From Insights May 2014 Vol.17 Is. 2 | John Davis, M.D., Associate Professor, Urology Director,

Urosurgical Prostate Program MD Anderson Cancer Center Houston, Texas 

In Part I of this two-part article, we introduced the concept of genomic testing in prostate cancer and established how such new information can augment what we can already estimate with basic clinical information. A common need in prostate cancer is to take patients diagnosed with early prostate cancer, and determine whether or not they have disease with significant lethal ability. Another clinical need is to look at patients who have undergone a radical prostatectomy (any technique) and found to have high-risk pathology features that may indicate they need post-operative radiation therapy or other treatments.

Diagnosis: High-risk pathology after Radical Prostatectomy

Choices: Observation or post-treatment radiation?

As shown in part 1 the biopsy pathology is a powerful clinical predictive tool [1]. After a prostatectomy, the complete prostate pathology is even more powerful. There are essentially 5 key components to interpreting a radical prostatectomy report to predict future disease recurrence:

  • Gleason Grade—this time more accurate compared to a biopsy since the whole gland is sampled. It could be higher, lower or the same as the biopsy.
  • Extra-prostatic extension—the finding of tumor cells invading the layer of fat around the prostate capsule. This is denoted as pathological stage pT3a.
  • Seminal vesicle extension—the finding of tumor cells invading these adjacent structures off of the base of the prostate. This is denoted as pathological stage pT3b.
  • Positive surgical margins—the finding of tumor cells touching the inked margin of the tissue the surgeon removed.
  • Positive lymph nodes—if a lymph node dissection was performed.

How often does this occur and what does it mean? There are nomograms that are available on the internet* that can predict these findings from the pre-treatment clinical information, as well as the next step of predicting the odds of biochemical recurrence (rising PSA) after a radical prostatectomy. In our own series, for example, we looked at men with commonly treated intermediate risk prostate cancer such as a normal DRE, PSA < 10, and Gleason 7, and found that their radical prostatectomy findings are quite diverse. Approximately 50% of men will have organ confined cancer that is Gleason 7, while the other half could have any single or combination of elevated risk such as positive nodes (10%), pT3a or pT3b (25%), or increase in Gleason to 8-10 (5%). Recently, the American Urological Association (AUA) and the American Society for Therapeutic Radiation Oncology (ASTRO) produced a joint guideline on post-operative radiation therapy [2]. A key message is that men who have one or more of the features of a positive surgical margin, extra-prostatic extension, or seminal vesicle invasion should be counseled on the benefits of adjuvant (immediate) post-operative radiation to reduce future risk of PSA recurrence, local recurrence, and clinical progression. In our series of 879 recently performed robot-assisted radical prostatectomies, we found indications for post-operative radiation therapy in up to 35% of cases. However, most of our patients prefer to observe their PSA in the case of high-risk pathology and selectively use radiation for a detectible and rising PSA—often called the “salvage” radiation approach. Yet we do not have good clinical trial information as to whether this strategy is as good as the adjuvant—such a trial is nearing completion in the U.K [3].

In this area of clinical uncertainty, the Prolaris test can look at the same CCP** genes in radical prostatectomy tissue and give an estimate of 10-year biochemical recurrence [4-5]. The final prediction will incorporate what is known from the clinical information, and the same descriptors of less, more, or equal to average risk can be summarized.

Example A: Radical prostatectomy pathology shows Gleason 4+3, pT3a, a positive surgical margin, and normal lymph nodes. The PSA was 10 pre-treatment. Based upon clinical features alone, biochemical recurrence could be predicted at up to 57%, and AUA/ASTRO guidelines would strongly encourage immediate (meaning between 4-6 months post-operative) radiation. The CCP score however was a negative 1.1—much less aggressive than average risk, and the revised risk of biochemical recurrence was 31%. This patient selected observation.

Example B: A patient’s radical prostatectomy pathology showed Gleason 4+3, seminal vesical invasion, and positive surgical margin. His CCP score was positive 0.4, which is more aggressive than higher risk. His combined risk of biochemical recurrence is 87%, and was clearly recommended to consider adjuvant radiation.

Moving Forward: Key Questions and Critiques for the Future of Genomic Medicine in Prostate Cancer

Let us continue with some simple terminology to help understand the role of Prolaris and other genomic tests. In addition to Prolaris, other commercially available genomic tests that use prostate cancer tissue as their source include Decipher from GenomeDX [6] and Oncotype Dx from Genomic Health [7] (other companies have additional products in testing). The Decipher test is only for post-operative testing and focuses on the highest risk patients who might develop early bone metastasis. The Oncotype Dx test is a biopsy test for only very favorable biopsies to determine if patients are at risk for upgrading/upstaging. Therefore the key questions to ask when reading about a novel genomic test are:

  • What tissue can be tested? Biopsy versus radical prostatectomy or both.
  • What questions are addressed? The most common are the surveillance question and the post-operative radiation question. With further study, perhaps more questions could be addressed.
  • What information is returned to the patient and physician? In general, a genomic test will have its own unique scale and different cut-points for descriptive results such as the “more-aggressive than average risk.” These scales will have a learning curve for use by patients and physicians. Some tests are stand-alone predictors while others may incorporate the clinical data.
  • What are the validation methods? Because prostate cancer is often slow growing, all of these diagnostic companies have tested their product on retrospective tissue banks. Therefore, the information they can give you depends upon what clinical information was known about the test subjects. For Prolaris, for example, the biopsy test was done on two observation cohorts in the U.K and the result is a mortality prediction. For the post-operative test, only the biochemical recurrence information was known.
  • How many tests are informative? This will be worked out with more research. The concept to stress is that there is always a chance that a genomic test will return a result that is equal to clinical information and not really be that helpful, but other than it does rule out the most aggressive result.


Genomic testing is now a reality in prostate cancer. There are several choices available and each test has its intended question to address. The critiques of current genomic testing are that we need more information on the cost savings of testing select patient populations and subsequently making higher impact decisions. We need more evidence on “clinical utility,” meaning how many clinical decisions are clearly changed based upon the findings of a genomic test. In addition, none of the current tests are therapy linked, meaning we do not have hard evidence that altered therapy decisions based upon genomic tests lead to better outcomes. Such information may come with time as it has been established in other areas such as breast cancer.