Deviations from the inverse relationship between tumor and normal tissue dose and RBE in fast neutron data are compatible with the presence of hypoxia, with possible applications to hypo-fractionated proton therapy.

Aim of study: Fast neutron data from Hammersmith (UK) showed that the inverse relationship between dose and RBE (Relative Biological Effectiveness) is not maintained in many Methods: A new model uses LQ formulism with radiosensitivity changes increasing with LET and with LET-related reductions in hypoxia. Inequalities are defined for the parameter ratios used to estimate RBEs in oxic and hypoxic compartments, with the overall RBE being also determined by the initial hypoxic fraction and the compartmental surviving fractions with increasing dose. Extrapolation from neutrons to protons is assumed by using a 90% RBE reduction for the excess of RBE beyond unity. Also, the separate influence of the dose rate effect on RBE at higher doses is investigated. Results: For normal skin, the C3H mouse mammary carcinoma and the Ehrlich ascites cancer, three quite different RBE responses are fitted by the model within practical statistical bounds. The overall RBE initially follows the oxic curve but shifts with increasing dose to the higher hypoxic RBE curve, the response depending mostly on the pretreatment hypoxic fraction and other hypoxia parameters. Such patterns of RBE changes are also obtained for protons, but cannot be obtained from the dose rate effect at higher dose, although the DNA repair rate will increase RBE with dose. Conclusions: This tentative but complex radiobiological modelling study shows that the Hammersmith neutron RBE results can be simulated by the presence of two physiological compartments and that similar RBE changes may occur for protons. Further experimental research is suggested for determining if protons do show effects at similar dose ranges, which could be relevant for hypo-fractionated or single fraction schedules.