The update of the national Guidelines on 23 December 2025 marks a turning point for cancer prevention in Italy. For the first time, the national panel, adopting the GRADE methodology, issued a conditional recommendation in favor of using an AI-based triage strategy for reading mammograms within breast cancer screening programs.

Question 3: a strategic innovation

The most significant change concerns so-called Question 3: the use of AI systems to decide, case by case, whether a mammogram should undergo single or double human reading, as an alternative to systematic double reading. This triage strategy makes it possible to optimize the allocation of clinical resources without compromising safety.

The evidence: clinical and economic

The Italian panel’s recommendation is grounded in high-quality independent studies:

• MASAI (Mammography Screening with Artificial Intelligence trial) – A randomized clinical trial conducted in Sweden and published in The Lancet Digital Health, showing an increase in cancer detection compared to standard double reading, together with a reduction in radiologists’ workload.
• Lauritzen et al., 2024 – An observational study conducted in Denmark and published in Radiology, reporting an increase in cancer detection and a reduction in workload, alongside a decrease in false positives compared to traditional reading.

From an economic perspective, independent analyses (consistent with international estimates) suggest that AI-based triage strategies can be cost-saving, delivering net savings while improving the overall efficiency of screening pathways.

BreastNegative’s contribution

Aligned with this approach, BreastNegative is Health Triage’s technology designed to automatically identify mammograms with a high probability of being negative. BreastNegative is currently being evaluated through a retrospective study and is intended to support radiologists by optimizing workflow and strengthening the quality of care.

As stated in the Guidelines, the radiologist remains the sole final decision-maker: AI does not replace clinical expertise, it supports it. The recommended approach signals a paradigm shift—not automating diagnosis, but making screening more efficient, sustainable, and safe, with tangible benefits for healthcare systems and for the women who participate in prevention programs.