Test for Postoperative Radiation Response in Prostate Cancer

Alexander M. Castellino, PhD

October 27, 2016

For the first time in prostate cancer research, a 24-gene molecular signature can predict response to radiation therapy following radical prostatectomy, say the developers.

The 24-gene signature, called PORTOS (Post-Operative Radiation Therapy Outcomes Score) was developed by a team of academic researchers led by Felix Y. Feng, MD, University of California, San Francisco, working in collaboration with GenomeDx Biosciences (headquartered in Vancouver, British Columbia, Canada).

PORTOS is able to predict which patients are likely to respond to postoperative radiation with a significantly reduced risk for metastatic recurrence at 10 years, the team says. At present, it is available to clinicians who order the company's Decipher test, which is a prognostic test for prostate cancer.

The team used a training set to develop the signature in a retrospective cohort of patients. They then validated the test in an independent cohort of patients in a blinded manner. Details were published online October 12 in Lancet Oncology.

"This work is a milestone because the gene-expression based PORTOS is the first to provide clinicians with a molecular signature to predict response to postoperative radiotherapy in prostate cancer," Michel Bolla, MD, of Grenoble University Hospital, France, writes in an accompanying commentary.

Current Practice, Future Trend?

In an interview with Medscape Medical News, Dr Feng explained that the tools that are currently used in clinical practice, such as the postsurgical Cancer Prostate Risk Assessment Score, based on clinical and pathologic parameters, are only prognostic.

"Most tools tell us which cancers are aggressive or nonaggressive but do not predict response to a specific treatment," he said.

As more high-risk prostate cancer patients undergo prostatectomy, there is a greater need to tailor adjuvant and salvage therapies, he explained. "Among the different adjuvant or salvage treatment options, postoperative radiotherapy offers the greatest opportunity, to date, to increase cure," he said.

However, the use of adjuvant radiation therapy is declining, Dr Feng noted, although data from three randomized trials (SWOG 8794, EORTC 22911, and ARO-9602) demonstrate that this treatment approach reduces biochemical relapse after surgery.

"We need better tools to determine who should receive radiation after surgery. This 24-gene molecular signature did just that and performed well," Dr Feng said.

"The 24-gene signature has the potential to personalize patient selection for postoperative radiotherapy and thus could improve outcomes and spare unnecessary toxicity in men with prostate cancer," he said.

Dr Feng indicated that every clinician would have to decide whether to wait for further validation. "There will be early and late adopters," he said.

He also pointed out that the Oncotype DX test for breast cancer was validated retrospectively from samples from clinical trials, as was this 24-gene signature. "But it was used in clinical practice long before it was prospectively validated," he said.

But is the 24-gene signature ready for clinical practice? Some say yes, others disagree.

"This promising signature should be regarded as translational research. It requires validation in phase 3 randomized trials investigating adjuvant or salvage radiotherapy," Dr Bolla comments in his editorial.

Medscape Medical News contacted other authors associated with the study and received mixed opinions.

"It will affect some of our decision making, and I view it as an additional tool in the armamentarium of oncologists," said Adam P. Dicker, MD, PhD, chair and professor of radiation oncology at the Sidney Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, Pennsylvania.

Robert Den, MD, from the Department of Radiation Oncology, Sidney Kimmel Medical College, regards the findings as "quite provocative."

He is in the same camp as Dr Bolla and would ideally like to see this validated. Although he will get this information with the Decipher GRID report, he does not anticipate the information altering his clinical practice.

A press release from GenomeDx Bioscience explains that the PORTOS signature is currently available on the Research Use Only GRID Tumor RNA Expression Profile, which is available to physicians who have ordered the Decipher Prostate Cancer Classifier test.

"This is a very exciting study and demonstrates a true predictive marker for radiation response. I would like to see more validation in either a retrospective or prospective fashion prior to its routine use in clinical decision making," Ashley E. Ross, MD, from Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins School of Medicine, Baltimore, Maryland, told Medscape Medical News. He believes the independent validation will come during the next year.

"This study has not yet changed management, as it is new and we are still fully evaluating it, but I think this will have a major impact in how we manage our patients," Dr Ross said.

Stephen Freedland, MD, is not awaiting further validation.

Dr Freedland is editor-in-chief of the journal Prostate Cancer and Prostatic Diseases, professor of surgery (urology), and holds the Warschaw Robertson Law Families Chair in Prostate Cancer at the Cedars-Sinai Medical Center, Los Angeles.

"Further validation is always ideal and reassuring. However, we need to live in the present and must treat patients today while we wait. As this study included both a training cohort and validation set, it is already validated," Dr Freedland said. "I do think we can use this today to help counsel patients," he added.

"When a physician orders Decipher, PORTOS is included at no charge to the physician. As such, whether you want it or not (and I do want it), you can get it," he said.

Dr Freeland explained that the 24-gene signature has not yet changed the protocol at his institution, but it is part of his decision making.

Many patients are hesitant to undergo radiation therapy after surgery, owing to concerns about side effects and the current inability to determine exactly where the tumor is, he explained. "Thus, while we radiate where the prostate used to be, we aren't guaranteed that is the only spot where the tumor is," he said.

Dr Freedland explained that for men with a high Decipher score (which would indicate a high risk for metastases) but who also have a high PORTOS (which would indicate a likely response radiation therapy), the argument in favor of radiation becomes quite compelling.

For men with a low Decipher score (indicating a low risk for metastases) and/or a low PORTOS (indicating that a response to radiation therapy is unlikely), radiation becomes a harder sell and may not make sense, he further elucidated.

"Thus, I do think this paper results in actionable information that we can — and are starting — to use in the clinic today," Dr Freedland told Medscape Medical News.

