Evaluation of 3D Monte Carlo-based scatter correction for 99mTc cardiac perfusion SPECT.

UNLABELLED Cardiac SPECT images are degraded by photons scattered in the thorax. Accurate correction for scatter is complicated by the nonuniform density and varied sizes of thoraxes and by the additional attenuation and scatter caused by female patients' breasts. Monte Carlo simulation is a general and accurate method for detailed modeling of scatter. Hence, for 99mTc we compared statistical reconstruction based on Monte Carlo modeling of scatter with statistical reconstruction that incorporates the more commonly used triple-energy-window scatter correction. Both of these scatter correction methods were used in conjunction with attenuation correction and resolution recovery. METHODS Simultaneous attenuation, detector response, and Monte Carlo-based scatter correction were implemented via the dual-matrix ordered-subset expectation maximization algorithm with a Monte Carlo simulator as part of the forward projector (ADS-MC). ADS-MC was compared to (i) ordered-subset expectation maximization with attenuation correction and with detector response modeling (AD); (ii) like (i) but with scatter correction added using the triple-energy-window method (ADS-TEW). A dual-detector SPECT system equipped with 2 153Gd scanning line sources was used for acquiring 99mTc emission data and attenuation maps. Four clinically realistic phantom configurations (a large thorax and a small thorax, each with and without breasts) with a cardiac insert containing 2 cold defects were used to evaluate the proposed reconstruction algorithms. In these phantom configurations, we compared the performance of the algorithms in terms of noise properties, contrast-to-noise ratios, contrast separability of cold defects, and robustness to anatomic variation. RESULTS Noise was found to be approximately 14% lower in the ADS-MC images than in the ADS-TEW and AD images. Typically, the contrast obtained with the ADS-MC algorithm was 10%-20% higher than that obtained with the other 2 methods. Furthermore, compared with the other 2 algorithms, the ADS-MC method was less sensitive to anatomic variations. CONCLUSION Our results indicate that the imaging performance of 99mTc SPECT can be improved more by Monte Carlo-based scatter correction than by window-based scatter correction.