Expert Financial Analysis and Reporting

Introduction

Adam Feuerstein just published a blog that attacks the results and conclusions drawn from the paper summarizing results from the 51 patient information arm of the DCVax-L phase 3 trial. This marks the 22^nd consecutive negative blog by Feuerstein on Northwest. I am growing tired of rebutting his articles, but as long as he continues to publish what are clearly one sided, contrived arguments, I feel compelled to introduce facts that balance out what is consistently his negative view of anything that Northwest accomplishes. I cannot let his blogs stand unchallenged.

Rebutting Specific Statements

I have read through his recent blog and focused on some of his statements which I believe ignore counter arguments, cherry pick and misrepresent data or are spun out of whole cloth (i.e. a story with no basis in fact). I have separated out some of the Feuerstein comments for rebuttal. Let me start with what is clearly the worst of the worst of his comments in his blog.

Feuerstein: Northwest Bio announced an interim efficacy analysis of the study to begin in late 2013 but then never disclosed the results.

SmithOnStocks: This is just blatantly wrong. Dr. Curt Furberg, the Chairman of the Data Monitoring Committee publicly stated last August that there had been no efficacy analysis. He then took the highly unusual step of going public and issuing a sharp rebuke to Feuerstein for his repeated statements that accused Northwest of “burying the results”.

Dr. Furberg said that there had been no interim efficacy analysis. He went on to say “The DMC adheres to established clinical trial monitoring procedures and does not release any data while the trial is ongoing.  This is an important issue, and it is surprising and troubling to see inaccurate claims being made by commentators who seem to lack a fundamental understanding of clinical trial monitoring.  I have been on the DMCs for more than 60 clinical trials, and I have never experienced this type of attack.”

Feuerstein is undeterred by these facts. He continues to repeat the incorrect allegation in this blog and several previous ones that an efficacy analysis was done and that Northwest hid the results. This  can only be seen as a blatant misrepresentation of the facts. Feuerstein represents that Northwest is lying about its data and that through exposing these lies; he is protecting the small investor. Hmmmmm.

I think that if Feuerstein wants to be seen as an objective reporter, he should acknowledge that he is wrong in saying that results of interim analysis were suppressed by Northwest. An apology is also called for.

Feuerstein: The problem is the data were compiled and analyzed outside of a clinical trial, which leaves conclusions susceptible to biases, intentional or not.

SmithOnStocks: This shows a lack of understanding of how the information arm data was collected and analyzed. Northwest Biotherapeutics reported data at the Immunotherapy of Cancer Conference (ITOC) in Munich Germany on March 26, 2015. This data dealt with 51 patients whom investigators originally intended to enroll in the ongoing 348 patients phase 3 trial of DCVax-L in newly diagnosed glioblastoma multiforme. They became ineligible when their cancer progressed before they could enter the trial. However, Northwest decided to treat them virtually identically to the 348 patients enrolled in the trial and follow them to see how long they survived.

These 51 patients were treated by the same doctors at the same institutions in the same way as the 348 patients. All interpretation of images and classification of patients’ status (as rapid progressors or indeterminate or unclassified) were made by the independent central lab that is responsible for determining patient status, not by the Company. Also, all data were collected and maintained by the independent CRO managing the phase 3 trial, not by the Company.

Feuerstein: Of the 51 patients, 20 were classified as rapid progressors. Unfortunately for Northwest Bio, 59% of its patients couldn't be classified, which makes survival comparisons difficult to interpret, at best.

SmithOnStocks: Feuerstein implies that Northwest went through the data and decided which patients would be classified as rapid progressors. This is incorrect. The independent central lab made the determination based on established criteria.

The 20 patients designated as rapid progressors were patients who clearly met commonly accepted criteria explained in the ITOC paper. The 25 patients designated as indeterminate also met some of the criteria but not all. Many or all of these patients could well have been rapid progressors. The remaining 5 patients could not be classified due to lack of available images.

The 20 patients were designated rapid progressors because they met all of the criteria. These patients had the most rapidly progressing and were the sickest of the sick in this group. In the interest of viewing the data in the most conservative way, only these 20 were designated rapid progressors. Their OS was 15.3 months. If the indeterminates who were probably rapid progressors were included in the analysis were included, results would have looked much better as their OS was 21.5 months. Feuerstein is silent on this issue.

The essential fact is that the sample of 20 rapid progressors was made independently, on the basis of clear criteria, and compared to results for comparable patient samples in the literature – and on this basis showed an impressive improvement over expected survival of 7 to 10 months as predicted by six smaller studies.

Feuerstein: Take a look at the top two GBM patients in the "indeterminate" chart represented by the blue lines. Both patients died in 10 months or less. Based on the six published studies found by Northwest Bio, it's certainly logical to consider these patients rapid progressors.

