Factors Impacting Clinical Protocol Optimisation and SUVmax uncertainty in [68Ga]-PSMA PET Imaging

2525 Objectives To measure the pixel standard deviation (SD) of SUVmax and to assess scanning parameters to determine an optimal protocol for whole body [68Ga]-PSMA PET scanning. Methods Following injection of 100-150 MBq of [68Ga]-PSMA, externally triggered gated (ETG) WB PET scans were acquired at 60 minutes and 2 to 2.5 hours on a ToF PET/CT (Siemens Biograph mCT64x64). Data were reconstructed into 6 gates using ToF 3D-OSEM with CT-based AC, SC and PSF recovery. Mean, SD and coefficient of variation (CoV) images were generated for both early and delayed imaging. Regions-of-interest were placed over areas of assumed homogeneity (e.g., liver), blood pool, right normal parotid gland and bladder. For each ROI, SUVmax and SUVSD (SD of the single pixel containing the SUVmax) values were determined, and image coefficient of variation (CoV%) and liver target:background, SUVmax for both imaging timepoints. Results A total of 21 patients (delay n=13) with elevated PSA, mean injected dose 137 MBq (107 to 172 MBq) and weight 77.6 kg (64.7 to 106 kg). As previously seen, higher the uptake in tissue lower the uncertainty in SUVmax with normal parotid and bladder showing lowest noise estimates, increasing on delayed imaging. CoV liver early and delayed was 20% and 42% respectively, higher than determined for FDG 4.2%1. Liver:background was 3 and 3.5 respectively (c.2 to 5,6; c.1.8 to 7.4). Conclusions Methodology directly measures the uncertainty in reconstructed pixel values due to instrumentation and technical factors. Further uncertainties may be present due to physiological factors. Higher noise estimates on delayed imaging increases the uncertainty in SUVmax measures, most likely due to the sort half-life of Ga-68. ETG acquired data allows assessment of varying parameters to achieve an optimum between injected dose and scanning duration.