Genomic Health's CEO Reviews Positive Clinical Validation and Development of the Oncotype DX Prostate Cancer Test (Transcript)

| About: Genomic Health, (GHDX)

Genomic Health, Inc. (NASDAQ:GHDX)

Review of Positive Clinical Validation and Development of the Oncotype DX Prostate Cancer Test Conference

May 08, 2013 4:00 pm ET

Executives

Kimberly J. Popovits - Chairman, Chief Executive Officer, President and Member of Non-Management Stock Option Committee

Steven Shak - Chief Medical Officer and Executive Vice President of Research & Development

G. Bradley Cole - Chief Operating Officer and Member of Non-Management Stock Option Committee

Emily Faucette

Mark Lee

Analysts

Rafael Tejada - BofA Merrill Lynch, Research Division

Kevin DeGeeter - Ladenburg Thalmann & Co. Inc., Research Division

David C. Clair - Piper Jaffray Companies, Research Division

Isaac Ro - Goldman Sachs Group Inc., Research Division

Vamil Divan - Crédit Suisse AG, Research Division

Doug Schenkel - Cowen and Company, LLC, Research Division

Daniel L. Leonard - Leerink Swann LLC, Research Division

George B. Zavoico - MLV & Co LLC, Research Division

Tycho W. Peterson - JP Morgan Chase & Co, Research Division

Operator

Good afternoon. My name is Roya, and I'll be your conference operator today. At this time, I would like to welcome everyone to Genomic Health Investor Conference Call from AUA. [Operator Instructions] I would now like to turn the call over to Kim Popovits, President and CEO of Genomic Health. You may begin your conference.

Kimberly J. Popovits

Good afternoon, everyone, and welcome. I know we have a good crowd here in the room. So thanks for the opportunity to again share some exciting data. This has, of course, been a milestone week for Genomic Health, for our collaborators and most importantly, for prostate cancer patients. So we're delighted to be able to have this extra time with all of you to go through the data again to answer questions.

I will mention upfront here with our Safe Harbor statements that we may be making some forward-looking statements here today. So keep that in mind and refer to our SEC filings for -- especially in regards to our risk factors around our business.

So let's get right into this. We have been joined today by various esteemed panel of guests, starting with Dr. Caroll who will be joining us on the phone to walk through some slides here followed by Dr. Cooperberg who spoke this morning. I hope many of you saw that and took copious notes and got every bit of the detail down in those 5.5 minutes, whatever they were. So you'll get an opportunity again to talk to Dr. Cooperberg and get some questions answered there. And then very importantly, Dr. Klein is joining us to talk about the development work that we did leading into this very important study. So I think you'll find that very interesting.

We will start though with the conclusions upfront since we know that we want to get the punchline out there really quickly for you, and Steve's going to walk through the patient report and just the positioning and some of our plans around launch. And I will just end my comments by saying we were delighted this morning around 8:00 to see the first sample be ordered online and to Genomic Health, quite a celebratory moment there as a company and look forward to helping many more patients in the coming weeks.

So with that, I'll turn the call over to the folks at UCSF, and we can get going. Thanks.

Peter Carroll

Well, thank you very much. If we can go to my next slide. So prostate cancer in 2013 is at a real crossroads. We have about 260,000 deaths worldwide. It's the sixth leading cause of cancer death in the world, actually the second-leading cause of cancer death in the U.S. Interestingly, over the last 2 decades, we've seen death rates decline in high-income North America, but increase in all other regions. Actually, it's increased over 41% worldwide.

What accounts for this is that in developed countries, with the introduction of PSA, this a simple blood test. And it's clear that screening of healthy, young, well-informed men with this blood test reduces significantly the risk of dying of prostate cancer, somewhere between 21% and 44%. And actually here in the U.S., we've seen a 39% reduction in mortality, and many people don't realize that this accounts for 20% of the overall reduction in cancer mortality in men.

If I could have my next slide. And this is a very good example. You can see cancer death rates in the U.S. lung and bronchial cancer declining as a result of smoking cessation. And you see these other remarkable declines in prostate cancer death rates starting in the early to mid-'90s at the time we reduced PSA. This is clearly due to early detection and improved treatment methods. But it's been a profound reduction, and that reduction continues.

Next slide. You go can go ahead and advance that. So further facts. The lifetime risk for U.S. men, men born in the U.S. each year, their lifetime risk of prostate cancer is 1 in 6. About 240,000 cases are diagnosed each year. It is the most common cancer. 31% will die each year, again, the second most common cause of cancer death. But if you look at the ratio between the number who die and the number of diagnosed, most men who get prostate cancer will not die of it. So despite the fact that it's a deadly disease for some, many more men will live with the disease than die of it. And the reason for this is using current method of diagnosis, again, PSA testing, we diagnosed many men with very small slow-growing cancers, which would never present a risk to them if left undiagnosed or untreated. This is a phenomenon called over-detection. It's not unique to prostate cancer, but it's probably most apparent in prostate cancer compared to some other malignancies.

Next slide. So the problem we are trying to solve is this one. Detection and treatment phenomenon called over-treatment are currently too tightly linked in the U.S. So in this country, if you get diagnosed with prostate cancer whatever risk you most uniformly treated. That is not the case in Europe where they do something called active surveillance, a little bit more common than here. So clearly, too many men undergo treatment for a disease, which would never become apparent if left untreated, again not unique to prostate cancer. Now there may be many reasons for this, but one, I think, very prominent reason is the worry about patients and their physicians that they harbor higher risk disease. By that I mean that they, in the biopsy, the PSA shows what appears to be lower risk disease, but they're fearful that they're harboring higher risk disease, which was not adequately diagnosed, kind of a tip of the iceberg phenomenon. And I think this is has been a major barrier, at least one of the barriers to active surveillance.

Next slide. So I think this is a good example here. Each one of those gray dots represents 1,000 men -- 1,000 U.S. men. The red dots hollow and filled in the lower-left corner represent the number of men who would die of prostate cancer if we didn't have screening. The hollow dots represent the number saved with screening. So, again, PSA reduces death rates. The problem is you diagnose many men with the disease. So in that large rectangle there, currently we'll diagnose about 90 to 120 men with PSA testing, and yet not all of those you can see here are at risk of dying of disease. And we would estimate probably that there's some controversy about this, but at least 20% to 40% of these patients, again diagnosed each year, would be diagnosed with what we call low-risk disease. Low grade, not very aggressive-looking, low volume confined to the prostate likely with a very low growth rate. So again we're diagnosing a lot patients to save lives, and this is a big problem in the United States.

Next slide. So this study was designed to address this issue, which is a really significant issue in my opinion in men's health. So the problem to solve was patients with low-risk prostate cancer concerned that they may have more advanced disease, which, if not treated immediately, would compromise their outcomes. Again, this fear that they harbor something more significant than what's been identified using current technology. So the solution, the hypothesis here and why we did this study was to answer the question, "Can gene expression analysis of the biopsy, which was done for diagnosis -- again, this is being done on the same biopsy we've done for diagnosis, can it predict the risk of more advanced disease better than current methods? Now the consequences if this is true is that more men would choose active surveillance in lieu of immediate treatment.

Next slide. If we could -- so I'm going to give you the conclusions here, but this will be delivered by Dr. Cooperberg and addressed by Dr. Klein. In patients with low/intermediate risk prostate cancer, the GPS has prospectively been validated as a biopsy-based predictor of adverse pathology. Again harboring more significant disease. Dr. Cooperberg will show this. It also adds independent predictive information beyond all standard clinical and pathologic data currently, currently available to patients, and again, he will show that.

And as will be addressed by Dr. Klein, I think, you will find out that this assesses the underlying biology from a very small amount of tumor. Again, so we can detect the signature in a very small amount of tissue, which is a major hurdle, and it does help address this whole issue of tumor heterogeneity and biopsy undersampling to more accurately predict overall disease aggressiveness. And, again, this will be reviewed by Drs. Cooperberg and Klein.

