NEW YORK – Two new studies are highlighting the utility of Exact Sciences' Oncotype DX Genomic Prostate Score (GPS) assay as an independent predictor of disease outcomes in two different populations: patients with unfavorable intermediate (UFI)-risk prostate cancer, and African American men.
In a study published this month in Urology, researchers from Case Western Reserve University, Genomic Health, Kaiser Permanente, and elsewhere described their analyses of 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. The analyses included 299 intermediate-risk prostate patients, 175 of whom had UFI-risk disease — 103 from the Kaiser Permanente Northern California cohort and 72 from the Center for Prostate Disease Research cohort.
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. UFI patients with a GPS score of greater than 40 had outcomes consistent with high-risk disease, whereas UFI patients with a GPS score of 40 or below 40 had outcomes similar to favorable intermediate-risk patients, the researchers said.
Oncotype DX GPS, a 17-gene panel, is currently recommended for men who've been diagnosed with clinically low-risk prostate cancer to determine whether their disease can be conservatively managed through active surveillance rather than treated with definitive surgery or radiation therapy.
According to the National Comprehensive Cancer Network's (NCCN) clinical practice guidelines on prostate cancer, men with low or favorable intermediate-risk disease can consider the use of the tumor-based molecular assays like Oncotype DX GPS for active surveillance.
The problem for men with UFI-risk disease is that until now, the lack of definitive knowledge about this particular subtype has made it almost impossible to opt for active surveillance, despite the fact that some of them end up with less aggressive disease.
"We had already shown GPS to be an effective tool to prognosticate bad outcomes in very low-, low-, and favorable intermediate-risk patients. But no one had focused specifically on unfavorable intermediate. And that's important because there is a gray zone, a very murky area, in knowing how aggressively to treat that disease, because we don't know if it's going to behave more like high-risk disease or more like favorable intermediate-risk disease," said Jennifer Cullen, the Urology study's lead author and associate director of cancer population sciences at the Case Comprehensive Cancer Center. "So, there was a gap of knowledge about what kind of treatment decision-making should take place with patients diagnosed with that type of risk straight on."
Splitting the scores at the 40-point level made the GPS tool very effective in the UFI population, which could help clinicians and patients in the treatment decision-making process, she added.
All of these patients in the study underwent surgery or a radical prostatectomy. The question would be whether they'd also undergo external beam radiation therapy or any irradiation modality, and/or hormonal therapy.
"You are always trying to strike a careful balance between preserving oncologic outcome and sparing patients decrements in quality of life," she added. "So, there's this very careful discussion and balance that patients and doctors have to strike in deciding what treatment course to pursue for men in that gray zone."
Though Cullen noted that the results of the Urology study are likely not immediately clinically actionable, there's cause for additional studies to examine these results in larger cohorts, as the data are definitive that the GPS score is a powerful indicator of how UFI-risk disease will behave.
Cullen also noted that the results of this study are somewhat analogous to the landmark TAILORx study, which showed how doctors should interpret and act upon an intermediate result using the Oncotype DX breast cancer risk test.
"The TAILORx study was a little bit larger and had more centers involved. This particular study focused on the [US Department of Defense] Center for Prostate Disease Research Cohort, as well as the Kaiser Permanente Medical Center in Northern California. So, there's not quite the scope of sample size as in TAILORx, but it's extremely analogous in that there is a gray zone of patients," Cullen said.
According to Exact Sciences Chief Medical Officer Steve Shak, the results of this study show that Oncotype can fit into the diagnostic toolbox for prostate cancer, even for subtypes that are harder to interpret.
"NCCN and ASCO both acknowledge the use of genomics within unfavorable intermediate-risk patients, so we're hopeful that this new publication will support inclusion of the Oncotype GPS test in guidelines for this patient population," he said. "We estimate the current total addressable market for Oncotype DX GPS to be about 135,000 men in the US, so adding this new patient population would increase that."
He further cited statistics from the American Cancer Society that indicate 43,000 UFI-risk patients are diagnosed in the US each year, out of about 192,000 prostate cancer cases overall.
