial intelligence company.


What’s unique about the research is that the algorithm was shown to be able to calculate bone density and identify existing compression fractures using CT scans that had already been performed for other purposes, thus offering the prospect that clinicians could leverage existing CT studies to find patients at risk for osteoporosis.

That would reduce the need for additional procedures, reducing overall cost and radiation exposure for patients.

The results are also significant because of the prevalence of osteoporosis, the cost of treating the health issues and the risk of complications from bone fractures. Researchers estimate it costs nearly $18 billion annually in the U.S. alone to treat. Other research suggests that one in three women and one in five men older than 50 will suffer an osteoporotic fracture. In addition, compliance with standard bone density screening is low, so many patients with the condition are not diagnosed or pre-emptively treated.

“We have always been aware that this is a common disease with a high risk of fatality, and even though there are some treatment options for prevention, we don’t identify at-risk patients early enough,” says Eldad Elnekave, MD, CMO at Zebra Medical Vision. “We found that all the information we need pertaining to at-risk patients already exists in CT tests done for other reasons. It’s time to utilize that information.”

The study, published recently in Nature Medicine Journal, helps illustrate an example of how analytic capabilities and big data techniques can help aid public health.

“There are hidden medical insights in imaging studies that the human eye cannot capture, but have the potential to save lives,” says Ran Balicer, professor and director of the Clalit Research Institute. “We are on the verge of developing a system that will present improved fracture prediction risk scores to all Clalit physicians and patients."

There’s potential to expand the use of the analytics to screen larger populations for osteoporosis, researchers add.

“This AI capability is now proven at scale to be able to massively increase the screened population for osteoporosis without exposing the patients for more imaging scans or radiation,“ adds Eyal Gura co-founder and CEO at Zebra Medical Vision.

The article on the research can be found here.

More for you

Loading data for hdm_tax_topic #better-outcomes...