Elusive Goal of Targeting Tumor Hypoxia for Therapeutic Gain

Abstract

In the article that accompanies this editorial, Janssens et al report the outcome results of a phase III trial of accelerated radiotherapy with carbogen and nicotinamide (ARCON) versus accelerated radiotherapy alone in the treatment of 345 patients with stages T2 to T4, any nodal stage, M0 laryngeal cancer. Despite the sound biologic rationale for accelerated radiotherapy plus carbogen and nicotinamide (ARCON) and the promising results from a large preceding nonrandomized study, no evidence of a therapeutic gain was seen in the definitive phase III trial. Specifically, no significant differences were observed in laryngeal preservation rates, disease-free survival, overall survival, treatment-related toxicity, or functional outcomes. The authors did note an improved regional nodal complete response rate in the ARCON arm as assessed clinically, but this did not translate into any improvement in disease-free survival after “salvage” neck dissection. Unfortunately, both the article and the protocol are unclear as to whether a neck dissection was stipulated for patients who failed to achieve a regional complete response by a certain time after treatment. If this was the case, the required neck dissection would normally be considered part of the definitive treatment package, and patients undergoing surgery would not be scored as having sustained a regional failure. In a translational research substudy involving 79 patients, the investigators estimated the hypoxic fraction of tumor cells in biopsies of the primary tumor by using pimonidazole staining. They reported that regional (but not local) control was significantly improved by treatment with ARCON in patients with more hypoxic primary tumors (defined as 2.6% pimonidazole staining) but not in patients with better oxygenated tumors. However, this conclusion is based on small numbers of neck failures and appears to take no account of the original neck node classification of these patients. It can therefore be considered no more than a hypothesisgenerating observation. Although these findings are consistent with the notion that any benefit from a hypoxia-targeting therapy will be restricted to the population with demonstrable hypoxia, an important caveat is that they did not actually assess hypoxia in the nodes. The hypoxia status of the primary and nodes has been reported to be discordant in studies that used hypoxic positron emission tomography (PET) imaging. ARCON is the latest in the growing list of interventions targeting hypoxia that have failed to realize their predicted potential in the treatment of unselected populations of patients with head and neck cancers. Strategies thathavepreviouslybeensubjectedtorandomizedclinical trials includehyperbaricoxygen, hypoxiccell radiosensitizers(misonidazole, etanidazole, and nimorazole), fast neutron therapy, concurrent chemotherapy with mitomycin C, and the hypoxic cell cytotoxin tirapazamine. Although a recent meta-analysis of trials of hypoxic modification did show a small but significant survival benefit, only nimorazole has entered clinical practice, and then only in Denmark, in combination with radiation without chemotherapy. It is an axiom of rational targeted therapy that the intervention can be successful only if the target is present and cure limiting. In the case of therapies targeting hypoxia, this requires not only that tumors harbor hypoxic clonogenic cells before treatment, but also that effective reoxygenation does not occur during treatment. Another important consideration as to whether hypoxia is cure limiting, at least for oropharyngeal tumors, is a tumor’s human papillomavirus (HPV) status. Recent data suggest that HPV-associated oropharyngeal cancers have a similarly high incidence of hypoxia compared with HPVnegative head and neck cancers but nonetheless retain their favorable prognosis. Furthermore, there are data from exploratory analyses of three randomized trials of hypoxia-targeting therapies that suggest that any benefit from these therapies may be restricted to patients with HPV-negative tumors. Several strategies have been used to detect tumor hypoxia before treatment, albeit with inconsistent correlations with outcome. These include oxygen electrode measurements, histochemical markers, plasma markers, functional imaging, and gene expression profiling. Use of such tests may enlarge the patient population potentially able to benefit from hypoxiatargeted therapy. Retrospective analyses of the Danish Head and Neck Cancer Study (DAHANCA) 5 trial have suggested that some of these tests may be predictive of benefit from nimorazole, and we have previously reported that patients with PETimageable hypoxia may benefit from tirapazamine. Nevertheless, such approaches have not been incorporated into phase III trials, at least in part because of the lack of a validated, reproducible test of tumor hypoxia that has been demonstrated to be suitable for use in multi-institutional clinical trials. The efficiency of reoxygenation during treatment is more difficult to measure and interpret. There are limited data from three small [18F]-misonidazole PET imaging studies about the persistence of hypoxia during treatment and correlation with outcome. It is noteworthy that in the Trans Tasman Radiation Oncology Group JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 30 NUMBER 15 MAY 2