Evaluation of transmission source collimation for SPECT attenuation compensation

We designed and constructed an axial collimation system for a fixed, 40 cm long Gd-153 line source used in an offset fan beam (FB) transmission computed tomography (TCT) imaging system to evaluate the effect of axial slat spacing on TCT image quality and its impact on non-uniform attenuation compensation in SPECT. The axial spacing of customized, thin tungsten slat collimators can be varied in 1.00 mm increments. Using low-scatter and high-scatter rod phantoms in axial and sagittal orientations, we evaluated the effects of different axial slat spacings. OSEM reconstructed TCT image resolution primarily degraded sagittally (along the camera axis) up to 40%. There was nominal contrast degradation transaxially, but up to 60% sagittally. Dose rate measured in a lung and on the surface of an anthropomorphic phantom increased linearly with coarser spacing and differed by an order of magnitude in going from 1 to 10 mm spacing. Measured nonuniform attenuation coefficients for all slat spacings in the lung, spine and myocardium of a phantom were constant, with a slight improvement in standard deviation of the mean value due to the improved statistics that come with having higher transmission fluxes. All slat spaced non-uniform TCT maps yielded identical and improved compensated SPECT images of a myocardial defect. In these studies, while finer slat spacing results in overall better transmission image quality, various axial slat spacings had no effect on attenuation compensation in SPECT. Moreover, fewer axial slats for FB TCT provide a lower cost line source shielding fixture, lighter weight burden for the line source holder, and potential for extending the source lifetime by regularly increasing slat spacing to modulate transmission flux as the source decays.