So in conclusion, incorporation with GPS enables more accurate identification of a larger population of patients who can more confidently choose active surveillance as an initial treatment strategy.

And I think it's time for Matt to go ahead and give the actual validation study.

Matthew Cooperberg

I'm going to be representing some of the information that was given this morning, hopefully at a slightly slower pace. But for those of you that were there this morning, this will be a bit of a repeat.

So this is our validation study design, which was, we looked at 395 men at UCSF who met our criteria for active surveillance but chose to go on to immediate surgery. So a couple of comments about the cohort first. This is a prospectively collected tissue repository at UCSF. So all men are sort of treated by the same protocol, and we banked their tissues prospectively in terms of diagnosis and then they are treated for clinical practice. There's a number of selection factors here. We weren't doing this in the context of a clinical trial. These are the real world patients, which we're seeing every day, referred -- mostly referred in from the outside actually.

And again this was done on needle biopsy tissues on a paraffin-embedded, needle biopsy is obtained again through routine biopsy, routine pathologic processing. The other point to make is about the endpoint. So we'd looked here at this end point of adverse pathology upgrading, upstaging at time of surgery. So remember anytime we do a prostate biopsy, we're sampling prostate even if do an extended core biopsy 12, 14 cores or more, there's always a known rate of undersampling, meaning that we take what we think is a low-grade, low-volume low-stage cancer, when we actually go to do the surgery, what we find is higher-grade and/or higher-stage, and it's known from multiple reports, multiple centers that this rate is typically around 25% to 35%. And that fear, the fear that the biopsy of my prostate, it looks low-grade, but what if there's something worse that you missed. That is a clinically very relevant end point, which is one of the reasons we thought that if the decision is about surveillance versus treatment, then it's a very clinically meaningful end point. It's also an available end point. Mortality study, which Dr. Klein will talk about later, we need much, much longer follow-up on these patients.

These really are contemporary patients treated in the last 2 years at UCSF. So we started out with 514 patients who met clinical criteria, and we intentionally chose a relatively broad definition of active surveillance for 2 reasons -- I'm sorry, broad definition of inclusion here, which included low risk and men who are at the lower end of the intermediate risk range.

2 reasons for that. First is that philosophically we believe that UCSF has a criteria for surveillance really should be broader than the most other centers, and that many men who have, for example, low volume, pattern 3+4 or higher volume 3+3, actually often do quite well with the surveillance. And the other reason is that you want to prove the validity of this test, you do need some biologic diversity.

So of these patients, we had adequate tissue for review and about 85% of them. Another test requires 1 millimeter of tumor, which is a tiny, tiny volume, but there were an handful of cases that had even less than that. So they were excluded. We had 412 patients who met the full set of clinical and pathologic criteria. And the first point to make is that the excess of the tests in terms of at least giving us a genetic readout about a good quality was extremely good considering how small these samples are, let's not forget the lower risk of the tumor, the less tumor there typically is available. And this have been one of the major hurdles, which has prevented markers from actually seeing successfully validated in this setting before.

So 96% of the biopsies yielded a successful valid score. And if we've got at least 10 nanograms of RNA out, which corresponds to about 2 millimeters worth of cancer, say, it was close to 100. So from that pure sort of analytic validity of the test ratio of reliable success.

So this is the score, and Dr. Klein is going to go over the derivation of these genes. But this is the previously described 17-gene assay, the GPS score, which is based on genes in the androgen signaling, stromal response, proliferation and cellular organization pathways, which are weighted. Each of you -- you get a domain score for each of the scores. Those are then weighted and normalized to a set of reference genes, which are basically associated with metabolic housekeeping or not expected to vary to anything that's going to take a stand in the presence of cancer or not.

So these are multiple different pathways, which are all associated with cancer, and again these were drawn from a large candidate set of genes, which you'll hear a little bit more about later, and these were specifically optimized for genes that first of all, work in setting of heterogeneity and prostate cancers would sometimes have a mix of low-grade and high-grade disease and again, genes which can be identified in very small tissue volumes.

This is a brief look at our cohorts. Again, this is not a strictly low risk cohort. We do include some men with low volume 3+4 and some men with what some would consider to be a higher stage, and for the reasons that I gave above. But for the most part, this is a low or the low-end of intermediate risk. We don't have many men who are kind of at the higher end of what we would consider intermediate risk. For example, nobody with a 4+3 of diagnosis, nobody with a high volume of 3+4, and this is the rates of upgrading upstaging. So 18% had high-grade disease and 22% pT3, 31% either of those 2 endpoints. This is a little bit higher than a typical upgrading upstaging report because we have broader criteria for surveillance than you would see in a typical low-risk upgrading upstaging report.

So the first question is are we are going to get an independent signaling, and one of the reasons -- and another of the reasons for the biomarker research has really not been successful over the last several years even most of the studies are not set up properly. Specifically how the studies are designed, that you've got a pathologic test, which predicts the Gleason Score. But we already know Gleason score, so it's not particularly helpful. So the question is if it's going to tell something new, above and beyond what we already know.

So we represent what we already know with the CAPRA Score. This is the score that we developed at UCSF in 2005. It's been validated all over the world and about 35,000 prostrate cancer patients treated with surgery with radiation therapy, it predicts recurrence and also long-term disease mortality. This represents the PSA, the Gleason score, primary and secondary, the stage, of course, positive and patient's age. And it's as good as the best nomograms or better in terms of its accuracy. So we use this as our base model. It's not the only gold standard, but it is a gold standard in terms of what can be done with the clinical information.

And it's fairly accurate, and one of the issues here is that the bar is actually pretty high to prove that we can do better than what we can already do today. So the first question is would we see -how tight of a correlation that we'd see between clinical risk and the genetic scores. And if you imagine trying a regression line through these dots here, these are the GPS scores at each separate level. So lowest risk, the highest risk. So there is a difference. I mean, on an average, these genetic scores are slightly higher and higher levels of clinical and pathologic risk, but there's a tremendous amount of spread as well.

So you're picking up patients at each of these CAPRA score levels who have lower, middle and higher GPS scores. This is important. Have we seen too much clustering? It would have been an early sign that we are relatively unlikely to find an independent signal. It's all about what can we do above and beyond clinical information. It's supposed to be an independent prognostic value.

So this was set up as a multinomial regression. I showed you the little 6 cell design this morning. So basically, you have a mix of possible outcomes, upgrading and/or upstaging. And those are independently important. It was set up as a multinomial regression where you can look at all those combination of outcomes against remaining completely or confined and low grade. And the prespecified analysis was to look at this multinomial regression approach of the GPS score initially just adjusting for the Gleason score because that's a kind of our representation of pathologic risk. And in fact, this is kind of the first no go -- go/no go analysis. And in fact, the GPS was statistically significantly predictive controlling for the Gleason score. Now, this is not the end of the story. Kind of the score itself is not interpretable but this at least was kind of the prespecified endpoint.

This is much, much more important. Okay. So this is actually looking at the score now, controlling for -- what I said before, controlling for everything that we already know about the clinical and pathological risk. And we did this 3 ways. So we've adjusted the CAPRA Score, we looked to the NCCN risk groups as well and we built a new model kind of fitting in all these individual variables de novo. There's always a risk here that you can kind of overfit a model, but we wanted to make sure that we weren't misrepresenting this information by focusing too tightly on the CAPRA Score.

So adjusting for the CAPRA Scores, such as -- in other word, PSA and all the pathologic variables, there's a twofold increase in the risk of adverse pathology high-grade disease and/or high-stage disease every 20 point increase in GPS. So GPS remember is a 0 to 100 score. The 20 number is totally arbitrary from a statistical standpoint.

