As a long-term healthcare investor, I’m used to seeing some ridiculous stuff. As I write this (11/15/18), I think I just saw one of the top screw ups, in my opinion, for 2018. If I was on the board of directors of Blueprint Medicines (BPMC) some heads would have to roll. Let me explain why.
Blueprint and GIST background
BPMC is a targeted therapy company, focusing on designing small molecules to inhibit specific kinases (described here). Their lead compound, avapritinib, is claimed to be a very specific c-KIT inhibitor (and close cousin PDGF-a D842V) (pipeline chart here). GIST (gastrointestinal stromal tumor) is a sarcoma that is almost always driven by mutations in c-KIT (review here). If you might recall, c-KIT is made famous by Novartis’ Gleevec, which was one of the first small molecules to show clear efficacy against a mutated kinase – in this case CML and the Philadephia chromosome (see review here). In retrospect, it wouldn’t have been surprising, but after it was discovered that GIST is a KIT driven tumor, Gleevec was rapidly tested and shown to be highly efficacious. Guidelines recommend that Gleevec should be the first line treatment, particularly since it is no longer on patent (overview here).
In GIST, Gleevec efficacy does not last forever (certainly not as long as CML). The GIST can become resistant to Gleevec. Fortunately, several pharma companies have tested other small molecules to address these GIST resistance mutations. The two approved drugs, Sutent (label) and Regorafenib (label). As you might expect, it would make sense that, for the most part, Sutent and Regorafenib can inhibit GIST which are resistant to Gleevec (see data in labels above). Unsurprisingly, the GIST can mutate around both Sutent and Regorafenib relatively quickly, and over time, a patient with GIST will develop and accumulate multiple mutations in KIT which convey resistance to all 3 approved drugs (example study here). I think these facts should not be surprising, as this situation arises is most analogous oncology conditions when using small molecules.
Drug Development-101
I think it should be obvious, if you are a small molecule oncology development company, when you look at indication with several approved drugs, the first thing you look at is the area of unmet medical need. At a minimum, it makes the most sense to make a drug that can combat the resistance mutations which arise from the approved drugs. So if I was the CEO of BPMC, I would think that if I would wanted to go after the GIST market, I would tell the R&D team to make a drug which can address all the common resistance mutations which arise from patients who are treated with the 3 approved drugs used to GIST. It makes far more sense to make a drug which can cover the existing known resistance mutations, thus being able to address an unmet medical need than to make a “me too” drug. A drug which can cover the resistance mutations is much “easier” to develop, since you can get approval by running a trial which takes in patients who have no treatment options – which is faster (enrollment won’t be impacted by the availability of an approved therapy) and have a better risk/benefit equation (in a trial for patients with no options, the risk/benefit equation is easier than a trial where you are going head to head with an active comparator). Also, from the marketing perspective, I would rather market a drug which works in later lines, as it would provide confidence in using the drug in earlier lines. Also, typically, after you get approval in patients who have no options, you then subsequently run trials to move up in the line of therapy.
Blueprint’s Failure
Based on what I saw, I think the Blueprint management and board made a colossal blunder in drug development. BPMC presented an update on avapritinib (slides are downloadable are registering to hear presentation). See slide 19 below:
What the company shows in their 4L (fourth line which is the unmet medical need) arm (also is the vast majority of their treated patients) is that 2 mutations (V654A and T670I) basically do not respond to avapritinib (colored bars). While GIST is a disease that can generate many mutations, it is notable that it seems that 25% of 4L GIST patients have either of these mutations - a number the company acknowledges in the call.
How did the company management and board choose a molecule that doesn’t work in about 25% of the unmet medical need population - and if I heard correctly on the call, about 25% of all GIST patients? This is, frankly, baffling to me. If these guys are so smart, wouldn’t you have demanded that the declared candidate should work in the known MAJOR mutations AT THE START OF THE PROGRAM. I would think that 2 mutations which compose 25% of unmet medical need population should have been high on the list of targets. Jeff Albers, the BPMC CEO, stated on the call, “At Blueprint Medicines, our mission is to deliver transformative, precision therapies to patients with cancer and rare diseases. By leveraging our scientific platform, we aim to rapidly and reproducibly design highly selective kinase medicines with improved potency and safety…” I find it difficult to reconcile his statements when he didn’t demand that his R&D team make a molecule which can hit all major mutations found after the approved therapies. Did Jeff tell the team to make a molecule which can work in just 75% of the known resistance mutations? For a smart team, this seems to be a pretty low bar. The impact of this “hole” resulted in a median PFS (mPFS) in the 109 4L study of just 3.7 months. 3.7 months mPFS for a targeted agent is pitiful, in my opinion (see slide 20).
