AI can be used safely in breast cancer screening – Lancet study

Artificial intelligence can be safely used to improve the accuracy of screening of mammography results, according to a new study published in The Lancet Oncology.

More than 80,000 women between the ages of 40 and 80, who were eligible for mammography screening were randomly allocated to AI-supported screening, the intervention group, or standard double reading without AI, the control group.

The randomised, controlled, population-based trial was conducted at four mammography screening sites in Sweden and funded by the Swedish Cancer Society, Confederation of Regional Cancer Centres, and the Swedish governmental funding for clinical research.

The research found that mammography screening supported by AI resulted in a similar cancer detection rate compared with standard double reading, while resulting in a “substantially lower screen-reading workload”. The false positive rate in both groups was just 1.5%, the study found.

The study’s authors concluded that the use of AI in mammography screening is safe.

The authors said that although there have been retrospective studies of the use of AI to improve breast cancer screening, they were not aware of a previous randomised trial.

Radiology and imaging are among the leading areas where AI is already being used in healthcare, with the technology touted as a potential way of reducing the workload for radiographers and other clinicians.

AI is already being piloted to help improve breast cancer screening in the UK. Last month, Digital Health reported that Leeds Teaching Hospitals NHS Trust is trialling new AI software from Canadian company Densitas that can evaluate the quality of mammograms and spot any positioning errors.

Earlier this year Leeds launched the first-of-its-kind LIBRA (Leeds Investigation of Breast screening AI) study. That study trialled an AI reader to work alongside two human readers to analyse mammograms for signs of cancer.