Expert Financial Analysis and Reporting

Introduction to Report

I was closely watching the Amercian Society of Hematology (ASH) meeting held during December 3 to 6 for any meaningful new data on the CAR-T products: Kite’s KTE-C19, Juno’s JCAR-015 and JCAR-017 and Novartis’ CTL-019. There wasn’t anything really earthshaking that came out of ASH, but there were a few tidbits of information I found interesting. I have gone over these in this report and I have also discussed in some detail which company (ies) may gain first mover advantage in the CAR-T space.

In addition, I have included a section that discusses which company will be the first mover in the CAR-T space. Kite has been adamant that it will be the first mover and I believe this is the expectation of the investment community. This view ignores the presence of Novartis which potentially could gain first mover advantage. I would urge you to carefully read this section

Key Points of this Report

• I thought that Kite might release some data at ASH on the CR rate at six months for r/r DLBCL patients in its ZUMA-1 trial, but they didn’t. Kite says they will release six month CR data in a press release in 1Q, 2017. The FDA has asked Kite to present six month data in its BLA filing.
• Kite previously presented data that shows that at three months, the CR rate for 51 r/r DLCBL patients was 33%. Kite claims that this CR rate of 33% will be retained at six months and even goes beyond that. Kite has suggested that these patients are effectively cured.
• A CR rate of 33% means that one out of three patients treated may benefit. This is much better than the approximate 8% CR rate of other available therapies. This seems low to me given the very high cost and severe side effects, but other find this more compelling.
• But what if the CR rate at six months slips to say 25% or even lower? I think it would raise serious questions on how widespread usage might be KTE-C19 is approved.
• Novartis is poised to file a BLA for CTL-019 in r/r pediatric ALL in 1Q, 2017. Kite is guiding analysts that it will file in r/r DLBCL in mid-2017. Novartis started its phase 2 JULIET trial five months after Kite started ZUMA-1. It is unclear when Novartis might file a BLA in this indication, but they seem close behind Kite.
• There are many permutations on who will gain approval first and gain first mover advantage. Could it be Novartis with an approval in r/r ALL? This seems probable.
• Will Kite be the first to gain approval in the much more important r/r DLBCL market? Possibly, but because the Novartis trial is so close in timeframe to Kite, the FDA many want to consider the BLA applications and grant approval at the same time.
• I continue to believe that Juno will drop development of JCAR-015 in r/r adult ALL and focus its resources on JCAR-017 in r/r DLBCL. If so, it is two years behind Kite and Novartis. I think this will make JCAR-017 irrelevant.

Investment Opinion on Kite and Juno

I continue to recommend selling Juno. They are two years behind Kite and Novartis in gaining approval in r/r DLBCL. Juno’s management says that JCAR-017 will be the best in class and will be more effective than the Kite or Novartis products and some investors and analysts accept this representation. However, Juno is making this claim on the basis of very limited phase 1 data. I have not yet seen any data that convinces me that any of the three CAR-T products are meaningfully differentiated. Juno’s pipeline beyond JCAR-017 is still in earl stages of phase 1 trials or in the pre-clinical stage; it is difficult to assign much value to the pipeline at this time. Some bulls on the stock still believe that the trial of JCAR-015 will be restarted. If, as I expect, development is dropped I think it will be a meaningful downside event for the stock.

I am avoiding Kite. There is the expectation by management and bulls on the stock that the CR rate in r/r DLBCL at six months will remain at the 33% rate seen at three months. Moreover, they maintain that patients who experience a CR at three months are cured. In most cancer treatments, the CR rate declines over time so that this is not a bet I am eager to make. Kite also says that they will have first mover advantage in the r/r DLBCL market. This is possible, but it is probably that Novartis’s CTL-019 is approved in r/r pediatric ALL before KTE-C19 is approved in r/r DLBCL. It may also be the case that Kite’s KTE-C19 and Novartis’ CTL-019 are approved at the same time.

If management and the bulls on the stock are correct that the CR rate at six months will remain at 33%, the stock could rally. However, it they are wrong and the CR rate drop meaningfully, there could be very substantial downside. We will know the outcome in 1Q, 2017. The outcome on whether kite has first mover advantage in r/r DLBCL may not be resolved until 2H, 2017 or even early 2018. I am not involved in the stock at this time because of the fundamental uncertainties and the $2.3 billion market capitalization which seems high to me.

Comments on KITE

KITE originally released interim data on 51 r/r DLBCL patients treated in the ZUMA-1 phase 2 trial in a September 26, 2016 press release.  This same data set was just presented at ASH. You will recall that I believe that the critical data needed to support filing a BLA is a strong CR rate in r/r DLBCL patients. In September, KITE reported that at three months after dosing of cells that 33% of 51 r/r DLBCL patients had achieved a CR. Management and bullish analysts on the stock maintain that patients who experience a CR at three months are effectively cured and so the CR rate at three months and six months will be the same. There is really no meaningful data to support this extremely important hypothesis. I remain quite cautious that management and the bullish analysts may not have this right.

