Liquid Biopsy in Prostate Cancer: Circulating Tumor Cells and Beyond

Abstract

Prostate cancer is a common malignancy impacting countless men without curative options in the advanced state. Numerous therapies have been introduced in recent years improving survival and symptom control, yet optimal methods for predicting or monitoring response have not been developed. In the era of precision medicine, characterization of individual cancers is necessary to inform treatment decisions. Liquid biopsies, through evaluation of various blood-based analytes, provide a method of patient evaluation with potential applications in virtually all disease states. In this review, we will describe current approaches with a particular focus on demonstrated clinical utility in the evaluation and management of prostate cancer.

1. Introduction

Prostate cancer is the most common solid malignancy among men worldwide [1]. The prevalence of prostate cancer in combination with a relatively protracted clinical course creates significant need for biomarkers to inform management decisions. Localized prostate cancer treatment options, including active surveillance, surgical excision, or targeted radiation, are made based on individual risk stratification including pathologic characteristics from prostate biopsy, prostate-specific antigen (PSA) level, imaging, and other patient factors [2, 3]. Despite increasing use of advanced imaging modalities, improved biopsy techniques, and development of novel systemic therapies, prostate-specific antigen (PSA) remains the dominant biomarker in clinical use for prostate cancer monitoring. Though controversial in its application for population-based screening due to concerns regarding overtreatment of otherwise indolent cancers, PSA is utilized for monitoring prostate cancer at all stages. Unfortunately, PSA levels often fail to accurately reflect disease burden or activity [4], and multiple therapies impact patient survival and symptoms without corresponding changes in serum PSA levels [5, 6]. As such, an urgent need exists for improved biomarkers that reflect therapy response; as importantly, tumor heterogeneity necessitates effective molecular profiling techniques to guide appropriate therapy selection. “Liquid biopsies” comprised of analytes from a peripheral blood draw offer an appealing modality for comprehensive cancer analysis. These techniques are simple, safe, and easily repeatable throughout disease course and can serve as prognostic and predictive biomarkers as well as ready tissue sources for molecular profiling. Findings from liquid biopsy have capacity to inform treatment decisions at all phases of cancer care from screening to advanced disease states. Liquid biopsy analytes including circulating tumor cells (CTCs), plasma cell-free genetic materials such as cell-free RNA and DNA (cfRNA, cfDNA), as well as extracellular vesicles harboring unique cancer-specific materials have each been evaluated in prostate cancer and continue to be developed. Here we examine current applications of these analytes to the evaluation and management of prostate cancer (Fig. 1).

Fig. 1. Liquid biopsy.

