(Editors' Note: This article covers a micro-cap stock. Please be aware of the risks associated with these stocks.)
Another round of fire has begun in the dendritic cell vaccine war. The latest casualty has been ImmunoCellular Therapeutics (IMUC). After the market closed on December 11, ImmunoCellular, the cancer immunotherapy biotech, announced the results from a Phase II study of ICT-107. The company tried to sugarcoat the news in its press release attempting to highlight what little good news there was to report and burying the bad in the second paragraph, but in the end, the treatment did not show an increase in overall survival for the intent to treat population, which was the trial's primary endpoint.
The responses and reactions have come in a flurry, and ImmunoCellular's dendritic cell vaccine counterpart, Northwest Biotherapeutics (NWBO), has once again been dragged in to the controversy. IMUC's failure is apparently bleeding into NWBO's stock price.
When the market opened on December 12, ImmunoCellular stock had plummeted to $1.07, losing over 60 percent of its value as a result of the missed endpoint. It now sits even lower, around $0.90. Within the same time frames, NWBO plummeted from just shy of $4 a share to a little over $3.30 and now sits at $3.35. Northwest released no data, yet suffered a similar fate. Investors believe the failure of ImmunoCellular's ICT-107 equates to a similar outcome for Northwest's DCVax-L.
The latest round of discussion centers around three authors. One is Josh Ginsburg at Seeking Alpha, who defends DCVax-L by accurately differentiating it from IMUC's ICT-107. Two is Adam Feuerstein at TheStreet, who lambasts Northwest as he has been doing for some time, calling its Phase I results misleading, just as IMUC's turned out to be. Three is Larry Smith, one of NWBO's staunch, thorough, and consistent defenders. While Smith does his usual good job at covering (almost) all the bases and taking Feuerstein's attacks in stride, I feel that Smith does not go after Feuerstein's accusations and attacks directly enough. Neither does Ginsburg. Feuerstein's reasoning with regard to DCVax-L's "misleading" Phase I results is strong, and needs to be dealt with head on. In this article I will try to sort out the thorniest of the issues involved here.
Ginsburg vs. Feuerstein, Who's Right Regarding What?
I will precursor this section with a statement not intended to offend, but simply to get us, as investors, to admit something. The vast majority of us are not scientists. Even those who are will likely be unable to claim they have a deep understanding of either the technology involved with the manufacture of dendritic cell vaccine or the science involved with the immune response to a brain tumor. We can only gather whatever information is out there and use it to the best of our ability, trying to check our bias at the door as much as possible.
That said, Ginsburg correctly points out that ICT-107 and DCVax-L are NOT the same, for two crucial reasons. First, ICT-107 uses 6 synthetic antigens in priming its dendritic cells. The synthetic antigen approach has been used many a time before in attempts to overcome a myriad of diseases. It rarely works. There are hundreds of different antigens on any given pathogen, and picking a small handful (6 in IMUC's case) of these is like shooting a giant with a BB gun loaded with 6 rounds, hoping somehow you'll hit a vital system and the giant will fall. Also, as Ginsburg points out, "There is absolutely no reason to be assured that products of a synthetic nature can be effectively identified and responded to by the immune system." In other words, the BB gun may not even fire in the first place.
Second, and as Ginsburg points out, the antigens picked up by dendritic cells in DCVax-L are from the patient's actual tumor. Nobody knows - not even the inventors of DCVax-L itself - exactly what or how many antigens are being picked up by these cells during the DCVax-L manufacturing process. It's all going on under a microscope and it's all a big mystery. It is unknowable. Asking how many or which antigens dendritic cells pick up in DCVax-L is like asking how many angels can dance on the head of a pin, or in this case, on the head of a brain tumor. However, if I may humbly and hypothetically put myself in the terrible and terrifying position of a GBM sufferer, being explained the differences between the DCVax-L and ICT-107 manufacturing processes, I would take my chances with DCVax-L. At least it loads the gun with a lot more bullets, bullets that match my own tumor indigenously.
