Follow Us GraphicFacebook IconTwitter IconLinkedIn Icon
Search Graphic

Expert Financial Analysis and Reporting

Northwest Biotherapeutics: DCVax-L Viewed Through the Eyes of Dr. Linda Liau, Lead Investigator on the Phase 3 Trial of DCVax-L (NWBO, $0.35, Buy)

Principal Takeaway Messages  from This Report

Dr. Linda Liau believes that use of DCVax-L can result in 25% of glioblastoma patients living four years or longer. This therapeutic effect is comparable to that seen with the checkpoint inhibitors Opdivo and Keytruda in certain refractory solid tumors and in her opinion should result in approval. She does express some concern about the phase 3 trial design, but believes the product could be approved even if it doesn’t reach the primary endpoint of median progression free survival.

Relying heavily on Dr. Liau’s feelings on DCVax-L, I roughly estimate that there is a 63% chance that DCVax-L will be approved based on phase 3 results, a 25% chance that the trial fails and a 12% chance that a new confirmatory trial will be required for approval. I would caution you that there are a lot of estimates that underlie these numbers and I would take them with a grain of salt. Still, they support my intuitive belief that the chances for approval are meaningfully better than 50%.

The current market valuation of the stock is roughly $72 million. Approval of DCVax-L could drive the valuation to well over $1 billion.

Who Is Dr. Linda Liau

Dr. Liau is the principal investigator for the pivotal phase 3 trial of DCVax-L in glioblastoma. She has been working on DCVax-L for nearly 20 years and has vastly more experience with the drug than anyone else on the planet. Her academic credentials are extremely impressive as one might expect from a skilled brain surgeon.

She is currently: (1) the Vice Chair of Academic Affairs, Department of Neurosurgery at UCLA; (2) Professor, Department of Neurosurgery, UCLA School of Medicine; and (3) Professor and Director of the UCLA Brain Tumor Program. Her clinical expertise is in intra-operative functional brain mapping and use of intra-operative imaging for resection of brain tumors (gliomas, meningiomas, and metastatic tumors). Her research efforts are focused on the molecular biology of brain tumors, gene therapy, immunotherapy, and brain cancer vaccines. She has treated many patients with DCVax-L.

Dr. Liau’s academic credentials are extremely impressive:

  • Undergraduate degree from Brown University with B.Sc. in biochemistry and B.A. in political science
  • MD degree from Stanford University.
  • PhD. in neuroscience from UCLA
  • MBA from UCLA

If you are interested in going over her long and impressive list of publications, follow this link.  Dr. Liau recently spoke about her experience with DCVax-L and thoughts on the current phase 3 trial in a presentation at the Seattle Science Foundation on December 15, 2016. You can see her speech at this link This offers a unique opportunity to understand her thoughts on DCVax-L and the potential for approval based on the phase 3 trial that is nearing completion. Rarely can investors gain such detailed understanding of the thoughts of a lead investigator on a critical trial. This note summarizes the key points that I took away from her speech. If Dr. Liau is correct, DCVax-L will be approved on the basis of the phase 3 trial.

Key Points Made by Dr. Liau in Her Presentation

  • The phase 3 trial of DCVax-L stopped enrollment at 331patients in November 2016. Roughly two thirds of these patients initially received DCVax-L and one third were given a placebo.


  • The trial will be stopped and data analyzed when 248 events (either disease progression or death) have occurred. These 248 events have not been reached so the trial is ongoing.



  • The trial remains blinded so that neither the Company nor investigators know which patients received DCVax-L and which placebo. Very encouragingly, Dr. Liau says that the overall results of the 331 patients indicate that patients are doing better than would be predicted by historical studies. However, until the trial is unblinded there is no way of determining if there is a difference between the DCVax-L treated and placebo patients.


  • Based on experience gained through working with DCVax-L for nearly 20 years, Dr. Liau believes that DCVax-L produces a long term survival benefit in which as many as 25% of glioblastoma patients live four years or more. If so, this would be a strikingly positive result. A long term effect of this magnitude would be similar to what has been seen with other immuno-oncology drugs like Opdivo, Keytruda, Tecentriq and Yervoy in other difficult to treat tumors that has so electrified oncologists and investors.



  • She sounded a cautionary note that the proscribed endpoints for the trial which are median progression free survival or mPFS (primary endpoint) and median overall survival or mOS (secondary endpoint) may not be the right endpoints for the trial. They don’t capture the profound benefit that occurs long after medians have been passed. Endpoints based on medians are more appropriate to chemotherapy, targeted therapies or monoclonal antibodies (which have different mechanisms of action) and may not be able to capture the long term effect that characterizes immuno-oncology drugs. She worries that this could be an issue for the trial.


