Agios Pharmaceuticals, Inc. (AGIO) Q3 2019 Earnings Report, Transcript and Summary

Good morning and welcome to Agios' Third Quarter 2019 Conference Call. [Operator Instructions] Please be advised this call is being recorded at Agios' request. I would now like to turn the call over to Kendra Adams, Vice President, External Communications and Investor Relations.

Thank you, Kevin. Good morning, everyone, and welcome to Agios' Third Quarter 2019 Conference Call. You can access slides for today's call by going to the Investors section of our website, agios.com. With me on the call today with prepared remarks are Dr. Jackie Fouse, our Chief Executive Officer; Dr. Chris Bowden, our Chief Medical Officer; Darrin Miles, our Senior Vice President of U.S. Commercial and Global Marketing; and Andrew Hirsch, our Chief Financial Officer and Head of Corporate Development. Dr. Scott Biller, our Chief Scientific Officer, will also be available for Q&A. Before we get started, I would like to remind everyone that some of the statements we make on this call will include forward-looking statements. Actual results could differ materially from those expressed or implied by any forward-looking statements as a result of various risks, uncertainties and other factors, including those set forth in the Risk Factors section of our most recent Form 10-Q filed with the SEC and any other filings that we may make with the SEC. With that, I'll turn the call over to Jackie.

Thanks, Kendra. Good morning, everyone, and thanks for joining us on our third quarter 2019 results call. Q3 was an exciting one for us as we achieved several of the ambitious goals we set for 2019 across our research, clinical and commercial activities. As Chris will describe in detail, we achieved several clinical milestones within our oncology portfolio that sets the foundation for multiple opportunities. In September, at ESMO, we shared the first full data from our positive Phase III ClarIDHy study of TIBSOVO in previously treated IDH1 mutant cholangiocarcinoma. We believe TIBSOVO has the potential to be the first targeted therapy for these patients, and our team is now focused on preparing that supplemental NDA. During the quarter, we've obtained guidance from the FDA and EMA on the endpoints for our Phase III trial of vorasidenib in low-grade IDH mutant glioma. We are now preparing for trial initiation. And earlier this week at the triple meeting, we presented data from the dose escalation portion of our Phase I study of AG-270 in a variety of MTAP-deleted tumors. The trial has advanced the dose expansion and patients are now enrolling in 2 combination arms evaluating AG-270 plus standard of care in non-small cell lung and pancreatic cancers. Within our rare genetic disease portfolio, we published updated data in the New England Journal of Medicine from the Phase II DRIVE PK study of mitapivat in PK deficiency. These data represents the first publication of clinical data in adults with PK deficiency and demonstrate the clinical benefit of mitapivat in this chronic anemia. On the commercial side, TIBSOVO U.S. revenue grew by 27% during the third quarter, and Darrin will get into more commercial detail later in the call. Outside the U.S., our team in Europe is executing on a gated build of our EU commercial infrastructure in preparation for potential approval of TIBSOVO in relapsed/refractory AML in the second half of next year. Finally, I would like to highlight the tremendous work happening within our research organization, with more than 15 active discovery programs, spanning target discovery to drug candidates, our talented team of scientists is making important discoveries every day in order to assure in the next wave of Agios' medicines. We will share more about these programs as we move them along. Heading into year-end, we are focused on executing against our remaining key milestones. These include completing enrollment for both of our Phase III ACTIVATE and ACTIVATE-T studies for mitapivat, submitting a supplemental new drug application for TIBSOVO in previously treated IDH1 mutant cholangiocarcinoma, advancing our Phase III INDIGO trial of vorasidenib in IDH mutant low-grade glioma and achieving proof-of-concept in our Phase II study of mitapivat in thalassemia. These milestones, combined with our remaining data presentations this quarter at both SNO and ASH, are critical to realizing the value creation potential across both our oncology and rare genetic disease portfolios in 2020 and beyond. I will now turn the call over to Chris to provide our clinical and regulatory updates.

