Expert Financial Analysis and Reporting

Investment Opinion

Kite will be reporting interim results of the phase 2, ZUMA-1 trial of its lead product KTE-C19 by the end of September. The most bullish analysts are projecting a very positive outcome in this trial that will result in approval in 2017 and sales of as much as $1 billion of sales by 2020 and are using terms such as “medical miracle” to describe KTE-C19. There is a lot riding on ZUMA-1. Kite is expecting dramatically positive results from this interim analysis. It has suggested that it will file a BLA by yearend based on the interim data and that KTE-C19 could be approved in mid-2017. I am more skeptical and I would not own the stock going into the imminent release of this interim data.

I would point out this is the first trial that Kite has ever conducted. It is a small, open-label phase 2 trial without a control group that will enroll 112 patients. This type of trial design is often used to create information to allow the design of a larger, controlled, randomized phase 3 trial. However, the FDA has approved drugs based on similar trial designs if they showed striking efficacy in aggressive cancers without effective drug treatments. For example, the checkpoint modulators Opdivo and Keytruda were approved for some indications based on similar types of trials.

Despite keen investor anticipation, the press release on interim results will give only limited information on data coming out of the trial. The full data set will be presented at the American Society of Hematology (ASH) in December 2016. I think that the most important number that investors will focus on in this press release will be the CR rate at three months post infusion of KTE-C19. So what are expectations for this data point?

The trial is designed on the expectation that 40+% of patients will achieve a CR at about six months after cell infusion. Kite indicated that the CR rate for seven patients treated in the phase 1 part of ZUMA-1 was 57% at three months, but a more rigorous analysis suggests it was 38%. I looked at data from an NCI trial for a drug very similar to KTE-C19 and also at data from a trial by Novartis with its directly competitive product CTL 019. I concluded that these small single center trials suggested a CR rate of perhaps 35% to 40% might be expected. However, I think that the interim results for the multi-center ZUMA-1 trial will not be as good. I am projecting a 25% to 35% CR at three months after cell infusion. I think that if the CR at three months is 35%, investors will view it as positive or slightly disappointing. If the CR is 25% over less, it would be very disappointing.

The CR rate has to be looked at in conjunction with duration of response. The CR rate at three months may be significantly higher than at six months or a year. My view is that the duration of response will have to be maintained six months at a minimum for the CR to be considered clinically important. I believe that a minimum, FDA will want to see the CRs maintained for six months. As previously noted, the press release will report the CR rate at three months. Along with most investors, I would expect the rate of CRs to go down over time. I do not expect anything meaningfully new to be reported on side effects.

Summing it all up, I think the data that will be reported in the press release may have frustratingly little meaningful information. We will have to wait for ASH to get more informative data on the trial. I see a strong possibility that the CR rate could be viewed as disappointing.

Overview of This Report

I have been working on a report on Kite for over three months. I have decided to publish this research in a series of reports instead of overwhelming my subscribers with one huge report. This is the fifth in the series of reports.  In my last report, I discussed my expectations for the imminent release of interim results from the critically important ZUMA-1 clinical trial of KTE-C19. This report goes through the lengthy research that underlay the conclusions presents in that report. In doing this work, I grew ever more concerned about the lack of transparency of Kite management in regard to an NCI conducted phase 1/2 trial. The data from this trial was the basis for the design of the ZUMA-1 trial.

Kite management is suggesting to investors that based on an interim look at efficacy in ZUMA-1, that its lead product KTE-C19 could be approved by mid-2017. Investors should have been given detailed data on the NCI trial in order to assess the probability of success in ZUMA-1, but this is not the case. In doing my research on Kite, I found that the information that Kite has released on the NCI trial to be incomplete and confusing. I have come to see management as being highly promotional.

I was fortunate that one of my physician contacts had detailed knowledge of the NCI trial. This provided the basis for a much more balanced view of the details of as well as plusses and minuses of the trial than what Kite has afforded. It also gives a more sober, less rosy assessment than what management has provided in its press releases and management presentations.

This note traces data from Kite on the trial released primarily in chronological order primarily in the form of press releases and financial regulatory filings. Very importantly, this is supplemented by information from my source which quite frankly provides a much clearer insight into the NCI trials. For subscribers, I am concerned that you will find this to be complex and difficult note to follow because there are so many numbers; it will require a lot of time to wade through. In order to make for easier reading and hopefully better understanding, I start with my brief investment thesis. I then go through a long list of my key points from the note. I will leave to you the option as to whether to read just the investment thesis and key points or the note in its entirety.

The preponderance of investors at this point in time have bought into the idea that CAR-T therapy has the potential to be a major advance in the treatment of hematological cancers and I agree. They have also bought into the investment thesis that Kite management has created and which I have problems with. Management says that KTE-C19 will demonstrate efficacy in an interim look of the ZUMA-1 trial; efficacy as so defined would require showing a Complete Response (CR) rate of 40% or more. Kite has guided that this will allow them to file a BLA by the end of 2016 and receive approval by mid-2017. I think that if ZUMA-1 will achieve this efficacy endpoint and that this could trigger a rally in a stock even though I think it is currently over-valued. My own expectation is that the CR will be on the order of 25% to 35%. See my report Kite Pharma Part 4: A Preview of Highly Anticipated Results from Interim Look at ZUMA-1 (KITE, Avoid, $54.95)

Kite Pharma Part 4: A Preview of Highly Anticipated Results from Interim Look at ZUMA-1 (KITE, Avoid, $54.95)

My analyst antennae are up on Kite because I believe that management has been over-promotional. This has created expectations that the most optimistic of scenarios is highly probable. To reiterate, I am concerned by the lack of transparency on their part in reporting results from the NCI phase 1/2 trial that is the basis of ZUMA-1. Press releases from Kite have given only a partial picture of the results in the NCI trial and in my opinion, over-emphasize and overstate efficacy results and understate the occurrence of life threatening side effects.

I would also emphasize that it is Kite that has established the expectation that a 40% complete response rate in the interim look at ZUMA-1 will be sufficient for approval. The FDA may not sign off on this. It may want more supporting data on duration of response following a CR than will be available at the interim look. There are also issues with commercialization in that the price of this therapy is extremely expensive and that Kite will be competing with Juno and Novartis for the same market. I will address these issues in a later report.

