San Francisco, CA USA (UroToday.com) Eight international sites participated in a large validation study to test the potential of a recently developed method to identify men with localized prostate cancer who are appropriate candidates for active surveillance. Experience with a novel CCR score was reported at the 2016 ASCO Genitourinary Cancers Symposium. Here, Michael Brawer, MD, from Myriad Genetics, Inc., an author on the study, responds to inquiries from UroToday about the potentials of a new strategy his company has developed and validated the new CCR, that combines cell cycle progression (CCP) with the well-known CAPRA.
Q: What was the thinking behind the development of Prolaris—a tool to aid clinicians and patients in making a decision about deferring treatment for localized prostate cancer?
There are controversies in prostate cancer today that include over-detection, the indiscriminate use of prostate-specific antigen, and reflex reactions to treat newly diagnosed prostate cancer. This has led to the concern of over-treatment, and [the suggestion] that many men don’t accrue a benefit from such treatment. As the co-chair of the PIVOT trial1 a number of years, ago, we showed that overall there was no benefit of radical prostatectomy as compared to conservative management—or watchful waiting.
On the other hand, we know that 33,000 American men will succumb to prostate cancer this year. So, on the one hand, we have over-detection and over-treatment, but on the other hand, in some instances, we have a problem of under-treatment, or late detection.
There are a number of ways that we can address this over-detection. One of the most important is to provide patients with more accurate, more precise, risk assessment of their cancers such that potentially, we can more likely refine the [characteristics of] the men that we treat [separating] the men that will need aggressive treatment, and put more men with more indolent cancer into a less aggressive regimen such as active surveillance.
That represented the genesis of the development of Prolaris, which is tissue-based. It’s done on the diagnostic prostate biopsy or on a radical prostatectomy specimen, where we utilize reverse transcriptase—RTPCR. We quantitate the amount of a specific type of gene—cell cycle progression (CCR) genes—that tell cancer cells to divide. Unregulated cellular proliferation is perhaps the quintessential hallmark of cancer.
Steve Stone and his colleagues at Myriad a number of years ago began looking at these genes to see if they could more accurately predict outcomes from prostate cancer. At Myriad, we’ve always used real oncologic endpoints, such as failure by the most commonly biochemical recurrence after radiation after radical prostatectomy or the development of metastases after these treatments, or absolute disease specific mortality. At Myriad, we have published nine validation cohorts, showing in each case the Prolaris is the dominant predictor of outcome when compared to other clinical/pathological parameters such as Gleason score, PSA, extent of disease clinical stage, etc.
In the poster presented at the 2016 ASCO Genitourinary Cancers Symposium by multi-institution researchers we performed an analysis to try to determine an endpoint – and actual cut-off—that could help clinicians guide appropriate patients into conservative management and surveillance versus more aggressive treatment such as radical prostatectomy, radiation therapy, etc.
To address this, the researchers asked themselves what would be reasonable criteria based on clinical-pathologic parameters, for consideration of active surveillance?
What we decided on was Gleason 3+4 or less cancer, 25% or fewer biopsies. Patient subjects had to have PSAs less than 10, and clinical stage T2A or less.
Q: Did your decisions about the criteria to adopt correspond to or align with any other standard clinical staging guidelines?
There is no real standard of clinic-pathologic criteria for active surveillance. Virtually every group has used Gleason 3+3. Some are using 3+4 (such as Peter Carroll [UCSF]). Most have used PSA <10. Most have used T2A or less. Among all of the studies on active surveillance, there are differences; there is no overall consensus.
Q: So, there is no accepted standard definition for active surveillance?
Some groups restrict [patients who are eligible for active surveillance] to 3+3. UCSF allows 3+4: there is no overall consensus. That is why, when we set our parameters [for the active surveillance threshold], we tried to [include] the range of active surveillance cohorts currently being done.
Q: And as I understand, it you applied, or tested, this validated AS (active surveillance) CCR threshold with a series of prostate cancer tumor samples.
A CCR threshold was previously developed in a training cohort of men. We tested this [active surveillance threshold] on 505 recent patients. We asked, “What was the [patient’s] combined clinical risk, consisting of their expression of cell cycle progression [CCP] genes. What was the number of transcripts, the number of Mrna of these 31 genes plus the CAPRA score, which is the most well validated way of amalgamating clinicopathological parameters? In all of nine validation cohorts, we ‘fixed’ the relationship between the CAPRA and the CCP to get what we call the combined clinical risk score.
