Quantitation in positron emission computed tomography: 3 Effect of sampling

Effect of insufficient transverse sampling on quantitative positron emission computed tomography (ECT) was investigated with computer simulation and measurements on parallel bar and line source phantoms. Aliasing artifacts were observed and were found to be dependent on both the configuration and the location of imaged objects. Images of parallel bar phantoms were found to have aliasing artifacts similar in characteristics to aliasing on one-dimensional signals. In line source images, aliasing effects were manifested as variations in amplitude and full width at half maximum resolution (FWHM) for sources at even slightly different locations in the field of view. It was found that employing sampling distances smaller than one-third of the intrinsic detector FWHM eliminated noticeable aliasing artifacts. Image resolution was also found to be affected by the sampling distance. For a sampling distance equal to one-half of the intrinsic detector FWHM, the imaging FWHM is about 10% worse than the intrinsic FWHM. Selection of sampling distance in noisy environments is discussed. Parallel bar phantoms are shown to have advantages over line sources in the evaluation of sampling and resolution performance of ECT scanners.