Precision Medicine's Progress: Reports From The Front Lines

A roundtable of industry experts discuss how advances in biomarkers and personalized medicine will impact pharma R&D decisions and ultimately benefit patients (free with registration).

Advances in biomarkers and precision medicine were in the spotlight at a session at Informa's BioPharm America 2015 conference in Boston in September. Christine Grygon, executive director of business development and licensing for prescription medicines at Boehringer Ingelheim GMBH, Ruth Phillips, oncology clinical manager for commercial effectiveness solutions at IMS Health Inc., and Cecilia Schott, head of personalized health care, corporate development and ventures at AstraZeneca PLC discussed these topics in a conversation moderated by "The Pink Sheet" senior writer Joseph Haas.

In this edited version, the panelists address what drug companies look for in a partnership to develop a companion diagnostic, opportunities for partnering in biomarkers and precision medicine, issues around next-generation sequencing, the possibilities resulting from basket and umbrella trials and the issues that arise when a patient's data indicate more than one appropriate route of treatment.

Joseph Haas: Where would you say we are in the process of bridging the gap between biomarkers and companion diagnostics?

Christine Grygon: If the biomarker’s not predictive, you're not anywhere near a companion diagnostic. The road is very, very long for therapy areas such as cardiovascular and CNS and other diseases where we just don’t have as much of a genetic understanding as we do for tumor biology. It’s a little bit shorter for oncology. But I don’t think we’re very far down the path at all, especially when you think of the number of therapies out there and the number of diseases out there, and the fact that we only have 18 companion diagnostics.

Cecilia Schott: For AstraZeneca, when I first started in personalized health care about seven years ago, our goal for 2015 was to get about 20% of our pipeline with an embedded personalized health care strategy. Now, 80% of our pipeline has a precise health care strategy. I think it’s now part of the fabric of drug development in big pharma. And we're moving beyond oncology into other therapeutic areas, which is very exciting.

JH: What can we say about the value of collaborations between drugmakers and test makers? Has it changed? How has it changed?

CG: As with any partnership and any collaboration, what's inherent in the value proposition is the risk level. Just a couple of years ago, putting a companion diagnostic into the marketplace was a much more challenging endeavor than it is today. So for instance, if you have identified a predictive biomarker and you have a mechanism for a test that can be approved by the FDA, and it can follow the pathway that leads to a companion diagnostic, then that’s a very different risk profile than, say, a new technology, where we don’t actually know how it’s going to pan out, and what it’s going to look like when it gets put through a rigorous clinical trial. There's still a range of partnerships between pharma companies and diagnostic companies. And I think we’re going to see that continue to evolve over the next few years.

CS: It also depends on how you also manage your value. If you come from a diagnostic company that has a platform, and you're going to be adding to the menu of that platform, that's adding value to it. If you're looking to the pharma company to bring the drug to the market, that's adding value to it. If you're looking to the patients getting access, there's value.

Christine Grygon, Boehringer Ingeheim

JH: It’s clear that in biopharma, some companies like Novartis and Roche have expressed a preference to keep their diagnostic work in house. Can any of you speak to what are the comparative values of developing companion diagnostics in house or through external collaboration?

CG: To merge a diagnostics business with a pharmaceutical business is really, really difficult under the same roof. [You have to assess] what are your strengths, what is your core business, what are your capabilities, and then what are you going to look outside for? In the case of BI, we’re not a diagnostics company. We don’t feel that that was something we were going to change our business model for, so we pursue diagnostic partnerships as part of our mandate when we need them.

Cecilia Schott, AstraZeneca

CS: We take exactly the same approach. We’re not a diagnostics company. So when you compare us with Roche, which clearly is the leader in the space, they have access to multiple technologies and that gives them a lot of flexibility. But we feel that by being able to choose a different technology and technology providers, we have a lot of flexibility to access innovation. I would say the Novartis model is a little bit different because Novartis is not a diagnostic company.

We do have a very strong translational science group and a precise health care and biomarkers group. Once we identify a biomarker, we have some assays that we develop internally for exploratory purposes and we get familiar with the technology. So when we do partner, we know where we are so we can speak as equally as you can from that perspective. So yes, I think it serves you well to have the flexibility to access innovation, unless you are a Roche that has very robust diagnostic expertise.

