Walldiver-Thank you for the very informative post on yahoo. I will try this board for awhile instead.
As I further dig into provenge and the trials I’m becoming more and more convinced that approval has a good chance. There is still a lot I don’t understand.
In trial 9901 overall survival was not an endpoint, primary or secondary yet the patients were followed for 36 months, presumable to measure the survival rates. It’s not clear to me what difference it is to the FDA if survival was or was not an official endpoint. Survival after all is the ultimate endpoint. Perhaps patients are followed differently after progression, if survival is an official endpoint in the trial protocol. I’m trying to understand how important the FDA will view this non-official endpoint for trial 9901 and how the survival rates could have been any different?
Trial 9902a seems to be under-powered and unbalanced to known prognostic factors. Both to the detriment of the trial. The decrease in p-value=.33 from log rank calculation to .02 by cox regression is dramatic, yet the analysis was pre-specified. What is dramatic is how many patients with poor prognosis ended up in the provenge arm while so may good prognosis patients ended in the placebo arm. It seems that there should have been some attempt to randomize based on prognostic factors (by Gleason score perhaps). And I guess that’s the rational for trial 9902b. In any case it seems that the pre-specified cox regression analysis precludes any charge of data-mining or retrospective analysis. And the results are what they are. As far as you know, has there has been any fallout after presentation of these data suggesting inappropriate analysis? The bottom line is that trial 9902a does show that there is beneficial activity with the treatment of provenge, and that will support the findings in trial 9901.
I have yet to research the phase 2 trials, but I suppose there is supporting data (perhaps compelling data) there as well.