Abstract ID: 176 Geant4 implementation of inter-atomic interference effect in small-angle coherent X-ray scattering for materials of medical interest

Abstract

Advanced applications of digital mammography such as dual-energy and tomosynthesis require multiple exposures and thus deliver higher dose compared to standard mammograms. A straightforward manner to reduce patient dose without affecting image quality would be removal of the anti-scatter grid, provided that the involved reconstruction algorithms are able to take the scatter figure into account [1]. Monte Carlo simulations are very well suited for the calculation of X-ray scatter distribution and can be used to integrate such information within the reconstruction software. Geant4 is an open source C++ particle tracking code widely used in several physical fields, including medical physics [2,3]. However, the coherent scattering cross section used by the standard Geant4 code does not take into account the influence of molecular interference. According to the independent atomic scattering approximation (the so-called free-atom model), coherent radiation is indistinguishable from primary radiation because its angular distribution is peaked in the forward direction. Since interference effects occur between x-rays scattered by neighbouring atoms in matter, it was shown experimentally that the scatter distribution is affected by the molecular structure of the target, even in amorphous materials. The most important consequence is that the coherent scatter distribution is not peaked in the forward direction, and the position of the maximum is strongly material-dependent [4]. In this contribution, we present the implementation of a method to take into account inter-atomic interference in small-angle coherent scattering in Geant4, including a dedicated data set of suitable molecular form factor values for several materials of clinical interest. Furthermore, we present scatter images of simple geometric phantoms in which the Rayleigh contribution is rigorously evaluated.