Let me now go through key points that are the basis of my thinking on the outcome for KTE-C19 in the ZUMA-1 trial. In my first report on Kite I provided a glossary of terms and acronyms in my report. Kite Pharma: Part 1- Glossary of Terms and Acronyms You may want to refer to that report at times.

Key Points to Understand the Design of the ZUMA-1 Trial:

• Kite is now conducting the ZUMA-1 phase 1/2 trial which is enrolling patients with r/r DLBCL, r/r PMBCL and TFL. The phase 1 part of the trial was to determine safety in order to advance the trial to phase 2. Phase 1 demonstrated acceptable safety allowing phase 2 to begin. There was one death in phase 1 which investigators said was not attributable to KTE-C19. There were also troublesome and some life threatening side effects that are common and expected with CAR-T infusions and the chemotherapy pre-conditioning program that is used in conjunction.

• The phase 2 part of ZUMA-1 is an open label trial with two cohorts. One will be 72 patients with r/r DLBCL. The second cohort will be 40 patients with r/r PMBCL and TiNHL. According to ClinTrials.gov the study began in January 2015 and is scheduled to complete in March 2017

• The primary endpoint of ZUMA-1 is to demonstrate a CR rate of 40% or more. There will be no control arm to compare against but for statistical design purposed, Kite is estimating based on historical experience that the expected rate is 20%. Hence, the trial is designed to show statistical significance against this historically determined 20% rate.

• Kite did an extensive meta-analysis which suggested that the CR rate in r/r DLBCL from currently used treatments is about 8%. This is much less than the 20% estimate as just explained.

• Obviously this is not a controlled trial. This is a design more often used for a phase 2 trial that the will inform the design of phase 3 registrational studies. Still, the Kite trial design can be the basis for approval in relapsed and refractory cancers. Recently, this has been the case with the checkpoint modulators Keytruda and Opdivo, as well as other drugs which gained approval in certain severe cancers.

• An interim look at efficacy in the first 50 patients enrolled in the r/r DLBCL cohort should be completed shortly. Kite management is guiding investor to expect that the results will be quite positive and support a filing of a BLA for r/r DLBCL by the end of 2016 and to expect approval by mid-2017. Management’s expectation is that the rate of CR will be 40%. The complete trial calls for six months of more duration of response. I am not completely clear what Kite expects for duration of response at this interim look I think that it will only have CR data for three months, not six.

• Kite has never completed a clinical trial with KTE-C19 and for that matter it has never completed any clinical trial with any product. ZUMA-1 will be the first Kite conducted trial.

• The remarkable genetic engineering used to create KTE-C19 was essentially developed by the NCI. That agency also studied a predecessor drug to KTE-C19 in phase 1/2 trials. Kite uses a different manufacturing process than NCI. Sometimes in living cell products, a change in manufacturing process can produce meaningfully different results. Data created so far suggests that KTE-C19 and the NCI product lead to similar outcomes, but this has not been established.

Key Points on Clinical Data from NCI Studies That Was Used to Design ZUMA-1

• The clinical data upon which ZUMA-1 was created in the phase 1/2 trials done at NCI. In doing my research on Kite, I became concerned about the lack of transparency on the part of management in discussing results of this trial. Based on the press releases and other data provided by management, I came to believe that management was massaging the data and possibly overstating efficacy of CAR-T therapy while minimizing information on life threatening side effects.

• All biotechnology managements are promotional, but I find Kite’s management to be extremely so. This has allowed them to create an extremely optimistic and enthusiastic investor base and has resulted in a huge market capitalization. This has served them well and allowed the company to raise huge amounts of cash. Now they have to deliver on these expectations.

• Efficacy data from my source who is not associated with Kite provided me with interim data from the NCI trial in a form that I believe is not available to other analysts. As of August 14, 2014, 17 patients with r/r DLBCL, r/r PMBCL and TFL had been treated and in that group the complete response rate was 35%. The median duration of response was 7.5 months. Remember that the ZUMA-1 statistical design assumes a 40% CR in DLBCL. From this data, the CR in DLBCL can’t be determined.

• As of August 14, 2014, there were 89 grade 4 (life threatening) events reported in 31 patients that were treated with the pre-conditioning chemotherapy regimen. (There were an additional 14 patients treated with different cancers than the 17 described in the previous key point.) The leading grade 4 side effects were: lymphopenia (30 cases), neutropenia (21) leukopenia (22) and thrombocytopenia (16). I would note that oncologists expect are familiar with these side effects and know how to address them based on long experience.

• I want to emphasize that the side effect data includes some patients who were treated with much a much higher dosage of cells, a more intense pre-conditioning program and IL-2 infusion which will not be used in the ZUMA-1 trial. These cause more severe side effects so that the grade 4 side effects reported in the two previous bullet points are likely to be much higher than what actually will be seen in ZUMA-1. However, there is a grade 4 side effect risk with both pre-conditioning and cell infusion. This is why patients were hospitalized after pre-conditioning chemotherapy in both the NCI and ZUMA-1 trial.

• Kite management usually only mentions side effects associated with cell infusion and doesn’t mention side effects due to pre-conditioning. I think that investors have to look at the combined side effect profile of pre-conditioning chemotherapy and cell infusion.

Key Points: What Kite Has Said About the NCI Data

• What investors want to know about the NCI trial is the response rate and side effects in r/r DLBCL, r/r PMBCL and TiNHL patients because these are the patients who are enrolled in ZUMA-1. In the following bullet points, I discuss information that Kite has put out at various points in time. To me, the data is not presented in a transparent way and looks to be massaged to make the data look better than it actually is.

• Remember that Kite is saying that a an interim efficacy look at 50 DLBCL patients will lead to approval if the CR is 40+% and duration of response is longer than 6 months. The next bullet points look at data in the NCI trial at various points in time.

• As of November 14, 2014 information from my proprietary source showed that the complete response rate one month after cell infusion in 17 DLBCL and PMBCL patients was 35%. The median duration of response was 7.3 months. The data for DLBCL patients alone was not available.

• Kite provided an update on the NCI trial in the 2015 10-K which showed data as of August 2015 on 26 patients with r/r DLBCL, r/r PMBCL TFL and r/r CLL. There were 14 patients with DLBCL and PMBCL which was three fewer patients than the 17 from a year earlier. What happened to those three patients? Kite reported that the CR rate was 43% (six patients) out of 14 DLBLC and PMBCL patients.

