No. The FDA document on Adaptive Trials clearly states that it is a “guidance document and not a regulatory requirement.” How convenient you left out that important fact.
Furthermore, and most importantly, the primary purpose behind “prespecifying” the trial design (per the FDA document) is to avoid arriving at erroneous conclusions from the data. However, there is no way to fake, mathematically or otherwise, the extraordinary OS data.
Personally I think adding the external controls makes a stronger statistical case as to the potency of the result…..unless, of course, the external controls were cherry picked…..which I don’t think they were because those external controls are the same or worse than the Stupp data, which is exactly what you expect if Stupp is the SOC.