What do you mean by "compare?"
Also, you have two directions to answer the questions once you decide what comparison makes sense. The first approach is the extract a scalar value, like ED50, You won't be able to apply statistical inference, though, unless you have more than one independent curve for each sample (drug or control). Fit Curve as @Thierry_S suggested provides methods to directly compare parameter estimates or test equivalence. The second approach is to use the entire curve. The Functional Data Explorer will convert each curve into a set of 'functional principle components' that represent the salient features of the curves and can be used as a new response variable. You can them use it with the drug and control categorical variable to compare.
