Point taken my critique was clearly off target in terms of censoring as you pointed out patients had to have PD'd to cross over. However analyzing a crossed over patient pretending it was the same as a de novo randomized patients violate the principle of randomization due to selection bias. I highly doubt that was in the apriori sap. Cheers!
I do agree with him it could potentially bias treatment sample because it added more "healthy" progressive disease patients from IC arm to 011 arm which is why the reported OS benefit is very important IMO DESPITE crossover.
Market Data 
Markets 
AI
