Estimation of transition doses for human glioblastoma, neuroblastoma and prostate cell lines using the linear-quadratic formalism

Purpose: The introduction of stereotactic radiotherapy has raised concerns regarding the use of the linear-quadratic (LQ) model for predicting radiation response for large fractional doses. To partly address this issue, a transition dose D* below which the LQ model retains its predictive strength has been proposed. Estimates of D* which depends on the a, β, and D0 parameters are much lower than fractional doses typically encountered in stereotactic radiotherapy. D0, often referred to as the final slope of the cell survival curve, is thought to be constant. In vitro cell survival curves generally extend over the first few logs of cell killing, where D0-values derived from the multi-target formalism may be overestimated and can lead to low transition doses. Methods: D0-values were calculated from first principles for each decade of cell killing, using experimentally-determined a and β parameters for 17 human glioblastoma, neuroblastoma, and prostate cell lines, and corresponding transition doses were derived.Results: D0 was found to decrease exponentially with cell killing. Using D0-values at cell surviving fractions of the order of 10-10 yielded transition doses ~3-fold higher than those obtained from D0-values obtained from conventional approaches. D* was found to increase from 7.84 ± 0.56, 8.91 ± 1.20, and 6.55 ± 0.91 Gy to 26.84 ± 2.83, 23.95 ± 2.03, and 22.49 ± 2.31 Gy for the glioblastoma, neuroblastoma, and prostate cell lines, respectively. Conclusion: These findings suggest that the linear-quadratic formalism might be valid for estimating the effect of stereotactic radiotherapy with fractional doses in excess of 20 Gy