Oncothyreon (ONTY) represents one of the few opportunities for investors to take a high probability, statistically-backed, event-driven investment with a payout of many multiples of their money. ONTY’s drug for Stage 3 Non-Small Cell Lung Cancer ("NSCLC"), Stimuvax, is nearing the end of its pivotal Phase 3 trial. Our analysis has indicated an overwhelming 90%+ probability of success. On the success of the Phase 3 trial, we believe that current ONTY investors are likely to at least triple their investment. Furthermore, there is a high chance of investors making five to 10 times their money based on additional cancer indications for Stimuvax. Lastly, over the long run, ONTY has a once in a lifetime return profile from future drugs developed out of its wholly-owned immunotherapy platform. The economics from the pipeline are far greater since ONTY owns all future drugs being developed and will not be limited to a royalty stream, as is the case with Stimuvax.
Current Stimuvax Phase 3 Trial: Very Likely to Be Successful
First, some background on Stimuvax and the Phase 3 trial. Oncothyreon’s Stimuvax is an immunotherapy, commonly called a cancer vaccine. It trains a patient’s immune system to respond to cancer cells that express MUC1, a protein antigen widely expressed on common cancers. MUC1 is over-expressed in lung cancer, breast cancer, prostate cancer and colorectal cancer. Through a series of agreements, ONTY partnered with Merck KGaA for the worldwide development, commercialization, and manufacturing of Stimuvax in exchange for a royalty stream.
Stimuvax is currently in a Phase 3 study for the treatment of unresectable Stage 3 NSCLC. A Phase 3 trial means that this is the last trial typically undertaken in advance of submitting the drug to the FDA for approval. Stage 3 NSCLC means the lung cancer has spread to other parts of the lung but hasn’t yet metastasized to other parts of the body. The Phase 3 trial, named START, was initiated in February 2007. Patients were randomized 2:1 (treatment patients: control patients) in a double-blind fashion with target enrollment of 1,322 patients (link here). In order for the trial to be successful, it needs the drug to show at least a six-month survival benefit for patients on Stimuvax with statistical significance (a one-sided p-value of 0.025, based on the efficacy side of a standard two-sided 95% confidence interval). The trial assumes the control arm will have a median overall survival of 20 months.
The START trial has two built-in interim looks, which are points at which an independent committee examines unblinded data for three options:
- Futility: The drug isn’t working; take everyone off.
- Efficacy: The drug is working; put everyone on.
- Inconclusive: Continue the trial as long as safety is okay.
Both interim looks and the final closing of the trial are triggered by the number of events (death). The first interim look is at 353 deaths (50% of total events), the second is at 529 deaths (75%), and the trial ends at 706 deaths. At each analysis point, the data monitoring committee determines a median overall survival for the treatment arm and the control arm, and compares it for statistical significance. The p-value is based on the standard O’Brien-Fleming method and gets progressively easier to meet as the trial progresses. We have estimated that the p-value for the first interim look is very small (p < 0.0026), the p-value for the second is a little higher (p < 0.011), and the final p-value is p < 0.024. The initial hurdle is extremely high and gets lower over time since a greater number of events provides greater statistical power.
The only certain way to know if patients are living longer is to know how long each trial participant lives and whether each patient is in the treatment arm of the trial or the control arm. While we don’t have this level of detail, we do know when most of the patients were enrolled, how long the control arm should live, and the number of people that have not died thus far. With this information, it is possible to create a statistical model to back into how the treatment arm of the trial is probably doing based on the trial timeline and the timing of events, and to therefore make a highly accurate prediction regarding the outcome of the START trial. Let’s review what we know about each of the datasets:
- When the patients were enrolled: We can pinpoint with reasonable accuracy when most of the patients were enrolled based on a poster published by Merck KGaA (MKGAF.PK) at the 2010 ASCO conference, the upper-right hand corner of which shows the enrollment curve for the first 1,075 patients enrolled out of the original targeted enrollment of 1,322. We also know that enrollment in the trial was suspended for three months between March 2010 and June 2010 due to a single patient death in another Stimuvax trial and that patient enrollment then resumed in June 2010. Due to the three-month hold, 165 patients who were on Stimuvax for less than six months had to be replaced, and total enrollment was increased to 1,476. The clinical hold did not affect patients on the trial for longer than six months since the trial protocol allowed patients to miss two doses over the course of treatment. Based on our calls to enrollment centers as recently as last week, we know that the trial is still enrolling new patients. A list of enrollment sites is here. Once the START trial resumed, ONTY and Merck provided information on the number of patients that had to be replaced, the timing of restart, and how long it took for them to regain enrollment momentum. Putting all of this information together, we constructed our best estimate of the enrollment curve, which can be found in our statistical report (see below).
