Abstract
1798
Objectives PET/CT has been recently proposed to visualize 90Y-microspheres after SIRT for liver metastasis. The aim of this study was to assess the ability of PET/CT for accurate quantification of 90Y and hence for reliable 3D patient-specific dosimetry in SIRT.
Methods Data were acquired on a GE Discovery LS PET/CT in 3D mode, 1 bed position. Images were reconstructed using FORE+OSEM (5i, 32s). Spatial resolution was measured using a 1-mm line source. Accuracy of absolute quantification was assessed in terms of recovery coefficients (RC) using the Kyoto phantom with the liver and both intra-hepatic, 20 and 40mm 90Y-filled spheres. A sphere-to-liver activity ratio of 4 was chosen without background activity. Total activity was 1.2 GBq to mimic a patient treatment. Activity data from the 30-min 90Y PET/CT were used to calculate 3D absorbed dose distribution (3DAD) by Monte Carlo using the dosimetry package 3D-RD. Resulting mean absorbed doses (MAD) were corrected using RC. 3DAD were calculated using data from a 90Y PET/CT centered on the liver of a patient with diffused metastases treated with 1.0 GBq of 90Y-labelled SIR-Spheres.
Results The spatial resolution was 9.0 mm (FWHM). RC of 0.40±0.14, 0.73±0.21 and 1.03±0.33 were obtained for 20- and 40-mm spheres and normal liver, respectively. MAD to both spheres and normal liver derived from 90Y PET-based 3D-RD calculation were in good agreement with MAD obtained with OLINDA using true activity. For the patient study, the MAD to the whole liver was 34 Gy, while OLINDA yielded 35 Gy. For a well-defined tumor that showed good response comparing pre- and post-treatment 18F-FDG PET (ΔSUV=-66%), a MAD of 90 Gy was found.
Conclusions This study supports the use of 90Y PET-based dosimetry after SIRT to investigate the dose-response relationship and potentially adjust subsequent therapy