In this interview, Daniel W. Lin, MD, discusses the practical use of currently available molecular and genomic tests, cost and reimbursement considerations, the role of MRI, and what the future holds for biomarkers.
The discovery of new prostate cancer biomarkers has led to a number of tests that have begun to alter the diagnosis and management of the disease. In this interview, Daniel W. Lin, MD, discusses the practical use of currently available molecular and genomic tests, cost and reimbursement considerations, the role of MRI, and what the future holds for biomarkers. Dr. Lin is chief of urologic oncology and professor of urology at the University of Washington, Seattle. Dr. Lin was interviewed by Urology Times Editorial Consultant J. Brantley Thrasher, MD, professor and chair of urology at the University of Kansas Medical Center, Kansas City.Dr. Lin has research projects and/or has been an investigator for several products mentioned in this article, including Prolaris, Oncotype DX, Decipher, PCA3, and 4Kscore Test.
Please give us an overview of some of the more commonly used biomarkers for prostate cancer.
The world of biomarkers in prostate cancer has virtually exploded in the last 5 or so years, particularly the ones that are molecularly or genomically oriented. The easiest way to separate the available biomarkers is by the clinical situation in which they are used. First are the markers that are used in the screening and detection of prostate cancer. Second are those that are used after the diagnosis, perhaps for making a decision about the various treatment options. And third are those that are used after treatment or at the time of failure of prostate cancer therapies.
That’s the first division I use. Within those categories, there are various markers. For example, some markers in the screening and detection realm are almost always used as an adjunct to PSA. In other words, we are not using these markers in a vacuum; rather, they are almost always used in a population of men who are being referred for elevated PSA or more likely have had a biopsy for an elevated PSA, the biopsy was negative, and we’re trying to determine whether a repeat biopsy is needed.
Next: What's the practical application of these tissue genetic markers?
Let’s talk more about the second group. Once I provide a tissue diagnosis, a lot of my patients ask for a genetic profile that might give them a better idea of whether or not the tumor is aggressive. That’s being mixed with active surveillance protocols and MRIs. What’s the practical application of these tissue genetic markers?
The major tissue genomic and molecular markers used after the diagnosis to determine which way to go-active surveillance or radical therapy like surgery or radiation-provide different data outputs to the patient. One of them, the Oncotype DX test (Genomic Health, Inc.), will give the patient and physician the likelihood of favorable pathology at time of prostatectomy. Another test, the Prolaris (Myriad Genetics, Inc.), provides a patient with the risk of mortality from prostate cancer with observation and the risk stratification within a group of other men like him-essentially a percentile score within his risk group, ie, compared to other men like him.
There are three main situations in which I use the tests in this category. One is in a patient who needs more data; he may understand where he lies within the risk stratification but just needs more data to push him one way or the other. That is a very common use. Where I tend to use them even more is in the situation of discordant clinical data. For example, a man may have a small focus of low-risk disease, yet his PSA is rather high. The clinical data don’t quite make sense, and the use of a molecular test may give insight into tumor biology.
The third use is in a man who needs confirmation of his decision to opt for a particular treatment strategy, active surveillance for example, and wants to be sure that’s what he really wants to do. In general, that is how I’ve used those tests in a man who already has a diagnosis of prostate cancer.
You’re not saying that every man with a low-volume Gleason 6 cancer who might be on active surveillance needs a molecular or genetic profile, correct?
You’re exactly correct. We already have level one clinical data-data from the U.S. PIVOT trial and the Scandinavian Prostate Cancer Group trial of prostatectomy versus watchful waiting-showing that in the typical low-risk man, particularly those with low-volume Gleason 6 disease and particularly in men over age 65 or 70 (in the Swedish trial), there did not seem to be any benefit of surgery over watchful waiting. In addition, guidelines from the National Comprehensive Cancer Network, the AUA, and the European Association of Urology indicate that at this point in time, for the very low-risk or low-risk man, active surveillance is a bona fide therapeutic option, and indeed a preferred option in some cases, that does not need further confirmation through molecular testing.
Next: Among your intermediate-risk patients, are you using active surveillance in any with Gleason 3+4 disease?
Among your intermediate-risk patients, are you using active surveillance in any with Gleason 3+4 disease?
