Predicting the effect of temporal variations in PO2 on tumor radiosensitivity.

PURPOSE

Tumor hypoxia is associated with less effective radiation-mediated cell killing, increased metastatic potential, and poorer prognosis. Transient variations in hypoxia, with characteristic periodicity on the order of 1 to 10 min, have been observed in animal models. This article explores the effect of these temporal variations in PO(2) on the oxygen enhancement ratio, effective radiation dose to the tumor, and tumor control probability.

METHODS AND MATERIALS

PO(2) over a 50-60 min period was determined at multiple sites in rat fibrosarcomas, 9L gliomas, and R3230Ac mammary adenocarcinomas. Using a correlation derived from the data of Elkind et al. (1965), PO(2) data are converted into oxygen enhancement ratios (OERs.) A tumor is assumed to consist of 10(3)-10(4) independent oxygenation subvolumes, each with a randomly chosen starting point on the OER-time curve. The effect of temporal variations in OER is examined for three cases: conventionally fractionated external beam radiotherapy (EBRT), stereotactic radiosurgery (SRS) and intraoperative radiotherapy (IORT). The oxygen effective dose (OED) for a subvolume is calculated from the dose to that subvolume modified by the OER. In turn, the distribution of OED for a tumor is analyzed for each treatment case and representative tumor control probabilities (TCPs) calculated.

RESULTS

Oxygen enhancement ratio varied from 1 to 3 over the range of PO(2) measured in this study. Mean OER ranged from 1.6 to 2.6, and the variation in OER vs. time was greater with decreasing PO(2). In EBRT, the standard deviation in OED was small, <2%. In contrast, the standard deviation in OED was much higher for both SRS and IORT, typically ranging from 3 to 6%, with the greatest variation at the lowest PO(2)s. Compared with a tumor with equal mean OED and uniform PO(2), TCP was minimally poorer for either EBRT or well-oxygenated tumors. However, for both SRS and IORT, temporal variations in more hypoxic tumors can produce a significant decrease in TCP.

CONCLUSION

Temporal variations in tumor PO(2) can produce significant variations OER, particularly at low PO(2), resulting in decreased TCP for hypofractionated treatment regimens.