Fair pushback. If the hold was FDA-initiated and globally applied, the German enrollment pattern might just reflect site logistics rather than an independent regulatory judgment about equipoise.
The point about limited OS data at Dresden is also well taken since you'd need meaningful event accrual to draw the kind of conclusions Gary is describing, and one site over one year probably doesn't get you there. The JAMA paper not mentioning it is a reasonable evidentiary standard to hold this claim to.
I think the broader point that early signals may have existed in the data still has some plausibility, but the specific German intervention narrative is harder to defend without documentation.
Market Data 
Markets 
AI
