Optimization of the width of the photopeak energy window in the TDCS technique for scatter correction in quantitative SPECT

Transmission-dependent convolution subtraction (TDCS) is a promising technique in quantitative SPECT for subtracting the scatter components from emission images. Usually, a 20% photopeak energy window is used in SPECT acquisitions. To date, no investigation has assessed the effects of energy windows in the subtraction of scatter components from emission images with the TDCS technique. To evaluate the energy dependence of the TDCS technique, we analyzed photopeak energy windows of 5%, 10%, 15%, 20%, 25%, 30%, and 35% using /sup 99m/Tc radionuclide. The scatter fractions for a point source placed in a water phantom were estimated using the triple-energy-window (TEW) technique. These estimates were used to establish the parameters of the TDCS equation as a function of transmission. The estimated parameters were applied to the TDCS technique to subtract the scatter components from data of corresponding energy windows using nonuniform and uniform phantoms. All data were reconstructed with the OSEM algorithm. The energy window dependence of the TDCS technique was verified by comparing the coefficients of variance (COV) of the data from the uniform phantom of each energy window. Ten Poisson deviates were generated from each projection's data from the uniform phantom. The nonuniform phantom results showed that the estimated parameters effectively estimated the scatter component from the projection image for each energy window. The COV for the 5%, 10%, 15%, 20%, 25%, 30%, and 35% windows were 2.10%, 1.76%, 1.44 %, 1.39%, 1.34%, 1.37%, and 1.61%, respectively. In the TDCS technique, any photopeak energy window from 15% to 30% may work almost as well as judged by the COV measured for a uniform phantom in this study. The best window by that measure was the 25% window.