So Ginsburg is correct on this issue, and Feuerstein is wrong when he bluntly states carte blanche,
Northwest's DCVax is not different from ImmunoCellular's ICT-107. DCVax will fail just like ICT-107 failed because dendritic cancer vaccines are too weak to overcome cancer's innate ability to overpower or evade the body's immune system," Feuerstein does have a very strong point when he talks about both Northwest's and ImmunoCellular's Phase I trials.
While DCVax may yet fail, he is wrong about why, because the two are different, and neither he nor anybody else can possibly know what's going on with the DCVax-L dendritic cells dancing on the head of a tumor and if they are "too weak" just because the whole dendritic cell approach is not to his liking.
As regards his attack against the dendritic cell vaccine approach in general, another company working on a dendritic cell vaccine, Agenus (AGEN), just recently announced the results of a Phase II study involving 41 patients for Prophage. Overall survival was 11 months. There was no control group so estimations of statistical relevance are problematic when comparing to standard of care, but median overall survival was 4 weeks longer than for Avastin, which is already approved, and Prophage has a much more benign side effect profile. Agenus' approach is almost identical to Northwest's, except that the former uses heat shock proteins to aid antigen recognition. While Feuerstein seems focused on IMUC's Phase II failure to bash dendritic cell vaccines, what about Agenus' recent developments with Prophage?
That aside, Feuerstein does have a very strong point when he talks about the anomalies with regard to the "spectacular results" surrounding IMUC's Phase I trial for ICT-107. How could it be that ICT-107 showed a median overall survival of 38.4 months in Phase I, but failed to even statistically outpace standard of care at all in Phase II with median overall survival? His answer is very damning for Northwest and DCVax-L. He says that in Phase I, patients are more easily handpicked for health and probability of long survival times, and he is right. On top of that, the 16 patients involved in ICT-107's "spectacular" Phase I trial may have even been handpicked for expressing the antigens attacked by the vaccine. From the press released linked above:
According to information presented at the scientific meetings, all eight long-term survivors had tumors with at least five antigens, 75 percent had tumors with all six, and 100 percent had tumors with at least four antigens associated with cancer stem cells - cancer-originating cells that appear to enable tumors to resist radiation and chemotherapy and even regenerate after treatment.
What does this mean for DCVax-L? It means that its Phase I data is also likely based on a handpicked sample of relatively healthier patients, and therefore data is also likely skewed by some margin.
The crucial question, however, is this: By how much is it skewed?
Yes, companies cherry pick patients for Phase I, if inadvertently. They all do it. Those who don't realize this are missing a crucial part of trial stage biotech investing. If you think that DCVax-L Phase III will show the same 36-month median overall survival that it did in Phase I, then there's a bridge in California I want to sell you, or however the saying goes. It won't. Any big money NWBO investor knows this.
Feuerstein himself quotes Larry Smith in admitting, and warning about, just this. I will reproduce the quote here:
The Phase I trials of DCVax-L enrolled a small number of patients in non-randomized trials performed at one clinical site. Such trials are subject to investigator bias, albeit inadvertent, as investigators have a tendency to select for younger, healthier patients. They are reluctant to enroll older, sicker patients with poor prognosis who might die regardless of treatment. No one wants to see a new drug be judged ineffective because it is given to patients for whom there is very little hope for survival. In Phase II and III clinical trials, however, such patients do find their way in. There is an axiom in cancer drug development that Phase III results are never as good as Phase II results and Phase II results are never as good as Phase I results. Comparing results from different trials is also viewed suspiciously because protocols and the characteristics of the patient populations can vary widely, especially when comparing small Phase I to large Phase III trials.
