Purpose: We have previous proposed a volume-of-interest expansion approach to include the spill-out region for estimation of the total organ activity for quantitative SPECT/CT images. We have found that a fixed volume-of-interest expansion size can almost capture the full spill-out activity for VOI (spheres) of different sizes. The aim of this study is to study the potential effect of different factors that can affect system resolution, which may affect this expansion size.
Methods: The clinical quantitative SPECT/CT reconstruction usually includes OSEM with CT attenuation correction, scatter correction, and resolution recovery. With a given clinical SPECT/CT, the factors that can affect the system resolution are pixel size and reconstruction algorithms (mainly resolution recovery). We acquired three sets of SPECT/CT data using Jaszczak SPECT phantom: one data set of concurrent data acquisition with Philips Brightview which include both 128x128 and 256x256 data acquisition. We then acquired two regular 128x128 SPECT/CT data using Philips Brightview and GE Discovery 670 SPECT/CT system. For the concurrent data set, the total counts for the 128x128 and 256x256 projection data sets are the same. We reconstructed the SPECT images into 128x128 and 256x256 image arrays. For the other two phantom data sets, we reconstructed images using OSEM with and without resolution recovery.
Results: We evaluated the visualization of cold spheres and rod segments of different sizes in the reconstruction image. There is no significant difference for the SPECT reconstructions of different pixel sizes. The images reconstructed with resolution recovery appears smoother than those reconstructed without resolution recovery. But the effects on the visualization of the cold spheres and number of rod segments that can be distinguished are very minor.
Conclusion: Our preliminary results suggest that the fixed expansion size for quantitative SPECT/CT can be robust for different quantitative reconstruction packages.
Not Applicable / None Entered.