Minimal risk, easily repeatable, low cost, feasible in all patients

2. Circulating Tumor Cells

2.1. Identification and Enrichment

CTCs are disseminated from a primary or metastatic tumor sites and circulate in the vasculature with potential for distant seeding [7, 8]. These have been identified in the context of virtually all solid malignancies, typically in the advanced state, and conversely are absent in healthy patients [9]. Though clearly essential to the development of metastatic disease which accounts for the majority of cancer-related mortality, the specific mechanisms which drive and enable the proliferation of CTCs remain poorly understood [8, 10]. Although CTCs have been demonstrated to have metastatic potential, not all CTCs are destined to form metastases and additional contributing factors are needed [11]. Nevertheless, the biological and potential clinical value of CTCs is established. CTC capture is technically challenging given their scarcity: usually between 0 and 100, relative to the billions of red and white blood cells within a blood sample [12]. Indeed, though initially recognized almost 200 years ago, only recently has consistent enrichment, identification, and even capture been made possible through technologic advances enabling better understanding of the physical characteristics and phenotypes of CTCs in comparison with other circulating cells [13]. In the evaluation of prostate cancer, the CellSearch system (developed by Janssen Diagnostics, LLC and recently acquired by Menarini-Silicon Biosystems) is the most clinically studied platform for CTC enrichment. CellSearch uses ferrofluid nanoparticles linked with antibodies directed toward epithelial cell adhesion molecule (EpCAM) to separate EpCAM+ cells from the buffy coat of centrifuged blood. This is followed by staining for cytokeratin (CK) and CD45 (leukocyte-specific antigen) to identify CTCs (EpCAM+, CK+, CD45−) [14]. Despite growing appreciation of CTC heterogeneity highlighting potential limitations of EpCAM-dependent identification [15, 16], CellSearch remains the most established platform in the setting of prostate cancer and is the only FDA cleared device for the detection of CTCs in metastatic prostate cancer [17]. It is also approved in the settings of metastatic breast [18] and metastatic colorectal [19] cancers. Alternative platforms isolate cells independent of marker status and rely instead on physical qualities such as size, deformability, or bioelectric properties [20–22]. In an effort to circumvent issues with CTC enrichment platforms have leveraged high-resolution scanning and automated detection algorithms to identify CTCs within whole blood smears or following RBC separation [23, 24]. These systems may prove to better represent total CTC population without regard for physical characteristics or surface antigen expression but limit manipulation and recovery of live cells. It is essential to recognize that all clinical studies must be interpreted with respect to the method of CTC enrichment and identification utilized as this directly impacts the population of cells collected with potentially significant implications. For instance, systems dependent on EpCAM expression fail to capture cancer cells that have undergone epithelial-mesenchymal transition (EMT), a population marked by increased aggressiveness and advanced disease [25]. For the most part, these nuances remain to be explored.

2.2. Clinical Applications of CTCs in Prostate Cancer

2.2.1. Enumeration: CTC Numbers as a Biomarker of Disease Activity

CTC enumeration has been extensively evaluated in localized and advanced prostate cancer states. In localized disease, initial hopes that identification of CTCs may predict disease recurrence were not realized. Davis et al. examined men with localized prostate cancer undergoing radical prostatectomy. Less than 5% had greater than 2 CTCs, and no correlation was found between CTC count and tumor volume, pathological stage, or Gleason score [26]. In another study utilizing the CellSearch enrichment platform, CTCs were detected in just one of twenty patients with high-risk localized prostate cancer (no CTCs were identified in healthy controls) [27]. Overall, CTC detection rates have varied between 5 and 52% in various studies of men with localized prostate cancer [28–31]. No study to date has demonstrated a significant correlation between CTC enumeration and Gleason score, tumor stage or PSA in the pre- or early postoperative time frames. Though enumeration in the localized setting has not proven clinically beneficial, further characterization of CTCs, when identified in this setting, may offer relevant clinical applications as discussed later [30].

More extensive evaluations have been performed in the setting of metastatic prostate cancer wherein ostensibly greater numbers of CTCs may be expected due to increased volume of disseminated cancer. Okegawa et al. found 55% of men with metastatic prostate cancer prior to androgen deprivation therapy to have ≥5 CTCs/7.5 ml blood. These men responded to androgen deprivation (by PSA) for just 17 months in comparison with men with ≤5 CTCs who responded to ADT for 32 months (P = 0.007) [32]. Another study examining this same population by CellSearch found a positive correlation of CTC count with LDH and alkaline phosphatase but not with PSA or testosterone levels. Patients who developed castrate resistance had a median CTC count of 17, while those who did not develop castrate resistance had a median CTC count of 1. On multivariate analysis, only baseline CTC counts were predictive of progression to castration resistance (P < 0.001) [33]. Recent evidence supporting the early administration of chemotherapy with hormonal therapy has caused a paradigm shift toward more aggressive initial management of metastatic prostate cancer [34, 35]. Given the demonstrated capacity of CTCs to predict time to castration resistance in this disease state, CTCs may prove beneficial for identifying those patients most likely to benefit from early chemotherapy.

