Abstract
1756
Objectives: The quality of single-photon emission computed tomography is heavily dependent on the collimator design, which is always a trade-off between resolution and sensitivity. In addition to the collimator performance, final image quality is also affected by the reconstruction method. Statistical reconstruction algorithms can incorporate resolution recovery into the reconstruction procedure, which has potential to improve the resolution-sensitivity trade-off. The objective of this study was to investigate the effect of resolution recovery on optimal collimator design in cardiac SPECT. Methods: Seven parallel hole collimators designs with different hole length, hole radius and septal thickness combinations were used to generate projection data sets of a realistic mathematical thorax phantom. The projection data sets were reconstructed using ordered subsets expectation maximization algorithm with and without resolution recovery. The performance of the collimators was evaluated by comparing the myocardial recovery coefficient and contrast between myocardium and ventricle. Results: The results showed that, without resolution recovery, high-resolution collimator is the optimal collimator for cardiac SPECT, whereas with resolution recovery collimator with approximately two times poorer resolution (~four times higher sensitivity) than the high-resolution collimator gave the best contrast and recovery coefficient values. Conclusions: These findings demonstrate that resolution recovery has large impact on collimator design and collimators with relatively high sensitivity should be preferred in cardiac SPECT if resolution recovery is applied in reconstruction.
Research Support (if any): Supported by a grant from Japan Society for the Promotion of Science.
- Society of Nuclear Medicine, Inc.