Validation on an anthropomorphic phantom of FORE optimization in 3D PET

A study for validating a proposed optimization of Fourier Rebinning (FORE) is presented. FORE is the most widely used rebinning algorithm for 3D PET data and operates into sinograms Fourier Transform domain. The frequency-distance relation (FDR) on which FORE is based is valid at high frequencies; low frequencies are usually rebinned with single slice rebinning (SSRB). A square low-frequency region is commonly defined on empirical basis and implies an abrupt transition of rebinning strategy. We proposed a simple index able to map the validity of the FDR in the sinogram frequency space and a consequent gradual transition mask for an optimized FORE. In this study we tested the method on an anthropomorphic phantom filled with activities simulating the in vivo 18F-FDG uptake and containing spherical lesions of different radii. Both abrupt and gradual partitions were implemented with different parameters corresponding to a different extension of the low-frequency SSRB region. Results show that, in the considered condition, the performances of the two methods are similar. Theoretically, a larger robustness of the gradual transition is foreseen on more complex structures and in more critical conditions characterized by appreciable discrepancies between Fourier Transform of rebinned sinograms obtained with FORE and SSRB.