Then, because I wanted to extra-conservative, I said “Don’t give me the cutoff of the 505 patients, what their highest level was just give me the 90%, so were [especially conservative]. It turns out that this cutoff was a CCR of .8.
One of the strengths of Myriad [Myriad’s study design] is that we have three cohorts of men with clinically localized prostate cancer who were tested with Prolaris. They were initiated with conservative management—watchful waiting—and were followed for at least a decade. The endpoint was disease-specific mortality, the most valid endpoint and the one we have to be impacting, or we’re really doing nothing with prostate cancer.
We had the luxury of having the ‘endpoint of all endpoints’ [disease-specific mortality]. With that cutoff, we looked men who came out of two of the series where we had done biopsies. We did the Prolaris [gene expression] testing on the biopsies and followed the patients for a decade, with an endpoint of death from prostate cancer.
It turns out that none--not one—of the patients who were below the threshold of a CCR of .8, died from prostate cancer at 10 years. Men who were over that threshold--.9 or above—and were followed for 10 years, 21% died of the disease.
------------------
Not one of the patients who were below the CCR threshold of .8 died from prostate cancer at 10 years.
------------------
Because of the confidence intervals and the statistics that were involved in this, we would have predicted that a few patients who were below the threshold would have died. But when we looked at this, we had none.
Q: You are saying that the difference in the CCR score between 0.8 and 0.9 was the difference between life and death at 10 years?
What it means is that we now have a threshold that was devised rationally, using Prolaris combined with CAPRA, and we had an observation [period] of a decade of conservative management when there were no deaths. Then we asked the obvious question. We know we need to increase active surveillance, or conservative management in patients treated today, but we need to treat the right patient. Patients who are not good candidates should be treated aggressively initially, because the best time to treat a patient is the first opportunity.
If we think a man based on Gleason score, PSA and TNM is a good candidate for active surveillance, but a molecular marker like Prolaris tells us that they are not [a good candidate], and we don’t [test] this marker, they may progress and miss the opportunity for care.
Q: I believe that when you then undertook this analysis, you had in mind to take it to a larger, and/or different population for testing.
What we want to do is get the right treatment to the right patients. So, with this threshold and evidence in our validation [study with Prolaris] and it worked 100%, then we said, what would happen when applied to a current US population?
For this we looked at a subset of patients up to the point of this analysis—patients who had been tested with Prolaris at Myriad—a total of 11,665 men. We asked what percentage of these patients could be considered now for active surveillance because they were below the threshold of combined clinical risk, of .8. It turns out that of 11,665 men, 62.8%, qualified for active surveillance [and] are undergoing commercial testing today [with Prolaris].
What is particularly interesting to note is that 45% of those men would not have qualified for active surveillance based on the clinical and pathological parameters alone. But because they had a low Prolaris score—a low number of transcripts of these genes—it would move them to left in the aggressiveness scale in prostate cancer.
Q: Are you saying patients who looked, clinically, like they had aggressive disease, but who had a low score on the Prolaris test, did not have aggressive disease?
We could initiate almost two-thirds of men, based on the commercial testing with Prolaris, on conservative management. And, every bit as important as this, the men who may have looked like they were good candidates for conservative management, but based on the cell cycle progression expression—Prolaris score—were not considered good candidates for active surveillance and would be move on to curative therapy.
What is comes down to be that a lot more men can be candidates for active surveillance than before. But equally important is that men we might not realize have more aggressive disease would have moved on to definitive therapy early on. So, you really are giving the right treatment to the right patient at the right time when you add the increased precision of the prognostic information that Prolaris provides.
------------------
You really are giving the right treatment to the right patient at the right time when you add the increased precision of the prognostic information that Prolaris provides.
------------------
Q: With this strong evidence for clinical value, has or will Prolaris be considered for expert recommendations—so to be brought into wider clinical use?
Initial interactions with CMS have resulted in agency approval for Prolaris and [reimbursement] for use in low-risk men with localized prostate cancer. We are now increasingly getting the private companies to follow along. We are working to expand the coverage through additional investigations, and discussions with many of these payers. [Modification of] NCCN [guidelines] is still in draft form, but they have seen the potential benefit of molecular testing and have revised their recommendations in the initial assessment. I should add that Prolaris is indicated across all groups—low intermediate and high-risk disease. The NCCN has said in the initial evaluation of prostate cancer patients with clinically localized disease that molecular markers can a useful addition to a standard testing. Other similar groups are looking at changing their recommendations.