Thanks, Jackie. As we laid out at the beginning of the year, our clinical development strategy for each of our programs is to expand across multiple disease areas where we have the potential to make an impact on the lives of patients. For our IDH inhibitors, we've established a clear benefit of these medicines in AML with approval of a single agent in relapsed/refractory and newly diagnosed disease. Our Phase III combination studies and newly diagnosed AML patients continue to enroll and progress as planned. Over the last 10 years, we have been leading the scientific and clinical understanding of IDH mutations in hematologic malignancies. And at the ASH Annual Meeting in December, we will share clinical and translational data for our IDH program that underscores our unique insight into this target and its implications for the AML treatment landscape. An area of great focus for Agios scientists [ who's ] expanding our data on mutational claims and minimal residual disease from patients who achieve a response with TIBSOVO. At ASH, we will have the opportunity to highlight some of this data in frontline AML patients who have received TIBSOVO plus VIDAZA in our ongoing Phase I trial. We'll share more about ASH presentation, when abstracts go online next week on November 6. Beyond AML, we reopened the myelodysplastic syndrome arm of the TIBSOVO Phase I study in hematologic malignancies with the goal of enrolling up to 25 additional patients and generating sufficient data to pursue a potential regulatory filing in this indication. This quarter, we made exciting progress toward advancing our solid tumor oncology programs. A ClarIDHy study of TIBSOVO and previously treated IDH1 mutant cholangiocarcinoma was the first randomized study in this patient population. As [ a ] discussion [ with ] Dr. Ian Chau at ESMO pointed out, these positive results helped to establish the practice-changing role that this targeted therapy can claim the treatment paradigm for patients with an IDH1 mutation. The study shows that TIBSOVO reduced the risk of disease progression or death by 63%. In addition, TIBSOVO significantly improved progression-free survival compared to placebo with notable 6- and 12-month PFS rates of 32% and 22%, respectively. By contrast, no placebo patients were free from progression at 6 months or beyond. While preliminary, the overall survival data are supportive of the therapeutic benefit of TIBSOVO and are especially impressive when adjusting for crossover. In addition, TIBSOVO was well tolerated [ in ] the safety profile [ as ] consistent with previous records. We are on track to submit a supplemental new drug application for this indication by year-end. On the heels of this positive Phase III study and [ an ] aggressive rapidly progressive solid tumor, we're increasingly confident in the potential for an IDH inhibitor to have a meaningful impact in IDH mutant low-grade glioma for patients who desperately need new treatment options. Vorasidenib is our brain-penetrant, pan-IDH inhibitor that we will study in patients with IDH mutant, low-grade gliomas postsurgical resection who are in a watch-and-wait setting, a clinical approach often employed in this disease to avoid the toxicities associated with chemotherapy and radiation. Vorasidenib has previously been evaluated in the Phase I study demonstrating 77% stable disease with a median treatment duration of 22 months in nonenhancing glioma. In addition, a perioperative study of vorasidenib and ivosidenib showed the ratio of tumor-to-plasma drug concentration is much higher for vorasidenib and is associated with more consistent 2HG suppression, consistent with our preclinical work. We will share updated data from the perioperative study at the Society of Neuro-Oncology Meeting in November. Today, I will walk you through the vorasidenib global Phase III study design. The study that we have named INDIGO will evaluate 366 patients with IDH mutant Grade II nonenhancing glioma in a one-to-one double-blind randomization to either 50 milligrams of vorasidenib once daily or placebo. Patients enrolled in this study must have had at least one prior surgery for glioma with the most recent surgery at least 1 year and no more than 5 years before the date of randomization and no other prior anticancer therapy, including chemotherapy and radiotherapy. The primary endpoint is progression-free survival as assessed by a blinded independent review committee. A number of secondary endpoints will be included such as safety and tolerability, tumor growth rate [ is ] assessed by volume, seizure control, time to next intervention, neurocognitive function and quality of life. Crossover from placebo to vorasidenib will be permitted. The team is actively preparing for initiation of the Phase III INDIGO study later this quarter. Continuing with solid tumors, I'll move to our AG-270 program in MTAP-deleted tumors. We reported the first data on 39 patients from the completed single-agent dose-escalation portion of our Phase I study at the AACR-NCI-EORTC Meeting earlier this week. The goal of this portion of the Phase I study was to establish a recommended dose based on the safety, pharmacokinetics and pharmacodynamics of MAT2A inhibition. On Sunday, Dr. Rebecca Heist from Massachusetts General Hospital given oral presentation followed by a poster presentation on Monday that showed AG-270 generates reductions in the biomarkers, plasma SAM concentrations and in levels of tumor SDMA at well-tolerated doses. The average reductions in plasma SAM concentrations were approximately 60% to 80% within the range associated with maximum tumor growth inhibition in preclinical models. The maximum tolerated dose was determined to be 200 milligrams daily. A confirmed partial response was observed in a patient with high-grade neuroendocrine carcinoma on the lung, 2 additional patients experienced prolonged stable disease, including 1 patient with a sex cord-stromal tumor with stable disease after 12.8 months of treatment with AG-270 once daily and 1 patient with the bile duct cancer with stable disease after 6.5 months of treatment with AG-270 once daily. Based on these clinical data as well as preclinical data that were also presented at the triple meeting that support our combination strategy, we are enrolling patients in 2 combination arms of the Phase I study. One arm will test AG-270 in combination with docetaxel in up to 40 patients with MTAP-deleted non-small cell lung cancer who have had no more than 2 prior lines of cytotoxic therapy. The second arm will test AG-270 in combination with nab-paclitaxel and gemcitabine in up to 45 patients with MTAP-deleted pancreatic ductal adenocarcinoma who have had no more than one prior line of cytotoxic therapy. The goal of each arm is to gather sufficient data in a homogeneous patient population to better understand AG-270's clinical profile when combined with standard of care, which will allow us to determine the next steps in clinical development. Moving to our PKR activated program where we are currently studying mitapivat in several hemolytic anemias. Our initial area of focus has been pyruvate kinase deficiency where mitapivat activates a mutated PKR enzyme. Earlier this year, we expanded clinical development into thalassemia and sickle cell disease where mitapivat has potential to provide therapeutic benefit through the activation of wild-type PKR. Beginning with PK deficiency, we have 2 ongoing Phase III clinical studies, ACTIVATE and ACTIVATE-T, where enrollment is expected to complete by the end of the year. In September, we announced that updated data from our Phase II DRIVE PK study were published in the New England Journal of Medicine and showed that hemoglobin responses were maintained in 19 patients in the extension phase of the study for a median treatment duration of almost 30 months. Additional follow-up data for DRIVE PK have been accepted for presentation at ASH as well as findings from the natural history study aimed at better understanding the burden of disease in adults with PK deficiency. Enrollment for the thalassemia Phase II study is going well, and we are on track to have sufficient data internally to establish proof-of-concept by year-end. In sickle cell disease, we are utilizing an IST strategy, which includes a cooperative research and development agreement with Dr. Swee Lay Thein at the National Institutes of Health. Dr. Thein expects to enroll up to 25 patients and to date, interest in the study has been high. As the first PKR activator in clinical development, we continue to have confidence in the important role that mitapivat can play for patients with chronic debilitating hemolytic anemia. I'll now turn the call over to Darrin to review our commercial activities.

