PATTERNS AND VARIABILITY OF TUMOR OXYGENATION IN HUMAN SOFT-TISSUE SARCOMAS, CERVICAL CARCINOMAS, AND LYMPH-NODE METASTASES

Purpose: The validity of tumor pO(2) measurement as a predictive outcome assay depends upon demonstrating that intrapatient pO(2) variation is less than interpatient variation. No consensus exists regarding the appropriate distance between individual measurements. This distance could affect the calculation of the hypoxic fraction (% pO(2)s <5 mm Hg) and the assessment of intra/interpatient heterogeneity. This study was performed to evaluate tumor oxygenation and to assess the effects of two different measurement intervals on pO(2) heterogeneity in three different sets of patients. Materials and Methods: Fifteen patients with soft tissue sarcoma, nine patients with cervical carcinoma, and eight patients with squamous carcinoma metastatic to lymph nodes underwent pretreatment polarographic pO(2) measurements. Two grossly distinct sites were studied in each tumor, and 2-3 linear tracks were measured at each site. Track lengths varied from 20-36 mm. Distance between measured points was either 0.7-0.8 mm or 0.4 mm. Mean pO(2), median pO(2), and hypoxic fraction were calculated for each track. Data for each patient were also averaged across all tracks obtained for that patient. Track-specific data were used to evaluate intrapatient variation. The range of average values for each patient was used to assess interpatient heterogeneity. The ratio of these measures provided an assessment of within- vs. between-patient heterogeneity. Results: The median number of pO(2) measurements/patient was 200 (range: 88-356). The average length of hypoxic regions varied from 4.5-5.6 mm. Median tumor pO(2)s for the cervix, lymph node, and sarcoma patients were 4.5 mm Hg, 12.6 mm Hg, and 18.0 mm Hg, respectively (p = 0.07). Median hypoxic fractions were 0.61, 0.36, and 0.31, respectively (p = 0.07). Intrapatient heterogeneity was less than interpatient heterogeneity for all parameters in all patients, except for mean pO(2) for the cervix patients measured at 0.7-mm increments (1.51). Assessment of oxygenation was not affected by the distance between samples. Conclusions: Heterogeneity of tumor oxygenation within tumors is less than that between tumors. Both 0.4 mm and 0.7-0.8 mm sampling increments provide similar data. Longer term follow-up of large numbers of uniformly treated patients is required to define the value of tumor oxygen measurement as a predictor of treatment outcome.