SmithOnStocks: This is clearly wrong. First, it overlooks the fact that the determination of rapid progressors was made by an independent central lab, applying established criteria – and they determined that the patients Feuerstein proposes to cherry pick and selectively re-classify patients as rapid progressors did not meet the criteria.

Feuerstein’s argument fails to recognize that in any sample there are deviations around the median. There will be some patients who do better and some that do worse in any sample; this does not mean that the classification of such patients can or should be changed. Some rapid progressors will do better than some indeterminates and some indeterminates will do worse than rapid progressors. This is because there are deviations around the median.

Feuerstein: To be clear, I'm not accusing Northwest Bio of intentionally excluding rapid progressors to improperly boost the reported median overall survival.

SmithOnStocks: At first glance, this seems to be a change in style for Feuerstein, a somewhat less vitriolic attack. In the past he has often accused the Company of misrepresenting data and has said that they withheld efficacy results from an interim analysis as was previously described; this is the type of statement that we are more used to seeing from him.

The plain fact is that Northwest cannot be accused of intentionally excluding patients in order to boost median survival numbers because the evaluations of the patients were done by an independent lab.  However, Feuerstein is himself doing exactly what he wrongly implies Northwest might have done. His own analysis intentionally cherry picks patients who were determined by an independent lab to meet all of the criteria established to be included in the rapid progressors sample. This improper analysis is clearly done in an attempt to make the data look worse.

Feuerstein: There is little clinical value in the DCVax presentation last Friday because the patient survival data come from outside of a clinical trial.

SmithOnStocks: This more than anything shows the amazing inconsistency of Feuerstein’s positions. He has written favorably on the results for the CAR-T companies- Juno and Kite. The results for their drugs were obtained from investigator sponsored studies. These probably do not have the rigor of phase 1 trials and certainly do not have the rigor of the phase 3 trial conditions used to conduct the trial and analyze the data in the 51 patient information arm.  The Kite and Juno trials were open label, not randomized and were compared to historical controls. Importantly, neither Kite nor Juno on their own has completed a phase 1 or phase 2 trial for any of their products. They have all been done by the NCI or academic centers.

The data for the 51 patients comes from study circumstances that are certainly as reliable as  academic investigator studies and in my opinion much more so. The 51 patients were treated in the same way as those in the controlled phase 3 trial.  The evaluation of the patients was done by an independent central lab, which is not usually the case in academic investigator sponsored studies. The collection and maintenance of the data was done by an independent CRO, which is also not usually the case in an academic sponsored study.

Like the CAR-T trials, the information arm results were compared to historical controls. This does not give the confidence that comes from a placebo controlled trial. However, in the case of both the CAR-T cells and DCVax-L, they do give an extremely important insight into how DCVax-L works in a refractory patient population that has no effective therapy.

I would argue that the Northwest data is of higher quality and greater reliability than that of the CAR-T companies. By the way, this[LP3]  is not to minimize in any way the outstanding results reported in relapsed/ refractory acute lymphocytic leukemia and other leukemias. I am enormously excited about the potential for CAR-T therapy. It is also not to say that on some absolute level the DCVax-L results are better than CAR-T. I don’t know how to compare results obtained in relapsed/ refractory acute lymphocytic leukemia to relapsed/ refractory glioblastoma multiforme.

I can see Feuerstein being positive on data creation as I have described from both Northwest and the various CAR-T clinical trials, or being negative on both. It seems inconsistent to be positive on the CAR-T results and critical and dismissive of those from Northwest based on trial design and conduct and how they capture data.

Feuerstein: The need to enroll additional patients and changes made to the number of progression events required to trigger a final analysis, disclosed by the company last August, was an ominous sign that DCVax was headed for failure.

SmithOnStocks: The change in the trial increased the number of patients to be enrolled from 312 to 348 and the original trial design prospectively allowed for an increase in the number of patients. Furthermore, at the time of the changes, the Company clearly explained substantial, positive reasons for the expansion of the trial, including taking account of a key variable affecting patient survival that was not known when the trial started.  No judgment on trial results can be made from this increase in trial size.

Other Points Feuerstein Did Not Address

Feuerstein studiously ignored some of the other key positive points from the survival data reported by Northwest.  For example, he has made no mention of the 25 indeterminate patients’ median OS of 21.5 months, which compares to 15 to 16 months or so reported in several large trials for “regular”, less sick GBM patients treated with standard of care. Given that the indeterminate patients are progressing more rapidly than the regulars and presumably have a more aggressive disease, this is an impressive result.

He has ignored the impressive, long survival tail, which is every bit as important as the improvement in median overall survival. In the rapid progressor group of 20 patients, one survived for 32 months and another for 37 months (this patient is still alive). In the indeterminate group of 25 patients, 11 patients survived for 27 months; 8 patients have survived for 30 months (of these 6 are still alive); and 2 have survived for 41 months (both alive). The literature suggests that 50% of rapid progressors are dead at 7 to 10 months and that 50% of “regular” healthier GBM patients are dead at 15 to 16 months.