Optimization of iterative reconstructions of 99mTc cardiac SPECT studies using numerical observers

In this work we investigate the use of a numerical observer to optimize ordered-subset expectation maximization (OSEM) reconstructions for the detection of coronary artery disease (CAD). The parameters optimized are the iteration number and the amount of 3-D Gaussian postfiltering. The numerical observer employed in the optimization was the channelized Hotelling observer (CHO). We used channels as previously determined to rank tumor detection accuracy of different reconstruction strategies, that showed good agreement with the rankings of human observers. The intent of this work was to determine if this CHO could also be employed for the detection of CAD. Results indicate that when quantized images are used, the CHO optimization results correlates well with human observers. However, when using floating-point images, the CHO maintains detection performance even when images are excessively filtered, a result that is at odds with our human observer study. Modeling the additive internal noise that quantization introduces, requires further investigation