At least 20,000 Australian women are diagnosed with breast cancer each year, and more than 3,300 die from the disease. Early detection remains the most powerful intervention, yet Australia's screening program has fundamentally relied on the same technology for decades. A new study published in The Lancet Digital Health suggests artificial intelligence could change that.
Developed using mammograms from nearly 400,000 women screened at BreastScreen Victoria and then tested on data from almost 96,000 Australian women, the results were then confirmed in an independent Swedish population of over 4,500 women.The AI-based risk score, called BRAIx, estimated breast cancer risk more accurately than the factors doctors have traditionally relied on, such as breast density and family history.
The findings are substantial.Among women ranked in the top 2 per cent of risk by BRAIx, nearly one in 10 were diagnosed with breast cancer within four years despite being given the all-clear. This matters becauseabout 25% of women with cancer are diagnosed between the biennial screens, and these 'interval cancers' are often aggressive and, unfortunately, more likely to be fatal.
Here's where the pragmatism comes in.The BRAIx project demonstrated that when used independently, AI produced strong results but didn't outperform the standard of care delivered through the multi-human reader system currently used in Australia, with detailed simulations highlighting AI as an effective tool for supporting human readers to improve accuracy and efficiency. The research is clear: this is not about replacing radiologists.When used in conjunction with human readers, AI did help improve cancer detection and reduce unnecessary recalls.
The sceptical observer rightly asks about implementation.BRAIx is now being trialled as part of the BreastScreen Victoria program to help read mammograms, and other states are already using and evaluating different AI tools for reading mammograms. This patchwork approach has a downside.Experts suggest it may be time for Australia to conduct a national, independent review of these new tools, which as part of a more risk-adjusted approach to breast screening, could save lives.
The centre-right case for AI is straightforward: it offers fiscal efficiency and improved outcomes with limited implementation burden. Yet the evidence requires caution.While this study found BRAIx more accurately predicts breast cancer risk compared to assessments based on breast density, the breast density data was collected using a different tool to those used by the Breastscreen program, so this finding should be interpreted carefully. Real-world validation matters before scaling.
While more studies are needed before it is considered for use in routine care, the BRAIx team is on track to achieve its goal of reducing the number of deaths from breast cancer to zero. Australia's strength lies in its evidence-based approach to health policy. The research case for AI in screening is growing. What's needed now is rigorous, transparent assessment before national expansion. Speed and caution can coexist.