Estimating Spect Count Densities, Scatter Fractions, and Statistical Noise

Abstract

To provide a quantitative understanding for the interdependence of radiopharmaceutical uptake, detector efficiency, and reconstruction parameters on SPECT statistical accuracy, a unified approach to estimating %rms noise has been developed. The procedure consists of the following steps: 1) Determine an acceptable geometric model for the organ system including radionuclide concentration Qs. 2) Select desired transverse and axial resolutions. 3) Select an acceptable imaging time Tscan. 4) Compute the total number Nt of expected gamma photons in slice using: Nt = π·Rs·Reff·Lslice·Qs·ϵ·Tscan·Abody where ϵ is the detection efficiency (measured in air for a point source), Rs is the physical source radius, Reff is the reduced source radius resulting from self-absorption, Lslice is the slice thickness, and Abody is the attenuation factor of surrounding body tissue. 5) Compute the %rms noise using an equation that includes the effects of spatial filtering and attenuation compensation. Values of Nt and %rms noise predicted by the model are compared with experimental data obtained with one research and four commercial SPECT systems. An expression is derived to estimate the average scatter fraction SFavg for cylindrical sources embedded within a surrounding attenuating medium. SFavg is calculated as a function of source radius and effective attenuation coefficient, and the results are compared with Monte Carlo simulations. The results indicate that the mathematical model is useful in evaluating SPECT performance, providing guidance in the selection of acquisition and reconstruction parameters, improving SPECT quantification, and estimating the usefulness of proposed SPECT radiopharmaceuticals.