I was curious to see if KITE would update this data at ASH and we would see if the CR rate is indeed maintained at six months. It was not, but management said that an update would be released in a press release in 1Q, 2017. There is no hard and fast algorithm that would show that this 33% CR rate is or is not sufficient for approval, but I think that a 33% r/r DLBCL rate is somewhat borderline given the dangerous side effects and extremely high price of this therapy. The SCHOLAR-1 retrospective study of 635 patients treated with chemotherapy indicated that 8% of r/r DLBCL patients treated with chemotherapy achieved a CR and that median overall survival was 6.3 months.  My guess and it is just a guess is that if the CR rate in r/r DLBLC drops to 25% or below at six months that analysts will begin to question whether KTE-C19 has much commercial potential or even if it is approvable. Hence, I am eagerly awaiting the afore-mentioned press release.

Kite did present some new data at ASH. Of interest, there was data presented on an additional 24 patients with r/r DLBCL that showed that the CR rate at one month was 33%, the same as for the previously reported 51 patients at three months. However, for those 51 initially reported patients some achieved the CR in the one to three months after cell infusion. Still other patients initially achieved a CR and lost it at three months. Hence the CR rate for these new 24 patients at one month could go higher or lower at three months and beyond.

Comments on Novartis

In a poster presentation, Novartis showed for the first time that CAR-T cells can be effectively manufactured in a single centralized facility serving patients across four different continents (North America, Europe, Asia and Australia) at 25 treatment sites. The manufacturing failure rate of 6% was reasonable. In addition, another presentation showed that the 82% CR rate seen in the phase 2 r/r pediatric ALL trial confirms the 90% previously reported in the single-center, phase 1/2 trials at the University of Pennsylvania.

The actions of Novartis in folding its CAR-T program into the corporate R&D structure has raise questions as to whether its interest in the CAR-T program is waning. Still, NVS believes it continues to lead in the CAR-T space and stated that it will be first to provide regulators with global, phase 2 data in 1Q, 2017 in r/r ALL. I assume that this will be the basis of its initial BLA filing.

Novartis has not provided much information on its JULIET phase 2 trial of CTL-019 in r/r DLBCL and how its structure and timelines compare to ZUMA-1. I have compared data for the two competing trials as reported in ClinTrials.gov.

• ZUMA-1 started in January 2015 and JULIET started in July 2015. By this measure, Kite might have five month lead in the conduct of the trial.
• ZUMA-1 is scheduled to enroll 142 patients of whom roughly 100 are r/r DLBCL patients with the remaining being TFL and r/r PMBCL patients. JULIET is scheduled to enroll 160 patients, but I am not sure how many of these are r/r DLBCL patients and how many are TFL or r/r PMBCL patients.
• I am not aware of any plans to report interim data on JULIET.

Comments on Juno

Juno presented data from its phase 1 trial of JCAR-017 in r/r DLBCL patients. This trial is ongoing and the Company plans to begin a phase 2 registrational trial in r/r DLBCL later in 2017. Kite’s ZUMA-1 trial began in January 2015 and interim data was presented in September 2016. Full topline results from ZUMA-1 could be reported in 1Q or 2Q, 2017 and management of Kite predicts approval in late 2017. If the JCAR-017 trial follows the same time line, we could see interim data in 3Q, 2018, final data in 1Q, 2019 and approval in late 2019 or 2020.

At ASH, Juno presented data on its phase 1, dose escalating trial of JCAR-017. The Company reported that at three months, 8 of 19 DLBCL parents or 42% had achieved a CR. Juno management is claiming that JCAR-017 will be best in class based on efficacy and safety. In my judgment, there is no meaningful data that shows any difference between JCAR-017, Kite’s KTE-C19 and Novartis’s CTL-019. If there is a difference, it will take many years of experience to make that judgment. At this point, I consider the claims of Juno management that JCAR-017 is best in class to be unsubstantiated.

Comments on bluebird bio (BLUE, No Opinion, $68.55)

On November 30, 2016 bluebird.bio announced interim data on its ongoing Phase 1 clinical study of bb2121. This data was not presented at ASH, but it caused a pop in the stock and raised a few questions so I thought I would comment on it.

Bluebird bio reported early results from its phase 1 trial of its CAR-T product bb2121 in multiple myeloma; it is collaborating with Celgene on the development of this product. This is a CAR-T product that targets the BCMA antigen that is heavily expressed in multiple myeloma. Recall that Kite’s KTE-C19, Novartis’s CTL-019 and Juno’s JCAR-01 target CD-19 that is found in ALL, DLBCL, TFL, PMBCL and other hematological cancers. BCMA is a new target in a new cancer.

