Based on a previous study,16 we estimated that a sample
size of 257 patients per group in each study would
provide 93% power (two-sided α of 0·05) to detect an
absolute diff erence of 15% in complete responses in the
delayed phase between the rolapitant and active control
groups, assuming 50% of patients achieved complete
responses in the active control group. The primary
population for the analysis of effi cacy results was a
modifi ed intention-to-treat population, which included
patients who received at least one dose of study drug at a
GCP-compliant site.