In the castration-resistant setting, the prognostic capacity of CTCs has been most well established. In one of the seminal clinical studies of CTCs in prostate cancer, De Bono et al. enumerated CTCs using CellSearch. A total of 276 men with metastatic castrate-resistant disease prior to initiation of new therapy were enrolled with subsequent evaluation of 231. CTC counts were categorized as favorable (<5 CTCs/7.5 ml blood) and unfavorable (>5 CTCs/7.5 ml blood). CTCs were identified in 219/231 (95%) men at baseline demonstrating the prevalence of CTCs in this advanced disease state. Stratified in this manner, unfavorable initial CTC counts were associated with shorter overall survival (median 11.5 vs. 21.7 months) and outperformed PSA algorithms at all time points. In addition, patients with initially unfavorable CTC counts who converted to favorable counts with treatment experienced similar improvements in median survival (6.8–21.3 months) demonstrating the capacity for CTCs to reflect response to therapy in these patients [17]. Subsequently, the prognostic efficacy of CTCs has been explored in the context of various other systemic treatments for advanced prostate cancer including docetaxel, abiraterone, enzalutamide, and various combinations consistently demonstrating an association with overall survival that rivals or surpasses that of PSA monitoring alone [36–38].

CTC enumeration has not been quickly adopted to clinical practice. Though effective in prognostication, no study to date has demonstrated ability to directly inform management and thereby alter patient outcomes. Increasingly, however, the potential to use CTC counts as surrogate endpoints in clinical trials where they may facilitate shorter trial duration and associated costs is being explored. In evaluating results of COU-AA-301, a large phase III trial of abiraterone plus prednisone versus prednisone alone in patients with mCRPC, Scher et al. found CTC count in combination with lactate dehydrogenase (LDH) level predictive of overall survival at two years [38]. These two measures together met all Prentice criteria required for use as a surrogate endpoint for overall survival [39].

2.2.2. Liquid Biopsy in Practice: Identifying Phenotypes Through CTC Characterization

Simple enumeration of CTCs, while correlated with disease burden, response to treatments, and prognostic among men with advanced prostate cancer, does not capitalize on the nature of CTCs as components of relevant, viable tumor tissue. Characterization of these cells therefore has the capacity to illuminate aspects of individual patient molecular profiles and guide treatment choices even in the localized setting where simple enumeration lacked prognostic significance. Pal et al. performed immunohistochemical staining on enriched CTCs for CD133 (a putative stem-cell marker) and E-cadherin (a marker of epithelial-mesenchymal transition) finding an association between expression of these markers of aggression and biochemical recurrence at one year following prostatectomy in the localized setting, a disease state wherein total enumeration of CTCs was not predictive [30]. Goldkorn et al. evaluated telomerase activity in CTCs in a corollary study of SWOG 0421 (docetaxel with atrasentan versus docetaxel alone in CRPC patients). CTC telomerase activity was prognostic of overall survival in this setting. Though many potential targets for CTC characterization exist, clinical studies have focused on the androgen receptor (AR) given its central role in hormonal therapies for advanced prostate cancer. Immunofluorescent staining of AR on CTCs to determine cellular localization (nuclear vs cytoplasmic) has been linked to chemotherapy response and clinical disease progression on abiraterone [40, 41]. Likewise, classification of AR as “on” vs “off” by immunofluorescent staining identified differences in CTC profiles of hormone-naive patients and those who had progressed to CRPC [42]. More recently, ligand-independent AR splice variants have been found to play a role in resistance to second-generation antiandrogen therapies enzalutamide and abiraterone [43]. Antonarakis et al. used quantitative reverse-transcriptase polymerase chain reaction (PCR) to evaluate AR-V7 expression in CTCs of men with CRPC. Men with AR-V7 expression had significantly lower PSA response rates, progression-free survival (PFS), and overall survival when treated with enzalutamide or abiraterone, suggesting a possible means of predicting response to these therapies through CTC profiling [44–46]. The association is not entirely clear, however, as other groups have demonstrated clinical response to abiraterone or enzalutamide despite AR-V7 positivity in CTCs [47]. Greater depth of characterization by identifying cellular location of AR-V7 protein may enhance predictive capacity as demonstrated by Scher et al. In their study, nuclear localization of AR-V7 protein was the strongest baseline factor influencing overall survival even in comparison with AR-V7 positive cells without localization [48]. Heterogeneity among these cells and tumors including alterations in specific signaling pathways have been shown to contribute to variable responses through RNA sequencing of single CTCs [49]. Whole-genome and exome sequencing of CTCs from men with prostate cancer has been completed demonstrating a strong capacity for CTC evaluation to recapitulate primary tumor mutations, thus furthering potential clinical applications of CTC analysis [50, 51]. Ongoing work to capture pure CTC samples and better characterize clinical implications of specific CTC characteristics in light of known heterogeneity will enable meaningful application of CTC analysis to patient care (Table 1).

