KECSKEMÉT, Hungary - In a dark ward 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 the mammogram of a patient.
Two radiologists had previously said that the x-ray did not show any signs that the patient had breast cancer.
However, Ambrózay was carefully looking at various areas of the scanner marked in red, which the artificial intelligence software had flagged as
potentially cancerous
.
Dr. Éva Ambrózay, a radiologist at Bács-Kiskun County Hospital with more than two decades of experience, has been using Artificial Intelligence software to help look for signs of cancer that doctors might have missed.
Photo Akos Stiller for The New York Times
"This is something," he said.
He immediately ordered the woman to be called back for a biopsy, which will take place next week.
Advances in AI are starting to make a breakthrough in breast cancer screening by detecting the signs that doctors miss.
So far, the technology is demonstrating 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 strong breast cancer screening program, is one of the largest testing grounds for the technology on real patients.
In five hospitals and clinics that perform more than 35,000 checkups a year, AI systems went live starting in 2021 and now help detect signs of cancer that a radiologist might 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.
Possible abnormalities in a breast cancer screening highlighted by Artificial Intelligence software.
Photo Akos Stiller for The New York Times
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. , sometimes with worrying results.
Breast cancer detection technology, created from a similar way used by chatbots that is modeled on the human brain, shows other ways that AI is seeping into everyday life.
According to doctors and AI developers, the widespread use of this cancer detection technology still faces many obstacles.
More clinical trials are needed before the systems can be more widely adopted as a second or third automated reader for breast cancer screening tests, 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.
In addition, the technology must prove capable of recognizing more complex forms of breast cancer and reduce the number of
non-cancerous
false positives , according to the radiologists.
Peter Kecskemethy, founder of Kheiron Medical Technologies, and his mother, Dr. Edith Karpati, who was a radiologist, with X-ray data being fed into AI models.
Photo Akos Stiller for The New York Times
AI tools have also sparked debate over whether they will replace human radiologists, with the developers of the technology facing regulatory scrutiny and resistance from some doctors and healthcare institutions.
For now, those fears seem overblown, with many experts saying the technology will only be effective and trusted by patients if it is used in collaboration with
qualified doctors.
And ultimately, AI could save lives, says Dr. László Tabár, one of Europe's leading mammography experts, who says he was convinced by the technology after looking at the performance of various providers in breast cancer screening .
What artificial intelligence can do is "a breakthrough," says Dr. András Vadászy, director of the MaMMa Klinika.
Photo Akos Stiller for The New York Times
I dream of the day when women go to a breast cancer center and ask:
"Do they have AI or not?".
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 cartoons," he told The New Yorker in 2017.
"You're already on the edge of the cliff, but you haven't looked down yet. There's no ground below."
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 on which his system is based, called a neural network, is inspired by the way in which 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 what enables Siri and Alexa to recognize the words people say.
Neural networks have also fueled the new wave of chatbots, such as
ChatGPT.
Mr Kecskemethy, left, with Kheiron co-founder Tobias Rijken, said Artificial Intelligence should help doctors.
Photo Akos Stiller for The New York Times
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 believes that replacing radiologists will be as easy as 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.
Kecskemethy grew up in Hungary spending time in one of the largest hospitals in Budapest.
Her mother was a radiologist, which gave her first-hand experience of the difficulties of finding a small malignant tumor on an image.
Radiologists often spend hours a 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," says Dr. Edith Karpati, Kecskemethy's mother, who is now director of medical products at Kheiron.
"You can't stay
focused
."
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 with special software that teaches the AI to detect a cancerous tumor based on its shape, density, location and other factors.
From the millions of cases the system receives, the technology creates a mathematical representation of normal mammograms and those showing cancer.
With the ability to see each image in more detail than the human eye, it compares that baseline to detect abnormalities on every mammogram.
Last year, after testing more than 275,000 breast cancer cases, Kheiron reported that its AI software matched the performance of human radiologists when acting as second readers for mammograms.
In addition, it reduced the workload of radiologists by at least 30%, since it decreased the number of radiographs they had to read.
In other results obtained last year in a Hungarian clinic, the technology increased the cancer detection rate by
13%
because more malignancies were identified.
Tabár, whose techniques for reading a mammogram are commonly used by radiologists, tested the software in 2021, retrieving several of the most difficult cases of his career in which radiologists missed signs of developing cancer.
In all cases,
the AI detected it.
"I was surprised at how good it was," says Tabár.
He said he had no financial relationship with Kheiron when he first tested the technology and has since received consulting fees to improve the systems.
Systems it tested from other AI companies, including South Korea's Lunit Insight and Germany's Vara, have also yielded
encouraging detection results.
Trials in Hungary
Kheiron's technology was first used on patients in 2021 at a small Budapest clinic called the MaMMa Klinika.
After the mammogram is done, two radiologists review it for signs of cancer.
The AI then agrees with the doctors or points out areas that need to be checked again.
Across the five MaMMa Klinika centers in Hungary, 22 cases have been documented since 2021 in which AI identified a cancer missed by radiologists, and about 40 more are under review.
"It's a breakthrough," said Dr. András Vadászy, director of the MaMMa Klinika, who met Kheiron through Karpati, Kecskemethy's mother.
"If this process saves one or two lives, it will have been worth it."
Kheiron says that technology works best with doctors, not instead of them.
It will be used by the National Health Service for Scotland as an additional mammography reader in six centres, and by the end of the year it will be present in some 30 breast cancer screening centers run by the National Health Service in England.
Oulu University Hospital in Finland plans to use the technology as well, and this year a bus will tour Oman for AI breast cancer screening.
"An AI plus doctor should replace the doctor alone, but an AI should not replace the doctor," says Kecskemethy.
The National Cancer Institute estimates that approximately 20% of breast cancers go undetected during screening mammograms.
Constance Lehman, a professor of radiology at Harvard Medical School and chief of radiology and breast imaging 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, Ambrózay was skeptical at first, but was quickly convinced.
He showed him the X-ray of a 58-year-old woman with a small AI-detected tumor that Ambrózay had trouble seeing.
The AI saw something, he said, "that seemed to come out of nowhere."
c.2023 The New York Times Company
look too
To prevent cancer, more women should consider having their fallopian tubes removed
How some cancer cells hide
