Abstract ID: 92 Skin model and its impact on Mean Glandular Dose in digital mammography

The Mean Glandular Dose (MGD) is the main physical quantity used in mammography dosimetry [1] . The skin plays an important role, absorbing part of the incident radiation and, its traditional thickness is overestimated with recent findings [1] . This work investigates the influence of breast skin models on the MGD using the PENELOPE (2014) + penEasy (2015) Monte Carlo code. The simulated geometry includes: a source 66 cm above the detector, a compression and a breast support paddle (PMMA 2 mm thick), the breast (a semi-cylinder with 8 cm radius), and the a-Se detector. The total thickness of the breast varied from 2 to 9 cm and its central region is composed by a homogeneous mixture of glandular and adipose tissue with different proportions [1] . Four breast skin models were implemented: 4 mm skin tissue (I), 5 mm adipose tissue (II), a combination of 1,45 mm skin plus 2 mm adipose (III) and 1,45 mm skin (IV) [2] . Simulations were made with monoenergetic beams from 8 to 60 keV and polyenergetic beams with different Anode/Filter combinations (Mo/Mo, Mo/Rh, Rh/Rh, W/Rh, W/Ag, W/Al) and a potential from 22 kV to 35 kV. For a 2 cm breast, a maximum MGD difference of 80% between the skins models (I)–(II) at 8 keV, decreasing as energy increases. For the polyenergetic spectra (Mo/Mo), a mean difference of 28%, 16% and 16% in comparing the models (II), (III) and (IV), with the model (I), respectively. The MGD differences between the models decreases while increasing tube potential. The composition and thickness of the breast has an impact up to 10% and 15% on the results, respectively. The results highlight the importance of the modelling the skin for accurate dosimetry in mammography.