Equivalent Uniform Dose (EUD) has been proposed as a means of comparing treatment plans which include large PTV dose heterogeneity. Moreover, it has been suggested that EUD is insensitive to the underlying cell survival model parameters used in its calculation, and that arbitrary values might be used. This dependence of EUD on radiobiological parameters has been studied using different cell survival models (exponential and linear quadratic) and a range of artificially generated and clinical (lung and prostate) treatment plans. EUD was found to be strongly dependent on the variable SF2, but not on α/β, or number of fractions. When clinical treatment plans were ranked in order of EUD, a dependence of rank order on SF2 was observed, suggesting that arbitrary SF2 values are not appropriate for treatment plan evaluation. These results suggest that tumour specific SF2 values should be used for EUD calculation, and this is discussed.