– he is very funny. I am very pleased that this indicates that the study was well designed and timely. It has received significant attention and may therefore form the basis for a future recommendation on artificial intelligence in mammography at the international level. “I am happy that we contributed to the development,” says Christina Lange, a breast radiologist and cancer researcher at Lund University.
Double review standard
In Sweden, women between the ages of 40 and 74 are regularly screened with mammography to detect breast cancer. It's called screening.
According to European standards, today every X-ray from mammography examinations is reviewed by two radiologists. Double review makes evaluation safe and reduces the risk of missing a cancer case. In this study, Christina Lange and her colleagues investigated whether it was possible to replace a doctor with an artificial intelligence machine.
– The goal was to reduce the workload of radiologists and find more relevant cancer cases early.
Discover more cases of cancer
100,000 women who participated in mammography screening in South West Scania were included in the study. They were divided into two groups. In one group, radiographs were reviewed by two radiologists according to the usual standardized process. In the second group, the images were reviewed by a doctor and an artificial intelligence. But if the AI detects a high cancer risk, the images are also reviewed by a second radiologist. Doctors' screening was also supported by artificial intelligence that identified suspicious findings in the image.
– The results exceed our expectations. We were able to reduce the work of radiologists by 44 percent and still detected 20 percent more cancers than we could with standard duplex review.
The results exceed our expectations.
The study received great attention in large parts of the world and in many scientific journals. This study is one of the first to address the importance of artificial intelligence in healthcare on a large scale. The results show that artificial intelligence can free up time for radiologists so that they can devote themselves to more complex tasks closer to the patient. But most importantly, the study indicates that artificial intelligence in screening can detect more cancer cases in a safe way.
Christina Lange also hopes that the study will have significance for patients by detecting more serious cancers at an early stage with the help of artificial intelligence.
– This research is entirely dependent on support from the Cancer Foundation and those who donate money to cancer research. I am very grateful for that!
Join and support cancer research.
Give a vital gift
More research is underway
The research project is now continuing with further analyses, and in 2024 researchers will follow up on what type of cancer cases are detected with the help of AI and whether AI can lead to fewer cases of so-called interval cancers that occur between two occasions of presentation.
Publication: Raising the level of artificial intelligence in cancer screening O'Leary K. Nat Med. 2023 Dec;29(12):2972. doi: 10.1038/s41591-023-02701-0.
“Extreme tv maven. Beer fanatic. Friendly bacon fan. Communicator. Wannabe travel expert.”