Okay, the statistical significance, which was very good, it's going to be the same no matter what your per unit increases. This just give you something that's clinically interpretable with respect to an odds ratio. Okay. So again, this is about a twofold increase with every 20 point increase in score, you could even have done this at the 10 unit increase and we'd have come up with 60% increase in risk with every 10 unit increase in score. And this was chosen because it was the median EPS score for the highest and lowest court house. The patients, it's a convenient way to do it, but recall that this is arbitrary. There's no magic to the 20-unit decision.

And this is, I think, a really nice way of sort of thinking about how this will actually work clinically -- how this could actually work clinically. So this is again, the likelihood of favorable pathologies. Again, I've been talking so far about predicting adverse pathology. This is the inverse now. This is saying what's the likelihood of not having adverse pathology. So having a tumor, which truly is organ-defined, truly is low grade.

So, again, the clinical risk alone is actually not bad in terms of doing these predictions. Using CAPRA, you can see pretty fairly spread -- I'm sorry, fairly widespread in terms of the likelihood of favorable pathology. If you now incorporate the GPS together with the CAPRA score at each level of clinical risk, you can really spread out these predictions quite substantially both up and down, okay. So for any one baseline clinical risk, you can find patients who are at lower risk or at higher risk. Now this is the distribution of the patients across these various CAPRA Scores.

So what this tells us is, for example, almost 50% of the patients had at least a 5% shift up or down. There's a mix of those went up and down. This is an important point okay? Because we think that there's patients who are baseline at low risk, some of them are going to be up shift, and this is equally important, okay? So when we think we know men who we'd know their at low risk, if we can tell them really in a really low risk, it's great. They can make more a confident decision, but it's also equally important that we identify the man who frankly should not be on active surveillance because they're out there. I mean, we always say the number of men who will die because they want an active surveillance instead of getting treatment is not 0. We think that it's very, very low, but it's not 0. And that fear of potentially being in that category is what drives a lot of men away from active surveillance, not frankly being to tell them it wasn't you, you're not one of the patients move down this curve, we think will give a lot of reassurance and conversely identifying the rare patients who really do have a significant shift, well, that should be very important to their clinical care.

And that's just another kind of example way of thinking about this, if we, again, arbitrarily set a bar of an 80% prediction of favorable pathology. If we did this with clinical information alone, you have a cap of 0 so we only get 5% of the patients whom we can get that information whereas if they combine the clinical and genetic information, this goes up to nearly 25%.

We can do the exact same thing with the NCCN risk groups. Again, you can shift these curves up and down for the patients who are very low -- low end to intermediate risk. And, again, increasing the proportion of patients now who can be identified as at least 80% likely to be a favorable from 10 to 26. And there's a significant number of patients who shift to these categories.

So in summary, I'd like to share -- I've already kind of read this, I won't read this to you again, but I think the important points to remember is that this really is, first of all, the first test that works in biopsy tissue. And this really is truly an independent validation of the score, okay? This is something new under the sun in prostate cancer, the biopsy test that contributes statistically significant and clinically highly meaningful information above and beyond everything that we know today. And it's frankly the first one that we've been able to do that, and we really do hope that this will be a piece of the puzzle in solving the screening problem. Dr. Caroll alluded to this already. One of the reasons that there's been so much controversy in the last few weeks over these screening guidelines, it's not so much that the screening doesn't work, it's the measure, assessing the harms and benefits of screening. And we can drastically reduce the harms of screening if we reduce the harms of overtreatment. Of course, the key to doing that is identifying a better precision which men do and do not need treatment and improve their confidence in their decisions to treat or not to treat. So we really do think that this just maybe an important step in the right direction.

Eric Klein

All right. Let me focus on clinical context for you. As a clinician when we have a patient who's newly diagnosed with prostate cancer, we only have a certain number of parameters, currently clinical parameters that help us arrive at the right decision for the patient, if the patient treatment or no treatment. We have the patient's age and general medical status, and if we have grade stage and PSA. And we have a little bit of biopsy material that a pathologist who was well trained but is subject to the subjectivity of reading something that's based on training and tell us how aggressive the cancer is.

What we don't have with biopsy is a complete sampling of all the tumor in the prostate, and I'll show you why that's the problem. So the goal here was to take the 1 piece of unexploited information we have at diagnosis, that's the biopsy and extract additional information out of it beyond the grade that the clinician -- that the pathologists assigns. That's really what we're trying to do here.

So there are technical challenges in doing that, and there are biologic challenges to doing that. The technical challenge is that this is the size of a biopsy core of a prostate, and generally we do 12 cores at diagnosis, and so we have 12 of these little things that might be 1 centimeter long or 2 centimeters long at most, and a few millimeters wide. That's not much tissue to work with. I won't go through the technical details on part because I don't know them all, but the Genomic Health people can tell you that. There were a number of technical validation studies that showed us that on a single millimeter of tissue from one slide, we can get a good biological signal that captures what we think what the true biology of the whole prostate is. That's the key breakthrough here.

Now here's the biological challenge. We know -- this is just 1/4 of one prostate that's been removed by radical prostatectomy. We know that if you do 12 biopsies in various areas in the prostate, only one of those 12 cores might show cancer. The reason for that is that it doesn't sample everyplace. It's not possible to sample the whole prostate. That way we do not have good imaging technologies that can see prostate cancer. It's usually ISO intense on ultrasound although there's some interest in MR now or even on MR on many of these tumors that would be eligible for surveillance are invisible.

We also know that if you take out a prostate even if only one biopsy showed cancer, on average, there are 6 or 7 little cancers scattered throughout the prostate, and we know that there is heterogeneity in the biology of those tumors, and that's what's illustrated here. So what you see again is just 1/4 of a prostate that's been removed. The pathologists using a magic marker has outlined 4 separate areas of tumor here. There's a red area here that is normal tissue. And that doesn't apply to what we're talking about today.

But each of these areas may have a different biology as measured by something called the Gleason pattern. That's the grading system that we use, and the primary Gleason patterns that we see in patients who are candidates for surveillance are pattern 3 or pattern 4. It's generally believed that pattern 3 tumor is relatively indolent, and that pattern 4 tumor is the one that can be a killer although it's very clear that not all pattern 4 tumor is lethal.

So when a needle strikes the prostate, and if it only hits one area, one of these tumors, let's say the low-grade tumor. As Matt pointed out, we don't know when you're looking at a patient and looking at the pathology report whether or not we have missed more biologically significant. That's the biologic challenge that we had to address here, and I'll tell you how we did that.

So this was the series of studies that we did, and we started out, as I mentioned, with the technical feasibility studies to demonstrate that we could get a good signal, that we've got high quality RNA out of these samples so we can get a good signal with low volume tumor and so forth. And I think we did 2 or 3 of those before moving onto what we call the gene discovery study. What we did here was to take using a cohort sampling design, 2,600 men who had their prostate removed and importantly, for whom we had up to 20-year clinical follow-up. We knew whether these men got metastatic cancer. We knew whether they lived or died and whether or not they died of their prostate cancer.

So the first question we asked was a simple one that we thought we knew the answer to, and that was are their genes in the tumors of these radical prostatectomy specimens that can tell us who got metastatic disease and died. And the answer to that was yes, and so we started with a little over 700 genes that we picked in part based on what's published in the literature about what we think is important in prostate cancer, and part based on some observations that Genomic Health had from some other studies and had seen some genes that were important in other cancers. And we went ahead and did that prostatectomy, and we did, in fact, show that there are genes that can predict metastasis or death.

Once we did that, we selected out the best-performing genes to do a second study, which was to see if we could answer the same question, are there genes that predict outcome on biopsy material, those little slivers of tissue there, part of that this was technical feasibility, but part of that was an important biological question, again, because of this heterogeneity here. And the way we assess that was to do this. We had our pathologist micro dissect a low-grade tumor and throw that in a tube and look at the gene expression in that area of low-grade tumor separate from the areas of high-grade tumor. And that's the key biological design, experimental design that distinguishes GPS from other tests that are being developed and that are on the market already.

