On the stochastic interpretation of cell survival curves