ENERGY DEPENDENT SYSTEMATIC ERRORS IN DUAL-ENERGY X-RAY COMPUTED TOMOGRAPHY

Abstract

The energy-dependent systematic errors in the image of attenuation coefficient distribution acquired from dual-energy x-ray computed tomography (DECT) have been studied in this M.Sc. project. In the study, attenuation coefficient images were generated via computer simulation of a DECT system and compared with the exact attenuation coefficient distributions. The systematic errors in the DECT image were analysed. The errors were especially significant when a high atomic number element such as iodine in a contrast agent was present in the region of interest. The results derived from the analysis agreed with a simple mathematical model formulated to analytically quantity the errors in terms of an equivalent DECT system using two monoenergetic spectra. This shed new light on the underlying basis of the errors. With a better understanding of the problem, a robust correction method was formulated to reduce the errors by numerically transforming the basis coefficients into those of a more desirable basis materials set. This method is known as basis materials coefficient transformation (BMT). By incorporating the correction method into the usual DECT protocol, the errors were found to be reduced significantly. This has important clinical value when accurate quantitative measurement is required. One potential application of the DECT(BMT) technique lies in the attenuation compensation in accurate quantitative SPECT imaging. An investigation was therefore carried out to assess the accuracy of the radionuclide density distribution image in a SPECT system which employed DECT(BMT) for attenuation compensation in the presence of a contrast agent (containing high atomic number element such as iodine) located in the body. The results were further compared with those obtained from SPECT systems which used DECT(BMD) or the conventional x-ray CT instead to acquire the attenuation map for attenuation correction. It was concluded that of the three DECT(BMT) was the most promising method for use in attenuation compensation in SPECT imaging.