Examining tumor hypoxia using MRI (BOLD/TOLD) in correlation with [18F]FMISO PET

1138 Objectives Hypoxia is an important biomarker for tumor progression and response to therapy. [18F]-fluoromisonidazole ([18F]FMISO) is the most widely used PET radiotracer for imaging tumor hypoxia. Various imaging techniques are developed to assess tumor hypoxia, but many are invasive, inconvenient, or impractical in the clinical settings. Oxygen sensitive MRI is an attractive alternative option. Blood Oxygen Level Dependent (BOLD) and Tissue Oxygen Level Dependent (TOLD) MRI offer potential insight into tumor hypoxia. Here, we compared oxygen sensitive MRI in prostate tumor bearing rats with [18F]FMISO. Methods Five rats bearing subcutaneous Dunning prostate R3327-AT1 tumors (1.4 to 2 cm3) were examined with respect to oxygen breathing challenge using interleaved BOLD (T2*-weighted) and TOLD (T1-weighted) sequence, which is sensitive to paramagnetic deoxyhemoglobin and tumor tissue oxygenation (pO2) respectively. Immediately following the administration of the tracer (570-600 μCi), dynamic PET imaging was performed up to 90 mins using a Siemens Inveon® PET/CT system. Imaging data (0-90 min) were reconstructed into eighteen frames (5 min intervals), and the time activity curve was generated. Results BOLD and TOLD responses averaged 6.7%ΔSI and 5.3%ΔSI respectively. The average tumor to muscle ratio by PET was greater than 1.8 at 90 min indicating hypoxia. Each technique showed heterogeneity. PET imaging showed maximum [18F]FMISO uptake in the peripheral regions and MRI showed lower BOLD response in the same region consistent with hypoxia. The results were further verified by performing immunohistochemistry on the tumor tissue. Pimonidazole accumulating in hypoxic regions and Hoechst stain in well perfused regions were observed. Conclusions PET imaging showed highest [18F]FMISO uptake in the tumor periphery suggesting a strong influence of perfusion. However, the regions also corresponded to generally lower BOLD signal responses consistent with hypoxia. Immunohistochemistry confirmed the presence of hypoxic regions. Research Support Supported in part by NCI R01 CA13904 and U24CA126608.