Tumor hypoxia is important in radiotherapy, but how should we measure it?

Abstract

The issues surrounding tumor oxygenation and radiotherapy used to be simple. Hypoxic cells are radioresistant and are present in a large proportion of human solid tumors; in addition, overall tumor oxygenation (as determined by a commercially available oxygen electrode, commonly referred to as the Eppendorf electrode) correlates with response to radiotherapy. Because of its correlations with clinical outcomes (particularly in head-and-neck tumors), the Eppendorf electrode has become “the gold standard” against which all other techniques are to be compared. Another appealing feature of the Eppendorf electrode is its ability to provide a quantitative distribution of real oxygen levels in tumors. However, it is not without problems. First, it can be applied only to superficially accessible tumors. Second, it requires considerable operator experience for proper use (something that is rarely discussed in the literature), and in many instances, image guidance is required for proper placement of the electrode. Most problematic, however, is that it cannot distinguish between tumor tissue and necrosis, which can bias the relevant oxygenation values to spuriously low levels in tumors with extensive necrosis. Because of these problems with the Eppendorf electrode, there are a number of ongoing efforts to find alternative methods of estimating tumor oxygenation. Two of the main candidates that have been used clinically over the past 2 to 3 years are nitroimidazole hypoxia markers and the socalled “endogenous markers” of tumor oxygenation. The first of these requires injection of a compound (a nitroimidazole such as pimonidazole or EF5) some hours before tumor resection or biopsy for immunohistochemical detection of the bound nitroimidazole derivative that occurs only in hypoxic tissues. The second of the alternatives to the Eppendorf is the use of endogenous markers, or proteins induced by hypoxia, such as the transcription factor HIF-1 or genes transcribed by HIF-1, such as carbonic anhydrase 9 (CA9). The levels of both of these proteins are increased under hypoxic conditions, and both can be detected in tumor sections by immunohistochemistry. The huge advantage of this latter approach is that levels of these proteins can be assessed on archival materials, thereby allowing rapid correlation to treatment outcomes. In addition, this approach requires neither the injection of foreign material nor any additional invasive procedure beyond that of taking a tumor biopsy at diagnosis. The overriding goal of all of these studies is a simple, widely usable, and robust method of assessing tumor oxygenation that will correlate with clinical outcomes. But here is where it gets more complicated. First, there are at least two distinct forms of tumor hypoxia. The first of these, described by Thomlinson and Gray 50 years ago, is one in which hypoxic cells were postulated to occur beyond the diffusion distance of oxygen from blood vessels and adjacent to necrotic areas. This is now known as chronic hypoxia. The second is acute hypoxia. It has been shown both in experimental animals and in human tumors that tumor blood flow varies over time, and (at least in animal tumors) this gives rise to hypoxia caused by the temporary reduction in flow or closure of certain blood vessels within the tumor. It must be borne in mind, however, that acute and chronic hypoxia are likely to be the extremes of a continuum caused by the dynamic nature of tumor blood flow, which causes variations in the diffusion of oxygen in at least some of the tumor vessels. The reason why having two forms of tumor hypoxia is a complicating factor is because we do not know whether both of these forms of hypoxia are equally important for radiotherapy outcome or whether one is more important than the other. The second reason for the increasing complexity of the field is that the different methods of assessing tumor oxygenation measure acute and chronic hypoxia to different extents. In the case of the Eppendorf electrode, the mea-