Physics in Medicine and Biology · 2005 · 96 citations · 39 references
MedicineNeuroscienceNuclear MedicineBiomedical ImagingDiagnostic ImagingMolecular ImagingBrain ImagingComputed TomographyImage ReconstructionDetector Spatial ResolutionExpectation-maximization AlgorithmResolution RecoverySingle-photon EmissionSingle-photon Emission Computed Tomography ReconstructionSpatial Resolution AnalysisSpatial Resolution MeasurementTomography Systems
This study simulates a multi-pinhole single-photon emission computed tomography (SPECT) system using the Monte Carlo method, and investigates different multi-pinhole designs for quantitative mouse brain imaging. Prior approaches investigating multi-pinhole SPECT were not often optimal, as the number and geometrical arrangement of pinholes were usually chosen empirically. The present study seeks to optimize the number of pinholes for a given pinhole arrangement, and also for the specific application of quantitative neuroreceptor binding in the mouse brain. An analytical Monte Carlo simulation based method was used to generate the projection data for various count levels. A three-dimensional ordered-subsets expectation-maximization algorithm was developed and used to reconstruct the images, incorporating a realistic pinhole model for resolution recovery and noise reduction. Although artefacts arising from overlapping projections could be a major problem in multi-pinhole reconstruction, the cold-rod phantom study showed minimal loss of spatial resolution in multi-pinhole systems, compared to a single-pinhole system with the same pinhole diameter. A quantitative study of neuroreceptor binding sites using a mouse brain phantom and low activity (37 MBq) showed that the multi-pinhole system outperformed the single-pinhole system by maintaining the mean and lowering the variance in the measured uptake ratio. Multi-pinhole collimation can be used to reduce the injected dose and thereby reduce the radiation exposure to the animal. Results also suggest that the nine-pinhole configuration shown in this paper is a good choice for mouse brain imaging.
39
MicroPET: a high resolution PET scanner for imaging small animals
Simon R. Cherry, Yiping Shao, Robert W. Silverman et al. · IEEE Transactions on Nuclear Science · 1997 · 615 citations
High-resolution SPECT using multipinhole collimation
N. Schramm, G. Ebel, U. Engeland et al. · IEEE Transactions on Nuclear Science · 2003 · 299 citations