Effect of Object Size on Scatter Fraction Estimation Methods for PET—A Computer Simulation Study

Abstract

Scatter fraction ( <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SF</i> ) for PET scanners is typically estimated by making a single measurement using a solid cylindrical phantom with a line source radially offset from the center. The radial displacement of the line source is expected to give a value for scatter fraction that is representative of a typical PET scan for a scanner. A range of phantom sizes suitable for small animal and whole-body PET scanners is investigated. For whole-body imaging, we simulate phantom diameters ranging from 15 to 42 cm, whereas for small animal scanners, we simulate phantom diameters ranging from 2.5 to 15 cm. We find that the line source displacements suggested by the NEMA NU 4-2008 for three phantoms results in a scatter fraction very similar to the one that would arise from uniformly activated phantoms of similar size. On the other hand, the 20 cm phantom used for count rate performance assessment for wholebody scanners is shown to overestimate by about 25% the <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SF</i> of the corresponding uniform phantom, a result that agrees well with that reported by the NEMA committee for the NU 2-2001 standard protocol. Combining the results obtained with small animal and whole-body scanners, we show that the optimal displacement of the line source for estimating the scatter fraction of an equivalent uniformly filled phantom is well approximated by a linear function of the phantom radius and is only weakly dependent on scanner size or detector material. The optimum radial displacement position appears to be at approximately four-fifths of the phantom radius from the center.