Science Spotlight

Signature required: predicting prostate cancer aggressiveness

Prostate Studies (Public Health Sciences)

About one in seven men will be diagnosed with prostate cancer during his lifetime. While many prostate cancer patients are unlikely to progress even if left untreated, other patients will have aggressive tumors that become life-threatening.

Current prognostic information for prostate cancer includes a biomarker, prostate-specific antigen (PSA), and a pathological measure of tumor grade called the Gleason score.

PSA is a protein produced by cells of the prostate gland; blood levels of PSA very low in healthy men but are often highly elevated in men with prostate cancer. The PSA test was originally approved by the Food and Drug Administration in 1986 to monitor the progression or recurrence of prostate cancer in men, and then in 1994 was approved for use in conjunction with a digital rectal exam as a diagnostic for prostate cancer in asymptomatic men.

One of the best prognostic measures, the Gleason score, serves as a system of grading prostate cancer tumor tissue based on how it looks under a microscope. A low Gleason score (e.g. ≤6) means the cancer cells retain some similarity to those in normal prostate tissue and the tumor may be indolent and less likely to spread; a high Gleason score (e.g. ≥8) means the cancer tissue appears to be very different from normal prostate cells and the tumor is more likely to spread.

Although current clinical and pathological measures provide important prognostic information, they do not accurately predict an individual patient's risk of progression, and better prognostic markers are needed.  Epigenetic signatures, or patterns of DNA methylation in tissue, provide a remarkable breadth of information about cells, including their chronological age, developmental history, and differentiation potential. There is substantial heterogeneity among prostate tumor DNA methylation profiles, and these epigenetic changes may provide insight into cancer aggressiveness.

To investigate the utility of using an epigenetic signature for predicting prostate cancer prognosis, Drs. Milan Geybels and Janet Stanford (Public Health Sciences Division) and colleagues generated an epigenetic signature by comparing patients with high versus low Gleason scores. They then tested this signature for its ability to predict recurrence in a validation cohort of patients with clinically localized prostate cancer who had had radical prostatectomy as primary therapy. The results of their study were recently published in Clinical Epigenetics.1

Graph depicting recurrence-free survival by quartiles of the epigenetic signature.
Epigenetic signature of Gleason score and prostate cancer recurrence in the Fred Hutchinson cohort. Recurrence-free survival by quartiles (Q1–Q4) of the epigenetic signature. The vertical dashed line shows the recurrence-free survival rate at 10 years after diagnosis. Top panel is for all patients and bottom panel for patients with Gleason score 7 tumors. Figure provided by Dr. Janet Stanford

The researchers used genome-wide tumor DNA methylation data from 333 prostate cancer patients from The Cancer Genome Atlas (TCGA). The elastic net method, a method for analyzing high-dimensional data, was used to generate an epigenetic signature by contrasting tumors from TCGA patients with high (8–10) versus low (≤6) Gleason score. They then tested this novel signature in a cohort of 523 patients. Samples taken from the primary tumor were used for DNA methylation and mRNA expression profiling. Clinical information, vital status, and cause of death were collected from the Seattle-Puget Sound Surveillance, Epidemiology, and End Results (SEER) Program cancer registry. Patient follow-up surveys and medical record reviews were used to determine prostate cancer recurrence status in the FH cohort over an average follow-up period of 8 years.

In the FH cohort, higher levels of the signature were associated with poorer recurrence-free survival (hazard ratio per 25% increase = 1.78; 95% confidence interval 1.48-2.16). Also, the signature significantly improved the area under the curve (AUC) for prostate recurrence compared to clinical-pathological parameters alone, particularly among patients diagnosed with Gleason score 7 tumors (0.64 vs. 0.76, P = 1.34x10−4).

Dr. Geybels describes the importance of these results, “The signature also predicted disease recurrence in patients diagnosed with intermediate Gleason score 7 tumors, suggesting that it may have potential as a molecular tool to further improve the prognostication of these patients who typically have a variable prognosis.”

“This work was made possible by funding from the Pacific Northwest Prostate Cancer Specialized Program of Research Excellence (SPORE) grant that partially supported the ongoing follow-up of this cohort of prostate cancer patients who were initially enrolled in studies conducted years ago by myself and my colleagues,” explains Dr. Stanford.2,3 Baseline and follow-up clinical and epidemiological data as well as biospecimens, including primary tumor tissue, were available from patients diagnosed in 1993-1996 or 2002-2005 and treated with radical prostatectomy. According to Dr. Stanford, “Key factors making the discovery of this methylation signature of tumor aggressiveness possible were the availability of the primary tumor tissue and the long-term patient outcomes data on adverse events such as recurrence and metastatic progression.”

Funding for this study was provided by the National Cancer Institute, National Institutes of Health; Illumina, Inc. supported the generation of tumor DNA methylation and gene expression data.


  1. Geybels MS, Wright JL, Bibikova M, Klotzle B, Fan JB, Zhao S, Feng Z, Ostrander EA, Lin DW, Nelson PS, Stanford JL. 2016. Epigenetic signature of Gleason score and prostate cancer recurrence after radical prostatectomy. Clinical Epigenetics. 8(1):97.
  2. Stanford JL, Wicklund KG, McKnight B, Daling JR, Brawer MK. 1999. Vasectomy and risk of prostate cancer. Cancer, Epidemiolgy, Biomarkers, & Prevention. 8(10):881-6.
  3. Agalliu I, Salinas CA, Hansten PD, Ostrander EA, Stanford JL. 2008. Statin use and risk of prostate cancer: results from a population-based epidemiologic study. American Journal of Epidemiology. 168(3):250-60.