Evaluation of the Image Contrast with a New Depth-Dependent Collimator Resolution Iterative Reconstruction Method

Abstract

Iterative reconstruction methods are well suited to improve image quality and clinical performance in SPECT imaging in nuclear medicine. This is done by incorporating internal modeling of the imaging physics in SPECT reconstructions to correct for the major factors affecting image quality. These factors include degradation of spatial resolution with increasing distance of the source from the collimator, Compton scatter and photon attenuation. Since noise suppression approaches degrade the image resolution, a good balance of resolution recovery and noise suppression is desirable. The Astonish software package provides powerful control of the resolution recovery technique and an application-specific optimization of noise suppression. Objectives: The aim of the study is to compare the contrast value of the images resulting from filtering with different available iteration methods with the new depth dependant collimator resolution iteration technique (Astonish) in SPECT imaging. Materials and methods: The study was performed on a Forte dual head gamma camera, using the SPECT phantom. The contrast value is calculated for a specified hot sphere within the cold background by applying the studied reconstruction iterative methods (MLEM, 3D-OSEM and Astonish) with iteration number from 1-15 iterations. The resultant calculated contrast values are then compared for each iteration method. Results: Regarding the MLEM iteration method the average contrast value was [0.8002 ± 0.0304] and for the 3D OSEM iteration method the average contrast value was [0.8662 ± 0.0068]. As regards to the Astonish iteration method the average contrast value was [0.907 ± 0.0394]. There was a statistical significant difference in contrast value between the studied iteration methods (p=4.2E-12). Conclusion:The MLEM iteration method increases the sphere count with increasing the iteration number; and also it increases the background counts with about 3 fold that of the 3D OSEM method, and about 4 folds that of Astonish method. Although both 3D-OSEM and Astonish methods give sphere counts that are almost within the same range but incorporating the depth dependant resolution recovery parameter in the Astonish method leads to decrease in the background count with increasing the iteration number, which in turn enhances the contrast of the resultant image from Astonish iteration method rather than that of both MLEM and 3D-OSEM methods.