Purpose: Tumour hypoxia is linked to treatment resistance. Positron emission tomography (PET) using hypoxia tracers such as fluoroazomycin arabinoside (FAZA) may allow identification of patients with hypoxic tumours and the monitoring of the efficacy of hypoxia-targeting treatment. Since hypoxia PET is characterized by poor image contrast, and tumour hypoxia undergoes spontaneous changes and is affected by therapy, it remains unclear to what extent PET scans are reproducible. Tumour-bearing mice are valuable in the validation of hypoxia PET, but identification of a reliable reference tissue value (blood sample or image-derived muscle value) for repeated scans may be difficult due to the small size of the animal or absence of anatomical information (pure PET). Here tumour hypoxia was monitored over time using repeated PET scans in individual tumour-bearing mice before and during fractionated radiotherapy.
Methods: Mice bearing human SiHa cervix tumour xenografts underwent a PET scan 3 h following injection of FAZA on two consecutive days before initiation of treatment (baseline) and again following irradiation with four and ten fractions of 2.5 Gy. On the last scan day, mice were given an intraperitoneal injection of pimonidazole (hypoxia marker), tumours were collected and the intratumoral distribution of FAZA (autoradiography) and hypoxia (pimonidazole immunohistology) were determined in cryosections.
Results: Tissue section analysis revealed that the intratumoral distribution of FAZA was strongly correlated with the regional density of hypoxic (pimonidazole-positive) cells, even when necrosis was present, suggesting that FAZA PET provides a reliable measure of tumour hypoxia at the time of the scan. PET-based quantification of tumour tracer uptake relative to injected dose showed excellent reproducibility at baseline, whereas normalization using an image-derived nonhypoxic reference tissue (muscle) proved highly unreliable since a valid and reliable reference value could not be determined. The intratumoral distribution of tracer was stable at baseline as shown by a voxel-by-voxel comparison of the two scans (R = 0.82, range 0.72-0.90). During treatment, overall tracer retention changed in individual mice, but there was no evidence of general reoxygenation.
Conclusion: Hypoxia PET scans are quantitatively correct and highly reproducible in tumour-bearing mice. Preclinical hypoxia PET is therefore a valuable and reliable tool for the development of strategies that target or modify hypoxia.