Simon and Maitournam
[24] and Maitournam and Simon [25] considered a phase III randomized
trial setting to evaluate molecular targeted treatment using a binary
endpoint or continuous endpoint. The evaluations were done under the
assumption that the assay was predictive of a subset of patients who
benefited from the targeted therapy. They compared the all-comers
design where all the patients are allowed to participate in the trial
with the enrichment design with restricted subset population. They
compared the operating characteristics based on average screened and
randomized sample size.
For a binary endpoint, the relative efficiency of the targeted design versus the untargeted design can be expressed ...