A comparison of scatter compensation methods in SPECT: subtraction-based techniques versus iterative reconstruction with accurate modeling of the scatter response

The authors compare two subtraction-based techniques, dual window and convolution subtraction, to iterative reconstruction incorporating an accurate model of the scatter response. Two phantoms, consisting of a cylindrical attenuator with circular cross section and hot or cold rods, were used. Projection data were simulated using Monte Carlo techniques. They were reconstructed using the maximum-likelihood-expectation maximization (ML-EM) algorithm with a projector/backprojector that accurately models photon attenuation and the spatially variant geometric and scatter responses. In addition, the data were scatter-compensated using the subtraction techniques and then reconstructed using ML-EM with a projector/backprojector modeling only attenuation and collimator response. It was found that the images reconstructed with the accurate scatter model are superior in terms of mean square error, standard deviation, and signal-to-noise ratio to data reconstructed following scatter subtraction. The quantitative accuracy of the images reconstructed with subtraction techniques depends critically on the selection of subtraction parameters