Print BCTV: Cancer Diagnostics -- New Enterprise Associates, Genomic Health on diagnostic conundrum

Cancer Diagnostics

Transcript of BioCentury This Week TV Episode 120




Dr. David Parkinson, Venture partner at New Enterprise Associates, Menlo Park, Calif.

Kimberly Popovits, Chairman, President and CEO of Genomic Health Inc., Redwood City, Calif.



Centers for Medicare & Medicaid Services (CMS), Baltimore, Md.

Genomic Health Inc. (NASDAQ:GHDX), Redwood City, Calif.

New Enterprise Associates, Menlo Park, Calif.

National Institutes of Health (NIH), Bethesda, Md.

Nodality Inc., South San Francisco, Calif.

Oncotype DX



Steve Usdin, Senior Editor




STEVE USDIN: We're supposed to be in the age of personalized medicine. But are outmoded payment policies and regulations killing investor interest in advanced diagnostics? I'm Steve Usdin. Welcome to BioCentury This Week.


NARRATOR: Your trusted source for biotechnology information and analysis. BioCentury This Week.


STEVE USDIN: The path to new, better cancer therapies is clear. Precision targeting based on advanced diagnostics will get the right therapy to the right patients. But developers of 21st-century molecular diagnostics face two roadblocks.


Regulators like FDA are increasing their demands for evidence of effectiveness. And payers like Medicare treat genomic diagnostics as if they were the simple blood tests of the past. Can this conundrum be fixed? We'll hear today from a sophisticated investor and a cancer diagnostic CEO.


First, David Parkinson's perspective. He's fought cancer at the bedside, in the lab, in a CEO's chair, and now at the venture capital firm New Enterprise Associates.


Later, we'll get the ground-level view from Kim Popovits. She runs Genomic Health, a biotech that produces diagnostics for breast and colon cancer, and is preparing to launch a new test for prostate cancer.


But first, David Parkinson's perspective. Dr. Parkinson's a partner at the venture capital firm New Enterprise Associates. He was president and CEO of Nodality, a cancer diagnostics developer, for seven years. He's served in senior positions in large biotech companies, and at NIH's National Cancer Institute.


David Parkinson, I'm glad to be joined by you today.




STEVE USDIN: I'd like to start by asking you about the promise-- the potential-- of advanced diagnostics for cancer.


DAVID PARKINSON: Well, I think-- and I speak as somebody with a background in clinical oncology and drug development-- I think what brought me to this whole area is the realization of the limitations of current therapy, and the inefficiencies and the difficulties in developing new therapies.


So, with the advent of all of these new technologies-- DNA, RNA, protein, imaging, et cetera-- that we've seen, we now have the ability to characterize patients in ways that are much more accurate biologically, and perfectly suited to make much more targeted all of these new biological therapies which are being developed.


So, the promise is there. I've followed that promise in my own career-- moving from being a clinician, to a drug developer, to a molecular test developer-- trying to find that space linking the biology of the individual patient with the application of appropriate therapeutics. I had no idea that it would be as difficult as it is.


STEVE USDIN: And that's what I want to talk about-- the difficulty.




STEVE USDIN: So, in order to do anything-- well, probably in any endeavor, but certainly in biomedicine-- you have to have a clear regulatory path.




STEVE USDIN: You have to know what the rules of the road are.




STEVE USDIN: And you have to know that you're going to get paid-- and, hopefully, what you're going to get paid.




STEVE USDIN: Those are kind of the two preconditions. Where do we stand on those two preconditions when we're talking about advanced diagnostics for cancer?


DAVID PARKINSON: Well, I think when I went into the small company environment, with the technology that I actually thought could be used to accurately link patients with therapeutics, I actually thought the greatest challenges would be technological, and taking those technologies to a level that would meet regulatory scrutiny.


And those are, in fact-- it is a complex task. It is difficult to do all of that. But, it can be done. What's more difficult is trying to deal with the real world kind of issues that you just raised, which is, first of all, if these technologies are going to be used to match patients with therapeutics, then, first of all, they need to meet clinical quality standards.


Secondly, they need to be associated with levels of evidence that justify their use, and show the limitations of their use. And the FDA has certainly recognized this, and over the last series of guidances which have emerged in the last couple of years, have indicated that, first of all, they intend to regulate this area, which they have not traditionally done.