  • She also notes that patients who were started on placebo and whose disease progressed were allowed to switch to DCVax-L. This confounds the overall survival analysis as 67% of patients initially received DCVax-L; 19% of control patients received DCVax-L at a later time after their disease progressed; and 14% did not receive DCVax-L. This could confound the analysis of overall survival.



  • Results for 20 rapid progressors in the information arm of the DCVax-L trial are quite encouraging. These indicate that the mOS for 20 rapid progressors was 15 months as compared to 8 to 9 months that would be expected based on historical results. Note that results from the information arm are not included in the phase 3 trial.


  • Results for 25 psuedoprogressors in the information arm were extremely impressive. In this arm, mOS was 21.5 months and ten of 25 patients (40%) are alive at four years with no disease progression. It is believed that psuedoprogressors experience longer than normal survival which is 15 months, but there are no studies that indicate what mOS should be expected for that group.


  • Liau cited data from 21 other studies done apart from NWBO involving 430 patients who were given autologous dendritic cell vaccines for glioblastoma. The median overall survival in these studies ranged from a low of 15 months to a high of 53 months. This compares to expected survival from standard of care of 15 months. This provides support from unrelated third parties that dendritic cell cancer vaccines do produce a meaningful therapeutic effect in glioblastoma.



  • She reiterated that the phase 1/2 data that encouraged NWBO to conduct a phase 3 trial was compelling. It showed that DCVax-L produced mPFS of 25 months in 20 newly diagnosed glioblastoma patients which compared to historical results of 7 to 8 months. The mOS was 36 months versus 15 to 16 expected with standard of care and 55% of patients were alive at three years versus 16% expected for standard of care.


  • The 331 patient phase 3 study will be extremely information rich. It will look at many different sub-groups of patients based on genetic mutations and other biological markers in an effort to find biomarkers that prospectively identify patients most likely to benefit. It could be the case that DCVax-L produces a striking benefit in some genetic mutations that cause glioblastoma while having little or no effect in others.



  • There is still another trial connected to but separate from the phase 3 trial and the information arm of the phase 3 trial. From 2012 to 2013, thirty two psuedoprogressors were randomized 2:1 versus placebo. Unlike the information arm this trial is blinded. It is intended to determine the effect of DCVax-L specifically on psuedoprogressors. This could be a very important trial that would support the phase 3.


  • Liau along with many key opinion leaders is enthusiastic about the potential synergy between the checkpoint inhibitors and DCVax-L. A phase 1/2 trial of Keytruda combined with DCVax-L was just started in metastatic colon cancer. Dr. Liau let slip that a phase 1/2 trial of Opdivo with DCVax-L will start soon; NWBO has not yet announced this trial publicly.

My Assessment of the Investment Significance of Dr. Liau’s Conclusions

Commercial Opportunity is Huge

If Dr. Liau is correct that DCVax-L leads to three to four years of survival for perhaps 25% of glioblastoma patients and it is approved, DCVax-L would be a major medical breakthrough and a commercial blockbuster. Here is why I say that. The checkpoint inhibitors Opdivo, Keytruda, Yervoy and Tecentriq have shown comparable survival benefits (15% to 20% of patients surviving for two to three years) in several refractory cancers with little or modest benefit in others. This has led to great excitement in the treatment of such cancers with these drugs. Glioblastoma is comparable to such refractory cancers in the inability of physicians to effectively treat it. It is a rapidly growing cancer that is confined within the skull and as it grows, it crushes parts of the brain resulting in severe adverse effects and short survival times. There is a very significant unmet medical need for effective therapy.

The commercial success of the checkpoint inhibitors has been extraordinary. The first such drug (Yervoy) was only introduced in 2011 and sales of these four drugs exploded to about $6 billion in 2016. Wall Street projections are that sales of checkpoint inhibitors and agonists (none yet approved) are projected to reach $33 billion in 2022. I would expect a similar explosive sales trajectory for DCVax-L if it is approved in glioblastoma. I think that US sales of $1 billion within three to four years would be a reasonable expectation.

The Mechanism of Action of DCVax-L and Checkpoint Inhibitors

A therapeutic cancer vaccine like DCVax-L has a very different mechanism of action from checkpoint inhibitors. It vastly increases the number of killer T-cells that are directed against antigens in a tumor. The checkpoint inhibitors block cancers from downregulating T-cells in the tumor. There is growing belief that there will be a natural synergy between the two mechanisms of action and indeed, a phase 1/2 trial of DCVax-L with Keytruda in metastatic colon cancer is underway and another phase 1/2 trial with Opdivo will start shortly. Dr. Liau is quite optimistic that these trials will be successful. Very importantly, from both a commercial and medical standpoint, the mechanism of action of DCVax-L like the checkpoint inhibitors applies to most solid tumors. Unlike targeted therapies, they are not just applicable to a small number of tumors that express a specific antigen. This makes for huge commercial opportunities.

