Background: The concentration of osteopontin (SPP1) in plasma is associated with tumour hypoxia. The DAHANCA 5 trial found that the hypoxia radiosensitiser nimorazole significantly improved the outcome of radiotherapy for patients with head and neck cancer compared with placebo. However, whether all patients benefit from such modification of hypoxia is unclear. We aimed to assess whether the concentration of plasma osteopontin could predict response to the hypoxia radiosensitiser.
Methods: Plasma concentrations of osteopontin were measured by use of ELISA from stored samples of 320 patients randomised in the DAHANCA 5 trial. Samples were grouped into tertiles according to high (167-1382 microg/L), intermediate (69-166 microg/L), or low (0-68 microg/L) concentrations of plasma osteopontin, and analysed for locoregional tumour control and disease-specific survival at 5 years.
Findings: Overall, locoregional tumour failure and disease-specific mortality were more frequent in patients assigned placebo than in those assigned nimorazole (relative risk [RR] 0.51 [95% CI 0.32-0.79] and 0.54 [0.35-0.85], respectively). Locoregional tumour failure was more frequent in patients with high concentrations of osteopontin assigned placebo than in those with high concentrations assigned nimorazole (0.19 [0.08-0.44]), as was disease-specific mortality (0.25 [0.11-0.59]). However, neither locoregional tumour failure nor disease-specific mortality differed between groups for patients with low concentrations of plasma osteopontin (0.79 [0.26-1.70]) and (0.69 [0.31-1.51]) or for those with intermediate concentrations (0.90 [0.41-1.98] and 0.89 [0.41-1.96], respectively).
Interpretation: High plasma concentrations of osteopontin are associated with a poor outlook after radiotherapy for patients with head and neck cancer, but can be improved by use of nimorazole. High concentrations of osteopontin can predict clinically relevant hypoxia, and might identify patients who will benefit from modification of hypoxia during radiotherapy.