Feasibility study of multipinhole collimators for high resolution small animal imaging

Multipinhole SPECT has the potential to overcome the inherent low efficiency of single pinhole gamma cameras typically used in small animal radionuclide imaging. By increasing the detector efficiency, multipinhole SPECT cameras allows to use reduced radionuclide dose and shorter scanning times than single pinhole gamma cameras. Reduced scanning times are advantageous for longitudinal studies where small animal models must be imaged several times. In addition, a spatial resolution of 1 mm or better is often required in drug delivery research and myocardial perfusion studies. In this article, we evaluate multipinhole collimator designs to obtain the best pinhole configuration that will increase the detector efficiency and maximize the use of a recently developed CZT with 128 times128 pixels. Additionally the multipinhole design minimize the overlap of the projection data to avoid artifacts in the tomographic reconstructions. For this purpose, we simulated the acquisition of projection data from point sources, microspheres phantoms, and resolution phantoms, by using a in house developed Monte Carlo program which accounts for the gamma photons interactions with the pinhole and detector crystal. The generated projection data was reconstructed using a standard ML-EM algorithm adapted for multipinhole radionuclide imaging. We determined the multipinhole collimator efficiency, resolution, and signal to noise ratio for several possible multipinhole configurations. A design of five pinhole arrays allowed high efficiency imaging (3.93 fold) while achieving a spatial resolution of 1 mm or better. Reconstruction of hot rod phantom data showed that a contrast of 90% between the hot rods and cold background was obtained after 45 iterations of the ML-EM algorithm.