Clovis Oncology – The more you ignore me, the closer I get

Earlier this week at the EORTC meeting, Clovis provided an update on rociletinib (CO-1686) and rucaparib. Not only do the data prove that both drugs are highly efficacious in the relevant patient populations, they also provide key distinguishing factors (safety profile and patient selection) relatively to competing programs. The negative market reaction exemplifies the discrepancy between the progress Clovis is making and its stock behavior.

The stock is down ~20% in 2014 despite the fact Clovis is finishing the year with 3 high-quality oncology programs with clear routes to market and a differentiated value proposition. With the market ignoring the long-term potential for Clovis’ programs, the stock appears undervalued.


Safety overhang removed, no longer the underdog

Rociletinib is Clovis’ lead program, expected to reach the market in 1H16. The drug has impressive activity in lung cancer patients with a specific mutation (T790M) that confers resistance to first-generation EGFR inhibitors (Tarceva and Iressa).  Clovis is in a tight race with AstraZeneca (AZN), as both companies are developing similar drugs intended for the same patient population (lung cancer with T790M mutation).

Both drugs are highly efficacious and are expected to reach the market by mid-2016. Consequently, both companies are trying to differentiate their drugs in order to optimally position them upon approval. Given the comparable efficacy (based on cross trial comparisons that are unreliable), focus now shifts to the safety/tolerability where the two drugs are quite distinguished.

At ASCO this June, both drugs were presented back-to-back, demonstrating phenomenal activity (60% response rate) in T790M+ patients. Preliminary progression-free survival was numerically better for rociletinib but AZD9291’s safety profile appeared superior. The primary safety issue with rociletinib was an increase in glucose blood levels (hyperglycemia) whereas AZD9291 frequently induced skin toxicities.

In other words, Clovis appeared to have a potentially more efficacious drug with an inferior safety profile. This pushed the sentiment in favor of AstraZeneca as investors feared that the hyperglycemia induced by rociletinib would not be manageable and lead physicians to prefer AZD9291.

Since then, updated results were presented for both drugs. On the efficacy front, the two drugs appear similar with median PFS of 10.4 and 9.6 months for rociletinib and AZD9291, respectively (Cross trial comparisons are unreliable and inaccurate but that’s what we have for now). Clovis may have had a slightly tougher patient population based on higher percentage of patients who had progressed immediately after Tarceva/Iressa and non-asian patients.

Rociletinib - PFS

While AZD9291’s safety profile remained relatively unchanged, new data for rociletinib proved that the hyperglycemia concerns were overblown as it can be easily controlled with metformin, a common generic anti-diabetic drug (No patient from the relevant cohorts required insulin).

Although rociletinib’s efficacy does not appear better than that of AZD9291, the better than expected safety profile should allow Clovis to capture a significant market share in the T790M segment, estimated to represent a $1B opportunity.

If anything, Clovis now has a slight advantage in my opinion given the lack of skin toxicities. The initial patient population for both drugs will be patients who progress on 1st generation EGFR inhibitors (Tarceva, Iressa), which cause significant skin and GI toxicities. These toxicities are rarely life threatening but they are quite bothersome to patients and have a major impact on quality of life. When given a choice, I expect patients who were on Tarceva for 10 months to go with the drug that has no EGFR-related side effects at the “cost” of an asymptomatic side effect that is easily manageable with a commonly prescribed oral drug. Therefore, assuming the two drugs are equally efficacious, rociletinib should have a better market share (~60%).

Opportunity beyond T790M+

Clovis also presented intriguing data in patients with T790M-negative tumors, which were surprisingly positive with a response rate and PFS of 36% and 7.5 months, respectively. This is numerically better than AZD9291, which demonstrated a response rate and PFS of 22% and 2.7 months, respectively. This comparison is highly speculative given the small number of patients and potential misclassification of T790M+ tumors as T790M-, however, the difference can be explained by rociletinib’s effect on other EGFR mutations and/or IGF1R. It also bodes well for potential use of rociletinib as 1st line treatment although it is still unclear whether treating patients upfront instead of sequentially after Tarceva/Iressa leads to better outcomes.

