A comprehensive Cochrane review presented at the 2026 ASCO Annual Meeting suggests that commonly used breast cancer risk prediction models provide only modest accuracy in identifying which women with a family history of breast cancer will ultimately develop the disease (Abstract 10545). Although some models performed better than others, investigators concluded that none were sufficiently accurate to fully support personalized risk-based decision-making.
Women with a family history of breast cancer are frequently offered formal risk assessments to estimate their likelihood of developing the disease. These estimates help guide important clinical decisions, but uncertainty has remained regarding which models perform best in this higher-risk population.
To address this question, researchers from Trinity College Dublin, St James’s Hospital, and collaborating institutions conducted what they describe as the most comprehensive review to date of breast cancer risk models used in women with a family history of the disease. The investigators analyzed 45 studies evaluating the performance of risk prediction models and assessed how accurately they predicted future breast cancer risk.
Key Takeaways
Four models had sufficient published data to allow detailed evaluation: the Gail Breast Cancer Risk Assessment Tool (BCRAT), Tyrer-Cuzick (IBIS), BOADICEA, and BRCAPRO. Among these, the BOADICEA model demonstrated the most balanced overall performance. Both BOADICEA and the Gail model generated risk estimates that were generally close to the number of breast cancers observed in the analyzed studies. In contrast, the Tyrer-Cuzick model tended to overestimate risk, whereas BRCAPRO generally underestimated risk.
The review also assessed discriminatory accuracy—the ability of a model to distinguish between women who did and did not subsequently develop breast cancer. Although all four models demonstrated some predictive value, their discriminatory performance was only modest, and none approached the level of accuracy needed to reliably personalize care.
Lead author Sarah McGarrigle, MD, noted that the findings clarify both the strengths and limitations of current approaches to breast cancer risk prediction. “These tools are already widely used in clinical practice, and we now have a clearer picture of their accuracy in women with a family history of breast cancer,” she said. “Our findings suggest that these tools have value in supporting risk assessment and that is encouraging, but we still have a long way to go.”
The investigators emphasized that improving risk prediction remains an important goal, particularly for women at elevated familial risk. Senior author Elizabeth Connolly, MD, stated that accurate risk estimates are critical because they influence decisions about screening intensity and preventive interventions. She noted that while progress has been made, more work is needed to ensure that clinicians and patients can rely on the most precise information available when making these consequential decisions.
DISCLOSURE: For full disclosures of the study authors, visit coi.asco.org.