That, secondly, they are expecting levels of clinical evidence that start to look like those kind of levels that a drug developer has to top.


STEVE USDIN: So, on the one hand, maybe you could say it's reasonable to have the same level of evidence as a drug, because you're giving a test that's going to make life and death decisions. On the other hand, if you're going to do that, are diagnostic companies getting the returns that you need to be able to generate that evidence, and to have that level of uncertainty, which you haven't really had traditionally in diagnostics?


DAVID PARKINSON: Well, there's the rub. I agree with you that it is appropriate to regulate this area. If these tests are as important as we believe they can be, then they need to meet the same kind of standards that therapeutics do. What's missing is the reward system that exists for therapeutics.


And because of the absence of that reward system-- and here I'm talking about payment issues, about reimbursement issues. About the fact that we are now expected, in the diagnostics world, to generate levels of evidence that start to look like those of therapeutics.


But the reward systems-- current Medicare standards for payment-- relate more to time and materials, and don't really recognize the kind of value that's recognized on the therapeutic side, which is, listen. It takes time, and skills, and resources to develop these tests that accurately predict whether a drug's going to work or not.


And, in the absence of a payment system at the other end which recognizes this, it's not clear who's going to develop these tests, because the business model for small startup molecular diagnostic companies-- the kind of companies with the technology to produce these tests that we all believe, as a community, are important to the future of cancer medicine, as well as other types of medicine--


It's not clear how those companies will be sustained, and how those tests will be developed, and how they will be introduced into general clinical medicine.


STEVE USDIN: OK, thanks. We're going to continue our discussion of the opportunities and challenges posed by advanced cancer diagnostics in just a moment.


NARRATOR: BioCentury. Named the 2012 Commentator of the Year by the European Mediscience Awards, for excellence in communications and clear, concise commentary. You're watching BioCentury This Week.




STEVE USDIN: We'll be joined shortly by Kim Popovits, CEO, Genomic Health. Now, we're back with David Parkinson.


David, I want to propose a scenario to you. You've got a company, you've got a great new technology for some advanced cancer diagnostic. You're going to do the diagnostic tests in your own lab-- it's called a CLIA Certified Lab. Is there a clear regulatory pathway now for what you need to do to get this diagnostic on the market?


DAVID PARKINSON: Well, the path would depend on the perception by the regulators as to the importance of the test and the risk around the test. Low-risk tests, the path is quite clear-- that's, you produce it in your CLIA Lab, you can't advertise it, you can't make claims. There are no levels of evidence around it, necessarily. You could just commercialize it.


STEVE USDIN: And low risk means something where the consequences of the test being wrong aren't serious.


DAVID PARKINSON: That's correct. But what we're really talking about here today are the tests that we care about, which are tests that are going to be--


STEVE USDIN: High risk.


DAVID PARKINSON: Absolutely. The ones that are related to determining whether a therapeutic should be used or not.


STEVE USDIN: Life or death.


DAVID PARKINSON: Life or death, high risk-- "advanced class risk," to the FDA-- the expectation is that those tests require FDA approval. There have been relatively few tests examined and approved under this approach. There are new guidances which have emerged-- and others, I'm sure, will emerge in the next couple of years.


But not a lot of precedent. In general, the guidances that have emerged are quite reasonable. And particularly to someone who's been involved on the therapeutics development side, there are many parallels. So, again, we're talking about an approach which starts to look like therapeutics development.


STEVE USDIN: But it sounds like it's still uncertain. You can't say, this is what the pathway is, this is what you're going to have to do to get your diagnostic on the market.


DAVID PARKINSON: On the therapeutic side, we have the advantage of having many, many drugs gone through-- some successfully, some unsuccessfully. We've had the advantage of advisory committees, public discussion about what's appropriate and not appropriate.


This is a new area. The regulators are feeling their way through it. The test developers have never had to do this kind of thing before. So this is a new world for diagnostic companies.


STEVE USDIN: And are they prepared to do it? Are test developers-- do they have the expertise and the experience to be able to produce the kind of data that's going to be needed to get these tests on the market?


DAVID PARKINSON: Speaking generally, no. The diagnostics industry, in general-- from my perspective-- has been an industry where people bring a technology, use it to develop the analytical validity of tests--


STEVE USDIN: Analytical validity meaning, does it do what it says it's going to do?


