Flip, I basically agree with you.
PFS very strong, full approval on such alone is reasonable to hope for.
If PFS is only modest (whatever exactly that means), they will need to conform with OS. If OS is already stat sig (case 1), then they get full approval then. If trending well (case 2), then AA with follow up. And yes, IS can be followed post unblinding.
My reason for considering case 2 very unlikely is the numbers. With this long of data in the book, they will need to thread a needle to hit case 2.
WRT how the FDA looks at OS data that is "claimed" to have been blurred by cross-over, they will not give any slack on this. There is no way to prove post hoc that the cross-over was the problem. So this trial needs to hit OS in the ITT groups, regardless of crossover status.
Enjoy the day and please do not reply :-)
P.S., to be clear, this is discussing spesifically your post of reigstrational pathways. I have other issues as I have stated elsewhere.
Market Data 
Markets 
AI
