Page 164 - 89Zr-Immuno-PET:Towards a Clinical Tool to Guide Antibody-based Therapy in Cancer
P. 164

                                Chapter 8
PET imaging study design
For PET imaging, RG7356 (Roche, Basel, Switzerland) was labeled with 89Zr (BV Cyclotron VU, Amsterdam, the Netherlands) according to Good Manufacturing Practice (GMP) standards, as previously described (11-13). Information on the quality control of the radiolabeled tracer can be found in the Supplementary Data. Patients received ~37 MBq 89Zr-labeled RG7356 (1 mg) within two hours after administration of a variable dose (range 0-675 mg) of unlabeled RG7356. Pre- loading with unlabeled mAb was preferred from a logistical perspective (infusion of the unlabeled mAb in a phase I Oncology unit), assuming no difference in uptake between pre-loading and co-infusion.
Whole body PET and low dose CT (ldCT) scans were acquired on a Gemini TF-64 PET/CT scanner (Philips Healthcare, Best, the Netherlands) and scheduled at 1, 24 and 96 h p.i.. Images were reconstructed as described previously (14). After completing the imaging procedure, patients continued in the main study and received RG7356 in the highest dose cohort that was cleared for safety.
89Zr-RG7356 PET analysis
Visual assessment of biodistribution and tumor uptake was performed by a nuclear medicine physician. Tumor uptake of 89Zr-RG7356 was defined as focal uptake exceeding local background. PET scans were considered positive if at least 1 tumor lesion showed 89Zr uptake at 96 h p.i..
Volumes of interest (VOIs) of liver, spleen, kidney, lung, bone marrow, blood pool and brain were delineated to derive mean activity concentrations (ACmean in Bq·mL-1). For lung and brain, VOIs were semi-automatically defined on the ldCT and projected on the PET images. The thresholds for lung VOI’s were based on Houndsfield units (lower limit -1000, upper limit -400), therefore excluding tissue with higher Houndsfield units (e.g. tumor localisations in the lung). VOIs of the liver, spleen and kidney were manually delineated on the PET images themselves, using the ldCT as reference. Fixed sized VOIs with volumes of 8.6 and 2.9 mL were placed on lumbar vertebrae and aortic arch (on ldCT) to estimate ACmean in bone marrow and blood pool, respectively. Tumor lesions were manually delineated and peak activity concentrations (ACpeak) were derived per tumor VOI (15). Assuming that tumor uptake is only due to non-specific uptake (blood volume fraction), the tumor blood volume fraction can be estimated by dividing the AC in the tumor by the AC in the blood pool.
For all VOIs, SUV’s were calculated by dividing AC (decay corrected to time of injection) by injected dose (ID in Bq) corrected for body weight. Additionally,
162


























































































   162   163   164   165   166