which selectively sensitizes hypoxic cells to ionizing radiation by replacing oxygen in the chemical reactions that lead to the production of DNA damage [30]. Imaging patients for a priori selection of hypoxic tumors will be of the utmost importance of determining the effect of hypoxia-modification strategies.
PET-Based Hypoxia Imaging
There are 2 main tracer classes for imaging of hypoxia, the 18F-labeled nitroimidazole-like compounds and the Cu-labelled diacetyl-bis(N4-methylthiosemicarbazone) analogues (Cu-ATSM) [10]. Since 18F-FDG-PET visualizes both aerobic and anaerobic glucose consumption, this tracer is not suited for visualization of the oxygenation status of tissue [14, 31].
Mechanism of Nitroimidazole Uptake in Hypoxic Tissue
After injection into the bloodstream of the patient, the radiolabeled nitroimidazole derivatives spread throughout the whole blood volume. The lipophilic nature of the compound leads to passive diffusion into cells [32]. Within the cells nitroimidazoles combine with an electron, which has a high affinity for nitroimidazole, especially the -NO2 part. The electron and nitroimidazole form a radical anion: -NO2-. Although electrons have a high affinity for nitroimidazole, they have an even higher affinity for oxygen. In the presence of oxygen, the electron in the radical anion can be taken over by the oxygen and the nitroimidazole returns to its parent state and can exit the cell again [33]. In the case of hypoxia (pO2 <10 mm Hg), binding to an electron is not irreversible and the reduction continues. The reduction product remains trapped in the cell. Reduction of nitroimidazoles relies on the presence of active tissue reductases, which means that only viable hypoxic cells accumulate the reduction products and apoptotic or necrotic cells do not [23]. The accumulation of reduction products in hypoxic cells is inversely proportional to the local pO2[10]. When nitroimidazoles are radiolabeled, the radioactivity also accumulates in the hypoxic cell, which enables visualization and quantification of vital tissue hypoxia by PET imaging.
Different Nitroimidazole-Based PET-Tracers
Of the different nitroimidazole-based PET-tracers, 18F-fluoromisonidazole (18F-FMISO) was the first and still remains the most extensively studied [32, 33]. Although there is only limited clinical experience with 18F-FMISO, there are some useful reviews describing the proven, assumed, or hypothesized differences between different nitroimidazole-based tracers [10, 23, 32]. In validation studies, a good correlation between 18F-FMISO, visualized by autoradiography, and immunohistochemical visualization of 2-nitroimidazole derivatives was found [34]. Furthermore, preclinical work showed that changes in hypoxia could, to a certain extent, and based on tissue architecture, be visualized and quantified with this compound [35].
18F-FMISO has the limitation of slow pharmacokinetics (about 2 h minimum between administration and imaging) and a poor hypoxic versus normoxic tissue contrast (20–40% difference in uptake). Therefore, other nitroimidazole-based tracers were developed in order to overcome these shortcomings. The different nitroimidazole-based tracers show (small) differences in uptake and imaging characteristics. In general, compounds that are more hydrophilic have a higher clearance and therefore higher tumor-to-blood ratio (T/B) [23]. Overall, theoretical benefits of other tracers, as compared to 18F-FMISO, only lead to relatively small advantages in clinical practice. 18F- fluoroazomycinarabinofuranoside (18F-FAZA) is one of the second-generation nitroimidazole compounds and, due to its reduced lipophilicity, it has a faster washout of normoxic tissue, leading to a higher tissue-to-muscle ratio (about 2.0 at 2 h postinjection) [32]. 18F-HX4 is one of the other more hydrophilic tracers with a faster clearance than 18F-FMISO. Other examples of nitroimidazole-based tracers are 18F-EF3 and 18F-EF5, which have a more complex but also more stable labeling chemistry. Unfortunately, in practice the potential advantage over 18F-FMISO turned out to be limited [32].
Today, many different nitroimidazole-based tracers are under investigation and, while they all have a slightly different profile, it is difficult to predict which one has the best overall potential to succeed 18F-FMISO as the leading hypoxia PET tracer.
In preclinical studies, tracers labeled with other positron-emitting radionuclides have also been tested. For example, 68Ga has the advantage that it can easily be produced with a 68Ge-68Ga generator, which makes the imaging independent of the availability of a cyclotron. The disadvantage of 68Ga is the shorter half-life compared to 18F (68 min vs. 110 min), while uptake of the nitroimidazole-based tracers takes hours. These differently labeled tracers have not (yet) proved to be advantageous over 18F-FMISO.
Imaging of Nitroimidazole-Based Tracers
In general, PET imaging is performed 2–4 h after the administration of the 18F-labeled nitroimidazole-based tracer [33]. This period between administration and imaging is required for sufficient uptake and accumulation of the tracer in the hypoxic tissue. With 18F having a half-life of 110 min, the signal increase due to higher levels of the tracer in hypoxic tissue is partly counteracted by the decay leading to rapid lowering of the PET signal. The detected signal is also influenced by blood perfusion, the distance of passive diffusion, and the image acquisition protocol. Hypoxic tumors are often hypoperfused, which can lead to underestimation of the hypoxia based on imaging of the tracer uptake [10], while imaging within a short interval after administration leads to overestimation from perfusion artifacts in a well-perfused tumor.
Ideally, a parameter representing the pO2 would be calculated from the acquired data. Unfortunately, this is methodologically challenging due to the abovementioned confounders. Therefore, in general the SUV, T/B, and tumor-to-muscle ratios (T/M) are applied for analysis. Alternatively, the hypoxic volume and the hypoxic fraction can be used as parameters for the quantification of hypoxia [36]. The hypoxic fraction is the hypoxic volume, defined as the tumor volume with T/B or T/M above a certain threshold (e.g., >1.2 to >1.4 for 18F-FMISO [32, 37]), divided by the total tumor volume. Some groups use more demanding pharmacokinetic modeling methods to improve quantification by maximal correction for perfusion; however, this is a procedure where the total imaging time is substantially longer, which is cumbersome for patients [38].
