This is the sixth in a six-part series exploring how artificial intelligence is transforming medical research and treatment.

When 58-year-old Will Studholme had an accident and emergency at an NHS hospital in Oxford in 2023 with gastrointestinal symptoms, he did not expect to be diagnosed with osteoporosis.

This disease is closely related to age and causes bones to become fragile, increasing the risk of fractures.

It turned out that Mr Studholme had severe food poisoning, but early in the investigation of the disease he underwent a CT scan of his abdomen.

The scan was then run using artificial intelligence (AI) technology and identified collapsed vertebrae in Mr Studholme's spine, a common early indicator of osteoporosis.

Further tests followed, and Mr. Studholm was not only given a diagnosis but also a simple treatment: annual injections of an osteoporosis drug that would hopefully improve his bone density.

“I feel very lucky,” Mr. Studholme said. “I don’t think any of this would have been possible without AI technology.”

It's not unheard of for radiologists to notice something incidental in a patient's imaging—an undetected tumor, a focus on a specific tissue or organ—beyond what they initially examined.

But applying artificial intelligence in the background to systematically comb through scans and automatically identify early signs of common preventable chronic conditions that may be brewing—regardless of the reason the scan was ordered in the first place—is new.

Clinical applications of artificial intelligence for opportunistic screening or opportunistic imaging are "just beginning," said Perry Pickhardt, professor of radiology and medical physics at UW-Madison and author of these algorithms. One of the developers.

It was considered opportunistic because it took advantage of imaging that was already done for other clinical purposes - whether it was suspected cancer, chest infection, appendicitis or abdominal pain.

It has the potential to detect previously undiagnosed diseases in their early stages, before symptoms appear, when the disease is more easily treatable or prevents progression. Professor Pickhardt said: "We can avoid many of the previously missed situations where preventive measures were inadequate."

He added that these diseases often go undetected by regular physical exams or blood tests.

There's a lot of data in CT scans about the body's tissues and organs, but we don't really use it, says Miriam Bredella, a radiologist at NYU Langone who also This field develops algorithms.

While radiologists could theoretically take measurements to complete the analysis without AI, this would be extremely time-consuming.

The technology also has benefits in reducing bias, she noted.

For example, diseases like osteoporosis are thought to primarily affect thin, older, white women, so doctors don't always consider looking beyond that group.

Opportunistic imaging, on the other hand, does not differentiate in this way.

Mr. Studholme's case is a case in point. Because he suffers from osteoporosis, is relatively young, is male, and has no history of fractures, it is unlikely that he would have been diagnosed without AI.

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

While the primary focus is CT scans (such as those of the abdomen or chest), we also work to gather information from other types of imaging, including chest X-rays and mammograms.

Experts stress that the algorithms are trained on thousands of labeled previous scans and that if the technology is to be deployed across diverse populations, it is important that the training data include scans from a wide range of ethnic groups.

And there should be some level of human review - if the AI ​​finds something suspicious, it's sent to a radiologist for confirmation and then reported back to the doctor.

The artificial intelligence technology used to examine Mr Studholme's scan belongs to Israeli company Nanox.AI, which is one of a handful of companies working on artificial intelligence for opportunistic screening - with more companies focused on using artificial intelligence to assist with accurate, Quick diagnosis of the specific conditions under which the scan is actually performed.

Nanox.AI offers three opportunistic screening products designed to help identify osteoporosis, heart disease and fatty liver disease respectively from routine CT scans.

Oxford NHS Hospital began trialling Nanox.AI's osteoporosis product in 2018 and officially launched it in 2020.

Results from Oxford University Hospitals show that six times more patients are diagnosed with vertebral fractures than the NHS average and can be tested for osteoporosis and start treatment to fight the disease, Professor Qassem Javid Kassim Javaid, an expert in osteoporosis and rare bone diseases at the University of Oxford, who pioneered the algorithm.

Further trials of the algorithm are also currently underway at hospitals in Cambridge, Cardiff, Nottingham and Southampton. Professor Javaid said: "We want to build the evidence so it can be used in the NHS."

Sebastien Ourselin, professor of healthcare engineering at King's College London and director of the Center for Artificial Intelligence in Value-Based Healthcare, said that while the technology could benefit individuals, the wider implications also need to be considered.

A big balancing act, he noted, is the number of additional patients that could result from the technology's use. "This increases demand on the health care system, not decreases it," he said.

First, people flagged by opportunistic screening as likely to have a disease may require further confirmatory testing, which requires resources. And if AI is inaccurate or overly sensitive, it could lead to a lot of unnecessary testing.

Services then need to be provided for those additional individuals who are ultimately diagnosed.

Professor Javaid acknowledges that the extra load is a challenge posed by the technology, but there are solutions.

In Oxford, patients with confirmed fractures are sent to follow-up facilities where nurses provide fracture prevention services to avoid overburdening doctors. "AI does force you to change your path," he said.

Professor Javaid believes that identifying more people with early-stage osteoporosis and receiving the preventive treatment they need will save the NHS money in the long term. "Fractures are one of the main reasons people go to hospital," he said.

Mr Studholme has seen firsthand the serious consequences of osteoporosis: it caused his mother to break both hips. In the past, he said, it was considered a disease of the elderly and nothing could be done about it. "I feel very privileged that I could do something before my bones turned to chalk," he said.