Comparative validation of the BOADICEA and Tyrer-Cuzick breast cancer risk models incorporating classical risk factors and polygenic risk in a population-based prospective cohort

Purpose The Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA) and the Tyrer-Cuzick breast cancer models have recently been extended to include polygenic risk scores (PRS). In addition, BOADICEA has also been extended to include reproductive and lifestyle factors, which were already part of Tyrer-Cuzick model. We conducted a comparative validation of the extended models including a recently developed 313-variant PRS in a population-based prospective cohort. Methods We used data from a nested case-control sample of 1,337 women of European ancestry (619 incident breast cancer cases) aged 23-75 years from the Generations Study. Models were evaluated for calibration of five-year absolute risk and risk discrimination. Results The extended BOADICEA model with risk factors and PRS was well calibrated across risk deciles: expected-to-observed ratio (E/O) at the highest risk decile = 0.97 (95% Cl = 0.51 to 1.86) for women younger than 50 years and 1.09 (0.66 to 1.80) for women 50 years or older. Adding risk factors and PRS to the BOADICEA model improved discrimination modestly in younger women (Area Under the Curve (AUC): 69.7% vs. 69.1%) and substantially in older women (AUC: 64.6% vs. 56.8%). The Tyrer-Cuzick model with PRS had similar discrimination as the extended BOADICEA model for both age groups; but showed evidence of overestimation at the highest risk decile: E/O=1.54 (0.81 to 2.92) for younger and 1.73 (1.03 to 2.90) for older women. Conclusion The extended BOADICEA model identified women in a European ancestry population at elevated breast cancer risk more accurately than the Tyrer-Cuzick model with PRS. These analyses can inform choice of risk models for risk stratified breast cancer prevention for women of European ancestry.