Inter-calibrazione di scanner tramite fantoccio al Ge68 per la valutazionequantitativa del SUV in esami PET

PET has always been used in a qualitative way in the staging, restaging, therapy response and follow-up of many of oncological patients, but intrinsically it is a quantitative technique. The purpose of this work is to verify that errors due to instrumentation could be reduced using a thorough procedure for PET cross-calibration with a long half-life isotope filled phantom within a multi-center clinical trial. A NEMA NU-2 image quality phantom filled with 68Ge in an epoxy matrix was made and compared to a standard 18F-filled phantom in term of absolute activity concentration and recovery curves, that are used for evaluating the sensitivity losses of PET scanners due to its limited resolution. The possibility of acquiring high counts statistics images with the 68Ge phantom permitted to: 1) test several metrics (maximum, hot average, peak and averaged activity) measured in 2D and 3D regions as the best estimates of the reference activity concentration in 68Ge phantom 2) evaluate the repeatability and reproducibility of the PET scan acquisition 3) study the dependency of iterative and filtered back-projection reconstruction algorithm parameters on recovery curve and on noise The effect of varying reconstruction algorithm parameters and acquisition duration was analysed on four different metrics for noise: standard deviation, image roughness, background variability and ensemble noise. Two additional wall-less 68Ge-filled spheres were added the usual six plexiglass spheres, all of them filled with an activity concentration four times higher than the surrounding background. They were used to evaluate the influence of the thick plastic wall on recovery coefficient. After the complete characterization of the 68Ge phantom in Cuneo, it was delivered to several PET centres in Italy for the cross calibration of the PET scanner and the optimization of the reconstruction algorithm.