Differential risk assessments from five hypoxia specific assays: The basis for biologically adapted individualized radiotherapy in advanced head and neck cancer patients.

PURPOSE Hypoxia adversely relates with prognosis in human tumours. Five hypoxia specific predictive marker assays were compared and correlated with definitive radiotherapy. PATIENTS AND METHODS Sixty-seven patients with advanced head and neck carcinomas were studied for pre-treatment plasma osteopontin measured by ELISA, tumour oxygenation status using pO(2) needle electrodes and tumour osteopontin, hypoxia inducible factor 1alpha (HIF-1alpha) and carboxyanhydrase 9 (CA9) by immunohistochemistry. The primary treatment was radiotherapy and the hypoxic radiosensitizer nimorazole. Loco-regional tumour control was evaluated at 5 years. RESULTS All five markers showed inter-tumour variability. Inter-marker correlations were inconsistent. Only plasma osteopontin inversely correlated with median tumour pO(2), (p=0.02, r=0.28) and CA9 correlated with HIF-1alpha (p<0.01, r=0.45). In Kaplan-Meier analysis high plasma osteopontin, high HIF-1alpha and high proportion of tumour pO(2)2.5mmHg (HP(2.5)) related significantly with poorer loco-regional control, whereas CA9 and tumour osteopontin failed to predict loco-regional control in this set dataset. When analyzing Hb, stage, and the five markers by competing risks HP(2.5) was the strongest variable to predict for loco-regional tumour control. CONCLUSION There was diversity and lack of correlation among five different hypoxia assays within individual tumours. High plasma osteopontin, high HIF-1alpha and high proportion of tumour pO(2)2.5mmHg (HP(2.5)) related significantly with poorer loco-regional control, whereas CA9 and tumour OPN failed to predict local control.