Range of Normalized Glandular Dose for Mammography Using Patient-Specific Glandular Fractions

Breast cancer is the most common cancer among women, and mammography remains the primary modality for early detection. Because mammography uses ionizing radiation, accurate estimation of normalized glandular dose (DgN) is important for risk assessment. Recent breast dosimetry models, including TG-282, incorporate population-based glandular tissue distributions; however, patient-specific glandular distributions remain unknown from conventional mammographic projections. In previous work (Smith, Dey et al., 2025), glandular fraction (GF) maps were estimated from a single mammographic projection. While these maps determine glandular path length along each projection ray, they do not uniquely define glandular tissue depth. In this work, we propose a framework for estimating a patient-specific range of DgN from a projection-derived GF map. Using simulated data, glandular tissue was distributed to the top, center, and bottom of the breast volume using Siddon ray-tracing. These configurations preserved the GF map while producing maximum, intermediate, and minimum DgN values. Monte Carlo simulations were performed, and DgN was normalized to entrance air kerma. DgN varied by up to a factor of three solely due to differences in glandular tissue depth, despite identical GF maps and visually indistinguishable projection images. Using randomized realizations derived from TG-282 glandular distributions for Cranio-Caudal (CC) and Medio-Lateral Oblique (MLO) views, dose ratios were calculated relative to central glandular placement. Central placement overestimated DgN by less than 5 and 15 percent on average for MLO and CC distributions respectively, whereas centroid-based placement underestimated dose by up to 25 percent. These results indicate that patient-specific bounds on DgN can be estimated from limited mammographic information and that central placement provides a conservative dose estimate.