New Genetic Test May Give Early Warning of Aggressive Prostate Cancers

October 2016, Vol 7, No 9

A novel 37-gene assay showed potential for classifying prostate cancer into 3 subtypes based on aggressiveness and likely response to therapy, reported Michael R. Freeman, PhD, Director, Division of Cancer Biology and Therapeutics Research, Cedars-Sinai Medical Center, Los Angeles, CA, and colleagues (You S, et al. Cancer Res. 2016;76:4948-4958). Validated in more than 4600 prostate cancer specimens, each prostate cancer subtype was associated with different degrees of activation of specific genes and signaling pathways. The most aggressive subtype, known as prostate cancer subtype 1 (PCS1), comprised approximately 66% of castration-resistant and metastatic tumors.

Tumor Subtype Can Predict Outcomes

Patients with PCS1 tumors had the poorest prognosis, even if their tumors had low Gleason scores. In addition, a study of circulating tumor cells (CTCs) showed that the expression or activation signature of PCS1 was enriched in a majority of patients with resistance to enzalutamide. PCS2 was associated with intermediate risk, and PCS3 conferred the lowest risk, although samples in PCS3 were still distinct from benign prostatic disease. “Our findings provide evidence for biologically distinct forms of prostate cancer that are independent of Gleason grade, [which is] currently the gold standard for clinical decision-making,” Dr Freeman and colleagues noted. “PCS subtyping may improve accuracy in predicting the likelihood of clinical progression and permit treatment stratification at early and late disease stages,” they added. Recent studies have associated specific molecular lesions with aggressive prostate cancer. In addition, expression signatures related to the lesions have been developed for predicting patient outcomes. However, previous studies involved a limited number of patient cohorts and prostate cancer specimens. Dr Freeman and colleagues performed one of the most extensive investigations to date into genetic and molecular signatures associated with prostate cancer. They began by analyzing pathway activation signatures in 1321 human prostate cancer transcriptome profiles (ie, the discovery cohort). The analysis focused on 22 pathway activation signatures with established relevance to prostate cancer. The analysis honed the number of activation pathways to 14, and allowed the investigators to develop a novel classification system comprising the 3 prostate cancer–related subtypes:

  • PCS1 was associated with high activation of EZH2PTENPRFESAV, and AR-V
  • High activation of the ERG pathway predominated in PCS2, accompanied by high activation of ARFOXA1, and SPOP
  • PCS3 exhibited high activation of RASPN, and MES, along with low activation of AR and AR-V.

The investigators validated the classification system in 10 independent patient cohorts and 19 laboratory models of prostate cancer, which included cell lines and transgenic laboratory animal models. The results show that tumors classified as PCS1 or PCS2 represent luminal subtypes, and PCS3 represents a basal subtype.

The New 37-Gene Test

After demonstrating that the PCS1 subtype progressed more rapidly to metastatic disease, the researchers developed a 37-gene panel that accurately classified individual tumors to 1 of the 3 subtypes. Using 5 cohorts, they evaluated the classification system’s performance with respect to metastatic progression, prostate cancer–specific mortality, and overall survival. Overall, the PCS1 subtype emerged as the most aggressive of the 3 subtypes. As another test of the classification system’s robustness, the investigators evaluated the gene panel’s performance in 77 intact CTCs obtained from 13 patients. The results showed low expression of PCS1-associated genes in 67 CTCs, and high expression of PCS1-associated genes in 10 CTCs. PCS1-associated genes were highly expressed in 7 of 12 CTCs from patients with enzalutamide-resistant tumors. PCS3-associated genes were not detected, or they exhibited very low expression across all the CTCs. “Collectively, the results demonstrate that the 37-gene classifier has a potential to assign individual prostate cancers to PCS1 using both prostate tissues and blood CTCs, suggesting that the classifier can be applied to subtype individual prostate cancers using clinically relevant technology platforms, including by noninvasive methods,” the investigators wrote. “The classification scheme we describe predicts the risk of progression to lethal prostate cancer in patients with a diagnosis of low-grade localized disease. It is possible that in these cancers, pathway activation profiles are independent of Gleason grade and that pathways indicating high risk of progression are manifested early in the disease process….PCS subtyping might also assist with the selection of drug treatment in advanced cancer,” they concluded.

Related Articles