Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
I just don't see them failing to achieve full enrollment within the next 6 months or so. There has been enormous publicity, is is now hard to believe anyone treating prostate cancer has not heard of Provenge.
Admittedly, the story is now mixed, but on balance, given the 17 - 0 panel vote on safety, I would think most doctors would reason that given the relatively mild side effect, there is little downside, unless you want to go straight to Taxotere, and most men don't, do they? And Petrylak's research suggests that if you are planning to go to Taxotere eventually, there is at least some evidence Provenge may help.
I don't see the FDA action as a strong suggestion to not to enroll in the trial, approvable just means, not proven yet, but taken together with the panel vote, doesn't this all add up to, there is some, maybe even substantial evidence of efficacy, the FDA choose not to approve because their standards are a bit higher. But do you think someone dying of prostate cancer really cares whether the p-value on the primary endpoint was .051 after adjustment or .049 or even, horrors, .083?
From an investment decision standpoint, I don't see much in this debate, as I don't see the stock popping if and when complete enrollment is announced. Of course, who knows what the stock will do, and who knows whether a failure to move if and when they announce complete enrollment means the market does not understand why this is good, or that the market assumed complete enrollment was a forgone conclusion. Perhaps you disagree, as you seem to think failure to enroll is a real possibility.
No, that is not exactly what you have been saying, not even close. You said to David:
"It's not clear to me why you continue to perpetuate the myth that the Provenge AC voted for (marketing) approval and the FDA overruled the panel. The AC panel most assuredly did NOT vote for approval."
In my opinion, they did vote for approval, they knew the vote was the standard for FDA approval, and they voted in favor of that, just not 13-4, rather, say 10 - 7. Some were confused, or perhaps more accurately, torn, in some cases because they felt compelled to answer the literal question (is there substantial evidence of efficacy) but yet apparently thought the FDA should not approve. But as I read the transcript of the votes, a clear majority thought the FDA should approve. In your opinion, as I understand it, they voting for an approvable letter, what was issued.
At this point, this is argument for the sake of argument, since this is history and I don't think it affects much going forward -- except as a counterexample to those who will claim after the next panel vote, that the FDA almost always follows the panel. Now there should be a big EXCEPT to that statement.
I need to move on to other stocks so I will try and wean myself off from posting on Dendreon, I am contributing to what I suspect is getting a little tiresome for others. But in response to your post, I would read or read the transcript of the vote at http://www.fda.gov/ohrms/dockets/ac/07/transcripts/2007-4291T1.pdf , start at pager 375. The "vote" was not 13-4, it may have techincally been that and recorded that way, but if you examine what was said, and whether panelists were really votiing for approval (i.e., should the drug be immediately available), it was much closer to even. Still favorable, but not 13 - 4. This is a point David Miller made in his newsletter, and I agree.
Some would cite the publication of Scher's and Hussain's letter in The Cancer Letter as evidence of a turf war. It certainly seems a bit of an odd way for panelists to trying to influence the FDA.
Back to the role of the panel, if you analogize to a jury trial, with the panel being the fact finders, typically there would be some detailed instructions from the judge (FDA) as to the proper standard. Reasonable doubt means.... or in a civil trial, "preponderance of the evidence means...". I don't see any of that in the FDA transcripts I have read, this is the first one where there was any discussion at all of the standard, and this was obviously done in a random and ad hoc manner.
I also wonder what the point of many of these panels is. If the questions are to be large determined based on matters of biostatistics, why not convene a panel of biostatisticians, rather than 17 random scientistics who know little on the subject. At least that would be more honest. And it might be more comprehensible in some ways to hear a real debate on the subject rather than trtying to understand what the FDA expert alone is saying.
Since "approvable" in this context means, wait for the results of the 9902B trial, I guess non-approval must mean, the drug is so worthless, don't bother completing the trial, and that panelists who voted no were voting to discontinue the 9902B trial. Somehow, I can't believe that is so. You are saying that there was never even a vote on approval, that the panelists were not offered the opportunity to vote on that question -- that it was withdrawn after the first three initial votes.
