Theoretical simulation of oxygen tension measurement in the tissue using a microelectrode: II. Simulated measurements in tissues.

BACKGROUND AND PURPOSE The objectives of this study were to make a computer simulation of tissues with different vascular structures and to simulate measurements of oxygen tension using an Eppendorf-like electrode in these tissues and to compare the response to radiation of the tissues with the real oxygen distributions (called input distribution) with the response to radiation of the tissues in which the oxygen distribution is given by the results of the simulated measurements (called output distribution). MATERIALS AND METHODS The structure of various tissues and the measurements of oxygen tension using a microelectrode were simulated using a computer program. The mathematical model used combines the description of a gradient of tissue oxygenation and the electrode absorption process. RESULTS We have compared the oxygen distributions resulting from diffusion (input) with those obtained from a simulation of measurements (output) for various tissues in the same points. Because the electrode measurement is an averaging process, the calculated oxygen distributions are different from the expected ones and the extreme high and low values are not detected. We have then calculated the survival curves describing the response to radiation if there is a small fraction of truly hypoxic cells (expected values) or a large fraction of cells at intermediate values (observed results) in order to determine the differences between them. CONCLUSIONS The results of our study show that oxygen electrode measurements do not give the true distribution of pO(2) values in the tissue. However, our results do not contradict the numerous empirical correlations between the Eppendorf measurements of tumour oxygenation and the outcome of treatments. Measurement results will be misleading for modelling purposes since they do not reflect the actual distributions of oxygen tensions in the measured tissue. Decisions based on such modelling could be very dangerous, especially with respect to the clinical response of tumours to new treatments.