The Company reported data on 11 patients from a phase 1 dose ranging trial. It said that in six patients treated at the highest dose there were 2CRs or a CR rate of 33%. Across all dosages, the CR rate was 2/9 or 22%. Patients treated in the trial had had a median of six prior treatments including transplants, proteasome inhibitors (Velcade, Kyprolis) and IMiDs (Revlimid, Pomalyst and Thalomid).

There are also studies underway at NCI and UPENN on CAR-T products targeting BCMA. bluebird bio has a leadership position in this treatment area. Management stated that it had optimized all components of its bb2121 cell therapy (antibody, gene delivery, cell processing, etc.)

Being First to Market Will Be Extremely Important

Importance of Being First to Market

In any competitive marketing situation, being first to the market (or a fast follower) is almost always critical to success. As a generality, in drug marketing the first product to the market may get about a 60% market share, the second 25% and late comers split the remainder. I think this likely will be the case for the first generation of CAR-T products targeted at hematological cancers, i.e. Novartis’ CTL-019, Kite’s KTE-C19 and Juno’s JCAR-017. Here are things to consider when addressing the first to market issue.

Addressable Market for CAR-T Drugs is Small as Measured by Number of Patients and Treating Medical Centers

The first market to be addressed will be the r/r ALL indication for which there are probably less than 1000 patients. A somewhat more important opportunity will be the r/r DLBCL patient population. Kite estimates that the addressable patient population is 7,500 r/r DLBCL patients and a few hundred (I guess 750) r/r PMBCL and TFL patients. In the case of r/r DLBCL only about one-third of patients respond to therapy. Physicians will try to identify which patients will benefit and in a perfect world in which these patients could be prospectively identified (not yet the case) this would mean that 2,500 patients would be treated. Will physicians treat all 7,500 to benefit 2,500? I doubt it.

These patients will be treated at a limited number of sophisticated medical centers that can quickly be reached as the companies bring their drugs to market. Being first to market will allow the companies to quickly form strong relationships with a very limited number of institutions and physicians who will account for almost all prescribing of CAR-T cells. Subsequent companies entering the market would be at a significant disadvantage in that strong relationships had already been formed.

The Role of the Sales Rep May Be Very Different from the Way Drugs are Traditionally Sold

Novartis has indicated that it plans to use a different type of sales strategy to market its CAR-T drug. It will use the model often employed for medical devices rather than the traditional drug detail approach. In this approach, the sales rep is extremely sophisticated and functions as a resource to the physician, often sitting in on surgeries. They help train and educate the physician initially and then keep them abreast of advances in the understanding and use of the therapy as the physician universe moves up the learning curve. They work hand in glove with the medical centers on reimbursement and logistics issues. This approach can result in extremely strong relationships and great loyalty to the first mover.

http://www.mmm-online.com/commercial/novartis-considers-new-sales-model-for-experimental-cancer-therapy/article/496556/

Reimbursement Issues Will Be Quite Significant

There are three components to the cost of CAR-T therapy. The first is the pre-conditioning chemotherapy that involves the cost of the drugs and their administration. Then the patient is hospitalized (in the clinical trials this was for a week), given the CAR-T cell infusion and treated for side effects if needed. Some of these side effects can be serious grade 3 and grade 4 life threatening side effects that can require heroic and very expensive treatment to manage.

Kite seems to be pointing the investment community to a price range for CAR-T cells of $300,000 to $500,000 and I would lean toward the upper end of that range. There will be additional costs from the chemotherapy pre-conditioning program and the costs of managing side effects associated with both pre-conditioning and cell infusion. Kite has been opaque in discussing side effects, but I think that between the chemo pre-conditioning and the cell infusion that 30% to 50% patients may experience at least one grade 3 or 4 side effect. These can be extremely expensive to treat. If KTE-C19 is priced at say $400,000 the all-in cost of treatment might be $600,000 or more when these other costs are added in.

During the recent analysts meeting when asked about reimbursement, Kite answered simply that medical centers would somehow find a way to gain reimbursement for a breakthrough product like KTE-C19. I am more skeptical. Given the complexity and cost of this therapy, gaining reimbursement is going to be time consuming and difficult in my opinion. In any event, it is going to take a lot of time and effort by the manufacturers to help the institutions get reimbursed by insurers. Once this is in place, it may be a disincentive to go through the process again.

