Tumour hypoxia confers poor prognosis in a wide range of solid tumours, due to an increased malignancy, increased likelihood of metastasis and treatment resistance. Poorly oxygenated tumours are resistant to both radiation therapy and chemotherapy. However, although the link between radiation therapy and hypoxia is well established in a range of clinical studies, evidence of its influence on chemotherapy response is lacking. In this study, a panel of human tumour-derived xenografts that have been characterised previously for in vivo response to a large series of anti-cancer agents, and have been found to show chemosensitivities that correlate strongly with the parent tumour, were used to address this issue. Immunohistochemistry was carried out on formalin-fixed, paraffin-embedded sections of xenograft samples to detect expression of the intrinsic hypoxia marker Glut-1 and adducts of the bioreductive hypoxia marker pimonidazole. Glut-1 scores correlated significantly with T/C values for CCNU sensitivity (r = 0.439, P = 0.036, n = 23) and showed a borderline significant correlation with dacarbazine T/C (r = 0.405, P = 0.076, n = 20). However, there was no correlation between both Glut-1 and pimonidazole scores and T/C obtained for the bioreductive drug mitomycin C. The use of human tumour-derived xenografts offers a potentially useful way of using archival material to determine the influence of hypoxia and other tumour-microenvironmental factors on chemosensitivity without the direct use of human subjects.