Medscape Medical News attempted to contact GenomeDx Biosciences through its website and its public relations firm to determine how it intends to roll out the test, but did not get a response.

Anthony V. D'Amico, MD, PhD, professor of radiation oncology at Harvard Medical School and chief of genitourinary radiation oncology at the Brigham and Women's Hospital and Dana-Farber Cancer Institute, Boston, Massachusetts, who was not associated with the study, had this to say.

"There are many men where clinical factors alone (eg, prostatectomy stage, grade, and margin status) puts them at very high or very low risk for metastasis after surgery and in whom additional testing and information is not needed," he told Medscape Medical News.

"However, there remains an intermediate-risk group where this test may have clinical utility," he added.

Dr D'Amico noted that to date, this test has not been used at his institute. However, he said that it could now be used as a prerandomization stratification factor in randomized studies evaluating the impact on time to metastasis of novel postoperative adjuvant treatment in comparison with standard of care in men with prostate cancer.

"Such an approach could provide the information needed on how to use this test wisely in clinical practice," he said.

Details of the Training Set

For the training set, researchers used an exact matching (1:1) of patients from the Mayo Clinic I cohort (1987-2001). Matching was conducted on the basis of Gleason score, preoperative prostate specific antigen level, surgical margin status, extracapsular extension, seminal vesicle invasion, lymph node invasion, and androgen deprivation therapy.

Clinical outcome was known for each of the 196 patients whose samples were used in the training set (98 received radiation therapy; 98 did not receive radiation therapy).

To develop the 24-gene signature, the researchers used archived formalin-fixed tissue samples to isolate RNA and used an Affymetrix exon-array platform that consisted of 5.5 million probes that could interrogate 1.4 million RNA transcripts.

The assay was performed in a Clinical Laboratory Improvement Amendments–certified laboratory by GenomeDx Biosciences.

Of the 1800 unique DNA repair and radiation response genes available on the microarray platform, the researchers were able to rank each gene as to how it performed with respect to treatment, as determined from a "univariate p-interaction value."

They used the ranked genes to develop a model with distant metastasis as the endpoint; the variables were treatment and "the interaction terms of radiotherapy and expression of each gene."

Among the 93 patients in the training cohort who experienced a metastatic event, they identified the 24-gene PORTOS predictive molecular signature.

In the training set (n = 196), among patients with high PORTOS (n = 39), 10-year metastatic rate was 5% if the patients received radiation and 63% if they did not (hazard ratio [HR], 0.12; 95% confidence interval [CI], 0.03 - 0.41; P < 0.0001). Among patients with low PORTOS (n = 157), the 5-year metastatic rate was higher if patients received radiation therapy (57% vs 32% for no radiation therapy; HR, 2.5; 95% CI, 1.6 - 4.1; P < .0001).

With a PORTOS of 0 representing no difference in outcomes, a high PORTOS (>0) indicated that radiation therapy was beneficial, and a low PORTOS (≤0) indicated no benefit.

Although the training set was small, it included patients with more adverse clinical and pathologic features than the validation cohort, the researchers explain. "The high metastatic event rate allowed us to train with a smaller number of patients," they indicate.

Details of the Validation Cohort

The validation cohort (n = 330) came from the Mayo Clinic (2000-2006, different from the training cohort), Johns Hopkins University (1992-2010), Thomas Jefferson University (1999-2009), and Durham VA Medical Center (1991-2010).

Patient outcomes were blinded to the researchers. Patients with a high PORTOS showed a 10-year metastatic rate of 4% if they received radiation therapy; it was 35% for those who did not (HR, 0.15; 95% CI, 0.04 - 0.60; P = .0020).

For patients with a low PORTOS, the 10-year metastatic rate was 32% regardless of whether or not they received radiation.

The researchers indicate that the validation cohort is not a perfect match with the training cohort and emphasized that that is the reason why the model may be generalizable.

"Validation of PORTOS suggests that it will be robust despite the limitations of the data," the study authors write. "We are working on acquiring additional cohorts, which will allow us to increase the patient numbers for further investigation," they add. Potential validation cohorts could be provided by future trials, such as RADICALS and RTOG 9601, they note.

The researchers also note that the population of patients who received radiation therapy was pooled, and the study does not distinguish between use of adjuvant and salvage radiation therapy. However, the combination of PORTOS and Decipher tests allows identification of patients who may benefit from radiation therapy in the adjuvant or salvage setting, the authors explain.

"As we try to personalize therapy, we need to use genomic tools that are designed to perform well in the clinical context in which they were designed," Dr Feng told Medscape Medical News.

"PORTOS has the potential to identify patients with prostate cancer who may benefit from postoperative radiation, while Decipher is a prognostic signature, which identifies those patients at risk of metastases independent of treatment," he explained.

"Validation of PORTOS should be extended to the low-risk and intermediate-risk, localised prostate cancers, which are increasingly frequently diagnosed by needle biopsy during individual screening," Dr Bolla commented in the editorial.

"If these patients are treated by definitive external beam radiotherapy, the challenge will be to select which patients to offer definitive image-guided radiotherapy with dose escalation to protect their quality of life and sexual health; will PORTOS be of use in this situation?" he asks.

The study was carried out in collaboration with GenomeDx Biosciences, which commercialized Decipher, the prognostic test for prostate cancer. Most authors receive grants and/or personal fees from industry. Some authors are employees of GenomeDx Biosciences. Several authors have a provisional patent filing for PORTOS pending with the United States Patent and Trademark Office. Dr Bolla has disclosed no relevant financial relationships.

Lancet Oncol. Published online October 12, 2016. Abstract, Commentary

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....