- The median lifespan of the control group: Unfortunately, people with Stage 3 NSCLC are almost certain to die. Five-year survival is 15% for patients with NSCLC. Multiple studies on Stage 3 NSCLC patients have been done, and based on current treatments, the median expected survival is about 20 months from the start of treatment. There is a range of median OS from different trials, but the range is relatively tight. It appears that the trial’s assumption of a median OS in the control group of 20 months is reasonable, and we believe the control arm is closely tracking that timeline. We discuss this in more detail below.
- How many people must be alive: This is the easiest metric to measure, because the START trial was designed with predetermined interim looks based on the number of deaths. The analysis of the first interim look, triggered on the 353rd death, was announced on December 22, 2010. Based on our understanding of the speed with which these announcements are made, we think it is likely that this look was reached in October 2010. The second interim look will take place when 529 deaths have occurred and there has not yet been an announcement that the second look has taken place. Based on the two-month estimated delay, this means that as of mid-May fewer than 529 people had died.
Taking all of this data, we created a model of what we believed was happening in the trial and what the implied differences were between the control and treatment arms. We know when people enrolled and how long they were in the trial; we know how many people are in each arm due to the 2:1 randomization scheme; we know about when the 353rd death happened; and we know how many people from the control arm should have died. The remaining deaths must therefore have occurred in the treatment arm. Finally, based on the enrollment curve, we estimated how long patients in the treatment arm are likely to be living.
While our work showed that the drug was highly effective and was almost certainly working, we wanted to be as certain as possible that this is the case, so we hired a biostatistics expert to consult with us and to conduct independent analysis based only on the public information (much of which we have links to in this document). The report was clear that Stimuvax is almost certainly showing statistically significant effects. A link to the report is here.
The only certain conclusion is that the number of events in the trial is taking an incredibly long time given the patient profile. There are two possible explanations: either everyone on the trial is living longer than expected, or the treatment arm is living significantly longer than expected. Based on the timing of the first interim look, the biostatistician calculated that the median overall survival time was 32-39 months, assuming no difference in survival curves between the control and treatment arms. We know the control arm is receiving a placebo, and have multiple studies that have documented their survival patterns. It is possible, but highly unlikely, that they are exhibiting survival times this long. Our consultations with oncologists suggest that the chances of this happening in a large patient population such as the START trial are extraordinarily low. The only reasonable explanation is that the Stimuvax arm is living significantly longer than expected. Our current estimate is that the control arm is exhibiting a median OS in the low to mid-20 month range, and the treatment arm has a median OS around 40 months. This would represent a quantum leap forward in lung cancer treatment.
FDA approval for Stimuvax is the next step. Fortunately the START trial will be helped by both its Special Protocol Assignment (SPA) and Fast Track status. In December 2006, ONTY reached an agreement with the FDA on a SPA for the START Trial. The SPA essentially designates that the FDA has reviewed and approved the Phase 3 trial’s design, clinical endpoints, and statistical analyses. If START reaches its endpoints, meets all compliance measures, and correctly files its NDA, the FDA will approve Stimuvax. In February 2011, the FDA reviewed the changes made to the trial as a result of the clinical hold and amended the SPA, so we can be assured that the clinical hold will not affect the FDA approval process. Furthermore, the START trial has Fast Track designation, which means the study results will be reviewed by the FDA within six months of filing and that Merck can file its NDA on a rolling basis. We expect commercialization of Stimuvax to begin in 2012.
