Optimisation of simultaneous emission and transmission measurements in PET

Sinogram windowing of rotating rod sources allows the possibility of simultaneously acquiring emission and transmission data. While transmission signal to noise ratio (SNR) is improved, mainly due to rejection of random coincidences, they account for a large fraction of emission prompts, severely limiting emission SNR. The authors have investigated three methods for optimising SNR in these measurements. Firstly, transmission SNR can be traded off by reducing rod source activity, resulting in a reduced randoms rate. Secondly, noise in the randoms histogram cap be reduced using Casey processing. Thirdly, segmented attenuation correction (SAC) can be applied to simultaneously acquired transmission data to minimise propagation of noise. Rod activity can be reduced by a factor of 4 without loss of SNR compared with a non windowed transmission study. As rod activity decreases, the incremental improvement in SNR decreases using Casey processing and increases using SAC. Using low activity rods and SAC, there is less than 10% difference in SNR between simultaneous and separate measurements. Initial results of total body PET studies are presented employing the methods described combined with accelerated EM reconstruction