Could a virtual bowtie filter improve image quality? A breast simulation study

Abstract

ABSTRACT Introduction: During CT scanning, anatomical shapes and body contours cause variation in the transmitted x-ray intensity reaching the detector. This can compromise image quality since the detector has a limited dynamic range. In principle, suitable bowtie filters can reduce transmitted X-rays through thinner parts of the body so that it is uniform at the detector surface.Aim: This study explores the possibility of using mathematical data modelling of transmission data to act as a virtual bowtie filter. The virtual filter is applied after the acquisition of projection data and can be based on body shape. In this study, the virtual bowtie filter is constructed from breast contour information and applied to transmitted x-ray beam profiles produced in egs_cbct Monte Carlo simulations.Methods and materials: The Breastsimulator software was used to generate suitable breast phantoms. Egs_cbct simulated X-ray transport through these phantoms. One simulation included a bowtie filter that accounted for breast shape as a control. Projection images were acquired over 360 degrees without the bowtie filter. We designed a bowtie filter and corrected the transmitted X-ray profiles based on its shape.Results: Image contrast indicated any improvement in image quality. Results indicated that using a virtual bowtie correction method improved image quality as well as a reduction in cupping artefacts.Conclusion: It is a useful method to improve the image quality without the need to incorporate a real bowtie filter during measurements.