Northwest Biotherapeutics: Is There Any Chance for DCVax-L to be Successful in the Phase 3 Trial? (NWBO, Buy, $0.24)
Acronyms Used in this Report
- SOC: The standard of care in newly diagnosed glioblastoma patients is surgical resection followed by treatment with radiation and the chemotherapy drug temozolomide
- PFS: Progression free survival is the time from which the surgery is performed until the tumor begins to regrow
- mPFS: The median time at which half of patients in a clinical trial have progressed
- OS: Patients who remain alive from the time of surgery regardless of cause of death
- mOS: The median time at which half of patients in a clinical trial have died
Yes, there is a reasonable chance for success.
Key Investment Issues
Phase 3 Trial is Nearing Completion
The phase 3 trial of DC Vax-L in newly diagnosed glioblastoma is nearing the time at which the data base will be locked, scrubbed and analyzed. We might see the results at ASCO in May, but this is frankly just an outright guess. So what are the chances that the trial will lead to regulatory approval? The stock sells at $0.24 which results in a micro-cap market capitalization of $55 million and the cash position is distressed raising questions as to how long the Company can remain solvent. An investor looking at the company for the first time would almost certainly conclude that stock market participants believe there is virtually no chance that the trial will be successful.
There are Reasons for Skepticism/ Concern
Cancer vaccine technology has produced a long string of failures with only two products approved- Dendreon’s Provenge and Amgen’s Imlygic. Neither has made much of an impact on cancer treatment. Still, the cancer vaccine technology on which DCVax-L is based holds the promise to be a major advance in cancer treatment. In my opinion, it could be every bit as dramatic as checkpoint modulation (think Opdivo and Keytruda) and chimeric antigen receptor T-cells (CAR-T cells). I have discussed this hypothesis at length in prior reports, and it is not the subject of this report. It is also important to understand that almost all cutting edge biopharmaceutical technologies can take years (decades) to perfect and the companies leading the early discovery efforts may not be the ultimate winners.
I recognize that there is a high risk of failure in the phase 3 trial. This could be because the technology does not work or needs to be improved in some way. Perhaps this will entail combining it with checkpoint modulators like Opdivo and Keytruda; indeed phase 2 combination trials with each of these drugs are actually underway). Also, trial design issues can often create a negative outcome in a trial even if the drug ultimately is shown to be successful.
But Patients Appear to be Living Longer in the Trial
Because the results of the trial are blinded to the Company and investigators, we can only guess at what is going on in the trial. The most encouraging observation is that patients appear to be living longer than would be expected for SOC. The trial has reached the number of disease progression events called for in the statistical design, but not for overall survival. I had originally thought that the latter might occur in 2H, 2016, but deaths in the trial seem to have been less than expected and this has prolonged the trial.
I have listened careful to the lead investigator on the trial, Dr. Linda Liau, who is also the innovator of DCVax-L. Dr. Liau gave an extensive speech at the Seattle Science Foundation on December 15, 2016 about DCVax-L. I summarized her remarks in a report called DCVax-L Viewed Through the Eyes of Dr. Linda Liau, Lead Investigator on the Phase 3 Trial of DCVax-L If you wish to listen to her speech in its entirety, here is the link. It is worth noting that she has purposely taken no financial interest in the Company to avoid a financial conflict of interest. She truly seems to be searching for truth although her deep involvement with the drug could cause unintended bias.
Dr. Liau believes that patients are doing better (living longer) than would be expected if DCVax-L were no better than SOC. However, until the trial is unblinded there is no way of determining if she is correct. Based on her experience from working with DCVax-L for nearly 20 years, she believes that DCVax-L produces a long term survival benefit in which as many as 25% of glioblastoma patients live four years or more. Historical data suggests that only 8% of patients treated with SOC are alive at four years. If both estimates are correct this would be a major medical advance. Excitement with Opdivo and Keytruda has been based on what is called a long survival tail in which 15% to 20% of patients treated experience long term survival. Dr. Liau believes that DCVax-L may also have such a long tail effect.
Endpoints of the Trial Could Produce Equivocal Results but This Might Not Preclude Approval
Dr. Liau cautions that the phase 3 trial design for DCVax-L is based on traditional measures intended to determine efficacy for chemotherapy drugs. There is a focus on tumor shrinkage and preventing progression that they may not fully capture the long term tail effect seen with immuno-therapy. The primary endpoint for this trial of median progression free survival may be inappropriate (this is discussed in detail later). Moreover, the cross over design of the trial allowing patients who progress on SOC to receive DCVax-L may confound analysis of the secondary endpoint of median overall survival.
