Abstract
2237
Objectives Fully automated software based algorithms for gating provide the possibility of gating in research and the clinic without changes to routine image acquisition procedures, and thus with minimal cost. In this study, we extend methods we previously presented in human PET to small animal PET (μPET), and report their implementation in large populations and with diverse radiotracers.
Methods 78 rat μPET scans ((18F)-FDG,-DMDPA,-NaF,-ML10, (11C)-DMDPA,-choline, (13N)-NH3) were acquired. Respiratory gating triggers were retrospectively determined and inserted into the listmode files. For each dataset, non-gated, software gated, and in some cases hardware gated images were reconstructed using vendor reconstruction. To assess the effectiveness of gating, total sum signal contained in respiratory frequencies was quantified as a measure that correlates with qualitative observations and distinguishes correctly from incorrectly gated data.
Results Static (non-gated) PET acquisitions were reprocessed to form gated images. The majority of images qualitatively exhibited structure motion for both software and hardware gated images. Summarizing the signal over the field of view in the gated scans, the mean/range of signal in the 1 cycle/period frequency as % of the non-varying DC component was 31/26-43 for the software gated images. Corresponding values for the randomly gated images were 26/26-27. Scans with non-thorax field of views showed appropriately little motion on gated images. Less motion was detected for scans having very low count scans (cases were gating is less appropriate).
Conclusions Our results indicate that fully automated gating algorithms can be extended from human PET to μPET, that they can be applied retrospectively to populations of data, and to scans acquired using diverse tracers. The ease of application of these gating methods supports their potential for use in research and the clinic.
Research Support USIEF Fulbrigh