I didn't miss anything as the original endpoint was a statistical end point estimate, but the OS data provided itself, and the proof is that DCVax-L works.
The original endpoints were PFS and OS between the randomized arm. Both failed per the disclosed results.
I have no idea what you are talking about on statical estimate. Technically all the endpoints (original and modified) are statistical estimates of the actual clinical results seen. No difference at all between OS and PFS in this regard.