Thanks, Chris. As I shared last quarter, our top priorities are execution and expansion. In the third quarter, we maintained our focus on continued execution support of TIBSOVO's frontline and relapsed/refractory labels, preparing for the potential launch of TIBSOVO for the treatment of cholangiocarcinoma and advancing prelaunch preparation activities in support of a potential approval of mitapivat for PK deficiency. With respect to our third quarter results, I'm pleased to report that the strength we saw in Q2 continued through the third quarter resulting in TIBSOVO net revenue of $17.4 million. This represents an increase of 27% over Q2 bringing year-to-date TIBSOVO revenue to $40.2 million. Growth in Q3 is largely due to strong gains in both the newly diagnosed and relapsed/refractory AML settings. In the frontline setting, our research suggest that the majority of treating physicians now believe in the importance of IDH inhibitors for newly diagnosed patients, and this belief is translating into a substantial increase in adoption in the frontline. Though we only promote approved uses of TIBSOVO monotherapy, we've observed an increase in frontline use in combination with HMAs. We've also seen a corresponding improvement in median duration on therapy between 4 and 5 months continuing a steady upward trend over prior quarters. Overall, we're making meaningful progress expanding utilization across new users and practice settings. [ Our new ] prescribers continue to grow up 27% over Q2 and up 65% since Q1. We continue to see strong double-digit growth in both the academic and community setting. While the academic setting represents the largest treatment volume, increased adoption in the community setting has been exceptionally strong and has increased its proportion of overall TIBSOVO volume relative to the first half of 2019. This is important as we observed more patients being co-managed between academic and community practices. I'm quite proud of the accomplishments of our team so far, successfully executing 2 launches in less than a year and tirelessly advocating each day for IDH mutation-positive patients. I'll look forward to sharing even more progress on our next call as we continue to execute and expand. I'll now turn the call over to Andrew to discuss our third quarter financials.

Thanks, Darrin. Our third quarter results can be found in the press release we issued this morning, which I'll summarize. More detail will be included in our 10-Q filing later today. Total revenue for the third quarter was $26 million, which consisted of $17.4 million of net U.S. sales of TIBSOVO, $5.9 million (sic) [ $5.5 million ] of collaboration revenue and $2.7 million of IDHIFA royalty revenue. Revenue increased compared to Q3 2018 by $10.8 million, which was largely driven by an increase in net U.S. sales of TIBSOVO of $13 million. TIBSOVO revenue growth of 27% compared to the prior quarter was largely driven by underlying demand. Gross-to-net and inventory levels remained consistent with the prior quarter. Cost of sales for the quarter was $393,000. Turning to operating expenses. R&D for the third quarter was $101.7 million, an increase of $19.1 million compared to the same period of 2018. The year-over-year increase in R&D was largely driven by startup costs for the planned Phase III INDIGO study of vorasidenib in low-grade glioma, clinical trial activity for mitapivat in PK deficiency, thalassemia and sickle cell disease and cost of the ongoing TIBSOVO combination Phase III trials in the frontline AML setting. Selling, general and administrative expenses were $33 million for the quarter, representing a $1.9 million increase over the third quarter of 2018. We ended the quarter with cash, cash equivalents and marketable securities of $540.5 million. A cash burn of $83.5 million for the quarter. We expect that this cash balance in addition to expected product revenue, expense reimbursements and royalties, but excluding anticipated program-specific milestone payments, [ we'll ] fund our current operating plans through at least the end of 2020. With that, operator, please open the line for questions.

[Operator Instructions] Our first question comes from Anupam Rama with JPMorgan.

. Congrats on all the progress. Just a couple of questions on ASH. So for the DRIVE PK extension study, is there going to be any new analysis we should be thinking about? Or is the focus really here on duration? And then for proof-of-concept in the thalassemia program, maybe you can walk us through what a win scenario looks like for establishing proof-of-concept?

Anupam, it's Chris Bowden here. For DRIVE PK, the main thrust of that abstract is really different people have a sense of the durability of responses and what safety looks like over that extended period of time. Given that -- given the activity, we've observed to date and the way we're thinking about treatment for this disease, long-term safety is really important as well as durability of the hemoglobin response. So really DRIVE PK is a pivotal trial in terms of the foundation support for that promise and then our ACTIVATE studies would continue to validate that would be our expectation. And with regards to the thalassemia, just let me back up for a minute and remind people on the call that this is an open-label Phase II study with 17 patients -- approximately 17 patients where we're studying 50 to 100 milligrams over a treatment period of several months and then patients can go into an extension. So what we're looking to understand is the percentage of patients who have a 1 gram or higher increase in hemoglobin over the treatment period. So success for us will really be predicated on the percentage of patients who have a hemoglobin response, how robust that hemoglobin response is. We also want to look at markers of ineffective erythropoiesis, it would be really important to understand whether we are able to cool off the bone marrow if you will and make it more efficient. This is the first trial where we're testing in the clinic our ability to activate wild-type PKR in a hemoglobinopathy. So those are all the questions we're asking. In terms of success, we're going to look at the response rate. We're going to look at the safety overall and then think about what that and what we see in terms of what things we think are coming in this disease in terms of other treatments that are coming for patients, talk to investigators and then that will -- those data and the totality of those data will be what will inform our decision.

Our next question comes from Michael Schmidt with Guggenheim.

