Monday afternoon, December 10, 2012, the FDA announced it had expanded the label for Johnson & Johnson's (JNJ) Zytiga (abiraterone acetate) for late-stage prostate cancer into the pre-chemo space - a space best recognized as occupied by Dendreon's (DNDN) Provenge. (In fact, Zytiga already was being prescribed off-label here) Readers will recall I earlier had questioned whether or not the agency would dance to J&J's tune and approve the company's sought-after label expansion. In fact, using history as a guide, I thought the odds did not favor Zytiga approval in the pre-chemo space. I was wrong. But it's instructive to review the data upon which I drew my conclusion and compare it to the data cited in the FDA press release because somewhere along the way, 'a miracle occurred.'
As you will recall from last March, "[T]he Independent Data (Safety) Monitoring Committee (IDMC) overseeing the pivotal Phase 3 Zytiga trial (known as COU-AA-302) unanimously recommended unblinding the study based on a planned interim analysis in which differences in radiographic progression-free survival, or rPFS, overall survival, or OS, and secondary endpoints were observed that constitute evidence of clinical benefit as well as continued evidence of favorable safety in patients receiving abiraterone acetate plus prednisone as compared to those receiving placebo plus prednisone. Based on these results, the IDMC also recommended that patients in the placebo arm be offered treatment with Zytiga." In fact, Zytiga reduced the risk of prostate cancer worsening or death - the definition of progression-free survival - by 57% compared to a placebo. And the chances of the results being a chance occurrence were less than one-hundredth of 1%.
After an initial burst of enthusiasm for the drug by the medical community and the press - and a sharp 7% drop in the price of Dendreon's stock - some members of the media pressed JNJ for more details regarding one of the two primary endpoints: OS. As it turned out, the trial had not achieved statistical significance (stat sig) for overall survival. Here are the data:
rPFS: The data demonstrate a statistically significant improvement in rPFS in the abiraterone acetate plus prednisone arm (Zytiga arm) of the study compared to the placebo plus prednisone (control) arm. The median rPFS in the control arm was 8.3 months, but a median had not yet been reached in the Zytiga arm because progression events were occurring more slowly in the Zytiga arm compared to the control arm (N=150 vs. 251, respectively). The Hazard Ratio (HR) equaled 0.43, there was a 95% confidence interval (CI): [0.35, 0.52], and the p-value was <0.0001.
OS: Treatment with Zytiga plus prednisone resulted in an estimated 33 percent improvement in survival (median overall survival in the Zytiga arm was not reached and was 27.2 months in the control arm; HR=0.75; 95% CI: [0.61, 0.93], p=0.0097). At the time of these interim analyses, the pre-specified p-value of 0.0008 to achieve statistical significance was not reached. With the strict requirement of achieving less than 0.04 at the final look, there is the possibility the trial still could achieve stat sig OS. (According to a Commenter on Pharmalot, the final alpha (upper bound for p-value to be considered successful) for the '302' trial is 0.0389.) However, with patients in the placebo arm now crossing over and taking Zytiga, the ability to demonstrate stat sig OS was problematic.
Regardless of whether or not the trial should have been stopped, several highly respected people in the industry called it a failure. Among them was David Miller, writing for Minyanville, who stated: "Well, the drug certainly has a benefit for patients. But the AA-302 trial was a failure. Don't take my word for it. Look up the presentations on the American Society of Clinical Oncology-ASCO-Web site (you may have to pay) and listen to [Duke University's] Dr. Susan Halabi's discussion of the data. She called it a failure, too."
What were they talking about? If you take a look at around the 15-minute mark of the Zytiga ASCO presentation, you will see one of the strangest Kaplan-Meier analyses I have ever seen. The treatment and placebo arms of the study do NOT separate for nearly 18 months! Here's the critical graphic (courtesy Janssen Biotech, Inc.)
One might posit the bizarre behavior of the Zytiga K-M analysis can be explained by a very long median PFS and OS for the patients in the control arm. These indicators suggest the patients enrolled in the trial must have been very healthy. So, it would seem plausible that the long period taken for the survival curves to separate may mean Zytiga provided no benefit to them while their health remained good.
Eventually, the curves for the Zytiga Phase 3 K-M analysis did separate, but they did so beyond the point when patients got off the treatment (14 months). Specifically, no survival effect was seen until month 18…4 months beyond the point where the patients had been taken off the drug (and prednisone).
