Exelixis' Halloween Revisited
[This post is from our Million Dollar Portfolio boards, a premium service. Click here to see all our premium offerings and find which service is right for you.]
Board: MDPS: Exeilixis (members only)
Despite many posts about Exelixis moving forward with the “306 trial” without FDA SPA agreement, I think that many interesting points and questions have yet to be raised on these boards. I’ll try to organize the facts that are publicly known, to hopefully rekindle the discussion. Exelixis is a pretty key holding for me, and I’ve been following the company closely for years now. I can offer my insights, and hopefully you’ll share yours in return. I also think this will be a good lead-in to get us thinking about what we should hope to hear at the Dec. 1 R&D Day.
While I still consider myself generally an Exelixis bull, this event, combined with the repayment of the last tranche of the Glaxo SmithKline (GSK) debt in shares (rather than cash), have me feeling more hesitant than ever. That said, the stock price – currently close to $4 – has a lot of risk already built into it.
Here is Exelixis’ press release associated with the announcement.
Here is the conference call transcript where management discusses their decisions.
There is an SEC filing that covers a lot of the same ground as the press release and conference call transcript, which includes some additional information not found in the other sources. See the 8-K dated 10/31/11 here.
Here is a recent Ohad Hammer post on Seeking Alpha regarding Exelixis. The impression I have of Hammer is that he is very biotech savvy. He seems to be a very good stock-picker, although I am less trusting of his financial analysis. I trust Charly (and Karl) first and foremost, but I also pay attention to Ohad…
I don’t want to go too far back in time, but I think that it is important to remember that two different “big pharma” companies (GSK and Bristol-Myers Squibb (BMS)) have held the rights to cabozantinib (a.k.a. cabo; a.k.a. XL184), only to return the drug to Exelixis. Many investors look at this as damning evidence of a drug with no future. In both cases, though, the drug company in question had a wholly-owned drug hitting similar targets, and would have faced significant potential payments to Exelixis (either immediate or profit-sharing) had they chosen to retain rights and further invest in this drug. While it seems reasonable for an investor to allow some doubt to linger, both “rejections” can be viewed as business decisions. All things being equal, each pharma company’s own drug would have been cheaper for them to develop, and more profitable for them, if approved, than cabozantinib. However, any negative surprises are sure to stir up memories of this legacy.
Speaking of legacy, one legacy of BMS’ involvement with cabozantinib is the randomized discontinuation trial (RDT), which was BMS’ suggestion for quickly assessing cabozantinib’s strengths. It was the RDT that helped Exelixis identify metastatic castration-resistant prostate cancer (mCRPC) as an indication where cabozantinib had something special to offer.
The Strengths and Weaknesses of What Makes Cabozantinib Unique in Fighting mCRPC
I have highlighted before that the drop-out rate in the RDT was disconcerting, but not evenly distributed across all the cancers in the study. mCRPC had a lower drop-out rate because something unique was happening – bone scan results, which were “known” to be impossible to improve, were improving (and for many patients, radically improving). Since late-stage mCRPC patients often die “a skeletal death”, bone improvement is a significant outcome – potentially both in terms of quality and quantity of remaining life. Doctors in cabozantinib’s mCRPC trials discovered another interesting trend. mCRPC patients on cabozantinib were inclined to decrease their narcotics intake. As bone scans improved, the debilitating pain associated with metastatic bone lesions also improved. Exelixis saw two results as potentially differentiating cabozantinib from an increasingly-crowded field of drugs targeting mCRPC: bone scan resolution; and pain alleviation.
The strengths of cabozantinib have “weakness” aspects as well. “Pain” is difficult to measure rigorously, especially in multi-language global studies. And “bone scan resolution” is an unprecedented measure – a measure with which the FDA has no regulatory experience.
Original Study Design – Overview + Expectations
Exelixis set out to design Phase 3 mCRPC clinical trials that would highlight cabo’s strengths. For the “306” study, they originally envisioned dual primary endpoints of pain alleviation and bone scan resolution. I think that management set expectations that this would be a relatively quick study which, if regulatory approval was granted based on the study, might lead to a limited label. Bone scan changes were seen within weeks of starting on cabozantinib, and the prospect of a speedy Phase 3 trial had investors giddy.