Why is Dr. Liau So Positive?

Dr. Liau bases her conclusions on: (1) her long personal experience with DCVax-L, (2) the results seen in the phase 1/2 trial of DCVax-L, (3) the results in the information arm of the phase 3 trial and (4) studies with autologous, dendritic cell vaccines performed at other centers. In addition and importantly, the phase 3 trial has been ongoing for eight years and while she is blinded to individual patient results, she has seen some patients who experienced long survival. This is presumably due to DCVax-L but we won’t know until the trial is unblinded.

She Expresses Concerns about the Phase 3 Trial Design

Dr. Liau has expressed concern that the endpoints of the trial which are based on mPFS (primary endpoint) and mOS (secondary endpoint) are not appropriate for measuring the long term benefit of DCVax-L. These median endpoints are reached in months and the effect of DCVax-L is established over two to three years. At the median times, its effect may not be apparent.

She also expressed a concern that allowing patients in the control group to switch to DCVax-L after their cancer has progressed could confound the analysis of median overall survival.

Could the Phase 3 Trial Fail to Reach Its Proscribed Endpoints?

There are many examples of drug studies which fail because of the trial design even though the drug is eventually shown to be effective against a disease. There is a possibility that the phase 3 trial of DCVax-L might not be successful because of these trial design issues.

Could DCVax-L Still Be Approved if the Trial Fails to Reach its Proscribed Endpoints?

If DCVax-L does not achieve its primary endpoint of median progression free survival and if the cross-over allowed in the study obscures the effect on overall survival, does this mean that DCVax-L would not be approved?  This is a very real possibility. The FDA is exceedingly stringent on insisting that proscribed endpoints be reached in a trial in order to gain approval. It virtually never allows a Company that has conducted a failed trial to go back and look at how different subsets of patients fared or whether the trial would have been successful if some other endpoint had been used. This type of analysis is called data dredging and the FDA in the past has in almost every such case rejected attempts to gain approval on this basis. The agency almost always insists on doing a new trial with new endpoints.

Is there a possibility that the FDA might break precedent and approve DCVax-L even if it fails to meet the primary endpoint of mPFS and mOS is unclear? Based on my experience, the answer would be a resounding no. However, Dr. Liau believes that this is within the realm of possibility. Here is how the argument to the FDA would go.

The FDA is well aware that immune-therapies are very different from chemotherapy, monoclonals and targeted therapies and the argument that traditional clinical trial endpoints are not appropriate. They seemed to take this into account in the approval for checkpoint inhibitors in certain refractory tumors. In some cases, they have granted approval on the basis of relatively low objective response (tumor shrinkage) without evidence of long term survival. Moreover, these approvals have sometimes been based on small, open label trials without a control arm. I believe the reason was that the FDA based on anecdotal evidence and prior experience believed there would be a long term effect and patients with these types of refractory cancers had no other options. In these cases, the FDA seemed to be making every effort to get a potentially life-saving therapy on the market even if it meant bending the rules.

The phase 3 trial of DCVax-L is going to be very information rich. It will give some insight into sub-groups that may suggest that it is particularly useful in some genetic mutations causing glioblastoma while not in others. Even though such data might not be statistically significant, it could still be compelling and lead to a (conditional) approval.

In regard to survival being confounded by the cross over allowed, I would note that several well conducted trials show that the expected median overall survival for standard of care is about 15 to 16 months. The agency will look closely at the length of survival in the 67% of patients given DCVax-L at the start of the trial and the 19% of patients who were given DCVax-L after their disease progressed. Hopefully this data will be compelling when compared to expected median survival of 15 to 16 months. Moreover, because of the very long time (eight years) taken to conduct this trial, there may be some compelling data on long term survival.

There is one other aspect to DCVax-L that works very much in its favor and that is an exceedingly safe, even benign side effect profile. Almost all other cancer therapies, checkpoint inhibitors included, come with life threatening side effect potential. The major side effects of DCVax-L are mild injection site reactions and a fever that can be treated with Tylenol. Because drugs are approved on the basis of benefit to risk, having almost no risk is a major positive.