• Also as of August 2015, Kite reported that 15 of the 26 patients remained in response with 11 having durations of response of greater that one year. Three patients were retreated and their duration of response ranged from 18 to 52 months. This are impressive results, but what we really want to know is what the duration of response is for DLBCL, PMBCL and TFL patients. Kite omits this data and this is another, among several examples of how Kite seems to massage the data.

• In a June 6, 2016 press release, reported on 22 patients which was less than the 26 described in the 10-K. There was no explanation about why 4 patients were dropped. Kite reported that CRs were seen in 12 of 22 patients (55%).

• Kite further said in the June 6 report that there were 19 patients with DLCBL and PMBCL, and that 9 of 19 patients (47%) achieved CRs, which are all ongoing with a duration of 7 to 20 months. These are encouraging results, but there is no breakout between DLBCL and PMBCL. Because of the small numbers, investors would like to see the CR for each group. Why does Kite aggregate the data? Could it be that the DLBCL results show a CR of less than 40%? There is no way of knowing.

• Also in the June 6 press releases, Kite said that reversible grade 3 or 4 neurotoxicity including confusion, dysphasia, encephalopathy, and gait disturbances was observed in 55% of treated patients. This is the first time that the Company was specific on side effects. It does not speak to cytokine release syndrome side effects nor does it break down the percentage of grade 4 side effects. I would note that these side effects are specific to cell infusion. The press release also does not speak to severe side effects that are the result of the chemotherapy pre-conditioning program. You may recall from my earlier discussion that these are more numerous and more likely to cause grade 4 toxicities.

• I trust that in going through the key points that I have made that Kite releases on the data seems to have been massaged to give it the most positive spin. It may be the case that the CR rate in DLBCL patients could be less than 40% which is the primary endpoint of ZUMA-1. The investment consensus that it is almost a given that ZUMA-1 will be stopped after an interim look showing a 40% CR. I suspect the CR could be around 40% based on the NCI data. I don’t see this as a slam dunk.

Key Points: Data from Phase 1 Component of ZUMA-1

• Once again, I was able to obtain data on the phase 1 component of ZUMA-1 from a non-Kite source who provided data as of August 15, 2015. At that time, eight patients had been enrolled in the trial and six had been treated; presumably the other two enrolled patients were subsequently treated. On June 6, 2016 Kite reported results on seven DLBCL patients from ZUMA-1. The question is wat happened to the eighth patient. Kite doesn’t say but there was one patient who died 16 days after cell infusion and they might have been dropped from the analysis so that Kite reported on the seven patients.

• In describing efficacy and side effects, Kite put the most positive spin on the results. As of August 15, 2015 the data was not mature enough for my source to comment on efficacy. Kite reported on June 5, 2016 that there were four CRs in seven patients for a rate of 57%. This was significantly better than the 40% expected in the phase 2 part of the trial.

• Kite said that three of the patients who had achieved a CR were continuing to be in remission at nine months. They did not say what the duration of response was for the fourth patient who achieved a CR. If the duration was short in this fourth patient and we don’t know if it was, there could be a case made that the CR was not meaningful and they should not have been included. In that case, the CR rate would have been 43% (three of seven).

• In the most conservative way of looking at the data we could say that on an intent to treat basis, three meaningful complete responses were achieved in eight patients so that the CR rate was 38%. However, I want to emphasize that even a 38% rate is actually close to 40% rate that what Kite was hoping for. Still, it bothers me that the data was still presented in the most positive possible way.

• Kite said that KTE-C19 related adverse events consisted predominantly of cytokine release syndrome and neurotoxicity, which were generally reversible. Grade 3 or higher CRS side effects were observed in 14% of patients and neurotoxicity side effects in 57%. All were reversible except in one patient with dose-limiting toxicity who died. The investigator determined that the death was not related to cell infusion. However, I think that it could have been caused or exacerbated by the chemotherapy preconditioning.

• Kite did not talk about side effects related to the chemotherapy regimen. It talked only about the percentage of patients who had grade 3 or greater side effects and was silent on grade 4. My non Kite source stated that in six patients treated there were two grade four side effects related to the chemotherapy and two related to cell infusion. Once again there is a lack of transparency on Kite’s part.

Key Points: Novartis and Kite Are Racing to Become the First Mover in the CAR-T Space

• Novartis has announced plans to file a BLA for its lead CAR-T drug CTL019 in 2017 for pediatric r/r ALL. I have found no guidance on the timing of when the BLA might be filed.

• Novartis presented data at the American Society of Hematology meeting in December 2015 on CLT019 in r/r DLBCL. Seven of fifteen patients (47%) experienced a CR three months post cell infusion.

• Novartis began a study on 118 patients in r/r DLBCL in July 2015. The primary efficacy endpoint is ORR. The phase 2 part of ZUMA-1 began on July 7, 2016. Novartis has not given an indication on when the trial might complete or when a BLA might be filed.

• It appears that completion of the Novartis trial and ZUMA-1 should complete at about the same time. Novartis has not guided on whether there will be an interim look as Kite is doing in the ZUMA-1 trial.

• Kite is hoping that it will gain approval for KTE-C19 based on the interim look at the phase 2 part of the ZUMA-1 trial. We don’t know if Novartis has the same strategy. If Novartis does not and if the FDA accepts the BLA filing for KTE-C19 based on the interim look, Kite could be first to market in the r/r DLBCL indication.

• If the FDA does not accept BLA filing for KTE-C19 based on interim data, Kite and Novartis could gain approval for r/r DLBCL at the same time. Approval of CTL019 in pediatric r/r ALL would likely be sooner than this.

MAIN REPORT

The NCI Phase 1/2 Study That Provided the Basis for Designing ZUMA-1

NCI Trial Design Overview

Kite funded a phase 1/2 trial at NCI which used the same CAR T construct (CD-19) that Kite is using in the ZUMA-1 and other trials. This trial enrolled patients with a number of B-cell malignancies such as CLL, DLBCL, PMBCL and TiNHL. Patients enrolled in this study were refractory to chemotherapeutic and other treatment options. This meant that their cancers were progressing. Also enrolled in this study were patients who had relapsed after a stem cell transplant.

