Head and neck cancer: The importance of oxygen

Objectives: To use recently introduced polarographic technology to characte rize the distribution of oxygenation in solid tumors, explore the differenc es between severe hypoxia and true necrosis, and evaluate the ability to pr edict treatment outcomes based on tumor oxygenation. Study Design: Prospect ive, nonrandomized trial of patients with advanced head and neck cancer, co nducted at an academic institution. Methods: A total of 63 patients underwe nt polarographic oxygen measurements of their tumors. Experiment 1 was desi gned to determine whether a gradient of oxygenation exists within tumors by ex amining several series of measurements in each tumor. Experiment 2 was an analysis of the difference in data variance incurred when comparing oxyg en measurements using oxygen electrodes of two different sizes. Experiment 3 compared the proportion of tumor necrosis to the proportion of very low ( less than or equal to 2.5 mm Hg) polarographic oxygen measurements. Experim ent 4 was designed to explore the correlation between oxygenation and treat ment outcomes after nonsurgical management. Results: No gradient of oxygena tion was found within cervical lymph node metastases from head and neck squ amous cell carcinomas (P > .9), Tumor measurements achieved with larger (17 mu m) electrodes displayed smaller variances than those obtained with smal ler (12 mu m) electrodes, although this difference failed to reach statisti cal significance (P = .60), There was no correlation between tumor necrosis and the proportion of very low (less than or equal to 2.5 mm Hg) oxygen me asurements. There was a nonsignificant trend toward poorer locoregional con trol and overall survival in hypoxic tumors. Conclusions: Hypoxia exists wi thin cervical lymph node metastases from head and neck squamous carcinomas, but the hypoxic regions are distributed essentially randomly. As expected, measurements of oxygen achieved with larger electrodes results in lowered variance, but with no change in overall tumor mean oxygen levels, Polarogra phic oxygen measurements are independent of tumor necrosis. Finally, oxygen ation as an independent variable is incapable of predicting prognosis, prob ably reflecting the multifactorial nature of the biological behavior of hea d and neck cancers.