We have been using it in select patients. If you look at the NCCN guidelines, it does say in a footnote that active surveillance is an option in “select” patients with intermediate-risk disease. Select patients would include perhaps those who were already under active surveillance and have a small amount of Gleason 3+4 disease. The nuance here on a practical level is that there are Gleason 3+4 cancers with a minor component of secondary pattern 4 versus those cancers that are 3+4 with a lot of pattern 4 disease. I think that nuance really has been lost. It may be difficult to practically apply this concept because it relies on a fair amount of collaboration with our pathology colleagues. But we do know that select patients with low-volume Gleason 3+4 could be offered active surveillance, per various guidelines.
Do you use molecular or genetic testing in those patients to make you feel better or give the patient some assurance?
I do think that would be a prime situation to use molecular testing, again in select patients with low-volume Gleason 7. If one looks at the current NCCN guidelines, in patients with Gleason 7 disease and a life expectancy of over 10 years, active surveillance is not offered as a treatment option. Molecular testing could provide confirmation of the biology of that tumor that goes beyond what we can see under the microscope. Thus, again in select cases, I have been offering molecular testing for those men when they want to have active surveillance despite having some 3+4 disease.
Let’s say a patient is on active surveillance and his PSA continues to go up, but you’ve only found low-volume Gleason 6 disease. How have you incorporated biomarkers and/or MRI in those patients?
First, MRI at this point in time is not a standard of care, certainly according to guidelines, and still cannot take the place of a biopsy. In that clinical scenario, which happens on a day-to-day basis, we have been incorporating MRI on select patients to understand whether there’s a lesion that might be in an abnormal location or a location that’s more difficult to biopsy, such as the anterior zone of the prostate. We’ve also been using molecular markers with or without MRI in that situation because most of the molecular marker data have shown that, anywhere in the prostate, the molecular signature reflects other sites of the prostate as well. So there might be a biopsy of Gleason 6 cancer, but it should reflect the biology of perhaps a higher grade cancer that’s in another area of the prostate.
Next: Which of these genetic profiles do you prefer and why?
Which of these genetic profiles do you prefer and why?
It depends on the endpoint of interest for the patient and physician. As I said, the Prolaris test will give the patient and me an idea of where within their risk group he lies, as well as his mortality at 10 years if he did nothing. Those are the two major points it gives. It provides the patient with a percentile risk within a group. In the low-risk group, for example, if it’s the 50th percentile, that means he’s right in the middle of the low-risk group. If it’s the 80th percentile, that means that he’s higher in the low-risk category. Then it gives a mortality risk for over 10 years, which is a very “hard” endpoint, and it’s usually low for a low-risk man.
Conversely, if a man wants to know, “if you take out my prostate tomorrow, what will be my risk of having bad pathology” such as primary pattern 4 disease, the Oncotype DX provides that answer. It not only tells the patient and physician how he lies in his risk, but it also tells him his chance of adverse pathology. Sometimes that information can be rather difficult to interpret because it is a percentage number of risk, and it may be difficult to understand where to draw the line. As a side note, many of these markers give various various percentiles or a percentage chance of risk, and knowing where to draw that threshold can be very difficult to interpret. In my opinion, that is one of the disadvantages to all the markers.
I have some patients with Gleason 8, 9, or 10 disease who want to have a genetic profile done. I tell them, based on the current available knowledge, that is not going to make much difference in what I’m going to recommend for their treatment. Do you use these genetic tests in any high-risk cancers?
I agree with you. I am not using them in any high-risk cancers for a couple reasons. One is, if you look at the clinical trial data, observational cohorts, case series, or SEER data, you will see that patients with Gleason 8, 9, and 10 cancers have suboptimal outcomes no matter what. There are certainly outliers, but how many tests do you have to perform to find that one outlier and also justify the cost? Also, when one looks at the published data on either the Prolaris or Oncotype test and how the results affect men in the high-risk group, men are usually still getting intervention. There are few men in the high-risk group who might actually switch from an intervention to a non-intervention due to the molecular test.
Next: What do you tell the patients about cost and insurance coverage for these tests?
Since you mentioned cost, what do you tell the patients about cost and insurance coverage for these tests and whether they’re really a return on their investment?