"To merge a diagnostics business with a pharmaceutical business is really, really difficult under the same roof . . . In the case of BI, we’re not a diagnostics company, so we pursue diagnostic partnerships as part of our mandate when we need them." – Christine Grygon, Boehringer Ingelheim

JH: Can any of you provide examples of successes or failures in partnerships around biomarkers or companion diagnostics?

CG: I can maybe point out two, and one of them might be a little bit unusual. We have a non-small cell lung cancer product on the market, Gilotrif [afatinib], which is an EGFR inhibitor, and it requires a companion diagnostic test. And we have a very successful partnership with Qiagen [Qiagen NV], which developed the test for us. On the flip side, thinking about the commercial side of that and how you reach the right customer in order to make sure that that test and its corresponding drug are sort of enabled in the customer population, we partnered with patient advocacy groups to set up something called Let’s Test, which is a patient education platform around understanding the value of what it means to get tested and why that’s so important. And this is really an effort into patient education which we have taken very seriously and are really, really proud of. So that’s kind of a different partnership. It’s not with the diagnostic company, but it’s around something that I think is really important. Making sure that the end customer understands the value of what they're getting is really critical.

CS: I come at it from a different perspective. I always think that the diagnostic succeeds, and what fails is the drug. The chances of the diagnostic succeeding and measuring what it’s supposed to measure is 99%. The hypothesis on the drug and the pathway is what becomes very challenging. And when I say the drug fails, it doesn’t mean the drug doesn't work, it just doesn’t work on the pathway that we thought it did. So can I give an example? I have several examples, unfortunately. But you learn, and you also learn how to put your agreements in place to account for that.

As a consequence of that, you try to de-risk. If we go back to the first collaboration we ever did in the companion diagnostics space, we were working with our partner and trying to put together what was going to be this companion diagnostic in Phase I, Phase II. And it was absolutely perfect: met all the regulatory requirements, all the bells and whistles. And then the drug didn’t make it. You know, the efficacy wasn’t there for thatparticular pathway. So we learned that when it’s early on, we don’t make that much of an investment into developing [an assay]. It’s a robust assay that meets regulatory requirements, but there's a different investment level.

"We are closer to ensuring success of the drug with a test, but we’re definitely not saving money in the development. We’re just decreasing the number of failures. So in a way you are saving money. But the development process is not inexpensive." – Cecilia Schott, AstraZeneca

Audience question: I just wanted to invite the panel to consider the example of Iressa [gefitinib], AstraZeneca’s drug, which was brought back onto the market with the advent of a test. Is there any work being done within their companies or in the industry where you might look at a marketed product that doesn’t have a companion diagnostic with the thought that perhaps it might benefit from precision medicine?

CS: Absolutely. I think Iressa is a good example. But the way I reflect on that is looking to different indications – you can have the same compound looking to different indications. And you can make a predictive diagnostic to use for that. I’m trying to think if there's one on the market right now. I don’t think I can come up with one. But the answer to your question is yes. And do we consider it? Yes. And that’s part of the life cycle management not only of the drug, but also of the diagnostic. So one of the things that we’re looking into is if you have a specific diagnostic that is looking to a single mutation, can you actually look into a panel of genes and say a specific mutation in one tissue is the same in another tissue? Can you do life cycle management of the test that would help predict the drug response in other areas? And the answer is yes, and we’re very open to it. The challenge is how much do you make in terms of an investment, and do you have the time to do it, and do you have the resources to do it. It’s not inexpensive to develop targeted therapies. We are closer to ensuring success of the drug with a test, but we’re definitely not saving money in the development. We’re just decreasing the number of failures. So in a way you are saving money. But the development process is not inexpensive.

CG: I fully agree with that. And I don’t have an example that comes to mind either, but I think it’s one of the things we look toward for the future, when we have volumes of data that we can tap into and start to understand why there's been a failure or where that failure has come from. A lot of the things are on the market now; we may not have the data packages to be able to go back and say what was right and what was wrong about the genetic sequence of a certain biology that could tell us, give us more information on why there was a failure. But as we go forward and keep collecting data, more and more often we are setting up a clinical plan where we are thinking about what data do we want to collect in the event that we might need the data later on for such an evaluation.

