Genomic Signature Predicts Prostate Cancer Metastasis Risk

Kate Johnson

April 07, 2014

VIENNA — A DNA signature obtained from biopsies from untreated prostate cancer patients can differentiate between patients at low or high risk for metastasis even when clinical predictive factors appear equal, according to new research presented here at the European Society for Radiotherapy and Oncology (ESTRO) 33 annual conference.

"If validated, this could move things forward in terms of changing treatment choices," lead researcher Robert Bristow, MD, PhD, told Medscape Oncology." We were able to find 4 distinct genetic subgroups whose 5-year biochemical failure rate ranged from 50% to 90% in patients that radiation oncologists currently treat exactly the same," he said.

Presence of the genomic signature could guide patients toward intensified treatment when they would otherwise not get it on the basis of clinical factors, and similarly, the absence of it could allow deintensification in patients who would otherwise be treated more aggressively, explained Dr. Bristow, a senior scientist at the Ontario Cancer Institute and professor of medical biophysics at the University of Toronto, in Canada.

"If all goes well, this will lead to a new test for cancer patients that can be turned around in 3 days that we will quickly put it on a platform that can be used in hospitals to tell doctors which patients will do well with local treatment alone and which will need additional systemic treatment such as androgen deprivation therapy or novel agents," he said.

"These are very promising results because, if borne out in other patient cohorts, they would demonstrate that some patients who we currently think are at high risk for treatment failure are actually not ― and vice versa," session chair Joseph Deasy, PhD, from Memorial Sloan-Kettering Cancer Center in New York City, told Medscape Medical News.

Different From Available Tests

Although there are 2 commercially available genetic tests for predicting prostate cancer metastasis (Prolaris and Decipher), both are intended for patients who have already undergone radical prostatectomy, said Dr. Bristow. "I call them the horse-out-of-the-barn tests. The way those tests are being used currently is to decide whether the patient should get any extra therapy after surgery. So we wanted to develop a practical thing that a physician can do right up front to triage someone into surgery or radiotherapy and decide whether to add hormone therapy."

The commercially available tests also analyze RNA, whereas the test from Dr. Bristow's group analyzes DNA because "perhaps DNA is more stable than RNA in terms of intraprostatic heterogeneity," he said.

Results From Several Studies

The research involved several studies. A genomic signature for biochemical treatment failure was developed on the basis of copy number alterations within individual men's genomes in pretreatment biopsies from 126 patients who were eventually treated with image-guided radiotherapy. This study was then followed by a validation study of the signature in postsurgical specimens from 154 patients treated with radical prostatectomy. Finally, scores based on percent of genome altered (PGA) were calculated.

On the basis of posttreatment results from the patients, the researchers found that the hazard ratios (HRs) for biochemical failure in the biopsy group and the surgery group were 4.53 and 3.51, respectively, in patients whose PGA was above the median of 7.9 compared with those below the median. "For every 1% increase in PGA, there was a 5% increase in failure, and high PGA also correlated with development of metastasis in this cohort (P < .004)," said Dr. Bristow.

He added that just looking at patients' Gleason scores, there was a slight trend toward increasing PGA with increasing Gleason score, but patients with low Gleason scores had high genetic instability, and patients with high Gleason scores had low genetic instability. "So maybe this is why some of our Gleason-score-6 patients do poorly and some with high Gleason scores do well," he said.

"We then used cluster analysis within these patients who are otherwise homogeneous with respect to their clinical prognostics and found genetic heterogeneity," he explained. "We were actually able to put them into 4 different genetic clusters," he said. "We would argue that if validated, these are interesting subgroups to utilize for prognosis because actually, what we found is that patients who have very quiet genomes in fact do very well, and patients with many alterations in their genome actually fail rapidly. There's a difference of between 50% to 90% in terms of biochemical free relapse rates."

Refining the signature further to maximize the differentiation between responders and nonresponders, the researchers were able to show a difference of 58% to 90% in risk for biochemical free relapse rates (HR, 6.07; P = .0015).

"What's important is that this signature might pick out those patients who might fail with a Gleason score of 6 or less ― patients that might have otherwise gone into active surveillance ― and in patients whose clinical factors suggest they have high-risk prostate cancer and who would normally have hormone therapy as well as radiotherapy ― if the genetics say it isn't high risk, you can stop giving hormones to those patients," he said.

In a secondary study, the group analyzed hypoxia levels in the biopsy and postsurgical samples and found that high levels of hypoxia were associated with poorer outcomes.

Combining the data, they found that patients with both an adverse genetic signature plus high levels of tumor hypoxia had an HR of 4.5 for biochemical failure.

"This is the first report of biopsy-driven, DNA-based indices which is independent of treatment and, most importantly, the first report for radiotherapy patients," concluded Dr. Bristow. "Genetic instability is highly prognostic for treatment failure and can be combined with hypoxia indices to triage patients in adverse prognostic subgroups."

In a written statement, professor Vincenzo Valentini, president of ESTRO and a radiation oncologist at the Policlinico Universitario A. Gemelli, Rome, Italy, said: "This is exciting research because an accurate and quick test that can predict which men are most likely to need extra treatment to reduce the risk of a recurrence of their cancer is urgently needed. If the utility of this genetic signature is confirmed in further research over the next few years, it could become an important tool for helping us to better target-appropriate treatment according to the genetic makeup of each man's tumor."

Dr. Bristow, Dr. Deasy, and Dr. Valentini did not disclose any relevant financial relationships.

European Society for Radiotherapy and Oncology (ESTRO) 33. Abstract OC-0139. Presented April 5, 2014.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.