When I initially started my research on Kite, I looked at its press releases and regulatory filings. These were confusing and inconsistent and at no point gave a detailed description of the trial design and outcomes of this important NCI trial. Information provided by Kite was more of a series of random snap shots that did not tie together. I was fortunate to find another source not related to Kite with detailed information on the trial as of August 14, 2014. I am going to start with information this information and then follow with an analysis of information provided by Kite. It appears to me that Kite has massaged the data.

Details on Design of the NCI Study

I have not seen a detailed description of the NCI study in any of the Kite releases so let’s start with this. It was designed as a single arm, open label, trial. As of November 14, 2014 it had enrolled 32 subjects broken down as shown below:

• 19 (59%) had r/r DLBCL or r/r PMBCL;
• 7 (22%) had r/r CLL ; and
• 6 (19%) had TiNHL.

Remember that the phase 2 component of the ZUMA-1 trial has one arm enrolling 72 r/r DLBCL patients and a second arm enrolling 40 r/r PMBCL and TiNHL patients. Note that no CLL patients will be enrolled in ZUMA-1. This was my first look at the data and since August 14, 2014 the number of patients treated has expanded from 32 to 41 or more.

All subjects were treated with a pre-conditioning chemotherapy regimen and were then hospitalized in preparation for a single infusion of the CAR-T cells. The hospitalization was necessary because both the pre-conditioning chemotherapy regiment and cell infusion can cause life threatening grade 4 side effects. The period of hospitalization was generally a week or so. The purpose of pre-conditioning is deplete normal T-cells which enhances the proliferation of CAR-T cells after they are infused. Initially in this trial, a few subjects were also treated with IL-2 following the cell infusion; this was also done with the purpose of stimulating the proliferation of the CAR-T cells.  IL-2 therapy was abandoned as the trial progressed due to toxicity issues. Retreatment with a second infusion of CAR T cells was allowed if the patient initially achieved a CR or PR and then the disease progressed.

The trial protocol was originally designed as a dose escalation study. However, as the dose of cells was increased dose limiting toxicities developed causing protocol amendments aimed at limiting the toxicity. As a result there is a great deal of heterogeneity in dosage of the preconditioning chemotherapy, the number of and cells infused and the use of IL-2. Eventually three groups of patients with meaningfully different treatment regimens were enrolled as follows.

Group 1 consisted of 8 subjects, including one subject who was being retreated. The chemotherapy pre- conditioning program consisted of cyclophosphamide of 60-120 mg/kg for 2 days followed by fludaribine (25 mg/mm²) for 5 days. After the cancer pre-conditioning program, all subjects were hospitalized in preparation for a single infusion of CAR-T cells. The CAR-T cell dose ranged from 3 x 10⁶ through 30 x 10⁶ cells/kg.  Note that the dose in Kite’s ZUMA-1 trial is 2 x 10⁶ cells/kg so that some of these patients were given a much larger dose than patients will receive in ZUMA-1. Some subjects were also treated with multiple doses of IL-2 after the cell infusion to stimulate CAR-T cell proliferation.  Retreatment of a second dose of CAR-T cells was allowed if the patient initially achieved a CR or PR, but then relapsed.

Group 2 consisted of 15 subjects, including 2 from Group 1 who were retreated. They were given 1 x 10⁶ to 5 x 10⁶ CAR-T cells/kg. They received the same preconditioning regimen as group 1, but no IL-2.

Group 3 enrolled 11 subjects who received a conditioning regimen of a higher dose of cyclophosphamide (300 mg/m²) and a higher dose of fludarabine (30 mg/m²). The schedule of dosing was changed as the drugs were given together for three consecutive days. The first 7 patients received 1 x 10⁶ cells and four received 2 x 10⁶ cells. No IL-2 was used.

Efficacy Results for all Patients as of November 14, 2014

Results were judged on the basis of shrinkage of the tumor. B-cell tumors are metastasized throughout the body. In order to assess tumor shrinkage, investigators measure the size of a few of the largest tumor masses before treatment and then re-measure as treatment progresses. A complete response (CR) is defined as absence of any sign of the tumor at these locations with the exception of scar tissue. A partial response (PR) is defined as shrinkage of >50% of the tumor mass. The sum of CRs and PRs is called the objective response rate (ORR). A CR does not necessarily mean that the tumor has been eradicated as there may remain small metastases. However, CRs and PRs confer therapeutic benefit and most physicians believe that just keeping the tumor mass stable (SD) is important therapeutically.

As of November 30, 2014, there were 34 treatments given in groups 1, 2 and 3 but because two of these patients were retreated there were only 32 patients to evaluate for safety and at that time only 29 could be evaluated for efficacy. The efficacy was defined as the response rates one month after cell infusion. Missing in this data is the critical information duration of response, because if the tumor initially shrinks and then begins to grow, the treatment has failed. It is highly subjective as to what is a meaningful duration of response but it is probably six months as a minimum.

This trial included a mixed group of cancer patients with DLBCL, PMBCL, and CLL and TiNHL. Remember that the ZUMA-1 trial has one cohort of 72 DLBCL patients and a second cohort of 42 patients with PMBCL and TiNHL so in looking at the NCI trial we are most interested in results in those patients. ZUMA-1 does not include CLL patients.

Efficacy Results in DCLBCL and PMBCL Patients as of November 14, 2014

As of November 14, 2014 there was also data on combined results of DLBCL and PMBCL patients. This is frustrating as we would like to know the results for each cancer. Remember that DLCBL and PMBCL patients are enrolled in different arms of ZUMA-1. I have not heard Kite say why this was done, but presumably it is because outcomes may differ. The efficacy results were as follows;

• 17 of 19 patients were evaluable for disease response (1 subject was not evaluable and 1 subject had not yet been evaluated). The general practice is to include the patient lost to follow-up in the analysis on an intent to treat basis; this would have made the results slightly worse by increasing the denominator to 18.
• Among these 17 patients, 11 (65%) 6 (35%) achieved a CR. We don’t know what the CR was in DLBCL alone.
• The median duration of response was 7.3 months.

Efficacy Results in TiNHL Patients as of November 14, 2014

Efficacy was also reported for five patients with TiNHL.

• Five of the 5 evaluable subjects (100%) had an ORR.
• One of five subjects (20%) achieved a CR.
• The median duration of response was 18.8 months. Five subjects (100%) remain in response with 2 subjects responding greater than 45
• Based in this limited sample, it appears that even though TiNHL patient show a lesser CR, they have much longer duration of response than DLBCL/ PMBCL patients.