We're able to assess or deal with this issue of heterogeneity, and the reason is that we've discovered, truly a discovery, that there's a subset of genes that can predict the outcome whether they're measured in the low grade or the high grade. Now there are genes that are independently expressed in low grade and high grade. They each have their own biology. But there's a commonality to it. We can refer to it as a common underlying biology between these 2 high-grade tumors, and so that allows us to have the confidence that even if the needle only strikes the low-grade cancer that we have captured the whole biologic potential of the prostate.

That's a key issue and talking around the country about this and here at the AUA and so forth, I'm not sure that everybody who has heard this has really understood that this is the key thing that differentiates this test from others. It's an important point.

So we looked at 81 genes in the biopsy study. And the end points of the biopsy study were precursor to the UCSF study. It was the same design. It was meant potentially candidates for surveillance who had biopsies and who also had their prostates removed at the clinic, and we designed the experiment to see if gene expression in the biopsies could predict adverse pathology, and the answer was yes. And then so we refined what genes were best. If you think of this like a funnel, we poured 727 genes in the top, and we got 17 for the most important of genes at the bottom, and that's what forms the basis of the test.

So this is how the prostatectomy study was done, the specimens were rereviewed for assignment of stage and grade. Again, the key issue of manual microdissection of low- and high-grade tumor then the RNA was extracted, and then the data analysis that we saw.

And what we identified initially were that there are 6 gene families that have genes in them, 6 different biologic pathways that have genes in them that can predict the outcome. And we found that of these 17 genes that we picked, they fell into these 4 families over here. So this is called a [indiscernible]. Are you used to looking at these? Yes, okay. So a gene that is over -- that is not overexpressed in the tumor or is overexpressed and has no impact on the outcome fall the dot for the hazard ratio, would fall right on that black line.

What you can see here, genes in the stromal group and the proliferation group on overexpressed connote a higher risk of metastasis or death. Genes in the androgen group and the cellular organization group if they're overexpressed for a lower risk of metastasis or death. It turns out that the same genes that can predict metastasis or death also predict adverse pathology. So we used that sort of.

Those were the important findings here, and so Matt's already shown you the answer of the validation study. All of these development studies allowed the development of an optimized technology to analyze the needle biopsies. The small amount of tumor we get there, we identified genes that predict disease aggressiveness and address the issue of tumor heterogeneity, a key point. Reestablish that multiple biologic pathways are highly significant, and we studied the relevant patient population.

And I mentioned all of this was developed to deal with the patient who's sitting in our office and will be diagnosed prostate cancer before he has any other treatment, and those were all the challenges that we received and the bearded baldheaded guy, and that's the -- it's like looking in the mirror. That's the guy who's typically in the office, and we put that in there just to emphasize.

In the end, this is all driven by patients who have to make this very difficult decision. They feel like it's a life-and-death decision about whether or not they should be treated or not. And as Matt said, this will reclassify patients into lower risk who are really suitable candidates for surveillance. Does that mean that every patient who fits that will fit surveillance? No, they won't because there's just too much anxiety for some patients.

It will also reclassify some patients into higher risk who are not candidates for surveillance but we think this will give them more confidence that they've made the right choice. Whatever side effects they have to put up with are now worth it because they don't have this doubt after the fact that, gee, now I know I have some incontinence or some erectile dysfunction, and I'm not sure that my cancer needed to be treated. We think it will improve the confidence of our decision-making. And that's how I see it as a major step forward in our ability to practice what's called precision medicine, the idea that you make a clinical decision on treatment or no treatment or what kind of treatment based on some unique biologic characteristic of the patient or their tumor. But this is really step forward that decision.

Steve, I'll turn over to you.

Steven Shak

Okay, thank you. So the work you've heard today are built on the experience we've had of delivering now our breast and colon assay. Again, over -- I think the number now is 350,000 patients we've served in more than 75 countries, and what I'm going to highlight in terms of now, the application of this test into clinical practice. There are aspects of these that clearly for those of you who've been following the story of the work we've done that are quite similar to what we've done in breast and colon. I think actually some of those points of similarity we just highlighted, genes matter, methods matter and answering the critical question, notice that the critical question here for prostate cancer patients was all around, is there the harboring of adverse pathology that I lack today insight into that this test can help me see?

There are also some aspects of the test and certainly the report you'll see in a second is different than the report that we've provided for breast and colon cancer. And again what's different here is that for the urologist, this all needs, to ultimately as outlined by our experts, needs to be translated into do I have the confidence in the fact that the clinical criteria, as well as the GPS are in the range that would mean that I can do active surveillance or is there evidence based on a combination of the clinical and the GPS that indeed it looks like this might be more aggressive. And as Dr. Klein just pointed out, that, again, the right thing for me is immediate treatment. And so this is what the Oncotype DX prostate cancer report looks like. Pretty small here on the screen. So I'm going to quickly blow up the different pieces of it so that you can see it.

With regard to the top, again, that typical patient information that we are provided, we get provided back to the ordering physician. With regard to the GPS result and the clinical experience, again the scores getting here this is an example score to go from 0 to 100. And again this describes the clinical experience in the UCSF validations studies that will allow them patient the result.

In this particular example, we actually have the AS [ph] on the requisition that the clinical and pathologic NCCN classification get provided on the record . Again, I said that what we will be interpreting is the combination as was performed by Dr. Caroll and Dr. Cooperberg in the UCSF study, their analysis combined, right, the NCCN score or the CAPRA Score on those beautiful [indiscernible] plots and the actual GPS score that was on the factor. So in this case, this is a patient with clinical NCCM low risk that would place them in this area.

And in terms then of the expected range of where they would start from based just on the clinical and pathologic criteria, it would be in this range. What, again, they want to know is would the GPS score now lead them to be confident that there's not a surprise that if the prostate was taken out that they would stay with not high grade and not spread beyond the organs or would it move in this direction? In this particular case with a score of 8. It does give them confidence that indeed that they would stay there at a relatively low risk.

Notice that the percentage of patients that are very low to start with is about 10%. The percentage of patients like this particular patient that are low are about 50%, and the percentage of patients that are here at the outset are about 40%. We know that in terms of this whole group, the likelihood of mortality is quite low. But right now, in many cases although in Dr. Carroll's practice many of these patients are considered for active surveillance with confidence. In other practices, they might just be in this range. And so this is as described by Dr. Cooperberg increase the proportion of patients who can be confidently there from around 5% to 10% to potentially up to 24%.

Notice though that in this particular case, you can guess the GPS. So some of the patients actually might have moved in this direction even though they started at low risk. And that's for this particular patient. The report is actually encouraging. You could confidently choose active surveillance. And thus, you can see that the report will provide for each physician and patient that combination of their NCCN clinical risk, Very Low, Low, and Intermediate and where they start and then the GPS continuous biologic risk, which is determined by the 17-gene assay that performs on the very tiny amounts of [indiscernible] for the Genomic Health lab.

Question-and-Answer Session

Kimberly J. Popovits

Room, let's start.

Rafael Tejada - BofA Merrill Lynch, Research Division

It's Rafael Tejada from Bank of America Merrill Lynch. Just wanted to see if you could provide us with any initial feedback that you're hearing regarding the clinical validation study from individuals and doctors at the conference today, and I have a couple of other follow-ups on the development of the study.

Kimberly J. Popovits

Thanks. So I would say the overall feedback that, I think, most of them, I've heard is quite positive but I'll ask, they're both in the panel. Maybe starting with Dr. Cooperberg.

Matthew Cooperberg

I would agree. I mean, I think the new discoveries have been very favorable. There's been a lot of Twitter activity about the presentation, and I think the feeling has been quite exciting, actually. It's hard to quantify that, obviously.