I actually had one on IDHIFA and obviously, acquisitions can be disruptive sometimes, but I was just wondering what your sense is how committed Bristol is to potentially maximizing the commercial potential to -- of IDHIFA. For example, what are their plans to initiate additional label expanding enabling studies in AML and other indications?

Michael, it's Jackie. So look let me start with this way. I mean we can't predict what Bristol, Celgene's going to do nor exactly how they're thinking about this. I mean they certainly had a lot on their play. The partnership that we have with Celgene, it will transfer to [Audio Gap] continues to go well. We have had a great partnership with them for the last 9 years, and I expect we'll continue to have a great partnership with them. In terms of the ongoing trials for IDHIFA, I mean as I think you probably know there is the IDHENTIFY trial that's ongoing for the drug in combination with azacitidine. And then there is HOVON trial that both drugs, TIBSOVO and IDHIFA, are being studied in, in the intensive therapy treatment eligible patient population. So those are ongoing. Those will contribute to label expansions for both drugs over time. And as far as the rest of it, let's let Bristol and Celgene get their deal closed and do their portfolio prioritization and see what happens.

Okay. Great. And then a question on PKD. I mean when we talk to physicians about this, it sounds like the range of symptoms can be from very mild to severe. And I guess, maybe 2 questions. I guess so how is your work going in terms of patient identification at this point? And then based on that, I guess, what percentage of PKD patients do you believe are potential treatment candidates for mitapivat assuming success in trials next year?

Okay. Great. It's Chris here, Chris Bowden. So our patient ID efforts are ongoing globally, and I think we've done a really good job and gained momentum as we've been going, as we started this program, however many years ago. This started out as a learning experience for Agios in the rare disease space. And as we brought in more expertise and put more resources on it, we now have people distributed throughout the world, interacting with hematologists and physicians in terms of identifying patients. I would say one piece of tangible evidence around that is our ability to approve more patients than we initially anticipated in the ACTIVATE-T study. So that's something that we're going to continue to do because one of the challenges is when you have a rare disease where there are no existing therapies is there is not much of a reason -- there is not a lot of motivation to identify patients because there is nothing that you can do for them. So as we continue to move forward, education and awareness will be a really important component of that. Now the second part of your question, there is a number of ways you can approach and that is the percentage of patients who are potential treatment candidates. And you start from the premise that all patients who have a diagnosis and have anemia are potentially eligible to be treated. We are thinking about what their symptom burden is, what their disease burden is and that's not just the -- their anemia, but if over time, we were able to develop -- we're able to demonstrate improvements in other aspects of the overall disease burden and physicians may take that into consideration and, say, think about treating a patient. If you take a more strict view and you think about our trials, we've selected patients who have -- in the ACTIVATE study, who have a hemoglobin less than 10 and then there are various genotype profiles, you have to have at least 1 missense mutation, which represents about 80% of the adult population. On the ACTIVATE-T side, again, we put the same genotype aspects on, so we're looking at about 80% of those patients. So I think you have to balance those and hold those 2 perspectives in terms of thinking about what the percentage of potentially eligible patients are. And I think that assuming successful trials, I think, there is a possibility that clinicians may even consider patients with 2 non-missense mutations if, in fact, we haven't characterized the activity of our drug against those mutations. They might give them an empiric course of treatment for several weeks and if they don't respond, then stop. So it's going to be definitely an evolving picture in terms of the percentage of eligible patients based on what you pointed out a spectrum of disease severity, molecular features, what our data looks like and sort of physicians' comfort with empirical treatment regimens.

Our next question comes from Kennen MacKay with RBC Capital Markets.

Congrats on the operational and commercial quarter here. Maybe on that note, just a question on what we're seeing in terms of trends of TIBSOVO usage in frontline AML. Specifically, I was wondering around some of the conversation around the frontline combo with HMAs and I was wondering specifically if that combination with HMAs was what was driving the increase in duration that was mentioned on the call from 4 to 5 months versus the frontline patients who are essentially unfit for otherwise therapy and might otherwise move into hospice. And on that note, I was just wondering sort of what data stood out or what you're hearing from physicians for justifying use there? What really is incentivizing patients to use that combination?

Ken, thanks for the question. This is Darrin here. I think as I mentioned on the call, I think we've seen nice growth in both the frontline and the -- and in the relapsed/refractory setting. Given the makeup of our distribution channel, it's -- we have limited visibility into exact patient composition characteristics in the -- or talk about the first-line versus new onset [Audio Gap] presetting, but what our market research indicates to us is that we're seeing continued growth in both newly diagnosed and relapsed patients. This includes also the observation that roughly half of the use in the frontline setting is in combination with HMA. So I think it's important to note that we only promote monotherapy use in both frontline and the relapsed setting. But I think it's reasonable and actually based on the feedback that we're getting from our advisers and additional antidotes, I think what the adoption of combo in the frontline setting reflects is growing awareness and comfort with the data coming out of MD Anderson, evaluating this -- the combination of both ivo and HMA, which we've updated a couple of times this year. We'll have additional updates [ and ] publication in the future. So I think that's what's driving the adoption in that setting. And I think what those data reflect is significant observation in terms of complete response rate and CRH rate as well in those newly diagnosed patients. And I think that ultimately, since efficacy rules the day and a fairly well-tolerated safety profile, that's what's driving the utilization in the frontline setting.

And then maybe just going back to ASH last year, there was a big conversation about IDH inhibitors sort of versus something like venetoclax, especially in some of these earlier-stage patients who might be more tolerable of more -- of drugs with a more challenging side effect profile. I guess wondering your thoughts on going into ASH this year, is there any chance we could see data from the combination of the IDH inhibitors with VENCLEXTA. And again, on that same topics, wondering any sort of feedback you're hearing in terms of how physicians are making the choice between like an IDH combo with HMAs versus VENCLEXTA combo with HMA?

