Multi Gene Predictive Signature to Discriminate Aggressive from Indolent Prostate Cancer May Allow Men to Avoid Unnecessary Treatments
September 19, 2012 -- After men undergo a needle biopsy of their prostate gland and receive a positive diagnosis of cancer, they often face a difficult question: should they consent to aggressive therapy, with surgery or radiation, both of which may cause serious side effects, or should they opt for a program of active surveillance that monitors men over time for signs of cancer progression. Doctors know that the majority prostate cancers are slow-growing and non-lethal, but pinpointing exactly which cancers fall into this category has been historically difficult. Thus many men and their physicians err on the side of caution.
Now, however, a new molecular test may prove useful in predicting which men may have increased confidence that they are eligible for active surveillance and which men have more aggressive disease and should consider immediate treatment.
Genomic Health, Inc., a cancer diagnostics company based in northern California that specializes in discovery and development of genomic markers are used to personalize medical care for cancer patients, announced topline results today from a study conducted with UCSF that demonstrates that a multi-gene signature can be used at time of initial biopsy of the prostate to predict the probability for high-risk cancers—even when initial biopsy shows low risk cancer.
This finding has broad and promising application, not only for men who are to test “positive” for the high-risk multi-gene signature, and for whom aggressive treatment may be the best option; but also, this may allow men who test “negative” for the high-risk multi-gene signature to more confidently opt for a program of active surveillance of their disease, deferring surgery and radiations and potentially avoiding side effects that can be attributable to those treatments such a incontinence or erectile dysfunction.
The new prostate cancer-specific multi-gene signature is a measurement of the amount of RNA expressed by various genes—with some genes making large amounts of RNA and others making little.
The process of finding this predictive “signature” began with a series of studies done at the Cleveland Clinic that looked at the cancerous tissue taken from some 700 men whose prostates had been surgically removed. The men were followed over time. In men who went on to develop metastatic prostate disease, or advanced cancers, a pattern of gene signature expression became apparent where certain genes were either predictably up-regulated, making excess RNA, or down-regulated, making too little RNA. That work was presented this past spring in Chicago at the annual ASCO meeting with Dr. Eric Klein of the Cleveland Clinic as the lead author.
“Rather than focusing on differences in how genes are expressed in different areas of a patient’s tumor, we searched for genes whose expression could provide clinically relevant information regardless of which area of tumor was sampled,” says Dr. Mark Lee, vice president of Oncology Development at Genomic Health and collaborator with Dr. Klein’s group on the studies that finalized the predictive gene signature.
The original studies winnowed down a list of 700 candidate genes to a predictive “signature” of less than 20 genes that could easily be tested in a large clinical validation study. In addition to precise gene expression patterns, the Genomic Health and Cleveland Clinic researchers identified a limited number of biological pathways that are prominently involved in their gene signature. A final, clinical grade assay based on this signature was then tested in a large clinical validation study conducted at UCSF in patients undergoing initial biopsy.
The patients in the UCSF clinical trial all had what looked to be low-risk prostate cancers at initial biopsy, says Dr. Matt Cooperberg, an investigator on the trial along with lead investigator, Dr. Peter Carroll, also at UCSF. Those patients opted to have surgeries to remove their prostates within six months of initial biopsy. When post-surgical pathology was compared to biopsy results, the researchers found that patients who turned out to have higher risk tumors also tended to test “positive” for the high-risk gene signature.
“In these cases where we showed an increased grade or stage of tumor at the time of surgery, these were probably cases not where the cancer had progressed in such a short period of time, but that higher-risk cancer was always present but we’d missed it at the time of biopsy,” says Cooperberg.
Because a needle biopsy--even when taking up to 12 different tissue samples from various areas of the prostate--only measures less than 1 percent of the entire gland the potential to miss areas of high risk tumor is great.
However says Dr. Lee, when you can test for a gene signature that is not dependent on what you sample from the gland, even in areas that under the microscope look to be low risk, you substantially increase clinicians’ ability to more accurately assess disease aggressiveness. “This test has the potential to address the major challenge of sampling error that occurs with biopsies of the prostate,” says Lee.
All the RNAs in the aggregate gene signature are coding RNAs, meaning they build proteins. Thus, eventually, apart from a predictive signature to guide clinical practice, these genes could become druggable targets for anti-cancer interventions.
In the here and now, says Dr. Howard Soule, chief science officer of the Prostate Cancer Foundation, this new signature validation study has the potential to address a critical treatment decision—whether or not to opt for active surveillance when low grade cancer is found at initial biopsy—faced by hundreds of thousands of men each year around the world.
Cooperberg is also optimistic saying that this signature should give clinicians like him better confidence in their risk stratification. “With standard clinical information we are now about 75 percent accurate,” says Cooperberg, “but while that’s not bad, it is not good enough and this should build on our accuracy.”
At the 30-thousand foot perspective, says Dr. Jonathan W. Simons, CEO of the Prostate Cancer Foundation, is the fact that predictive gene signatures are an outcropping of the enormous accomplishment of the Human Genome Project that catalogued all of our 28,000 genes. “Gene signature tests are multiplexed informatics; no single gene can give us enough data to do this, yet multiple gene patterns can encapsulate and group data in the manner of bar codes that give manufacture date, shipping date, price, expiration dates, etc., that together add up to a neat package of usable systems information,” says Simons. In the world of patient care, such information can translate to countless surgeries and side effects spared.
Genomic Health has announced the complete data will be submitted for presentation at the 2013 ASCO Genitourinary Cancers Symposium this winter. The company plans to offer their test—the multi-gene Oncotype DX Genomic Prostate Cancer Score—to clinicians and their patients in 2013, following the presentation of complete data.