Modeling SPECT acquisition and processing of changing radiopharmaceutical distributions

Abstract

The accuracy of SPECT images is compromised and artifacts may be produced when the radiopharmaceutical distribution changes during image acquisition. Optimization of SPECT acquisition protocols for changing tracer distributions can be difficult not only in patient studies (undesirability of performing repeat studies on the same patient) but also in phantom studies (difficulty of emulating the changing distributions). This study proposes a simulation that allows computer modeling of both tracer kinetics and different acquisition schemes. /sup 99m/Tc Teboroxime (Bracco Diagnostics) is used as a model. SPECT acquisition of a software phantom (NCAT, UNC Chapel Hill) is simulated with photon attenuation, collimator resolution, Compton scatter, Poisson noise, and changing tracer distribution. Short-axis uniformity is used to assess the severity of artifacts in the myocardium. The simulation produces similar artifacts to those found in patient studies with /sup 99m/Tc Teboroxime. This simulation methodology can provide a valuable tool for testing novel acquisition and processing techniques and to facilitate the optimization of SPECT images of changing tracer distributions. Summed fanning (back and forth) acquisitions have been tested and artifact reduced short-axis images obtained. Image restoration techniques are proposed to further improve the image quality. Furthermore, the simulated studies can be compared to the simulations with assigned low liver uptake and no tracer clearance from the myocardium to detect and resolve artifacts through variations in the acquisition and processing schemes.