Table 1.

Select examples of AR-V7 detection in clinical studies

References	Analyte	Patients analyzed (n)	Method	Findings
Antonarakis et al. [44, 45, 97]	CTCs	62 and 37	RT-PCR to evaluate AR-V7 transcripts in CTCs among men receiving second-generation hormonal therapy or taxane chemotherapy [45, 97]	AR-V7 expression associated with lower PSA response, PFS, and OS among men receiving enzalutamide or abiraterone. Men found AR-V7+ better response to chemotherapy in comparison with antiandrogens
Scher et al. [46, 48]	CTCs	161	Immunofluorescent staining of CTCs for AR-V7 protein [46] with additional evaluation for nuclear-specific signal localization [48]	CTCs expressing AR-V7 found in 34 (18%) samples using nuclear-specific criteria, 56 (29%) without nuclear criteria. AR-V7 associated with resistance to hormonal therapy, decreased PFS, shorter OS among those with nuclear localization of AR-V7+ indicating role for chemotherapy selection
Liu et al. [76]	RNA and CTCs	46	Comparison of PAXgene preserved RNA versus leukocyte depletion and CTC analysis	AR-V7 detected in 68% of samples. Increased expression associated with receipt of second-line hormonal therapies
De Laere [73]	CTCs	30	Low-pass whole-genome sequencing of ctDNA and targeted sequencing of AR gene. Splice variant analysis from CTC RNA	AR alterations identified in 25/30 patients and associated with PFS. AR-V7 negativity more prevalent among poor responders
Del Re et al. [92]	Exosomes	36	Exosomes isolated and RNA extracted for analysis	39% of patients AR-V7+. AR-V7 associated with longer PFS (20 vs. 3 months, p < 0.001) and OS (8 months vs. not reached, p < 0.001)

3. Plasma Cell-Free Genomic Materials

3.1. Identification

Fragments of DNA circulating freely in the bloodstream are termed cell-free DNA. In the presence of malignancies, the fraction of cell-free DNA (cfDNA) derived from cancerous cells (primary tumor, metastatic sites, or CTCs) is alternatively identified as circulating tumor DNA (ctDNA). cfDNA may be released through a variety of natural and pathologic processes including apoptosis, necrosis, or even physiologic release from viable cells [52–54]. Healthy individuals have cfDNA levels of 1–10 ng/ml [55, 56], whereas cfDNA levels, though greatly variable, are consistently elevated among cancer patients [57, 58]. The role of benign processes in impacting cfDNA levels is significant as intense exercise alone can increase cfDNA levels as can trauma, infections, and inflammatory conditions [59]. cfDNA has been identified in almost all bodily fluids including urine [53]. These DNA fragments can be quantified and analyzed to offer insights regarding prognosis, response to therapy, and tumor mutational status. Identification of the source of DNA fragments is difficult given varied possible sources. Therefore, isolating small fractions of ctDNA within total cfDNA demands highly sensitive approaches targeting specific gene alterations, chromosomal abnormalities, epigenetic alterations or other characteristics to identify the cancerous source [60]. Digital droplet PCR (ddPCR) and associated methods have proven sensitive and can perform well in absolute quantification of ctDNA detecting point mutations at low allele frequencies [54, 58–61]. More recently, next-generation sequencing (NGS) of circulating DNA fragments has allowed comprehensive genomic profiling [62]. The short half-life of cfDNA (<2.5 h) allows accurate characterization of real-time tumor profiles [58, 63]. Extensive clinical evaluations are ongoing to better understand cfDNA and relevance to cancer care. Evaluation of circulating RNA has been performed as well but requires special approaches to collection due to generally quick physiologic degradation of cfRNA. Early studies evaluating cfDNA and ctDNA in the setting of prostate cancer have demonstrated potential for promising clinical applications.