Efficacy Results in TiNHL Patients as of November 14, 2014

Even though CLL patients are not being enrolled in the ZUMA-1 trial, it is interesting to look at results.

• Six of the 7 evaluable subjects (86%) with CLL had a response with 4/7 subjects (57%) achieving a CR.
• The median duration of response was 22.2 months with 4/7 subjects (57%) still in response including 3 subjects with ongoing responses of greater than 27 months.
• It appears that these CLL patients did better than the DLBCL/ PMBCL patients.

Data from Kite on the NCI Trial

All of the previous data came from my proprietary source. I next turned to data on the Kite website as presented in the scientific papers and press release sections. I list these in chronological order.

November 11, 2010 scientific abstract:  This was the first paper in the publications section of the Kite website that discussed the anti-CD19 CAR-T cell therapy. The study was in a mouse model for B-cell malignancies. Encouraging efficacy was seen, but the paper said that efficacy was critically dependent on irradiation of mice before cell infusion. This radiation depletes the natural T-cell population and allows the CAR-T cells to differentiate more rapidly when infused. Subsequently, the NCI implemented a pre-conditioning chemotherapy for the same purpose.

May 10, 2013 scientific abstract: The paper reported that effective clinical treatment with anti-CD19 CAR T cells was first reported in 2010 after a patient with advanced-stage lymphoma treated at the NCI experienced a partial remission of lymphoma and long-term eradication of normal B cells.

The article went on to say that while anti-CD19 CAR T cells are at an early stage of development, the potent antigen-specific activity observed in patients suggests that infusions of anti-CD19 CAR T cells might become a standard therapy for some B-cell malignancies.

November 15, 2013 scientific abstract: The NCI reported that they had administered 23 total CAR-T-cell infusions to 20 patients. It said that the first 9 CAR-T-cell treatments were reported by Dr. Kochenderfer in articles in Blood in 2010 and Blood in 2012.This article summarized the data to that point in time and said that the NCI study was the first report of successful treatment with anti-CD19-CAR T cells of chemotherapy-refractory primary diffuse large B-cell lymphoma (DLBCL) patients, mediastinal B-cell lymphoma (PMBCL) and other unspecified aggressive lymphomas.

The article said that there were 8 treated patients with either DLBCL or PMBCL who were chemotherapy-refractory. It said that 5 of these 8 patients (63%) obtained either a CR or PR in this trial. One patient died suddenly of unknown etiology 16 days after infusion of CAR T cells. The abstract did not discuss other types of cancers treated.

This is essentially topline reporting that emphasizes the objective response obtained was 63%. The breakdown between CRs and PRs was not given nor was duration of response. It also did not give a breakdown between CRs in DLBCL or PMBCL. There was no mention of side effects.

August 25, 2014 Kite Press Release:

Kite put out a press release on 13 patients treated in the phase 1/2a NCI trial in patients with advanced B-cell malignancies. The published clinical trial results relate to patients in the second cohort in the NCI's Phase 1-2a clinical trial. You will recall that data from my source that 15 subjects had been treated in cohort 2 as of November 2014 and that two of these were retreatments of patients first treated in Cohort 1. The article does not clarify this point, but I think that the 13 patients were those who were treated for the first time.

Twelve of 13 patients (92%) had an objective response rate. This was broken down as 8 patients (62%) with CRs and 4 (31%) with PRs. Kite further reported that four of 7 (57%) patients with DLBCL achieved CRs of which 3 were ongoing and ranged from 9 to 22 months. In this press release Kite sometimes uses DLBCL to describe all aggressive lymphomas and not just DLBCL for some reason. In the footnotes of the release, Kite noted that one of these 7 patients with DLBCL had transformed from chronic lymphocytic leukemia, and three patients had primary mediastinal B cell lymphoma (PMBCL). This means that there were at most 3 patients with DLBCL in this 7 patient group.

We don’t know the CRs or duration of responses for the 3 DLBCL patients, 3 PMBCL patients or the transformed CLL patient. We are just given the aggregate data. This omission is troubling. Recall that in ZUMA-1 there will be 72 patients with DLBCL patients treated in cohort 1 and 40 patients in cohort 2 with either PLBCL or TiNHL. Very importantly, Kite has suggested that the ZUMA-1 trial will likely be halted after at an interim analysis after 30 DLBCL patients have been treated and that this will support a BLA filing leading to approval in 2017.

Given the importance of the DLCBL cohort, I would like to have seen this press release separate them from the 3 PLBCL patients and the transformed CLL patient. Again there is lack of transparency on Kite’s part. There was no data reported for the first and third cohort. As you will recall, the chemotherapy pre-conditioning and cell infusion doses in cohort 2 are closer to that used in ZUMA-1 than cohorts 1 and 3. This could explain the focus on cohort 2, but in the name of transparency the comparable data on cohorts 1 and 3 should also have been presented.

October 13, 2014 Kite press release:

Kite reported that 14 of 20 pediatric or young adult patients (70%) with relapsed or refractory ALL experienced a complete response. Remember that the ZUMA-1 trial is in DLBCL, PMBCL and transformed indolent NHL patients. It is not in ALL patients. Twelve of the 20 ALL patients (60%) achieved a minimal residual disease (MRD)-negative complete response.

The most exciting data from Kite, Novartis and Juno has been in ALL where the CRs have been more dramatic than in DLBCL, roughly 80% to 90% versus 35% to 50% in DLBCL. I think that some investors have not separated the results seen in ALL from DLBCL. These ALL results are not predictive of what will happen in ZUMA-1

February 20, 2015 scientific article

The NCI issued another update on 15 patients with advanced B-cell malignancies. Nine had diffuse large B-cell lymphoma (DLBCL), two had indolent lymphomas, and four had chronic lymphocytic leukemia. I think that this is the same data discussed in the August 25, 2014 press release which spoke of 13 patients from Cohort 2. At the time Cohort 2 was made up of 15 patients and I speculated that 2 patients were omitted because they were retreated with cell infusions. This appears to be the same data as was discussed in the August 25, 2014 press release but I can’t tell for sure.