So we all know Phase I is a little biased. But what we all want to know is what numbers will the Phase III show? Investors don't need 36 months or anything close to that in order to win big. Let's not forget that Dendreon's (DNDN) Provenge was approved on just over four months of survival extension over control. Four months. That's it. Median overall survival for standard of care GBM is 11.2 months. That means 15 months will do it for DCVax, so even if DCVax-L's Phase I results are skewed by a factor of 2.5, DCVax-L still has a good chance of being approved. And all sides agree that DCVax-L's side effect profile will most likely come in benign, as side effect profiles tend to do for other autologous vaccines.
But what about ICT-107? It went from 38.4 months in Phase I to statistical insignificance in Phase II. Can't that happen for DCVax-L as well? Theoretically yes, but I wouldn't bet on that. Why not? To paraphrase Bill Clinton, "It's the antigens, stupid!" As we know from that glaring red flag in ImmunoCellular's last press release about ICT-107's Phase I, all of the patients had at least 4 antigens attacked by ICT-107 on their tumors, and all 8 long term survivors had at least 5. One of the reasons that Phase II didn't pan out so well probably has to do with patients not matching those synthetic antigens as well. Lack of antigen matching cannot happen with DCVax, as the nature of manufacturing it means that the antigens necessarily match because they are tailor fit for each patient.
And once again, how many antigens dance on the head of DCVax-L? I'm willing to bet it's more than 5. And they are not synthetic copies, and they are never the same ones for each patient.
Other differences between ICT-107 and DCVax-L Clinical Trials
The next major difference between the two technologies is the design and execution of the respective clinical trials. Northwest's trial involves over 310 patients. ImmunoCellular's trial involved just 124. The Medicines and Healthcare products Regulatory Agency [MHRA] in the U.K., the Paul-Ehrlich-Institut [PEI] in Germany and the FDA in the U.S. have approved Northwest's trial. Only the FDA had approved ImmunoCellular's trial. Northwest's trial has a built in margin of error, being powered to reach a p value of 0.01 with an 82% probability of success. Finally, Northwest's trial has three interim reviews, with a possibility to increase, and the company has the option to increase the trial size at the third review. In short, the DCVax trial is much more robust, and its design-assuming, of course, that the treatment actually works-will make achieving statistical significance more likely. Earlier studies have shown that both treatments work to some extent at least, and in light of the recent results released by ImmunoCellular, it is not unreasonable to suggest that a better designed trial might have resulted in ICT-107 achieving statistical significance.
Other Value Differences
Further differences lie in the operations of the two companies and the value these operations offer investors. Take a look at Northwest's pipeline and you will see DCVax-L is well into its Phase III trial and DCVax-Direct is well into its Phase I/II trial. Take a closer look and you will see the company's operations expand out of the U.S., across Europe. Both of these trials, locally and internationally, create value for Northwest, and as results are released throughout the course of 2014 there could be a number of major catalysts that present upside potential to investors. ImmunoCellular only has the one trial under way and only in the U.S.
ImmunoCellular holders will likely despise me for pointing this out, but it goes without saying that there is a large risk involved with taking a position in any development stage biotech. Development stage biotech companies face a long and costly road when they set out to get anything approved by the FDA, and experimental brain cancer treatment is likely more difficult, and more expensive, than many other endeavors. Having said this, even in the risk department Northwest is superior. The aforementioned diversification, both in terms of pipeline and locations of regulatory authorities, serve to dilute the risk associated with the company's stock.
All said, I believe the dip in Northwest's stock is collateral damage, and an emotional response to articles like Feuerstein's. He makes some good points, but neglects the differences between ICT-107 and DCVax-L that address his points if considered in context. DCVax-L is superior to ICT-107 in almost every way except the fact that DCVax-L is much more expensive to manufacture. The quality of the technology and treatment, the design and progress of the clinical trials, the fundamental operational aspects of the company, and most importantly the antigens. Northwest stock is cheap now, and the company just announced that the first interim analysis of the DCVax-L Phase III is due in 6 to 8 weeks. Keep your eyes peeled. We'll finally have the beginning of the answer we've been waiting for nearly 10 years.