I don’t agree with the rest of your statement other than that they were still blinded when they submitted their revision, but the reality is, with comments back and forth, you’d want to have it completed before unblinding, and they delayed data lock until they apparently were ready.
The interesting difference in your argument from the past is, and I take this as a concession, they were not unblinded because of their interim blinded analysis, but the reality is that they knew even before that that most patients had crossed over, and there would be issues with the placebo arm, because of that, and that they had a pseudoprogression issue.
I do not believe they needed the interim analysis to know those details at all. They had a compassionate arm trial to address some of those issues as well, in fact.
So yes, they can and are allowed to learn things while still blinded. This is the nature of adaptive trials, which the regulators approve of and even encourage with new, living and dynamic drugs that are only more fully understood when they are given to humans under the actual intended ultimate circumstances of illness or injury.
Market Data 
Markets 
AI
