Skip to Content

Business and Markets

Opdivo: Euphoria to Panic

We’ve gotten used to seeing one success after another in the cancer immunotherapy field, so this morning’s news is a bit of a shock: the Bristol-Myers Squibb flagship antibody, Opdivo (nivolumab), seems to have completely missed its endpoints in a trial in non-small cell lung cancer patients. And this was in over 500 patients who had already been selected for PD-L1 expression. That’s an endogenous ligand for the PD-1 receptor (the target of the antibody), so you’d think that these patients were set up for the therapy to show effects. Clearly, that’s not enough by itself.

The stocks of the two companies are undergoing pre-market moves that are rarely seen for these market caps: BMY was down as much as 19% and MRK up as much as 11%, which represents tens of billions of dollars sloshing around. We’ll see what happens during the day – personally, I think it’s a bit early to conclude that since one PD-1 antibody ran into trouble in this indication, the other PD-1 antibody won’t, but whatever. And in the end, the main thing these results tell us is that PD-1 antibodies can’t do it on their own in every indication, so we have to wait for combination data. Both a selloff in BMY and a rise in MRK would seem to be premature.

Update: one problem seems to be that the Opdivo trial may well have set the cutoff for PD-1L expression too low. Merck’s trials for Keytruda in this indication targeted higher-expressing patients only, and there have certainly been indications that expression level can be a critical variable. Bad trial design, then?

24 comments on “Opdivo: Euphoria to Panic”

  1. Bruce Grant says:

    Interesting that this news should break on the same day that STAT leads with a story on novel Cuban-developed immunotherapies that target the ligand(s) not the receptors.
    https://www.statnews.com/2016/08/05/lung-cancer-cuba-biotech/

    1. Anon8 says:

      @ Bruce-Thanks for this article. Indeed amazing if it can work in all patients! Though we do not get to read many article on science from Cuba, it is commendable what they are able to accomplish with limited resources.

      1. Imaging guy says:

        Though STAT article is mostly anecdotal, there is one paragraph about a recently published randomized trial. It says, “a Cuban study of 405 patients published earlier this year in the journal Clinical Cancer Research found that those who were given CimaVax lived about three months longer than those in a control group who got standard care only. More importantly, after five years, 23 percent of the patients who took the vaccine were still alive. None in the control group survived that long”. I have read the original article and found the second sentence misleading. That figure is from an exploratory analysis of EGF concentration (which is the target of the CimaVax vaccine) and the authors clearly state that “the evaluation of the EGF concentration as predictive or prognostic biomarker was classified as exploratory”. All you can take from the original article is the fact that “five-year survival rate was 16.62 % for those vaccinated patients that received 4 vaccine doses vs. 6.2 % for non-vaccinated patients”. Please note that the trial was not double blinded although it could have easily done.

    2. Barry says:

      as the Statnews article explains: “Gonzalez and her colleagues devised a biological compound that could effectively interfere with the binding process”of EGF to EGFR. I.e. CimaVax is not a cancer-targeted immunotherapy at all (although it might have advantages over targeting EGFr on genetically-unstable cancer cells

  2. bhip says:

    The level of PD-1 expression is an important factor in predicting activity when attacking that system. However, ligand expression is likely necessary but not sufficient for activity in certain tumors.
    Other factors beyond PD-1 & CTLA-4 play into the suppression of the immune response to tumors. Some examples include tumor IDO1 expression (local tryptophan depletion suppressing CTL function), adenosine production by tumors (increases cAMP in CTL, suppressing their function) & influx of myeloid-derived suppressor cells which impair T cell function within the tumor microenvironment.
    It’s important to remember that the efficacy of the PD-1 therapeutics in certain tumors, while astonishing, only approaches 30-50%.

  3. Ruud says:

    BMS got caught for playing cute on the biomarker front- with their “complementary diagnostics” strategy and the cutoff to include as many patients as possible. Business greed trumps science in a very near-sighted fashion.

