Influence de la densité et du numéro atomique moyen du milieu dans la correction d’atténuation par TDM en TEP : une étude sur fantômes

Abstract

Aim The aim of this work is to study the influence of medium density on the CT or external source attenuation corrected images, by simulation on a phantom, with various positron emission tomographs. Material and method A series of experiments on a cylindrical phantom filled with water marked with [ 18F]-FDG, containing six vials filled per pair with mediums of different densities or solutions of KI, CaCl2 and saccharose with various densities, was carried out under comparable conditions on three different tomographs. In only one of the vials of each pair, an identical radioactivity of [ 18F]-FDG was added, three to five fold the surrounding activity. The reconstructions and attenuation corrections suggested by the manufacturers, were carried out under the usual conditions of each site. The activity of each structure was estimated by the methods of profiles and regions of interest, on the non attenuation corrected images (NAC), the images corrected by CT (CTAC), and/or external source (GPAC). Results With all three tomographs, the activities estimated on the NAC images present an inverse correlation to the medium density (important absorption by dense material). On CTAC images, we observed with only two of the three tomographs, an overestimation of the activity in the “radioactive” vials, depending on the medium mean Z number and density (over correction), and a artefactual “activity” in the denser “cold” vial (incorrect attenuation correction. The dense saccharose solutions, with non elevated Z number, do not affect the CT attenuation correction.