I have been a strong supporter of Provenge. However, I also firmly believed FDA shouldn't approve Provenge back in 2007. It's very simple, FDA should NEVER approve a drug that failed predefined primary endpoint. Having worked in pharma/biotech for a long time, I can tell you if FDA opens this flood gate, every drug that failed primary endpoint can find post hoc analysis that it met other endpoint or met primary endpoint for a subset patients. The reason is very simple, every drug in phase III works for some patients. ...
With Provenge it was not a subset, or just 'some' other endpoint.
Had the 2002A(?) confirmed the OS, I expect that the FDA might well have given a thumbs up on efficacy.
As it was, the review presented to the panel implied the FDA staff were at least undecided.
The FDA decision on Provenge was very reasonable, especially with the CRC issue and 2002B well under way.
But I suspect the issue is still open. If a drug misses of a PR/CR type primary but CLEALRY hits on OS, it could get approved.