IN A GROUNDBREAKING study published this month, researchers from the University of Lübeck, Germany, have demonstrated the effectiveness of AI-supported mammography screening in real-world, nationwide settings. Their findings indicate that integrating AI into mammography screening processes leads to increased breast cancer detection rates without significantly raising recall rates.
The study, led by Dr Eisemann and colleagues, evaluated AI-supported double reading against standard double reading of mammograms. The PRAIM (PRospective multicentre observational study of an integrated AI system with live Monitoring) study involved 463,094 women aged 50 to 69, screened across 12 sites in Germany between 2021 and 2023. Among these participants, 260,739 were screened with AI assistance.
AI-supported double reading detected more cancers (6.7 per 1,000) compared with standard double reading (5.7 per 1,000). Recall rates were slightly lower in the AI group (37.4 per 1,000 versus 38.3 per 1,000), and positive predictive values (PPVs) for both recall and biopsy were higher. These results underline the potential of AI to enhance diagnostic accuracy and improve outcomes in breast cancer screening.
“Our findings substantially add to the growing body of evidence suggesting that AI-supported mammography screening is feasible and safe and can reduce workload,” the researchers noted.
While retrospective studies have highlighted AI’s promise in radiology, prospective studies like PRAIM are essential for validating its real-world effectiveness. The study also demonstrated that radiologists using AI spent less time interpreting exams flagged as normal by the AI system. A post hoc analysis suggested that this could lead to a 56.7% reduction in workload, alongside a 15% decrease in recall rates and a 16.7% improvement in cancer detection rates.
The authors emphasised the need for future research to examine the long-term impacts of AI-supported screening, including interval cancer rates and the distribution of cancer stages at diagnosis. They also called for urgent integration of AI systems into mammography screening guidelines to realise their full potential.
This study marks a significant step toward enhancing breast cancer screening programmes with AI, potentially improving early detection and reducing strain on radiology departments.
Reference
Eisemann N et al. Nationwide real-world implementation of AI for cancer detection in population-based mammography screening. Nat Med. 2025;DOI:10.1038/s41591-024-03408-6.
