Nationwide Screening Program Shows Ai Can Spot Breast Cancer With High Accuracy

A groundbreaking study published in Nature Medicine has found that the use of artificial intelligence (AI) in breast cancer screenings significantly increases the detection rate of the disease, while also reducing the workload of radiologists. The study, conducted by researchers from the University of Lübeck in Germany, analysed data from 461,818 women who underwent breast cancer screening between July 2021 and February 2023 as part of a national programme targeting asymptomatic women aged 50-69.

The AI tool used in the study was developed by Vara, a company co-founded by one of the researchers. The tool not only visually labels scans deemed unsuspicious as “normal” but also issues a “safety net” alert when a scan rated suspicious by radiologists is deemed normal. This alert highlights the area of the scan that merits further scrutiny.

The study found that the detection rate was 6.7% higher in the AI group, with one additional case of cancer spotted per 1,000 women screened when AI was used. The researchers also found that the rate at which women were recalled for further investigation as a result of a suspicious scan was approximately the same in both groups.

Professor Alexander Katalinic, lead author of the study, stated that the results are “better than expected” and demonstrate that AI can improve breast cancer detection rates without increasing harm to patients. The tool’s “safety net” feature also helped to reduce the number of missed diagnoses by radiologists who had not reviewed scans deemed normal.

The study’s findings have significant implications for the NHS, where there is a current shortfall of radiologists. The Royal College of Radiologists has estimated that there is a 29% shortage of radiologists in the NHS, and any tool that can boost accuracy and productivity is welcome.

However, Dr Katharine Halliday, president of the Royal College of Radiologists, cautioned that the deployment of AI into the NHS must be done carefully, with expert oversight. “While the potential benefits are significant, so are the potential risks,” she said.

The study’s results also highlight the need for long-term follow-up to fully understand the clinical implications of integrating AI into mammography screening. Dr Kristina Lång of Lund University noted that while the findings are encouraging, more research is needed to ensure that AI can detect clinically relevant cancers at an early stage, where early detection can meaningfully improve patient outcomes.

Overall, the study provides strong evidence for the potential benefits of using AI in breast cancer screenings and highlights the need for further research to optimize this approach.