In fact, let's look at what Howard Scher thought he had voted on, as he explained his letter to the FDA published in TCL. It beings, "I am writing to express concerns about the recent review of Sipuleucel-T at the FDA Advisory Meeting on March 29. 2007. These concerns are: a recommendation for approval based on data that fall short of regulatory requirements..." Note the words, "A recommendation for approval".
Also The Cancer Letter, April 27, quotes Richard Schilsky, former chairman of the Oncologic Drugs Advisory Committee, and president elect of ASCO. His interpretation of the vote: "If the FDA follows the advice of the advisory committee (comprised largely of non-oncologists) and approves Provenge, they will establish a precedent for approval of drugs based on minimal evidence of effectiveness... If they reject the advice of their advisory committee, they will likely be criticized for doing so and for preventing patients from receiving a potentially beneficial therapy."
Moving to the actual language of the question, here is some of what the FDA said, taken from the transcript on the FDA website:
DR. WITTEN: Excuse me, Dr. Mulé? Yes. Maybe we should try to rephrase it as
-- I mean, the question is really asking for
you, you know, on the advisory committee, do
you believe that this product works, that
it's efficacious. I mean that's really what
we're asking. So if it's somehow some of
the words are not clear, that's what's
intended. We want to know whether you
believe, as individuals, that this works, that they've shown that it works."
Now somehow that is supposed to be interpreted, "we are asking you whether you think the drug should either be denied or merely deferred for a few years?" That seems unlikely to me. The precise regulatory standard the FDA then went on to quote, "substantial evidence," is, of course, the statutory standard whereby the agency can deny marketing approval if the standard is not met. It has nothing to do with whether the drug is "approvable."
I have a few difficulties with this approach, and maybe some of this is easier for professionals than for amateurs. However,
(i) the more complex the trade, the bigger the spreads. I don't recall the exact number, but the bid-ask on the vertical spread I put on was something like .90 - 1.40. I think this is because the spread in the options themselves are fairly wide -- .20 -- so when you execute a spread, it is twice as wide. And I think the spreads compared to the price increase as you move to the higher strike prices. I use thinkorswim and it allows you to put in a bid on a complex option position, but I can tell you that after a day waiting to for a limit order at the lower end of the spread to hit, I gave up, held my breath, sold the higher strike price option first, and then put in a limit order to buy the lower price call. It worked, in the sense of reducing the cost of the trade, but at the cost of risk -- the kind of risk that generally pays off, just as picking up nickels in front of a steamroller generally works (a description I read of the trading style of Long Term Capital).
(ii) I think the above is a function of the fact that the option makers cannot hedge their positions using the usual delta hedging tools. The spreads and high premiums reflect that inability to hedge, and you are paying all of those costs.
(iii) These positions trade one kind of risk for another. Now, in addition to judging the probability of the binary event, you also want to judge the post approval price. If you judge incorrectly, you can win the approval bet, but lose on the option position if the post approval price does not move as expected. Admittedly, if you own the stock, you are also betting on the stock price post approval as well, but you don't lose everything if the stock pops to just below the strike price on the option position you were considering, your return is just reduced a bit. In addition, In the case of DNDN, I played the May option, taking the chance that the FDA would stick to the PDUFA date. But that was a risk as well. I could have moved to the next strike, but that increases costs.
I think all of this is a useful exercise for investors because it forces you to think about the risks you are willing to take. I just think that in the end, it may make more sense to just own the stock, risking the amount you are willing to lose. And of course, by that, I mean the amount based on the stock price prior to the binary event, not the fool's examination of cost basis.
I'd add a couple of others:
6. The Company has a great relationship with the FDA, they are in constant communication, the FDA encouraged them to file -- well, I recall Genta claimed a great relationship with the FDA. I want an example of a company that acknowledges a crappy relationship with the FDA
7. The Company is hiring lots of people, they must know what they are doing. Well, people who are hired can be fired, and companies take lots of risks -- if they didn't, they would not make any money if things go well.
I agree cost basis is generally irrelvant to investment decisions, the contrary belief is often expressed in such fallacies as "I am only playing with the house's money" and seems to reflect a belief that the cash that could be obtained by selling a winning position is of less utility than the cash that comes from selling a break-even position. Put alternatively, if you would not establish the same position with cash from your bank account on any given day, you probably shouldn't maintain the position.