Of some concern, too, is the fact that 69% of patients in the treatment arm (Zytiga + prednisone) discontinued treatment at the time the trial was stopped (versus 84% of patients in the placebo + prednisone arm). Some ASCO participants found these results 'surprising,' though David Miller of Biotech Stock Research (personal communication, June 4, 2012) do not think they were out of line for a trial of this nature.
So, I asked, what will the FDA do? In the absence of statistically significant OS data, will Dr. Richard Pazdur, ODAC Chief and "Cancer Czar" of the FDA, compromise his principles, which require strict adherence to statistical rigidity and which define overall survival as the 'gold standard' for the approval of cancer drugs? Or will he and his panel hold their noses, look the other way, and dance to JNJ's tune, giving the nod to Zytiga for pre-chemo use?
The answer, obviously, was the FDA approved an expanded label for Zytiga. But - and here's the interesting part - according to the agency's press release: "[P]atients who received Zytiga had a median overall survival of 35.3 months compared with 30.1 months for those receiving the placebo."
Now, from where did these data derive? The data from the K-M curves above, which are the same data presented at this year's ASCO conference, clearly show a median overall survival in the control arm of 27.2 months. And again, there was no determination for the median OS benefit in the treatment arm.
Another interesting question is: what happened to the much heralded median 9-month OS benefit that was being touted by JNJ at ASCO, among other venues, to the detriment of Dendreon? For those with short memories, here's what J&J said earlier this year: "The trial also found that prednisone-only patients lived for a median of 27.2 months. Overall survival for patients treated with Zytiga had also not yet been reached, but J&J estimated that it improved their survival by 33 percent, or 9 months." This is more than twice the median 4.1 months OS benefit for Provenge, and so, is it any wonder Dendreon's stock was knocked down 7% upon the presentation of the Zytiga results at ASCO?
Alas, it wasn't to be. Zytiga is a good drug…it has much to recommend it in the pre-chemo space, especially where a man's disease load is increasing rapidly. But trial data do NOT support a median 9-month overall survival benefit. In fact, the benefit, at a median 5.2 months, is only 1.1 months greater than that of Provenge. And remember, no overall survival benefit was seen in the '302' study until 4 months after patients in the treatment arm had completed 14 months on Zytiga (at a cost of $5,495 per month, not including prednisone and monthly blood tests to monitor liver functions). In the pivotal Provenge Phase 3 studies, the K-M curves separated after 6 months.
As I said, somewhere between the unblinding of JNJ's '302' study and the approval of Zytiga for pre-chemo use, 'a miracle occurred.'
Finally, it is both interesting and instructive to note what Michael Meyers, J&J's development team leader for Zytiga, said upon the FDA's announcement: "Provenge is approved for a similar use and could be used with Zytiga and in sequence." In this regard, the Journal of Clinical Oncology has this to say:
"The practical dilemma of the appropriate sequence of use of the two new noncytotoxic agents (sipuleucel-T and abiraterone) is being addressed by trials that are under development. For now, given the broader window of applicability of abiraterone and the longer time required to develop an immune response with sipuleucel-T, if both agents are to be used, it seems reasonable to administer sipuleucel-T first with Abiraterone after additional disease progression. Biomarkers to help define the optimal use of immunotherapy are needed."
The Daily chart, courtesy StockChart.com, shows DNDN gapping up on the opening, a result, perhaps, of word leaking out of the FDA to the effect Zytiga would be approved to treat men with late-stage (metastatic) castration-resistant prostate cancer prior to receiving chemotherapy. When the announcement finally came, the stock continued to climb before settling back and closing just above $5. While it seems counterintuitive, the approval was viewed as a positive for Dendreon because instead of the earlier, much touted (estimated) median 9-month OS benefit, Zytiga's median OS benefit is just 1.1 months higher than Provenge's. Further, Michael Meyers' comments about sequencing Provenge and Zytiga indicated growing recognition and acceptance within the industry that treatment regimens should employ the co-administration or sequencing of multiple treatments. The stock, which left behind a small gap just below $5, is overbought. The MACD is positive.
The Weekly data tell a similar story. The Relative Strength is rising from its Oversold condition while the MACD has just turned positive.
Disclosure: I am long DNDN.
Additional disclosure: I am long both DNDN and the May 2013 $3 PUTs. I am not a registered investment advisor and do not provide specific investment advice. The information contained herein is for informational purposes only. Nothing in this article should be taken as a solicitation to purchase or sell securities. Before buying or selling any stock, do your own research and reach your own conclusion. Investing involves risks, including loss of principal.