I decided to go look at BPMC’s competitor, Deciphera (DCPH). The purpose of my missive is not to go into who is better, but DCPH’s corporate deck (here) has some biochemical resistance data regarding the IC50 for BLU-285 (avapritinib). Keep in mind that this generated by a competitor, but I think there is no reason to doubt this biochemical data (slide 11). What we can see already that V654A and T670I have poor IC50 (lower right hand box) – so there was no reason to expect avapritinib to work in the first place. It was dead on arrival for those mutations and for many others. In fact, if this data is true (and BPMC seems to have validated it), there are a host of known resistance mutations you can use to easily predict which patients would be resistant to avapritnib. Note that BLU-285 is avapritinib.
The company is now spinning a story where now they can remove these patients with genetic diagnostics. Of course, if you exclude the genotypes that your drug doesn’t work at all, avapritinib works well. Duh. Note to Jeff, maybe you should have told your development team to design a drug that hits a pair of mutations found in 25% of all GIST patients. In fact Jeff, maybe you should make sure that the drug candidate should hit, say 95% of the known resistance mutations. Also, if the company was so dedicated to patient safety, maybe they should test all the known avapritinib mutations found in the above chart and exclude them as well, since it seems obvious that a number of mutations will lead to avaprinitib resistance. It would seem the ethical thing to do, wouldn't it? Why would you only exclude two of the most frequent ones. If the above data is correct, there are at least a dozen of other known GIST mutations which are very likely to resistant to avapritinib, yet the company seems fine to give these patients an ineffective drug. This is a slippery rabbit hole BPMC is about to go down. I don't know what the frequency of the other mutations are, but if they are 1-2%, this could add up to another 15-25% of additional GIST patients.
There is another potential problem in having a drug that doesn’t work on a known common resistance mutation. The notion that you could simply exclude about 25% of the patients to get a good result, I believe, a fool’s errand. This is because these common mutations are likely to be just around the corner – the GIST patients, who are currently negative for V654A and T670I, that go on this drug will likely relapse with these known mutations (or others). It obviously is not difficult for the GIST to develop these, since these are common with existing therapies. In fact, since BPMC is already sequencing the KIT genes, doctors and the company will know, prior to treatment, what mutations are present. If you are going to go through the trouble of sequencing the KIT, you might as well use that information to exclude all the patients you know the drug won’t work (such as D820Y, N655S, etc.). It’s absurd that you are willing to sequence the whole gene and only remove the 2 big obvious resistant mutations, yet you hold in your hands the information to exclude patients who will not benefit from the drug, yet you plan on treating them anyway. This galls me.
Flunky Sell Side Comment of the Day
I think the sell side analysts, in general, are total tools of the investment banks. If you listen to the call, you hear their ridiculous fawning congratulations from a variety of sell side analysts. My favorite was Leerink’s Andrew Berens, who said, “it’s always nice to wake up to good data.” Come on, is he really that clueless? I don’t think he knew the day before that avapritinib wouldn’t work in 25% of patients he was modeling before the call that day. Talking about selling your integrity. This issue is so basic that he either is a fool or tool. Neither is good. Additionally, BPMC closed down almost 7% today, on a day that the XBI up 3.5%. This is massive underperformance. Good luck with the buy rating Andrew. Unfortunately, based on the questions on the call, all the other analysts are similar to that guy.
I believe that if you wake up and find that your eligible patient population is at least 25% smaller than the day before - this is a bad thing. You would think this would cause a very significant impact to one's target price. Sell side is absolutely not your friend.
Conclusion
I am stunned at the development criteria set by BPMC. How did the management agree to develop a compound which won’t work in 25% of the patients right out of the gate? Also, what they propose for screening, I believe, puts them on a slippery slope. That 25% ineligible is likely much bigger, given the number of resistant mutations found on the DCPH table. I think clinicians have an ethical duty to look at the mutations which are present and use their own judgement if the patient will respond to avapritinib. Thus, this 25% exclusion will be just the start. Overall, I think these results are a massive failure on the part of BPMC management and the board.
Comments
Log in or sign up to join the conversation.