The report said that patients received a conditioning chemotherapy regimen of cyclophosphamide and fludarabine followed by a single infusion of anti-CD19 CAR T cells. This pre-conditioning was lightly touched on in the previous reports in this section. Of 15 patients, eight achieved complete remissions (CRs), four achieved partial remissions, one had stable lymphoma, and two were not evaluable for response.

In this paper it was said that there were 9 patients with DLBCL (I think this means DLBCL plus PMBCL patients as I previously discussed) as opposed to 7 in the August 25, 2014 press release. It also included two transformed indolent non-Hodgkin’s lymphoma patients versus none and four CLL patients versus one. Hence the data sets are pretty different.

CRs were obtained by four of seven evaluable patients with chemotherapy-refractory DLBCL; three of these four CRs are ongoing, with durations ranging from 9 to 22 months.

Kite Update as of August 2015 as Presented in 2015 10-K

In its 2015 10-K, Kite presented data on results from 26 patients as of August 2015. The data from my source was as of August 2014 and presented data on 29 patients. I am not sure why the number of patients discussed dropped from 29 to 26. What happened to these three patients? Did they leave the study, move on to another treatment or die?

Kite doesn’t say and at this point I don’t know but this is discomforting as inclusion of these three patients might or might not have an impact on the data reported. The data I previously cited showed a CR rate of 35% in 17 evaluable DLBCL/ PMBCL patients. Kite presented data on 14 DLBCL/ PMBCL patients that showed a CR rate of 43%. However, Kite reported on three fewer patients. Again, we do not know what happened to these three patients. Here is the table as presented in the Kite 10-K.

These were the responses observed at one month. In terms of duration of response Kite said that 15 of 26 patients remained in response with 11 having durations of response of greater that one year. Three patients were retreated and their duration of response ranged from 18 to 52 months.

Unfortunately, Kite does not give information on duration of response in DLBCL/ PMBCL patients and TiNHL patients. This is an issue for me because based on information from my source, duration of response is longer with CLL and TiNHL so that this data may not accurately portray duration of response in DLBCL/ PMBCL patients.

June 6, 2016 Kite Press release:

In early 2016, Kite released additional data. The data was presented on 22 patients, which was four less than the patients reported on the Kite 2015 10-K. I don’t know what happened to these four patients. The data was based on the best response at time.

• In these 22 patients, there were 12 CRs or 54%.
• Of these 22 patients, 12 had on going remissions > 1year, 9> 2 years and 3 > 3 years.
• Again Kite does not provide duration of response data by DLCBL/ PMBCL, CLL and TiNHL patients. Hence the data is difficult to interpret relative to the two earlier data sets.

Subset of Data on Patients Treated with Kite Manufacturing Process

KTE-C19 cells being used in ZUMA-1 use the identical anti-CD19 CAR construct and viral vector that was used in the NCI clinical trials. There were some changes made in manufacturing that streamlined the process over that of NCI. In July 2014, the NCI submitted an IND amendment in order to use Kite’s streamlined and optimized process. Remember that in living cell manufacturing that changes in process have the potential to change the therapeutic characteristics of the cells.

Kite presented a subset analysis for 13 of these 22 patients who were treated with product produced by Kite's manufacturing process.

• In the 22 patient group, ORR was 70% at one month after infusion and the CR rate was 33%. For the 13 patients treated with Kite’s manufacturing process the ORR was 69% at one month after infusion and the CR was 46%.
• With longer follow-up the CR rate was 48% in the 22 patient group and 54% in the Kite manufacturing treatment group. This limited data sample suggests that the differences in manufacturing processes used at NCI and the one used by Kite is using in its clinical trials produced comparable results or perhaps superior results.
• Once again, Kite fails to disclose results broken out by DLCBL, PMBCL. CLL and TiNHL.

A Discussion of Side Effects in the NCI Trial

Overview

Investigators and regulators take into account all adverse events that occur after the start of the pre-conditioning chemotherapy regimen and on through for the period following cell infusion. The side effects seen with chemotherapy primarily result from effects on bone marrow and lymph nodes with the result that they can severely impact the production of white and red blood cells. The side effects from cell infusion are different and are attributable to cytokine release syndrome and neurotoxicity.

Before I discuss the incidence of side effects let me give you the definition of how side effects are graded:

Grade 5: Death

Grade 4: Life threatening requiring hospitalization

Grade 3: Serious

Grade 2: Moderate

Grade 1: Mild

Types and Incidence of Side Effects

Kite has provided virtually no data on side effects so that most of my information comes from the data provided by my source. It is as of November 30, 2014, 31 subjects had been evaluated for safety in the NCI study. I would emphasize that some of these patients received a more intense chemotherapy regimen, a higher dose of cells than was used in ZUMA-1. Some also received an IL-2 infusion. This probably means that the number and intensity of side effects in ZUMA-1 will be somewhat less what this section indicates.

• One patient died before evaluation and a second died after evaluation. These are grade 5 side effects. In both cases, the investigator concluded that the deaths were not related to treatment.
• All 31 had at least one adverse event of grade 4 and several had multiple grade 4 side effects.
• There were 89 grade 4 events in 31 patients that were attributable to the pre-conditioning regimen. These were: lymphopenia (30), neutropenia (21), leukopenia (22) and thrombocytopenia (16).
• There were 8 grade 4 events directly related to the CAR-T cells that were due to cytokine release syndrome and neurotoxicity. These were hypotension (3), speech impairment (2) and creatinine levels indicating serious kidney toxicity (2). Altogether there were 8 grade 4 events that were believed to be attributable to the cell infusion as opposed to the pre-conditioning regimen.

In 15 patients who received the highest doses of CAR-T cells, five patients (30%) required admission to the ICU for mechanical ventilation and/or vasopressors. It is believed that this resulted from cytokine release syndrome which is directly related to magnitude and rapidity with which the CAR T cells expand within the patient. The NCI hypothesizes that a lower dose of pretreatment chemotherapy would clear less space within the T cell compartment, causing slower or more limited expansion of the CAR T cells and reduce the intensity of this effect. It also reduces cell infusion side effects.

In regard to the two subjects who died within 30 days of the cell infusion. One died 18 days after infusion due to a cerebral infarction concurrent with viral pneumonia, influenza A infection, E coli infection, dyspnea, and hypoxia.  Another with PMBCL died 16 days after treatment.  No cause of death determined on autopsy and the autopsy report concluded likely cause of death was cardiac arrhythmia. Investigators concluded that neither death was due to treatment.

