As a generalization, nonlinear regression has some advantages over transformations into linear form (Lineweaver Burk, Eadie Hofstee, etc.). However, as a means of finding a preliminary estimate of the IC50, the linear transforms are at least adequate in my opinion.