1. The true status of the CMC issues can be easiiy cleared up by publishing a redacted form 483 to remove "proprietary" items and reporting when the FDA has accepted the required written corrective action plan that would have restatrted the PDUFA clock. My guess continues to be that the nature of the 483 problems are those that even when viewed in February would require a follow-up unanounced FDA inspection that taking into account the time for FDA acceptance of a written corrective action plan and the 6 month resetting of the PDUFA clock at that time would have taken DNDN into 2008.
2. I think that croumagnon's response is closer to what may actually have occurred. DNDN could have appealed the FDA CRL with little impact on the much delayed schedule caused by agreeing so rapidly to await the interim survival results and even taken the FDA to court for procedural irregularities under Federal Administrative law. There is probably no other action that could have strengthned the position of Provenge supporters in the FDA and at the NCI, but even if it removed the requirement for further efficacy data, the spotlight would be shifted on the time to resolve CMC issues.
3. There would have also been the possibility of reopening the TTP issue if DNDN hadn't rushed to reaffirm the interim look for survival as the next regulatory event, as allowed in any event by the existing SPA. Under the new paradigm for clinical trials of cancer vaccines and immunotherapies, an initial ramp up period for an immune reaction to become effective(suggested at three months) is allowed, which would have made 9901TTP statistically significant, but was not allowed in 9901(retrospective and long since passed) and would not have helped 9902a. 9902b TTP data, allowing six months from randomization, should be fully mature in early 2008, as should be some early survival data. The Prentice criteria for surrogacy would require that TTP be stat significant, that it be reasonably likely that survival will ultimately also be stat significant at the trial's conclusion and that the cause for the increase in TTP be the cause for increased survival. It seems to me that causation requirement could be met simply by testing long term survivors of 9901 and 9902a for continued reactivity to PA 2024 (OK that still raises the human PAP vs.PA2024 reactivity issue). Since a FDA Accelerated Approval based on such analysis would be equivalent to the conditional approval recommended by the AC, and with a Federal Court looking over the FDA's shoulder, this process would have allowed DNDN to skip the more risky interim survival look with a reduced alpha to await 2010 final data while allowing commercial sales.
4. Since we have so many fine statisticians posting today, a question. Without digging deeply into the reference material, my recollection is that the Petrylak Halabi based analysis of Provenge + Taxotere was based on 81 patients in 9901 and 9902a who took Taxotere and reported that median survival increased by 14 months. Dr. Small's Halabi analysis reported in 10/06 was that the increased median survival of the 9901 was 5.8 months. 202 9901 and 9902a patients took Provenge either initially or on crossover. Putting aside for simplicity's sake any non crossver control patient who may only have had Taxotere, approximately 40% of 9901 and 9902a enrollees would have had Provenge + Taxotere. Simplistically multiplying that 40% by the 14 months increase from Petrylak's data comes to an average for all treated patients of 5.6 months increase in MS per Halabi, which, in tirn suggests that the combination treatment may have been overwhelmingly responsible for increased median survival. DNDN reported that the average time interval between the completion of Provenge treatment and Taxotere for those patients who took both was 4 to 6 months and a full course of Taxotere is a 7 month process, which seems to tie into the period when the survival curves begin to separate. The point of this line of reasoning is that if increased survival is largely the result of the combination therapy, shouldn't that hypothesis be tested in a clinical trial and potentially yield greater statistical significance than the 9902b trial where the use of Taxotere after Provenge in the ITT and control group is uncontrolled? The combination therapy also happens to be much more similar to MOA ideas of the NCI's Drs. Niederhuber and Rosenberg, where chemo is used to deplete Tregs, and of course with the suggested protocol of Dr. Petrylak.
5. BTW, one does not have to be bitter to be disgusted with DNDN's management and to want the very best therapy available to cancer patients as quickly as possible. Insistence on management transparency is hardly a radical idea.