Manufacturing Logistics Are Complex

No less important than the physician/ rep relationship and help in obtaining reimbursement are the logistics involved in quickly and reliably manufacturing the CAR-T cells. The logistics of manufacturing CAR-T cells is a brave new world for the pharmaceutical industry which is used to making vast quantities of a medicine that can be used in all patients. CAR-T cells in sharp contrast are made specifically for one patient. T-cells are removed from a patient and shipped to an offsite manufacturing facility. There the cells are manipulated to create CAR-T cells in a process that takes as much as two weeks. They are then shipped back to the physician for infusion. You can see that this is a complex logistical situation and once the physician and medical center have gained experience with and confidence in the manufacturing cycle, it is going to be difficult for a second company to break into the center.

Which Company Will Be the First to Market?

If you listen to Kite management, they constantly stress that Kite will have first mover advantage in the CAR-T drug space. Key analysts supporting Kite echo this assertion. Both groups rightly dismiss Juno because the clinical hold in the ROCKET trial has delayed its possible entry, probably until 2020. Interestingly and importantly, Novartis seems to never be mentioned as a competitor by either Kite or Juno or analysts recommending those stocks. The reason is partly because of secrecy on Novartis’s part. I asked a knowledgeable member of KITE management about when Novartis might come to market and was told that Novartis was being very secretive and that Kite had no good sense as to when they might launch CTL-019.

Novartis has indicated that it likely will file a BLA for pediatric r/r ALL in 1Q, 2017. If so, they will clearly be the first to file for approval in any of the possible indications for CAR-T cells. Kite has said that it plans to file a BLA as early as mid-2017 in r/r DLBCL. Novartis has been mum on its filing plans in r/e DLBCL but its JULIET trial design is very similar to Kite’s ZUMA-1.  According to ClinTrials.gov Novartis started its trial in r/r DLBCL patients in July 2015 that will enroll 160 patients. Recall that Kite’s ZUMA-1 trial started in January 2015 and has enrolled 124 patients. Both trials are open label, phase 2 trials and use objective response rate as the primary endpoint. The secondary endpoints are similar. Novartis has not said if it will or will not conduct an interim look for efficacy as Kite is doing, but they could.

There are a broad number of scenarios that will determine when Kite and Novartis gain approval in r/r DLBCL in which one or the other could come first or they could come to market at the same time. It does seem highly probably that Novartis will gain approval of CTL-019 in r/r ALL before this as the data is outstanding. If they file for r/r ALL in 1Q, 2017, I think the product could be approved in 3Q, 2017 and marketing could begin in late 2017 or early 2018.This confers a meaningful advantage for Novartis as they approach the major centers and get a head start on building physician relationships and working out reimbursement and logistics issues.

I do not think that it is a given that CTL-019 and KTE-C19 will be approved in r/r DLBCL as the CR rate is modest and we do not yet know the duration of response; still I think they probably will be approved. Kite is giving a blow by blow description of its progress and has created high expectations for rapid approval. Novartis has been sphinx like so that we don’t know its filing plans. Kite started ZUMA-1 five months before Novartis JULIET. Does this mean that they will complete the trial five months earlier and file five months earlier? This seems to be the assumption of Kite management and bullish analysts.

As a note of uncertainty, I would point out that a BLA requires much more information than just clinical data. While the investment community rarely sees it, there is an enormous amount of data required on Chemistry, Manufacturing and Control (CMC). We as investors don’t know the quality of this CMC data, but it is not infrequently the reason for a Complete Response Letter denying approval. I have no way of knowing if this will be an issue for either company. However, I would point out that ZUMA-1 is the first clinical trial that Kite has ever run and it obviously has never filed a BLA. Does this mean that it is more likely to trip up over CMC issues than the highly experienced Novartis?

I would also raise an FDA issue. How will the FDA treat the Kite and Novartis BLAs? It is the same division and many of the same people reviewing the data. If I were the FDA, I would want to see and compare the data of each product on the same timeline. The BLAs are based on small numbers of patients and there is much to learn about KTE-C19 and CTL-019. Would the FDA move rapidly to approve the Kite BLA before it completes review of the Novartis BLA? I have no way of knowing, but I think that if the BLAs are deemed approvable, they would be approved at the same time. My guess is that the BLAs are approved in early 2018 and marketing begins by mid-2018. Kite is guiding to approval and marketing in late 2017 and that they will be first to market.

In the event that KTE-C19 and CTL-019 are approved at the same time, Novartis would have a meaningful advantage. Because of the expected approval of CTL-019 in r/r ALL before approvals in r/r DLBCL, it would have a head start on establishing relationships with key centers and physician and establishing reimbursement and logistics. How would this shake out in terms of market share? As a guess, I would say 60/40 in Novartis’ favor.

 

 

 

 

 

 

 

Tagged as CTL-019, JCAR-017, JUNO, KITE, KTE-C19, ZUMA-1 results + Categorized as Company Reports, LinkedIn