Kimberly J. Popovits

And I should -- let me also introduce a couple of folks that have joined us up here to answer questions and for those on the phone. So Dean Schorno, our Chief Financial Officer; Brad Cole, our Chief Operating Officer; Kathy Hibbs, who's our General Counsel and Head of Regulatory; and Mark Lee, who actually have led the development and the program of the prostate program at the Genomic Health. So, Mark, with that, maybe, you've probably gotten a lot of feedback this week. You want to comment on what you're hearing?

Unknown Executive

Definitely much appreciated. I think the sign of the validation study is focused on temporary [indiscernible]. On this specific clinical question, grade [ph] is relevant. I think those have been highlighted, and I appreciate it.

Kimberly J. Popovits

I'd add to the short time that I've been here to Dr. Klein's presentation, a lot of questions around how is this assay differentiated and what was it about this development that makes it different and really unique today. And clearly, the discussion around the -- addressing the issue of heterogeneity and undersampling is, I think, a very important topic of conversation.

Rafael Tejada - BofA Merrill Lynch, Research Division

Just on the development of [indiscernible]. Another competitor is also trying to develop a multi-gene test to answer a similar question. And looking at the genes that were selected, the final 17 genes that were selected for the assay, in fact, if I recall, the test is mainly looking at proliferation. So I was wondering if you could just speak to whether you were surprised or not surprised by the types of genes that were selected and kind of -- to take aggressiveness with prostate cancer.

Eric Klein

Overall, not surprised when we saw that. Let me add some detail here. We selected those 700 genes with some specific reason behind them. But after we selected those 700 genes, we were completely agnostic as to what the best predictors would be. So we let the genes do their thing and picked out the best performance. And when we looked at them, cell proliferation is important, the other test that you're talking about. But it's -- clearly, it's not the only pathway. So I think one of the strengths of this assay is that it surveys the biology of the cancer better because it surveys other pathways. And for a gene that only looks -- for a test that only looks at one pathway, if it's worrisome, then, of course, you should pay attention to that and be treated. But if it's not a worrisome score, you don't know if you've missed some other biology. Furthermore, the other test on the market did nothing to assess this issue of heterogeneity. So if again, that score is low, you don't know if you have missed the other underlying biology. And then Matt mentioned this, if we look at the gene, all of those genes are tightly known -- tightly linked to cancer -- most of them are linked to prostate cancer biopsy [indiscernible].

Kimberly J. Popovits

The other thing [indiscernible] said is that we've not seen this grade in 3 cancers. So in breast cancer and in colon cancer, the same issue occurred, multiple pathways were stronger than any single pathway. So third time that we've seen these different pathways across the 3 cancers, but important to look at multiple pathways. Did you have a...?

Rafael Tejada - BofA Merrill Lynch, Research Division

And just one last one. Just on commercialization for the test. Just wondering with the experience that you have had in marketing on [indiscernible] types for breast cancer, just wonder if you could just speak about what's going to be different as you market this test to urologists, what you're expecting?

Kimberly J. Popovits

Well, you can imagine we're going to leverage our experience in the success that we've had in the breast cancer arena. This clearly, though, is a different audience. So it's our first time in addressing oncologists. We've hired a small sales force, we have a great marketing team, we've hired a number of folks in our medical team and we have medical science liaisons that will be out there working to educate key opinion leaders with a similar approach to what we did in breast. So what we want to do is we want to go out and we want to address the high volume practices and the key opinion leaders to really kind of get this going, and we think we can do that fairly quickly. And then we'll begin to add to the sales organization as time goes on as we see traction taking place and as we move more towards payer reimbursement. Brad, any other comments?

G. Bradley Cole

I think that one of the things -- those launches that we're going to be in front of and have information available for patients in a way that we didn't have for our breast cancer patients is the early days, but this is well understood. And cancer information treatment versus available on the Internet, we want to be there with advanced [indiscernible] results.

Kimberly J. Popovits

So we launched our website today, which is great. You should all take a look at it. And the other thing that I would say that's a bit different from both breast and colon cancer is that we have recognized the immediate need for decision impact work. So we will be starting a decision impact study with prostate cancer immediately, and we think that will be information, as it was in breast and colon, that's going to be very important for payers. Any questions?

Kevin DeGeeter - Ladenburg Thalmann & Co. Inc., Research Division

Kevin DeGeeter from Ladenburg. I was hoping you could maybe clearly define for us the -- again, with prostate, for example, with breast cancer, the end game was respond, chemos, and a number of companies put forward various options to get there, guidelines speak for themselves. Genomic Health came up with the best test to get to that end game. What is the comparable end game with respect to prostate, recognizing this is a very exciting first or [indiscernible] in the prostate?

Kimberly J. Popovits

Steve, do you want to...?

Steven Shak

Yes. So clearly, this is a good patients in the study that we did, the validation study that it was already selected to be low risk or low intermediate risk based on abundant literature that UCSF and many other groups have highlighted, the long-term outcomes of that group of patients is generally favorable. And so the focus in this case is how to treat the right patients aggressively early and how to do active surveillance. And so the first step that we've reported today with the endpoint of adverse pathology does determine whether or not a patient should get active surveillance by looking at the combination now of clinical factors that determine risk, pathologic factors such as Gleason and now the GPS score. And I don't know if I would call this the end game, but the next stage of studies, and others here can comment on it, are going to be much -- also now going to be focused on how to do active surveillance with what intensity and whether to do it the same way in all patients who are active surveillance because it's very likely that we need precision medicine with regard to how we do surveillance.

Matthew Cooperberg

[indiscernible] that we'll -- actual studies that we'll do, but it's well acknowledged that active surveillance as we practice it today is not perfect, And with the serial biopsies that are required are associated with substantial risk of infection, [indiscernible] percent in time. Not huge, but you multiply that over the number of biopsies that needs to be done and it becomes a real risk and, of course, a real discomfort, inconvenience, psychological burden to the patient, et cetera. We know that there's a large reservoir of prostate cancers that are so indolent that we shouldn't call it cancer, but we do, we probably can follow them less intensely. And the idea -- and again, this is really overthinking, but the idea would be to have the test, put the information together, if everything lines up low risk, you don't need a biopsy. Something like that and obviously those will be done.

Steven Shak

There's another way it can impact how we do surveillance. One is eliminating biopsies. It's my own practice, before I consider a candidate for an active surveillance, I do 2 biopsies because the first biopsy misses the 20% of patients who have adverse pathology, so second biopsy. Well, I'm going to eliminate using this test and eliminate that second biopsy. And the second thing is that some men progress and get more aggressive tumors while on surveillance. We currently have no way of identifying that because we -- that's based on biopsy criteria. So if someone goes from a little low-grade cancer on initial biopsy, and a year later, they have a little 3 plus 4. We don't really know if that's undersampling, if a different pathologist read that biopsy or if it represents true progression. So we don't have any data on this and haven't done this study, but you could apply this same or similar test for that repeat biopsy and see if there's been a change in the biology and then intervene appropriately. And as Matt said, it may get to the point where you've done that once or twice on an older patient where you don't need to ever do anything to that patient again. You can give a conference call and don't worry about it.

Kevin DeGeeter - Ladenburg Thalmann & Co. Inc., Research Division

That's very helpful. And maybe sort of taking that question from a slightly different angle. You envision there being stages [indiscernible] NCCN Guidelines that would be -- as ultimately the goal, and how would sort of each of the clinicians kind of how would you sort of envision, based on at least this data that we've seen so far, what type of language do you think would be most appropriate for it?

Eric Klein

Matt touched on this, and this is tightly linked to the controversy over PSA screen. The preventive task force that recommended against screening didn't really recommend against screening because PSA screen doesn't save lives. It recommended against screening because men with non-lethal cancers get overtreated. And so that has put a damper on everything, and I think it's a very nihilistic approach. This is a proactive great approach to tackle the problem of overtreatment. I hope the guidelines will say now that we have a tool that prevents overtreatment, if the market adopts it, that we can recommend PSA screening for people because we've really reduced the harm. That's what I personally would like to see. [indiscernible] on the panel? [indiscernible]

Kimberly J. Popovits

Dr. Caroll, would you like to jump in there?