Fully cognizant that the “306”study’s “unusual” endpoints, Exelixis planned for the “307” study to focus on the classic endpoint of “overall survival” (OS). Had approval been granted based on “306”, a successful “307” trial could have helped broaden the label. Had approval not been granted based on the unusual “306” endpoints, successful results from “307” would have been likely to gain approval.
The “308” trial, to come later, would hopefully be able to expand cabozantinib’s label to pre-metastatic CRPC patients – a larger population with longer expected duration on drug. The previously-stated goal was proving that cabozantinib could stall the onset of bone metastasis longer than a comparator drug.
The “Whats” and “Whys” of a SPA
A spa is designed to help one relax and de-stress. An SPA – Special Protocol Assessment – has some of those same features. An SPA provides a pre-negotiated framework for clinical trial design. To a certain degree, it partially de-risks the trial because the FDA has already agreed that if the trial is run a certain way, and specific results are achieved, then approval will be forthcoming. An SPA is not required to run a trial with unusual endpoints. But it must be stress-reducing for management to have an SPA in place.
So… What Happened on Halloween Monday?
At a high level, Exelixis announced that SPA negotiations had broken down. Management described its plans for running the “306” trial without an SPA. The details disclosed included the number of patients to be enrolled, the choice of endpoints, the comparator drugs, and some hints toward expected duration. Some information was disclosed about the “307” trial too. Market reaction – a 40% single-day drop – was indeed Halloween-style frightful.
The number of moving parts here is somewhat staggering. Some of the issues have been raised on these boards, but I think some haven’t. My goal here is to try and objectively organize all the pieces, while adding some of my own personal opinion.
How Many “Moving Parts” Are There?
More than I can count… But let me try to touch on the ones that are most important. Some will be easier to explain, some more difficult, and others I’ll just have to let you decide how you feel about it…
1. Size and Reach of the “306” + “307” Trials
2. Trial Inclusion Criteria
3. Comparator Drugs
4. Endpoints of the Trials
a. Pain Alleviation
b. Bone Scan Resolution
c. Overall Survival
6. Drugs Fighting mCRPC
7. What is the Point of the “306” trial, and is it Worthwhile?
8. Thoughts on the “307” Trial
9. Current Thinking on the “308” Trial
10. My Thoughts on Exelixis’ Cash Situation
Size and Reach of the “306” + “307” Trials
The “306” trial will enroll 246 patients, entirely in English-speaking countries to minimize cross-language concerns regarding patient discussions of pain. The number of patients to be enrolled was increased marginally to support a secondary endpoint of overall survival, but I’ll discuss that later.
The “307” trial is not fully-designed yet, but current estimates are that 800-1,000 patients will be enrolled. This trial will enroll globally, except to the extent that there will be no overlap with “306” trial sites.
Trial Inclusion Criteria
I’ve already touched upon language constraints for the “306” study. Both the “306” and “307” studies will require that the patient’s CRPC must have metastasized to the bone, and further, the patient will have already failed both docetaxel and abiraterone, in no particular sequence. Patients may also have failed cabazitaxel, but that is not a requirement. For the “306” trial, patients must be in moderate-to-severe pain. This would be evidenced by a Brief Pain Inventory (BPI) score of at least 4, despite “optimized narcotic medication” (a combination of time-released and immediate narcotic dosage that cannot be increased during the trial without invalidating pain readings), while the “307” study will not include such a “pain” requirement. This set of inclusion criteria encompasses a large patient population, so enrollment can potentially complete quickly.
