In a dark room at Birch Kiskun County Hospital outside Budapest, Dr. Eva Ambroze, a radiologist with more than 20 years of experience, was staring at a computer monitor displaying a patient’s mammogram.
Two radiologists previously said the X-rays showed no signs the patient had breast cancer. But Dr. Ambrózay took a closer look at some areas of the scan circled in red that the artificial intelligence software flagged as possible cancer.
“This is something,” she said. She immediately ordered the woman recalled for her biopsy. A biopsy will be done next week.
Advances in AI are beginning to bring breakthroughs in breast cancer screening by detecting signs that doctors miss. So far, one of the most tangible signs of how AI can improve public health, according to his early results and radiologists, is that the technology is at least as good as human radiologists. It has shown an impressive ability to detect cancer as well.
Hungary, with a robust breast cancer screening program, is one of the largest testing grounds for this technology in real patients. Starting in 2021, his five hospitals and clinics, which perform more than 35,000 screenings a year, will be using AI systems to check for signs of cancer that radiologists may have missed. . Clinics and hospitals in the US, UK and European Union have also begun testing or providing data to help develop the system.
The use of AI is on the rise as technology is at the heart of the Silicon Valley boom. The release of chatbots like ChatGPT shows that AI has an amazing ability to communicate in human-like prose. Breast cancer screening technology built from similar forms used in chatbots modeled after the human brain shows another way AI is permeating everyday life.
There are still many hurdles to widespread cancer-detection technology, according to doctors and AI developers. Beyond the limited number of locations currently using the technology, additional clinical trials are needed before the system is more widely adopted as a second or third leader in automated breast cancer screening. The tool should also demonstrate that it can produce accurate results for women of all ages, ethnicities, and body types. The technology must also prove that it can recognize more complex forms of breast cancer and reduce false positives that are not cancerous, the radiologist said.
AI tools have also sparked debate about whether they will replace human radiologists, with the technology’s makers facing regulatory scrutiny and resistance from some doctors and medical institutions. , these concerns appear exaggerated, with many experts saying the technique is effective and trusted by patients only when used in collaboration with a trained physician. increase.
Ultimately, AI could save lives, said Dr. László Tabár, a leading European mammography educator, who was fascinated by the technology after reviewing its performance in breast cancer screening from several vendors. said.
“I dream of the day when women go to breast cancer centers and are asked, ‘Do you have AI,'” he said.
Hundreds of images per day
In 2016, Geoff Hinton, one of the world’s leading AI researchers, argued that the technology will surpass the skills of radiologists within five years.
“If you work as a radiologist, I think you’re like the Wile E. Coyote in the cartoon,” he said. told the New Yorker “You’re already over the edge, but you haven’t looked down yet. There’s no ground under it.”
Hinton and two students from the University of Toronto have built an image recognition system that can accurately identify common objects such as flowers, dogs, and cars. The technology at the heart of systems called neural networks models how the human brain processes information from various sources. It is used to identify people and animals in images posted to apps like Google Photos, and enables Siri and Alexa to recognize what people say. Neural networks have also driven a new wave of chatbots like ChatGPT.
Many AI evangelists believed that such technology could be easily applied to the detection of diseases and diseases, like breast cancer in mammograms. According to the World Health Organization, 2.3 million people were diagnosed with breast cancer in 2020 and 685,000 died from it.
But not everyone felt that changing radiologists would be as easy as Hinton predicted. Computer scientist Peter Kecskemethy, who co-founded Kheiron Medical Technologies, a software company developing his AI tools to help radiologists detect early signs of cancer, says the reality is more complicated. I knew it would be
Kecskemethy grew up in Hungary and spent time in one of Budapest’s largest hospitals. His mother was a radiologist, and she witnessed the difficulty of finding small malignancies on images. Radiologists often spend hours each day looking at hundreds of images in a darkened room to make life-changing decisions for their patients.
“It’s very easy to miss small lesions,” said Dr. Edith Karpati, Kecskemethy’s mother and now director of medical products at Kheiron. “It’s impossible to stay focused.”
Kecskemethy, along with machine learning expert Tobias Rijken, co-founder of Kheiron, said AI should help doctors. To train the AI system, he collected more than 5 million historical mammograms of patients with known diagnoses, provided by clinics in Hungary and Argentina and academic institutions such as Emory University. The company, located in London, uses special software that pays 12 radiologists to label images and teach AI to identify cancer growths by their shape, density, location and other factors. doing.
From the millions of cases fed into the system, this technology creates mathematical representations of normal and cancer mammograms. We can see each image in a way that is finer than the human eye can, so we compare that baseline to find abnormalities in each mammogram.
After testing more than 275,000 breast cancer cases last year, Kheiron report Its AI software matched the performance of human radiologists when acting as a second reader for mammography scans. It also reduced the radiologist’s workload by at least 30% because fewer of his X-rays needed to be read. Another result from a Hungarian clinic last year showed a 13% increase in cancer detection as the technology identified more malignancies.
Dr. Tavar, whose technology for reading mammograms is commonly used by radiologists, will try the software in 2021 in the most difficult case of his career when radiologists overlooked signs of cancer in progress. I got some of the In all cases the AI found it.
“I was pleasantly surprised by how good it was,” Dr. Taber said. He said he had no financial ties to Kheiron when he first tested the technology, and has since received advisory fees for feedback to improve the system. Systems he has tested from other his AI companies, such as South Korea’s Lunit Insight and Germany’s Vara, have also provided promising detection results, he said.
Proof in Hungary
Kheiron’s technology will be used on patients for the first time in 2021 at a small clinic in Budapest called MaMMa Klinika. After the mammogram is complete, two radiologists will check for signs of cancer. The AI then either agrees with the doctor’s opinion or flags the area to check again.
Five MaMMa Klinika sites in Hungary have documented 22 cases in which AI has identified cancers missed by radiologists since 2021, with about 40 more under investigation.
“This is a big breakthrough,” said Dr. András Vadászy, director of MaMMa Klinika, who was introduced to Kheiron through Kecskemethy’s mother, Dr. Karpati. “If this process saves one or two of her lives, it’s worth it.”
Kheiron says the technology works best with a doctor, not in place of him. Scotland’s National Health Service has used it as an additional reader for mammography scans at six sites, and it will be used at about 30 breast cancer screening sites operated by the UK’s National Health Service by the end of the year. Oulu University Hospital in Finland also plans to use the technology, and this year a bus will tour Oman to conduct AI-powered breast cancer screenings.
“AI plus doctors should only replace doctors, but AI should not replace doctors,” said Kecskemethy.
The National Cancer Institute Estimate About 20% of breast cancers are missed during mammogram screening.
Dr. Constance Lehman, professor of radiology at Harvard Medical School and mammography specialist at Massachusetts General Hospital, urged doctors to keep an open mind.
“We are not unrelated,” she said.
Speaking at the Batikiskun district hospital outside Budapest, Dr. Ambrozai said he was skeptical about the technology at first, but was quickly convinced. She pulled up her X-ray of her 58-year-old woman with a small tumor discovered by her AI that Dr. Ambrózay had trouble seeing.
The AI saw something, she said.