If approved and commercialized, we believe the opportunity for Stimuvax is vast. The FDA’s SPA and Fast Track designations reflect the urgent need for care for NSCLC patients. The only approved maintenance treatments for Stage 3 NSCLC are Tarceva and Alimta, both of which have shown limited improvements in survival and both of which have significant adverse side effects mainly related to toxicity. If Stimuvax shows a 17-month improvement in survival and a better safety profile than Tarceva and Alimta, it would become the standard of care for all Stage 3 NSCLC patients who achieve stable or better disease after induction CRT.
Valuation of ONTY Based on Stimuvax in NSCLC Alone
We began by valuing ONTY based only on its royalty stream in Stimuvax and assuming that Stimuvax only receives an indication for Stage 3 NSCLC. For a variety of reasons which we will detail below, these are very conservative assumptions. We built our valuation based on the size of the NSCLC lung cancer market, where we think Stimuvax will fit in as a treatment, how much we think Stimuvax will cost, and the appropriate multiple for ONTY’s royalty stream on these sales.
The first step in understanding the valuation under this framework is to define the size of the market. Lung cancer is the leading cause of cancer death for both men and women. More people die of lung cancer than of colon, breast, and prostate cancers combined. According to a report by the World Health Organization and International Agency for Research on Cancer, lung cancer (both non-small cell and small cell type) affects more than 1.6 million patients a year, with around 1.4 million deaths annually.
Of the approximately 1.6 million patients, 225,000 are in the US, 288,000 in the European Union, and 85,000 in Japan. These are the three major markets we will focus our analysis on. 85% of lung cancer cases are Non-Small Cell Lung Cancer. Of those, approximately 30% are in Stage 3. Of all Stage 3 patients, 95% are eligible for induction chemoradiation therapy (CRT) and approximately 75% of those reach stable or better disease following induction CRT. That gets to a total of 108,600 patients a year (40,900 in the US, 52,300 in Europe, and 15,400 in Japan) that are eligible to receive Stimuvax.
For Stimuvax penetration, we have assumed 40% penetration in the US and 25% in Europe and Japan. The blended global penetrate rate is 30%. Keep in mind that penetration rates are a function of efficacy, adverse side effects, and competition. There are only two competing drugs out there today, the efficacy of the competition is significantly lower, and side effects of Stimuvax are low since it is a targeted immunotherapy. We believe our penetration rate assumptions are laughably conservative.
For pricing, we assumed $85,000/patient in the US and $65,000 for the Europe Union and Japan. Merck KGaA has not made any indications on pricing strategy, so we used an estimate between the cost of Alimta ($120,000 in the US) and Avastin ($60,000 in the US). With Alimta providing an improvement of only one month and Stimuvax most likely showing at least 15 month improvement in OS, there are strong reasons to believe our pricing assumption is too low. Moveover, Provenge, the first approved cancer vaccine, provides a four-month improvement in median OS and costs $93,000. Pricing tends to be set based on the improvement a drug provides in OS. The rule of thumb in cancer drugs has been $20,000 per month of survival benefit. Our assumptions, which look quite conservative to us, get us approximately $2.5 billion in annual sales for Stimuvax.
In terms of the economics to Oncothyreon, the royalty agreement with Merck KGaA provides a mid-teens royalty on North American sales and a high-single digit royalty on sales in the rest of the world. For all US sales, ONTY pays out a mid-single digit royalty to Cancer Research Technology Limited (CTRL) for the rights to a MUC1 antigen. We estimate the net royalty rate to ONTY is 11% in North America and 8% in the rest of the world. On full penetration, we estimate $241 million in annual royalty payments net to ONTY. The current market cap of ONTY is $380mm on a fully diluted basis, assuming all options and warrants are exercised.
