Inside a dark room at the Bács-Kiskun County Hospital on the outskirts of Budapest, Dr. Éva Ambrózay, a radiologist with more than two decades of experience, stared at a computer monitor displaying a patient’s mammogram.
Two radiologists had previously said that the x-ray did not show any signs that the patient had breast cancer. But Dr. Ambrózay was taking a close look at several areas of the scan circled in red, which the artificial intelligence software had flagged as potentially cancerous.
“This is something,” he said. She soon ordered the woman to be called in for a biopsy, which will take place within the next week.
Advances in AI are beginning to lead to advances in breast cancer detection by spotting the 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 early results and radiologists, in what is one of the most tangible signs to date of how AI can improve. public health.
Hungary, which has a robust breast cancer screening program, is one of the largest testing grounds for the technology on real patients. At five hospitals and clinics that perform more than 35,000 screenings a year, artificial intelligence systems have been deployed starting in 2021 and are now helping to detect signs of cancer that a radiologist may have missed. Clinics and hospitals in the United States, Great Britain and the European Union are also beginning to test or provide data to help develop the systems.
The use of AI is growing as the technology has become the center of a Silicon Valley boom, with the launch of chatbots like ChatGPT showing how AI has a remarkable ability to communicate in human-like prose, to Sometimes with worrying results. Building on a similar form used by chatbots that is based on the human brain, the breast cancer detection technology shows other ways that AI is seeping into everyday life.
Widespread use of cancer detection technology still faces many hurdles, doctors and AI developers said. Additional clinical trials are needed before the systems can be more widely adopted as a second or third automated breast cancer screen reader, beyond the limited number of places now using the technology. The tool must also demonstrate that it can produce accurate results in women of all ages, ethnicities, and body types. And the technology must show that it can recognize more complex forms of breast cancer and reduce false positives that are not cancerous, the radiologists said.
AI tools have also sparked 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 seem overblown, with many experts saying the technology will be effective and trusted by patients only if used in collaboration with trained doctors.
And ultimately, AI could save lives, said Dr. László Tabár, a leading European mammography educator, who said he was won over by the technology after reviewing its performance in detecting breast cancer.
“I dream of the day when women go to a breast cancer center and ask, ‘Do you have AI or not?’” she said.
Hundreds of images a day
In 2016, Geoff Hinton, one of the world’s leading AI researchers, argued that the technology would eclipse the skills of a radiologist in five years.
“I think if you work as a radiologist, you’re like Wile E. Coyote in the cartoon,” he said. told The New Yorker in 2017. “You are already on the edge of the precipice, but you have not yet looked down. There is no floor below.
Mr. Hinton and two of his students at the University of Toronto built an image recognition system that could accurately identify common objects like flowers, dogs, and cars. The technology at the heart of his system, called a neural network, is based on how 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 it allows Siri and Alexa to recognize words people say. Neural networks have also fueled the new wave of chatbots like ChatGPT.
Many AI evangelists believed that such technology could easily be applied to detect diseases and illnesses, such as breast cancer on a mammogram. In 2020, there were 2.3 million breast cancer diagnoses and 685,000 deaths from the disease, according to the World Health Organization.
But not everyone felt that replacing radiologists would be as easy as Mr. Hinton predicted. Peter Kecskemethy, a computer scientist who co-founded Kheiron Medical Technologies, a software company that develops AI tools to help radiologists detect early signs of cancer, knew the reality would be more complicated.
Mr. Kecskemethy grew up in Hungary and spent time in one of the largest hospitals in Budapest. His mother was a radiologist, which gave him firsthand insight into the difficulties of finding a small malignancy within an image. Radiologists often spend hours every day in a dark room looking at hundreds of images and making life-changing decisions for patients.
“It’s very easy to miss tiny lesions,” said Dr. Edith Karpati, Mr. Kecskemethy’s mother, who is now director of medical products at Kheiron. “You can’t stay focused.”
Mr Kecskemethy, along with Kheiron co-founder Tobias Rijken, an expert in machine learning, said AI should help doctors. To train their AI systems, they collected more than 5 million historical mammograms from patients whose diagnoses were already known, provided by clinics in Hungary and Argentina, as well as academic institutions, such as Emory University. The London-based company also pays 12 radiologists to label the images using special software that teaches the AI to detect a cancerous growth based on its shape, density, location and other factors.
From the millions of cases fed into the system, the technology creates a mathematical representation of normal mammograms and those with cancer. With the ability to look at each image in a more granular way than the human eye, it then compares that baseline to find abnormalities on each mammogram.
Last year, after testing more than 275,000 breast cancer cases, Kheiron reported that its AI software matched the performance of human radiologists in acting as a second mammography reader. It also reduced the workload for radiologists by at least 30 percent because it reduced the number of x-rays they needed to read. In other results from a Hungarian clinic last year, the technology increased the cancer detection rate by 13 percent because more malignancies were identified.
Dr. Tabár, whose techniques for reading a mammogram are commonly used by radiologists, tested the software in 2021 by retrieving several of the most challenging cases of his career in which radiologists failed to detect signs of a developing cancer. In all cases, the AI detected it.
“I was really surprised at how good it was,” said Dr. Tabár. He said he had no financial connection to Kheiron and that other AI companies, including South Korea’s Lunit Insight and Germany’s Vara, have also seen encouraging detection results.
test in hungary
Kheiron’s technology was first used on patients in 2021 at a small clinic in Budapest called the MaMMa Klinika. Once a mammogram is complete, two radiologists review it for signs of cancer. The AI then agrees with the doctors or marks areas to check again.
Across five MaMMa Klinika sites in Hungary, 22 cases have been documented since 2021 in which AI identified a cancer missed by radiologists, with around 40 more under review.
“It’s a breakthrough,” said Dr. András Vadász, director of the MaMMa Klinika, who met Kheiron through Dr. Karpati, Mr. Kecskemethy’s mother. “If this process will save a life or two, it will be worth it.”
Kheiron said the technology worked best with doctors, not instead of them. It will be used by the National Health Service for Scotland as an additional mammography reader at six sites, and it will be at around 30 breast cancer screening sites operated 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 across Oman this year to conduct breast cancer screenings using AI.
“An AI plus doctor should replace the doctor alone, but an AI should not replace the doctor,” Mr. Kecskemethy said.
The National Cancer Institute has My dear that about 20 percent of breast cancers are missed during screening mammograms.
Constance Lehman, a 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 are not irrelevant,” he said, “but there are tasks that are better done with computers.”
At the Bács-Kiskun County Hospital on the outskirts of Budapest, Dr. Ambrózay said she had initially been skeptical of the technology, but was quickly convinced. She X-rayed a 58-year-old woman with a small AI-detected tumor that Dr. Ambrózay had a hard time seeing.
The AI saw something, he said, “that seemed to appear out of nowhere.”