DAVID PARKINSON: The actual performance of the test. The technical performance of the test. And they've kind of left the clinical evidence development to the clinical research community.


STEVE USDIN: Clinical evidence development meaning, does that result tell you something you can use to make a clinical decision?


DAVID PARKINSON: That's correct. Now, the world is changing. You actually can not, under these new guidances, commercialize a test that will be used in clinical decision making without developing that kind of evidence. So, this is the kind of world the diagnostic companies need to move into.


And the easiest way to do that would be in the setting of "companion" diagnostics-- a diagnostic company working in step with a therapeutics company. And there have been some recent successful examples of that happening.


STEVE USDIN: And just take a step back. A companion diagnostic is a diagnostic that's linked to a drug. You have to take that diagnostic in order to know whether you're eligible to take that drug or not.


DAVID PARKINSON: That's correct. I mean, that is really the future of therapeutics, to my mind. As drugs have become more and more biologically targeted, based on appropriate rationales related to tumor biology, the need to characterize patients in a parallel manner-- so that drug application to an appropriate biological context occurs-- is really important.


That's what we're talking about here. It's a new world. It has fantastic possibilities. But it comes with a lot of unknowns related to process, to determination and judgment of outcome, and to reimbursement of the process. And absent that, there aren't business models to sustain this.


STEVE USDIN: Well, there aren't business models for independent companies to do it. Obviously, though, every biotech company-- every pharma company that's developing an advanced cancer drug-- is trying to develop a companion diagnostic to go with it. So, what's wrong with that?


Maybe what happens is that-- and we've only got a few seconds left-- but maybe there isn't going to be a big diagnostics market, an industry like that. There'll just be companies developing companion diagnostics product by product.


DAVID PARKINSON: Well, Steve, I think you've put your finger on it. Everybody recognizes the importance of this. Everybody finds the problem with the standalone molecular diagnostic company. We're going to see some new business models. And I would predict that they will include tight integration of therapeutics and diagnostics development.


STEVE USDIN: Thanks very much. Now that we've discussed the landscape for new cancer diagnostics, we'll get a ground-level view from Kim Popovits, the CEO of Genomic Health. Here's a snapshot of tests her company provides to improve treatment decisions for cancer patients.




NARRATOR: Now, back to BioCentury This Week.


STEVE USDIN: We're joined now by Kim Popovits who leads Genomic Health. Kim, I want to start. There's a lot of talk in health care now about value. And there's different ways to perceive value. I think that the two kind of perspectives that are important are, on the one hand, patients. What's the value to patients? And the other, what's the value to society? In the United States context, it's usually, what's the value to payers?


How do you define the value, for example, of your breast cancer diagnostic? First, for patients. And second, for payers and society.


KIMBERLY POPOVITS: Well, I'll start with how we view value internally. So we think about value in terms of lives impacted by the work that we're doing. So today, when we speak to our own employees, value is determined by 300,000 women, families, patients that we've helped make critical treatment decisions. That said, when we go out to the world of payers, we realize that value has different definitions. And I think the most important one is, how are we affecting outcomes? So how can we, with the work that we're doing, demonstrate to the world, to the health care system, that our tests actually can increase value by delivering our resources more appropriately as we treat, in our case, specifically cancer?


And an example of that would be as the use of Oncotype DX, or the recurrent score for breast cancer and colon cancer patients increases, are we seeing the use of chemotherapy decrease? And so we monitor those things to make sure that we, in fact, can say-- we can quantify the value that we're delivering to the health care system.


STEVE USDIN: So maybe give us a few numbers. The idea of the test is that people who get it know whether their cancer is likely to be susceptible to chemotherapy. And if it has a certain level of range of scores, then they don't go on to get chemotherapy. If it does, they do. Have you any idea in terms of numbers or percentages of women who make those kinds of decisions based on your tests?


KIMBERLY POPOVITS: Yeah, so what's important to start with is that today we know-- and this isn't Genomic Health data, it's clinical data that's published-- both in breast cancer and in colon cancer, early stage cancer patients, so I want to emphasize early stage, most will not benefit from chemotherapy. So I think we all know somebody who's been through chemotherapy, touched by breast cancer or colon cancer, and what people often don't know is only 3 to 4 in 100 patients actually get any benefit from chemo.


