The article by Chen et al1– (see page 1031) is highly instructive from a number of perspectives. First and foremost, the authors (and the reviewers) failed to emphasize that this is a randomized phase II trial and not a randomized phase III trial. Phase III trials are true comparison trials and are sized so that clear inferences can be made about the superiority (or lack thereof) of any of the study arms tested. Randomized phase II trials are designed to choose the best option from among several with similar expected outcomes and to test the feasibility of randomizing across different modalities. They are, in reality, parallel phase II trials. The required numbers for these trials are far smaller than phase III trials. Their major use is in eliminating the problems inherent in interpreting the outcome of multiple phase II trials done at separate institutions. It is not statistically valid to draw conclusions based on comparisons between the study arms.2–3 Consequently, the inter-arm comparisons in this trial are invalid, although an actual phase III trial confirmed these findings.4 Despite that, there are several useful lessons from this trial.