AI-Supported Mammography Screening Matches Radiologists in Spotting Cancer – Promising Results in Largest Study
Artificial intelligence (AI)-supported mammography screenings have shown promising results in a large-scale study, matching the accuracy of radiologists in detecting breast cancer. The study, published in the Lancet Oncology, involved over 80,000 women in Sweden and demonstrated that AI could potentially reduce the screening workload by almost 50%. With an existing shortage of radiologists, the National Health Service (NHS) is exploring the use of machine learning to assist in checking mammograms.
Traditionally, mammograms are double-read by two experts in the NHS. However, due to a significant shortage of radiologists, the healthcare service is seeking innovative solutions to address the issue. The recently conducted randomized control trial directly compared AI-supported screening with standard care. The study found that computer-aided detection using AI technology could identify breast cancer at a similar rate to two radiologists. Additionally, the use of AI did not lead to more false-positive diagnoses.
During the trial, half of the mammograms were assessed by two radiologists (standard care), while the other half were evaluated using AI-supported screening before interpretation by at least one radiologist. The results showed that there were 244 cancer cases detected by AI-supported screening, compared to 203 cases recalled from standard screening. Moreover, radiologists performed 36,886 fewer screen readings in the AI-supported group, reducing their workload by 44%.
The lead author of the study, Dr. Kristina Lang from Lund University in Sweden, mentioned the importance of understanding the impact of combining AI with radiologists’ expertise in detecting interval cancers often missed by traditional screening methods. The NHS spokesperson expressed optimism about the potential of AI in breast screening and how it could contribute to quicker image analysis, earlier cancer detection, and ultimately, more lives saved.
Dr. Katharine Halliday, president of the Royal College of Radiologists, highlighted the benefits of AI in maximizing efficiency and supporting decision-making for clinicians. She also noted the complexity of mammogram interpretation and the shortage of radiologists in the UK, making AI integration challenging. Professor Fiona Gilbert, professor of radiology and head of the department at the University of Cambridge, emphasized the significant potential manpower savings that AI could provide, helping address workforce issues in the UK.
The findings of this study are expected to facilitate the testing and implementation of AI into the UK national breast screening program. However, it is crucial to maintain a balanced view and consider different perspectives and opinions. Continued research and exploration are necessary to fully understand the implications on patient outcomes, including the role of AI in detecting interval cancers. Nonetheless, the promising results demonstrate the potential of AI-supported mammography screenings and their ability to alleviate the burden on radiologists while maintaining accuracy and efficiency.
In conclusion, AI-supported mammography screening has shown promising results in a large-scale study, matching the accuracy of two radiologists in identifying breast cancer. The use of AI has the potential to significantly reduce the workload of radiologists and contribute to faster image analysis and early cancer detection. While further research is needed, the integration of AI into breast screening programs could have a transformative impact, ultimately saving more lives.
