NEW YORK (GenomeWeb) – Researchers have used GenomeDx Biosciences' data-sharing network Decipher GRID to identify a new gene expression signature that is predictive of prostate cancer patients' response to post-surgical radiation therapy.
A study describing the discovery appeared this month in the Lancet Oncology. A team led by investigators at the University of Michigan and the University of California, San Francisco showed that for patients with high expression of a certain pattern of genes, those treated with radiotherapy had a lower incidence of distant metastasis at 10 years than those who did not receive radiotherapy. In contrast, patients with low expression scores had similar outcomes regardless of whether they received radiation treatment or not.
This suggests that analyzing patients for this expression signature, dubbed PORTOS (Post-Operative Radiation Therapy Outcomes Score), could help more accurately pick out who should be considered for radiation treatment. Patients with high scores appear to benefit from radiation, while those with low scores don't appear to get the same added value.
PORTOS is the first predictor of its kind to emerge from GenomeDX's Decipher GRID (Genomics Resource Information Database), a database containing genomic information on patients who have received its 22-marker prognostic test Decipher, the company said in a statement.
GenomeDx launched Decipher GRID in May 2015. While Decipher itself only reports on the expression of 22 markers, it uses a 1.4 million-marker Affymetrix GeneChip. This means the company has access to a much broader range of tumor gene expression data on every tested patient. Decipher GRID allows Decipher users to access this broader data alongside other information on patients' tumors and outcomes.
All physicians who order a Decipher test can access the Decipher GRID data for their own patient on a research-use basis. Those who want to use more of the resource must sign user agreements, but Genome Dx President and chief scientific officer Elai Davicioni told GenomeWeb in an email that access is free for all non-commercial investigations.
GRID-based investigations have included a study published last year in European Urology in which University of Michigan researchers analyzed Affymetrix GeneChip expression profiles from more than 1,300 patients who had been tested using Decipher to identify three distinct molecular subtypes.
Other work is ongoing to identify markers and signatures predicting not just response to radiation as in PORTOS, but also hormone therapy and chemotherapy.
As of this June, GenomeDx said that investigators from 55 cancer centers had used the resource to conduct and publish studies, and that the database contained profiles from more than 10,000 patients, clocking 900 new profiles per month. However, none of these studies has yet yielded validation of a novel signature with clinical applications like PORTOS.
Davicioni said in his email that the PORTOS research group was one of the first to explore the use of the fuller breadth of genomic information collected on the Decipher array platform.
In the Lancet Oncology study on PORTOS this month, the researchers used Decipher GRID to develop and test a 24-gene expression signature, demonstrating that patients with high PORTOS scores who were treated with radiotherapy had a lower incidence of distant metastasis at 10 years than those who had high scores but were not.
The team first conducted a retrospective analysis using a training cohort of 196 prostate cancer patients who received radical prostatectomy with or without postoperative radiotherapy to define the PORTOS signature. They designed the set of targets, focusing on genes that are related to radiation or DNA damage response.
After training the signature in the initial set of retrospective samples, the team then validated it in a multi-institutional cohort of 330 patients.
In the validation cohort there were 82 patients with a high PORTOS score. The data showed that those who received radiation had a significantly lower incidence of 10-year distant metastases compared with those who did not. Four percent of the radiation-treated group had 10-year recurrences compared to 35 percent in the no-radiation group.
In contrast, among PORTOS-low patients (248 in total), 32 percent of both the radiation-treated and non-treated patients went on to have 10-year distant metastases.
Meanwhile, neither Decipher itself, nor other measures like CAPRA-S and a microarray version of the cell cycle progression signature, could predict response to radiotherapy with similar significance.
The study authors concluded that the results suggest PORTOS could be used to pick out a subgroup of prostate cancer patients who are likely to benefit from postoperative radiation.
The data also suggest the test could help reduce unnecessary toxicity in those who would not benefit significantly from treatment. In other words, since patients with a low score recurred at the same rates regardless of treatment, they potentially should avoid radiation since it offers no overall benefit.
However, the researchers wrote, the signature should be investigated further in additional independent cohorts.
According to GenomeDX, any physician who orders a Decipher test receives a report based on Decipher GRID's larger set of RNA expression markers. Therefore, physicians that access Decipher GRID can currently evaluate their patient's raw PORTOS score based on the model reported in the Lancet Oncology paper, as well as the percentile ranking of that PORTOS score in the overall GRID population, Davicioni said in his email.
The PORTOS researchers don't currently hold rights to develop their own standalone test, he added, but GenomeDx does plan to work toward offering PORTOS more directly as part of its clinical Decipher test.
"Our plan is to eventually offer PORTOS as part of the Decipher test report. As described in the paper, Decipher provides the metastatic potential of the tumor which is used to determine whether or not (and when) additional therapy after surgery such as radiotherapy is necessary. PORTOS further helps determine whether or not the tumor will be responsive to postoperative radiation," he added.
"Ultimately, we envision the field moving towards a biomarker-driven approach whereby prognostic signatures identified through the use of a genomic test, such as Decipher, are used to select patients with aggressive disease, and predictive signatures, such as PORTOS, are used to select specific therapies," GenomeDx CEO Doug Dolginow said in a statement.