Reduced inter-exposure time in dual-energy imaging: a technological and image processing cooperative approach

Abstract

Dual-energy imaging increases the possibility of pulmonary nodule detection by reducing the bone structure noise. Dual-shot techniques are limited by structural artefacts due to patient and natural movement during the switch of voltage between energies. A new acquisition approach for dual-energy imaging was envisioned in order to reduce this inter-exposure time. The idea is to keep the tube voltage constant, switch a filter in front of the patient and thus modulate the outgoing x-ray spectrum. The drawback of this method is a poorer spectral separation between low and high energy images leading to a higher sensitivity to noise. On the other hand, noise in the reconstructed image is mainly controlled by high-energy image noise, allowing the use of noise suppression algorithms without loosing high-frequency information present in the low-energy image. The first part of this paper is a simulation study presenting system optimisation that includes noise reduction in the HE image. Exposure times and filter thickness are chosen when optimising the signal difference to noise ratio (SDNR) and dose. Results show better SDNR (9 %) for similar dose than state-of-art dual-shot switching voltage technique. Thicker filters could lead to better results, but would demand more tube charge. In the second part is presented experimental validation and implemented noise suppression algorithm. As radiographs of anatomical phantoms are structured, anisotropic algorithm have been considered. Nodule and anthropomorphic phantoms were used to measure detail suppression after image processing. Results are shown in terms of noise suppression in the reconstructed image as well as in detail preservation.