NEW YORK – In tracking the utility of Exact Sciences' Oncotype DX Genomic Prostate Score test for patients with localized prostate cancer over the course of 20 years, a recent study has helped link the biology and the histology of the disease more closely, perhaps making it easier for clinicians and patients to make decisions on surveillance versus treatment, rather than having to rely on clinical indicators or gene expression alone.
The researchers found that the test provides an accurate assessment of the risk of long-term distant metastases, or DM, and prostate cancer-specific mortality, or PCSM, for patients with localized prostate cancer. If taken into account with data from other published and as-yet unpublished studies, these 20-year findings seem to indicate that the Genomic Prostate Score, or GPS, test can be used as an indicator of long-term risk in a dichotomous manner.
This data could also be used to eventually help Exact seek an expansion of the established guidelines as to which prostate cancer patients are best suited for diagnosis with Oncotype GPS.
In the study, which was published in JCO Precision Oncology, researchers at the Cleveland Clinic, Exact Sciences, and elsewhere analyzed the association between an Oncotype DX GPS result and long-term oncological outcomes following radical prostatectomy in a stratified cohort sample of 428 patients who were treated between 1987 and 2004.
Specifically, they evaluated the association of GPS results calculated from radical prostatectomy tissue with DM and PCSM risk over the 20 years following the procedures and found that GPS scores were meaningful indicators of long-term outcomes beyond clinical factors alone.
The 17-gene panel Oncotype DX GPS is currently recommended by the National Comprehensive Cancer Network's, or NCCN, clinical practice guidelines for men who've been diagnosed with very low-, low-, or intermediate-risk disease, to determine whether their disease can be conservatively managed through active surveillance rather than treated with definitive surgery or radiation therapy.
In the study, the researchers noted that the disease course for localized prostate cancer can be extended and the long-term risks unclear at first, making immediate treatment decisions difficult. Further, they said, calculating the risks based on clinical variables alone is imperfect and can lead to over- or undertreatment.
According to Eric Klein, chairman of the Glickman Urological and Kidney Institute at Cleveland Clinic and the study's lead author, prostate cancer specialists have learned that although Oncotype GPS was designed for active surveillance, the "sweet spot" for its use is in patients with intermediate Gleason scores and favorable tumor histology, or a lower Gleason score but with a higher-volume tumor.
He specifically highlighted one finding from the study that he believes will have the most immediate clinical impact for prostate cancer patients: "If you look at the graphs, there's an inflection point, right around the score of 29 or 30," Klein said. "In the past, we used the test to try and estimate relative risk for adverse pathology or some other bad outcome on a sliding scale. I look at this data and I say, 'Gee, if your score is below 29, you have a really small chance of developing metastatic disease or dying of prostate cancer."
Indeed, the researchers concluded that not only were GPS scores highly associated with risk of DM and PCSM, but that absolute risks of these outcomes showed a low-slope relationship with a GPS score from 1 to 29, with an inflection point evident at a score of 30 or higher. When the score went above 30, the risk of DM or PCSM increased substantially, reaching an absolute risk of 38 percent for DM and 13 percent for PCSM at a score of 60, they said. Further, they found that including the GPS results in models assessing risk of DM and PCSM at 20 years improved discrimination compared with using clinical variables alone.
According to Klein, this shows that a man with a score of 29 or under could be a good candidate for surveillance, whereas a higher score could indicate a molecularly aggressive tumor that should be more closely monitored. The exact clinical decisions after that would depend on a physician's philosophy of treatment, he added. Some may think a patient with a high score should always receive treatment while others may believe in leaving the patient under a closer surveillance regimen. "Eventually, I think we'll get to the point where we recognize that those patients [with high scores], if they're young and healthy and have a long life expectancy, generally can be treated," Klein said.
What's most exciting about the GPS test to Klein — and about the 20-year data, which proves the test's utility in localized disease — is that it provides independent information that isn't evident to a pathologist who might be looking at a biopsy through a microscope. "It provides additional evidence that you can take this sort of [genomic] information and add it to standard clinical parameters that we make decisions on, including Gleason grade, and make a more accurate prediction about whether or not the tumor is likely to cause harm to the patient," Klein said. "So, I think it's a new niche in this space and it should be used in a complementary fashion with all the other data we have."
