77. Normalized glandular dose coefficients in mammography, digital breast tomosynthesis and breast dedicated CT

Abstract

Purpose To compute, via GEANT4 based Monte Carlo (MC) simulations, normalized glandular dose (DgN) coefficients for mean glandular dose (MGD) estimates in 2D and 3D X-ray breast imaging. Methods We developed a Monte Carlo code for DgN calculation (MGD/air kerma, mGy/mGy) in mammography, digital breast tomosynthesis (DBT) and breast CT. The code is based on GEANT4 toolkit ver. 10.00 with the low-energy physics list Option4, and it was validated vs. data from the AAPM-TG195 report and vs. measurements. The geometry of the available scanners were replicated and the breast was modelled as a cylinder with a semi-circular cross section (in mammography and DBT), and as a pendant cylinder in breast CT. The skin thickness was 1.45 mm. The MC simulations were run for monoenergetic beams in the energy range usually adopted in clinic and research. The monoenergetic DgN curves were fitted with 8th-order polynomial curves, and then the polyenergetic DgN coefficients were calculated. Results We obtained a full set of monoenergetic DgN coefficients for mammography, DBT and breast CT. The monoenergetic DgN fitting curves showed a fitting R2 higher than 0.9998. The polyenergetic DgN coefficients were computed for a large cohort of spectra used in the routine clinical practice. The data were comparable with those present in literature. Conclusions A validated MC code for DgN estimates in 2D and 3D X-ray breast imaging has been presented. A breast model has been defined. Monoenergetic and polyenergetic DgN coefficients have been computed. A complete dataset containing monoenergetic DgN curve fit coefficients and polyenergetic DgN coefficients have been released. DgN differed, on average, 6% and 13%, in mammography and breast CT, respectively, from literature data in Refs. [1] , [2] .