K-Edge Imaging in Photon-Counting Spectral CT

Sammanfattning

The spectral capability of photon-counting computed tomography enablesK-edge imaging, that is, material decomposition with one or several highatomic number contrast agents with K-absorption-edges (K-edges) in thediagnostic X-ray energy range. This opens up for new simultaneous multi-contrast imaging applications. This thesis investigates the feasibility andlimitations of several dual-contrast imaging cases using different contrast-agent pairs, through simulations of a silicon-based photon-counting compu-ted tomography system performed using CatSim. Specifically, a maximum-likelihood three-material decomposition method based on an empirical ra-tional forward model was implemented, where water and the two contrastagents, investigated for the given case, were set as the basis materials. Themethod was then applied to simulations of slabs and analytical phantoms ofthe basis materials to evaluate the material decomposition performance. TheCNR in the resulting reconstructed material images was evaluated as themain figure of merit. The investigated cases in this thesis were: Iodine (K-edge at 33.2 keV) and gadolinium (50.2 keV), iodine and barium (37.4 keV), gadolinium and barium, iodine and xenon (34.6 keV), gadolinium and xenon,and iodine and krypton (14.3 keV). The material decomposition was suc-cessful with good performance for the cases with contrast agents that havewell-separated K-edges. The CNR in the material images was strongly affec-ted by the size of the simulated phantom and the imaging dose (mAs). Theresults for the iodine and krypton dual-contrast imaging case were especiallypromising for application in perfusion-ventilation CT to identify pulmonaryembolisms. The material decomposition performance for the cases with closeK-edges (iodine and barium or xenon) was much poorer, although with moreeffort, it might still be possible.