Monte-Carlo simulations for the identification of different contributions to scatter in whole-body PET

Summary form only given, as follows. Detailed Monte-Carlo calculations were performed to investigate the necessary assumptions for a scatter correction method in whole-body PET (positron emission tomography). Of particular interest was the extension of Bergstrom's scatter correction method to objects larger than the brain. The calculated scattered coincidences were subdivided into scattered in the gantry only (=gantry only scatter) and scattered in the phantom or in the phantom and in the gantry (=effective scatter). The sum of gantry only scatter and effective scatter gives the total scatter. For axial line sources in water cylinders of 20, 30, and 40 cm diameter the projections of the effective scatter could be described by single exponential functions. Thus, it is justified to extrapolate the effective scatter from the wings of the measured line spread functions into the area under the peak. The gantry only scatter was found to be proportional to the primaries. When added to the primaries there was no effect on the resolution. Therefore, only the effective scatter has to be corrected. These results allow the extension of Bergstrom's scatter correction method to large objects.<>