To my knowledge, almost all of the tests are covered by insurance for select patient types. It’s very clear that the scenario you presented of a high-risk patient might not be covered. Then we’re talking about a cost to the patient of $2,000 to $4,000, and I don’t think the return on investment is clearly evident there. For those who are within the low-risk or very low-risk group and soon possibly the low-intermediate-risk group, Medicare and most private insurance carriers have been providing reimbursement at a reasonable level, which in that case would cost the patient a couple hundred dollars. For the patient with a discordant pathology or who needs affirmation to relieve anxiety, I think that’s worth the money.
At least in our area, third-party payers are not allowing us to do an MRI when the patient walks through the door. They generally say the MRI won’t be covered if the patient hasn’t had a previous biopsy. However, I’ve heard that in places where they’re doing a lot of MRIs, including Europe, the MRI is done upfront. Do you see these biomarkers being done upfront along with a $500 MRI to potentially avoid a biopsy or do a more targeted biopsy?
At this point, I agree with the payers that doing an upfront MRI before any biopsy is not supported by the data. The Europeans and Canadians might do MRIs for $200 to $500. At our site, it’s well into the thousands. The great Achilles heel of MRI currently is the problem of interobserver and community-based transmissibility or applicability. The centers of excellence that you mentioned are just that; they have people dedicated to MRI who would tell you that the applicability to a community setting or other settings where there is interobserver variability is very problematic. Until we figure out how to read MRIs consistently and reliably, I think MRI will have some issues.
The reimbursement is a major issue. An MRI is several thousand dollars. Until the radiologists get behind doing limited MRI-perhaps just T2 and diffusion-weighted MRI instead of the whole multiparametric MRI-and lower the cost in this country, it will be a problem. Regarding doing biomarkers upfront, regardless of MRI, there are quite a few biomarkers that predict the likelihood of cancer-and in one biomarker specifically, high-grade prostate cancer-and these can and are being used upfront before biopsy.
Next: "Patients often say they don’t want another biopsy; they want a test that tells them with high negative predictive value that they don’t have cancer."
That allows us to change gears a little. One of the most common requests my patients make is for a 4Kscore Test or other serum test that would give them assurance that they don’t need a biopsy. Tell me how you’re using this type of test in your practice.
There are several major tests in this category available now. All are involved in the pre-diagnostic setting to lend information about decision-making for an additional or initial biopsy. Two are blood tests, one is a urine test, and one is a tissue test.
Patients often say they don’t want another biopsy; they want a test that tells them with high negative predictive value that they don’t have cancer. There are two tests that give high negative predictive value. One is the ConfirmMDx (MDxHealth), which looks at tissue methylation. The tissue from a just-completed biopsy is examined for methylation signatures. This test provides a negative predictive value of up to 90%. In other words, if the test is negative, then 90% of the time there is no cancer on repeat biopsy based on two clinical trials. Furthermore, there is some evidence that the negative predictive value for high-grade cancer may be 95%.
The other assay that has a high negative predictive value is the PCA3 test, a urine-based test that uses a urine sample given after a digital rectal examination. It is FDA approved and is probably one of the most widely used tests. If the test is negative-the PCA3 is less than 20-the negative predictive value is also 90%.
If results from either of those tests are negative, most patients would say they’re not having another biopsy.
The other two tests give different information. The 4Kscore (OPKO Lab) and the phi (prostate health index [Beckman Coulter]) are very easily obtainable and require only a blood draw. The 4Kscore gives the positive predictive value of high-grade disease, which is very attractive because we don’t necessarily want to diagnose a Gleason 3+3 lesion; we want to look for clinically relevant, high-grade disease. The 4Kscore actually gives a percentage of having high-grade disease, allowing the patient and physician to decide whether to proceed with a biopsy.
The phi basically does the same thing. Higher phi means a higher chance of any cancer. The result is also related to high-grade cancer; they just don’t necessarily give that as a primary output.
Of these tests, is there one that you use more frequently in your practice, and why?
I’m using the PCA3 for the negative predictive value aspect only because ConfirmMDx is just coming on the market, although I think ConfirmMDx is attractive as well. Both have high negative predictive value. I’m using the 4Kscore more than the phi primarily because I used it during the clinical trials, and it provides me the probability that a patient has a cancer is high-grade cancer, which is important because clearly those are the cancers that we want to detect.
Next: Do you talk to patients about cost and issues with insurance related to these particular markers?
Do you talk to patients about cost and issues with insurance related to these particular markers?