Assigning Probabilities to Possible Outcomes for the Phase 3 Trial

There are a very large number of possible outcomes, but I have limited the number to just four. I have then assigned a probability for each possibility. I would be the first to criticize this approach as simplistic (some would say simple minded) but it is the only way I can convey my thinking. Here goes:

Outcome 1: DCVax-L reaches its proscribed endpoint of median progression free survival and the secondary endpoint of median overall survival. Estimated probability is 20%.

Outcome 2: DCVax-L misses on primary endpoint of progression free survival but achieves secondary endpoint of median overall survival. Estimated probability is 20%.

Outcome 3: DCVax-L misses on both the primary endpoint of mPFS and secondary endpoint of mOS, but the data on mOS is strongly suggestive that both patients who were started on DCVax-L and those who were switched from placebo to DCVax-L are meaningfully benefitted. Estimated probability is 35%.

Outcome 4: DCVax-l misses on both the primary endpoint of mPFS and secondary endpoint of mOS and the data on overall survival is not convincing. Estimated probability is 25%.

What Will the FDA Do for Each Outcome?

I would point out that the review of the BLA for DCVax-L will be coming sometime after the Trump administration has appointed a new FDA Commissioner. One of the people being considered is James O’Neill; he is now a venture capitalist who has no background in science or drug approvals. He believes that drugs should be approved if they are shown to be safe and that their efficacy would then be determined in actual medical practice. I think that it is highly unlikely and undesirable that such a policy could be implicated. However, even considering this person signals that Trump may select someone who will make it easier for a drug like DCVax-L to gain approval. Trying to factor this into account, here are my subjective assumptions on what the FDA might do if faced with any of the above four outcomes:

  1. In the case of outcome 1 there is nearly 100% chance of approval.
  2. Outcome 2- I estimate an 80% chance of approval.
  3. Outcome 3: The chances are probably 35% that the FDA would ask for a new trial and 65% that the FDA would grant approval.
  4. Outcome 4: There is no chance for approval.

If you weigh the potential outcome of each possibility, it would suggest that the probability of approval is 63%, the probability of having to conduct another trial to gain approval is 12% and the probability of failure is 25%. I would take these numbers with a grain of salt, but they are reflective of my belief that the odds for approval are meaningfully better than 50%.


More In-depth Discussion of Dr. Liau’s Comments in Her Speech


Her Update on DCVax-L Phase 3

Northwest Biotherapeutics is nearing completion of a phase 3 trial in glioblastoma that is comparing the results of its therapeutic cancer vaccine DCVax-L added to standard of care (chemotherapy plus radiation) to standard of care. Glioblastoma is an extremely aggressive cancer for which current treatment is grossly inadequate. There have been no advances in treating this disease since 2005 when the chemotherapy drug temozolomide was approved. There is an urgent unmet medical need for more effective therapy. If DCVax-L gains approval based on this phase 3 trial, DCVax-L would likely and quickly become a drug selling well in excess of $1 billion in the US.

The trial was held in 50 sites throughout the US, Canada, UK and Germany. Key points about the trial were:

  • At the time of surgery parts of the glioblastoma tumor were harvested and sent to Cognate to begin preparation of the DCVax-L vaccine.


  • Following surgery, patients were treated for roughly six weeks with chemotherapy and surgery. They were then randomized to DCVax-L or a placebo.



  • Two of three patients were initially randomized to DCVax-L.


  • The remaining one third (control or placebo group) received an injection of mononuclear cells from their blood.



  • The primary endpoint was progression free survival.


  • A key secondary endpoint was overall survival.



  • The causes of glioblastoma are heterogeneous. The trial is stratified to look at results for several distinct mutations that may lead to different forms of glioblastoma.


  • Investigators and sponsors are blinded until the trial is completed and a data lock occurs.


Current Status of the DCVax-L Phase 3 Trial

The trial closed enrollment in November 2016 with 331 patients enrolled. Investors have been puzzled as to why enrollment was haled last August. Disappointingly, Dr. Liau offered no insight into this. The FDA at that time allowed patients already enrolled to complete the trial, but since then has not allowed any more patients to be enrolled. Hence the enrollment fell somewhat short of the 348 patients planned for. The trial will be stopped when 75% of (331) patients have had an event which can be either death or disease progression. Dr. Liau said that the 248 events required to stop the trial has not yet been reached. When this does take place, the trial will be halted and the data will be collected and analyzed.

Dr. Liau did not predict when the 248 events might occur and following that when investors might see the full results. Obviously, I have no insight on when the 248th event will transpire. I do think that once this happens, it could take several weeks to release the data. The Company will want to completely analyze the data, especially on the genetic mutation subsets.  If I were management, I would want to make complete and full of disclosure and leave as few ambiguities as possible. I would not rush out with information just on results on the primary endpoint of progression free survival. I am guessing that the reporting on comprehensive phase 3 results could occur in late 1Q, 2017 or 2Q, 2017 but this is just a guess.

