SPET refers to a computerized emission tomographic system that depends on that single photons (as distinct from positrons in PET). Single photons are detected singly rather than in coincident pairs (as in PET). Collimation – the trapping of emitted photons and their direction towards the detector – is required because single photons are scattered randomly, and this means that most photons are absorbed by collimators and thus go undetected. Thus only a fraction of emitted photons are counted by SPET detector systems and SPET resolution is achieved at the expense of SPET sensitivity. The sensitivity of SPET is the degree to which the system responds to an incoming signal measured as counts per second (CPS) per slice (megaBequerel per litre or MBq/L). The most frequently used detector systems in clinical practice are rotating .
Once acquired, SPET data are organized as slices, and reconstructed separately from projections spaced over a 360° arc of rotation about the subject. The SPET detector system behind the collimator is made of with tubes (PMTs) and SPET detector systems may have one large detector covered with many PMTs, or multiple detectors capable of higher count-rate detection and suitable for dynamic studies. Brain-dedicated SPET detector systems view the head from several angles simultaneously with separate scintillation detectors, while converging collimators increase the crystal surface area utilized for a given slice and thus maximize sensitivity. Reconstruction in transverse, coronal and sagittal planes is possible (Fig. 35.5). SPET images are collected over a much longer period of time and depend on many fewer photons than standard CT. Thus SPET images have more noise and less resolution than CT images. As with PET, ROI data analysis is still commonly used in SPET. However, SPM (see above) has recently been adapted for SPET. Ideal anatomical localization with SPET would require co-registration as with PET, but in contrast to PET, this is difficult to achieve with SPET.