Performance of current generation of three-dimensional positron emission tomography (3D PET) scanners will be discussed, in terms of both physical measures and data computation. Data correction methods will be described and the impact on quantification addressed. Recent improvements include the incorporation of a model-based scatter correction, randoms subtraction, and refinements to the fully-3D iterative reconstruction algorithm. The modifications improve both accuracy of quantification and image quality.

Although overall performance of PET has significantly improved in recent years, there are certain limitations especially for heavy patients where attenuation and scatter effects are increased. We have therefore begun investigations of new scintillators, scanner designs, and image processing algorithms in order to overcome these limitations and further improve imaging performance. The iterative algorithm reconstructs directly from the list-mode data which stores energy, time and positional information of the coincident events. In particular, we are studying the benefits of lanthanum bromide, a scintillator which has outstanding light output, energy resolution and timing resolution. The very high light output leads to excellent spatial resolution, and the excellent energy resolution allows better rejection of scatter and randoms through use of a high energy threshold. Very good timing resolution makes it possible to incorporate the time-of-flight information between coincident gamma rays into the image reconstruction algorithm, which is shown to improve the image signal-to-noise.

Using Monte Carlo simulations, we predict improved scanner performance for a system based on lanthanum bromide, using measures of noise-equivalent count-rate, and lesion contrast and noise of reconstructed phantom data. The simulations are based on detector performance modeled on experimental measurements of lanthanum bromide arrays, and the effects of time-of-flight are included.