Inside a dark room at the Doctor-Kiskun County Hospital outside Budapest, Dr Éva Ambrózay, a radiologist with more than two decades of experience, stares at a computer screen showing mammograms of the patient.
Two radiologists previously said the X-ray did not show any signs that the patient had breast cancer. But Dr Ambrózay is taking a close look at some of the scan areas circled in red where the artificial intelligence software has flagged as potentially carcinogenic.
“This is something,” she said. She quickly asked the woman to be called back for a biopsy, which would take place within the next week.
Advances in AI are beginning to yield breakthroughs in breast cancer screening by detecting signs that doctors miss. So far, this technology is showing an impressive ability to cancer in situ at least as well as human radiologistsAccording to early results and radiologists, this is one of the clearest indications to date of how AI can improve public health.
Hungary, which has a strong breast cancer screening program, is one of the largest trial sites for the technology in real patients. At five hospitals and clinics that perform more than 35,000 visits a year, AI systems have been deployed starting in 2021 and are now helping to check for signs of cancer that radiologists may have missed. Clinics and hospitals in the US, UK and European Union are also starting to test or provide data to help develop the system.
The use of AI is growing as the technology has become at the heart of the boom of Silicon Valleywith the release chatbots like ChatGPT shows how superior AI is to communicating in human-like prose — sometimes with worrisome results. Building on a similar form used by chatbots modeled after the human brain, breast cancer screening technology demonstrates other ways AI is pervading everyday life.
Doctors and AI developers say widespread use of cancer detection technology still faces many obstacles. Additional clinical trials are needed before these systems can be adopted more widely as an automated second or third breast cancer screen reader, beyond a limited number of sites. currently using this technology. The tool must also show that it can produce accurate results for women of all ages, ethnicities and body types. And the technology must prove it can recognize more complex forms of breast cancer and cut down on non-cancerous false positives, radiologists say.
AI tools also remind A dicussion about whether they will replace radiologists, with makers of the technology facing regulatory scrutiny and opposition from some doctors and medical organizations. economic. For now, those concerns appear to have been overblown, with many experts arguing that the technology will only be effective and trusted by patients if it is used in partnership with trained doctors.
And in the end, AI could be a lifesaver, Dr. László Tabár, a leading mammography educator in Europe, said he was captivated by the technology after reviewing its performance in screening. breast cancer from certain providers.
“I’m dreaming of the day when women come to a breast cancer center and they ask, ‘Do you have AI?’,” he said.
Hundreds of images every day
In 2016, Geoff Hintonone of the world’s leading AI researchers, argues that the technology will eclipse the skills of radiologists within five years.
“I think if you were a radiologist, you would be like Wile E. Coyote in the cartoon,” he says. told The New Yorker in 2017. “You’ve crossed the edge of the abyss, but you haven’t looked down yet. There is no ground below.”
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. Technology is at the heart of their system — called a neural network — modeled after the way the human brain processes information from different sources. It’s what’s used to identify people and animals in images posted to apps like Google Photos and allows Siri and Alexa to recognize the words people say. Neural networks also promote new wave of chatbots like ChatGPT.
Many AI propagandists believe that such technology can easily be applied to disease detection, such as breast cancer in mammography. According to the World Health Organization, in 2020, there will be 2.3 million breast cancer diagnoses and 685,000 deaths from the disease.
But not everyone feels replacing radiologists will be as easy as Mr Hinton predicted. Peter Kecskmethy, a computer scientist who co-founded Kheiron Medical Technologies, a software company that develops AI tools to help radiologists detect early signs of cancer, knows that Reality will be more complicated.
Mr. Kecskemethy grew up in Hungary and worked at one of Budapest’s largest hospitals. His mother is a radiologist, which gave him firsthand insight into the difficulty of finding a small melanoma in an image. Radiologists often spend hours each day in a dark room looking at hundreds of images and making life-changing decisions for patients.
Dr Edith Karpati, Kecskemethy’s mother, now director of medical products at Kheiron, said: “It’s easy to miss small lesions. “Can not concentrate.”
Kecskmethy, along with Kheiron co-founder Tobias Rijken, an expert in machine learning, said AI should aid doctors. To train their AI system, they collected more than 5 million historical mammograms of diagnosed patients, provided by clinics in Hungary and Argentina, as well as academic institutions. arts, such as Emory University. The London-based company also paid 12 radiologists to label images with special software that teaches the AI to detect cancerous tumor growth by shape, density, location and other features. its other elements.
From the millions of cases for which the system is powered, the technology generates a mathematical representation of normal mammograms and those with cancer. With the ability to look at each image in more detail than the human eye, it compares that baseline to find abnormalities in each mammogram.
Last year, after testing on more than 275,000 breast cancer cases, Kheiron report that its AI software matches the performance of human radiologists as second readers of mammogram scans. It also cuts the workload of radiologists by at least 30% because it reduces the number of X-rays they need to read. In other results from a clinic in Hungary last year, the technology increased cancer detection rates by 13% because more malignancies were identified.
Dr Tabár, whose mammogram reading techniques are commonly used by radiologists, tried out the software in 2021 by tracing some of the most challenging cases of his career , in which radiologists missed signs of a developing cancer. In any case, the AI detected it.
Dr Tabár said: “I was amazed at how good it was. He said he did not have any financial ties to Kheiron when he first tested the technology and has since received a consulting fee for feedback to improve the system. He said systems he has tested from other AI companies, including South Korea’s Lunit Insight and Germany’s Vara, have also yielded encouraging detection results.
Evidence in Hungary
Kheiron’s technology was first used on patients in 2021 at a small clinic in Budapest called MaMMa Klinika. After the mammogram is done, the two radiologists look for signs of cancer. The AI will then agree with the doctors or mark the areas for re-examination.
Across five MaMa Klinika sites in Hungary, 22 cases have been recorded since 2021 in which the AI identified cancers that radiologists missed, with around 40 more cases being viewed. review.
“It was a big breakthrough,” said Dr András Vadászy, director of MaMMa Klinika, who was introduced to Kheiron through Dr. Karpati, Kecskemethy’s mother. “If this process will save a life or two, it will be worth it.”
Kheiron says the technology works best with doctors, not for them. Scotland’s National Health Service will use it as a supplemental reading of mammograms at six sites and it will be available at around 30 breast cancer screening sites run by the Health Service The UK country is operational by the end of this year. Oulu University Hospital in Finland also plans to use the technology, and a bus will travel around Oman this year to perform breast cancer screening using AI.
“A doctor plus AI should replace doctors alone, but AI should not replace doctors,” Mr. Kecskemethy said.
The National Cancer Institute has estimate that about 20 percent of breast cancers are missed during mammograms.
Constance Lehman, professor of radiology at Harvard Medical School and chief of the department of breast imaging and radiology at Massachusetts General Hospital, urges doctors to keep an open mind.
“We are not related,” she said, “but there are tasks that are better done with computers.”
At the Doctor-Kiskun County Hospital outside Budapest, Dr Ambrózay said she was initially skeptical about the technology – but was quickly convinced. She took an X-ray of a 58-year-old woman with a small AI-detected tumor that Dr Ambrózay could hardly see.
She said that the AI saw something that “seemed to appear out of nowhere.”
