Design and use of computational models for contrast-enhanced spectral mammography

Abstract

INTRODUCTION: Contrast-enhanced spectral mammography (CESM) is a new technique for cancer investigations. The application of this technique to screen the breast would result in contrast improvement of the different breast lesions. This can be effectively ensured by the use of phantoms. Computational phantoms are one of the ways to investigate the characteristics of the received image and thus to evaluate the whole technique. AIM: The aim of this study is to validate and analyse the design of three different computational phantoms for CESM. MATERIALS AND METHODS: An in-house software tool was used to create three computational phantoms, consisting of iodinated inserts for the simulation of the CESM procedure. The inserts in the phantoms, modelled from Omnipaque, have a radius of 8-10 mm and varying height. The three phantoms are made of polymethyl methacrylate (PMMA), with different shape and composition. One of the phantoms is characterised by a heterogeneous background. For each phantom two x-ray radiographs were generated, one at x-ray energy of 20keV and one at 34 keV. The images were processed to obtain a recombined iodine image, which shows the iodine contrast agent and suppresses the surrounding background tissue. RESULTS: Simulated spectral images demonstrated a great improvement of the image quality compared to low-energy images of the phantoms. The simulations with the inhomogeneous model revealed that the heterogeneous background has been successively depressed while improving the visibility of the iodine inserts.CONCLUSION: The heterogeneous breast phantom might be used as a reference tool for information about the needed iodine concentration which needs to be inserted during the procedure to obtain significant enhancement in the suspicious area.