Phase 3 Trial Remains Blinded, But All Patients Seem to be Doing Better than Expected

The trial remains blinded to the Company and investigators at this point so that in common with investors, Dr. Liau has no idea as to how individual patients in the trial are doing, i.e. those who received DCVax-L or placebo. She can, however, see the overall results, i.e. progression free survival and overall survival for the entire 331 patients. Based on this, she says that the whole group is doing better than expected as compared to historical results, but didn’t quantify this. However, she doesn’t know whether the patients in the DCVax-L arm are doing better and if so how much better than control patients.

One possible reason as to why the whole group is doing better is that in order to encourage patients to enroll in the trial, patients who were first given the placebo were allowed to be switched over to DCVax-L if their cancer progressed. About 67% of patients initially received DCVax-L and 33% placebo. Another 19% of patients who initially received the placebo were later given DCVax-L after they progressed. Hence the trial results will compare 67% of patients who received DCVax-L initially, 19% who received DCVax-L at some later time and 14% of patients who did not receive DCVax-L at any time.

Information Arm of Phase 3 Is a Separate Study of 55 Patients

Northwest and Dr. Liau designed this trial almost eight years ago. In the interim there are important things that they have learned that will help to better design future trials. In designing the phase 3 trial, patients were excluded if their first MRI given about six weeks after surgery indicated that the tumor had increased in size. The rationale for excluding these patients was that their tumor was growing rapidly and they would not have enough time to mount an immune response.

All patients in which imaging showed an enlargement of the tumor were thought to be rapid progressors and excluded. However, investigators then learned that detection of early rapid progression is difficult due to the issue of psuedoprogression. This occurs when patients experience a very strong immune response which causes inflammation in the tumor that causes enlargement and can be confused with progression. This makes it appear that the tumor is progressing. In point of fact, these patients are probably the most likely patients to benefit. These are called psuedoprogressors.

In the early part of the trial, both rapid progressors (who had a poor prognosis) and psuedoprogressors (who had a good prognosis) were both excluded from phase 3 and put in the information arm. Because the vaccine had already been prepared, they were given DCVax-L, but their results will not be analyzed as part of the phase 3 trial. Patients were enrolled in the informational arm from 2008 until 2012. In and of itself, this is a separate and informative trial of 55 patients.

The decision to put patients in the information arm was based on imaging performed six weeks after radiation therapy had been completed. If there was a 25% increase in the size of the tumor or if a new lesion greater than one centimeter was detected, they were excluded from the phase 3 study and given DCVax-L. Patients were then re-imaged two weeks later to confirm actual disease progression. This resulted in a group of patients, some of whom were rapid progressors and others who were psuedoprogressors.

In order to sort out the effects of DCVax-L on rapid progressors from psuedoprogressors an independent imaging company was contracted to characterize the results. Imaging from these patients was all centrally reviewed to establish consistency. The criteria used and results were as follows.

  • 20 patients had further progression which was defined as another 25% increase in tumor size or new lesion greater than one centimeter at month 2. These were classified as rapid progressors.


  • 25 patients had stable disease or modest (less than 25%) progression or regression at month 2. These were indeterminate patients but possible psuedoprogressors.



  • 1 patent had all original 25% increase gone at month 2. This was a confirmed pseudo progressor.


  • 5 patients were unclassified due to lack of images.

Results from the Information Arm

Psuedoprogressors Results are Very Encouraging

Because the information arm was not blinded, investigators were able to analyze the data in real time. Here are the results for the psuedoprogressors:

  • The median overall survival was 21.5 months.
  • Ten of these 25 patients (40%) have reached or exceeded three years of survival.
  • Not only are they alive, they have not had progression. This is suggestive of a cure.

While these results are striking, there is a caveat and that is that there is no control group. Without this, we can’t really fully determine whether these results were largely or wholly due to DCVax-L or whether they were responding to standard of care and would have done just as well without DCVax-L. However, the results seem so striking that it is kind of hard to not think that DCVax-L was a major factor.

Rapid Progressor Results are Also Encouraging

These patients had recurrent glioblastoma. The results were as follows:

  • The median overall survival for the rapid progressors was 15.3 months.
  • Based on the literature, the expected median overall survival for recurrent glioblastoma is 8.3 to 10.6 months.

The lack of a control group creates uncertainty in analyzing these results.

