I've received quite a number of emails on Dendreon
A lot has been written about the upcoming interim look at the efficacy data from Dendreon's phase 3 IMPACT study for its prostate cancer treatment Provenge. Many analysts, like me, don't think the odds are very high that this look (which will occur later in 2008) will be a success. The reason is that the FDA sets a very high hurdle that drugs have to overcome in an interim data peek.
For example, in Dendreon's D9901 phase 3 trial, the p-value hurdle that it had to beat for that study's interim look at its primary endpoint was 0.001 -- that means the survival data from the two treatment groups needed to be different enough to show that there was only a 0.1% chance that the results from patients receiving the drug were identical to the results from those not receiving the drug. Because of this interim look, the FDA penalized Dendreon on its final data reveal and made the final data p-value that it had to beat 0.049 instead of the customary 0.050 without an interim look. Dendreon hasn't said what the interim or final p-value hurdle is for the IMPACT study, but if the interim p-value hurdle is anything like 0.001, then it will be tough for Dendreon to hit this mark on the study's overall survival endpoint.
The sample size in IMPACT is much larger than the past two phase 3 studies for Provenge (D9901 and D9902A) combined, so that's the good news for Dendreon investors betting on a positive interim look (a larger sample makes hitting a particular p-value easier since the randomness gets evened out). Even combining the excellent overall survival Provenge results from its previous two phase 3 studies though, the p-value for overall survival in those studies only reached the 0.011 significance level, so you can see how much of a hurdle getting to 0.001 may be.
The good news is that if Dendreon's Provenge does show an improvement in overall survival at the interim IMPACT data peek later this year, and the results are relatively clean, then the FDA will likely approve Provenge extremely fast.
Even a failed interim look could still be a very positive and exciting sign for Provenge. If Provenge fails at the interim, just knowing that the interim look failed is not the most important issue, what matters is how closely it failed at the interim. If the interim data is showing that the Provenge-treated patients are trending strongly toward an improvement in their overall survival versus the control group, then the odds of the final 2009 IMPACT analysis being a success will have gone up. This was one of the things that got me excited about GPC Biotech
Investors should also remember that there are real risks involved with the interim look besides just failing to beat the tough efficacy hurdle. If Provenge were to perform poorly enough at the interim look, and the data monitoring committee calculated that there was little chance of the drug outperforming the control group, then the study could be scrapped at this point.
Unfortunately for investors, I don't think we will get to know how closely Provenge will fare at the interim because that info will probably remain confidential, aside from the top-line "interim failed" or "interim succeeded" overall survival data announcement. There are good reasons for not revealing the hard numbers to accompany these top-line results, but hopefully Dendreon will give us some sort of data nuggets to chomp on even if the interim study results are not good (or bad) enough to end the trial early.