  4. Mol Biologist says:

    IMO what is going on with PD-1 & CTLA-4 is a good example of cognitive dissonance.
    http://www.reuters.com/article/us-health-cancer-fda-idUSKCN0YW15T
    There are too many failures but nobody want to think the therapy is failing due to altered cancer metabolism.
    http://symposium.cshlp.org/content/76/335.full
    M

  5. Xplodyncow says:

    I wonder if the selection of the primary endpoint has anything to do with it…? The current lung cancer indication for nivolumab is based on two trials with OS as the primary endpoint. I thought there was some evidence that immunotherapies can seem to cause tumor growth before there’s a response. But, eh, what do I know? I’m no scientician.

    1. AC says:

      Tumors increase in size do to the increased infiltration of immune cells in some cases.

      1. Xplodyncow says:

        So maybe if PFS was measured too soon, that would make the trial miss its primary endpoint? Just speculating.

  6. me better says:

    5% cutoff vs 50% cutoff

    1. ruud says:

      Granted that BMS’s diagnostics test (IHC) uses a different antibody than Merck’s, 5% cells stained positive vs 50% cells stained positive is quite different. BMS’s motive of getting as many patients as possible is quite apparent.

      1. BigSky says:

        Looked at another way, this could be a manifestation of the patient accrual game that companies play during clinical trials>> patients you accrue for your trial can’t be accrued by your competitor. Sometimes one of the unstated motives can be to cut off or stall your competition’s access to patients. This is the stuff the mob working on the ‘carpet side’ of the building comes up with.

        1. Eric says:

          Nobody does that. Way too expensive.

      2. me better says:

        all from Dako

  7. steve says:

    It shows what happens when marketing people influence scientific decisions. Confine the marketing guys to the post-approval process and keep them out of the drug development process and pharma will be much better off.

  8. Sirtainly Not says:

    Derek, you sound unaware of the results of Keynote-024 earlier this year.

  9. Mike B. says:

    New decade, another new tech that only prolongs life by a few months in many cases. This will just be like chemo–a ton more work needs to be done exploring and coming up with combination immunotherapies to achieve maximum efficacy. PD1,CTLA4, ….they can’t do it all alone in many cases because I can think of 6 or more other ways cancer cells can still potentially block immune response.

    1. zero says:

      Suppose there is a cancer such that patients have one month to live after diagnosis using current standards of care. If a therapy ‘cures’ 1 person out of 20 (who survives another ~10 years in remission) but the other nineteen people die in one month as expected, then that therapy extends life on average 6 months. For the nineteen people that see no benefit it was worthless, but for the one person who was effectively cured it is a major benefit.

      Is this treatment worth developing? I think yes. Dealing in terms of average benefit is essential for some parts of the process, but it is important to remember the details. A low-probability cure is much better *when it works* than a high-probability life extension, but both have a role to play. If it was me getting the diagnosis I’d rather try for the cure first and then decide whether to take the shorter extension.

      That’s all to say that a treatment with a three-month average benefit is better than nothing, and for responsive patients it could be much better than nothing if it produces long-term remission for even a small fraction of the population.

  10. steve says:

    Nivolumab – as a single agent – gave a 1-year 69% survival rate in advanced melanoma. Combined with Ipilumimab it gave a a 1-year survival rate of 85% and 2-year survival rate of 79% for advanced melanoma patients. If I was a patient looking at what used to be a death sentence from advanced melanoma I sure would be impressed. The Keynote-087 trial showed an overall response rate of 73-83% in advanced Hodgkin’s and a complete response rate in 27-30% (depending on the cohort). The Keynote-024 trial in non-small cell lung cancer that Sirtainly Not cited showed an overall response rate of 24.8% and progression-free survival of 6 months, with a median duration of 23.3 months. This was for pembrolizumab as a first-line, stand-alone treatment. For sure there is a lot of work to do on combinations to improve efficacy across a range of tumor types but I sure wish the cynics on this blog would do their homework before just posting negative comments. Cancer is a difficult disease and advances like these should be applauded, not cynically derided.

    1. anchor says:

      Steve: Great thoughts and am fully with you on this! A combo would work great and we are talking about another clinical trials. But, how else are you going to convince thoughtless and cancer free (!) stock holders who do not know science and how it works? They only know how to count $$$$!

Comments are closed.