I’m not aware of any rule that says AA can only be granted if the surrogate biomarker(s) overall mean to eos change is stat sig for the whole dosed trial population vs. placebo.
I do believe it is very likely there will be surrogate biomarker and clinically statistical significant change from the pre-specified subgroup analysis.
But recall my chance of AA is low, although much higher at that than doc’s.
If AA is granted from pre-specified subgroup analysis, then the confirmatory trial should be enrolling by selecting for the responder subgroup(s), which it can be designed to do and initiated before AA.
As mentioned before, this strategy seems optimal because the worst that could happen is that AA may fail, but the already initiated P3 PM trial is likely to lead to approval with the responder subgroup label some 3 - 4 years after the AA decision point.
Market Data 
Markets 
AI
