Assessment of the scatter fraction evaluation methodology using Monte Carlo simulation techniques

To evaluate scatter fraction and scatter pair spatial distribution, experimental methods are generally used. These methods make use of a line source, placed along the FOV axis, inserted in a cylindrical phantom filled with air or water. The accuracy of these experimental methodologies can be tested by the use of a Monte Carlo method. In fact, the simulation allows the shape of the scatter event projection and the scatter fraction to be defined. An example of this application is the simulation package PETSI (PET SImulation). In this paper the comparison between the predicted scatter fraction and the experimentally evaluated one, obtained using an ECAT III PT 911/02 double ring whole body scanner are presented. PETSI permits additional data to be obtained: a) the true and scatter component of the energy spectrum; b) the spatial distribution, in the FOV plane, of the detected scatter events at different energy thresholds; c) the scatter to total detected events ratio; d) the predicted scatter fraction at both energy thresholds and FOV diameters. This information is very useful for optimizing both energy threshold and FOV size and to improve the accuracy of the currently used methods for the scatter fraction evaluation. Preliminary results of the predicted scatter fraction in a uniform phantom are presented.