The Last Review?
CEL-SCI Corporation (NYSE:CVM) today announced its second Independent Data Monitoring Committee (IDMC) review this year for the IT MATTERS Phase III Multikine study for head and neck cancer. We believe that the longer than expected duration of the trial combined with the IDMC recommendation to continue its progress are positive for the safety and efficacy of the investigational drug.
We note a number of precedents of IDMC reviews being followed by the halt of a trial due to futility. Bioigen’s aducanumab, NuCana’s Acelarin and the Transgene/SillaJen’s Pexa-Vec Phase III trials were all halted in 2019. These terminations show the willingness of monitoring committees to shut down a study if it fails to show results. While many examples can be found for trials ending early for futility, it is difficult to find examples of a trial being stopped early for exceptional efficacy.
We see the latest IDMC meeting as a favorable milestone and believe the results justify an increase in our probability of success estimate, although there could be other unexpected factors that are delaying the report of the target number of events. The delay in reported events could be attributed to higher than expected dropouts, a greater mix of Stage III vs. Stage IVa patients enrolled, or a higher proportion of a less fatal subtype of head and neck cancer among other unknown and unexpected factors. A slip by Ergomed in a September 25th slide deck suggested that the IT MATTERS trial was 96% complete, and by extension lacks only an additional 11 or 12 events in the active and comparitor arm to achieve completion. While the statistic disappeared in a later version of Ergomend’s slide deck, we anticipate the 298th event to be just around the corner.
CEL-SCI provided some data in a September 10th slide presentation which updated data specific to the study population enrolled in CEL-SCI’s Phase III trial. In its original study filing with the U.S. National Library of Medicine eight years ago, CEL-SCI cited median 3-year overall survival (OS) for their patient population as between 52 and 55% and five year OS as 43%. These statistics are somewhat out of date, especially considering the advances that have been made in immuno-oncology since the trial began.
The updated data was sourced from the Surveillance, Epidemiology and End Results (SEER) program, which is part of the National Institute of Health (NIH) National Cancer Institute (NCI). The group collects and publishes cancer incidence and survival data that represents over a third of the US population. CEL-SCI commissioned an external statistical group to analyze the data and determine an OS estimate for the specific population Multikine is addressing. Below we provide a summary of the data as provided by CEL-SCI:
Exhibit I – Overall Survival (OS) Estimates for Multikine Study Population1
Compared to the initial data cited by CEL-SCI, OS is lower in the updated population, which seems to be at odds with the OS trend for cancer types in general. As mentioned above, the three year OS was from 52 to 55%, while the updated SEER2 data demonstrated an OS of 45 to 48%. At the five year mark, OS was cited as 43% in the U.S. National Library of Medicine filing, as compared to the updated data which indicates a 37% five-year OS. While it is difficult to say why head and neck cancer mortality rates have worsened over time, some ideas have crossed our path. One may be the dramatic decrease in smoking over the last decades, which is highly correlated with head and neck cancer. Changing the prevalence of this risk factor could have impacted the characteristics of the head and neck population. Other factors could include changes in underlying behaviors and disease such as alcohol use, Epstein-Barr Virus (EBV) and Human Papilloma Virus (HPV). In any case, as the population was randomized, both the active and comparitor arms should reflect the same characteristics and not impact the measurement of benefit from Multikine. Our goal with this updated SEER data analysis was to estimate how many deaths should have occurred if it matched the control arm.
If one assumes the median OS as indicated in the SEER data3, for the ~797 patients in the SOC and Multikine + ciz + SOC arms4, this suggests that there should be over 460 events so far. If we assume median OS in the SOC arm and a 10% improvement in OS in the Multikine arm, this suggests about 437 events in September 2019. Even under our conservative assumptions of patients in each cohort being enrolled at the end of the cohort period and at the highest bound of the 95% confidence interval for OS and a 30% improved efficacy, there should have already been over 60 events more than required to complete the IT MATTERS trial. While the longer than expected trial duration does increase the likelihood that Multikine is extending OS, it also suggests that there may be other factors also reducing the number of events as compared to what is expected based on the SEER dataset.
We place a lot of confidence in the IDMC team and their focus on safety and efficacy for the patient. The goal of any IDMC is to ensure that the interests of the patients are well served, the risk benefit ratio is appropriate and the scientific integrity of the trial is maintained. They have the responsibility to recommend early termination for a trial if interim data is sufficiently compelling for the medicine is either notably effective or harmful to the patients.