Still another Trial of 32 Pseudoprogressors Remains Blinded

In the period from 2012 until 2013, there was still another trial that was started and was comprised of 32 psuedoprogressors. These were patients who showed disease progression at time of enrollment but no further progression eight weeks later. These patients were randomized 2:1 to DCVax-L and placebo. This trial remains blinded.

Issues with the Phase 3 DCVax-L Trial Design

The design of a clinical trial can sometimes have issues that can affect the outcome of the trial even if the drug is effective. I would cite the example of Bristol Myers Squibb’s CHECKMATE-026 of Opdivo in first line non-small cell lung cancer which compared Opdivo as monotherapy to chemotherapy. The trial set as a cutoff patients with PD-L1 expression ≥5%; it failed to reach the primary endpoint of progression free survival. Merck ran a similarly designed trial, KEYNOTE-024 that used PD-LI expression ≥50% that was successful. The success of KEYNOTE-024 and the failure of CHECKMATE-026, in the opinion of many key opinion leaders, was because BMY used the much more aggressive PD-L1 expression cutoff. See my report Immuno-Oncology is Probably The Most Explosive Commercial Opportunity in All of BioPharma; My Investment Picks are Bristol-Myers Squibb (BMY, Buy, $57.14) and Agenus (AGEN, Buy, $4.28) for a more detailed explanation. Most KOLS believe that Opdivo would have been successful if it used the same trial design as Keytruda.

Endpoints of the DCVax-L Trial May Not Be Optimal for an Immuno-Therapy Drug

The DCVax-L trial began over eight years, long before the results with checkpoint modulators provided valuable insights about immuno-oncology. Dr. Liau says that if the DCVax-L phase 3 trial were to be designed today, it might be designed differently.

The primary endpoint of the trial is median progression free survival with median overall survival as a secondary end point. Historical data shows that progression free survival for patients treated with standard of care (radiation plus the chemotherapy drug temozolomide) is about six months. The DCVax-L trial is statistically designed to be successful (p<0.05) if median progression free survival of patients given DCVax-L is four months greater than for the control given standard of care. If the control group behaves as predicted by historical data (not necessarily a given), for the trial to be successful DCVax-L would have to achieve median progression free survival of ten months.

Dr. Liau pointed out one of her concerns with the trial design is that median progression free survival and median overall survival may not capture the true efficacy of an immuno-oncology drug. These endpoints have been designed to capture the efficacy of chemotherapy, targeted therapy and other drugs in which the effect of the drug is to quickly shrink the tumor, halt its progression and hopeful extend survival.

Immuno-oncology drugs work differently. They mount an immune response against the tumor that manifests over a much longer period of time than chemotherapy. She believes that their true benefit of immune-oncology drugs should not be measured at the median. This is the time at which 50% of patients have had an event (either progression or death depending on trial design). Using the checkpoint modulators Opdivo, Yervoy and Keytruda as examples, she emphasizes that benefits take longer to materialize with immune-oncology. In her opinion and most key opinion leaders, she believes that about 20% to 30% of patients treated with checkpoint inhibitors achieve a long lived response or even a cure that is not apparent until long after the medians of progression and/or survival have been reached. The true benefit is not at the median point, but at the tail end of the Kaplan-Meier curve.

In the case of DCVax-L and the checkpoint modulators the true benefit is that perhaps 20% to 30% of patients have long progression free survival measured in years, but appear less effective or ineffective when median progression free survival or median overall survival are the endpoints. There may not be much difference in effect at the median level, because this does not capture the true benefit which occurs much later at the tail end. It seems to be the case that 20% or more of patients given checkpoint inhibitors (Opdivo, Keytruda, Tecentriq) may have the chance of living three, four or five years. For perspective, key opinion leaders consider a four to five month in median overall survival to be a major advance for chemotherapy and cures are very rare.

Dr. Liau reiterated that current cancer trial designs are designed to measure the effects of chemotherapy and targeted therapy drugs. In the case of immune-oncology, companies and investigators are struggling to get these agents approved on the basis of inappropriate trial designs. Median overall survival at one year or even two years may not be the appropriate endpoint. It is really the long tail benefit. She says that we should really be computing the hazard ratio at three or four years for these trials.

Cross Over Confounds Interpretation of Overall Survival Results in DCVax-L Trial

Increasing the length of overall survival is ultimately the goal of DCVax-L therapy. This is a prospectively defined as a secondary endpoint of the phase 3 trial, but overall survival results could be difficult to determine. As mentioned earlier, patients on standard of care were allowed to be switched over to DCVax-L if their cancer progressed. About 67% of patients initially received DCVax-L. Another 19% of patients who initially placebo were then given DCVax-L after they progressed. Hence the trial results compare 67% of patients who received DCVax-L initially, 19% who received DCVax-L at some later time and 14% of patients who did not receive DCVax-L at any time. The cross-over design unfortunately may make it difficult to determine the survival benefit of DCVax-L since some of the patients in the control group also received DCVax-L.