So we don't expect to see an update on the ven plus ivo combination, which is a study being conducted by Courtney DiNardo out of MD Anderson. I will just note that study is being expanded to include newly diagnosed patients as well in combination with HMAs. And to your point, I think the interesting conversation that's going on in the community around the various options that are newly available for them. So I think we're just heading into the norming stage after a significant storming phase where physicians are just trying to figure out how to use all of these -- all these new agents and in what particular patients. I think in this setting, in particular, again, what rules the day is going to be safety and efficacy. I think the -- our strategy has been to focus monotherapy in those patients who are previous -- secondary AML patients, so those patients have been previously exposed to HMA or disease like MDS and that from what we can tell from the feedback that we get in our market research as well as from the community, that's largely where we see the monotherapy use. I think it's important to make sure that we focus on what's driving sort of choices here. Group of patients who -- these are very sick patients, right? Oftentimes, these are older patients. Many of them previously treated for another disease and so are unwilling or unable to withstand the cytopenias associated with some HMA -- with HMA or HMA-based combinations and what they're going to appreciate about our data is the complete response rate, CR/CRh rate and the durability of that response and the safety profile. What physicians, I think, often appreciate is the clinical data supporting the transfusion independence as well as the time to recovery on platelets and neutrophils. I think that is what we hear consistently from those who are steady users of ivosidenib in the frontline setting.

Super helpful color. And one final question, if I may. It seems like, in enrollment, to your point, in ACTIVATE-T has gone sort of well above expectations with that trial and actually getting upside, and I think a lot of confidence in that getting fully enrolled by year-end. If one of these trials, for instance, ACTIVATE-T, does complete enrollment by year-end, ACTIVATE is still enrolling, is there a chance these data readouts could be -- could come in sort of separate time points next year as these independently reached their Phase III endpoints? And just wanted to take your temperature on whether that's -- or the confidence in that completion of enrollment in ACTIVATE by year-end. Is there a chance that could be sort of early 2020? Or is it really working on track for year-end '19?

Ken, it's Chris here. We're guiding to completing accrual by the end of the year. We're working really hard to get patients who are screened, randomized and into treatment for ACTIVATE and ACTIVATE-T, as you've said, has been accruing nicely. It's certainly possible that the readouts for the 2 trials could be disengaged, but I think that when you think about -- when we look at them to see how they're accruing or -- and then you've got the time on treatment, follow-up periods and all that. They're more likely to be engaged, but yes, sorry, it's possible.

Our next question comes from Mohit Bansal with Citi.

So one question I have is on your glioma trial. Could you please walk us through why you're choosing this particular population? Asking for because IDH mutation tend to be for positive prognosis factor for low-grade glioma, so how should we think about this particular control -- this particular subset of patients and your expectations around control arm, how could that -- how would that arm perform? And I have a follow-up after that.

This has always been an interesting aspect of developing a drug -- it's Chris Bowden here, in this indication because you have a -- when you look at published data, whether it's natural history data or from randomized trials with active treatment, you will see a long results in terms of progression-free survival and overall survival that is far longer than what you're used to seeing in trials where patients would have metastatic advanced solid tumors. And the question has been what's the unmet need and what are the logistics of doing a trial on this patient population? So let's start with the unmet need. The majority of patients who are diagnosed with low-grade glioma or younger relative to how you think about many of the solid tumor trials that you look at or, for instance, see AML patients that Darrin was just talking about in the previous discussion, were around 40 years of age, 40 to 50. So you have patients who, on the one hand, have a survival that's measured in 10 years plus, if you look at the entire population. Yet at the same time, if you've given a diagnosis at the age of 40 of low-grade glioma, you undergo a resection. And then the unfortunate part is, for most patients, is a long, slow and inexorable decline. And that decline is typified by gradual progression of disease, the use of radiation therapy, chemotherapy, steroids and other supporting measures that carry with them an enormous treatment-related burden in addition to the problems of having a progressive disease in the brain, which brings with it neurocognitive deficits and other problems. So what has been a long use clinical strategy, given all of that knowledge, is discussions with patients once you've had your initial surgery as we can watch and wait until you have progression and then institute radiation and chemotherapy and other ends, more surgery and other disease controlling measures. So as we were developing on drug and showing long-term stable disease as well as evidence of tumor regression when we look at our waterfall plots, this population emerged in discussion with investigators on the basis of the fact that we have a well-tolerated drug that can be given once a day for long periods of time in a group of patients who would like to stave off the use of agents like chemotherapy and radiation therapy. So that set the stage for the INDIGO study where we have carefully characterized group of patients who will have had surgery only and then eligible patients will be randomized to daily vorasidenib versus placebo with the intent to delay progression and as a secondary endpoint delay the need for chemotherapy and radiation therapy and all the problems associated with that.

Got it. And then how do you -- how should we think about the control arm expectations here in terms of months of PFS?

So we disclosed some of the initial details around the trial design today, and we will certainly update you with the statistical information in the future. I think it's important to both know what the control arm -- what we expect the control arm performance to be as well as what we expect the performance to be in the active arm because that's how we can set up a sample size that allows us to accrue a trial in a reasonable period of time based on our feasibility discussion with investigators. So that information is forthcoming.

