INVESTIGATION OF THE USE OF X-RAY CT IMAGES FOR ATTENUATION COMPENSATION IN SPECT

This study investigates the general use of single-beam Xray computed tomography (CT) images for generating attenuation maps for compensation of photon attenuation in SPECT images. A 3D mathematical thorax phantom is used to simulate both emission and transmission projection data for monoenergetic (radionuclide) and polyenergetic (X-ray) sources, Polyenergetic transmission projection data are simulated for a standard X-ray spectrum and fan-beam geometry. The projection data are reconstructed using filtered backprojection to form an X-ray CT image which is then scaled to produce an estimate of the attenuation map at the energy of the emission radionuclide. Emission projection data are simulated for a fan-beam geometry at the energies of Tl-201 and Tc-99m, two radionuclides commonly used in cardiac SPECT. Detector response and scatter are not included in the model. Noiseless, emission projection data are iteratively reconstructed using the ML-EM algorithm with nonuniform attenuation compensation and attenuation maps derived from both the simulated X-ray CT image and from a simulated monoenergetic transmission CT image. The attenuation maps generated from the X-ray CT images accurately estimate the attenuation coefficient for muscle and lung tissues, but not for bone tissues, which show error in the attenuation coefficient of 21-42% for spinal bone and 34-58% for rib bone. However, despite the inaccurate estimate of bone attenuation, the reconstructed SPECT images provide estimates of myocardial radioactivity concentration to within 9% and show few artifacts.