乌鸦传媒

CHICAGO -- A prostate cancer gene test accurately identified patients who would ultimately develop distant metastases, die of prostate cancer, or die prematurely of any cause, a biopsy-based analysis showed.

When applied to archived pretreatment tumor specimens obtained up to 29 years ago, the genomic classifier (GC) showed that a higher score was associated with a 24% higher risk of distant metastasis, a 27% higher risk of dying of prostate cancer, and a 43% higher risk of death from any cause.

The study showed a 10-year distant metastasis rate of 29% among patients with high or intermediate GC scores as compared with 13% for patients who had low GC scores.

The results with the 22-gene test were consistent with an earlier validation study involving radical prostatectomy specimens. Collectively, the data suggest that the GC can help guide treatment decisions for newly diagnosed high-risk prostate cancer, reported Paul L. Nguyen, MD, of Brigham and Women's Hospital and Dana-Farber Cancer Institute in Boston, during the American Society for Radiation Oncology (ASTRO) meeting.

"High-risk prostate cancer accounts for two-thirds of the deaths from localized prostate cancer," Nguyen said in a recorded presentation. "Currently, the radiation management is a 'one-size-fits-all' scheme of radiation plus long-term androgen deprivation therapy. The use of genomic testing to stratify patients into cohorts with higher and lower risk of metastasis could allow for personalization of therapy, which is where we eventually want to go."

In an effort to take a step toward personalization, investigators evaluated the Decipher-22 GC, which assesses protein expression by 22 genes found in prostate tumors. The resulting expression signature is translated into a score, and a higher score reflects a more aggressive tumor, associated with worse outcomes.

Nguyen and colleagues applied the GC to stored tumor samples from three cooperative-group trials initiated as far back as 1992. The primary endpoint was the GC's prognostic performance for distant metastasis by multivariable analysis. Key secondary endpoints were prostate cancer-specific mortality (PCSM) and overall survival (OS). The analysis included GC results for 385 prostate biopsy samples, 265 of which met quality control criteria and had a median follow-up of 11 years.

By univariate analysis, every 0.1-unit increase in the GC score was associated with a 30% increase in the risk of distant metastasis, a 31% increase in the risk of PCSM, and a 16% increase in the survival hazard, all of which were statistically significant. Separate analyses of each of the three clinical studies yielded statistically significant increases in the hazard ratios for all three outcomes.

In the multivariable analysis, GC score remained statistically significant as a prognostic factor for distant metastasis, PCSM, and OS (P<0.001). Higher Gleason grade score was significantly prognostic for distant metastasis (P=0.002), but not PCSM or OS. Patient age, higher prostate-specific antigen value, and tumor stage were not prognostically significant for any of the outcomes.

"This is the first validation of any gene expression biomarker on pretreatment biopsy samples from prospective randomized trials," said Nguyen. "It demonstrates an independent association of the GC score with distant metastasis, prostate cancer-specific mortality, and overall survival. We know that high-risk prostate cancer is a heterogeneous disease, and the GC score can improve risk stratification to help personalize shared decision making."

The ongoing trial is investigating the GC score as an aid in optimizing treatment for men with newly diagnosed high-risk prostate cancer, he added.

GCs are the "future of our field for personalization of radiation therapy for prostate cancer patients and this is extremely exciting," said ASTRO invited discussant Sophia Kamran, MD, of Massachusetts General Hospital Cancer Center in Boston. "This study is basically showing that we can use a genomic classifier at the point of care ... and we can potentially use the genomic classifier to understand who we need to intensify treatment for and who we can de-escalate."

The NRG-GU009/PREDICT-RT trial affords an opportunity to "get additional answers and understand how we can move this into the clinic and how we can apply this now to our patients who are very much awaiting personalization of radiation therapy," she added. "We really need to further understand [how the GC works]. Genomic classifiers are here and they're coming to the clinic in due time."

All of the patients were classified as high risk when they were enrolled in the three trials, but as the results demonstrated, not all of the patients had high GC scores, said Rahul Tendulkar, MD, of the Cleveland Clinic, during a session highlighting key ASTRO presentations in genitourinary cancers. The prognostic value of the GC also held up in patients who received long-term or short-term androgen deprivation therapy, in addition to radiation therapy.

"This is the first validation of any gene expression biomarker in pretreatment biopsy samples," Tendulkar noted.

Comment