Cytokine Release Syndrome with CAR-T Cell Infusion

Cytokine release syndrome is induced by the activated T cells engaging the CD19 target.  Symptoms may include fever, febrile neutropenia, hypotension, acute vascular leak syndrome, elevated creatinine, renal failure, hypoxia, and pleural effusion. Twenty eight of 31 patients reported some grade 1, 2, 3 or 4 adverse events that could be attributed to cytokine release. Twenty four reported a ≥ grade 3 event and 6 experienced a grade 4 event. Clinical manifestations of CRS occurred typically in the first week after cell infusion and were less common in the subjects in with lower doses of cells.

Neurotoxicity with CAR-T Cell Infusion as of November 14, 2014

The symptoms of neurotoxicity include ataxia, confusion, somnolence and speech impairment.

Efficacy and Side Effect Data from Phase 1 Component of ZUMA-1

KTE-C19 ZUMA-1 Phase 1 Experience Based on Non-Kite Source as of August 14, 2015

Kite completed the phase 1 portion of ZUMA-1 in 2015. I was able to obtain information on the results from a source knowledgeable about the trial who was not associated with Kite as of August 14, 2015. They provided much more detail about phase 1 than Kite and so I will start there. At that time, there were 10 subjects screened, 8 were enrolled, of whom 6 were treated with KTE-C19. The 6 subjects enrolled and treated with KTE-C19 had r/r DLBCL that was either refractory to their last line of treatment or who had relapsed within 12 months of autologous stem cell transplant. Remember that the primary objective of the phase 1 component was to assess safety issues.

One of the six patients died. They were included in the safety analysis, but obviously not efficacy. The median follow-up time was 9.5 weeks after infusion. Treatment-emergent adverse events were defined as any adverse event that began on or after the start date of conditioning chemotherapy. There are different side effects associated with chemotherapy pre-conditioning and cell infusion. Some can be grade 4 life threatening side effects.

Prominent side effects associated with chemotherapy are suppression of white blood cells (lymphopenia, leukopenia and neutropenia) and suppression of platelets (thrombocytopenia). Cytokine release syndrome is specifically related to cell infusion. Among CRS related side effects are hypotension, cardiac failure, hypoxia, metabolic acidosis and acute kidney injury. Neurotoxicity is also related to cell infusion. Among neurotoxicity side effects are encephalopathy (swelling of the brain), aphasia (speech and communication impairment), tremor, somnolence, agitation, delirium and restlessness.

Including the pre-conditioning chemotherapy regimen and cell infusion, all 6 patients (100%) experienced side effects of grade 3 or higher. There were four grade 4 toxicities and 26 grade 3 toxicities in these 6 patients. The four grade 4 toxicities included one case of febrile neutropenia and one of thrombocytopenia that were almost certainly due to the chemotherapy pre-conditioning regimen. There was one grade 4 toxicity of encephalopathy (due to neurotoxicity) and one kidney injury (due to CRS) that were attributable to the cell infusion. In terms of grade 3 there were three cases of febrile neutropenia, two encephalopathy, two hypoxia and two somnolence. In addition, there were 17 other single occurring grade 3 side effects.

On an individual patient basis, the incidence of side effects ranged on the low end from one patient who had only grade 3 fatigue. The most severely affected patient had multiple side effects; these included grade 4 encephalopathy that lasted 16 days and required intubation, grade 4 acute kidney injury that lasted 11 days, grade 3 hypotension that lasted 17 days, grade 3 metabolic acidosis that lasted 16 days and grade 3 acute systolic heart failure that lasted 15 days.

One patient was included in the safety analysis set who died on study day 16 following KTE-C19 infusion due to an intracranial hemorrhage. This patient also had sepsis resulting from a pseudomonas infection which may have been caused or exacerbated by the chemotherapy. From a layman’s standpoint, I would be curious to know why the investigator dismissed any link between encephalopathy and intracranial hemorrhage.

All 6 subjects treated in the Phase 1 portion of ZUMA-1 received medications to ameliorate their CRS and/or neurotoxicity events.  Four subjects received both tocilizumab and steroids, and two received tocilizumab only.

KTE-C19 ZUMA-1 Results in Phase 1

Overview

For the first time in its brief corporate history, Kite created actual clinical data. This dealt with eight patients enrolled in the phase 1 component of ZUMA-1. As in the case of the NCI data, I am fortunate to have a source with detailed information on the trial. Again, I find Kite to lack transparency in their data release and to put a highly positive spin on the data. I compare information from my source to that released by Kite in this section.

KTE-C19 ZUMA-1 Phase 1 Experience Based on Non-Kite Source as of August 14, 2015

Kite completed the phase 1 portion of ZUMA-1 in 2015. I was able to obtain information on the results from a source knowledgeable about the trial who was not associated with Kite as of August 14, 2015. They provided much more detail about phase 1 than Kite and so I will start there. At that time, there were 10 subjects screened, 8 were enrolled, of whom 6 were treated with KTE-C19. The 6 subjects enrolled and treated with KTE-C19 had r/r DLBCL that was either refractory to their last line of treatment or who had relapsed within 12 months of autologous stem cell transplant. Remember that the primary objective of the phase 1 component was to assess safety issues.

One of the six patients died. They were included in the safety analysis, but obviously not efficacy. The median follow-up time was 9.5 weeks after infusion. Treatment-emergent adverse events were defined as any adverse event that began on or after the start date of conditioning chemotherapy. There are different side effects associated with chemotherapy pre-conditioning and cell infusion. Some can be grade 4 life threatening side effects.

Prominent side effects associated with chemotherapy are suppression of white blood cells (lymphopenia, leukopenia and neutropenia) and suppression of platelets (thrombocytopenia). Cytokine release syndrome is specifically related to cell infusion. Among CRS related side effects are hypotension, cardiac failure, hypoxia, metabolic acidosis and acute kidney injury. Neurotoxicity is also related to cell infusion. Among neurotoxicity side effects are encephalopathy (swelling of the brain), aphasia (speech and communication impairment), tremor, somnolence, agitation, delirium and restlessness.