Theoretically, approval in 1st line for rociletinib and/or AZD9291 could increase the opportunity to $2.5B depending on treatment duration (Tarceva and Iressa have combined annual sales of ~$2B). Preliminary results for AZD9291 in 1st line patients were somewhat disappointing with a response rate of “only” 63% vs. 70% that has been seen with Tarceva/Iressa. This casts a shadow on the value proposition of the new EGFR inhibitors as 1st line treatments although much more information is needed. Rociletinib’s activity in T790M- patients may imply better 1st line performance relatively to AZD9291 as most treatment-naïve patients do not have a meaningful level of T790M cells in their tumors.

Even if rociletinib and/or AZD9291 are not competitive as 1st line treatments, there is always the option to combine them with approved EGFR inhibitors in order to cover a wide spectrum of EGFR mutations. The first generation inhibitors will inhibit the primary activating EGFR mutations (exon 19 deletions, L858R etc.) and the second generation inhibitors will block the emergence of T790M-mediated resistance. In this scenario, Clovis has a clear advantage since rociletinib has no EGFR-related toxicities and is likely to be more combinable.

Clovis is evaluating rociletinib vs. Tarceva in 1st line patients with EGFR mutations in a phase 2/3 trial. The initial portion of the trial is open-label and should have initial data in 2015 to inform a go/ no-go decision.

Rucaparib – Initial validation for the BRCAness hypothsis

Clovis presented phase II data for rucaparib (PARP inhibitor) in ovarian cancer. In contrast to other PARP programs which focus solely on BRCA+ tumors, Clovis is putting a lot of effort on identifying tumors that do not carry the BRCA mutation but are still sensitive to PARP inhibitors (aka BRCAness). Clovis is using a diagnostic test designed in collaboration with Foundation Medicine (FMI) to identify a BRCAness signature, by looking at various DNA repair genes other than BRCA.

At the EORTC meeting, Clovis presented preliminary validation for the “BRCAness” approach in patients whose tumors are not BRCA-mutated. The response rate in the BRCAness+ patients was 32% vs. 8% for BRCAness-. This implies there is a significant group of ovarian cancer patients who could benefit from PARP inhibitors, irrespective of BRCA mutation or platinum sensitivity.

RucaparibClovis’ strategy with rucaparib is a great example of how a company can differentiate its product via smart patient selection. Obviously, other companies are likely to adopt the BRCAness strategy down the road but Clovis should have a clear head start with the only PARP inhibitor initially approved for BRCA-like+ ovarian cancer (assuming phase III success). BRCA-like ovarian cancer represents a significant portion of ovarian cancer patients (42% in Clovis’ study), which should translate to a meaningful commercial opportunity.

 It is still premature to assess the commercial opportunity in BRCA-like ovarian cancer but assuming prevalence of 20% of all ovarian cancer cases translates to 4500 patients in the US alone. In addition, the BRCAness signature may be applied to additional tumor types, which may significantly broaden utility of rucaparib and other PARP inhibitors.

Rucaparib’s data are also great news for Foundation Medicine, which will sell the companion diagnostic if rucaparib is approved. Foundation Medicine should also benefit from the overwhelming results Agios (AGIO) presented for its IDH-1 inhibitor (AG-120). Surprisingly, the market continues to ignore the tremendous potential represented by Agios’ IDH inhibitors and Clovis’ rucaparib to Foundation Medicine. As I recently discussed, the companion diagnostic angle represent a significant near-term upside that is not subject to reimbursement risks. Combined, these programs represent ~$100M in potential revenues in the US alone (assuming $3000 per patient), based on the incidence of AML and ovarian cancer. Expansion to other tumor types for these drugs could translate to a meaningful increase in FMI’s revenues as well.

