GEANT4 calculation of normalized glandular dose coefficients in computed tomography dedicated to the breast

Abstract

Introduction Normalized glandular dose coefficients (gN CT )for computed tomography dedicated to the breast with a polychromatic X-ray source are available for estimating the mean glandular dose in a breast CT scan. However, a number of assumptions and simplifications are introduced in their derivation using Monte Carlo (MC) codes, both on breast geometry and on photon tracking and energy deposition details. Purpose To investigate, via MC simulations, four different factors which affect DgN CT calculation: (1) skin thickness, (2) electron tracking energy cutoff, (3) source position and (4) glandular tissue distribution. Materials and methods The MC code was developed using GEANT4. The breast was simulated as a cylindrical homogenous mixture of fat and glandular tissue. Skin thickness was varied from 1.45 to 4 mm. Monochromatic (8–80 keV) DgN CT have been calculated and the polychromatic DgN CT were computed by weighting them for typical X-ray tube spectra adopted for breast CT. A heterogeneous breast model as a mixture of adipose and randomly disposed glandular tissue voxels was investigated. Results Default GEANT4 electron energy cutoff leads to a DgN CT overestimation up to 1.6% at 60 keV. The skin thickness affects DgN CT values by about 1% (30–80 keV). DgN CT values calculated for a heterogeneous breast differ by 2% with respect to a homogeneous tissue mixture approximation. The X-ray source position also affects the DgN CT determinations, and setup-specific dose coefficients should be adopted. Conclusion Some factors which influence DgN CT MC calculations have been investigated and new polychromatic DgN CT coefficients have been computed in breast CT. Disclosure Nothing to declare.