Biopharma companies typically trade at 5x-6x net royalties. Tarceva was purchased for 7x royalties. On a fully diluted basis, a 5x-6x multiple represents $26 to $31 per share. ONTY shares today are $8.00. Every 5% of worldwide penetration is a ± $4 per share change. Every 0.5x change in multiple is a ± $3 per share change. Every $10,000 chance in treatment cost is ± $3.50 per share change.
A numerical analysis of the above is as follows:
As stated earlier, MUC1 is over-expressed in many common cancers. Studies have shown MUC1 is over-expressed in 90% of all adenocarcinomas including breast, lung, pancreas, prostate, stomach, colon, and ovary. We believe the first addressable markets for Stimuvax outside of lung cancer are advanced stages of breast cancer, prostate cancer and colorectal cancer. Roche’s Avastin has established the roadmap for additional indications cancer when it was first approved for colorectal cancer in 2004 and then expanded into lung, breast, kidney, and brain cancers. In 2005, the first full year of Avastin commercialization, sales were $1.1 billion. In 2010, Avastin had sales of approximately $7 billion.
Without getting too much into details and hypotheticals, it’s safe to assume additional indications based on MUC1 expression would greatly expand Stimuvax’s potential. According to WHO/IARC, the worldwide incidence of breast cancer in 2008 was 1.384 million with 458,000 deaths a year. The worldwide incidence of colorectal cancer is 1.234 million with 608,000 deaths a year. The worldwide incidence of prostate cancer is 899,000 with 258,000 deaths a year. Given that Stimuvax would be targeted for similarly advanced stages of cancer as with lung, the number of deaths is a better proxy of the potential market. The total number of deaths per year from breast, prostate and colorectal cancers is 1.324 million. That’s about the same number of deaths per year from lung cancer, which was 1.375 million in 2008 according to the WHO/IARC.
In our discussions with ONTY’s management, they have told us that Merck has Phase 3 trials ready to roll out in additional cancers if Stimuvax’s START trial is successful. They mentioned breast cancer, colorectal cancer, and prostate cancer as areas of interest. This makes us quite hopeful that ONTY and Merck KGaA will ultimately be able to extend Stimuvax’s indication into additional cancers. With indications for Breast, Colorectal and Prostate, Merck KGaA can double Stimuvax’s potential sales.
Expansion into additional cancers results in a share price of $50-60. That’s 6.3x-7.6x from today’s $8.00/share and it doesn’t even consider stomach and ovarian cancers mentioned above, both of which also show MUC1 overexpression.
Our model also assumes a static 30% penetration rate on the target population, which will almost certainly be much too low if Stimuvax shows best in class overall survival. In addition, Stimuvax may prove to be even more effective for earlier stages of cancer, which would further expand the markets within and outside of NSCLC. Seeing 10x upside to ONTY on Stimuvax alone is not difficult.
Beyond Stimuvax: ONTY’s Technology and Its Next Drug Candidate: ONT-10
Lastly, while ONTY receives only a royalty on Stimuvax sales, it owns 100% of the rest of its drug pipeline. Of import is ONT-10, also known as Stimuvax 2.0. It is currently in preclinical stages and expected to begin a Phase I trial within six months. Compared to Stimuvax, ONT-10 has a larger, glycosylated antigen and contains a next generation synthetic PET-Lipid-A based adjuvant. Essentially, it is a better version of Stimuvax because it’s bigger and more powerful. Results from the preclinical animal trials were presented in April 2011 at the American Association for Cancer Research Conference. In the study, ONT-10’s larger size and additional sugars allowed it to outperform Stimuvax in identifying and targeting MUC1 expressing cells. ONT-10’s synthetic PET-Lipid A adjuvant also stimulated a more powerful immune response. (Link to the presentation is here.
It is important to note that Merck KGaA only has the right to first negotiation on ONT-10 if ONTY decides to take on a development or marketing partner. ONTY can choose to develop ONT-10 internally and not trigger Merck KGaA’s right. That’s unlikely since ONTY doesn’t have the requisite development, commercialization, or manufacturing expertise. However, a large, multinational pharmaceutical company with a strong oncology platform, like Roche (RHHBY.PK), Pfizer (PFE) or Bristol-Myers Squibb (BMY) could purchase ONTY and develop ONT-10 internally with no need to license anything to Merck KGaA. The acquirer would then be able to keep all the economics on ONT-10. ONT-10 effectively presents a huge risk to Stimuvax a few years down the road, so we believe that Merck KGaA will be forced into bidding for ONTY if Stimuvax is successful. There is simply no way Merck KGaA can allow a competitor to own the rights to a better version of its coming blockbuster drug. However, in order to win, Merck KGaA will probably have to engage in a bidding war. The value of ONTY’s technology is just too high for other pharmaceutical companies, many of which lack blockbuster compounds in their pipeline, to sit by and let Merck KGaA purchase ONTY on the cheap. This is our final and ultimate scenario.
We have used Stimuvax’s economics as a proxy to estimate the value of ONT-10. Assuming Stimuvax can do $2.5 billion of NSCLC sales (from our calculation above), a 50% pre-tax profit margin, 35% tax rate, and a 10x multiple on after-tax profits of $810 million, ONT-10 would be worth a minimum of $8.1 billion, or $173 per share to ONTY.
If we include additional indications for Stimuvax, the potential value of ONT-10 and therefore ONTY becomes outlandish, and we simply want to run through this exercise to highlight the potential value. Clearly, the success of ONT-10 is tied to Stimuvax. However, Stimuvax’s approval will trigger incredible optionality on the value of ONT-10 for both ONTY and Merck KGaA.
Risks to Investing in ONTY
We mentioned above that the there are two potential explanations for the timing of the trial. The main risk to our analysis is that the control arm is living substantially longer than expected. ONTY’s CEO, Dr. Kirkman, echoed that risk to us when he told us it was the one thing that keeps him up at night. If the control arm is living longer than expected, then the treatment arm is living shorter, and the difference in survival from Stimuvax may not be significant. To gauge this risk, we examined 13 previous studies on unresectable Stage 3 NSCLC patients. The list is below:
Most studies show median OS of this group in the 20-month range. One important aspect to note is that most of these trials start at the beginning of treatment. ONTY’s Phase 3 START trial begins after patients have undergone chemo and have reached stable or better disease. Essentially, ONTY’s trial has a healthier patient population since they have undergone induction chemoradiation therapy and have not experienced disease progression. Roughly 25% of patients drop out during or right after induction CRT. This indicates that the Stage 3 NSCLC population in the START trial should live longer than these studies. However, ONTY’s Phase 3 trial begins counting time of survival at a later point in time than these studies. Approximately four months will have elapsed from the start of induction CRT to beginning of Stimuvax treatment or placement into the placebo arm. Few studies examining overall survival for Stage III NSCLC patients have been done in a similar manner. We have spoken to doctors and they believe the four-month time difference should offset the fact that ONTY’s trial starts with a healthier patient population. An oncologist with experience as principal investigator for lung cancer clinical trials stated she would expect average survival for patients who reach stable or better disease after first line CRT to be 20-22 months.
Of the above studies, only the Kelly K., et al. Phase 3 study documented abnormally long survival times for the control group. The median OS of the control group from the point of randomization was 35.0 months. This is a cause for concern. However, the trial design was very different from the START trial. See the accrual chart of the Kelly trial below:
The Kelly study started with 571 eligible patients. All received concurrent induction CRT and were evaluated afterwards for disease progression. Of the 571 initial patients, 429 patients achieved stable or better disease. This patient population is most similar to the START population. The 25% drop-out rate is consistent with other studies. The patients then received a round of consolidation therapy. No other trials had this sort of trial design, which required two treatment steps before randomization. All of the other trials, including START, required only one chemoradiation treatment before randomization. After induction and consolidation therapy, only 243 of the initial 571 patients in the Kelly study achieved stable or better disease. It was at that point that the patients were randomized into the treatment arm (Geftinib) or the control arm (placebo). In the end, only 43% of the initial 571 patients were even eligible to be randomized. In the first restaging, 75% of patients (429 of 571) made it past induction therapy. In the second restaging, 56% of patients (243 of 429) made it past consolidation therapy. In effect, the study ended up with an extremely healthy patient group by preselecting for the 43% of the healthiest patients whereas Stimuvax’s START trial should have a much sicker population by accepting about 75% of the people who receive induction CRT.
It is also important to note that the median OS for all patients from the start of treatment was 19.0 months, and that the treatment arm (those that received the drug that was supposed to improve survival) had a median OS from time of randomization of 23 months. This further suggests the trial ended up with a control group that is unrepresentative of the greater Stage 3 NSCLC patient population.
Taking the Kelly trial data, we were able to estimate the survival time for the patients who had stable or better disease after induction therapy. From the initial 571 patients, 429 patients achieved stable or better disease after induction therapy. Again, that’s 75% of all comers and very similar to the START patient population. The median OS for all patients was 19.0 months. This includes deaths from the 142 patients that dropped out and deaths from the 429 patients that continued to consolidation therapy. By definition, the median OS was calculated from 286 deaths out of the 571 patient population. The 142 patients who dropped out are likely sicker and have substantially shorter survival times. In other words, nearly all of them likely died and were included in the 286 deaths. That means about 144 deaths came from the 429 patients who achieved stable or better disease. We estimate the 142 patients who dropped out had a median OS of 10-12 months, which would put the median OS of the 429 patients at 26-28 months from the start of treatment.
To make the 26-28 month estimate comparable to the START trial, we must subtract three to four months since the START trial starts counting time of survival after induction CRT and at the point of randomization. We must also account for a difference in induction CRT administration. All patients in the Kelly trial received concurrent CRT (chemo and radiation therapy at the same time). In the START trial, patients can receive concurrent or sequential CRT (chemo, followed by radiation). Concurrent CRT has a documented 3 month survival benefit over sequential therapy but comes with greater toxicity (see Furuse K., et al., 1999). Due to toxicity and different best standard of care protocols, not every patient can receive concurrent CRT. We believe the START population will be split evenly between concurrent and sequential CRT. Therefore, we subtract another 1.5 months from the estimated OS. That gets us to an estimated Median OS of 21 to 23 months for the patient population in the Kelly study most similar to the START trial, i.e. those that achieved stable or better disease after induction CRT, after normalization for time differential and therapy administration. See figure below:
The doctors and statisticians we have spoken with said the Kelly study was a complete anomaly. A similar occurrence could happen but is unlikely. Importantly, the START trial will have enrolled approximately 1,500 patients, nearly three times the number of patients in the Kelly trial. Nevertheless, a significantly higher median OS in the control group is a possibility and we believe there is a 1 in 50 chance of such an occurrence happening. We believe the odds are still substantially in our favor.
Using statistical analysis of available, public information, we were able to show that the START trial has an overwhelmingly high probability of success. We estimate that the treatment arm is exhibiting a 17 month survival benefit, or 40 months survival time versus 23 months for the control arm. If this is true, Stimuvax will be a blockbuster drug for Merck KGaA and ONTY. There has been no drug like this in the history of cancer treatment. We believe ONTY’s royalty stream on Stage 3 LSCLC is worth $26-31/share to ONTY.
Stimuvax also has substantial opportunity for expansion into additional cancer indications and use in earlier stages of cancer. This could push ONTY’s value to $50 per share or above. Furthermore, Stimuvax’s approval would highlight the value of ONTY’s wholly owned, developmental drug, ONT-10, which we believe will attract Merck KGaA and others to bid on ONTY. The value of ONT-10 to ONTY could be worth $173 per share.
There is risk that the START trial is unsuccessful in exhibiting an improvement in survival times. However, we believe it to be highly statistically improbable. Given the exponential potential return on investment and the overwhelming statistical evidence, we believe ONTY presents a once in a lifetime investment opportunity for investors.
Disclosure: I am long ONTY.