Aim: The spatial resolution of PET scanners can be improved by using smaller detector elements. This approach, however, results in poorer counting statistics of the reconstructed images. Therefore, the aim of this study was to investigate the influence of different acquisition parameters on the signal-to-noise ratio (SNR) and thus to optimize PET image quality.
Methods: The experiments were performed with the latest-generation whole-body PET system (ECAT Exact HR+, Siemens/CTI) using the standard 2D and 3D data acquisition parameters recommended by the manufacturer. The EEC whole-body phantom with different inserts was used to simulate patient examinations of the thorax. Emission and transmission scans were acquired with varying numbers of events and at different settings of the lower level energy discriminator. The influence of the number of counts on the SNR was parameterized using a simple model function.
Results: For count rates frequently encountered in clinical PET studies, the emission scan has a stronger influence on the SNR in the reconstructed image than the transmission scan. The SNR can be improved by using a higher setting of the lower energy level provided that the total number of counts is kept constant. Based on the established model function, the relative duration of the emission scan with respect to the total acquisition time was optimized, yielding a value of about 75% for both the 2D and 3D mode.
Conclusion: The presented phenomenological approach can readily be employed to optimize the SNR and thus the quality of PET images acquired at different scanners or with different examination protocols.