Inside a dark room at the Bács-Kiskun County Hospital outside Budapest, Dr. Eva Ambroze, a radiologist with more than two decades of experience, shows a patient’s mammogram on a computer monitor.
Two radiologists had previously said that the X-ray showed no indication that the patient had breast cancer. But Dr. Ambroise was looking closely at several areas of the scan circled in red that the artificial intelligence software had flagged as potentially cancerous.
“It’s something,” she said. He hastily orders the woman to come back for a biopsy, which is happening within the next week.
Advances in AI are beginning to be successful in breast cancer screening to detect signs that doctors miss. So far, the technology is showing an impressive ability to detect cancer at least as well as human radiologists, according to preliminary results and radiologists, one of the most convincing signs to date of how AI will benefit the public. Can improve health.
Hungary, which has a robust breast cancer screening program, has one of the largest test bases for the technology on real patients. The AI systems were rolled out in 2021 at five hospitals and clinics that perform more than 35,000 screenings a year and now help spot signs of cancer that radiologists may have missed. Clinics and hospitals in the United States, Britain and the European Union are also starting to test or provide data to help develop systems.
The use of AI is on the rise as the technology becomes the center of the Silicon Valley boom, with the release of chatbots such as Chatbot showing how AI has a remarkable ability to communicate in human prose – sometimes with worrying results. Built in a similar fashion to those used by chatbots modeled on the human brain, the breast cancer screening technology shows other ways AI is seeping into everyday life.
Doctors and AI developers said the widespread use of the cancer-detecting technology still faces several hurdles. Beyond the limited number of locations using the technology, additional clinical trials are needed before more widespread adoption of the system as an automated second or third reader of breast cancer screens. The device must also demonstrate that it can produce accurate results on women of all ages, ethnicities and body types. The radiologist said the technology has to prove it can spot more complex forms of breast cancer and reduce false positives.
AI tools have also prompted debate over whether they will replace human radiologists, with makers of the technology facing regulatory scrutiny and resistance from some doctors and health institutions. For now, those fears appear overstated, with many experts saying the technology will be effective and trust patients only if it’s used in partnership with trained doctors.
And finally, AI could be a life-saver, said Dr. László Tabor, a leading mammography educator in Europe who won over the technology after reviewing its performance in breast cancer screening.
“I’m dreaming about the day when women are going to a breast cancer center and they’re asking, ‘Do you have AI or not?'” he said.
hundreds of pictures a day
In 2016, Geoff Hinton, one of the world’s leading AI researchers, argued that the technology would eclipse radiologists’ skills within five years.
“I think if you work as a radiologist, you’re like Wile E. Coyote in the cartoon,” he told The New Yorker in 2017. “You are already at the edge of the cliff, but you haven’t seen the bottom yet. There is no land at the bottom.
Mr Hinton and two of his students at the University of Toronto built an image recognition system that can accurately identify common objects such as flowers, dogs and cars. The technology at the heart of their system – called neural networks – is based on how the human brain processes information from a variety of sources. It is used to identify people and animals in images posted to apps such as Google Photos, and allows Siri and Alexa to recognize the words people speak. Neural networks also drove a new wave of chatbots, such as ChatGPT.
Many AI evangelists believed that such technology could easily be applied to detect disease and illness like breast cancer in mammograms. According to the World Health Organization, in 2020, there were 2.3 million breast cancer diagnoses and 685,000 deaths from the disease.
But not everyone thought changing radiologists would be as easy as Mr. Hinton had predicted. Peter Kxkemethy, a computer scientist who co-founded Kheron Medical Technologies, a software company that develops AI tools to help radiologists detect early signs of cancer, knew the reality would be more complex.
Mr. Kxkemthy grew up in Hungary while spending time in one of Budapest’s largest hospitals. His mother was a radiologist, who gave him his first look at the difficulties of finding a small distortion within an image. Radiologists often spend hours each day in a dark room looking at hundreds of images and making life-changing decisions for patients.
“Small wounds are very easy to miss,” said Dr. Edith Karpati, Mr. Kxkemethy’s mother, who is now a medical product director at Khiron. “It’s not possible to stay focused.”
Mr Kexkemethy, along with Tobias Rijken, co-founder of Kheiron, who is an expert in machine learning, said AI should assist doctors. To train their AI system, they collected more than five million historical mammograms of patients whose diagnoses were already known, provided by clinics in Hungary and Argentina, as well as academic institutions such as Emory University. The company, which is based in London, also pays 12 radiologists to label the images using specialized software that teaches the AI to identify cancerous growths by its size, density, location and other factors.
With millions of cases feeding the system, the technology builds a mathematical representation of normal mammograms and those with cancer. With the ability to see each image in more fine detail than the human eye, it compares each mammogram to that baseline to find abnormalities.
Last year, after a test on more than 275,000 breast cancer cases, Khiron reported that its AI software matched the performance of human radiologists when acting as a second reader of mammography scans. It cut the workload of radiologists by at least 30 percent because it reduced the number of X-rays they needed to read. In other results from the Hungarian clinic last year, the technology increased cancer detection rates by 13 percent as more malignancies were identified.
Dr. Tabar, whose technique for reading mammograms is commonly used by radiologists, spent 2021 trying out the software by recovering several of the most challenging cases of his career in which radiologists missed signs of developing cancer. In every instance, the AI saw it.
“I was surprised to see how well it did,” Dr. Tabar said. He said he had no financial ties to Khiron and that other AI companies, including Lunit Insight from South Korea and VARA from Germany, also made encouraging discoveries. Results are given.
evidence in Hungarian
Khiron’s technique was first used in 2021 on patients at a small clinic in Budapest called Mama Klinika. After the mammogram is completed, two radiologists review it for signs of cancer. The AI then either agrees with the doctors or marks areas to re-examine.
Across five MaMMa Klinika sites in Hungary, 22 cases have been documented since 2021 in which AI identified cancers missed by radiologists, with approximately 40 more under review.
“It is a great success,” said Dr. Andrus Vadzei, director of MAMA Klinika, who was introduced to cucumbers through Dr. Karpati, Mr. Kxkemethy’s mother. “If this procedure saves a life or two, it will be worth it.”
Khiron said the technology works best with doctors, not in lieu of them. Scotland’s National Health Service will use it as an additional reader for mammography scans at six sites, and it will be in around 30 breast cancer screening sites run by England’s National Health Service by the end of the year. Oulu University Hospital in Finland also plans to use the technology, and a bus will travel around Oman this year to screen for breast cancer using AI
“An AI-plus-doctor should replace a stand-alone doctor, but an AI should not replace a doctor,” Mr. Kekskmethi said.
The National Cancer Institute estimates that about 20 percent of breast cancers are missed during screening mammograms.
Constance Lehman, professor of radiology at Harvard Medical School and chief of breast imaging and radiology at Massachusetts General Hospital, urged doctors to keep an open mind.
“We’re not irrelevant,” she said, “but there are tasks that are better done with computers.”
At the Bács-Kiskun County Hospital outside Budapest, Dr Ambroze said he was initially skeptical of the technique – but was soon won over. He x-rayed a 58-year-old woman who had a small tumor seen by AI that Dr. Ambroze was finding difficult to see.
The AI saw something, he said, that “seemed to appear out of nowhere.”