Dr. Liau cautions that the FDA may have to approach this data with more of an open mind than just basing its decision on achieving the primary or secondary endpoints. There may be other measures such as sub-group analysis that might indicate a prospectively defined group of patients who would meaningfully benefit from the drug. The FDA has shown flexibility in granting approvals for Opdivo and Keytruda based on small phase 2 trials without a control group in refractory tumors without evidence of an effect on survival. Indeed, Kite is banking on approval of its CAR-T drug based only on showing tumor shrinkage in about half of patients with DLBCL without evidence of a significant increase in survival.
The phase 3 trial of DCVax-L is larger and more data rich than many other trials that have led to approval. It is not unthinkable that it could be approved even if results for progression free survival and overall survival are equivocal based on statistical analysis. There would have to be other measures that suggest there is a long survival tail. The benign side effect profile of DCVax-L is a big plus if it comes to this.
Design of the DCVax-L Phase 3 Trial; How Do We Judge if the Drug is Effective?
What Kind of Outcomes Can we Expect with SOC
If DCVax-L is successful in this trial, it must improve meaningfully on standard of care in newly diagnosed glioblastoma. This is a combination of radiation therapy and the chemotherapy drug temozolomide following a surgical resection of this brain tumor. This standard of care was established in 2005 based on the Stupp trial. Since then there have been tweaks that have slightly improved outcomes, but no truly significant advances. Based on my research, here is what I think is the outcome for standard of care as determined by length of survival following surgical resection.
|Percentage of Patients Alive After Surgery|
|Time following surgery||% of Patients Alive|
Endpoints of the DCVax-L Phase 3 Trial
The primary endpoint of the trial is median progression free survival which extensive clinical trial data suggests is roughly 6.6 months. The trial is designed to show statistical significance if DCVax-L improves mPFS by four months to 10.6 months. Some investors have questioned if PFS is a hard endpoint and suggest that only OS will be accepted by the FDA. This is not a worry in my opinion. Glioblastomas expand aggressively in the confined area of the skull, impinging on surrounding areas of the brain and in some cases pushing parts of the brain into the brain stem. This impingement results in crippling effects. Delaying this progression would have a dramatically positive effect on quality of life. In highly aggressive cancers comparable in severity to glioblastoma, the FDA has approved drugs based on mPFS and in some cases just by showing they can shrink the tumor or prevent it from growing. If DCVax-L hits this mPFS endpoint, it would almost certainly be the basis for approval.
The secondary endpoint of the trial is median overall survival. I judge that SOC produces a mOS of 16.3 months. Generally speaking, a 4.5 month increase in mOS in almost any cancer type is considered a major advance. One of the issues with this trial design is that patients who progressed on SOC could switch to DCVax-L. The trial was originally designed to give two-thirds of patients DCVax-L and one-third SOC. Because of the cross over design, 84% of patients enrolled received DCVax-L at some point. This could complicate the analysis of the trial because only 16% of patients were on SOC. This could create complications in statistical analysis of the results.
The Stupp Trial Established Current Standard of Care in Newly Diagnosed Glioblastoma
The standard of care in treating newly diagnosed glioblastoma is based on a combination of radiation and the chemotherapy drug temozolomide. SOC was established based on a clinical trial that was reported in 2005 in the New England Journal of Medicine (NEJM). The previous standard of care was radiation. In the Stupp trial, radiation plus temozolomide was given to 287 patients and compared to 286 patients given radiation alone. For those who are interested, here is the link.
The median overall survival was 14.6 months for those treated with radiation plus temozolomide versus 12.1 months for radiation alone or a two month improvement. The median progression free survival was 6.9 months for SOC versus 5.8 for radiation. The two year survival rate as reported at the time was 26.5% for the now established SOC of radiation and temozolomide versus 10.4% for patients treated only with radiation. Follow-up data showed that the three year survival rate was 16% and that 10% were alive at five years.
Roche Trial Comparing Avastin Plus SOC to SOC
In February 2014, Roche reported on a trial which compared Avastin plus SOC (320 patients) to SOC alone (317 patients). Median overall survival for Avastin plus SOC was 15.7 months as compared to 16.1 months in SOC. There was no survival benefit for adding Avastin. Note that mOS for SOC increased to 16.1 months as compared to 14.6 months in the Stupp trial. In terms of mPFS Avastin plus SOC was 10.0 months as compared to 7.3 months with SOC which compared to 6.9 months in the Stupp trial.