Ruth Phillips: Recently ASCO, the American Society of Clinical Oncologists, along with NCI [National Cancer Institute] and a number of large pharma companies have partnered to look at basket clinical trials where they're looking across tumor types and specific targeted mutations that they can then maybe collaborate and make that cost that you were talking about a little bit less for each of the pharma companies. But they're also looking at umbrella trials, where they take one particular tumor type and look across that tumor type to see if there are different mutations that maybe they can target in the future. So there are a lot of different, novel things going on in oncology to try and help defray some of that cost, and yet get the most patients possible on some treatment in that targeted therapy area.

Ruth Phillips, IMS Health

JH: One of our goals when we planned this roundtable was to offer guidance to smaller firms and start-ups on partnering with big pharma and larger, more established companies in biomarkers and precision medicine. What do drug companies look for in a partner? What are the hallmarks of a successful collaboration? And what are the realistic or unrealistic expectations of such collaborations?

CS: I have my usual answer for this, but I’m going to change it a little bit because it really depends what stage of the drug development program you're looking at. If you're looking at the Iressa example, you really need a partner that’s able to distribute. But if it’s a completely new area that you have only a hypothesis for, then you ask is there a technology that would help develop the test, and who has that? If it’s a very innovative technology, the biggest challenge that I see from a BD perspective, is what is the global footprint and access to that platform once the test hits the market?

One of the things that we’ve learned, early on I would say, is we want a global footprint and we want regulatory expertise. And we want commercial distribution. But we have learned by working with Myriad [Myriad Genetics Inc.], our partner on Lynparza (olaparib), that not having the regulatory experience is not really a block because you do that together. The other example is the circulating DNA that’s now part of the Iressa label in Europe. Again, a very different way. It’s life cycle management of the test. So it depends where you are in the process of the drug, if that’s what you're looking for. We have compounds in Phase I, and we can go out and explore different technologies. When you get into your Phase II, you want a partner that is committed – I think that’s probably the one word that I have that comes to mind – that is really committed to developing that test and bringing it to the market and to patients.

"There are a lot of different, novel things going on in oncology to try and help defray some of that [drug development] cost, and yet get the most patients possible on some treatment in that targeted therapy area." – Ruth Phillips, IMS Health

RP: From a data perspective, IMS has a collaboration with Foundation Medicine [Foundation Medicine Inc.]. And so we can actually take the biomarkers that Foundation Medicine has and link that information with patient data. We can look at the medical claims, we can look at the prescription, we can look at hospital data and EMR data and get a complete picture, so that as you're developing your clinical trials, as you're doing your research and development or maybe you're looking at a patient journey so you understand what happens to patients with different biomarkers, you have a better picture of what to take to a diagnostic company so that you can develop those products. So there are different levels of collaboration that need to be done.

JH: Another topic that we wanted to discuss was the use of next-generation sequencing and how that can end up with different drugs or drug sponsors involved in treatment. That can raise a lot of issues. Why don’t we start by discussing the issues around the drive for precision medicine and selecting patients for appropriate therapeutic regimens?

RP: From a clinical perspective, I think that’s challenging. When – and this is purely clinical, what I’m saying here – when a patient comes in and they’ve had a full sequencing of their genome, things show up, and we don’t know today what to do with them. Is there truly a target that we can modify and have a successful outcome for that patient? Do we even have a therapy that's available? We have some today, but not many. I’ll be honest with you. I think we discussed 18 targets, maybe. But there are a lot of mutations, and when patients come in, they want answers, and we don’t have those answers, unfortunately. So there's a long way to go still in that precision medicine.

JH: We were looking at an article in the September 8 issue of The Journal of the American Medical Association. A couple of academics who, to quickly summarize their story, basically were saying precision medicine has overpromised and under-delivered. And I think what you're saying is if you look at it right now, possibly, but that’s not the final answer.

RP: They started the Human Genome Project in 1989, Just think, it’s taken almost 20 years to get to today. How long will it take to get to that next point? When I was at ASCO in June, they did say that we know that there are about 228 mutations that are actually causing cancer, and they kind of weeded out those other genetic mutations that they don’t believe are truly causative of cancer. But now what do we do with those targets? What do we do with those mutations? How do we act upon them? And what do they really mean? We know they cause cancer but in what time frame? In which patients? What environmental factor do we have to have combined with that genetic mutation to actually have that cancer develop? So there are a lot of questions that are still out there. It’s not, OK, we have 228 mutations; let’s find targets for or products that target all of these mutations. I don’t think it’s quite that simple.

