A breakthrough study has shown that AI-supported mammography screenings are just as effective as having two radiologists analyze the scans. This discovery could potentially reduce the screening workload by almost half, making a significant impact on a field facing a shortage of radiologists. Currently, the National Health Service (NHS) in the UK requires double reading of mammograms by two experts. However, with around one-third of radiologist positions vacant, the NHS is exploring the use of machine learning to aid in scan analysis.
The randomized control trial, published in the Lancet Oncology journal, involved over 80,000 women from Sweden with an average age of 54. The researchers found that computer-aided detection, utilizing AI, was able to identify breast cancer in mammograms at a similar rate to two radiologists. Unlike previous studies that retrospectively assessed scans already analyzed by doctors, this interim study compared AI-supported screening directly with standard care.
The trial involved dividing the scans into two groups. Half were assessed by two radiologists, which represented the standard care group. The other half were analyzed by the AI-supported screening tool, followed by interpretation by at least one radiologist. The results revealed that the use of AI did not generate more false positives, where a scan is incorrectly identified as abnormal. The false-positive rate was consistent between the AI group and the group assessed by radiologists, at 1.5 percent.
The significant finding of the study was the reduction in screen-reading workload for radiologists. In the AI-supported group, there were 36,886 fewer readings by radiologists compared to the standard care group. This resulted in a remarkable 44 percent decrease in the screen-reading workload. Lead author Dr. Kristina Lang from Lund University in Sweden emphasized the need to further analyze the impact on patients’ outcomes, specifically in detecting interval cancers that are often missed by traditional screening methods when combining radiologists’ expertise with AI technology.
The NHS has expressed interest in exploring how AI can facilitate breast screening by enabling rapid and large-scale image analysis. If proven effective, it could expedite diagnosis, detect cancers at earlier stages, and ultimately save more lives. Dr. Katharine Halliday, president of the Royal College of Radiologists, highlighted the potential of AI in maximizing efficiency, supporting decision-making, and identifying urgent cases, thereby alleviating the strain caused by the current shortage of radiologists in the UK.
Professor Fiona Gilbert, professor of radiology and head of department at the University of Cambridge, acknowledged the substantial benefits in terms of manpower savings, which could address workforce issues prevalent in the UK. These findings are expected to contribute to the planning, testing, and implementation of AI in the national breast screening program.
While AI-supported mammography screenings have shown great promise, it is crucial to maintain journalistic integrity by presenting a balanced view of the topic. Researchers and medical professionals continue to emphasize the importance of understanding the implications on patient outcomes and addressing potential limitations. By combining the expertise of radiologists with AI advancements, the potential for enhanced breast cancer detection and improved patient care becomes even more apparent.
