P036] Comparison of the image quality of virtual monochromatic images generated by two different algorithms

Abstract

Purpose One of the major applications of Dual-Energy Computed Tomography (DECT) technology is virtual monochromatic imaging, which are representations of an object as if the source of a CT used single energy photons (keV). The advantages with virtual monochromatic imaging is that the user can manipulate image quality by freely selecting the monochromatic energy. The aim of this work was to study the image quality of virtual monochromatic images generated by two different algorithms, one of which includes a noise-reduction feature (Mono and Mono+). Methods A modified commercial image quality phantom was scanned using a dual-source CT scanner (80/140 Sn kVp). Virtual monochromatic images were generated over a range of 40–170 keV in 10 keV increments using both algorithms. Uniformity, noise, the CT number of water, the CT number of iodine, the contrast-to-noise ratio (CNR) and linearity were evaluated in both image stacks using open-source image processing software. Results In general, the Mono+ images resulted in better image quality. A significant noise reduction was found at low keV using this technique, with a maximum reduction of 50% at 40 keV, compared to the Mono images. The Mono+ images also provided an increased iodine CNR at lower keVs, moving the optimum CNR from 60 keV to 40 keV. Conclusions The Mono+ algorithm for creation of virtual monoenergetic images was shown to be superior to the Mono algorithm, especially for low keVs.