Peter Carroll

Yes, I'm on the panel that's tackling the guidelines issue which will likely change may not be completely in line with the AUA. The NCCN Guideline already strongly recommends active surveillance in very low-risk patients. And yet as we've heard the uptake of active surveillance is not what we think it should be. So I do think this will be another way to drive patients and physicians to adopt a guideline which already recommends we consider active surveillance.

Emily Faucette

Questions in the room? Do we have any questions on the phone?

Kimberly J. Popovits

Start with our questions on the conference call, please.

Operator

[Operator Instructions] Our question comes from the line of Bill Quirk with Piper Jaffray.

David C. Clair - Piper Jaffray Companies, Research Division

It's actually Dave Clair here for Bill. I guess my first question would just be kind of on that active surveillance, getting the patients to comply with the test results here. I mean, how do you expect to drive compliance I guess basically with the outcome of the test?

Steven Shak

Physicians are really the key here. I don't think the issue is to risk the patients harm. It's to inform the patients with a more accurate prediction of what their risk is and let the patient make the choice. That's really what it's about. Some patients currently have cancers that we think need to be treated but refuse treatment because they are risk-averse as far as side effects go and want to be surveilled. Some patients who using current guidelines or current information are great candidates for surveillance but insist on treatment because they can't deal with the anxiety. This should give them the confidence that they can make a correct decision.

Matthew Cooperberg

And I should gently disagree that the physicians are the key. It really is just conversation between the physicians and the patients. And I think there are also situations where patients will actively seek us out once the word gets out there, and I think there are a lot of patients that are looking to avoid treatment if they can. And we know there are some urologists that just don't do a lot of active surveillance for a variety of reasons, and I think this would be a way that it just -- it shifts the balance of the conversation and I think it really reduces uncertainty and helps both sides of the discussion, the physicians and the patients, get a better sense of confidence that they're making the right decision at that time.

Kimberly J. Popovits

This is where too the decision impact work that we have planned will be really helpful. So we have the same questions that were raised when we launched the breast cancer test and the colon cancer test. How do we know that patients aren't going to use the test and then continue to do the same thing. And remarkably or perhaps not across the world with both breast and colon cancer, we've seen treatment decisions change on the order of 30% to 40% of the time. And that's why that work around the Genomic Prostate Score is also going to be important.

Steven Shak

And well, I'll make one other point and then we'll ask Dr. Carroll about this. The other group that's been incredibly helpful to us over the years has been the advocacy organizations that devote themselves to patients with prostate cancer, Howard Soule in the Prostate Cancer Foundation has been tremendously supportive, and clearly they play a big role also in being there as an advocate for helping patients to understand the test and how to incorporate it into their treatment decision-making. And Dr. Caroll?

Peter Carroll

So I'd agree that I do think the tumor [ph] in the prostate cancer is gone just beyond urologists. There's a lot of information in the web, support groups are becoming very, very powerful in this regard, medical oncologists are increasingly being turned to as so-called honest brokers in this issue about treatment. The other thing I'd like to point out, for many patients, they don't -- they need confirmation that they're at low -- very low risk. So a test, it tells them they are indeed at very low risk without telling them -- without moving the needle, one way or another, a lot I think will also have an impact. So I think when we have results like this that draws attention to this issue, I think that these tests will be meaningful. I think they will be found to be helpful. We're certainly getting a lot of interest from our patients here at UCSF.

Kimberly J. Popovits

Right, and I think that perhaps maybe the real thing to focus on here too is that these side effects that these men incur for their life are extremely life-changing and they're significant. So as the awareness also increases around that piece of it, I think again a tool to help men more confidently feel like they can choose active surveillance and avoid that is going to be important.

Peter Carroll

If I may make a comment, if I could, I think it's the reverse that's also important to understand lives are saved with treatment, and I think the test will identify some men who may not be good candidates for surveillance. And certainly our experience here at UCSF suggests that there are such men. So I think the test again goes both ways, both lower risk and higher risk.

Kimberly J. Popovits

Yes. So it's a very important point. Thank you. Other questions?

David C. Clair - Piper Jaffray Companies, Research Division

Yes, and kind of another one for me. So the 4% biopsies that were not yielding successful GPS results, do you -- what do you expect that to be in practice? Is it going to be similar to 4%, and is there anything that we can do to lower that?

Kimberly J. Popovits

Steve?

Steven Shak

So actually, the -- we were actually quite pleased given the size of these biopsies to have had actually a success rate that was 96%. So point number one, I think that this is -- was an amazing technological feat. The amount of RNA that's required for the prostate cancer assay is twenty- to thirtyfold less than the amount of RNA that's required in the assay for breast and colon cancer. That being said, being able to deliver a result to every patient is really important to us. And so as we gain experience with the samples that come in now over the next weeks, we will be studying and learning from that experience, and as we did with breast and colon cancer, looking to do what we can to optimize that success rate. So actually that's something -- that's a good question to hold on to when maybe to come back to it when we have more experience.

Operator

Our next question comes from the line of Isaac Ro with Goldman Sachs.

Isaac Ro - Goldman Sachs Group Inc., Research Division

Sorry I'm not there, but I wanted to ask a question, Kim, if I could about any commentary or conversation you guys have had with the payer community yet. Number one, have you engaged with payers on the technology? And if so, could you maybe weigh the importance that they're ascribing to, the discovery of incremental indolent cancers versus the sort of low risk patients? Just trying to understand how that dynamic is playing out with the payers. And if not, I have a clinically related question.

Kimberly J. Popovits

Yes, I'll comment and then maybe ask Brad to jump in. One of the things that we can say we can absolutely leverage here is our managed care sales team, which has been highly successful, of course, in breast and colon cancer is the same audience here. So we'll be using that team to -- they already are talking to payers, and I also think similar to breast and colon cancer, the need for a tool like this is not a big debate. The question for payers will be how will people use it. There is no question that prostate cancer is overtreated along with the side effects that they continue to have to pay to manage in terms of quality of life and just these lifelong issues. So convincing payer there's a -- payers that there's a need, not a problem. I think the next step is convincing them that decisions will be changing with a tool like this. And, of course, we'll be doing additional studies to answer other questions, including further validation on this one.

G. Bradley Cole

Yes, I think what Kim says is recognizing that all the things we did in breast are going to be necessary here, to demonstrate that the clinical community finds this a useful tool and that patients in their [indiscernible] areas are actually getting this test and that we're seeing decisions change. So the data and publications around that are going to be just as important as they were in breast. And it will take some time. We're saying that it's not going to happen tomorrow, it's going to take some time.

Isaac Ro - Goldman Sachs Group Inc., Research Division

Sure, that makes sense. And Dr. Carroll and the rest of the physician group, if I could ask a question maybe on just sort of postulating around sort of the test results here, the clinical data results. Just do you have a sense of, from a pathology standpoint, why the stratification between the indolent and low risk cancers weren't even higher using the gene signature? Was there a reason why some of the pathology here can't be described purely with gene expression? Any postulation there would be interesting, just as we think about, like in an ideal world, what future iterations of this test might look like.

Peter Carroll

I'll postulate that these are -- well, that the biology of cancer is very complicated and that looking at gene expression, there's only one way to capture the biology of the tumors. And this is the first generation assay. There are lots of new technologies, [indiscernible] sequencing and looking at subtle mutations and amplifications and that sort of thing that we didn't assess here. So clearly, this doesn't capture all the biology, but there are other mechanisms that drive cancer progression, and this will serve as a platform to build the goal on and add those other things to -- excuse me, that's an important future challenge for us is applying all the advances in the genomics of cancer to get those curves up closer to 100%.

Matthew Cooperberg

And so you get other information from things like the PSA and the number, of course, is positive that tell you something about the burden of disease and the extent of the cancer that you're not going to pick up with the genetic information alone. But the other point to make is that there's really no point in trying to replace all sort of information because we will always have the PSA and the Gleason score and these other variables on hand. So we always we wanted -- the whole point is to maximize the use of the available information and really thinking about all these variables together is the right way to do that.

