Adapting by calibration the sample size of a phase III trial on the basis of phase II data