JH: Moving on with this topic, between next-generation sequencing, companion diagnostics, and off the shelf tests, what differences do these options present for biopharmaceutical companies?

CG: We covered this a little bit perhaps already when we talked about partnerships. And I think it depends a lot on where you are in the development process and what the needs are. We look for partnerships and we look for collaborations all along the value chain. So a companion diagnostic again is a very specific beast, if you will. And it has a very specific development pathway. And so there are perhaps limited options associated with getting that done in a certain time frame. But there are a lot of options for other types of technologies, depending on where you are in the development process. An early-stage project may require biomarker identification. It may require tools to help identify specific biomarkers, and those are in a class of different types of partnerships. And maybe for a particular clinical trial, there are lab developed tests that are going to enable a particular clinical trial design. So I think they run the gamut.

CS: I think that the other point is when pharmas ask how do you develop a drug, go through the regulatory process and commercialize it, you can step back and look into the academic centers, where biomarker identification has already been done for many years. Institutions like Mass General [[Massachusetts General Hospital]], MD Anderson [MD Anderson Cancer Center], Memorial Sloan-Kettering [Memorial Sloan Kettering Cancer Center], they all run these solutions. They have whole genome, they have huge databases. A lot of the innovation comes from those academic centers. So by the time that [a candidate] gets to us, it’s a much more narrow application, or more actionable application. We're asking, OK, what can you do with that information, how are you going to act on it? Whereas looking to the academic centers is more exploratory. So that’s at the other level of partnership that you're working on, with the academic centers identifying these libraries that you have of information.

JH: When next-generation sequencing yields multiple markers for one patient, suggesting various treatment options, how do you choose the right one? That's a particularly prevalent issue now in oncology.

RP: That's a good question. How do you pick the right marker or the right target to go after? And I think the first thing you have to think of is, if we have 18 products, and 228 potential mutations, we only have a limited amount of targeted therapies currently available. So we’re going to pick a therapy that we have a target for. So EGFR, ALK. And then if there are two, heaven help us, then we have to look at the patient and realize that each of these target therapies has its own side effects, its own adverse events. Patients can tolerate products differently. Is this IV, is it oral, is it something that they have access to financially as well as just in a geographic means? When you think of patients that live – I call it out in the country – and they have limited ability to get to, say, an academic center, they may not have the ability to receive a targeted therapy at the same level as somebody that is closer to an academic center, just because maybe the community facility doesn’t have the staff or the equipment needed to provide that targeted therapy. Plus, if you have to drive 100 miles to get to your closest community center, how often are you going to want to do that? So that’s when you have to look at the patient. And really, when you're making decisions about genetic mutations and treatment, that’s what it comes down to. Who is the patient, what do they look like, do you have a targeted therapy to begin with, and then go from there. So if there are two [potential therapies], you do the best you can.

JH: Next-generation sequencing can lead to significant amounts of data. We know that. But can this wealth of data be used to inform drug R&D?

RP: Absolutely. From an IMS perspective, I think that's something that we feel very strongly about. That’s why we pursued this collaboration with Foundation Medicine so that we can take the data that we do have and link it to the various patient assets that we have so that we can look at which biomarkers are of value, which length of therapy, duration of therapy; we can look at patient journeys, we can ask is there a target for a particular product, or maybe a product that’s in development that we can start to guide down that pathway? You can even look at physician targeting more from a marketing perspective. Are there physicians that are ordering more biomarkers? Our oncology medical research [group] has additional biomarkers, not just Foundation Medicine, so we can pull that data and say, "Oh, Dr. Smith is ordering biomarkers on every patient, so maybe that’s a doctor we want to approach for our particular targeted therapy, or maybe for a clinical trial collaboration because we know that he’s very involved in that area."

Audience question: Are we going down an avenue where you don’t really need regulatory approval for the drug or the diagnostic? Does anyone have any observations on where the field is going, if this sort of thing becomes more prevalent?

