IMPROVING THE PRECISION AND ACCURACY OF MONTE-CARLO SIMULATION IN POSITRON EMISSION TOMOGRAPHY

Most of the gamma-rays generated following positron annihilation in subjects undergoing PET studies never reach the detectors of a positron imaging device. Similarly, simulations of PET systems waste considerable time generating events which will never be detected. Simulation efficiency (in terms of detected photon pairs per CPU second) can be improved by generating only rays whose angular distribution gives them a reasonable chance of ultimate detection. For many events in which the original rays are directed generally towards the detectors, the precision of the final simulation can be improved by "recycling" them without reseeding the random number generator, giving them a second or greater chance of being detected. For simulation programs which cascade the simulation process into source, collimator, and detection phases, recycling can improve the precision of the simulation without requiring larger files of events from the source distribution simulation phase. The speed, accuracy and ease of use of the simulations of PET systems can be improved by the use of tabulated values of the photoelectric and Compton absorbtion coefficients for all common biological, collimator and detector materials. These can now be generated from the chemical formula or the physical composition, and the density of the absorbing medium.