Lucitanib – de-risked following Ultragenyx’s anti-FGF23 data

Although Clovis has not provided new data for lucitanib (dual FGFR/VEGFR), I still view it as a promising and undervalued drug. Preliminary data from EOS (which Clovis acquired in November 2013) suggests lucitanib has robust efficacy in breast cancer with FGF dysregulation. Importantly, such activity is not seen with competing FGFR inhibitors, likely due to lucitanib’s unique selectivity profile. Data from a breast cancer trial in the US are expected in 2015 and could validate the initial efficacy signal observed by EOS. It is unclear whether Servier (Clovis’ partner) will present updated results from an ongoing European trial in San-Antonio Breast Cancer symposium next month.

Recent data from Ultragenyx (RARE) for its FGF23 antibody further de-risks lucitanib as it alleviates concerns that lucitanib is not a true FGFR inhibitor given the lack of increase in phosphate levels observed with other FGFR inhibitors (If lucitanib does nit inhibit FGFR in patients, the signal observed to date may be a statistical artifact). Clovis’ explanation for this property relied on the fact that lucitanib does not inhibit FGFR4 whereas other FGFR inhibitors inhibit all FGFR isoforms (1-4). Ultragenyx’s data in a rare genetic disease (X-Linked Hypophosphatemia) clearly show that inhibition of FGF23 leads to increase in phosphate levels. Since FGF23 is known to signal also via FGFR4, this indirectly validates Clovis’ hypothesis.

Portfolio update

Following the positive results for rociletinib and rucaparib, we are adding a second position in Clovis and a third position in Foundation Medicine. We are also initiating a position in Conatus (CNAT). We are selling all of our Ambit (AMBI) positions although in real life the right choice would be to hold the CVR. We are also selling one position in Genmab (GEN.CO) and a portion of our Morphosys (MOR.DE) holdings for a profit of 52% and 487%, respectively.

 Portfolio holdings – November 23rd, 2014

portfolio - 23-11-14 - after changes biotech etfs - 23-11-14

39 thoughts on “Clovis Oncology – The more you ignore me, the closer I get

  1. hi Ohad

    nice to see you added CNAT.

    i really do not get their low market cap, and continue to look for red flags. aside from patent position (a little unclear I think) and back then the trial halt with pfizer i cannot find anything else. their strategy seems sophisticated, and the results look good (ok i am not the expert for this anyway..)

    definitely it will be a very exciting year for CNAT, question is only whether north- or southwards *lol*


  2. Well done, Ohad! As of small number of T790m- patients presented at 500/625mg doses, you can make inference from responses in overall population: in 179 patients enrolled in CO1686 ph1/2 trial on therapeutic doses regardless of T790m status – slide 14 – ORR=46% DCR=84%, which implied quite good response and DCR rates from T790m- patients. This type of number could be approvable without T790m+ selection.


  3. Christian – I view CNAT as a very high risk investment but the risk is already captured by the low market cap. On top of the new MOA that has not been validated for liver diseases, teh issue of long term safety is a major overhang. What I like about them is the good biologic rationale and the clear PD marker data.

    JQ – Thanks. Right, that’s definitely an option going forward plus it obviates the need for patient selection. Given the lack of other treatment options for Tarceva failures, physicians should feel comfortable prescribing a drug like rociletinib that has very few side effects to all patients. Not sure whether Clovis will pursue this initially, could make things more complicated.



  4. Sukkeralf – Basically it’s a combination of rich valuations across the biotech sector and specific concerns I have regarding MOR’s CD19 program and both companies’ CD38 programs for which I am still optimistic but they are not as potent as I originally hoped.



  5. Ohad, can you elaborate on the case for CNAT a little further? A doctor friend of mine said that the capase activity biomarkers are no longer relevant for liver disease.

    Do you have any comment on ARQL’s presentations at EORTC?


  6. Ohad,
    I have been following you blog for a while. Do you have time to look at SGYP? They have a ph3 and a ph2 drug with quite positive clinical results however market vale was quite low. It seemed the market did not care about its success at all. Thanks.


  7. Ohad,

    Regarding CLVS,do you have any concern regarding the competing drug from ARIAD PHARM-AP26113.there data to date has been very good but doesn’t seem to get the press that the competing AZN drug gets.