We believe the threshold required to recommend stopping the trial early for efficacy is extremely high, even when data are supportive. According to a study published by Pak, et al.5, about 0.5% of trials that are terminated early are done so for success. When interim analyses suggest that patient health is at risk or the medicine under investigation is not working, the threshold is lower for recommending early termination as demonstrated in the examples above for Biogen, NuCana and Transgene/SillaJen. Therefore we see an asymmetric relationship between trials halted for superior efficacy and those halted for futility.
Using the SEER data analysis and some conservative assumptions6, we estimated the number of events that should have occurred assuming the control arm matches the SEER analysis and the Multikine arm demonstrates a 30% improvement in OS. We apply this to the upper bound (fewer deaths) of the confidence interval in Exhibit I and calculate a result of over 360 deaths as of September 2019.
The IT MATTERS trial has not yet ended, indicating that the 298 required events have not yet occurred. The events required for trial completion seems to be outside the confidence interval bounds suggested by the SEER data analysis, our own research cited in our initiation and other analyst research. It appears most likely to us that the group enrolled in the trial may be an outlier in terms of OS and Multikine is improving OS. Applying standard logic to the longer than expected trial and noting that that IDMC has allowed the trial to continue, it stands to reason that the longer the trial lasts, the more likely it is that Multikine is increasing OS. Combined with the IDMC’s recommendation that the trial continue until the appropriate number of events have occurred, we are incrementally more optimistic on the success of the IT MATTERS trial and increase our target price for CEL-SCI to $19.00 per share.
We remind investors of the key tenets that support ownership in CEL-SCI which are discussed in detail in our initiation:
‣ Compelling preclinical and clinical data supportive of Multikine’s effective mechanism of action
‣ Multkine is complementary to first line “intent to cure” SOC, in contrast to other monotherapies
‣ Multikine is administered prior to SOC, synchronizing with the preparation period prior to surgery
‣ Differentiated approach that employs multiple proteins for cancer cell identification & destruction
‣ Proprietary manufacturing process, patent protection and anticipated biologics exclusivity
‣ CEL-SCI maintains operation and control of its Multikine manufacturing facility
‣ Multikine source material is abundant human PBMCs
‣ Favorable drug safety profile with no reported drug-related adverse events
‣ Biologic eligible for 12 years of exclusivity in United States
‣ Global rights to intellectual property
‣ Pipeline includes LEAPS platform with additional indications
◦ Rheumatoid Arthritis
◦ Pandemic Flu
◦ Breast Cancer
CEL-SCI is in the final stages of the Multikine IT-MATTERS trial and we believe final events will occur in the next months, based on publically available commentary. As of September, all patients have been in the trial for a minimum of three years in the study and the latest IMDC review may be the final one prior to trial completion. If Multikine is able to meet or surpass its primary endpoint of 10% increase in overall survival as compared to standard of care alone, we see substantial upside to current levels. We increase our target price of $19.00.
DISCLOSURE: Zacks SCR has received compensation from the issuer directly, from an investment manager, or from an investor relations consulting firm, engaged by the issuer, for providing research coverage for a period of no less than one year. Research articles, as seen here, are part of the service Zacks provides and Zacks receives quarterly payments totaling a maximum fee of $30,000 annually for these services. Full Disclaimer HERE.
1. Source: CEL-SCI Corporate Presentation, September 10, 2019.
2. The SEER database estimates includes the same population enrolled in the IT MATTERS trial who received SOC. This includes oral cavity (anterior tongue [not base of tongue], cheek, floor of the mouth and soft palate. Over 3,000 patients were included in the study population to calculate 5 year survival.
3. With a whole host of simplifying, conservative assumptions. Numbers for year cohorts provided by company are all assumed to occur at the end of the period. Five year survival is applied to all patients who have been in the trial for five or more years
4. This assumes median OS for both arms and zero additional benefit from Multikine.
5. Pak, TR; Rodriguez, MD; Roth, FP; Why Clinical Trials Are Terminated. bioRxiv. July 2, 2015. https://www.biorxiv.org/content/10.1101/021543v1
6. We assume that all enrollees enrolled between 2011 and 2014 have been in the trial for five years, patients enrolled in 2015 have been enrolled for four years, and patients enrolled in 2016 have been enrolled for three years.