Study of the effects of whole body PET spatial sampling schemes on data SNR

The limited axial Field Of View of state of the art Positron Emission Tomographs, necessitates the movement of the patient bed at multiple axial positions for the acquisition of Whole Body studies. From the variety of the possible sampling schemes in the axial direction, the data acquisition method that effectively improves the axial uniformity of the data Signal to Noise Ratio and eliminates the resolution artifacts in Whole Body studies, is the continuous spatial sampling scheme. In this work, in addition to the SNR uniformity benefit from the continuous sampling scheme, a gain in the absolute SNR value of the data is investigated. The continuous axial sampling scheme implies that all the data points are sampled with all the detectors in the axial direction hence the efficiencies of the detector system in the axial direction are averaged. As a result the overall variance of the data after the correction for the detector efficiency variation is reduced. A method is developed, that theoretically estimates the overall variance of a data set corrected for detector efficiency, as a function of the normalization correction matrix. This method can be used to predict the relative SNR performance of any PET tomograph under different data sampling schemes. It can also be used for the comparison of the relative data SNR between two PET systems that have the same overall sensitivity but differ in the efficiency uniformity of their detector systems. When the method is used to theoretically predict the overall data variance improvement from the discrete to the continuous sampling scheme for an ECAT EXACT HR PET system, the theoretically calculated values closely match the measured improvements.