Yes. Most all the tests are covered, although sometimes only for certain subsets of patients. The PCA3 is covered in the repeat biopsy setting. The ConfirmMDx test is also covered, and I think most insurers cover the phi. For the low-risk group, the Prolaris and Oncotype tests are covered by most private insurance. Again, it’s important to have the conversation with patients who want added data. Also, as you implied before, there are still questions about how to use these tests with or without an MRI, and those studies are still being done.
In a patient who has a previous negative biopsy, comes back with a rising PSA, and is concerned, he will generally go straight to MRI. Are you saying now that you would prefer the biomarkers upfront instead of imaging, or are you doing both?
We are using both. We have the availability of both multiparametric MRI and MRI fusion. It does appear that the fusion technologies target the lesions better than simply obtaining an MRI and aiming needles where there appears to be a lesion. Fusion technology has greatly improved accuracy, and thus we’ve increased the utilization of fusion recently.
Have you read: PCa agent detects 97% of lesions when paired with PET
I give patients the options and explain that insurance carriers will sometimes not cover an MRI. When I tell them about PCA3 and the fact that it’s only a couple hundred dollars, they often choose that. I think what patients are being burdened with is starting to affect our practice patterns. In a very structured way, I leave it up to them to choose what they want. For example, do they want a high negative predictive value test or an MRI, which has been shown to have reasonably high negative predictive value for high-grade disease but generally a low positive predictive value?
Please tell us about the ongoing, important studies you think would give us better direction on when to use MRI and what new biomarkers and tissue markers might be available in the future.
First, I’ll point out that many of the published studies of the molecular markers have been done using radical prostatectomy and radiation therapy patients, but now in practice, these markers are being used oftentimes to determine whether a man might be eligible for active surveillance. To my knowledge, none of the markers have been tested in the active surveillance population, and studies have not been done in a longitudinal, prospective fashion to understand the relevant questions about how these tests relate to whether patients will need treatment later and whether worse disease will be found later. The studies that are ongoing now, many of which our group is participating in, are designed to study these markers within the intended-use population.
The other way that these markers are being used is in a clinical trial setting. Many of the clinical trials that are being designed, including those by the cooperative groups, may include some of the markers. For example, the Decipher test (GenomeDx Biosciences) may be used to provide entry criteria for some of the trials of advanced or locally advanced prostate cancer.
Finally, there’s a wealth of data on circulating tumor cells, cell-free DNA, and androgen receptor variants. We will be better able to understand, perhaps with next-generation sequencing, the baseline risk of each individual. In this way, I think the next stage of biomarker development will be personalization-clear individualization of markers for either certain genotypes or even certain clinical subsets of patients, for example, men with smaller prostates or men of a certain race or a certain age. I think that’s the direction that we are headed next.
Next: Use of markers for determining adjuvant therapy
You mentioned the Decipher test. There haven’t been many markers used in the adjuvant setting. Can you talk about the use of markers for determining adjuvant therapy?
Most of the tissue markers that we talked about are prognostic markers; they provide a prognosis for how a tumor might behave. The Decipher test clearly has the ability to provide the probability of metastatic disease after prostatectomy. Additionally and importantly, it has also been shown in a couple of small but growing studies that it can also be a predictive test. It can actually predict whether a man might respond better to adjuvant radiation after the diagnosis of high-risk disease or perhaps whether he can wait and get salvage radiation therapy. I think that is truly important, provocative, and novel. That’s really the Holy Grail of markers-those that predict response to a therapy, whether that be radiation in this setting or chemotherapy or hormonal ablation.
For the rural practitioner who doesn’t have an MRI and may not have access to some of these markers, what would be available to them that would be both pragmatic and cost effective?
Most of these marker companies have made the accessibility less of an issue. They are readily available as far as access to running an assay is concerned. The world is flat, and my sense is that the biomarker tests can be done anywhere.
More on PCa: A new castration-resistant prostate cancer entity?
The penetrance of MRI is more of an issue, particularly the fusion-guided technology. I’ve been telling rural providers across my region in the Northwest that the biomarker tests are available to them and the newer MRI technologies are coming their way.
Among our patients on active surveillance who undergo fusion biopsies, I don’t think we’re seeing as many as one out of four in whom we find something that changes what we would do, as some studies have shown. We are, however, finding a significant number of lesions with MRI that are poorly accessible with random biopsies-usually on the anterior base or anterior zone near the apex. Has that been your experience?
