Molecular predictor differentiates indolent, advanced prostate cancer

May 1, 2011

Researchers have identified a four-gene signature that may significantly advance the current clinical ability to differentiate indolent from aggressive prostate cancer in the mouse model and in human tissue samples.

Boston-Researchers at the Dana-Farber Cancer Institute and Harvard School of Medicine, Boston, have identified a four-gene signature that may significantly advance the current clinical ability to differentiate indolent from aggressive prostate cancer in the mouse model and in human tissue samples.

Today, pathologists can differentiate between indolent and aggressive tumors in biopsied tissue with between 60% and 70% accuracy. The Harvard/Dana-Farber research, which was published in Nature (2011; 470:269-73), suggests that accuracy may be raised to the vicinity of 90% or higher should its initial findings in animal studies and human tissue samples be translated to clinical application.

"Cancer is an enormously complex disease on the genetic and biological level," senior author Ronald DePinho, MD, told Urology Times. "Gleason scores certainly help guide cancer management, but in the early stages [of the disease] it can be hard for pathologists to figure out the future course of the disease based simply on morphological and clinical parameters. Identifying the genes that might accurately provide prognostic and therapeutic points of attack would be a significant advance.

Prevention of unneeded treatment possible

Dr. DePinho told Urology Times that if the study data fulfill their promise, it would enable the identification of post-prostatectomy patients who should receive adjuvant treatment and substantially reduce the number of unnecessary prostatectomies being conducted today.

He pointed to a large European study that found that 48 men were treated for prostate cancer for every one life saved (N Engl J Med 2009; 360:1320-8).

"Tens of thousands of men are subjected to unnecessary prostatectomies and radiation treatments that have significant morbidity and cost the country perhaps as much as a billion dollars per year," Dr. DePinho said. The test's technology and procedure have been licensed to Metamark Genetics Inc. (Cambridge, MA), which Dr. DePinho and co-author Lynda Chin, MD, founded. If subsequent study and testing supports initial findings, a practical clinical test may be available within 2 years.

To identify the genes that might be driving or permitting metastases, the researchers used computational biology technologies to identify genetic pathways that were activated in Pten knockout mice, a murine model of prostate cancer that is usually indolent. One pathway, the TGF -SMAD4 pathway, proved to be particularly hyperactive. When they looked at mice in which both Pten and SMAD4 genes were silenced, they found cancers that were exceptionally aggressive and mobile. Continued research into these murine models implicated two more genes (SPP1 and CyclinD1) as having a primary role in producing exceedingly aggressive tumors.

This four-gene signature (Pten, SMAD4, SPP1, and CyclinD1) proved to be highly predictive when used as a screen in 400 prostate tissue samples from the Physician's Health Study, which holds records on more than 15,000 physicians aged 55 and older.

"We found that these four genes were able to outperform Gleasons. When combined with Gleason scores, predictability moves from the low 80s to 90% or better," Dr. DePinho said.

"An important caveat is that we acquired data from radical prostatectomies, so it was known which tumors were aggressive and which were not," he pointed out.

"Nevertheless, there was a tremendous amount of genetic analysis, in-house biologic validation, and independent biologic validation. There is every reason to believe that this test is going to be useful at the very early stages of clinical management. But that remains to be proven."