Quantitative Imaging With a Pinhole Cardiac SPECT CZT Camera

This work develops methods to quantify absolute activity measurements with a multi-head pinhole cardiac SPECT camera that uses cadmium-zinc-telluride (CZT) detectors (Discovery NM530c, GE Healthcare). A central component of absolute activity quantification is the correction for Compton scattered photons. A modified Dual Energy Window (DEW) scatter correction (SC) method was developed that compensates for the presence of unscattered photons in the lower-energy window used to measure scatter. The DEW-SC method was validated using phantom experiments. The mean error in absolute activity measurement was 5 ± 4 % when correcting for attenuation, scatter and the partial volume effects. Clinical accuracy was assessed using 99mTc-tetrofosmin myocardial perfusion images obtained on the NM530c and comparing to DEW scatter-corrected images acquired on a conventional SPECT camera for the same patient. DEW-SC images acquired on the NM530c had increased noise but had summed perfusion scores that were not significantly different from those acquired on the conventional SPECT camera. To reduce noise in scatter corrected images and provide a more accurate estimate of scatter, a model-based SC method was developed based on the analytical photon distribution (APD) approach. Using physical phantom experiments and a set of five clinical studies, the accuracy of APD-SC was evaluated and compared to DEW-SC. Images generated using the model-based method agreed well with acquired data. APD-SC images had lower noise than DEW-SC images and provided a more accurate measure of cardiac activity in high-scatter scenarios. The developed methods provide an accurate means of correcting for scatter and thereby allow the absolute quantification of pinhole cardiac SPECT images.