DR News: To the Bone
By Will Briggs, MD, PhD
Radiology Today
Vol. 25 No. 5 P. 5

AI and opportunistic imaging could aid bone health care.

Depending on which estimates you use, osteoporosis and the fragility fractures it causes cost the United States health system between $50 billion and $100 billion each year, with the human impact even higher. For example, the 12-month mortality rate after an osteoporotic hip fracture can be as high as 30%, which is worse than many cancers. Compounding this is the fact that most cases of osteoporosis are only diagnosed after a fracture has occurred. Importantly, there are interventions that can be highly effective in reducing the risk of fractures. Despite all of the above, bone health isn’t anywhere near as high up on the agenda as heart disease, diabetes, or cancer, and with a rapidly aging population, that could be a catastrophic public health mistake. Radiology, however, has a unique role to play in improving the situation.

An Imperfect Standard
Dual-energy X-ray absorptiometry (DEXA) is the current gold standard for investigating bone health. DEXA is noninvasive, has relatively low radiation, and is widely established. However, DEXA measures bone mineral density, which is a metric of bone quantity. In contrast, bone strength is a function of bone quantity, quality, and morphometry, so DEXA provides an incomplete picture of bone health.

In the context of osteoporosis, what we really care about is preventing fragility fractures. These fractures are not just a consequence of poor bone health but are also associated with fall risk and other clinical factors. Therefore, we should not overly rely on DEXA, which provides an incomplete picture of one aspect of fracture risk and should instead take a more holistic view of patients.

Some significant issues exist around the technology and processes used for diagnosing and managing osteoporosis. We can broadly group these together as being “downstream problems.” If we look upstream, the question becomes, “Are the right people on this care pathway?” The US Preventive Services Task Force says if you’re female and over the age of 65, you should have a DEXA scan, which should potentially be repeated every two years. The problem is that there is a significant amount of guideline nonadherent care; in fact, only about 25% of women who should be screened for osteoporosis actually are. So, problem number one is getting people onto the care pathway, which is where opportunistic screening comes into its own.

A Big Opportunity
Hundreds of millions of X-rays are performed every year for a variety of reasons. Generally, when a physician orders an X-ray, they do so with a particular clinical question in mind. For example, “Mrs Jones fell on her hip; has she fractured it?” or “Mr Smith has a cough and difficulty breathing; does he have pneumonia?” This inherently biases the interpretation.

With opportunistic osteoporosis screening, if you’re coming in for any X-ray, we’re also going to check your bones; it’s as simple as that. It’s not predicated on the physician thinking about osteoporosis screening and juggling that with their other competing priorities. It’s not predicated on the patient responding to a letter asking them to come in for osteoporosis screening. It allows us to use X-rays already being performed to provide additional health value to the patient beyond their initial complaint. Ultimately, it will enable us to find people with signs of poor bone health and high fracture risk and get them onto the correct care pathway.

One question I get asked is whether using X-rays for opportunistic screening will save time and reduce the number of scans being performed. I do think it will save time (by getting people onto care pathways in a more automated way and preventing the downstream consequences of untreated osteoporosis), but I don’t think it will reduce the number of DEXAs being performed. Quite the opposite. DEXA is still the gold-standard diagnostic test. Opportunistically screening X-rays for osteoporosis will serve a prescreening role. If anything, it’s likely that more of the right people will receive a DEXA, which is good news for imaging centers and for those patients who need it.

In the future, do I believe that making an X-ray based assessment of somebody’s bone health to a level that meets or exceeds DEXA will be possible? Absolutely. But the technology and the products aren’t there yet.

The AI Question
There’s tremendous hype in health care surrounding AI, perhaps nowhere more so than in radiology. Undoubtedly, AI can be a powerful tool, but ultimately, it’s just a tool. It’s how that tool fits into the workflow and what problem it solves that really matters. I have yet to see any concrete examples where AI has created a “10x person,” let alone a 10x physician. Instead, there are plenty of examples where it has added incremental value to workflows and processes, which add up to a modest but meaningful impact.

At my company, we use a combination of proprietary biology, physics, AI, and traditional image analysis to uncover hidden signals in X-rays. This is an example of the role I think AI will play. It will not be the savior or scourge; it will be another useful tool in our fight against disease.

Changing the Status Quo
Maintaining the status quo of osteoporosis care is not a viable option. The harm it causes and the costs it imposes are already too high, and the challenge of an aging population will make the situation even worse. Although there are some imperfections regarding current downstream approaches, the most pressing issue that needs to be solved is getting the right people access to things like DEXA in the first place.

The intersection of proactive care pathways, opportunistic screening, and the appropriate use of technology could be transformative in managing osteoporosis. Integrating AI-enabled tools into regular radiology procedures offers a promising avenue for early detection and better overall management of bone health. I’m excited to see the impact we can make on this field.

Will Briggs, MD, PhD, is the CEO of Naitive Technologies.