This is the FDA's view, expressed in a letter to Sherman Frankel who made the same kind of point that you are making - "FDA does not require that a dose-ranging study demonstrate statistical significance between doses; showing a positive slope in the dose response curve, as did the Propecia study, is sufficient. You also object that there were no studies “down to 0.05 mgâ€. It is true that the dose-ranging study did not include a 0.05 mg dose. It did, however, include an even lower dose, 0.01 mg, in addition to 0.2 mg and 1 mg doses. The study showed the 0.01 mg dose to be ineffective, and the 1 mg dose to be more effective than the 0.2 mg dose and equally safe."