However, when facing a binary event, I don't think there is any free lunch. You cannot have the "very same" upside by risking $35,000 as by risking $125,000. In Dendreon's case, for example, before the FDA decision, the at-the-money call options for the binary event (i.e., May) were trading at about 6.5 or so, in the range of 1/3rd of the stock price. If I really could have all the upside of a $200,000 position with a 17% premium, I guess I would sell the at the money call for 33% and put on whatever magical trade gives me the upside with only 17% at risk. In my own case, I sold some 20 may calls and bought some 17.5 may calls for a spread of $1, in essence risking $1 to make $2.5; I judged the chance of approval rather greater than 40%. But I did not tell myself I had the same upside as holding the stock or buying the option outright.
With regard to your trading idea, there are those who want to believe that the "additional clinical data" requested is something other than 9902B. That may be holding up the stock, time will tell whether this is fantasy or not.
To be filed under the category of "silver linings", the FDA decision will help with enrollment and maintaining the validity of the 9902B trial (i.e., no worries about patients bolting the trial for the approved drug).
I understand that you were just trying to create a hypothetical, but usually, the endpoint is not survival at day X, but "overall survival" through day x, and usually the primary endpoint is the overall survival through the longest measurement period. The difference is this: treatment arm could show 50% survival at day 200, placebo 25% at day 200, but it could be at that day 199, placebo had 50% survival, then 25% of the placebo patients died on day 200. And on day 201, 25% of the treatment patients might have died. That's an extreme example, but if you look at survival curves, the "best" ones show a nice separation between placebo and treatment, increasing over time, while others show the arms coming apart, back together, crossing, etc. It is not always clear when a company refers to 3-year survival, whether they are referring to the overall survival through year 3, or the "landmark analysis" at year 3.
I'm not David Miller, but so far as I know, nobody, not the shorts or panel members who voted no such as Scher, doubt that the current 500 patient trial would continue. Whether this means the FDA would give an "approvable" letter, meaning, approvable contingent on successful results from 9902B, or simply no but with the understanding the company would refile based on 9902B if successful. The Scher position is that data so far as suggestive, but need to be confirmed.
The true short would add, that 9902B is likely to fail, whereas the true long believes 9902B will inevitably/almost surely succeed.
Looks like they are late with their 10-K. Now, if your real question is WHY are they late, that I don't know.
PART III -- NARRATIVE
State below in reasonable detail the reasons why Forms 10-K, 20-F, 11-K, 10-Q, 10-D, N-SAR, N-CSR, or the transition report portion thereof, could not be filed within the prescribed time period.
SpectRx is in the process of preparing and reviewing the report on form 10-KSB for the year ended December 31, 2006, and management does not believe the Form 10-KSB can be completed on or before the April 2, 2007 prescribed due date without unreasonable effort or expense. Management has no reason to believe that the audit will not be completed within the extension period. The company's Form 10-KSB for the year ended December 31, 2006 will be filed as soon as practicable and in no event later than the fifteenth calendar day following the prescribed due date.
See http://sec.gov/Archives/edgar/data/924515/000092451507000010/sp12b25032907.htm
It is accurate, http://biz.yahoo.com/prnews/070419/mo353.html?.v=1
I like this sentence: "The analysis is ongoing and entails employing an accurate statistical model that appropriately describes the data and provides accurate results."
Good thing they did not use one of them accurate models that provide inaccurate results.
But putting aside their difficulties with English, I guess the question is what does the folowing mean:
Neurochem has been advised by its external team of statisticians that adjustment to the initial statistical model, as set out in the statistical plan, would be necessary to provide accurate results. The procedure to arrive at a reliable model involves a detailed analysis of potential confounding factors such as the effect of concomitant medications, baseline characteristics of the study population or differences in clinical sites. Neurochem points out that potential refinement of the statistical model was discussed with the U.S. Food and Drug Administration before filing with the Agency and was anticipated in the plan filed.