  8. Richard – Actually I am still gathering information about CNAT. There are a lot of question marks, primarily the correlation between the clear effect on cCK-18 and clinical outcome. I like the low valuation and the broad clinical program.

    Re: ARQL, so far I didn’t see anything unusual. Focus now shifts to biomarker-defined populations and proteus syndrome for 092.

    JInyu – Sorry, not following SGYP.

    david – ARIA’s drug is not an EGFR inhibitor in patients and I am not aware of any activity in T790M patients. Activity in ALK+ NSCLC is promising, though.

    chris (MRTX) – My main concern about their drugs is lack of selectivity which will make it challenging for them to reach deep and durable inhibition of Axl and MET.



  9. BP – OMED are doing phenomenal science but clinical evidence to date is not compelling in my opinion as most of their data is either from single agent studies with limited clinical activity or single arm combination studies from which it is hard to draw any conclusions. In general, the cancer stem cell theory has yet to be validated in humans. Bottom line, too expensive for a price of 650M.

    btw, I thought the notch biomarker data they presented at EORTC was very interesting (but again, preliminary).



  10. Hey Ohad
    thanks for the update, I was going to follow you into CNAT but the prices shot up 15% just yesterday! Good timing on your part. You have a knack for timing your purchases! Same thing happened when you wrote abou AERI and SAGE!

    Re Clovis – Nice title, it seemed familiar but didn’t get the Morrissey quote until I read the posts. Thanks for the clarifications and the info regarding RARE and derisking (I wasn’t aware of that.) Yes, the market doesn’t seem to give (as per your title) any value to the recent update on Rucaparib, and seems very focussed on co-1686 PFS data compared to AZN compound. Regaring this head-to-head race, I read in a well-written article that the AZN study has 60% asian participants, (this particular population supposedly has better outcome than others). Asian population in CLVS study is at 11% (so big difference). Were you aware of this? Is this significant?
    When do you expect the market will stop ignoring CLVS?

    Thanks and nice reading you, as always


  11. Dan – Thanks. In the short run, it’s very hard to predict stocks’ behavior as they are very volatile. As you know, there are also opposite examples….

    Re CLVS – Astra did have significantly more patients in Asia compared to CLVS who had predominantly western population. You are right about the common perception that Asian patients respond better to EGFR inhibitors but this is true for 1st line inhibitors. For the new inhibitors we simply don’t have enough data to conclude.Since Astra has a large number of patients in its database they can certainly do the analysis but I haven’t seen anything from them. CLVS brought up this difference several times, it may or may not be the case.

    With respect to market sentiment, it’s hard to predict. ASCO will be very important obviously with updates likely for all 3 programs + update from AZD9291. there might also be a buildup in momentum as we approach mid-15 assuming CLVS maintains guidance to file by then.



  12. Is anyone else having trouble using the search function on this blog? When I type in something in the “search” box and click “search,” nothing happens.


  13. Ohad, BIND has gotten remarkably cheap after the Market showed extreme disappointment in the lung cancer results reported at EORTC. However, some analysts are saying that the Market was expecting too much and misinterpreted the results. The company is moving ahead in KRAS-mutant lung cancer. Their PSMA targeted drug may report out before the end of this year. Year-end tax-selling has also hurt–and may continue to hurt–the share price. The market cap is now a little over $100 million. With partnerships with Pfizer, Astrazeneca, Roche and now Merck, the Accurin story looks promising, but may need time to play out.


  14. Richard – This might be due to maintenance work carried out last week, let me know if you there are issues going forward.
    Re: BIND, I share the market’s lukewarm interpretation of the data. Activity is not impressive at all, hardly differentiated from what you would expect to standard docetaxel. The PK profile looks promising but it doesn’t translate to a real clinical benefit. I don’t get the rationale for pursuing KRAS mutants, looks like a decision based on classic data mining. The list of partnerships is impressive and may bring value down the road but in the near term future clinical results will matter the most.

    Christian – Thanks. Looks like VCs (TPG, Sofinnova) liquidating their initial investments. Although insider buying is always more positive, I don’t think this is a clear negative indication.



