Background: The present study was conducted to analyze the value of ([18F] fluoromisonidazole (FMISO) and [18F]-2-fluoro-2'-deoxyglucose (FDG) PET as well as color pixel density (CPD) and tumor perfusion (TP) assessed by color duplex sonography (CDS) for determination of therapeutic relevant hypoxia. As a standard for measuring tissue oxygenation in human tumors, the invasive, computerized polarographic needle electrode system (pO2 histography) was used for comparing the different non invasive measurements.
Methods: Until now a total of 38 Patients with malignancies of the head and neck were examined. Tumor tissue pO2 was measured using a pO2-histograph. The needle electrode was placed CT-controlled in the tumor without general or local anesthesia. To assess the biological and clinical relevance of oxygenation measurement, the relative frequency of pO2 readings, with values < or = 2.5, < or = 5.0 and < or = 10.0 mmHg, as well as mean and median pO2 were stated. FMISO PET consisted of one static scan of the relevant region, performed 120 min after intravenous administration. FMISO tumor to muscle ratios (FMISOT/M) and tumor to blood ratios (FMISOT/B) were calculated. FDG PET of the lymph node metastases was performed 71 +/- 17 min after intravenous administration. To visualize as many vessels as possible by CDS, a contrast enhancer (Levovist, Schering Corp., Germany) was administered. Color pixel density (CPD) was defined as the ratio of colored to grey pixels in a region of interest. From CDS signals two parameters were extracted: color hue--defining velocity (v) and color area--defining perfused area (A). Signal intensity as a measure of tissue perfusion (TP) was quantified as follows: TP = vmean x Amean.
Results: In order to investigate the degree of linear association, we calculated the Pearson correlation coefficient. Slight (|r| > 0.4) to moderate (|r| > 0.6) correlation was found between the parameters of pO2 polarography (pO2 readings with values < or = 2.5, < or = 5.0 and < or = 10.0 mmHg, as well as median pO2), CPD and FMISOT/M. Only a slight correlation between TP and the fraction of pO2 values < or = 10.0 mmHg, median and mean pO2 could be detected. After exclusion of four outliers the absolute values of the Pearson correlation coefficients increased clearly. There was no relevant association between mean or maximum FDG uptake and the different polarographic- as well as the CDS parameters.
Conclusion: CDS and FMISO PET represent different approaches for estimation of therapy relevant tumor hypoxia. Each of these approaches is methodologically limited, making evaluation of clinical potential in prospective studies necessary.