Great. And then one housekeeping kind of question. So you mentioned that in AML, you're currently more penetrated in the academic setting. Can you please give us some color on what percent of patients right now on [indiscernible] are coming from academic versus community? And what is the general internal population, what is the breakup in these 2 settings for your target patient population?

No problem. So the -- it's not -- it has been our practice to disclose the exact split between those patients coming from the academic and the -- versus the community setting. I think it's fair to assume now that a significant portion will come out of the academic setting. What's important to note what we observed in the third quarter continues to be the similar trend to what we've seen elsewhere, which is what we're having or seeing in prior quarters and that is that we've got good strong growth in both settings, both in terms of scripts, our new patient starts, but also the number of new treaters that are identified in both of those settings as well.

[Operator Instructions] Our next question comes from Chris Shibutani with Cowen.

If I could just ask a series of sort of upcoming go, no-go decisions, particularly just to clarify what time lines and what we should expect. But first, you mentioned for the proof-of-concept data for mitapivat in thalassemia. You said you will have that in-house by the end of the year. When can we expect you to share that information? That will be number one. Number two, in MDS, you're continuing to enroll now again in the Phase I, which you had mentioned, I believe, in a previous quarter as well, when might we expect some information there? And what kind of time line would that be for making that go-no decision? And then, finally, for AG-270. I believe now that we have the data disclosure that Celgene has a 5-month period for which they'll make a go, no-go decision about whether they'll continue to partner that. Jackie, you've talked a little bit about being able to contemplate now so many years into the relationship with Celgene thinking about possible ways that a reconsideration of how that relationship is structured, is that ongoing and is that at all tied to AG-270, that would be helpful to know again what your thinking is there?

Chris, thanks for the questions. I'm going to just jump in and start with 270 and then let Chris come back on the thal and MDS questions. So the -- we have triggered the opt-in period with Celgene. So we have submitted the data package to them under the contractual agreements that have up to 150 days. They don't have to take 150 days, but they can take up to 150 days to ask us for additional analysis or ask us for more data and then go away and evaluate that data package. So again, I can't speak for what Celgene and Bristol together are going to do. As you know, we have multiple components to the relationship. One of them is around IDHIFA, which is more of a commercial partnership and then we have the metabolic immuno-oncology research collaboration that is also ongoing. And we have this program with 270 where they potentially may opt-in to that program. So we have to let them do what they're going to do in terms of getting their deal closed and look at their various portfolio prioritization decisions. What I've said in the past is, I think, they've been great partners in the past. They continue to be great partners, and again, I think they will be great partners in the future and I only see opportunity associated with that ongoing relationship. I don't see any downsides to Agios from the Bristol-Celgene merger. That's all we can say on that, and I'll give it to Chris for the other 2 questions.

Okay. Thanks, Jackie. Chris Bowden here. So myelodysplastic syndrome, Chris, is up and running and sites are opened for accrual. I can't guide to when we think we're going to have that data and -- but hopefully, I guess we get a little more clarity around the ramp up and patients coming in, then we'll be able to provide that information at a later date. MDS is about IDH1 mutation-positive patients represent about 3% overall of the MDS population. We're going after relapsed/refractory patients who really have high unmet need because they've all seen HMAs and other therapies. So it's a relatively small percentage, especially if you think about it in the context of IDH1 and AML where we're looking more like 8% to 10%. Nevertheless, given that we're working with the same sites that did our initial study. They're all centers of excellence. We think there'll be lots of interest in getting patients on to the study. So more to come there. As far as the proof-of-concept timing goes, our goal is to have adequate information in-house to enable a decision of whether we've achieved proof-of-concept or not by the end of the year. When we choose to communicate that whether we have achieved that going on is something that we're still looking at whether it would be at the end of this year or sometime early next year. This is something that we're working on.

[Operator Instructions] Our next question comes with Tyler Van Buren with Piper Jaffray.

I had just one last follow-up on the mitapivat, beta thal proof-of-concepts. Specifically, can you state what minimum response rate you guys think you need to see to achieve proof-of-concept? And just as a quick follow-up,does time to response matter as well?

I think that's -- the interesting question is what do you need to see a minimum response rate as well as time to response. I'm just -- touch on some of the aspects I've commented on earlier. First is to say, open-label pilot study or a study where we're trying to understand does the drug work or not. And that's the primary efficacy variable in terms of how we're looking that and see increase in hemoglobin of 1 gram or higher. It's a great question, if you do -- how does it impact us if it takes a long time to get there versus a very short time in patients with pyruvate kinase deficiency. We see a relatively rapid onset of response and that has a lot of interesting characteristics about -- that are generally favorable. So -- but I wouldn't hang my hat on that. Just like if the other aspect of -- if you see, let's say, a high percentage of patients who have a 1 gram increase and a very small percentage of patients who have a higher increase that might impact in terms of how we look at, how we achieve proof-of-concept or not. And then duration and safety is also important. We know a fair amount about this drug now from our trials in patients with pyruvate kinase deficiency, and we don't expect to see any major changes in the safety profile, but assumptions you have to test and you have to verify them. So I can't emphasize enough how important that safety component is as well. So I'm not giving you a specific answer of what response rate that we need to see because there are number of components that come into that phrase that you hear me is a lot, which is the totality of data, and we tend to look at things that way, especially in early -- earlier phases of development in single-arm open-label studies whether it's in the oncology or rare genetic diseases.

Our next question comes from Salveen Richter with Goldman Sachs.

Just touching on the AG-270 data from the triple meeting, the maximal tumor reduction remodels, is it served within like a range of 60% to 80% magnitude of the SAM reduction. In your view, how translatable are these levels of production from the mouse models to the humans? And then I have a follow-up.

