Comparison between the comet assay and the oxygen microelectrode for measurement of tumor hypoxia

Background and purpose: Hypoxic cells are present in some solid tumours and are known to limit radiocurability. To compare two measures of tumour hypo xia, 25 patients with locally advanced disease and accessible tumours or me tastatic nodes were examined using an oxygen microelectrode and the alkalin e comet assay. Measurements and methods: For the comet assay, fine needle aspirate biopsie s were taken immediately following a dose of 5-10 Gy. Single cells were exa mined for radiation-induced DNA strand breaks, and the percentage of radio- resistant hypoxic cells within the population was calculated from DNA damag e histograms. For oxygen tension (pO(2)) measurements, multiple tracks were made using an Eppendorf oxygen microelectrode. The possibility that applic ation of the first method might influence hypoxic fraction measurement by t he second method was examined in a more controlled system by creating four tracks in murine SCC-VII tumours using an oxygen electrode, and measuring h ypoxic fraction at subsequent times. Results: For 28 tumours from 25 patients, hypoxic fraction measured by come t assay correlated with the percentage of pO(2) values < 5 mmHg (r(2) = 0.4 6, P < 0.001). The mean comet hypoxic fraction was 0.36 for five tumours wi th a median pO(2) < 10 mmHg. For the remaining 23 rumours with a median pO( 2) > 10 mmHg, the mean hypoxic fraction was 0.09. Advancement of an oxygen electrode through SCCVII tumours had no significant effect on hypoxic fract ion measured 5 min to 24 h later using the alkaline comet assay. Conclusions: Tumours defined as hypoxic based on a median pO(2) < 10 mmHg a ppear to contain more than 20% radio-biologically hypoxic cells as estimate d by the comet assay. In an animal tumour model, puncture of the tumour wit h an oxygen electrode did not influence hypoxic fraction measured using the comet assay, in agreement with the clinical data that the order in which t he two methods were performed was not important. (C) 1999 Elsevier Science Ireland Ltd. All rights reserved.