Past Trials of Other Autologous Dendritic Cell Vaccines are Also Encouraging; Dr. Liau Provides Some Data

Dr. Liau cited 21 published studies done by other investigators that used autologous dendritic cells loaded with tumor lysate to treat glioblastoma. This was the first time that I have seen this data and it is very encouraging. Altogether these 21 studies involved 403 patients on five continents. Of course, these were not randomized controlled studies.

These studies showed that autologous loaded dendritic cells increased overall survival by a low of 15 months and a high of 53 months). For comparison, median overall survival for standard of care is about 15 months.   To put this in perspective, I would note that any therapy that increases median overall survival by 4.5 months is deemed by KOLs to be a significant advance.

Dr. Liau’s View on Using Tumor Lysate to Load Dendritic Cells

In her research, she has looked at loading dendritic cells with many different antigens and combinations of antigens. However, she did not find any combinations that were as good as autologous loaded antigens from tumor lysate. She attributes this to the heterogeneity of glioblastomas which requires different types of killer T-cells to treat. This is not a “one sized killer T-cell” type of tumor.

How Do we Find Patients Who Will Respond?

Dr. Liau has long experience with dendritic cell cancer vaccines. She injected her first patient in 1999. In the phase 1/2 trials of DCVax-L which were the basis for the planning of the phase 3 DCVax-L trial, about 30% of patients are still alive. The question is how can these patients who apparently are so responsive to DCVax-L be identified prospectively?

Glioblastomas are Heterogeneous

She says that glioblastomas are heterogeneous as the disease can emanate from quite different genetic mutations which may determine which subset(s) of patients who benefit. She cited mesenchymal vs. proneural. MGMT, IDH, 1p19q, TERT as examples. There are also other predictors such as the degree of resection. The challenge is to understand what biomarkers characterize the 20% to 30% of patients who receive great benefit?

She said that one would expect that patients in the proneural sub-group would have better survival than mesenchymal. Surprisingly, this is not the case as mesenchymal patients actually did better. All of their long term survivors in phase 1/2 were in the mesenchymal subgroup. There is something in this sub-group that is producing better results and this is what they have been studying for the last five or six years.

Are Tumor Infiltrating Cells a Biomarker?

Tumor infiltrating lymphocytes (TILs) are killer T-cells that penetrate into the tumor mass. Better outcomes do seem to correlate with the presence of TILs. Interestingly she says that it is the level of TILs after vaccination rather than before that is key. This is probably a marker for a more immune responsive sub-group and is also more subject to pseudo-progression. Their tumors are more mutated which seemingly turns out to be better for DCVax-L because there are more antigen targets.

Potential Synergy with Checkpoint Modulators

She is also looking at ways to enhance the T-cell response. T-cells may get into a tumor but they are non-functional because of checkpoint biology that inactivates them. This suggests that a combination of DCVax-L with the PD-1 inhibitors Opdivo or Keytruda could be synergistic.

Trials of PD-1 inhibitors as single agents, have not shown that much activity in glioblastoma. The reason is that there are probably not sufficient numbers of killer T-cells present. If you don’t have T-cells in the tumor unblocking the checkpoints doesn’t do much. There need to be more T-cells in the tumor and this is what DCVax-L is designed to do. Hence, the combination could produce exciting results. .

Dr. Liau Pre-Announces Start of Combination Trial of DCVax-L with Opdivo

She inadvertently announced that NWBO had gotten FDA permission to begin a combination trial. DCVax-L will be given first to give rise to T-cells. This will be followed by Opdivo to prevent blocking of T-cell effects due to PD-L1 ligands released by the tumor. This trial will start just after the New Year. In animal studies combining the two does lead to increased survival.

Dr. Liau’s Conclusions

  • Brain tumor vaccines like DCVax-L are feasible and safe. They can potentially lead to a sub-group of long term progression free survivors (greater than five years).
  • Endpoints relevant for immune-oncology may be different than for other treatments such as chemotherapy and targeted therapy.
  • Heterogeneity of tumor/ patient needs must be taken into account: (1) mesenchymal versus proneural gene expression, (2) MGMT methylation (portends to psuedoprogression) , IDH mutation, 1p19q co-deletion, TERT and (3) imaging evidence of psuedoprogression
  • She think that about 25% of patients will experience a long survival benefit that is possibly a cure.