  15. Ohad
    did you look into MRNS? Validated GABAA mechanism with new MOA, good efficacy and safety/ tolerability. 90M company with derisked Phase III does not look expensive.


  16. andre – I am actually looking into MRNS these days given the points you bring up (agree to all). The fact their drug is an allopregnanolone drivative similar to SAGE-547 makes it very interesting. Given the huge difference in valuations, I am trying to find the catch…



  17. Ohad, My read of the potential “catch” is: SAGE547 is IV formulation for acute setting which not only is bigger unmet need but also will have much faster route to market while ganaxolone is oral formulation for chronic setting which not only is more competitive but also requires larger and longer duration of trials. Even if ph3 is positive 2H15, another ph3 is needed for NDA filing, thus time to market is several years away, plus ph2 result was good but not great. However, the two orphan indications, genetic epilepsy and FXS ph2 trial results 1H15, could change MRNS profile.


  18. Thanks JQ. Developing an IV version of ganaxolone shouldn’t be too hard, it’s usually the opposite route which is more challenging. If both drugs are based on almost the same active ingredient , you would expect them to have comparable efficacy across indications (assuming both have good brain distribution). I wonder if SAGE547’s orphan designation can block MRNS in SRSE.



  19. That’s possible as synthetic analog of allopregnanolone, with same chemical structure as allopregnanolone. Outside of SRSE, ganaxolone is ahead of SAGE217 (oral formulation) in orphan genetic epilepsy.


  20. Hi Ohad, a question regarding Syndax and there ongoing IPO efforts: is there still a promising perspective for entinostat (HDAC-inhibitor) despite palbociclib and other new agents in HR-positive breast cancer?

    It seams that palbociclib is everybodies darling, but… when i be right, then palbociclib had so far failed to demonstrate a statistically significant improvement in overall survival. entinostat does so, at least for now in a phase II trial.

    entinostat is in a PIII-trial under SPA an has a breakthrough therapy designation and and showed statistically significant improvement in overall survival (and PFS) in randomized Phase 2b trial.

    further: is there a reasonable hope for synergistic combined anti-tumor effects of entinostat and immune therapy (PD-1)?

    Thanks for all your work,



  21. Hubert – Broad spectrum HDAC inhibitors earned their bad reputation again and again across many indications. On the other hand Syndax’s p2 survival data are impressive, almost too good to be true for a drug that is predominantly a HDAC1/2 inhibitor. Agree also about palbo, the lack of OS benefit is disconcerting.

    roy – Although everybody has written off prostate cancer, the complete failure of COMET2 to demonstrate any pain benefit is very disappointing and indicates that the previous p2 trial is completely unreliable. Personally, this is an important lesson about the dubious nature of single arm trials. EXEL still has several catalysts next year with cobi+PDL1 and cabo in RCC data but I don’t see any reason to get back in given the debt issue and negative sentiment.



  22. Mike – It’s hard for to interpret the data as I can’t really say what is considered a clinically meaningful decrease in the PD markers they use (NT-ProBNP and proteinuria) and how much time it should take for a disease modifying agent to exert its effect. It’s also important to look at the kinetic of responses over time.



  23. hi Ohad

    ARRY MEK162 transition deal terms look really favorable in my opinion.

    but what is the compound worth as such – do you see a market for it in face of Mekinist?


  24. Is this good news for you ? As expected ARRY got it back.
    “Novartis will also provide Array continued access to several Novartis pipeline compounds including, but not limited to, LEE011 (CDK 4/6 inhibitor) and BYL719 (α-PI3K inhibitor), for use in currently ongoing combination studies, and possible future studies, including Phase 3 trials, with binimetinib.”
    — Is that good what is stated, Is ARRY a buy now ?



  25. Christian – This is great news for ARRY! Best case scenario on every front, especially the access to NVS’ drugs even for future p3 trials. Don’t know if they will shelf filanesib but they will certainly de-prioritize it.

    Ruhulla – This is excellent as it will allow smooth development of these combinations. Yes, ARRY is a buy now, the only overhang is an imminent fund raising.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s