In the “306” study, the comparator drugs are mitoxantrone and prednisone. These drugs are some of the older ones available for mCRPC. These drugs have pain alleviation in their label. One concern the FDA raised during SPA discussions was “inadvertent unblinding” – that patients might be able to conclude which drug they were on based on either pain alleviation or toxicity profile. Choosing comparator drugs that offer pain improvement helps on the first point. To mitigate concerns over toxicity, Exelixis chose to dose cabozantinib at 60 mg., which has shown improved tolerability over the 100 mg. dosage used in the RDT. The choice of prednisone as the comparator drug in the “307” study wasn’t described in much detail. On the one hand, prednisone is well-studied as a comparator. Both cabazitaxel and abiraterone – drugs relatively recently approved for CRPC – were involved in trials where prednisone was a comparator. It is also likely that prednisone is seen as a relatively low hurdle against which to prove an improvement in overall survival. My guess is that the study’s required improvement in overall survival will take into account results observed in cabazitaxel and abiraterone studies. CEO Dr. Morrissey had this to say about the choice: “We considered a range of options … including going head-to-head against docetaxel in first-line chemotherapy-naïve patients. After consulting with key opinion leaders … we believe …” [the current plan for 307 has] ” … a higher probability of success and increased chance of quickly moving cabozantinib to market in metastatic CRPC.”
Endpoints of the Trials
Three endpoints are important in the “306” and “307” trials. I’ve touched on each already, but I’ll go into a little more detail here, with emphasis on the original plan versus the FDA’s SPA feedback versus the amended plans.
Originally this was to be a primary endpoint of the “306” trial, and it remains a primary endpoint for “306”. During his prepared remarks, CEO Dr, Morrissey pointed out that the FDA had granted, during SPA discussions, “… that pain palliation can be appropriate as a primary endpoint …”. Later in those prepared remarks, he indicated that one of the issues in the feedback letter was “… that if we use pain response as a primary efficacy endpoint, we conduct two trials to demonstrate effectiveness.” The “307” trial should meet that requirement. Pain will be measured using the Brief Pain Inventory (BPI) method, which is a “widely-used … robust, reliable, and sensitive measure of pain …” (from the press release). A “responder” will have a 30% or better reduction in reported pain based on data collected at week 9 and confirmed at week 15.
Bone Scan Resolution
This was intended to be another primary endpoint of the “306” study. The FDA instead suggested that it be an exploratory endpoint, probably due to the unprecedented nature of “bone scan resolution”, as discussed above. Instead, Exelixis has chosen to make this a secondary endpoint of the “306” study. Chief Medical Officer (CMO) Dr. Schwab said, “Including bone scan as a secondary endpoint in the 306 trial represents a significant step towards validating bone scan response in CRPC and correlating bone scan response with other meaningful clinical outcomes. Moreover, the inclusion of bone scan supports our long-term plan for developing cabozantinib in multiple prostate cancer indications.” Bone scan readings are performed using the computer-aided detection (CAD) methodology at independent radiology facilities (IRFs). A “responder” will have a 30% or better decrease in the lesion area of all bone lesions, and no soft tissue progression.
Overall survival (OS) was originally, and remains, the primary endpoint for the “307” study. It has been added as a secondary endpoint to the “306” study, where success is defined as “no negative impact”. They increased the study size by a few patients to allow this endpoint to be tested with “80% power to detect a 50% difference.” I’m sure that the “307” study will set a higher hurdle for approval based on OS efficacy as a primary endpoint.
I think the market’s the expectation was for the “306” trial to be relatively brief compared to the duration of most Phase 3 cancer trials – bone scan results and pain improvement have been seen with cabozantinib in just a few months. CMO Dr. Schwab said, “… the “306” trial will initiate before year-end, the “307” trial shortly thereafter in the first half of 2012. Both will enroll in parallel … our modeling predict that we will have data that supports a filing in the 2014 timeframe. So both trials should approximately read out at the same timeframe.” This is longer than expected, and the longer a drug isn’t on the market, the more time competing drugs have to establish themselves, diminishing future market potential. I took a quick look at established and potential competitors in a recent post.
Drugs Fighting mCRPC
Please see this post for details
What is the Point of the “306” Trial, and is it Worthwhile?
