Molecular Imaging of Tumor Hypoxia: Existing Problems and Their Potential Model-Based Solutions

Abstract

Molecular imaging of tissue hypoxia generates contrast in hypoxic areas by applying hypoxia-specific tracers in organisms. In cancer tissue, the injected tracer needs to be transported over relatively long distances and accumulates slowly in hypoxic regions. Thus, the signal-to-background ratio of hypoxia imaging is very small and a non-specific accumulation may suppress the real hypoxia-specific signals. In addition, the heterogeneous tumor microenvironment makes the assessment of the tissue oxygenation status more challenging. In this study, the diffusion potential of oxygen and of a hypoxia tracer for 4 different hypoxia subtypes: ischemic acute hypoxia, hypoxemic acute hypoxia, diffusion-limited chronic hypoxia and anemic chronic hypoxia are theoretically assessed. In particular, a reaction-diffusion equation is introduced to quantitatively analyze the interstitial diffusion of the hypoxia tracer [¹⁸F]FMISO. Imaging analysis strategies are explored based on reaction-diffusion simulations. For hypoxia imaging of low signal-to-background ratio, pharmacokinetic modelling has advantages to extract underlying specific binding signals from non-specific background signals and to improve the assessment of tumor oxygenation. Different pharmacokinetic models are evaluated for the analysis of the hypoxia tracer [¹⁸F]FMISO and optimal analysis model were identified accordingly. The improvements by model-based methods for the estimation of tumor oxygenation are in agreement with experimental data. The computational modelling offers a tool to explore molecular imaging of hypoxia and pharmacokinetic modelling is encouraged to be employed in the corresponding data analysis.