1) How did they get away with changing the trial primamry after having unblinded about 1/6 of the patients? Would the FDA re-run the data w/o these events?
2) If I can count tics, the early unblinding after the 51'st patient was treated had 15 events. [Correction, it is 12 events] Not all that significant (OS HR was p=.15).
3) Where does the 42% OS gain come from? It's not on the poster for this data: