Hypertrophic cardiomyopathy ("HCM") is a genetic heart disease mostly caused by mutations in one of the many sarcomeric genes. The most prevalent mutations include genes encoding myosin heavy chain (~30% of the HCM patients), myosin-binding protein C (~20% of the HCM patients), and cardiac troponin T (~20%). The exact mechanism of how these mutations lead to HCM still remains unclear, but pathological pathways may include increased sarcomeric contractility, altered calcium cycling and sarcomeric calcium sensitivity, disturbed biomechanical stress sensing, and impaired cardiac energy homeostasis. Those impacted pathways lead to establishment of clinical manifestation including left ventricular ("LV") hypertrophy, deranged cardiomyocyte energetics, diastolic dysfunction, microvascular ischaemia, enhanced myocardial fibrosis, and multifactorial arrhythmias, which further result in physiological symptoms such as dyspnea, fatigue, chest pain, limited exercise capacity and increased rate of sudden cardiac death ("SCD"). Current or potential targets of pharmacologic therapies for HCM cover wide range of those pathways (Fig. 1), with each having their own limitations.
Figure 1. Targets of pharmacologic therapies for HCM. (Source: Internet)
Being a monogenic, autosomal-dominant genetic disease, HCM is the most common hereditary heart disease with an estimated prevalence of 1:500 in US, resulting in a 700k affected population. However, only a small portion (15%~20%) of HCM patients are clinically diagnosed, with vast majority lacking clear or self-aware symptoms. Current diagnostic methods include echocardiography (ECG), cardiac catheterization, cardiac MRI, genetic testing and trace of family history of HCM. HCM also has a high degree of phenotypic heterogenicity. In fact, it is further classified into two subgroups: obstructive HCM (oHCM) and non-obstructive HCM. Obstructive HCM is usually defined by HCM patients who have a resting or provoked peak left ventricular outflow track gradient (LVOT) of at least 30 mm Hg. Obstructive HCM takes about 70% of all HCM cases.
Myokardia’s leading asset, mavacamten is an allosteric inhibitor of cardiac myosin. It is under phase III clinical trial (EXPLORER) for oHCM patients. In its Phase II PIONEER-HCM trial, mavacamten demonstrated its efficacy in reduction of HCM symptoms and partial normalization of cardiac functions (Fig. 2), either as a monotherapy (cohort A, high dose) or co-administered with background medication like beta-blockers (cohort B, low dose). Although the PIONEER trial only enrolled around 20 patients, neither was it placebo-controlled, the market was intrigued by its unanticipated outcome with the stock price more than doubled on Aug. 7th, 2017. Now, with the release date of EXPLORER trial’s topline data approaching, the market gets active again, with soaring option trading volumes implying a binary outcome.
Figure 2. PIONEER-HCM trial outcome. (Source: Internet)
I. Phase III EXPLORER trial likely meets its primary endpoint
We think it’s quite likely that EXPLORER trail will meet its primary endpoint. EXPLORER is a larger scale placebo controlled trial and it tries to demonstrate a clinical response to mavacamten in HCM patients at week 30, defined by “improvement in New York Heard Association (NYAH) classification and more than 1.5 mL/min/kg increase in peak oxygen consumption (pVO2)” or “no worsening in NYAH classification but more than 3.0 mL/min/kg increase in pVO2”. If we applied the same endpoint towards PIONEER’s week 12 data, we find that around 60% of patients in Cohort A and 50% of patients in Cohort B had a response. The limiting factor for the response is actually pVO2 as 70% of patients in Cohort A and 90% of patients in Cohort B achieved improvement in NYAH classification. The EXPLORER trial started with lower initial doses similar to PIONEER’s cohort B while allowing subsequent dose increment based on the protocol. For the purpose of conservative estimation, we think the active treatment arm of EXPLORER could reach a 50% response rate, similar to PIONEER’s cohort B.
In order to estimate the placebo effect, Myokardia’s management leveraged Gilead’s LIBERTY trial outcome as well as their own MAVERICK trial, as mentioned during their earning call. Note that Gilead’s HCM candidate, eleclazine is a sodium channel inhibitor which works very differently from mavacamten. The LIBERTY trial also has a different design and very different endpoints. Estimation of EXPLORER’s placebo effect based on pVO2 readout from LIBERTY is not necessarily accurate, but we think a 25% estimated placebo response mentioned during their earning call is reasonable. LIBERTY recruited both oHCM and non-oHCM patients. If we assume a real world 7:3 ratio between those two sub-populations, and further extrapolate the data with MAVERICK’s non-oHCM placebo response of 21%, we get a derived oHCM placebo response rate of 27%.
Based on the above estimation, we get an 23% treatment effect (50% treatment response – 27% placebo response).
Potential risk factors for a negative outcome of EXPLORER trial include: 1, underestimation of the placebo effect; 2, safety concerns related to significant LVEF drop (below 45%) observed in 20% of the patients in PIONEER; 3, lower range of body mass index ("BMI") of recruited patients could reduce the treatment effect.
For 1, we think mavacamten’s efficacy on improvement of NYHA classification is apparent. As shown in low dose Cohort B in PIONEER trial (Fig. 3), 90% of patients experienced an NYHA classification upgrade after treatment at week 12, but then 60% patients did revert after stop taking mavacamten for 4 weeks, which clearly demonstrates that such improvement is indeed caused by the treatment rather than the inert dynamic nature. For pVO2 improvement in placebo group, we think this is a realistic uncertainty. One study has shown that exercise training could increase pVO2 by 2.2 ml/kg/min in heart failure with preserved ejection fraction (HFpEF) patients, which is above the clinical meaningful change threshold of 1.0 ml/kg/min. Had there existed drastic difference in life style changes between patients in treatment group and patients in placebo group, the treatment effect could be impaired.
Figure 3. NYHA functional class transition in PIONEER trial. (Source: Internet)
For safety concerns on abnormal LVEF drop, we think EXPLORER’s trial was designed intentionally to dodge this bullet. According to PIONEER’s published data, all patients who have mavacamten plasma concentration below 695 ng/ml maintained an LVEF of 50% or greater. Further, the researchers stated that, “Plasma concentrations above 1000 ng/mL were associated with an exaggerated decrement in LVEF beyond what is necessary to obliterate the LVOT gradient in 4 patients.” As a result, EXPLORER’s protocol which starts with lower doses and subsequently adjusts the dose after measurement of drug’s plasma concentration and LVOT gradient is actually a smart move. Further, all patients who stopped the dosing have their LVEF revert towards the baseline. Keep in mind that all these efforts are still likely to result in a drop-out rate close to 10% in the treatment arm, as indicated by MAVERICK trial which also implemented a similar individualized dosing method, even the company stated that “drop rate for EXPLORER is well below 15%”.
For the third point related to BMI, very few investors are actually aware of this potential source of variation. Most of the patients in PIONEER’s cohort B (so does EXPLORER) were on beta-blockers (BBs) as background therapy, and mavacamten shows significant synergic hemodynamic response (reduction in LVOT gradient on top of BBs). However, previous studies demonstrated that BMI is a strong predictor of patients who lack hemodynamic response to BBs (hazard ratio 2.03 per 1 kg/m2 increase; 95% confidence interval 1.2 to 3.4; p<0.05). The same study also shows that 67% of patient with BMI<25 completely normalized the post exercise gradient by just taking BBs. It is reasonable to expect that, for good BBs responders, i.e. those with a BMI less than 25, mavacamten’s add-on efficacy on LVOT gradient will be reduced. How this reduction in hemodynamic response would impact mavacamten’s improvement of functional measures such as NYAH classification or pVO2 is yet known, but positive correlation between those indicators should be expected in some degree. Further, the EXPLORER trial does not seem to exclude patients with smaller BMIs. The good news is that, the average BMI of the general HCM population seems to be around 29, similar to PIONEER’s recruited demographics. In other words, PIONEER did not intentionally recruit obese patients to manifest mavacamten’s efficacy on top of BBs.
As the company mentioned in their corporate presentation, EXPLOERE trial is 95% powered to show a 25% treatment effect. After taking all the above calculation and risks into considerations, we think there is 75% chance that EXPLORER will meet its primary endpoint.
II. Mavacamten’s real market potential.
In order to assess the mavacamten’s market potential (assuming a positive outcome from EXPLORER), we need to first examine its efficacy in oHCM subpopulation. Currently, the first-line therapy for HCM is BBs such as nadolol and bisoprolol. Similar to mavacamten, BBs are proven to improve functional capacity and quality of life in HCM patients. These benefits are mediated by sympathetic modulation of heart rate, ventricular contractility, and stiffness, leading to improved ventricular relaxation, increased time for diastolic filling, and reduced excitability. Specifically, it was shown BB could improve oHCM patients’ NYAH classification. BBs were also demonstrated to partially normalize post exercise LVOT obstrction in HCM patients..
At a first glimpse, there is quite a lot overlapping on clinical benefits between BBs and mavacamten. However, close examination revealed that mavacamten has some unique features that BBs completely lack of. Firstly, historical analysis indicates that there is a decrease in functional capacity as measured by pVO2 with individuals taking BBs. On the other hand, mavacamten seems to significantly increase pVO2 (Fig. 2). Why is functional capacity so important? pVO2 has been a strong predictor of death in patients with heart failure receiving beta-blockers. Further, a higher percentage of achieved age-gender predicted VO2 is demonstrated to be associated with better outcomes in HCM patients.
The second unique feature mavacamten has is reducing the resting LVOT gradient. BBs are not expected to reduce resting LVOT gradients; rather, their beneficial effect on symptoms stems from a reduction in LVOT gradient rise after exercise. Mavacamten, in contrast, is able to reduce resting LVOT gradient as well as post exercise LVOT gradient, potentially bringing more comprehensive clinical benefits.
In reality, physicians weigh more on functional improvement (i.e. NYAH classification) and obstruction (LVOT gradient) than other indicators, both of which are clearly demonstrated in PIONEER trial. We think mavacamten should be easily accepted by oHCM patients after approval.
With all above evidence, we think mavacamten could be a successful drug for oHCM patients. However, we think its path towards non-oHCM and further the much larger HFpEF population will unlikely to be smooth. The biggest reason is that MAVERICK trial for non-oHCM patients actually failed to meet its functional composite endpoint (Fig. 4). Under the ITT setting, there is hardly any difference between treatment group and placebo. Even under the on-treatment setting, only 35% of the patients had a response to the treatment, far less than the estimated 50% response rate in oHCM patients. Given those evidence, we think a large portion of mavacamten’s clinical benefit may due to its capability to reduce LVOT gradient, which is almost absent in non-oHCM patients.
Figure 4. The negative outcome of composite functional endpoint of MAVERICK trial. (Source: Internet)
Another reason limiting mavacamten’s market potential is that it does not tackle the atrial fibrillation ("AF"), which exhibits in about 20% of HCM patients. Further, there were 5 AF cases in PIONEER study which were believed to be treatment related, with one SAE resulted in treatment discontinuation. AF is known to be associated with HCM-related mortality and morbidity. It may be possible that mavacamten be advised to co-administer with AF related therapies, such as oral anticoagulant and antiarrhythmic medication (e.g. amiodarone).
Assuming mavacamten gets approved for oHCM, it is likely to carry a boxed warning, which limits its market acceptance. The main reason is that, a drop in LVEF after treatment could be confused with the more severe, end-of-stage HCM symptoms (Fig. 5). Due to ultra-contractility and other reasons, a small portion of HCM patients eventually develops myocardial fibrosis, stiffening the tissue and eventually leading to systolic dysfunction which causes LVEF to drop below the normal threshold (45%) and an increased rate of heart failure. It will be hard for physicians to tell whether such LVEF drop is indeed caused by mavacamten treatment, or it is the sign of the end-of-stage disease. So once LVEF drop occurs, one has to stop the treatment, hoping that LVEF would revert.
Figure 5. Stages of HCM. (Source: Internet)
In conclusion, we think it is hard for mavacamten to expand beyond the indication of oHCM.
III. Evaluation of the mavacamten pipeline.
There are around 70,000 diagnosed oHCM patients in US. For the base case scenario, we assume a 30% penetration and $50,000 annual cost, we get around 1B in peak sale. As mentioned above, we think its expansion to other indications will be hard. If we assume a 10% revenue boost from non-oHCM and other patient population, we get $1.1B peak sale resulting a roughly $2.7B evaluation (2.5 multiplication on the peak sale).
For the purpose of focusing purely on mavacamten, we are not going to discuss Myokardia’s other pipelines in this article.
Base case | Best case | Worst case | ||||
Value ($M) | Price/Share | Value ($M) | Price/Share | Value ($M) | Price/Share | |
mavacamten | 2,700 | 57.94 | 4,000 | 85.84 | 0 | 0 |
MYK-491 | 350 | 7.51 | 350 | 7.51 | 350 | 7.51 |
other assets | 100 | 2.15 | 100 | 2.15 | 100 | 2.15 |
Total | 3,150 | 67.60 | 4,450 | 95.49 | 450.00 | 9.66 |
We get a fair valuation of $68/share under the base case scenario if EXPLORER published its positive topline result. Investors should be aware that had EXPLORER failed its primary endpoint, the stock price could drop as much as 80%.
With the current stock price of $60, we think the market has priced in most of EXPLORER’s success. The option market exactly tells the same thing, with out of the money (OTM) call options much less expensive (~170% vol) than OTM put option (400%+ vol). Specifically, we saw a large trading volume of $22.5 puts, indicating that some investors are betting on the failure of EXPLORER trial. According to our analysts, we think there is 25% chance it could happen. That said, we think EXPLORER is likely to succeed and it should not push stock price too much.
Investors should be noted that Sanofi opted out the global rights / loyalties of mavacamten and sold it back to Myokardia for a total of $80M in early 2019, even after PIONEER’s success in 2017. Under the terms of their former agreement, Sanofi had its ex-US commercial rights and could receive 5% - 10% royalty from mavacamten’s US revenue. One could argue that Sanofi assessed the net value of mavacamten’s US pipelines to be definitely less than $1.6B based on its decision, which is way less than our baseline estimation of 2.7B.
The anecdotal reason for Sanofi's exit was that Sanofi wanted to get more involved in the US partnership, but the proposal was rejected by Myokardia. It would actually be a reasonable decision, as Sanofi may think, based on the clinical data, a successful launch of mavacamten in US requires Sanofi's strong sales and distribution team. Such level of collaboration would not happen if Sanofi was only getting a single digit loyalty. Apparently, Myokardia's management did not agree with Sanofi on such perspective.
This was by no means a good news to Myokardia. But we think investors should not overly emphasize on Sanofi's exit. All in all, Sanofi had spent around $230M in licensing fees and equity investment since their partnership in 2016. It got $80M cash back and sold its 4.2M shares at roughly $45/share, which summed up to around $270M at the beginning of 2019. Sanofi clearly was not cutting loss in the deal.
Myokardia has around $417M in cash reserve by the end of 2019. It’s burning around $50M per quarter, so its cash can likely sustain towards the end of 2021. However, it is possible that a positive EXPLORER topline data could incentify the company to offer another round of dilution, since it is probably the most important binary event for the company in the following year.
We are shorting the volatility of the stock by shorting the straddle.
(Author: Weiwei Wang)