Exelixis tried to answer this question from both a trial and a commercial perspective. CMO Dr. Schwab pointed out “… that oftentimes in large studies that are enrolling quickly, pain information is collected with less rigor. And so you end up with a lot of missing data. So that’s in part why we believe other compounds that have collected some pain information have not received a pain label.”. Newly-hired Chief Commercial Officer Scott Garland opined, “… this opportunity to deliver a package that allows us to both point to anti-tumor activity through overall survival, but also pain improvement, will be differentiable and will ultimately, if approved, be successful in the marketplace.” Without a pain indication on the label, Exelixis won’t be able to talk about that during sales calls.
My personal take is the pain data will have to be stellar, not just successful, and coupled with impressive overall survival improvement, for approval based on “306” alone. This is “quite possible”, perhaps even “not unlikely”, given the experience of earlier trials. But it is not an “assured” result. If “307” is also successful, a filing based on both trials may present a lower hurdle for a pain indication as part of the label.
Ignoring the FDA’s request that the bone scan resolution endpoint be “exploratory” rather than primary or secondary makes me think the FDA will want to scrutinize this data, and not make a hasty decision towards its appropriateness as a regulatory endpoint. This could also potentially complicate approval based on “306” alone, even if the pain data are stellar. Or it could lengthen the regulatory process for a combined “306” and “307” package. Clearly Exelixis sees this as an important differentiator and wants to push the issue. I doubt that the FDA will allow itself to be rushed with regards to this metric.
I think the “306” trial is a bit of a gamble on Exelixis’ part. I think they see this trial as their opportunity to present the data that makes cabozantinib shine brightest, even if the endpoints are unusual. The “307” study appears to be the safety net associated with that gamble. Mitigating the risk associated with the gamble is the relatively small size and somewhat shorter duration of the study. At 246 patients, and probably less than two years, it will likely be much less expensive than a typical Phase 3 cancer trial.
But I think it is clear that Mr. Market doesn’t currently value “306” very highly.
Thoughts on the “307” Trial
The “307” trial is of critical importance to Exelixis. I think that Exelixis could survive a “307” failure because (1) cabozantinib shows promise in other cancers, and (2) Exelixis still has out-licensed compounds moving forward, with hopes of milestone payments and royalty streams. But the investment that Exelixis has made in mCRPC is such that remaining shareholders would likely face significant dilution to be able to get the drug moving forward again, should “307” fail.
That said, I think a “307” failure is unlikely. As discussed above, Exelixis has chosen a “low hurdle”, based on the comparator drug. Given the bone scan responses seen in most mCRPC patients, and the link between bone morbidity and patient mortality, I expect success – perhaps even the “stop the trial early for ethical reasons” kind of success that some potential competitors have shown.
Current Thinking on the “308” Trial
I was fairly disturbed to hear CEO Dr. Morrissey’s response to a question about whether the “308” trial would be postponed until after the “306” and ”307” data were available: “… our view is really dependent on the regulatory landscape that evolves in that non-metastatic space for CRPC based primarily on what comes out of the DMAB 147 regulatory review. So again we’re not planning on moving forward until we understand with great clarity how that file plays out, because to a large degree that will define the boundary conditions for how we could move forward in that space.” My concern is that Amgen’s denosumab isn’t even really a cancer drug. It has been approved by the FDA for use in forestalling skeletal-related events in cancer patients with bone metastases, but the drug’s focus is on strengthening bone, not fighting cancer. So I find it disappointing that Exelixis is reluctant to engage in this space, and it leaves me wondering why (hopefully more answers will be forthcoming at R&D Day). My biggest concern is: “Do they think that cabozantinib wouldn’t prove superior to denosumab?” But I think (hope) that this is less likely the issue. I think that Exelixis is more concerned about its cash position.
My Thoughts on Exelixis’ Cash Situation
Last quarter, Exelixis reported $313 million in cash and investments. However, roughly $100 million of that amount is not “spendable” (see my 1Q11 earnings analysis for details). With cash outflow from operations running at ~$43 million per quarter, spendable cash will run out before these “306” and “307” trials are complete. In fact, I would expect the quarterly outflow to rise as the “306” and “307” trials begin, and as Exelixis works to build out its commercial organization. I think that the “cash is getting short” flag was first raised when Exelixis paid off the GSK debt in relatively cheap shares.