3.2. Clinical Applications of cfDNA in Prostate Cancer

Quantification of cfDNA, especially in the context of temporal changes, offers a means of evaluating tumor burden and response to therapy. A retrospective study analyzed the prognostic significance of cfDNA concentration among men with CRPC prior to initiating chemotherapy and found elevated cfDNA concentration to be associated with poor PSA response and to act as an independent predictor of overall survival on multivariate analysis [64]. Likewise, analysis of cfDNA levels in men with CRPC during chemotherapy in conjunction with serial imaging for treatment response monitoring demonstrated a significant positive relationship between cfDNA levels and tumoral activity as determined by PET/CT [65]. Just as CTC enumeration reflects disease burden and prognosis, ctDNA quantification can reveal disease burden with potentially even greater sensitivity. By applying an immunospot assay for CTC detection and a PCR-based analysis with a panel of 14 polymorphic markers for detection of allelic imbalances on cfDNA, researchers found a significant relationship between the presence of CTCs and ctDNA levels with meaningful correlation to both tumor stage and Gleason score [66].

The specific relationship of cfDNA and ctDNA concentrations to prognosis remains to be fully defined given evolving techniques for detection. However, in-depth evaluation of these analytes has provided meaningful insights into individual cancer biology with potential therapeutic implications. Examination of cfDNA can demonstrate copy number variations or mutational status of relevant genes for treatment selection. Heitzer et al. completed whole-genome sequencing of cfDNA from patients with advanced prostate cancer and identified multiple copy number alterations and gene rearrangements with potential clinical significance. Most importantly, sequencing of cfDNA demonstrated a capacity to recapitulate the “genomic landscape” of prostate cancer in a more comprehensive fashion than even evaluation of the primary tumor which showed variable copy number changes on multiregional sequencing consistent with multifocal disease [67].

As with CTCs, evaluation of the AR among prostate cancer patients is essential due to its critical role in current treatments and known clinical relevance of various mutations [68]. One study examining CRPC patients undergoing treatment with abiraterone found a significant association between gains of AR copy number or CYP17A1 gene and survival outcomes when evaluating cfDNA prior to treatment [69]. Another group using array comparative genomic hybridization for copy number analysis found AR amplification to be much more common among patients with disease progression on enzalutamide in comparison with those on abiraterone or other treatments. In addition, AR gene aberrations in pretreatment cfDNA were predictive of adverse outcomes including lower rate of PSA decline and shorter time to progression [70]. Likewise, Romanel et al. performed targeted NGS on cfDNA of patients with CRPC receiving treatment with abiraterone and identified plasma DNA AR copy number gains and point mutations that predicted resistance to abiraterone, overall survival, and progression-free survival [71]. Further, emergence of new AR point mutations was identified in 13% of patients progressing on abiraterone without AR copy number change [71]. Therefore, evaluation of AR status prior to treatments can be useful in predicting response to therapy.

cfDNA analysis, due to safety and easy accessibility, lends itself to temporal evaluations not generally feasible when restricted to traditional tissue biopsies. Comparative genomic hybridization for copy number evaluation and sequencing of the AR gene among men treated with enzalutamide revealed clonal selection over the course of treatment. At time of progression, AR mutations or copy number changes were identified in all patients [72]. As in earlier studies, AR amplifications and mutations correlated with worse progression-free survival. De Leare et al. completed comprehensive profiling of the AR from ctDNA of men with CRPC evaluating for any AR perturbation. They identified abnormalities in 25/30 patients and found an association between presence of any AR variant and progression-free survival (PFS). They also analyzed splice variant expression from CTC RNA. Of a minority of poor responding patients who were AR-V7 negative, most were found to have other AR abnormalities [73].

