Science Spotlight

Expressing the potential for prostate cancer progression

From the Stanford Lab, Public Health Sciences Division.

Prostate cancer is the most common non-skin cancer among men, and is also a leading cause of cancer-related death. Nearly one third of patients diagnosed with clinically localized disease will relapse within five years after their initial treatment. Predicting which tumors will recur or progress is an important clinical and public health goal, as it would help inform treatment and surveillance decisions. In a recent issue of Oncotarget, Drs. Min Jhun and Janet Stanford and colleagues in the Public Health Sciences Division developed a gene expression signature of Gleason score that may improve predictions of prognosis.

Several factors are associated with the likelihood of prostate cancer recurrence, including cancer stage at diagnosis, Gleason score, and prostate-specific antigen (PSA) level. The Gleason score, which grades the cellular appearance of a prostate biopsy, is considered the best predictor of prostate cancer aggressiveness. Patients with tumor Gleason scores of ≤6 typically have a favorable prognosis, while scores of 8-10 typically foretell a poor prognosis. Intermediate scores of 7 have a variable prognosis, and represent a heterogeneous subset of patients that require better prognostic markers in order to better determine their risk of recurrence.

In order to provide an improved prognostic tool, the authors developed a gene expression signature of Gleason score. The first step involved developing the gene expression signature using publically available data on 333 prostate cancer patients from The Cancer Genome Atlas (TCGA). Using gene expression data from TCGA on over 16,000 genes, the authors utilized elastic net logistic regression to build a gene expression signature of tumors with Gleason score 8-10 versus ≤6. This procedure led to an expression signature of transcripts representing 49 genes. Higher scores of this signature were associated with elevated expression of genes in cell cycle-related pathways, and decrease expression of genes related to oxidative phosphorylation, androgen and estrogen response, and apoptosis. Previous studies have also found these pathways relevant to prostate tumor progression, suggesting this gene expression signature may be capturing relevant biological differences.

Comparison of AUC curves using expression panel
Area under the curve for the prediction of metastatic-lethal prostate cancer for the gene expression panel, clinicopathologic features, or both. Image modified from the publication

The authors found that a 25% increase in the gene signature was associated with a hazard ratio of 1.51 for recurrence, which remained statistically significant after adjusting for age at diagnosis and pathological stage. Furthermore, this association also remained statistically significant, thought slightly attenuated to 1.44, after restricting the analysis to only those patients with a Gleason score of 7. The authors then calculated the area under the curve (AUC) to assess the predictive probability of the gene expression signature. For recurrence, the signature had an AUC of 0.68, representing a 3% improvement in prediction accuracy over models containing clinicopathologic factors alone. For metastatic-lethal progression, the signature had an AUC of 0.76, representing a 6% improvement. When the gene expression signature was combined with clinical measures, outcome prediction accuracy was improved even further (see figure).

Overall, the development of this gene expression signature provides a potentially useful tool to help improve the prognostication of prostate cancer patients. Other types of molecular biomarkers in prostate cancer tissue may also be relevant to understanding which genomic pathways influence prostate cancer progression. The authors continue to develop alternative and complementary methods for evaluating progression potential of primary localized prostate tumors, and hope these tools will provide value to clinical decision-making.

Also contributing to this project from the Fred Hutch were Drs. Milan Geybels, Jonathan Wright, and Suzanne Kolb.



Jhun MA, Geybels MS, Wright JL, Kolb S, April C, Bibikova M, Ostrander EA, Fan JB, Feng Z, Stanford JL. Gene expression signature of Gleason score is associated with prostate cancer outcomes in a radical prostatectomy cohort. Oncotarget 2017; 8(26):43035-43047.  doi: 10.18632/oncotarget.17428.



Funding for this study was provided by the NIH, Fred Hutch, Illumina, and the Dutch Cancer Society.