Peter Carroll

This is Peter Carroll here. If I could just comment, and I believe this addresses part of your question. It is clear that grading of cancer can vary based on the experience of the pathologists. And we've seen that over and over again here at UCSF. And my sense is that these genetic tests in the future may better describe the biology of these tumors, their aggressiveness, than maybe even a pathologist can. So again, in our study, we had an expert pathologist, we review all the slides, and I think that if you take an expert pathologist, this test may be actually even more valuable.

Operator

Our next question comes from the line of Vamil Divan with Credit Suisse.

Vamil Divan - Crédit Suisse AG, Research Division

So a couple of quick questions for the panel there, the urologists. One just in terms of PSA. I was just curious if you can kind of give your views on sort of how that has changed for you guys in terms of how many PSA test you're doing now versus maybe, say, a year or 2 ago as all this controversies heated up? I appreciate what you're saying about how adding this may generate more interest in PSA, but can you give me a sense of kind of how you've changed your practice, if at all, over the last couple of years?

Steven Shak

The task force was including non-evidence-based, politically motivated documents, so it had no impact on our practices. But in primary care community, we've actually really been on the front lines, trying to -- actually trying to get in front of this thing locally with the primary care organizations. It was not up to the urologists to screen most men. It's really up to the primary care physicians. And I think it's anecdotal, I don't think there are strong data out there yet. But there is a sense that PSA is being used less often, the concern is it's being used less often indiscriminately. Obviously, we know there's overscreening of older men with a lot of co-morbidity and things, but the way the messaging is out there from the task force is basically stop screening altogether. And I think in some cases, that's happening. Obviously, that's not the right answer. We've been really working to try to push this smarter screening message, and a lot -- a big part of that, of course, is better targeting of treatment to the men who are most likely to benefit from it. And I think urologists at a lot of the academic centers have really been out there trying to get this message onto the streets because ultimately, it's the primary care doctors make this decision. So I don't know that there's going to be a massive change in the epidemiology, but obviously it's something that we need to keep a very close eye on.

Peter Carroll

This is Peter Carroll. There is a movement of [indiscernible] the AUA and NCCN of revising their guidelines to try and address the issue of overdetection to try and limit the number of patients with very low risk disease to get detected. But even if they're fully implemented, the number of patients with low and very low risk disease will still be very, very large, and that's if urologists fully -- and primary care doctors fully implement the new guidelines. So I don't think this issue of overdetection is going to go away.

Operator

Our next question comes from the line of Doug Schenkel with Cowen and Company.

Doug Schenkel - Cowen and Company, LLC, Research Division

Is there [indiscernible] outcome data here on clinical value, for instance, how patients you classify as high risk, how do they respond to treatment? And how important is this type of data in your opinion?

Kimberly J. Popovits

So the question was outcome data in high-risk prostate cancer disease? Okay.

Doug Schenkel - Cowen and Company, LLC, Research Division

Yes, well, really just the predictive value of the test for [indiscernible].

Kimberly J. Popovits

The predictive value of this test?

Doug Schenkel - Cowen and Company, LLC, Research Division

For high risk -- yes, [indiscernible] I'm wondering if there is data available that would support the assertion that a patient that is classified as being at high risk will actually respond to treatment. And if you don't have that data, is that important in the context of talking to payers?

Steven Shak

Dr. Klein actually had in his study, it included patients who are at higher risk. You might want to comment about that in terms of how you think about this going forward now with the test available in clinical practice.

Eric Klein

Yes, it's a great question I think that we didn't related directly assess the development study. We were focused on answering the question, can genes predict outcome. It really wasn't designed to assess what the potential benefit of therapy is in men who have more aggressive tumor. So if you're asking do we have direct data that says if you have a high GPS score that you're more or less likely to be cured than someone whose GPS score is a little lower or higher. We just don't have that data. Mark, isn't that -- would you agree with that, Mark?

Mark Lee

For that -- to support the benefit of therapy, especially surgery, comes from 2 data sets [indiscernible] and the Scandinavian study. These randomized studies are very difficult to come by. The best evidence would suggest that higher-risk patients are the ones that derive the benefit.

Eric Klein

I agree. But I think the question was can we tie the GPS score as a marker for curability, and we did not assess that in any of our study.

Matthew Cooperberg

Right. You might have touched a point that I made this morning. Don't forget the genes were chosen from the [indiscernible] clinic cohort with the mortality endpoints. So these genes were originally picked to predict very distal clinical endpoints. So the fact that we're actually seeing the same signature predictive of earlier stage, more proximal endpoints, really probably that suggests that there is a signal here that's consistent across the whole progression of the disease. And that probably is relevant, I mean across the whole range of risk, low risk and high risk and in both before and after surgery.

Operator

Our next question comes from the line of Dan Leonard with Leerink Swann.

Daniel L. Leonard - Leerink Swann LLC, Research Division

I have a couple of questions. First, how do you envision -- for any of the clinicians on the panel, how do you envision this test being incorporated into active surveillance? Is it something that you would need to administer on a periodic basis to understand whether the biology has changed? Do you just administer it once? How do you do that playing out?

Steven Shak

Not a lot about, and I think there's a potential for this test and tests like it to tell us about whether or not a patient has biologic progression. So the initial motivation for this study was to develop a biopsy-based test that predicts biologic aggressiveness at diagnosis. But after we thought about that and we thought about how we do surveillance, it has occurred to us that you could also use it to deal with this other problem that I mentioned earlier. And a typical example, if someone has low-grade Gleason pattern 6 -- Gleason 6 on initial biopsy, low volume, and a year or 2 later has some grade 7 cancer. We don't know if that represents true biologic progression just based on grade, we don't know whether that grade 7 cancer was there initially and was missed by a needle biopsy. Those are 12-core biopsy samples, only 1/3,000 of an average-sized prostate. Or do we know that it was a different pathologist or even the same pathologist who didn't have some caffeine that morning and read it a little higher grade or slept at NBC suites [ph] or whatever that commercial is. So there is the potential for a test like this to compare the original gene expression in the original biopsy and the gene expression in a subsequent biopsy and see if the signature's changed, if the score has gone up. We have it, we -- honestly, we've talked about doing the study that needs to be done. We don't have any direct data on this yet. We also can't say with absolute certainty that if the score goes up, how high is high, what patient can still be safely followed, what's the variability test, retest variability, or whether we can cure the patients who are up higher. But we can really make the argument that this will be useful in monitoring patients. As Steve said, it's not only a test of who goes on surveillance, but it's also a test of how we do surveillance.

Operator

Our next question comes from the line of George Zavoico with MLV & Company.

George B. Zavoico - MLV & Co LLC, Research Division

This represents a little bit of a shift in the way the score is presented compared to your breast and colon cancer test. And I'm wondering how -- I guess, this question is directed more to the Genomic Health team, how you came about with the GPS from the Recurrence Score that you have in breast and colon, and whether there is a correlation between the GPS and the Recurrence Score test?

Kimberly J. Popovits

I'll ask -- well, I can say a lot input from many, many stakeholders here. But I'll let Steve comment on how we ended up evolving to the report we have.

Steven Shak

Yes. So first of all really, although the Recurrence Score is a multi-gene assay and is calculated by an algorithm, and prostate cancer score is a multi-gene assay and is calculated for -- from an algorithm, completely different genes, so -- and very different biology. And that's one of the things that led the score to be reported in a different way. The other thing that's different is the endpoint in prostate cancer was different in this study than the endpoint that breast cancer physicians cared about, which was, again, the absolute benefit of chemotherapy. So the report was really tailored, in that case, to show the relationship between the Recurrence Score and the benefit of chemotherapy. So in getting input on how to present this data, I would say Mark and the team just really thought about it a lot and then spoke to a lot of other people to get their input. And again, maybe, Mark, you want to just comment again on how this is all about -- isn't this all about can we confirm that our clinical and pathologic estimate is low risk? Can we confirm that, or is there something that might be harbored in the prostate that we need to know about that would change the way we think about our immediate management?

