Clinical and genomic analysis of metastatic prostate cancer progression with a background of postoperative biochemical recurrence - Beyond the Abstract

Decipher® provides insights beyond PSA after prostatectomy.

Further clarity is needed for the precise role a rising PSA, also referred to as biochemical recurrence (BCR), plays in the clinical management of prostate cancer after radical prostatectomy (RP). Increasing evidence is revealing that a rising PSA, the primary indicator of poor prognosis, is an insufficient surrogate for metastatic disease.

Nearly all patients with clinically determined metastatic prostate cancer will first experience a PSA rise after RP. However, a recent study published in European Urology by Ross et al. has shown using a prostate cancer natural history cohort that even though approximately 13% of intermediate and high risk prostate cancer patients exhibit a rising PSA at 10 years after RP, only 38% of those patients will go on to develop metastasis within the following 5 years [1]. This finding is also supported by estimates from the CISNET Prostate Cancer Modeling group (led by Dr. Ruth Etzioni) that indicate “over-detection” of recurrent disease is between 10-37% in men exhibiting rising PSA, depending on a patient’s age and Gleason grade at RP [2].

To enable improved care in the presence of rising PSA, it is critical to know which patients are destined to go on to develop lethal disease. Furthermore, determining the prevalence of PSA rise in patents with a favorable outcome is of particular importance because PSA level is relied upon so heavily, as a prognostic biomarker, treatment response indicator and as a surrogate endpoint in clinical trial design.

To understand the molecular basis of the disease recurrence “over-detection” problem Alshalalfa et al., using the tissue-based whole transcriptome analysis platform Decipher®, examined the expression profiles of three post-RP clinical outcome groups at the Mayo clinic, the same cohort on which the Decipher metastasis signature was originally developed. These groups comprised of patients with no evidence of disease (NED, n=108), those with rising PSA but no metastatic progression (BCR, n=163), and those who went on to develop metastatic disease (MET, n=196). An analysis of the data demonstrated that at the molecular level the BCR and NED groups were surprisingly similar. Of over a half million Probe Select Regions (PSRs) interrogated, only 18 were differentially expressed between the two groups with statistical significance. In contrast, the MET group displayed a unique genomic profile differing from both the NED and BCR groups, with 1054 and 761 PSRs, respectively at the same statistical threshold. Of these differentially expressed PSRs, 334 were common to both the NED and BCR group suggesting that the MET group had a transcriptional profile distinct from the NED and BCR groups.

The findings from this study indicated that detection of rising PSA alone after RP is likely insufficient evidence to predict the development of clinical metastasis. This result may also provide clues to the underlying reason for the findings of the CISNET group, suggesting that there are really two modes to PSA recurrence with only one leading to early metastasis. This study thus highlights the significance of using metastasis as an endpoint, rather than PSA alone, when developing molecular tests designed to predict disease recurrence. Metastasis appears to be a more robust endpoint for identifying those patients who are truly at risk for lethal disease. This was precisely the reason that the Decipher prostate cancer classifier, now commercially available, was developed using the MET group as test cases and the NED and BCR group as controls.

Decipher results from this study support the notion that the majority of men with prostate cancer who are treated with surgery have disease that can be effectively managed with local therapy approaches, regardless of whether or not they have rising PSA. If BCR does occurs this may signal disease progression that will follow a relatively indolent course that can be effectively managed with postoperative radiation; however, in the subset of patients where BCR precedes rapid onset of metastasis, earlier intervention with systemic therapy may be required to delay or prevent disease progression and early death from prostate cancer.

With further investigation, the data from this study may also provide insights into the altered biological pathways that manifest early in the primary tumors of these patients. A better understanding of this could lead to new therapeutic approaches that can prevent or further delay metastasis and lethal prostate cancer. This study suggests that the use of Decipher, a genomic signature of metastasis, can play a prominent role in treatment for patients with adverse pathology and PSA rise in order to better prioritize secondary treatment.

1. Ross, A.E., Yousefi, K., Davicioni, E., Ghadessi, M., Johnson, M.H., Sundi, D., Tosoian, J.J., Han, M., Humphreys, E.B., Partin, A.W., Walsh, P.C., Trock, B.J., Schaeffer, E.M. Utility of Risk Models in Decision Making After Radical Prostatectomy: Lessons from a Natural History Cohort of Intermediate- and High-Risk Men. European Urology. 2015; DOI: 10.1016/j.eururo.2015.04.016.
2. Xia J, Trock BJ, Gulati R, Mallinger L, Cooperberg MR, Carrol PR, et al. Overdetection of recurrence after radical prostatectomy: estimates based on patient and tumor characteristics. Clin Cancer Res. 2014;20(20):5302–10.

Written by:
Mohammed Alshalalfa, PhD and Heesun Shin, PhD
GenomeDx Biosciences Inc
Vancouver, Canada.

AbstractClinical and genomic analysis of metastatic prostate cancer progression with a background of postoperative biochemical recurrence