The references to "as set out in the statistical plan,: "discussed with the U.S. Food and Drug Administration," and "was anticipated in the plan filed" are all a little vague, a little short of flat out saying that the precise adjustment process was prespecified. And even if it was prespefied to the FDA's satisfaction, as what? As a secondary endpoint, was there an alpha allocation so that a lower p-value would have to be met? On balance, one gets the strong impression that this process is not going to yield results that would satisify FDA statisticians.
This is data mining, retrospective analysis, it provides nothing more than an interesting hypothesis to test. Is there any suggesting that JD had prespecified he would hit a homer after the first homer endpoint was triggered? (There are said to be baseball precedents) Did he enter into an agreement with the baseball commission's office defining "homerun" and the measurement thereof in advance? If the answer to all of these is no, the event is not worthy of discussion and should be disregarded.
Let me also add that there is some evidence that the baseball might have hit someone in the stands and so was unsafe and should not be approved for that reason.
I question whether Black-Scholes is a good model for valuing options in development stage biotech. The formula depends on the assumption that options can be replicated with dynamic hedging, which in turn assumes that the stock price is continuous. For example, as the stock price goes up from a point where the delta is .5, you replicate the option by buying additional shares so that the number owned takes into account the increasing delta (ultimately, the delta of a deep-in-the-money call approaches one). Dendreon is a particularly good example of a stock about to face a discontinuous stock price, at least, if you believe that sometime on or before May 15, a binary event will happen. Admittedly, CEGE is less binary than that.
It is something of a mystery to me what the DNDN stock price, in combination with the option pricing, tells us about the post approval stock price and chance of approval. It cannot be telling us that there is a large chance of failure or that the stock is likely to skyrocket if there is approval. For example, you could explain a stock price of $15 by saying that the market believes there is a 50% chance of success (to be followed by a $28 stock price) and a 50% chance of failure (to be followed by a $2 stock price). The mathematical expectation of those two outcomes is $15. But then the call would be worth .5 x (28 - 15) = $6.5, and the put would be equally valuable. But instead, the 15 put and call are trading for about $3.8. I think this all suggests that the market is assigning a relatively high probability to approval, and that as the approval is expected, a large price jump is not. expected. I am not saying I agree with that assessment, but it seems to me that is what the stock and option prices imply.
The may 11 party "may imply "substantial evidence" of over-confidence at this juncture"
No, I think it "establishes" it. Not to mention, it creates bad karma.
I am a lawyer, not a doctor, and I will avoid your question, except to say that in my view, you ask the wrong question when you say, "would you want to own the stock and hold it at a $7.50 adjusted cost basis?" I think your cost basis is irrelevant, putting aside tax considerations, which I think generally should be ignored (and I am tax lawyer, by the way). As I write this, you can sell the stock for $15.25, and that $15.25 will be paid in the same good ole US dollars whether your basis is $7.50 or $18.00. So far as I am concerned, if you own 1000 shares, you should ask yourself, if I held no shares but had $15,250 in my brokerage account, would I buy 1000 shares of DNDN? If the answer is yes, fine, but if the answer is, you would not risk that sum of money, then don't hold 1000 shares.
I understand, however, that many disagree...
Isn't it true that the standard way that ODAC phrases the question is to ask whether there is "substantial evidence"? See the questions below from the Zometa panel (http://www.fda.gov/ohrms/dockets/ac/02/questions/3827q1_final.pdf)
Now, I suppose your argument is, that when ODAC asks the question, everyone knows that they mean, "beyond a reasonable doubt", and I suppose the circumstances of the question change in the Dendreon situation made it clear that the CBER INTENDED the "substantial evidence" standard to be lower, getting to a place they would have not been had the question been phrased in terms of "substantial evidence" originally. On different note, I do note that in the Zometa case, the FDA managed to ask the "supportive question."
Questions to the Committee:
For new drug approval, "substantial evidence" of efficacy from adequate and well-controlled
investigations is required. Evidence from multiple clinical trials is usually submitted, but
robust results from a single multi-center trial have been accepted. In your deliberations of
the following questions, consider whether the results from trials fulfill the regulatory
requirement.
1. Study 010 in breast cancer and myeloma
In Study 010, 44% of Aredia patients had an SRE on study versus 46% of Zometa patients.
