Making treatment decisions once you have been diagnosed with prostate cancer is difficult. Often, it is impossible to accurately and reliably be able to predict the eventual course of the cancer. Prostate cancer, besides being one of the slower growing cancers, is also one of the deadliest. So good, informed decisions are vital.

The original standby prognostic markers for prostate cancer are a man’s PSA, the stage of the cancer and their Gleason grade. Given the poor quality of these markers, in the 2000s, more sensitive markers were developed, such as PSA density, % of positive cores, % of cancer in each core, and tertiary patterns were incorporated into our predictive models.

Today, in response to the still poor ability of even these newer markers, we have recently seen the incorporation of molecular markers into decision-making.

These newer prostate cancer tissue based molecular markers include Prolaris, OncotypeDx, and Decipher:


  • Prolaris is a tissue-based marker taken from a biopsy. It includes 46 genes, and provides 10-year probability of cancer specific mortality with conservative management. It measures the aggressiveness of the cancer based on how fast the cancer cells are dividing.


  • OncotypeDx is also a tissue-based marker taken from a biopsy. It includes 17 genes, and was developed specifically for men having low risk cancer (Gleason 3+3 and 3+4 disease with <=30% of the biopsy cores found to be positive). It uses the biopsy tissue taken from the tumor itself. It predicts the likelihood (odds) that the cancer will grow and spread.


  • Decipher is different then the prior two markers as it uses tissue from the actual prostatectomy in men with adverse pathology. It includes a 22-gene RNA assay and has been validated to provide a 5-year probability of a man eventually developing advanced or metastatic prostate cancer post surgery and within 3 years of a biochemical recurrence (PSA only).


Using liquid (serum or blood) as a marker is another newer approach. Not yet commercially available is a marker known as AR-V7. It is a predictive marker which looks at a splice variant of the androgen receptor where the ligand binding domain has been eliminated, but the receptor remains persistently activated. AR-V7 looks as if it also can be used to predict sensitivity to some of the drugs we use to treat advanced, metastatic prostate cancer.

Another new direction we have recently seen discussed in prognostic markers is the development of a novel classification system that includes identification of subtypes 1, 2, or 3. However, even within each subtype, there remains a degree of heterogeneity in aggressiveness. Despite this, models that incorporate these subtypes can predict progression to metastatic disease.

The bottom line is that prostate cancer is heterogeneous by nature. Understanding the heterogeneity can lead to better prognosis and treatment options. A clearer picture of prostate cancer heterogeneity is emerging, and we can take advantage of this to develop more predictive biomarkers.