Imaging hypoxia after oxygenation-modification: comparing [18F]FMISO autoradiography with pimonidazole immunohistochemistry in human xenograft tumors.

PURPOSE: Hypoxia is one of the reasons for radiation therapy resistance. Positron emission tomography using (18)F-labeled misonidazole ([(18)F]FMISO) is a non-invasive method of imaging tumor hypoxia. Aim of this study was to validate [(18)F]FMISO against the clinically most widely used hypoxic cell marker pimonidazole under different oxygenation conditions.
MATERIALS AND METHODS: One human head and neck squamous cell carcinoma (SCCNij3) and two human glioblastoma (E102 and E106) xenograft tumor lines were studied after injection of [(18)F]FMISO and pimonidazole. Control mice were compared with a second group breathing carbogen to reduce tumor hypoxia and with a third group with clamped tumors to increase hypoxia. Tumor sections were analyzed on a phosphor imaging system and consecutively stained immunohistochemically (IHC) for visualization of pimonidazole. Pixel-by-pixel analysis was performed and the hypoxic fraction, obtained after segmentation of the pimonidazole signal, was related to the mean optical density of [(18)F]FMISO and pimonidazole.
RESULTS: A moderate pixel-by-pixel correlation between [(18)F]FMISO autoradiography and pimonidazole IHC was found for the control tumors, after carbogen breathing and after clamping for SCCNij3. For E102 and E106, mean signal intensities for pimonidazole significantly decreased after carbogen breathing and increased after clamping, mean [(18)F]FMISO signal intensities increased significantly after clamping and a significant correlation between the hypoxic fractions and the mean [(18)F]FMISO signal intensities was found.
CONCLUSIONS: [(18)F]FMISO autoradiography and pimonidazole immunohistochemistry can both be used to visualize treatment induced changes in tumor hypoxia. However, the response to these modifications differs widely between xenograft tumor lines.