Both Shak and Cullen noted, though, that GPS scoring wouldn't be a replacement for PSA testing, which can sometimes result in unnecessary biopsies. Typically, PSA testing is used pre-diagnosis, and after treatment to understand if the disease is responding to treatment.
"That will not be replaced by any genomic or marker test. The GPS had been meant to complement clinical parameters, not replace them," Cullen said. "We're still going to see physicians rely on PSA, stage, and grade. Those are the status quo variables that will always be collected and examined in understanding what treatments are best. But the GPS is complementary and provides other key information beyond those three traditional parameters because we know the PSA, stage, and grade alone don't do a great job at forecasting what's going to happen in five, 10, or 15 years."
The GPS score is also unlikely to be useful in high-risk patients, she added. When patients are observed to have high-risk disease at the time of diagnosis, there generally isn't any debate about whether the treatment needs to be aggressive.
"The gray zone is really with the intermediate patients and even with the low-risk patients to some extent, because there are some low-risk patients who will have future aggressive disease," Cullen said. "And so, what you really want GPS to do is to inform the gray zone."
Bigger, and More Diverse
One way to help validate the results of the Urology study will be to replicate them in a larger cohort. The Canary Foundation is currently undertaking the Prostate Cancer Active Surveillance Study (PASS), a large tissue microarray biomarker project that's designed to address challenges in prostate cancer, including studying the men who are candidates for active surveillance, Cullen said. The researchers of the Canary PASS study are collecting GPS data as part of their work, so she's hopeful the effort will help inform how GPS can help prognosticate future bad outcomes.
Cullen also believes that the utility of GPS scoring must be studied in other racial ethnic groups, specifically African Americans compared to Caucasians, and then expanded out to other groups such as Hispanics and Asians.
At the American Urological Association's 2020 Annual Meeting in May, a group of researchers presented a poster on Oncotype GPS as a predictor of outcomes in African Americans with prostate cancer. The researchers, including Cullen — who was at the Center for Prostate Disease Research at the Uniformed Services University — performed a composite analysis of six independent cohorts of men with newly diagnosed prostate cancer treated with radical prostatectomy, including more than 200 African-American men. The results showed that tumor biology as measured by the Oncotype GPS was similar in African American and Caucasian men, and that the assay was similarly predictive of in both groups.
"Race was not an independent predictor of bad outcome. GPS was. And that accounts for a robust number of African Americans and a robust number of biochemical recurrence events in the study," Cullen explained. "What we found collectively is that GPS was distributed similarly across race and GPS predicted bad outcome. Race did not."
Importantly, she added, this data highlights a point in the longstanding debate as to whether the higher incidence and higher mortality due to prostate cancer for African Americans compared to Caucasians is driven by biology or driven by social determinants.
"I think what we see, especially in the military cohorts and in the DoD program … is that when the patients are diagnosed and treated and followed within an equal access health care system like the military, we don't necessarily see differences in outcome," Cullen said. "So, there may still be biological differences in why they get cancer in the first place, but how they're treated and how early they're detected are key variables that also predict bad outcome."
Further, it points to the importance of a comprehensive diagnostic panel like GPS. Though the literature is always being updated with new information about single genes or single mutations that inform outcome, the cancer field is moving more toward an understanding that comprehensive assays are needed, according to Cullen.
"There's not going to just be one single genetic factor that is the penultimate variable that matters in predicting bad outcome," she added. "What GPS does so nicely is that genes were selected in an agnostic manner as to which ones would predict metastatic disease."
In the discovery work that was done to determine which genes should be included in the assay, she noted, the candidates were assessed in clusters. There's a cluster for cell proliferation, one for androgen signaling, one for apoptosis, and so on, to address every meaningful mechanistic area of cancer progression.
What this study showed is that while biology matters greatly in cancer outcomes, social determinants cannot be underestimated. "There are clearly biological determinants in aggressive cancers of many types, and biology matters," Cullen said. "But I also think what's underplayed are the social determinants. There's certainly a number of cancers where if you don't have good access and you don't have good follow-up, you're more likely to suffer poorer outcomes in the future post treatment."