Second Roche Trial Comparing Avastin Plus SOC to SOC
The second trial of Roche enrolled 458 patients on SOC plus Avastin and 458 on Avastin alone. The mOS was 16.8 months in the Avastin plus SOC group and 16.7 months in SOC. The respective mPFS was 10.6 months versus 6.2 months.
Seventy-two percent of people treated in the Avastin plus SOC arm were alive at one year compared to 66 percent of people in the SOC arm. Thirty-four percent were alive at two years compared to 30 % in SOC.
Internet Literature Search
There are all kinds of publications on the web that address expected survival for newly diagnosed patients treated with SOC. One of the best that I found was at this link. This paper looked at several studies addressing how long patients with newly diagnosed glioblastoma might live. The studies were of varying quality and number of patients involved and some looked at patients before SOC was established in 2005 with the Stupp trial. My overall assessment with which you may disagree after reading this article is that it suggests the survival at various time periods of:
- 35% at two years,
- 15% at three years,
- 8% at four years, and
- 5% at five years
My Conclusions on Outcomes That Can Be Expected with SOC
I think clinicians improved results with SOC from the time when the Stupp trial was conducted to the time when the Avastin trials were conducted and I would put more weight on results reported from the Avastin trials. I would suggest that mOS for SOC is about 16.1 to 16.7 months. Similarly mPFS for SOC is about 6.2 to 7.3 months. One could conjecture that SOC results have improved even more since the Avastin trials, but there is no data to support that.
Putting together all the survival data, I would suggest that the death rate at various time points for SOC is as follows:
- One year: 66%
- 1 to 16.7 months: 50%
- Two years:30%
- Three years:16%
- Four years: 8%
- Five years:5%
Important Factors About the DCVax-L Trial
The primary endpoint of the trial is median progression free survival and the trial should produce statistically significant results if there is a four month improvement in mPFS for DCVax-L. As just explained, I estimate that mPFS for SOC will be 6.2 to 7.3 months so that mPFS for DCVax-L would have to be 10.2 to 11.3 months. The original plan was to stop the trial and analyze the data when there were 248 PFS events. The Company has indicated that this number has now been reached. The Company and investigators remain blinded and do not know the distribution of PFS events between the DCVax-L arm and the control arm.
The secondary endpoint of the trial is overall survival and this calls for 233 patient deaths. This number has not been reached and it appears that the company is waiting for the 233rd death to occur before locking and analyzing the database. This is being done even though the 248 PFS threshold has been reached. For anyone who might see something nefarious in this, the trial remains completely blinded. My judgment is that based on looking at the number of deaths in the trial (which is known) that they have concluded that overall survival with DCVax-L –plus SOC as compared to what might be expected with SOC looks very good.
The trial was originally planned to enroll 348 patients. However, the screening halt that the FDA put in place in August of 2015 came after 331 patients had been enrolled. The FDA subsequently lifted the hold on February 6, 2017 which would have allowed NWBO to complete the enrollment. However, this would have caused the trial to be prolonged significantly so that they elected to stop enrollment at 331.
The first patient in this trial was dosed nearly ten years ago in 2007. There were some fits and starts in the enrollment process based in part on financial issues that caused the trial to be suspended in 2009 and then restarted in 2011. When the trial was restarted, enrollment was initially slow because of a pervasive skepticism about immuno-oncology. This changed quickly as investigators became excited about the promise of immuno-oncology following the approval first of Yervoy and then Opdivo and Keytruda so that beginning in late 2013, enrollment took off.
It is not possible to figure out an enrollment curve based on this limited information. However, it indicates that a small, but meaningful group of patients were treated in the 2007 to 2009 time period and probably a small number in the 2011 to mid-2013 time frame. The bulk of patients were treated in the mid-2013 to mid- 2016 time frame. Remember that the FDA halted screening in August 2016 so I would guess that the last patient began treatment in 4Q, 2016.
This means that almost all of these patients have meaningfully surpassed the 16 month threshold that seems to be the best estimate for mOS that can be achieved with SOC. All of this adds to the argument that the trial has a more mature data set that should bias the outcome toward a higher death rate. Intuitively, 29% of patients remaining alive at this point in the trial suggests to me that DCVax-L may be having a positive effect on survival. It is suggestive of a long survival tail that is a trademark of immuno-oncology therapies.
