Switching the endpoint after an IA had already performed a formal efficacy analysis is a huge issue. Does anybody in their right mind think that had the IA said the trial had succeeded NWBO would have changed the endpoints?
Changing the endpoint from OS in a randomized trial to OS against host summary data from other trials is a huge deal..
Though I cannot prove it, they fairly clearly were performing additional IAs to see if the randomized OS worked. What does anybody think they were doing when they said they were waiting for 2 final OS events to unblind, then said they were waiting on more events? And before somebody notes thetye were blinded, efficacy IAs are commonly carried out while still blinded, and knowledge of the results does not unblind.
All these actions are a huge deal. That the trial was still formally blinded means absolutely nothing. What matters is that the change was not driven by knowledge of what happened in the trial, and even the 72 authors admitted the OS change from randomized to ECA was driven by trial results.
Market Data 
Markets 
AI