Specifically, he noted, Oncotype GPS helps the clinician go beyond the histology of a patient's disease and into the biology of it. Histology in prostate cancer is useful to a degree. For certain patients, histological profiling and PSA testing are clear-cut and genomic testing or gene expression profiling won't be necessary. But for others, the Gleason score is more ambiguous, and knowing the biology of the tumor is necessary to treating the patient. "The pathologist can't see whether or not there's a genomic instability, but this kind of test can," Klein said. "If you think about the progression of the cancer, you get changes at the genomic and the gene expression level first, and then you get histologic changes. And then you get tumors that appear on an MRI, and then you may get a PSA elevation. This [advances] our ability to see what's going on with the tumor."
Most importantly, however, other studies would seem to show that Oncotype GPS scores of 29 or 30 correlate with prostate cancer histologies that have been shown to be associated with worse outcomes. One such histology is called cribiform histology. Based on previous genomic studies that have been done on prostate cancer, cribiform histology looks and behaves like high-grade cancer, Klein said. In active surveillance cohorts, it has been shown to be associated with the worst outcomes. In a study Klein and his colleagues recently did, which has been submitted for publication, there were no patients with cribiform histology who had a GPS score below 29.
"That's kind of an independent observation around this number, around 29 or 30, that really indicates some biologic potential," he said.
Further, Klein pointed to a study published by researchers at the University of California, San Francisco and their collaborators in the American Journal of Surgical Pathology showing that stromogenic histology was also associated with bad outcomes. He also noted that it takes a lot of effort by a trained genitourinary pathologist using special stains to find stromogenic histology. Given that these histological features are subjective, and that molecular testing is both objective and reproducible, use of a molecular profiling test as an adjunct to standard histopathology assessment could lead to improved patient prognostication, the UCSF researchers concluded.
They found that both cribiform and stromogenic histology correlated with GPS scores, Klein explained. And because the majority of prostate cancer biopsies are read by nonexpert pathologists who don't report cribiform or stromogenic histology, “an underappreciated aspect of these [GPS] tests is that they measure biology that's either not apparent to the pathologist or that the pathologist isn't trained to report," he added.
While the histology itself is important, he added, it's not completely informative of the biology of the tumor. That's where gene expression profiling gives the clinician an added advantage. And for Klein, it's also validating to have the additional studies that independently confirm that Oncotype GPS scores of 29 or 30 can serve as an inflection point for clinicians wavering on whether to leave patients on surveillance or move on to active treatment.
It's especially helpful to have such long-term data, he added, to show the real-world value that such tests can have for patients.
"At the beginning, when these tests were introduced, there was a lack of understanding of what they were useful for," Klein said. "And so, the point of this long-term study was to show that not only do they predict and are validated to predict for adverse pathology, but that has real meaning in terms of what happens to the patient in the long term, with hard endpoints that no one could argue with. No one would argue that avoiding metastases and that avoiding mortality from prostate cancer is desirable. So, I think it solidifies that we're measuring real biology."
Exact Sciences' Matt Franklin, general manager for the company's precision oncology business, concurred. In an email to GenomeWeb, he noted that the 20-year outcomes study added to published clinical utility evidence on the test that has been collected in more than 5,000 men so far. The company is hoping to build "awareness of the GPS test and the significant value it can provide in informing treatment decisions" and will be "actively leveraging this published data in our interactions with the urology community," he added.
Franklin also pointed to a study published in Urology last June that showed Oncotype GPS could be used as an independent predictor of disease outcomes in patients with unfavorable intermediate-risk prostate cancer. In that study, researchers from Case Western Reserve University, Genomic Health, Kaiser Permanente, and elsewhere analyzed associations of GPS results with biochemical recurrence, distant metastases, and prostate-specific death in two cohorts of men with UFI prostate cancer who were treated with radical prostatectomy, and they found that a dichotomous GPS score of above 40 or 40 and below was a significant predictor of biochemical recurrence in UFI patients, and that GPS was a strong predictor of all three endpoints.
According to Franklin, Exact is leveraging recent published data, including the Urology study, to petition the NCCN to update its guideline to include use of Oncotype GPS test in unfavorable intermediate-risk prostate cancer patients. Further, the company is anticipating publication later this year of Oncotype GPS data in high-risk patients with localized prostate cancer and may seek to expand use of the Oncotype under the guidelines in those patients as well.