Mapping transient hypoxia from in situ activation of 15O by photon beams: A simulation study

Abstract

BackgroundQuantifying acute hypoxia is desirable yet challenging in radiation therapy. Here, we report a simulation study on the feasibility of mapping acute hypoxia, generated in a 2D simulation vasculature of a pancreas model, using in situ15O washout images. The influence of varying tissue types and photon beam energies (i.e., 20, 25, 30 and 50 MeV) was investigated. Method and materialsAn ex vivo autoradiography 64Cu-ATSM image of a rat tumor was used as a surrogate for the microvessel density map. It was then input to a reaction diffusion model to simulate a steady state partial oxygen pressure (pO2) map. The resulting in situ15O-PET image was generated using a high-energy photon beam and included both mobile and immobile 15O. An image of 15O washout was produced by subtracting the immobile decay portion from the overall decay curve in the in situ PET image, (i.e., fitting the sum of the two exponentials for mobile and immobile oxygen) using different decay constants. Finally, the pO2 map and in situ15O washout images were thresholded and compared. ResultsThe reaction diffusion model showed a heterogeneous distribution of pO2 with mean values of 27.4 and 16.8 mmHg, depending on microvessel permeability, tumor vasculature structure, and architecture. A 62% threshold of the mean intensity in the in situ15O washout image was found to agree reasonably well with the original hypoxia map using 5 mmHg as a cutoff value. These parameters yielded two sets of images with hypoxia fractions of 0.146 and 0.152, respectively. ConclusionsThis simulation study demonstrates the feasibility of mapping acute hypoxia using tissue in situ activation with high-energy photons. With recent advances in in-room or in-beam PET and more work being performed on thresholding the in situ15O washout, mapping acute hypoxia in real time may become a reality.