The problem, prior to our approach and to our test, was that if you don't know who the 3 to 4 are in the 100, you feel compelled to offer all the option of chemotherapy in their treatment. People hear the word "cancer," they're motivated by fear and anxiety. They think they need to be aggressive. So unfortunately, many patients will take chemotherapy despite the fact that it has very little benefit.


So what our test set out to do is, can we get more specific about who those four patients are? And clearly we didn't get to the four, but we discovered that 50% of women can be very comfortable not getting chemotherapy. They have very low-risk disease. And their disease is not likely going to respond to chemotherapy.


Conversely, there's a group of patients-- 25%-- that have relatively high-risk disease. And the good news for those patients is they are going to get a big benefit from chemotherapy. So the test sends patients in both directions.


And what's most important for payers is that we've been able to demonstrate, through multiple studies that have been very consistent, that when a physician and patient have a recurrent score that they can include into their treatment planning, treatment decisions are changing on the order of 30% to 40% of the time, which is a huge difference.


STEVE USDIN: So I want to go forward now. You kind of broke new ground when you developed the breast cancer Oncotype DX. Now you're working on the prostate cancer test. The need for it is really obvious. There's an enormous need for it. But you're doing it in a different environment. You're doing it in a different regulatory environment. You're doing a different reimbursement environment. What is it that's different now? And how do you have confidence that there is a regulatory path forward? And at the end of the day, that you're going to get paid for it?


KIMBERLY POPOVITS: Well, there are some differences on both sides. Some very positive.


When I think back to when we launched the test in breast cancer, we were out trying to explain to the world what we were doing. We're using tumor tissue that's been banked, and we're getting RNA out of f pep block. And the whole world was like, is this really an appropriate way to do clinical studies?


So one big positive today is that the world has accepted that the diagnostic approach is going to be key in the future. And certainly, has to be at the core of personalized medicine. If we're truly going to individualize therapies, then you have to start with a diagnostic.


And what we know is that cancer is a complex disease. It's likely not going to be answered by one gene and one drug. It's going to be a multitude of things. And we're going to have to make sure that we correlate and translate all of this information well. And that's what we're now moving forward and doing in prostate cancer. So I would say the positive is the approach that we're using, very well accepted. There are great standards around it today.


And the negatives are probably still around the area of reimbursement, and the appropriate pathways for these products to you evolve through in terms of regulation. We've made progress, but there's still some uncertainty there. And that sort of uncertainty is a bit of a handicap for business models coming out today in the molecular diagnostic space because of the reimbursement issues that surround these tests.


STEVE USDIN: So we've got just a few more seconds left. But exactly what do you mean by uncertainty about reimbursement? That you don't know whether you'll be reimbursed, for example, by CMS? Or you don't know at what rates?


KIMBERLY POPOVITS: We're getting closer to understanding what the criteria are for reimbursement. In fact, I think Genomic Health has been a real leader here in working specifically with CMS to identify, what are the appropriate criteria for which these tests should be reviewed? So that's good.


The rate, I think, is probably less of an issue if the evidence is there and the steps that payers want to see are followed. I do think that we've made some progress in reimbursement. The problem is, is historically, what we've done with diagnostics has been to reimburse them based on activity.


STEVE USDIN: In developing your prostate cancer diagnostic, I'm wondering about one thing in particular. For prostate cancer, tumors are very heterogeneous. And there's a concern, if you take a single needle biopsy. That might not reflect the entire tumor. And you might get a false result. How do you deal with that? How do you overcome that in the diagnostic you're developing?


KIMBERLY POPOVITS: Yeah, that's a very important point and good question. We spent, actually, several years-- two years, minimum-- looking at that very issue. So one of the things that we knew is that we were going to be dealing-- the most important thing in developing a diagnostic for prostate was that we'd be able to use the needle biopsy tissue. If you have to use the prostatectomy, then you've done the surgery.


STEVE USDIN: It's already over. That means you've already removed the entire prostate. You don't need a diagnostic anymore.


