AI supports earlier breast cancer detection in Dutch screening programme

AI can detect more and earlier-stage tumours in the Dutch breast cancer screening programme. Researchers led by Radboudumc analysed over 42,000 mammograms from the Utrecht region, comparing results from two radiologists with results from one radiologist supported by AI. The findings have been published in The Lancet Digital Health.

The AI-supported method detected more clinically relevant tumours, and often at an earlier stage. In some cases, AI flagged a potential tumour that radiologists did not recognise until a later scan confirmed its presence. “AI turned out to be right earlier,” said PhD candidate Suzanne van Winkel. These were often larger, invasive tumours that benefit from early treatment.

In Sweden, AI is already used to replace the second radiologist in breast cancer screening. The Dutch study suggests a similar approach is feasible in the Netherlands, potentially reducing healthcare costs by millions while easing radiologist workload. However, national implementation faces logistical and IT infrastructure challenges.

Read the full publication in The Lancet Digital Health: AI as an independent second reader in detection of clinically relevant breast cancers within a population-based screening programme in the Netherlands