But yet it is the CROSSOVER itself that is confounding variable. Regardless of the comparator. They couldn't use the internal comparator even if they wanted to as they had virtually no sample as it was confounded by the CROSSOVER too.
NWBO says the OS data is confounded as it is the same endpoint it is just now the primary endpoint. NOTHING else changed. As NWBO stated:
"The secondary endpoint was overall survival, or OS...Both of these endpoints were confounded..the OS endpoint by the “crossover” provision in the trial design, which allowed all of the patients in the trial to cross over to DCVax-L treatment after tumor recurrence"
The naive OS was changed from the secondary endpoint (original protocol) to the primary endpoint ("new" protocol) -- which was the only change. It was still confounded by the crossover.
News 
Market Data 
Discover 
