Image quality simulations in X-ray mammography

Abstract

The application of tunable monochromatic X-rays to mammography can improve the image quality and reduce the radiation dose to the patient. A new generation of quasi-monochromatic, high-flux X-ray sources is currently under development, based on Thomson scattering of photons produced by a laser on a highly focused electron beam. In this paper we investigate the effect of different spectral distributions on the image contrast on 2 cm, 4 cm, 6 cm, 8 cm breast mammographic phantoms containing details of different composition (tumor mass and calcification details) and thickness. In the simulations we assume a homogeneous breast composition and three compositions (25%; 50% and 75% glandular) are investigated. We simulate monochromatic, quasi-monochromatic and polychromatic X-ray sources in order to define the optimal energies that maximize the image quality parameter. In particular we compare the simulations of the Gaussian quasi-monochromatic spectra. having different energy standard deviations (1, 2, 3, 4, 5 keV) with traditional X-ray tubes spectra. Results show that Gaussian spectra having an energy spread of about 1 keV standard deviation call improve the image quality of about 30% with respect to traditional X-ray tube ones.