We’ve been using both multiparametric MRI and fusion technology for a while now, and we’ve found the percentage of involvement of cancer within individual cores from a lesion is higher than with systematic template biopsies. The template biopsies might hit that lesion, in part because we know where it is by the MRI, but it might not hit it with as much robust sampling. In other words, with MRI fusion, we may have almost a whole core be cancer from a Gleason 7 lesion, whereas my systematic biopsy might hit the cancer, although it might only be 20% to 30% of that core. So, is it adding information on an actionable level? Yes, but not all the time.
Next: "With fusion, I am 'forced' to biopsy an MRI lesion that also appears hypoechoic on ultrasound."
Dr. Ian Thompson said it best when he said: “I think I’ve gotten much better at reading ultrasounds just by using my fusion because I’m seeing things I didn’t see before now.”
I completely agree. With fusion, I am “forced” to biopsy an MRI lesion that also appears hypoechoic on ultrasound. We still perform template systematic biopsy following the lesion-directed, fusion cores, and I’m coming around to where the MRI lesion was and say to myself, “There it is again.” If I didn’t do the fusion biopsies, I wonder whether I would have paid so much attention to that hypoechoic area.
You talked about personalized medicine earlier, which is really the buzzword we hear in discussions about the future of biomarkers. What do you think the most exciting markers will be in the future?
I think it’s going to be direct tumor sequencing. Then the question is whether the circulating tumor cell (CTC) is a window into the soul of the primary tumor or the metastatic tumor, and one hopes that it will be. Sequencing the relevant pathways within a tumor is ultimately where we will be. It’s currently in the hands of our medical oncology colleagues, who are doing biopsies of bone metastases and directing therapies for advanced disease based on sequencing. One would hope we could translate that into a blood draw that looks for cell-free DNA or circulating tumor cells (CTCs) and then sequence these cells for insight into tumor behavior or for targets of therapy. I think the world of precision medicine will be in the tumor perhaps in addition to inborn genomics, such as single nucleotide polymorphisms (SNPs).
We hear often about BRCA and other germline mutations, and while those are definitely relevant, the number of patients who have BRCA mutations is less than 5%. You’re talking about screening many hundreds of people to find those few. Most of the ones that are very common in prostate cancer genomics, like the SNPs in 8q24, and others might be very prevalent, but the risks that they confer are low. At this point, my sense is that currently the tumor tissue is the relevant focus and in that tissue, we will be applying next-generation genomics and sequencing.
Next: "Eventually, it will be very affordable to have, perhaps not whole-genome sequencing, but certainly relevant pathway sequencing on the order of the cost of an MRI."
Sequencing could carry a potentially substantial cost. Is it going to be reasonable or solvent in the future, based upon what it might cost us today?
We can learn from the past. At one time, it cost millions of dollars to sequence a genome, then it went down to hundreds of thousands, and now it’s down to the ten thousand dollar level. Eventually, it will be very affordable to have, perhaps not whole-genome sequencing, but certainly relevant pathway sequencing on the order of the cost of an MRI. I clearly think that will happen.
Is there anything you would like to add?
I would point out that we’re probably a few years away from having more directed molecular markers on clinical guidelines. I am not talking about markers that are merely suggested or presented as options or in the footnotes, but more prescribed and recommended markers that are within the main flowchart of guidelines such as the NCCN. For example, looking at the NCCN guideline for certain breast cancers, it currently includes not only estrogen and progesterone receptor status, but also HER2 status. The Oncotype DX test for breast cancer is also included within the main flowchart of these breast cancer NCCN guidelines. They’re not simply listed on the guidelines, but prescribed pathways are provided; in other words, they indicate that if a patient has a certain test score, then they should or should not receive a certain endocrine therapy or chemotherapy.
While we’re a few years away from that for prostate cancer, if we get the clinical trials done, we will eventually have something more than an algorithm that says either do nothing with surveillance or do something radical with radiation therapy or surgery. We’ll get to the point of having a biomarker or biomarker panel, perhaps not in all patients but in certain patients, to truly prescribe the way that we go about treating our prostate cancer patients.
Subscribe to Urology Times to get monthly news from the leading news source for urologists.