On the relative performance of edge illumination x-ray phase-contrast CT and conventional, attenuation-based CT.

Abstract

This article is aimed at comparing edge illumination (EI) x-ray phase contrast computed tomography (PCT) and conventional (attenuation-based) computed tomography (CT), based on their respective contrast and noise transfer. The noise in raw projections obtained with EI PCT is propagated through every step of the data processing, including phase retrieval and tomographic reconstruction, leading to a description of the noise in the reconstructed phase tomograms. This is compared to the noise in corresponding attenuation tomograms obtained with CT. Specifically, a formula is derived that allows evaluating the relative performance of both modalities on the basis of their contrast-to-noise ratio (CNR), for a variety of experimental parameters. The noise power spectra of phase tomograms are shifted towards lower spatial frequencies, leading to a fundamentally different noise texture. The relative performance of EI PCT and CT, in terms of their CNR, is linked to spatial resolution: the CNR in phase tomograms is generally superior to that in attenuation tomograms for higher spatial resolutions (tens to hundreds of μm), but inferior for lower spatial resolutions (hundreds of μm to mm). These results imply that EI PCT could outperform CT in applications for which high spatial resolutions are key, e.g., small animal or specimen imaging.