Investigating the relationship between breast cancer risk factors and an AI-generated mammographic texture feature in the Nurses’ Health Study II

The mammogram risk score (MRS), an AI-driven mammographic texture feature, strongly predicts breast cancer risk independently of breast density, though underlying mechanisms remain unclear. Using data from the Nurses' Health Study II (292 cases, 561 controls), we validated MRS's association with breast cancer and evaluated its relationships with established breast cancer risk factors through observational analyses, polygenic score analyses, and Mendelian randomization. MRS was significantly associated with breast cancer risk before (OR=1.92 per SD increase; 95% CI:1.57 to 2.35; 10-year AUC=0.69) and after adjustment for predicted BI-RADS density (OR=1.85; 95% CI:1.49 to 2.30). Early life body size and adult body mass index (BMI) were inversely associated with MRS, while benign breast disease history and predicted BI-RADS density showed positive associations; after adjusting for density, associations between MRS and the other three risk factors were attenuated. Polygenic score analyses and Mendelian randomization consistently demonstrated significant positive associations between genetic predictors of breast density measures (dense area, percent density, predicted BI-RADS density) and MRS. After adjusting for predicted BI-RADS density and BMI, genetic predictors of higher waist-to-hip ratio were significantly associated with increased MRS. Our findings reveal robust associations between breast density measures and MRS and suggest a potential impact of central obesity on MRS. Future larger-scale validation studies are needed.