Sci-Thurs AM: YIS-01: Monte Carlo Simulation of a First Generation Scatter Enhanced CT System for in Vivo Breast Imaging

Abstract

Introduction: Breast CT is an imaging modality capable of reconstructing 3D images of the linear attenuation coefficient without breast compression. Significant levels of scatter can lead to image artifacts and contrast reduction, but may also be used to reconstruct electron density () images. Monte Carlo (MC) simulation was used to evaluate our ability to reconstruct images when all types of scattered radiation are present. Methods: We have designed a first generation scanner and developed an algorithm capable of reconstructing images out of measurements of single Compton scatter. The EGSnrc MC code was validated against analytical calculations of the number of single Compton scattered photons. Scatter produced in the breast for a complete CT scan was simulated and used to reconstruct using an iterative FBP algorithm. Results: The analytical calculation had a mean error of 1.1% with respect to the MC distribution whose uncertainty was of 3.2%. On average multiple scatter (MS) contributed 43% to the total scatter while Rayleigh scatter contributed 81% for angles less than 8°. The reconstructed image shows an increase in values towards the centre of the breast due to the increase of the relative contribution of MS. We introduce a post-reconstruction correction method that results in relative images of with errors of ±3% for most breast materials. Conclusions: Our results show the feasibility to reconstruct in vivo images of the breast with a prototype system using the same incident fluence as cone beam breast CT.