As noted, analysis of whole blood RNA is limited by quick degradation but has been performed with assistance of preservative tubes (e.g., PaxGene, Qiagen; RNA Streck, Streck Inc.). Microarray RNA profiles allowed creation of gene expression signatures which demonstrated prognostic utility in CRPC patients from blood collected in PaxGene tubes [74, 75]. Another study using PaxGene tubes examined select gene expression in men with CRPC by RT-PCR. Transcript detection predicted overall survival and when combined with CTC enumeration had a high concordance probability estimate suggesting potential additive value in liquid biopsy analytes. Liu et al. determined capacity to evaluate AR-V7 expression from PaxGene® preserved blood samples with a correlation to second-line hormonal therapies as demonstrated previously in CTCs [76]. Each of these studies required special attention to the prevention of RNA degradation. Micro RNAs (miRNAs) are small noncoding RNA segments with roles in gene regulation. Known to be altered in virtually all cancers and marked by great stability, miRNAs are now being explored as biomarkers in prostate cancer [77]. Lin et al. identified fourteen miRNAs associated with serum PSA or overall survival among patients with CRPC receiving docetaxel. Detection of high levels of the miR-200 family predicted non-response to docetaxel [78]. Despite these promising early results, further validation in clinical studies is needed prior to clinical applications.

Excitement regarding potential applications of cfDNA analysis must be tempered in light of significant remaining challenges. For instance, in one study modification of sequencing approach enabled identification of multiple new AR mutations with functional characterizations from cfDNA not originally identified highlighting the importance of analytic approach on subsequent findings [79]. In addition, prevalence of false-positive mutations using NGS for low-abundance mutations highlights the need for optimization and standardization of these tests prior to application for clinical decision-making [80]. However, like CTC analysis, cfDNA offers exciting avenues of continued exploration including early disease detection, screening, and disease monitoring. Recent FDA approval of a targeted mutation test from cfDNA (cobas® EGFR Mutation Test v2, Roche Molecular Systems, Inc.) to determine eligibility for specific lung cancer therapy exemplifies the potential applications of cfDNA and the steady drive toward clinical utility which will certainly reach prostate cancer in the near future [81].

4. Extracellular Vesicles (EVs)

4.1. Overview and Preclinical Studies in Prostate Cancer

Increasing clinical interest has been directed to the identification and evaluation of EVs in many disease states including prostate cancer. EVs range in size from 50 nm to 10 μm and are produced by virtually all cells [82]. Appreciation of the unique characteristics and functional roles of various EV populations has grown with mounting evidence demonstrating cell specific vesicular contents as well as roles in intercellular signaling [82–84]. Contrary to early descriptions of EVs as cellular waste, they are now recognized to represent a heterogenous population unique in size, content, and mechanism of cellular release with variable biologic functions. Exosomes, perhaps the most studied EV subtype to date, measure 50–100 nm in diameter and originate from fusion of multivesicular bodies with the plasma membrane [85]. Alternatively, ectosomes and large oncosomes (1–10 μm) originate as direct buds off the plasma membrane which is relevant to isolation and analysis techniques. Each of these vesicles harbors cargo specific to their cells of origin including DNA, RNA, proteins and lipids [86]. While exosomes may originate from both benign and malignant cells, large oncosomes are thought to originate preferentially from malignant cells [87]. Each may supply unique biomarkers due to their highly selective cargos.