Including the pre-conditioning chemotherapy regimen and cell infusion, all 6 patients (100%) experienced side effects of grade 3 or higher. There were four grade 4 toxicities and 26 grade 3 toxicities in these 6 patients. The four grade 4 toxicities included one case of febrile neutropenia and one of thrombocytopenia that were almost certainly due to the chemotherapy pre-conditioning regimen. There was one grade 4 toxicity of encephalopathy (due to neurotoxicity) and one kidney injury (due to CRS) that were attributable to the cell infusion. In terms of grade 3 there were three cases of febrile neutropenia, two encephalopathy, two hypoxia and two somnolence. In addition, there were 17 other single occurring grade 3 side effects.

On an individual patient basis, the incidence of side effects ranged on the low end from one patient who had only grade 3 fatigue. The most severely affected patient had multiple side effects; these included grade 4 encephalopathy that lasted 16 days and required intubation, grade 4 acute kidney injury that lasted 11 days, grade 3 hypotension that lasted 17 days, grade 3 metabolic acidosis that lasted 16 days and grade 3 acute systolic heart failure that lasted 15 days.

One patient was included in the safety analysis set who died on study day 16 following KTE-C19 infusion due to an intracranial hemorrhage. This patient also had sepsis resulting from a pseudomonas infection which may have been caused or exacerbated by the chemotherapy. From a layman’s standpoint, I would be curious to know why the investigator dismissed any link between encephalopathy and intracranial hemorrhage.

All 6 subjects treated in the Phase 1 portion of ZUMA-1 received medications to ameliorate their CRS and/or neurotoxicity events.  Four subjects received both tocilizumab and steroids, and two received tocilizumab only.

KTE-C19 ZUMA-1 Phase 1 Experience Described By Kite as of June 6, 2016

Kite presented data on June 6, 2016 that updated results from ZUMA-1. In this section, I compare results as described by Kite with those provided by my non-Kite source. Kite said that a total of 7 patients were treated in phase 1. This compares to my source stating that as of August 14, 2014 there were 8 patients enrolled and six had been treated.  What happened to the other patient enrolled in phase 1? I suspect that the patient who died 16 days after cell infusion was dropped, but I have not yet been able to confirm this.

Kite said that four patients had complete responses and that three had ongoing responses at nine months. It did not say how long the response was for the fourth patient. This would be important information for investors. Kite calculated the CR rate as 57% or 4 out of 7. If it had only counted the 3 with nine months duration of response the CR rate would be 43%. I don’t have the information to determine if the patient who died was dropped from the analysis, which would be inappropriate, but somehow one patient from the eight enrolled as of August 14, 2014 was dropped from the analysis. Using 8 in the dominator of the CR calculation the rate would have been 50% four CRs and using 3 it would be 38%

Kite said that KTE-C19 related adverse events consisted predominantly of cytokine release syndrome and neurotoxicity, which were generally reversible. Grade 3 or higher CRS were observed in 14% of patients and neurotoxicity in 57%; all were reversible except in one patient with dose-limiting toxicity. The latter patient died. Even though the study design assesses side effects in both the chemotherapy pre-conditioning and cell infusion, Kite only discusses side effects attributable to cell infusion which is misleading. It also only talks about grade 3 or higher side effects. It does not specifically break our life threatening grade 4 side effects. Again, this is misleading because we know from my independent source that there were two grade 4 side effects related to chemotherapy and two from cell infusion.

Novartis CAR-T Development Program

Overview

Novartis is a huge multinational pharmacy company and the 600 pound gorilla in the CAR-T room. Certainly for big pharma and indeed for all participants, it has been the first mover in the CAR-T space based on licensing in technology from the University of Pennsylvania and Children’s Hospital of Philadelphia. Novartis has an exclusive license to globally develop the lead CAR-T drug CTL019 and others in the pipeline

Novartis surprised me and most observers when it announced a restructuring of its CAR-T development strategy. It has had a separate CAR-T development group operating outside of its traditional research structure.  On August 31, 2016 it announced that it was laying off 120 employees out of 400 in its CAR-T groups and merging the remaining employees into its traditional research organization. Novartis explained this as a restructuring to become more efficient. However, I interpret this as a de-emphasis in the CAR-T and a vote of lack of confidence in the commercial potential of CAR-T drugs. Still they have an active development program in CAR-T.

CTL019 is the lead drug and is being studied in pediatric r/r ALL and r/r DLBCL. Novartis also has CAR-T drugs in development in multiple myeloma and acute myeloid leukemia. Pivotal studies for CTL019 in pediatric ALL started in 2015 and Novartis is planning on filing a BLA for this indication in the US and Europe in 2017.

CTL019 in Pediatric r/r ALL

Novartis recently presented updated results for its open label phase 2 study of CTL019 that enrolled 53 patients with pediatric r/r ALL. This study will be the basis of the BLA filing. The potential study showed that 93% had complete remission and that 76% were showing no progression at 6 months and 55% were experiencing no remission at twelve months. Eighteen patient were in remission beyond 1 year without further therapy. Cytokine release syndrome was observed in 88% of the patients but was reversible and controlled with anti-IL -6 therapy (tociluzumab).

CTL019 in r/r DLBCL and TFL

Early data was also reported data for patients with r/r DLBCL. At the 2015 ASH meeting it reported that seven of fifteen patients or 47% had a CR at with a median progression n free survival of 3.0 months. . For TFL, eight of eleven (73%) had a CR with a median progression free survival of 11.9 months.

Novartis began a study on 118 patients in r/r DLBCL in July 2015. The primary efficacy endpoint is ORR. It also has not given an indication on when the trial might complete or when a BLA might be filed. It appears that completion of the Novartis trial and ZUMA-1 should complete at about the same time. Novartis has not guided on whether there will be an interim look as Kite is doing in the ZUMA-1 trial.

Use of CAR-T in Solid Tumors

A CAR-T drug targeting EGFRvIII to treat glioblastoma has recently entered the clinic. Despite the efficacy shown by CAR-T drugs in hematological cancers, Novartis has cautioned that finding a safe target in solid tumors is challenging. Most antigen targets overexpressed in solid tumors also have meaningful levels of expression in normal tissue which can lead to severe side effects. Novartis cited the example of a CAR-T product against HER 2 positive breast cancer. The efficacy was impressive but HER 2 is also expressed in the tissue lining the lungs and resulted in unacceptable side effects.

 

 

 

 

Tagged as CAR-T Therapy, KITE, Kite Pharma, KTE-C19, ZUMA-1 + Categorized as Company Reports, LinkedIn