KIMBERLY POPOVITS: That's right. So knowing approaches that are out there today, we wanted to go a different path and say, no, we've got to be able to work out of that needle biopsy. So our research team spent a couple of years looking at this issue of heterogeneity, making sure that we were accounting for that as we chose the genes and the pathways to put into our test. And that's the second part that's very important, and we saw it with breast cancer and colon cancer, is cancer is complex.


And so what we know, that established pathways and prostate cancer-- for example, the proliferation pathway, are established as being important in prostate cancer. Well, we agreed that is. But what we also found in our initial work, in looking at this heterogeneity issue and the biology of prostate cancer, that multiple pathways do a better job of looking at aggressiveness of disease than just looking at one pathway alone. So those two things combined, the heterogeneity issue being accounted for and looking at multiple pathways, we think is giving us an approach that's never been used before, to really get at-- which is most would argue, one of the biggest needs in medicine today. And that's helping these prostate cancer patients understand better how aggressive their disease is.


STEVE USDIN: And I wonder another thing, which is that tumors change over time. Are you going to be able to take a single shot in time and be confident that that's going to be accurate for the rest of that man's life? Or are people going to have to get more than one test?


KIMBERLY POPOVITS: Well, that's an interesting question. So the first step that we're looking at is, can we take the tumor today, and especially for those Gleason 6, 7 patients, which is the majority, and give them a better idea of the aggressiveness of their cancer?


Now, if that patient went into active surveillance or watchful waiting-- they use both terms-- and had some sort of recurrence or had another biopsy, the question around whether we would do that second biopsy is something that we will be studying. So perhaps that would happen in the future, that you would be doing the same test on the second biopsy if another biopsy was necessary down the road.


STEVE USDIN: We'll be right back with Kim Popovits.


NARRATOR: Now in its 20th year, visit for the most in-depth biotech news and analysis. And visit for exclusive free content.




STEVE USDIN: We're back for some final thoughts with Kim Popovits. Kim, you were talking about billing being based on materials and the steps that are needed to create diagnostic tests-- that's the way that CMS thinks of things. Now how would you like to see payers think about diagnostic test value and payment?


KIM POPOVITS: What we're really focused on today is value based reimbursement for diagnostics. We think we should have value based reimbursement for everything in health care, quite frankly. And so what we've talked about is demonstrating that value through the work that we're doing to demonstrate the overall impact our tests are having. So I could call that value based reimbursement.


And what happens today is that diagnostics are actually reimbursed by activity levels. And so there are specific codes one can use to get a diagnostic reimbursed. And the example I would use there would be much like going to a drug manufacturer and saying that we're going to reimburse you for your drug, based on the cost of your manufacturing steps.


And it's just something, historically, that's been done for labs. And it really doesn't fit today with the type of diagnostics that we're developing. So what I say is these new molecular high complexity diagnostics, like Oncotype DX, really need to be looked at through a different lens, and through a different system. So we're trying to change the conversation a bit to say let's look at value delivered.


STEVE USDIN: I want to ask you, maybe starting back where we started from also in my previous conversation with David Parkinson, is Genomic Health the exception that proves the rule? Does all this uncertainty about the regulatory system, about reimbursement, would it be possible for another company to do-- starting today-- something similar to what you've done to create an advanced molecular diagnostic company for cancer?


KIM POPOVITS: I would say Genomic Health proves the future. It proves the value of what these tests can deliver. And it also proves that it's difficult. You know, this is not something that is going to take a traditional approach. And so what I would say to companies today, moving into the space-- because there's such a huge need for this sort of analysis-- everybody you talk today will say, it's wonderful that we can sequence a genome for less than $1,000.


But if we can't turn that data into something clinically meaningful-- and that's the translation piece, and that's what these diagnostics are about-- so taking that data from the sequencing of a genome. Understanding biology better. Creating tests like Oncotype DX, that are based on highly sophisticated algorithms, and delivering it to clinical practice. That's the future. But it doesn't happen quickly. It takes many years and a significant investment.


I think what people need to know is that the runway you need is longer than what we may have expected eight years ago. But in fact, the rewards are huge, for physicians, payers, and patients. So we believe this is the way health care should go in the future. And a great in a great way to answer some of the significant questions that we have today that are a problem.


STEVE USDIN: Great, thank you very much. That's this week's show. I'd like to thank David Parkinson and Kim Popovits. And thank you for watching. I'm Steve Usdin. I'll see you next week.