学术文献库

Purpose: Explaining why 90Y TOF-PET based equivalent uniform dose (EUD) using standard FDG reconstruction parameters has been shown to predict response. Methods: The hot rods insert of a Jaszczak deluxe phantom was partially filled with a 2.65 GBq 90Y - 300ml DTPA water solution resulting in a 100 Gy mean absorbed dose in the 6 sectors. A two bed 20min/position acquisition was performed on a 550ps- and on a 320ps- TOF-PET/CT and reconstructed with standard FDG reconstruction parameters, without and with additional filtering. The whole procedure was repeated on both PET after adding a 300ml of water (50Gy setup). The phantom was acquired again after decay by a factor 10 (5Gy setup), but with 200min per bed position. For comparison the phantom was also acquired with 18F activity corresponding to a clinical FDG whole body acquisition. Results: The 100Gy-setup provided hot rod sectors image almost as good as in 18F phantom. However, despite acquisition time compensation the 5Gy-setup provides much lower quality imaging. TOF-PET based sectors EUDs for the three large rod sectors were in agreement with the actual EUDs computed with a radiosensitivity 0.021Gy-1 well in the range observed in external beam radiotherapy (EBRT), i.e. 0.01-0.04Gy-1. This agreement explains the reunification of the dose-response relationships of the glass and resin spheres in HCC using the TOF-PET based EUD. Additional filtering reduced the EUDs agreement quality. Conclusions: Standard FDG reconstruction parameters are suitable in TOF-PET post 90Y liver radioembolization for accurate tumor EUD computation. The present results clearly rules out the use of low specific activity phantom studies to optimize reconstruction parameters.