Investigating the Average Glandular Dose and Exposure Parameters in Mammography Based on Compressed Breast Thickness...

The mammary gland is highly sensitive to ionising radiation, making the average glandular dose (AGD) the most appropriate metric for breast dosimetry. With increasing participation in breast screening programmes, cumulative radiation exposure remains a concern. This study aimed to investigate the correlation between AGD and exposure parameters under controlled conditions of compressed breast thicknesses and mammographic projections.

A retrospective, cross-sectional analysis evaluated data from 609 patients who underwent clinically justified digital mammography examinations with a single direct digital radiography system at King Abdulaziz Medical City, Jeddah, Saudi Arabia, between September 2023 and September 2025. The Shapiro-Wilk normality test indicated that AGD data were not normally distributed. Consequently, the Wilcoxon signed-rank test was used to assess differences between craniocaudal (CC) and mediolateral oblique (MLO) projections for the right and left breasts across CBT ranges. Spearman's rank correlation coefficient was used to evaluate associations between AGD and exposure parameters (compression force (CF), mAs, and kVp) under controlled conditions of CBT ranges and projections.

AGD increased with CBT, rising from 1.36-1.39 mGy at 30-39 mm to 2.25-3.05 mGy at 70-79 mm. MLO projections consistently showed higher AGD than CC projections, with greater differences from 50-59 mm. Statistically significant differences were observed in projections across CBT. Spearman's analysis demonstrated a significant positive correlation between AGD and CBT (p < 0.0001), which was strongest in the right MLO (ρ = 0.5082). Within the 50-59 mm range, AGD strongly correlated with mAs and moderately with kVp, but not with CF.

AGD increases significantly with CBT, particularly in MLO projections. Within the 50-59 mm CBT range, mAs showed the strongest influence on AGD, while kVp had a moderate effect and CF was non-significant. These findings support dose optimisation through exposure parameter control and alignment with DRLs while maintaining image quality.