Techniques for EV isolation must be tailored to the specific vesicle of interest and continue to be refined. Traditional methods have relied on differential ultracentrifugation, and this remains the gold standard means of isolating EVs. Ultrafiltration- and immunoaffinity-based approaches may prove beneficial due to less technical demands and capacity for improved throughput and scalability. Given recent findings regarding heterogeneity of EVs, appropriate identification and confirmation of particles of interest is essential [88]. Just as for CTC and cfDNA analysis, therefore, clinical results should be interpreted with an eye to method of isolation and purification.

In preclinical settings EVs have been found to exhibit promising potential applications as prostate cancer biomarkers. Examination of EVs from prostate cancer cell lines allows for study of the vesicles in a comparatively pure media as opposed to the diverse milieu of human blood. One group examining EVs from prostate cancer cell lines found high expression levels of surface markers of aggressiveness that were not found in less aggressive lines suggesting capacity to identify specific tumor-derived EVs as well as to function as biomarkers of disease aggression [87, 89]. Other groups have noted alterations in exosome abundance within cell lines following acquisition of resistance to chemotherapy indicating a role for exosome monitoring for detection of drug resistance in addition to potential functional involvement in that process [90]. As EVs are identified and purified with greater consistency, recognition of their differential content and make-ups will enable continued meaningful interpretation and translational applications.

4.2. Clinical Applications of Extracellular Vesicles in Prostate Cancer

Limited trials analyzing EVs in clinical studies have been completed, especially from plasma. Huang et al. performed RNA sequencing of exosomal RNA in men with CRPC and identified two exosomal miRNAs associated with overall survival that they subsequently validated in a larger cohort of patients. miR-1290 and miR-375 were each associated with poor overall survival (p < 0.004) demonstrating potential application for exosome analysis [91]. Detection of AR-V7 has been accomplished in exosomal RNA. Del Re et al. extracted RNA from plasma-derived exosomes of CRPC patients and assessed for the presence of AR-V7 by digital PCR. Thirty-nine percent of patients were found AR-V7+ with significantly shorter overall survival among that subset [92]. Validation and comparative studies are required, yet the significant potential of extracellular vesicles as a meaningful liquid biopsy analyte is clear. Many studies have focused on the utility of urinary exosomes or other urinary markers to facilitate improved prostate cancer diagnosis, staging, and prognostication with variable results as well [93–96]. In one study, a prognostic score was developed by comparing urinary exosome gene expression assay by reverse-transcriptase PCR with biopsy outcomes in men with elevated PSAs. The three-gene assay including ERG, SPDEF, and PCA3 was validated in a larger cohort and improved discrimination among Gleason scores and benign disease demonstrating a viable method for determining need of biopsy in that setting [96]. Continued evolution in the differential detection and analysis of EVs will enable further characterization of their roles in prostate cancer signaling, implications for prognosis, prediction of therapeutic response, and clinical application.

5. Summary

The absence of sensitive and accurate biomarkers in the prostate cancer arena has limited clinicians to depend on PSA for disease monitoring and biopsy from primary or metastatic sites for genomic characterization with clear shortcomings. Liquid biopsies are noninvasive, easily repeatable and harbor significant cancer and patient-specific data that, when captured, may provide a near comprehensive representation of individual patients’ disease state throughout therapy. Many barriers and challenges remain prior to common clinical use. Optimization of identification, recovery, and analytic approaches is ongoing, and scaling of these approaches and integration to clinical use while maintaining accuracy and precision are challenging. Prospective clinical studies will enable interpretation of detected abnormalities in context of traditional biopsy findings, and various therapeutic settings but optimal applications remain unclear at this time. As seen in AR-V7 evaluation, information gleaned through examination of various available blood analytes may overlap or offer additive value. Therefore, reconciliation through evaluation in the context of translational studies with attention to costs and technical issues is essential when assessing for broad clinical use. Great progress has been made toward establishing applications for liquid biopsy in the clinical care of prostate cancer. Though still limited at this time, liquid biopsies will play a significant role in the care of prostate cancer patients in the near future with direct applications from the level of screening to delivery of precision care in the advanced metastatic state.