>GNVC – Would it not also be the case that the KM curves very likely are crossed well into 12 months are so? And if so, that invalidates a log-rank test?<
I’m not sure what you mean by invalidates—please clarify.
>And the company elected to conceal the curve?<
Evidently they did, which is yet another reason to be skeptical.
>I doubt it matters how the final events play out, unless by some miracle enough early events come in favorable enough to change the picture (I have no idea if enrollment timeline even makes this possible).<
This speaks to the invalidation issue above. For the sake of discussion, let’s say the K-M curves at the final analysis cross early and the 12-18-month “bulge” favoring TNFerade is still seen. Under these admittedly implausible circumstances, I don’t see why the results would not come out statsig. Do you disagree?
Market Data 
Markets 
AI
