[Appropriate energy window width for gamma camera]

Scatter radiation is one of the major sources of error in nuclear medicine data processing. Different methods of scatter correction have been introduced in order to improve the quality of data. However the best method is to avoid recording of scatter photons in acquisition. The only difference between scattered and non-scattered photons is the energy. Pulse height analyzer is the only option available to discriminate primary photons from scattered ones. Energy resolution of the gamma camera is gradually improving consequently the energy window width has to be decreased accordingly. In this study we tried to determine the most appropriate energy window width for present gamma camera systems. Methods and Materials: Since it is not possible to retrieve the data spectrum from the most of the gamma camera systems, a simple method was developed to extract the data from the image of the energy spectrum. Using a scatter phantom different level of scatter and count rate were generated and corresponding energy spectrum data were analyzed. It was assumed that around the peak of the spectrum, the primary photons obey a Gaussian distribution. The data were analyzed using three different methods. All methods prove that the optimum window width regarding the present gamma camera energy resolution is 15%. At this level, the scattered radiation is decreased to 5%. In comparison to the conventional widow width of 20%, the sensitivity does not change dramatically. At the present, for most gamma camera, the energy window width of 20% is recommended. However occasionally energy window width of 15% and 25% are also used. In this study the energy spectrum at different levels of scatter were analyzed and the most suitable energy window width was found to be 15% for the gamma camera having approximate energy resolution of 11%. At this window setting the scatter decreases to 5% of the total counts recorded. Visually the quality of the images dose not improves significantly. However accuracy of data quantification improve significantly.