AI study: Revolutionary early detection of breast cancer in Germany!

A current study by the University of Lübeck and the Schleswig-Holstein University Hospital (UKSH) shows that the use of artificial intelligence (AI) can significantly improve the early detection of breast cancer. The results of this study were published in the journal Nature Medicine . During the investigation, data from over 460,000 women were evaluated between 2021 and 2023 in the mammography screening program (MSP) in Germany.

The study shows that AI can increase the discovery rate of breast cancer by almost 18 % without increasing the rate of false alarms or unnecessary additional examinations. About half of the mammographies were evaluated with AI, while the other half was traditionally examined by radiologists. It was found that KI per 1,000 women discovered 6.7 breast cancer cases, compared to 5.7 cases in the control group. These results indicate that additional breast cancer can be discovered by AI.

expectations and effects of the study

Alexander Katalinic, director of the Institute for Social Medicine and Epidemiology and Head of the Praim study, expressed optimistic about the improved forecasts for women with breast cancer. In Schleswig-Holstein in particular, statutory health-insured women between the ages of 50 and 75 have the opportunity to take a free X-ray examination of the breast every two years. This is particularly important because breast cancer is the most common cancer in women in Germany, with 78,000 new cases annually.

The use of AI could not only close diagnostic gaps, but also significantly reduce the workload for radiologists. Every year, radiologists evaluate around 24 million individual images as part of the screening program, which underlines the relevance of new technologies to support the specialists. According to Stefan Bunk, CTO of Vara, the use of AI offers great potential to improve global screening programs, and the number of re-orders could be reduced by up to 15 %, while recognizing unobtrusive AI accusations could be significantly increased.