RP: I think we’re still going to need regulatory approval. It would be lovely, but I don’t know that we’ll get away with that. I think that’s one of the values that ASCO and NCI are trying to bring to the table, using these umbrella studies and these basket studies, so that they can go to the FDA or some other regulatory body and say, in these studies, we know that Xalkori [crizotinib] impacts patients regardless of their tumor type, as long as they have a Ross 1 tumor. And then maybe the label doesn’t say the product is specific to lung cancer that’s ALK positive or Ross 1 positive. But maybe it says it’s specific to the gene mutation. That’s kind of where I personally see things going. Our labels will be a bit broader and will cover more the mutation than the specific tumors. Otherwise, you're going to have a long list of tumor types that could potentially be used for that targeted therapy.

CS: Well, yes, from a pharmaceutical company perspective, but also from a patient perspective, really, the regulatory process is here to ensure that we have safety and efficacy. We want to make sure that whatever we’re taking and giving to our families, it’s safe. You still need to have the patient safety at the heart of the drug development process.

"If you don’t have that linked ability to look at the whole patient, it’s very easy to miss a part of the picture." – Ruth Phillips, IMS Health

Audience: How do you combine the molecular level of information with the clinical information, molecular level simulation plus overlaying the gene expression data sets? Usually, there is no real collaboration between these partners. How do you see this trend developing in these organizations?

RP: I definitely think it’s increasing. I think it’s a very important future for big pharma and little pharma and everybody in between. If you don’t have that linked ability to look at the whole patient, it’s very easy to miss a part of the picture. Actually, they brought me into IMS to focus on that specific initiative, and to make sure that we have all the pieces of the patient picture so that we can understand what that patient journey is, both prior to developing a cancer as well as after they start treatment. And do they respond well based on different genomics, or different treatments that they receive, and how long does that treatment last? And are there things that can be done at different points in a patient’s journey that can maybe keep a patient on treatment longer, because obviously if you're on a particular treatment, that benefits pharma for longer periods of time instead of maybe a short course, and then having to switch to another product. For me, understanding the whole patient is how we’re going to move forward in curing cancer.

JH: Our last topic is basket and umbrella trials. And specifically, how can basket and umbrella trials be used to help with partnering, finding a combination of drugs to serve a specific need?

RP: Well, I think you have to partner in those situations because if you look at the umbrella trials, where you're looking at one tumor type with multiple gene mutations, you're going to need different targeted therapies. And not one company is going to be able to have all of those targeted therapies. So you have to partner to be able to get that accomplished. How easy is it? I don’t know. I do think that there are some companies that maybe see the investment in partnering as a positive. But it’s challenging because again, there's that risk. What happens if somebody else’s product works better? There's lots of risk. And who’s willing to take that risk?

CS: I’ll give my perspective from AstraZeneca's point of view. It's a huge opportunity to actually share the cost of clinical trial development. It’s a good problem to have when you have a robust pipeline and you can put your compounds into a basket trial, which we have. It also helps you expand – there are just so many trials that we can do internally. There's only so much we can develop internally by ourselves. So it gives that opportunity to actually do combinations as you said.

We're going to see more and more of that because if you look into the immuno-oncology space, it’s going to be all about combinations. And the way to test that hypothesis is these kinds of trials. Otherwise, it becomes such a complex and investment intense process that you're not going to be able to leverage the wealth of information and the wealth of potential for a drug. With basket trials, pharma companies really do collaborate because it’s the way to get the drugs to the market.

CG: Yes, I think it’s one of the biggest trends of the future. It is clear that a lot of oncology therapy is going to require combinations of drugs. We all recognize that and working through how those partnerships need to evolve and how they need to work is definitely a big challenge, and is definitely something that’s really at the forefront of our thinking as well.

JH: For biopharmaceutical companies, what are the advantages or disadvantages of participating in basket or umbrella trials run by ASCO or the National Cancer Institute compared with the potential advantages or disadvantages of running those trials themselves?

CG: Gone are the days where you could, as a pharma company, invest in every clinical trial you could think of around a therapy. In oncology, the potential breadth and depth of a treatment is difficult to comprehend sometimes. And I think that when you have to make an investment in a lot of different types of cancers, a lot of different therapy areas, a lot of different stages of cancer, whatever clinical trials you can think of, the investment becomes really nearly impossible to make. Going forward, as we think about cost control measures and things associated with pharmaceutical development, then that just becomes a burden that's really, really difficult for any company to take on.

Editor's note: A video of the complete BioPharm America panel is available at "The Pink Sheet" DAILY.