In a recent 8-K, NWBO stated that 421 patients were treated with DCVax-L in the course of running the phase 3 trial, which is 90 more than the 3331 patients who were enrolled. This number includes 55 patients in the Information arm and 32 pseudo-progression patients. These patient were prospectively excluded from the trial because they appeared to progress before they received DCVax-L. In addition, there are a considerable, but as yet unknown number of patients who have received DCVax-L on a compassionate use basis. Results in these patients could provide meaningful additional information about DCVax-L.
Trial is Running Longer than Anticipated
Management said that based on the pace of deaths occurring during the last six to eight months, they anticipate that it could be several months until the trial reaches 233 deaths. I am guessing that this means we might see the topline data as a late breaker at ASCO in May 2017 or perhaps in the summer
Potential Complexities in Analyzing Data from the DCVax-Trial
I think that when the data is analyzed, there are going to be a number of complex issues involved in assessing the efficacy of DCVax-L. Let’s think first about the primary endpoint of progression free survival. This is measured by using MRI or CT imaging to detect changes in the size of the tumor. If the tumor is getting larger, this will be picked up on the scans and judged to be progression. However, immune therapy like DCVax-L is intended to create an immune response against the tumor by lymphocytes, particularly T-cells. These cells release chemokines which in attacking and destroying the cancer cell result in inflammation which can be mistaken for progression of the cancer.
It is not always crystal clear as to whether the tumor is progressing or that the apparent enlargement actually means that DCVax-L is working. The latter is a situation is called pseudo-progression. Paradoxically, these are probably the best responding patients, but they can be mistaken for patients experiencing rapid progression. There are algorithms that have been developed to distinguish between pseudo-progressors and rapid progressors. These have been refined meaningfully over the ten years that this trial has been in progress. However, it may be the case that some unknown number of patients were improperly categorized. This might have been more the case early in the trial.
This raises a number of questions. Could patients who were psuedoprogressors perhaps been taken off the drug mistakenly. Could investigators have mistaken this and classified patients as rapid responders and even taken them off therapy. For the purpose of regulatory submission, the scans will all be analyzed by a central group of blinded physicians so that the criteria used to classify them will be uniform. This could lead to changes in the assessment of mPFS in some patients? How would the FDA respond to data that was interpreted as showing that the patient progressed but who in actuality was responding positively to the drug?
As I have previously mentioned, the cross over design will complicate the analysis of survival. Based on the comments by Dr. Liau, 84% of the 311 patients in the trial have received DCVax-L. The trial design was to randomize two thirds of patients to DCVax-L plus SOC (about 208) and one third to SOC (about 103). However, 84% of 311 patients or 261 received DCVax-L. This indicates that 208 patients were started on DCVax-L; 50 received only SOC and 53 were started on SOC and after progression of their tumor were given DCVax-L and temozolomide. This could cause issues with the statistical analysis.
Patients treated in this trial all suffered diagnoses with glioblastoma, but they were far from a homogeneous group. There were different baseline characteristics such as age, sex, degree of resection (how much of the tumor could be safely removed from the brain) and other issues. As with all tumors, specific mutations that can lead to meaningfully different outcomes. There is also the possibility that some tumors were actually astrocytomas which have a much better outcome and were mis-diagnosed as glioblastomas.
MGMT methylation status is the most predictive variable for survival correlating with the use of temozolomide. MGMT is a DNA repair enzyme that essentially undoes the work of temozolomide which damages the DNA of cancer cells and interferes with DNA replication. Methylation of the MGMT gene reduces the amount of the MGMT enzyme produced so that it cannot undo the work of temozolomide. Clinical data indicates that methylated MGMT patients have a median overall survival of 22 months while unmethylated MGMT patient had much shorter survival of 13 months.
Another important mutation occurs with The IDH1 gene that encodes for the enzyme isocitrate dehydrogenase 1. The mechanism of action of this enzyme is too complex for me to try to explain in this note. Please take my word that glioblastomas with high expression more readily disperse into surrounding healthy tissues increasing the invasiveness.
Possible FDA Action
You can see that the analysis of the trial data may be complicated. I am skeptical that the topline data will be the sole basis to determine that the trial was or was not successful. Whether or not the primary endpoint of mPFS is achieved, the FDA will want to see confirmation or support from mOS data. Because of the complications in this data arising from the cross over, there may not be a clear statistical distinction between DCVax-L plus SOC and SOC.
The FDA will also want to look at outcomes subgroups such as MGMT methylation and IDH expression in an effort to identify is some patient types respond better than others.
Tagged as DCVax-L Phase 3 trial in newly diagnosed glioblastoma, Northwest Biotherapeutics Inc., NWBO + Categorized as Company Reports, LinkedIn