The Dendreon/Provenge Story

This summary owes credit to Ocyan, Walldiver, Wolfeman, Casey, P3, Clark, Batgirl, and many others. Please keep in mind this is a brief summary of an epic story that I believe is still in its early stages.

In Jan. 2000, DNDN started the enrollment of the first of two Phase 3 trials for Provenge in men with prostate cancer (PC) that had progressed after primary hormone deprivation therapy. Several months later, the second Phase 3 trial began enrollment with the same enrollment protocol as the first Phase 3. Provenge, which is also known as Sipuleucel–T, is an active immununotherapy. It contains mature, autologous antigen presenting cells (APCs). APCs are obtained from the patient via a standard leukapheresis procedure, which is a blood collection process that isolates a patient's white blood cell approximately two days before each scheduled infusion. The patient's APCs are then transported to a Dendreon manufacturing facility where they are co-cultured with a recombinant fusion protein containing prostatic acid phosphatase (PAP).

The activated, antigen-loaded APCs (now sipuleucel–T) are then delivered to the physician's office (infusion site) for infusion into the patient. Sipuleucel–T is then infused into the patient, where it can potentially stimulate a T cell response against prostate cancer cells. The process is performed three times over the course of a four-week period, upon which treatment is completed.

Side effects vary from slight dizziness to 48 hours of flu-like symptoms (fever/chills).

Men in this stage of PC live on average 15-19 months. The primary endpoint of the trials was time to disease progression (TTP). Enrollment in the first trial (9901) was completed in Sept. 2001 and preliminary results were available by mid-2002. Comparison of the Provenge treated group to the placebo group using the Kaplan-Meier method revealed a clinical benefit in the Provenge treated patients (p-value = 0.085) that approached but did not achieve the pre-specified primary endpoint of the study (p-value = 0.05). Further analysis of the data showed that patients with lower Gleason Scores (GS), the accepted measure of the aggressiveness of a patient's tumor, had significant delay in TTP. Where the median TTP for the group was 9 weeks, the treatment subgroup of patients with GS <= 7 had a median progression of 16 weeks, a 78% advantage, with a highly significant p-value of 0.002. Although DNDN gained important information concerning the effectiveness of Provenge, these retrospective analyses are not suitable for FDA approval. After discussions with the FDA, DNDN amended the protocol for the second phase 3 trial (9902) to enroll only patients with GS of 7 or less, the population shown to benefit most from Provenge treatment in the 9901 trial. The company planned for 9902, as amended, to serve as the pivotal trial for seeking marketing approval for Provenge. So as to maintain statistical clarity the 9902 trial (now 9902A) was stopped and further enrollment became a third separate trial 9902B. In June 2003, the FDA gave 9902B a Special Protocol Assessment (SPA), a signed contract indicating that the new trial would serve as the basis for a Biologics License Application (BLA). In Sept. 2003, DNDN was notified by the FDA of fast track status, meaning the company could file the BLA in sections as the data became available. Then something very interesting happened. After following the patients from 9901 for 3 years, the percentage of surviving patients treated with Provenge was 3-fold higher (P = 0.0046) than those who received placebo. The overall survival p value was a statistically significant 0.01. These results were independent of GS. The median survival benefit of 4.5 months was greater than that observed with any type of treatment in any published Phase 3 study in late-stage prostate cancer. A Cox regression analysis, which is a tool used by FDA statisticians to even out trial imbalances between arms, further strengthened 9901's p value to 0.002. It should also be noted that as the 9901 TTP matured the P-value for the entire data set dropped to a near miss of 0.052. These findings were significant mainly because survival is considered the "gold standard" for late-stage cancer drugs. TTP is used because it is considered a surrogate for survival and results can be finalized much earlier than waiting years for the survival data to mature. In 2005 the survival data from the shortened 9902A was released. The percentage of patients alive in the Provenge-treated group was substantially greater than the percentage of patients alive who received placebo. However, as expected from the smaller trial the results were not statistically significant (p=0.332). The results did however follow the same trend as 9901 (median survival of 19.0 months vs 15.7 months, 3-yr survival 32% vs 21%). Using the more powerful statistical tool of Cox regression that adjusts for prognostic factors known to influence survival, and using the same five stat sig prognostic factors from 9901's Cox regression analysis, the 9902A survival data met the criteria for statistical significance (P = 0.023). In addition, this same Cox analysis found a P-value of 0.0006 for the pooled data (9901 + 9902A). One of the key factors is that patients with bony lesions were more prevalent in the Provenge arm of both trials, but expecially 9902A. This may have worked against the drug.