Using the conservative two-95% confidence interval method, FDA calculates that Zometa
retains at least 49% of Aredia' s efficacy (demonstrated historically in comparison to
placebo).
a. Do other studies (011 and 039) provide supportive evidence for Zometa's efficacy in
breast cancer and myeloma?
b. Is there substantial evidence from adequate and well-controlled investigations of
Zometa (4 mg) efficacy in breast cancer and myeloma?
2. Study 039 in prostate cancer
a. Zometa studies 010 and 011 have evaluated Zometa efficacy in predominantly lytic
metastases. Can results from these studies provide supportive evidence for Zometa's
efficacy in prostate cancer, which produces predominantly blastic bone metastases?
b. Is there substantial evidence of Zometa (4 mg) efficacy in prostate cancer from
adequate and well-controlled investigations?
"You have two separate trials that gave conflicting outcomes on survival and TTP (or, arguably, conflicting outcomes on survival and two failures in TTP). If that doesn't need a tie-breaker, I don't know what does."
I would not call the results in the first trial (9901) "conflicting" on survival and TTP. Survival was stat sign and TTP was about .084 or .054 after some errors were corrected on audit. For want of .004 you say the null hypothesis wins? That seems awfully pedantic.
Obviously 9902A was less successful but even there it trended towards success. I agree that it is a close call whether 9902A is supportive; it was interesting that that was not discussed at the panel. But I would not go so far as to say 9902A data is conflicting.
Obviously there was no "subset analysis" for the basic overall survival analysis. That was statistically significant whether you examine 9901 or 9901 and 9902A combined. David's argument in rebuttal to Scherr's argument that the placebo itself mightt have a negative effect is a subset analysis, but given the design of 9902B -- the subject of an SPA -- I suspect we will never run a trial of Provenge v. inert placebo. If the FDA approves, obviously that won't happen, and if they require waiting till 9902B results are in.
1. On the first point, consider Mr. X, who enrolled one month ago. He does not know whether he received treatment or placebo. Maybe he suspects it was placebo as he did not any fever. He wants to quit and enter CU, where he has a 100% chance of receiving treatment, not a less than 2/3rds chance of treatment (recall his suspicion he in the placebo arm) to be followed by a crossover, which he views as too little, too late.
2. And if the answer to Mr. X is tough, if you quit the trial you get nothing from us, then the company is discriminating against men who have entered their trial. Maybe that can be done, but it feels "wrong." If the answer to Mr. X is OK, then isn't there a strong incentive to leave the trial? Again, as time passes and men progress and crossover, that incentive strongly diminishes.
3. Let's say that the FDA believes given the totality of the evidence, the "true" p-value is .04. I.e., a 96% chance the results were not due to chance. I think they have a strong interest in confirming this result and reducing the chance of a type I error as much as possible. Why are the considerations materially different than those that obtain where approval is based on a surrogate endpoint followed by a phase IV study? (I understand the REGULATORY framework is different).
4. As to waiting till 2010, that is what they said at the panel presentation, and they also said that the SPA provides for termination upon 350 or 360 deaths (I don't recall which of those is the number, but it one of them).
You say, "And although enrollment can be negatively impacted in many cases, i think that reason is a red herring in this particular example. The enrollment in 02B has clearly picked up, and arguably it will be fully enrolled very soon. At the stage 02B is at, an enrollment delay is not a reasonable argument to nix the possibility of a CU program. "
But, of course, enrollment is not the end of story. The patients need to be kept in the study as well. Would the CU program be closed to people who quit the trial? Is it ethical for the company to treat men who have not enrolled in the trial better than men who have not?
Now, I think it is possible to exaggerate this problem for both approval or CU, because men progress pretty fast -- after 4 months or so, something like 80% of the men on the 9901 placebo arm had progressed. At that point, the men presumably crossover and take Provenge from frozen. Further, at some point, men progress enough so that they would not fit within the proposed label (asymptomatic). So if the FDA is really worried about approval invalidating the results of 9902B, why wouldn't an althernative be, approval conditional on full approval plus, say, 6 months. Conditioning approval on the RESULTS of 9902B (i.e., till 2010) does not seem justified by a desire to maintain the statistical purity of 9902B, such as it is.