Q: Is there potential for Prolaris to have a role in the treatment of high-risk prostate cancer?
If we ask, where do we get the “biggest bank for the buck”—we know that the low-risk patient can only move to higher-risk disease. A high-risk patient can only be moved to lower-risk. An intermediate patient can go in both directions. About one-third of intermediate patients, with Prolaris, will be shown to have less aggressive cancer than we think, and about one-third of patients will have higher-risk disease.
Overall, Prolaris will shift the risk assessment across low, intermediate, and high-risk patients about half the time. About half the time, men with high risk [disease] will remain at high risk. About 45% of the time they are going to move toward intermediate, or low risk. About 40% of the time, low-risk men with move toward intermediate or higher risk. Intermediate patients move in both directions.
We are going move most patients in the intermediate group. Pragmatically, the biggest question today is “Is my patient a good candidate for active surveillance?”, and in that setting, more patients are being tested at low risk.
Specifically referring to the question about high-risk: we know that about 40% of patients in a high-risk group are going to behave like low- or intermediate-risk patients. By identifying them, we can tailor the therapy more precisely.
For example, there are a number of papers at this [ASCO GU] meeting that talk about multi-modality therapy in high-risk prostate cancer. We know that the failure rate of radical prostatectomy of external beam radiation therapy or radiation therapy alone, is probably too high in that setting. The data are best for the combination of androgen deprivation therapy and external beam radiation therapy in high-risk patients.
But what if, with a more sophisticated molecular test, we can show that one-third or 40% of high-risk patients look more like intermediate risk patients. Then we may be able to decrease the utilization of multi-modality, with its concurrent increased morbidity, because the patients are unlikely to accrue a benefit.
Conversely, patients who are at intermediate risk, and without Prolaris, would perhaps be treated with single modality or radical and radiation. But, if a Prolaris [test] is done and the patient is shown to have higher risk features with a higher CCP score, then that intermediate patients may well accrue a benefit to receiving addition therapy added to a single modality.
Q: This seems to be a form of personalized medicine. I wonder then whether these gene signatures might somehow, someday, be linked to certain types of targeted therapies.
This is a very good question. In fact, we have a very large study going on with Jonathan Tward, at the Huntsman Cancer Institute, where we are looking at intermediate risk prostate cancer treated with external beam radiation therapy, where about half the doctors today in that patient group, will add hormonal ablation—androgen deprivation therapy. Half won’t, and there is really no process for making the decision. We are hoping, and our study is powered to show, that we may be able to identify which man—by conventional parameters—benefits from androgen deprivations therapy, and in which man we can avoid the side effects and the costs of hormonal ablation in the setting of external beam radiation therapy.
That would be the first time we may be able to see whether Prolaris can help to change our therapeutic approach.
But I think that more specifically, what you are getting at is that a lot of excellent papers are looking at selection of targeted therapy [based on] sophisticated molecular markers. There are a number of studies in place to [investigate the question] Can we more precisely identify which agent may be beneficial to Mr. Jones and less beneficial to Mr. Smith, based on some expression of a molecular marker.
That is the next wave, and in this sense, Prolaris in its current form is not likely to be a major addition, because the genes in Prolaris are so universally fundamental in cancer, they are going to be expressed across the board.
A really good example of where a difference can occur is that we’ve seen a paper in the New England Journal of Medicine by de Bono and his group,2 with the PARP inhibitor, olaparib, that expression of the mutated form of the BRCA gene very dramatically identified a subset of men with castrate-resistant prostate cancer who benefitted from this particular PARP inhibitor. This is an example of when a precise molecular marker and a BRCA inhibitor could change what therapeutic regimen a clinician and patient decide to choose. This is an exciting wave of the future.
Written By: Barbara Jones
1. Scardino P, Cuzik J, Stone S, et al. Application of active surveillance threshold to series of samples submitted for commercial testing. ASCO Genitourinary Cancers Symposium. January 7-7, 2016, San Francisco, CA. Available at:
http://www.nejm.org/doi/full/10.1056/NEJMoa1506859?af=R&rss=currentIssue
2. Mateo J, Carreira S, Sandhu S, et al. DNA-Repair Defects and Olaparib in Metastatic Prostate Cancer. N Engl J Med. 2015; 373:1697-1708.