[en] The lack of a reliable, and preferably mechanistic, method to predict radiosensitivity parameters impedes ongoing efforts to develop biologically optimal radiation therapy treatments. The development and application of predictive assays is also hampered by the apparent lack of a systematic relationship between in vitro and in vivo intrinsic radiosensitivity. As a first step towards better quantifying the relationship between in vitro and in vivo radiosensitivity, we re-analyzed a large number of datasets using the Linear Quadratic (LQ) survival model with and without the generalized Lea-Catcheside dose protraction factor G. With G = 1 (no correction for dose rate effects), the analysis of cell survival data for acute irradiation conditions (dose rate > 45 Gy h^-1) yields α/β ratios of 8.17 and 8.33 Gy for CHO 10B2 and C3H 10T1/2 cells, respectively. When the dose protraction factor is computed using a 2 hour repair half-time, reanalysis of these same datasets gives an α/βof 7.2 Gy for CHO 10B2 cells and 7.6 Gy for C3H 10T1/2 cells. For repair half-times from 0.2 to 5 h (G ∼ 0.65-0.99), α/β ranges from 1.8 to 7.8 Gy for CHO 10B2 cells and from 3.2 to 8.0 Gy for C3H 10T1/2 cells. Global fits to low, intermediate, and high-dose-rate survival data indicate that the optimal α/β is 1.96 and 7.61 Gy for CHO 10B2 and C3H 10T1/2 cells, respectively. These studies demonstrate that the α/β cannot be reliably estimated from survival data for acute irradiation conditions. To better quantify intrinsic radiosensitivity, both low and high dose rate survival data are needed. Regardless, the analysis of cell survival data should always use the LQ with corrections for dose rate effects. A range of repair half-times should also be used to help estimate uncertainties in α , β, and the α/β ratio