Purpose: To investigate the dependence of the alpha/beta ratio determined from in vitro survival curves on the dose ranges.
Methods: Detailed clonogenic cell survival experiments were used to determine the least squares estimators for the linear quadratic model for different dose ranges. The cell lines used were CHO AA8, a Chinese hamster fibroblast cell line; U-373 MG, a human glioblastoma cell line; and CP3 and DU-145, two human prostate carcinoma cell lines. The alpha, beta, and alpha/beta ratio behaviors, combined with a goodness-of-fit analysis and Monte Carlo simulation of the experiments, were assessed within different dose regions.
Results: Including data from the low-dose region has a significant influence on the determination of the alpha, beta, and alpha/beta ratio from in vitro survival curve data. In this region, the values are poorly determined and have significant variability. The mid-dose region is characterized by more precise and stable values and is in agreement with the linear quadratic model. The high-dose region shows relatively small statistical error in the fitted parameters but the goodness-of-fit and Monte Carlo analyses showed poor quality fits.
Conclusion: The dependence of the fitted alpha and beta on the dose range has an impact on the alpha/beta ratio determined from the survival data. The low-dose region had a significant influence that could be a result of a strong linear, rather than quadratic, component, hypersensitivity, and adaptive responses. This dose dependence should be interpreted as a caution against using inadequate in vitro cell survival data for alpha/beta ratio determination.