Following this DNDN again obtained a Special Protocol Assessment (SPA) from the FDA to change the protocol of the still enrolling 9902B trial. The enrollment criteria were changed to allow all GS and some disease related pain. Furthermore, survival would be the primary endpoint with TTP as the secondary endpoint. The FDA also agreed that the Cox regression would be the primary method for data analysis.

Analysis of the trial data continued. An analysis of prostate-cancer-specific survival for the 9901 Phase 3 showed a median survival of 35.2 months for patients randomized to PROVENGE compared to 23.5 months for patients randomized to placebo, a difference of 11.7 months and a 50 percent reduction in prostate cancer-specific mortality (P = 0.002). In other words, 13 patients in the Provenge arm of 9901 died of causes unrelated to their prostate cancer, while only 4 patients in the placebo arm died of these unrelated causes.

In Nov. of 2006 DNDN announced preliminary results from their P-11 or PROTECT trial for earlier stage PC. There was a delay of approximately 27 percent (HR = 0.73) in the time to distant metastasis for patients randomized to receive PROVENGE compared to patients randomized to receive placebo. This study will continue to be followed.

In Jan. 2007, DNDN received notification from the FDA that the BLA for the completed trials was accepted and that it would receive priority review, meaning that the final FDA decision would be due by no later than May 15, 2007. The company presented their case before the Advisory Committee to the CBER arm of the FDA on March 29.

March 29^th was, for investors such as myself, one of the most dramatic days in memory. It began with the company’s presentation, included some moving testimony from prostate cancer survivors, and concluded with wild swings in sentiment; it appeared the committee would not vote in favor of efficacy, but the wording of the question was changed (to the question that should have been asked in the first place) and in a dramatic vote the committee followed their unanimous vote in favor of safety with a 13-4 vote in favor of efficacy. None of the four dissenters were immunologists.

The stock more than tripled by the open the following day and reached a high of 25 before the FDA decision. During that period several key events occurred. Three letters were written to The Cancer Letter, two by Dr.’s Scher and Hussain who were dissenters on the Panel, urging the FDA to ask for more data before approving Provenge. On May 9 the FDA sent the company a Complete Response Letter (CRL) asking for more data and the stock price plummeted. There also remained manufacturing issues, which led (along with a substantial stock sale by CEO Gold) to a lawsuit by board poster Rancherho that is still pending. The company claims these issues will be resolved and were not a factor in the CRL.

This action has led to a grassroots effort to overturn the decision, spearheaded by Kerry Donohue and his crew at Care to Live. Their heroic and exhaustive public information and legal campaign for Provenge has included protests, billboards, and a lawsuit that has challenged the FDA all the way to the top. Please click on the Care to Live link at the top of the IV DNDN page for more. You can also get thousands of hours of reading in at Cliffhucker’s provenge.blogspot link as well.

Since then we have waited for the interim analysis, which can form the basis of a new application for approval. On the day I am writing this the company has announced that the results of that interim analysis will be announced in October of ’08. Estimates vary widely on the expected success or failure of this look, although “better than 50 percent” has come from very knowledgeable board stat guys. The company claims that the interim “is reasonably powered for success.”

In addition, the company has initiated two new Phase 2 trials: NeoACT in earlier stage PC prior to surgery and ProACT to test different concentrations of the antigen in late stage PC.

So either the company will skyrocket again in two months or we will wait another year for the final results in the second half of next year. If it does hit, the drama will continue as the question will change from “will this company survive” to “will this company be the next big biotech?” Questions that will determine that include:

It appears that the drug works best with chemo…can chemo dosages be cut and the same synergy preserved?

How quickly, if at all, will the drug move to earlier stage use?

Will CEGE’s drug GVAX succeed? If so, will the two drugs work side by side, synergistically, or winner take all? Will another drug competitor emerge?

Will DNDN be able to expand their cassette technology to the treatment of other cancers?

I welcome comments on this summary, keeping in mind that the complete story is not my goal; we’ll leave that to Ming and Jerzee. Now that would be a collaboration worth waiting for!

-Frogg