Performance evaluation of 4D reconstruction methods for gated cardiac single photon emission computed tomography in obese patients

Abstract

The purpose of this study is to evaluate 4D reconstruction methods for the processing of gated cardiac single photon emission computed tomography (SPECT) images from obese patients.Clinical gated SPECT projection data were reconstructed using the ordered-subsets expectationmaximization (OS-EM) IntroductionMyocardial SPECT imaging is frequently the imaging technique of choice used to visualize perfusion defects of the heart. Cardiac gating adds a vital dimension to myocardial imaging, enabling the diagnostician to visualize cardiac motion and wall thickening, and assess cardiac function quantitatively by means of chamber volumes calculated by clinical software packages. A principal problem with gated SPECT images is the increase in noise due to the division of acquired counts into time frames, and this issue is magnified in the case of obese patients, due to additional attenuation from the larger body size. The images can be noisy enough for clinicians to discount 4D data altogether in this population of patients.Our hypothesis is that 4D reconstruction methods applied to gated studies in this population will result in more usable images with less noise degradation and no loss of temporal resolution as compared to images smoothed with 2D and 3D Fourier filters. The rescaled block-iterative maximum a posteriori (RBI-MAP) algorithm is a modified version of the iterative RBI-EM algorithm, closely related to the ordered-subset, expectation-maximization (OS-EM) algorithm [1]. The former incorporates useful features of the MAP reconstruction method (particularly by the use of Gibbs priors [2], which are designed to retain edge features while smoothing noise, thus contributing to better temporal resolution when applied in the time dimension [3]). It is not known whether 4D reconstruction methods result in improved image quality in reconstructed gated SPECT images from obese patients, as compared to 3D reconstruction methods. Thus, the aim of this study is to evaluate this question by means of computer simulation tools described herein. MethodsAnthropometric data was used to obtain a distribution of body dimensions across different populations. This information was used to determine an appropriate body size to use in building a standard obese male and female non-uniform rational B-spline-based (NURBS) cardiac torso (NCAT) phantom. The NCAT phantom is a computerized phantom which includes models of cardiac and respiratory motion, allows for the modeling of perfusion and motion defects of the myocardium, for the modeling of different organ and body sizes, as well as for the modification of radioactivity uptake ratios and attenuation maps [4], [5]. It is thus a very powerful tool that provides the researcher with a gold standard on which to test the efficacy of different reconstruction