How AI can detect diseases doctors aren’t looking for

This is the sixth feature in a six-part series looking at how AI is changing medical research and treatment.

When 58-year-old Will Studholme arrived at Accident and Emergency, an NHS hospital in Oxford, with gastrointestinal symptoms in 2023, he didn’t expect to be diagnosed with osteoporosis.

This disease, which is strongly associated with age, causes bones to become weak and fragile, increasing the risk of fractures.

It was discovered that Mr Studholme had a severe case of food poisoning, but early in the investigation into his illness, he was given a CT scan of the abdomen.

That scan was later run through artificial intelligence (AI) technology, which identified a collapsed vertebra in Mr Studholme’s spine, which is a common early indicator of osteoporosis.

Further tests began, and Mr. Studholme came up with not only his diagnosis, but also a simple treatment: an annual infusion of an osteoporosis drug that is expected to improve his bone density.

“I feel very fortunate,” says Mr Studholme, “I don’t think this could have been lifted without AI technology.”

It’s not unheard of that a radiologist may see something incidental in a patient’s imaging – an unknown tumor, a concern related to a particular tissue or organ – outside of what they were originally examining.

But applying AI in the background to systematically comb through scans and automatically identify early signs of common preventable chronic diseases that may arise – regardless of the reason the scan was originally ordered. Has gone – is new.

The clinical use of AI for opportunistic screening, or opportunistic imaging, as it is called, “is just beginning,” says Perry Pickhard, a professor of radiology and medical physics at the University of Wisconsin-Madison, who was among those developing the algorithm. Are from.

It is considered opportunistic because it takes advantage of imaging that has already been done for another diagnostic purpose – whether it be suspected cancer, chest infection, appendicitis or abdominal pain.

This has the ability to catch previously undiagnosed diseases at an early stage, before the onset of symptoms, when they are easier to treat or prevent them from progressing. “We can miss a lot of the prevention gaps that we missed before,” says Professor Pickhardt.

He says routine physical or blood tests often fail to detect these diseases.

Miriam Bredella, a radiologist at NYU Langone who is also developing algorithms in this area, says there’s a lot of data in CT scans about body tissues and organs that we don’t really use.

And while it can theoretically be analyzed without AI by the radiologist doing the measurement – ​​it will take time.

She says technology also has benefits in terms of reducing prejudice.

For example, diseases like osteoporosis are thought to mostly affect thin, elderly white women — so doctors don’t always think to look outside that population.

Opportunistic imaging, on the other hand, does not discriminate that way.

The case of Mr. Studholme is a good example. Being a relatively young male with no history of broken bones for osteoporosis, he is unlikely to have been diagnosed without AI.

In addition to osteoporosis, AI is being trained to help opportunistically identify heart disease, fatty liver disease, age-related muscle loss and diabetes.

While the main focus is on CT scans, for example of the abdomen or chest, other types of imaging, including chest X-rays and mammograms, are also opportunistically gathering information.

The algorithms are trained on thousands of tagged previous scans, and it is important that the training data includes scans from a wide range of ethnic groups if the technology is going to be deployed on a variety of people, experts said. Insisted.

And there is supposed to be a level of human review – if the AI ​​finds something suspicious it will be sent to a radiologist to confirm before reporting it to doctors.

The AI ​​technology used to examine Mr Studholme’s scans belongs to the Israeli company Nanox.AI, one of a handful of companies working on AI for opportunistic screening – to aid in accurate and fast diagnosis. There is more focus on using AI for Scans are actually conducted for specific circumstances.

nanox.AI offers three opportunistic screening products aimed at helping identify osteoporosis, heart disease and fatty liver disease respectively from routine CT scans.

Oxford NHS hospitals began testing Nanox.AI’s osteoporosis-focused product in 2018, before officially launching in 2020.

Results from Oxford hospitals show the number of patients with vertebral fractures has increased by up to six times the NHS average – patients who can be screened for osteoporosis, and started treatment to combat the disease. , says Professor Qasim Javed, an expert in osteoporosis and rare bone diseases at the University of Oxford who led the introduction of the algorithm.

Further trials of the algorithm are now also underway at hospitals in Cambridge, Cardiff, Nottingham and Southampton. “We want to build the evidence for using it in the NHS,” says Professor Javed.

Yet while the technology could benefit individuals, Sebastian Orcelin, professor of healthcare engineering at King’s College London, who heads the AI ​​Center for Value-Based Healthcare, says it has wider implications that need to be considered.

He said a major problem that needs to be balanced is the additional patient population that may arise from the use of technology. “This is increasing, not reducing, the demands on the health care system,” he says.

First, people who are flagged as potentially having the disease by opportunistic screening will need further confirmatory testing, which requires resources. And, if the AI ​​is wrong or too sensitive, it could result in a lot of unnecessary testing.

Then services are needed for additional people who are diagnosed.

Professor Javed admits the extra weight is a challenge that comes with technology – but there are solutions too.

In Oxford, patients with confirmed fractures are largely referred to a nurse-led fracture prevention service to avoid burdening doctors. “AI forces you to change your path,” he says.

Professor Javed believes that in the long run the NHS will save money by having more people with early-stage osteoporosis identified and getting the preventative treatment they need. “The main reason people end up in the hospital is fractures,” he says.

Mr. Studholme has seen the devastation of osteoporosis firsthand: His mother broke both of her hips because of it. He says that earlier it was considered only an old person’s condition and nothing could be done. “I feel very fortunate that I can do something before my bones turn to chalk,” he says.