Association between elevated serum VEGF and polarographically measured tumor hypoxia in carcinomas of the head and neck

Purpose: Clinical investigation of a potential relationship between VEGF concentration in serum (sVEGF) and polarographically measured tumor oxygenation in patients with squamous cell carcinoma of the head and neck (SCCHN). Patients and Methods: In 56 patients with SCCHN we estimated the classical tumor parameters, the sVEGF concentration (immunoassay) and the tumor oxygenation (EppendorF pO(2) histograph). The platelet count and the tumor volume were evaluated simultaneously. Results: In a unifactorial analysis the total volume (132 cm(3) vs. 38 cm(3)), the hypoxic subvolume (HSV a total volume multiplied with the relative frequency of values less than or equal to 5 mm Hg/63 cm(3) vs. 10 cm(3)) and the platelet count (380 10(9)/L vs. 271 109/L) were significantly higher in the patient group with a sVEGF level > 707 pg/ml compared to the group with a sVEGF below this threshold. The multifactorial analysis confirmed significant effects for the hypoxic subvolume and the platelet count. Regarding hypoxic subvolume and sVEGF as continuous parameters a significantly positive correlation was found. This correlation remained somewhat weaker but significant after inclusion of the platelet count as covariate. Conclusion: On base of our data a clinical association between elevated sVEGF and polarographically measured tumor hypoxia could be confirmed. This was possible considering not only the relative grade of hypoxia but also the absolute amount of hypoxic regions. The VEGF released from platelets during blood clotting influences the sVEGF level essentially, however, the hypoxia effect was not completely deleted. Due to the platelet effect an estimation of sVEGF is not able to substitute polarographical measurement of tumor pO(2). Therefore in an ongoing study we investigate whether VEGF values estimated in plasma are better correlated with the polarographically measured tumor pO(2) than serum VEGF levels.