Replies to post #139562 on Biotech Values

[iwfal]

I think the recent ridaforolimus briefing docs indicated that technically they could have stopped at the 2nd interim but the DSMB elected not to. Paradoxically, the advantage of rida over the other arm at that point was 7 weeks - likely enough for approval.

Interesting find - which I confirmed in the FDA briefing docs:

Although the primary endpoint, PFS, technically met the boundary for the secondary interim analysis for efficacy, the overall survival estimates are premature, limited by a 40-week median follow-up time. Therefore, it is not yet clear that overall survival is supportive. The PFS advantage of 7 weeks is not of sufficient clinically meaningful benefit, especially considering the toxicity, and is insufficient to suggest early termination for efficacy without at least a positive trend on overall survival

But note that the meeting minutes aren't up yet - did the AC really focus in on the difference in medians being only 3 weeks in the final look? Or was it the lack of material OS benefit that caused rejection by the AC? If it was the former the AC is seriously flawed - since the 3 week difference in median PFS was clearly an artifact of the schedule for progression checkups. 7 weeks is clearly more reflective of reality. And this is a clear example of why you should use HR, not medians. When the minutes are up it will be an interesting read - e.g. if it was the difference in medians that tripped up the AC why didn't the statistician say anything?