A team of researchers has identified a gene signature that could be used to predict a prostate cancer patient's metastasis-free survival at the time of diagnosis.
Using genomic information obtained from tumor samples via Affymetrix's OncoScan FFPE Express service paired with an internally designed algorithm, the researchers claim they can provide clinicians with a metastatic potential score to determine whether aggressive treatment of prostate cancer is necessary.
The approach was described recently in the Journal of Probability and Statistics.
Lead author Alexander Pearlman told BioArray News that the team's ultimate vision is seeing the prognostic signature through clinical adoption as a US Food and Drug Administration-cleared in vitro diagnostic.
"The finish line is pretty clear; it is to develop an FDA-approved kit for the purpose that is highlighted in the paper to predict prostate cancer metastases and to alleviate men of needless surgeries and potential radiation treatment," said Pearlman. While he noted the "hurdles to getting there are still huge," he said that he and his colleagues are "determined" to reach that goal.
According to Pearlman, the study is now being sponsored by Albert Einstein College of Medicine, where he is an instructor. He said that the project started eight years ago when he was a graduate student in Harry Ostrer's laboratory at the New York University School of Medicine. Ostrer, who was director of NYU's human genetics program, has since become a professor of pathology and genetics at AECOM.
Ostrer is a co-author on the new JPS paper, as are investigators from Baylor College of Medicine, Johns Hopkins University School of Medicine, and the Stanford University School of Medicine.
In the paper, the authors described the development of a semi-supervised clustering method using tumor genomic DNA copy number alterations to classify patients into inferred clinical outcome groups of metastatic potential.
The study's data set consisted of 294 primary tumors and 49 metastases from five independent cohorts of prostate cancer patients, according to the paper, and the alterations were modeled based on Darwin's evolutionary selection theory. Genes overlapping these altered genomic regions were used to develop the metastatic potential score for a prostate cancer primary tumor.
The authors then used a Cox proportional hazards model to evaluate the metastatic potential score alongside other clinical predictors available at diagnosis, such as PSA status, tumor grade, and the like. They claim in the paper that their proposed score was the "only significant predictor of metastasis-free survival." Based on these findings, they believe that the metastasis gene signature and associated score could be applied directly to copy number alteration profiles from patient biopsies positive for prostate cancer.
"This predictor might be important for correctly categorizing men at the time of diagnosis and could predict whether surgery, radiation therapy, or watchful waiting was warranted," the authors wrote in the paper. They added that the method could be applied to other cancers, such as breast cancer, that "exhibit variation in the metastatic potential of the primary tumor and have similar difficulties in collecting tumor samples with long-term clinical outcome data."
Pearlman referred to the JPS study as the "methods paper" for the clinical diagnostic the team is developing. He said that he and colleagues are looking at regulatory issues and discussing with the FDA how to develop the signature into a test. They are also working with collaborators, software developers, and other institutions on collecting new cohorts to be used in additional studies.
NYU has filed a patent application describing the signature with the US Patent and Trademark Office and is "very interested in seeing it through," said Pearlman. He added that Albert Einstein would become partial owners of the resulting IP should the university's application be approved.
The commercialization path for the test is still unclear. Pearlman said that he, Ostrer, and other colleagues have a "vested interest in moving this test forward," and have looked into licensing the signature to a larger diagnostics company, or working with a smaller company where the investigators might play a larger role.
The test is currently run on Affy's microarray platform. Pearlman said that the team selected the OncoScan assay because it is the "most robust" for dealing with archival formalin-fixed, paraffin-embedded tumors. "It is capable of obtaining data off of about 75 nanograms of starting material," said Pearlman. Though he said that Affy currently provides the "best platform" for looking at such material, Pearlman said the researchers are "not married to any technology." That being said, they see working with OncoScan as the "easiest way forward."
Affy launched the latest version of OncoScan via its research services laboratory last year. The 335,000-marker, whole-genome array offers high-resolution coverage of cancer-relevant genomic regions, as well as integrated analysis of whole-genome copy number changes and copy-neutral loss of heterozygosity (BAN 4/5/2011).
According to an Affy spokesperson, a new version of the OncoScan product is in development, and the company intends to launch it in the second half of 2013. The spokeperson told BioArray News that the Santa Clara, Calif.-based company is "working closely" with Ostrer and Pearlman, and noted that Pearlman will present data at Affy's seminar at the American Society of Human Genetics meeting in San Francisco next month.
In terms of the platform's use as a future diagnostic, the spokesperson commented that the company expects OncoScan will be used for signature development for cancer as a primary application.