No one “failed”, they ran out the adaptive trial a long time because they knew they OS numbers were super. They knew they would be short placebo patients because patients virtually all crossed over, there was so much interest in getting what was perceived as a very successful drug. And as I said elsewhere, with survival, it’s a concrete and true endpoint rather than a surrogate attempted prediction of survival, which is what PFS often falsely is, ignoring pseudoresponse, which it often ends up being.
So it’s good they got OS numbers, numbers that the regulators not only prefer, because they are concrete, real and the most preferred gold standard measures of success, but because survival for a deadly disease is virtually impossible to refute.