Abstract
PURPOSE: Therapeutic administration of (90)Y-loaded microspheres is routinely used for primary and secondary liver tumours. For activity-based therapeutic prescription the activity must be within 10% of the intended activity. Previous studies reported significant discrepancies between manufacturer-declared vial activities and both experimental and Monte-Carlo assessments, greater than 10%, for resin/glass (90)Y-microspheres. The objective of this work was to investigate whether these discrepancies were also seen in patients. METHODS: We analysed patient (90)Y-PET reconstructions (99 glass and 15 resin microspheres) from 4 different institutions and 4 different systems. We considered tail-fitting background scaling (TFBS) and absolute scaling (ABS), for scatter correction. Residuals after therapeutic injection were measured. Eighty-one patients were imaged with PET/CT and 33 with PET/MR. The PET measured activity (A(PET)) was assessed in the whole liver. The ratio A(PET)/A(calibrator) was calculated for each patient, where A(calibrator) was the injected activity measured by the dose calibrator corrected for residual and lung shunt. RESULTS: Quantification ratio between calibrators and PET was significantly different from 1, regardless of the scatter correction used. In glass microspheres, the mean A(PET/CT)/A(calibrator) was 0.84 ± 0.06 for TFBS and 0.90 ± 0.06 for ABS (0.66 ± 0.09 and 0.76 ± 0.07 for (A(PET/MR)/A(calibrator))). The mean A(PET/CT)/A(calibrator) ratio for resin microspheres was 1.16 ± 0.09 for TFBS and 1.30 ± 0.12 for ABS. CONCLUSIONS: We observed in patients similar activity discrepancies as reported for vials, with a relative difference of 44 ± 16% between glass and resin (90)Y-loaded microspheres. In (90)Y hepatic radioembolization, the 10% accuracy prerequisite on knowing the administered therapeutic activity is then unlikely to be met.