How will Exelixis raise cash? There are five ways that I can quickly think of, and I’ll mention them in order from least attractive to most attractive. First would be another secondary offering. With current share prices much lower than the time of the last secondary offering, this would be an unattractive option, although I fear that some amount of future dilution may still be inevitable. Second would be debt (or convertible debt). Debt often comes with onerous terms that could limit future options, but it could be effective to raise cash to see Exelixis through to the point that cabozantinib is approved for a major cancer indication. Third is to out-license selective rights to cabozantinib. This has been talked about on and off, with the most recent news being that Exelixis chose to “extend the timeline” on a deal that was anticipated to close during 2011, surrounding Asian rights to cabozantinib. Still, the size of such a deal alone would be unlikely to fund Exelixis for very many quarters. The fourth source of income, and most attractive of the realistic options, would be milestone payments. These would be realized when out-licensed assets reach certain regulatory, developmental, or clinical thresholds. These are attractive because Exelixis has already performed the work, and would just be reaping benefits. However, the amounts and timing of milestone payments is hardly ever disclosed in advance, so outsiders have little insight into when these flows may occur. The most attractive form of income would be royalty payments from out-licensed assets. However, since none of those assets have progressed beyond Phase 2 yet, Exelixis will certainly not see any royalty flows anytime soon.
I believe that Exelixis’ decision to move forward without the SPA agreement was largely a timing decision. It is difficult to know how long it would have taken to reach agreement, but CEO Dr. Morrissey used the phrase “protracted discussion”. With new drugs recently having entered the mCRPC space, and additional drugs on the horizon, Exelixis management probably sees an urgent need to get differentiating clinical data into the FDA’s hands. If cash burn is an increasing concern for management, this would likewise push them towards taking the swiftest approach possible.
If I try to think like Mr. Market (ugh… my head hurts), I think there were five main concerns that led to the sell-off and continued lower prices. First, the new timeline for potential FDA approval probably exceeded what the market was expecting, reducing Exelixis’ appeal as a near-term holding. Second, the delayed timeline implies the potential for worse-than-expected dilution to provide enough funding to see Exelixis through the longer time horizon. Third would be the increased risk associated with the trials. Based on the trials as originally envisioned, the risk and benefit was probably viewed as a bit more evenly split between the “306” and “307” trials. Now, the market seems to think that “306” is riskier due to lack of an expected SPA, and “307” is riskier because it is now almost an “all the eggs in one basket” trial. Fourth, if selective out-licensing of cabozantinib rights is Exelixis’ main source of funding until its approval for a major indication, how many rights will Exelixis be able to retain? Perhaps fewer rights will remain than previously-expected. Finally, and I suspect this is less important because it was far in the future anyway, Exelixis’ ability to compete in the pre-metastatic CRPC market (a larger market with longer-lived patients than mCRPC) has been called into question by Exelixis’ own management by choosing to “wait and see” regarding the denosumab trials.
Less tangibly, there seemed to be a lack of cohesiveness in the presentations by management. CEO Dr. Morrissey presented the “306” and “307” trials in more complementary fashion during his prepared remarks, while CMO Dr. Schwab was insistent that each was capable of registration on its own during the Q&A. While Dr. Schwab described, as quoted above, “multiple prostate cancer indications” as part of the “long-term plan” during her prepared remarks, Dr. Morrissey seemed to undercut that with his Q&A comments about waiting for denosumab results before considering the “308” trial. This may have also played a role in the market’s “discomfort” with this announcement.
Personally, I continue to hold all my shares. I have neither sold nor purchased shares recently. Selling at current prices seems unattractive to me. But increasing my already-large exposure to this biotech likewise seems unattractive. I hope we learn more at R&D Day about Exelixis’ plans for the future.
I’m sorry for such a long (and delayed) post. I do hope you find it to be of value.
TMFDatabaseBob (long: EXEL)
Peace on Earth