Evaluation of isoeffect formulae for predicting radiation-induced lung damage

Abstract

An experiment has been performed in which fractionated irradiation was given to the whole thorax of Sprague-Dawley rats with schedules chosen so that doses per fraction and overall treatment time were changed independently. Damage was monitored by lethality. The data have been analyzed to yield dose per fraction and time parameters using multiple non-linear regression analysis. The results show that a linear-quadratic cell survival formula, extended to include an exponential time component to account for proliferation or slow repair during the treatment, can predict isoeffective doses to within 7% accuracy over a wide range of times (3.5–49 days) and doses per fraction (1.8–10.2 Gy). Other isoeffect formulae based on the linear-quadratic and empirical power law functions were also evaluated. A linear-quadratic formula with a time dependent alpha parameter fitted the data particularly well. This result suggests an alternative underlying mechanism and requires further investigation.