The "new" primary endpoint is for naive OS. That means they were in the original DCVax-L arm and many of those had psuedo-progression or actual progression and were given the another rescue DCVax-L dose. So many of the original naive GBM treatment patients got a second dose and may have responded. That confounds the ability to read how a single dose of DCVax-L in naive responded.
The external comparator trials were based on standard dosing and DCVax-L had much of their sample receiving multiple doses -- not a 'well-controlled" clinical trial if you have patients get extra doses. That is the confounding and the FDA will clearly understand the issue and the questions it raises.