Kinetic modeling of [¹⁸F]-FMISO microPET data and its correlation with image-guided pO₂ measurements

Objectives Hypoxia increases resistance to treatment and may predict tumor progression. Non-invasive methods to determine intra-tumor hypoxia, such as PET, are being explored and validated against direct methods such as partial oxygen pressure (pO₂) probe measurements and immunohistochemistry. The current study validates ¹⁸F-fluoromisonidazole ([¹⁸F]-FMISO) PET imaging in the assessment of tumor hypoxia by demonstrating the expected correlation between microPET-derived [¹⁸F]-FMISO kinetic parameters in rat tumors with spatially registered pO₂ measurements.

Methods Sixteen Dunning R3327-AT prostate tumor-bearing nude rats were immobilized in a custom-fabricated whole-body mold, injected iv with FMISO, and imaged for 105 minutes on a Focus 120 microPET. Maintaining anesthesia, animals were then transferred in situ to our robotic system for image-guided intra-tumoral pO₂ measurements (Oxylite probe). A removable registration plate with 4 fiduciary markers was used to align the robot and microPET coordinate systems.

Results A voxel-based 2-tissue compartment model of the microPET-derived FMISO kinetics in tumor yielded a significant inverse correlation between pO₂ and the hypoxia-dependent FMISO “trapping” parameter (k₃). Specifically, scatter plots exhibited a sigmoidal relationship between the pO₂ vales and the corresponding FMISO k₃, flux and Patlak (K_i) parameters with correlation coefficients (mean + SD) of 0.72 ± 0.29, 0.75 ± 0.28, and 0.80 ± 0.25, respectively.

Conclusions Image-guided pO₂ probe and measurements validated FMISO PET imaging of tumor hypoxia, but also illustrated statistical and resolution limitations in the correlation between pO₂ values and FMISO uptake parameters