Tagged as , , + Categorized as Company Reports, LinkedIn


  1. Larry,

    There is one more scenario which I believe has the most chance of occurring, and that is positive Primary Endpoint (PFS shows statistical significance), however, OS secondary endpoints do not show statistical significance. The reason being is that in the majority of similar immunotherapy trials, PFS has shown significant differences from SOC (ICT-107, DCVax-L Ph 2, etc…), while OS has been the difficult endpoint to achieve, which is due to immunotherapies effectiveness on specific patients (i.e. DCVax-L with 25% of patients, etc..) and something that will be looked at with the new Phase 2.

    However, with the OS, like you mentioned, there could just be 1 month difference between placebo and vaccinated groups, but if the median OS values are significantly greater than the average SOC of 15 months, that puts a lot of pressure on the FDA to acknowledge this as positive data. I believe that PFS endpoints will be met, its the OS endpoints that I am not sure of.

  2. Thomas Cunningham says:

    The company you really need to check out is Affimed AFMD now they have a cutting edge product that will lead to the best response of killer T cells against tumors

  3. Thank you Larry for making so much sense out of the presentation by Dr. LL. Did you then look at Dr. P’s presentation that was also posted from the same conference? He had plenty of information and also spoke about 2 BP joining in a trial with DCVAX. Have you ever tried to contact the company and try and find more clarity about issues they publicly raised, but then issues were raised by the dark pools? So, I hope by ASCO 2017 we do see the un-blinded results and an in-depth discussion of them. This is a new and revolutionary field of treatment for cancer. It will take years and many trials and the cooperation of authorities like the FDA to launch a successful attack against the many form of cancer that attack the bodies of both humans and animals. Could Germany or the UK approve “L” while the FDA holds back? That is an issue you did not raise. Any comments on that might be helpful. Have a happy New Year. “staying tuned”

  4. OKAY, I DID MORE NUMBER CRUNCHING AND FEEL MORE CONFIDENT THAN EVER PFS ENDPOINTS WERE ESTABLISHED!!! Okay so from Linda Liau’s presentation, we know that screening was suspended no later than Aug 2015, and that it takes 3 months for the first treatment, so Nov 2015. So lets say that since the last patients treatment it has been atleast 1 year (12 months).

    Total 331
    Placebo 110
    DCVax-L 221
    Total DCVax-L 285 (86% of 331)

    PFS Events
    Placebo 64 (285-221)
    No DCVax-L 46 (14% of 331)
    ALL 186

    So we know that at minimum, 58% of the placebo group has had a PFS event

    Placebo PFS 58% (64/110)

    Now lets use some SOC data. From a data set of 287 patients, the data set found that @ 12 months, at least 30% of the population died (OS event). Lets assume only 30% die of the placebo group, but lets only take 30% of the No DCVax-L pop (46 patients).

    OS (@ 12months) (0.30*46) = 14

    Now lets add the OS events that would have occurred prior to any PFS to the 14% that didn’t get any vaccine, presumably because they are still alive or they died before it was seen that the tumor progressed:

    Total Placebo Events 78 (64 + 14)

    Placebo PFS & OS – 71% (78/110)

    DCVax Events 108 (75% of 248 events – 78)

    DCVax-L PFS 49% (108/221)

    So, using these assumptions, we can easily see that well over 50% of the placebo group would have had a PFS event, meaning that the median value would have already been determined, while only 49% of the DCVax-L population would have had a (PFS or OS) event, so there is a good chance that the DCVax-L median PFS will be greater than 12 months at minimum since that was the last time any patient would have received a vaccine.

    Now, even if we just looked at the 122 other events, this would be comprised of the DCVax-L population, AND the 46 placebo patients that didn’t crossover. So that’s only 48% of 221+46 patients. No matter, the signs are good from this that the PFS endpoints are very likely to be met.

  5. Sorry, my bad Larry, I was rechecking my work and the 75% statement Dr. Liau used in her presentation was for the endpoint for the 331 patients, which is 248, so there isn’t the ability to measure the amount of events of each group.

    However. I still believe that the longer the time for the endpoints events to occur, the better it is for DCVax-L PFS endpoints, because we do know that ~ 64/110 ~ 58% of placebo events have happened, and a good portion of the 46 placebo patients that either had progression and didn’t cross over, or died (which we don’t have the number on), will probably account for some of the endpoint events. This means that the longer the trial doesn’t hit its endpoint, the more likely this is due to an extended PFS of the DCVax-L population.

    We know for certain, that the median PFS value of the placebo group has occurred. We do not know if the DCVax-L group has reached a median PFS value.

  6. My question is simply this – how is NWBO going to finance their operations moving forward?


You must be logged in, or you must subscribe to post a comment.