Yes. This is Scott Biller here. Yes, that's the range that we achieved and see effects in preclinical models, but we see a range of activities within that lowering of SAM to some really sensitive models that have -- that we see actually tumor regression, but most models are actually -- we're seeing effectively stable disease or cytostatic effects. Translating that to now a set of very highly pretreated patients and rapidly progressing patients is really difficult. And it was always our intention to get into earlier lines of therapy in combination because our preclinical data on that combination was so very compelling. So that's -- we're sticking with our strategy and those are the clinical trials that are starting up now.

Great. And then if you could just comment on any updates or additional work being done around diagnostic testing in cholangio, specifically the ongoing efforts with Incyte to promote diagnostic testing ahead of their launch?

So -- this is Darrin here. So, as you may know Thermo Fisher is our partner for the companion diagnostic that would be approved at the time of the cholangio indication approval, which we expect by the end of next year. The -- in terms of collaborations, the -- I think the opportunity here is the sort of rising tide lifts all boats. You've 2 -- now 2 new effective treatments in a setting where there is little to no effective options for these patients and there is an opportunity to both educate on the importance of testing for both mutations, education on mechanisms -- mechanism of disease and then following the approval of both products, the education on the clinical profiles of the products. So I'm not prepared to discuss any work that we -- that could potentially be done in collaboration with Incyte at this time. I would say that we would -- I would expect that we both be in the community, discussing those pre-approval disease education with -- for our respective products.

Our next question comes from Mark Breidenbach with Oppenheimer.

I'll keep this quick. So given the trends you're seeing in TIBSOVO use in newly diagnosed AML patients, I'm just wondering, if there is a chance we might get an interim readout from the HOVON study sometime in 2020. And a real quick follow-up is, do you have plans or can you give comments on what you see as the best course for potential label expansion into newly diagnosed cholangiocarcinoma patients assuming a successful sNDA process once that's under review?

So we -- there should be no expectations that there'll be any information around the readout or efficacy from the HOVON study. That trial is a large Phase III trial that's been conducted by cooperative groups internationally. So we're very excited about the trial. And now we're working really hard with the groups to get this -- all the sites up and running and really get momentum going around the accrual. So there will be -- you should have no expectations around any interims on that next year. And then the second question around what to do in the newly diagnosed patients with IDH1 cholangiocarcinomas, one that's of great interest to us. We're in active discussions both internally with investigators on what the best study designs are thinking about feasibility and some of those aspects as we consider moving into that setting.

Our next question comes from John Newman with Canaccord.

Just wondered, if you have disclosed or talked about the dose levels at all for AG-270 in the combination study [ of ] nice initial data at AACR, partial response, stable disease, which was great. I think you mentioned maximum tolerated dose is 200 milligrams. Just curious if you're going to be looking at range of AG-270 doses in combination or if you'll have a set dose there?

Yes, John, it's Chris here. It's always -- it's up a little bit of a challenge, I'd say, when you move from monotherapy into combinations with chemotherapy. Both of those regimens that we're testing docetaxel as well as ABRAXANE, gemcitabine in pancreatic cancer bring with them a fair amount of toxicities such that when given alone, just those agents stand-alone, there is a lot of guidance in the label in terms of dose reductions whether [ it's ] for its -- for hematologic or nonhematologic toxicity. So that's the first component that one needs to keep in mind as you're getting ready to add another drug, in this case, AG-270 to those combinations. We're starting this with 100 milligrams daily. And if you think back to the comment -- the question that Scott Biller answered earlier is because at that dose, the data that we showed on -- that Dr. Heist showed and that we discussed on Sunday night shows that we've reached our target in terms of SAM reduction in that 60% to 80% range. And we'll explore that initial aim in the first patients that we treat. If we could go up to 200 milligrams in combination with those 2 regimens that I mentioned, but we're very comfortable that we can operate in that range and get the exposure that's associated with the pharmacodynamics or modulation that we're targeting. We're not going to do a whole lot of dose exploration. We don't think we need to do that based on the data that we got out of our Phase 1 trial.

Our next question comes from George Farmer with BMO Capital Markets.

On vorasidenib, I'm curious as to why you're not thinking about moving forward into later-stage glioma like in glioblastoma where you possibly can get a faster readout?

Yes. It's Chris Bowden here. So if you go back and you look at some of our Phase I data, we -- first, let's just talk about the data and then a few comments on why that may be. We did study some patients with GBM in our initial Phase I trial with TIBSOVO, and we've published that data a couple of times now. And when you look at it, patients progressed pretty rapidly. So that's the first thing. If you think about the frequency of IDH1 mutations in GBM, it's pretty low and those tend to be secondary. So what they are is there is a percentage of patients with low-grade glioma who are then nearing the end of their life unfortunately transform into GBM and that's a very difficult situation to handle. And that transitions to why is that because the molecular profile of the tumor at that point has changed pretty remarkably and the amount of drive to that tumor that's coming from IDH1 versus resistance mutations that have come up as a function of time, as a function of treatment with alkylating agents, as a function of treatment with radiation therapy is the ability to impact with this monotherapy [ in ] IDH1 inhibitor makes a probability of technical success pretty low. And we're not alone there. Think about all of the drugs that have gone into treatment in the GBM setting, and there is not a lot of new treatments there. So that's a really high unmet medical need, and if -- our data spoke to us in that way we would certainly go there because you're right, you can go faster. So as we think about the low-grade glioma population based on, [ A, ] our findings in the clinic and a lot of the work that we've also published in terms of making some historical comparisons of our data against similar data sets that we've been able to pull out from collaborations with academic institutions like MGH and Dana Farber and UCLA and others, we feel like that biologically it makes more sense to go there, and we put together a trial that we're confident can be done and demonstrate clinical benefit in a patient population that wow, yes, it has a longer survival than in patients with GBM, but that doesn't mean that there is not a high unmet need there. So with that, we're going there because we thin we have a highest probability of clinical and regulatory success in there where there is an unmet need with the use of a single-agent drugs. Oral therapy can offer some pretty significant benefits for these people.

