Also, one shouldn’t hold a single metric in isolation. For example, response showed a clear dose dependence from lowest to highest where the lowest was in line with placebo, middle dose showed response signals, and the highest administered dose was clearly superior to the placebo. Yes, the trial wasn’t powered as a pivotal study (sarcastic “no shit” here), but three points on a graph are sufficient to see a very promising dose dependent efficacy curve.
One can’t and shouldnt ignore all other data and only focus on the low subject numbers.