Using artificial intelligence in breast cancer screening increases the chances of detecting the disease, researchers have found, in what they say is the first real-world test of the approach.
AI helps radiologists detect breast cancer earlier, shows a new study PHOTO Archive
Numerous studies have suggested that artificial intelligence (AI) could help medical professionals detect cancer, either by identifying abnormal growths in CT scans or signs of breast cancer in mammograms, The Guardian reports.
However, many studies are retrospective – meaning that AI is not involved from the start – while prospective studies often have small sample sizes. Important and larger studies do not necessarily reflect use in real conditions.
Now researchers say they have tested AI for the first time in a national screening program, demonstrating that it provides benefits in a real-world setting.
How Artificial Intelligence is used
Prof. Alexander Katalinic, co-author of the study from the University of Lübeck, Germany, said: “We were able to improve the detection rate without increasing the risk for women participating in breast cancer screening”adding that the approach could also reduce the workload of radiologists.
Katalinic and colleagues analyzed data from 461,818 women in Germany who participated in breast cancer screening between July 2021 and February 2023 as part of a national program targeting asymptomatic women aged 50 to 69.
All women had their scans independently reviewed by two radiologists. However, for 260,739 of the women, at least one of the experts used an AI tool for support.
The AI tool visibly labels the scans it considers “normal” and issues a type alert “safety net” when a scan they rate as suspicious is deemed non-suspicious by the radiologist. In such cases, the AI also highlights the area of the scan that should be examined.
Thousands of cases detected with the help of AI
In total, 2,881 women included in the study, published in the journal Nature Medicine, were diagnosed with breast cancer. The detection rate was 6.7% higher in the AI-assisted group. After adjusting for factors such as the women’s age and the radiologists involved, the researchers found a 17.6 percent increase in the detection rate in the AI group: 6.70 per 1,000 women, compared with 5.70 per 1,000 in the standard group. In other words, the use of AI led to the detection of one additional case of cancer for every 1,000 women examined.
The team pointed out that the rate of women recalled for further investigation following a suspicious scan was about the same.
“In our study, we achieved a higher detection rate without having a higher false positive rate”said Katalinic. “It’s a better outcome with the same level of risk.”
The team stated that the type alert “safety net” was triggered 3,959 times in the AI-assisted group, resulting in 204 breast cancer diagnoses. In contrast, 20 cases of cancer in this group would have been missed if doctors had not reviewed the scans labeled as “normal” by AI.
Fewer false positives
Stefan Bunk, co-author of the study and co-founder of Vara, which developed the AI tool, said the technology has increased the speed at which radiologists review scans marked as “normal”.
He also added that even if these scans had not been reviewed by experts, the overall breast cancer detection rate would have been higher and the recall rate lower than without the tool. This means fewer false positives for women and a reduction in workload for radiologists.
Dr. Katharine Halliday, president of the Royal College of Radiologists, said the organisation’s latest report showed a 29% shortage of radiologists in the NHS.
“Any tools that can increase our accuracy and productivity are welcome. But while the potential benefits are significant, so are the potential risks”she said. “It is vital that the implementation of AI in the NHS is done carefully, under expert supervision.”
