Prediction of tumor hypoxia using paired images of tumor perfusion and glucose metabolism in rodent breast cancers

Objectives Imaging tumor hypoxia is very important not only to predict prognosis but also to select the appropriate therapy. The goal of this study is to estimate hypoxic areas in vivo using images of tumor perfusion and glucose metabolism.

Methods Rats bearing syngeneic mammary tumor (RMT) in their interscapular subcutaneous region were prepared. One of the well-established hypoxia markers, pimonidazole (60mg/kg) was injected i.p. six hours before sacrifice, and 14C-deoxy-d-glucose (10μCi) and a perfusion marker, Hoechst 33342 (10mg/kg) were injected i.v. two hours and one minute before sacrifice, respectively. We removed the tumors and made sections of three or four different parts of each tumor. Digital autoradiography for 14C and immunofluorescent staining using a confocal microscope were performed with the same frozen sections. The images of the distribution of the three tracers were quantitatively assessed for density on tumor sections using ImageJ. Correlation among the tracers was analyzed. Furthermore, the data was classified into groups according to Hoechst 33342 and 14C-deoxy-d-glucose density, and 3D graphs were created to analyze the correlation among the three agents.

Results Pimonidazole and Hoechst 33342 (flow) showed a mild negative correlation (r=-0.29, p<0.000001). Hoechst 33342 and 14C-deoxy-d-glucose uptake showed positive correlation at low flow up to a middle level of Hoechst 33342 uptake, but at the highest flow, glucose uptake plateaued. Pimonidazole and 14C-deoxy-d-glucose showed mild positive correlation (r=0.33, p<0.000001) after exclusion of possible necrotic areas. The tissues with low Hoechst 33342 (flow) and high glucose uptake showed the highest pimonidazole accumulation in the comparison of the three tracers.

Conclusions Our results suggest that the combination of perfusion and glucose metabolism images is more helpful to predict tumor hypoxia than either metric alone and that the most hypoxic tissues are those with a signature of high FDG accumulation and relatively low tumor blood flow