Recent

% | $
Quotes you view appear here for quick access.

Exelixis, Inc. Message Board

  • erniewerner erniewerner May 3, 2013 7:33 AM Flag

    Aveo's Tivoznib

    You will notice that AVEO came down in 2 steps. The stock took a hit when the briefing documents were released and then again when the ODAC recommendation was voted on.

    This is interesting. The pivotal trial was open label Tivo vs Sorafenib (Nexavar). Several TKI's are approved to treat RCC and most were approved on the basis of PFS. Tivo bested Sorafenib with a statistically significant improvement in PFS, a better response rate, longer duration of response and a lower SAE rate. Sounds like a slam dunk approval----but here is the catch. Until the briefing document was released, it was not public knowledge that Tivo had shown a trend toward a worse survival outcome. Median Nexavar was 29.3 months and Tivo 28.8 months. I have said on multiple occasions that what is not said is often more important than what is said. Aveo shareholders were blindsided by a management team that cherrypicked the data to be made public. A .5 month difference in median survival doesn't sound too bad, but the HR was 1.25 and the p value was .11. That means that the unpublished survival curves are skewed pretty consistently against Tivo. Further muddling the situation was a crossover that allowed progressing Nexavar patients to receive Tivo.

    Rolling into the Cabo MTC approval decision, I was worried that unpublished Cabo vs placebo OS data might have killed the NDA and I was gratified to see that CRPC OS data is being presented at ASCO.

    SortNewest  |  Oldest  |  Most Replied Expand all replies
    • Not quite the case. The negative OS was well known and presented. What was not known:

      1) The FDA recommended a second trial pre-BLA submission
      2) The cross-over, despite the SPA, was not communicated to the FDA; it was devised post-SPA agreement

      • 2 Replies to xwavve
      • @xwave:
        For 2) You mean that after AVEO failed to receive an SPA, they ironed out some acceptable trial design with the FDA which they subsequently changed to the single arm cross-over design that they used without informing the FDA?

      • As always, Thx for correcting my appraisal of the situation.

        2) The cross-over, despite the SPA, was not communicated to the FDA; it was devised post-SPA agreement

        Never let a good act go unrewarded. For contrast, here is how EXEL handled the SPA issue after they had EXAM topline results. "Exelixis will consult with the FDA to determine whether the trial conduct should be changed as a result of these data in conjunction with the SPA. "

    • "Rolling into the Cabo MTC approval decision, I was worried that unpublished Cabo vs placebo OS data might have killed the NDA and I was gratified to see that CRPC OS data is being presented at ASCO. "

      Ernie, are you assuming the OS data for MTC was part of the reason the FDA approved Cometriq and that it will translate to CRPC? Do they have to dramatically beat the existing CRPC data for OS? Does the recent collapse of the stock reveal a leak or is it just part of the safe hedge bet? I suppose we'll have to wait and see.

      Sentiment: Strong Buy

      • 1 Reply to enabeler
      • "Ernie, are you assuming the OS data for MTC was part of the reason the FDA approved Cometriq and that it will translate to CRPC?"

        We do not know what was in the complete filing. If there had been an ODAC meeting, all of that would be public now. There was an interim OS analysis done at concurrent with the PFS analysis and the result was inconclusive with no indication which direction it may have leaned. We do not even know if the DMB shared the OS analysis with EXEL. From a logical perspective, it is likely the OS data improves with time. VEGF inhibitor SAE's tend to show up in early in the treatment cycle. MTC was done with the understanding that an approval could be given for a PFS endpoint as long as there was not an OS disadvantage as occurred with tivozanib.

        "Do they have to dramatically beat the existing CRPC data for OS?"

        It is a pass/fail test and the statistic that determines pass/fail is the p value. The required p value that determines if the result is statistically significance is usually set at p=.05 or less for trials with a single analysis. Since Comet 1 has at least one interim analysis, each of those analyses will have a preset p value (already spelled out in the trial protocol). If the p value is met, then the result (regardless of the numeric values of the median comparison) is said to be statistically significant. So no, the result does not have to be dramatic, but it does need to be statistically significant.

        "Does the recent collapse of the stock reveal a leak or is it just part of the safe hedge bet?"

        My opinion on this is no better or worse than anyone else's. When we took that 8% drop, the only potential leak that popped into my head was an, uh-oh I hope they aren't floating another secondary.

        "I suppose we'll have to wait and see." Yes, but the puzzle is slowly coming together. I'm anxious to see if any business development news is announced and the ASCO abstracts come out in less than 2 weeks.

    • Very good point about the harbingers for AVEO. It looks like selling off last August might have been prudent with the disclosure of the OS curves. I don't know the researchers who investigated Tivozanib that well. Would you say that they were top flight doctors who have a consistent track record or fledgeling postdocs who needed a publication?

      • 2 Replies to wildbiftek
      • "Would you say that they were top flight doctors who have a consistent track record or fledgeling postdocs who needed a publication?"

        They had two of the best, Robert J. Motzer, M.D., of MSKCC and Dan George of Duke. The investigators of a trial would prefer to lay out all the data in an unbiased manner. That said here is what was presented at ASCO 2012 about the Tivo-1 trial:

        "Overall survival (OS) data are not yet mature. In TIVO-1, 53% of patients randomized to the
        sorafenib arm of the trial went on to receive subsequent therapy, nearly all of whom received
        tivozanib after sorafenib. Based on an early, interim analysis, 81% of these patients achieved one
        year OS. In comparison, only 17% of patients randomized to tivozanib went on to receive a
        subsequent therapy, and 77% of these patients achieved one year OS. Mature data are expected
        to be presented in 2013."

        That desire to be unbiased is tempered a bit when it is a pivotal trial paid for by the sponsoring company. The 2012 ASCO blurb was a weak warning that the OS data was leaning the wrong way, but the explanation attributing it to the crossover was a good one. Still, I have never seen FDA or ODAC give an extra credit bonus to companies for incorporating a crossover. When the final data came out the OS hazard ratio was 1.25 in favor of Nexavar. Aveo agues that you are comparing Tivo vs Nexavar/Tivo because of the crossover and perhaps they are right, but again, no extra credit for crossover.

        We may never know what results were achieved at the first Cabo OS analysis in the EXAM trial. When the topline results were released EXEL did approach FDA to see if they wanted the trial to continue blinded with no crossover or unblind with Cabo made available to placebo patients. The trial has continued blinded and as a result the MTC OS data will not have that confounding influence.

      • The ad hominem aspect:

        Of course, who does the research doesn't necessarily change the veracity or content of the results, but I think it's reasonable to dig extra deep if there might be a lot of reason for the disclosure to be skewed because of career building or stakes in the company etc.

        I also noticed that there was also heavy insider buying of the shares of AVEO in the past year. The CEO owns 500,000+ shares so it seems like he went down hard with his shareholders. In this case, the insider buying actually flew in the face of their fully disclosed results and was probably a ploy to whip up investor excitement, or maybe a desperate vote of confidence in their life's work. That said, I'm glad that our management has some skin in the game as well, although Scangos' large stake in EXEL is for me the most heartening insider holding.

        Whether AVEO is a buy around $2 is another story I'll have to look into but that can certainly wait since even companies w/ approved products (EXEL or NAVB) might have their share price depreciate these days.

 
EXEL
4.54-0.21(-4.42%)Feb 5 4:00 PMEST