Our next question comes from Andrew Berens with SVB Leerink.

I have a question on the glioma program. And then, if you allow, question on the mitapivat program. When we looked at the glioma program before and low intermediate grade really seem like it could take a while for a clinical benefit to be realized in a drug that's primarily stable disease and a disease that's primarily indolent. Can you give us some general idea of when we might see a benefit from the trial? How long that could take to show?

I'm not going to -- we're not going to provide guidance on the approval time lines or time to approval -- submission and approval in the setting of a positive study. I think as per some of the previous comments, we've done a lot of work to understand what the outcome will be in a control arm in terms of progression-free survival in a disease that once you do progress in this watch-and-wait setting, you're getting chemotherapy and radiation therapy. And in fact, our -- in our discussions with patients and investigators, they don't characterize that as an indolent occurrence, and so that's one of the reasons why we're setting the trial up the way it is. I think the other aspect that I just want to remind everybody on the call is that our trials are ongoing and will be updating the perioperative data at SNO and that will, of course, give you some insight -- further insight into that trial that we've previously published and then we'll, of course, update some efficacy data there as well.

Okay. And I guess, in the mitapivat program and beta thal specifically, can you give us a little color, I guess, on what type of patients you're enrolling in that trial? Are they going to be transfusion independent -- transfusion-dependent? And are there any biomarkers, specifically, that you're looking at that could potentially show a benefit for mitapivat versus some of the other options they have?

Yes. They are adults who are not regularly transfused, who have a hemoglobin less than 9 and as per some of the work you've seen us publish with mitapivat in pyruvate kinase deficiency and other chronic hemolytic anemia. We're interested in effects on bilirubin indirect, we're looking at erythropoietin and other markers of ineffective erythropoiesis, reticulocyte counts, LDH, a whole slew of things. We've also done some pretty neat, elegant work over the years in PKD, you'll recall, and volunteer as we saw nice changes in 2,3-DPG and ATP. When we translated that and started to look at that data in patients, we saw a trend in 2,3-DPG. The ATP data wasn't quite as clear and our group -- our metabolism group did some really nice work in terms of looking at flux through that pathway in patients who responded versus those who did not and that was very helpful for us in terms of further validating that. [ On ] equivocal clinical effect we were seeing was driven by the mechanism of action of the drug.

Okay. Are there any specific biomarkers that you may look for to enrich the population for mitapivat? Or is it just generally all beta-thal patients that are not transfusion dependent?

Yes. That's a really interesting question perspective because in pyruvate kinase deficiency, we're activating a mutated enzyme. The reason why we've always been interested in both thalassemia and sickle cell is that if you think about that battery slide that we sometimes show, by activating wild-type PKR in patients with thalassemia or sickle cell, your -- one hypothesis is by providing more energy through increased activity through the pathway you're generating more ATP and allowing those cells to have better defense mechanisms from the stress [ of their ] underlying hemoglobinopathy. So at this point, we do not have any data that would suggest there are nuances within wild-type PKR that we would do further selection.

Our next question comes from Michael Schmidt from Guggenheim.

Just a question on the AG-270 combination studies. We noticed that in your slides, the lung cancer cohort is now, I guess, focusing on frontline on small cell lung cancer patients, looking at AG-270 in combination with docetaxel and we were just wondering if that's a change relative to your prior plans? And I guess to -- I guess docetaxel is routinely used in frontline lung cancer, we're just wondering if there is a particular, I guess, rationale behind looking at those patients as opposed to later line setting.

Yes. Well, docetaxel was approved a long time ago in the second-line setting. And that's where we expect the majority of patients who will be coming through. They will have had probably at least platinum-based therapy and they will also probably in this DNA age have a high probability of having had a checkpoint inhibitor. So we'd expect that they would have had several prior therapies. And it's being pointed out to me there is an error on our -- on the slide we showed. So think about docetaxel in the context of its current indication and how it's used in the clinic as second-line therapies. You can have had 2 or 3 [Audio Gap]. Honestly, I think that if there is just a pretty good treatments in the frontline setting now that clinicians are not going to reach for docetaxel as first-line treatment for patients where that wouldn't be feasible.

Yes, cool. Cool. Thanks for clarifying. That makes a lot of sense. And then, I guess, how fast do you expect those combination arms to enroll?

Well, the faster the better, we don't -- they're open now and we'll be able to provide guidance as we get some momentum going and start to get some numbers in that we feel we can guide to, but nothing at this point.

Ladies and gentlemen, this does conclude the Q&A portion of today's conference. I'd like to turn the call over to Jackie Fouse for closing comments.

Thank you, operator. 2019 is an important year for Agios as we work to execute across our broad oncology and rare genetic disease portfolios. I would like to wrap up by thanking all the tremendous employees at Agios for their dedication and passion to making a difference for our patients. I also want to thank all of the patients, caregivers and physicians who participated in our clinical trials. Without them, we couldn't do what we do. Thank you for joining us today. We'll see you soon.

Ladies and gentlemen, that concludes today's presentation. You may now disconnect, and have a wonderful day.