fyi per page 14 of the 2016 ASM presentation, the number of subsequent immunotherapy patients was 96 (16% of total 597) and per the OS Forest plot, bavi plus D was better than control. Overall hazard ratio for both groups was .844. They didn't break out the groups for HR.
So assuming basically even split, approx 48 patients each arm. Enough to get some stat significance. And given the immunotherapy option was prob US based, better chance of follow up IMO.