Impact of injection dose, post-reconstruction filtering, and collimator choice on image quality of myocardial perfusion SPECT using cadmium-zinc telluride detectors in the rat

Abstract

Abstract Background The aims of this study were (1) to evaluate the impact of injection dose, post-reconstruction filtering, and collimator choice on image quality of myocardial perfusion single-photon emission computed tomography (SPECT) using cadmium-zinc telluride (CZT) detectors and (2) to determine how these factors affect measured infarct size in the in vivo rat. Methods Twenty-four healthy and eight myocardial infarct (MI) rats underwent myocardial perfusion SPECT imaging after injection of various doses (25 to 200 MBq) of 99mTc-tetrofosmin using a standard (STD) five-pinhole collimator and high-sensitivity (HS) five-pinhole collimator. Image quality score, contrast-to-noise ratio, sharpness index, coefficient of variation (CV), and measured defect size were assessed as measures of image quality. Results The image quality score increased and CV decreased as a function of injection dose. The contrast-to-noise ratio increased and sharpness index decreased as a function of Gaussian kernel size. When STD and HS were compared, HS tended to show higher image quality score and lower CV than STD. The use of post-reconstruction filter significantly improved image quality score and lessened CV. The reproducibility of defect size measurements, as assessed by intraclass correlation coefficients (ICC), between the collimators was poor-to-moderate (ICC = −0.31~0.57) with low (25 MBq) injection dose and with no or light (1.5-mm kernel size) filtering, whereas it was good-to-excellent (ICC = 0.75~0.97) with high (200 MBq) dose or low dose with heavy (2.5-mm kernel size) filtering. The filtering-related reproducibility was poor (ICC = −0.18~0.17) for STD with low injection dose, whereas it was good-to-excellent (ICC = 0.79~0.89) for HS. Furthermore, there was a filtering-related underestimation of defect size particularly with the use of heavy smoothing. Conclusions Appropriate imaging setting is important to obtain high quality images and thereby reliable measurements using a preclinical myocardial SPECT in the rat. When only a low injection dose (25 MBq) is allowed, we would recommend to use HS with light (1.5-mm kernel size) filtering.