Mark Lee

Okay. This report came about through a lot of discussions both with urologists in the academic community, but also importantly, many urologists in the sort of community practice. We also had the opportunity to speak to many of the patient advocacy groups and talked through sort of different concepts of how this information should be relayed. But, Steve, you're absolutely right. I think at the end of the day, what this is talking about going from population-based risk estimation to more about individual risk estimates and looking at that overlap between different groups and providing that kind of confidence that patients are on the lower end of the risk spectrum and be able to make decisions based upon that.

Steven Shak

And maybe one other thing is that as was pointed out earlier, finding the small number of patients who indeed move to a less favorable assessment of risk is very important, and that can be clearly seen on this report and will lead again to again a different conversation between the physician and the patient.

George B. Zavoico - MLV & Co LLC, Research Division

So the way it works, and I mean the urologist and -- first scores the patient as low risk, then runs, I guess, the GPS score and sees whether it's above or below the original presumption?

Steven Shak

Correct.

Kimberly J. Popovits

Right. Or it may just confirm.

Steven Shak

Or may confirm it, which is very important as well.

George B. Zavoico - MLV & Co LLC, Research Division

Or confirm, yes. And so there's no correlation whatsoever to a Recurrence Score?

Steven Shak

Correct.

Emily Faucette

We have a question from Derik De Bruin from Bank of America coming in from the webcast, and his question is what is the estimated cost savings to the medical system if you do active surveillance instead of surgery?

Matthew Cooperberg

Well, it's actually -- I mean, to compare with surgery, surgery is relatively inexpensive compared to some of the other treatments out there. It's actually a difficult question to answer with precision because active surveillance is actually not necessarily that cheap because serial biopsies are actually quite expensive and because somewhere between 1/3 and -- around 1/3 of the patients will ultimately wind up getting treated within 3 to 5 years anyway. So I think the question really comes down to how many tests can be done to avoid one amount of IMRT, for example, or proton-beam therapy. And those numbers actually start to look quite favorable. Or, like we talked about earlier, if this test can be used to replace biopsies, biopsy itself is quite expensive. By the time you're figuring in the clinical costs and the local anesthesia and the pathology cost in particular, and I think the question is really how we can make active surveillance still more cost effective. I mean, the raw numbers, radiation is $50,000, proton's $100,000, surgery's $125,000. Active surveillance is harder to pin down because it depends on the frequency of the biopsies, et cetera.

Kimberly J. Popovits

And then there's the cost of managing side effects as well.

Matthew Cooperberg

Yes, of course.

Operator

Our next question comes from the line of Tycho Peterson with JPMorgan Chase.

Tycho W. Peterson - JP Morgan Chase & Co, Research Division

Just wondering if you wanted to comment on competitive positioning, in particular against Myriad, which really doesn't address heterogeneity. So can you just talk a little bit about how you'll be positioning competitively in the market?

Kimberly J. Popovits

Well, I'll start here and others can jump in. I think that is a very key differentiating factor. But it would also just go to the clinical question that we're asking and how it's relevant to the decision around moving more patients into the active surveillance or probably more specifically moving the right patients to active surveillance, because it is very important that we're seeing these patients move in both directions here. So it's really optimizing the treatment and reducing the overtreatment and undertreatment in prostate cancer. So just want to -- and I guess maybe one thing I would add too is the multiple pathways, I think, on top of the heterogeneity is certainly important as we've seen over and over again these cancers that matters in predicting the disease.

Eric Klein

I might add that this is a patient population that's treated like we treat patients in the U.S. So it's a relevant patient population for us to think about the patients. And biopsy samples is a real relevant question.

Kimberly J. Popovits

Right.

Eric Klein

It's where we started.

Tycho W. Peterson - JP Morgan Chase & Co, Research Division

And then can you maybe just comment also how you think about other technology approaches, exome sequencing in particular being competitive over time?

Steven Shak

Well, there's certainly tremendous interest in the general group of technologies that would be classified next generation or high-throughput sequencing. But I think it's also now very clear that what matters with that technology again is what data matters, how do I use it, how do I put it into a clinical context, and again that's the critical question. So I would say if I was a patient today, would I rather have a GPS score and a Gleason score and a PSA or information on my whole genome to make the decision around active surveillance, I think it's clearly the former. But clearly, from a research and development point of view and again in terms of our organization looking forward in terms of our pipeline, clearly the high-throughput technologies like NGS open up the lens and give us a chance to look at molecular alterations that we had never considered before in an incredible way. And so as we think about the future products that we develop, clearly we're going to be using those kinds of technologies and learning from them. It's very exciting.

Mark Lee

I agree with Steve. It's very exciting. I would point out that the sexiness of a new way of looking at biology is interesting in its own right. But unless it adds some unique power to making a decision beyond what's already available, it may not be useful at all clinically. That doesn't mean it's not important biologically. It just means it may not be able to be applied to a clinical situation in a meaningful way. So when you have a new biomarker, it's always not does that biomarker replaces what we already know is useful in grade-stage PSA and now GPS, but what does it add to those. That's really the question.

Kimberly J. Popovits

This is what payers will be looking for.

Eric Klein

Yes.

Operator

And I turn the floor over to the room.

Unknown Analyst

On just cost that we should be expecting in the future. So can you help us think about the number of follow-up trials that are necessary, number of sales people and specialists that might be coming onboard, and what your expectations are for 2013, a couple of years out? So can you help us just frame all those costs associated with the launch?

Kimberly J. Popovits

Dr. Cooperberg? Brad?

G. Bradley Cole

I think it ought to be free.

Matthew Cooperberg

Yes [indiscernible] get together.

G. Bradley Cole

It would make our job so much easier.

Steven Shak

We have an online portal so feel free to order anytime. From a commercial perspective, we have a group of about 15 people dedicated to it today, a sales team of 9 or 10 and supporting medical marketing team associated with that. Much like we did in breast, we will use that small group to be very focused under large community practice groups academic centers to get started. And as we see adoption and more importantly reimbursement, we'll add to that investment. So we're going to spend, I think, we've said this in previous calls, upwards close to $10 million on launch and commercial activity this year. And I imagine it'll grow in future years, it'll be high traded with progress in the market. There is more work to do clinically, and I think we're going to expand the market beyond the 50% of patients we think are addressable today. Maybe Steve can talk to that.

Kimberly J. Popovits

The research and development budget as a percent of our budget, now prostate, is a big portion of that. So we have a number of other studies planned and continue obviously to make a significant investment.

Steven Shak

I'll highlight again, although this study focused on adverse pathology as the primary endpoint for all the reasons that were outlined, the previous studies by Dr. Klein emphasize that this same set of genes actually did predict clinical recurrence in the long run. And obviously, as we do additional studies, we'll be looking not only at adverse pathology but also looking at those longer-term events as well in order again to reemphasize that point and reinforce that point going into the future.

Kimberly J. Popovits

Great. Well, thank you, everybody, for again your interest. It's been a very exciting week. I think we're -- probably many of us are running on adrenaline at this point. So we'll call it a day here in San Diego. We look forward to know your follow-up questions. You know where we all are, and I just would be remiss to not thank our panelists here, especially our great colleagues that have joined us in this wonderful effort in creating a great result for patients this weekend. Dr. Carroll, thank you for hanging on the line with us, and we'll see you soon.

Operator

Thank you. This concludes today's teleconference. You may disconnect your lines at this time, and thank you for your participation.

Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to www.SeekingAlpha.com. All other use is prohibited.

THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.

If you have any additional questions about our online transcripts, please contact us at: transcripts@seekingalpha.com. Thank you!