Clinical utility of current biomarkers for prostate cancer detection.

Although prostate-specific antigen (PSA) remains the most used test to detect prostate cancer (PCa), the limited specificity and an elevated rate of overdiagnosis are the main problems associated with PSA testing. Over the last three decades, a large body of evidence has indicated that PSA screening methods for PCa are problematic, although PSA screening significantly reduces PCa-specific mortality. A number of novel biomarkers have been introduced to overcome these limitations of PSA in the clinical setting. These biomarkers have demonstrated an increased ability to select patients for biopsy and identify men at risk for clinically significant PCa. Although a number of assays require further validation, initial data are promising. Forthcoming results will ultimately determine the clinical utility and commercial availability of these assays. Extensive efforts have recently been made to identify and commercialize novel PCa biomarkers for more effective detection of PCa, either alone or in combination with currently available clinical tools. This review highlights the role of existing and promising serum and urinary biomarkers for the detection and prognostication of PCa before prostate biopsy. © The Korean Urological Association, 2021.

INTRODUCTION

Prostate-specific antigen (PSA) is widely accepted as a tumor marker for screening, diagnosis, monitoring, and risk prediction of prostate cancer (PCa). However, PSA has limitations as a screening biomarker owing to its low specificity and lack of sensitivity. Its positive predictive value (PPV) was shown to be only approximately 25% in a pooled meta-analysis [1], which leads to a large number of false-positive results. As a result, up to 75% of patients undergo unnecessary prostate biopsies [2]. Up to 30% of PCa cases, and among these 10% of aggressive PCa cases, are identified in patients with a normal range of PSA [3]. Conventional prostate biopsy has inherent discomfort, cost, and problems with undersampling [4], although its morbidity is low. In particular, the rate of unnecessary prostate biopsy is increased in patients with PSA levels in the range of 2.5–10 ng/mL, the so-called gray zone [25].

Screening for PCa remains a controversial issue, although the European Randomized Study of Screening for Prostate Cancer (ERSPC) and recent analyses from the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial have provided evidence that PSA-based screening can significantly decrease PCa-specific mortality [67]. A large proportion of PCa is latent, that is, it is never destined to progress or affect a patient's life. Approximately 50% to 60% of newly diagnosed PCa cases are at low risk for progression [8]. However, a substantial number of patients with low-risk PCa undergo some form of aggressive treatment regardless of risk. This has led to concerns regarding overdiagnosis and overtreatment. Active surveillance (AS) has emerged as an alternative treatment option for men with early-stage disease, in which any kind of definitive treatment is delayed and applied only if there is evidence of progression. Notwithstanding, detecting disease progression in a patient selected for AS remains a continuing challenge. Moreover, some patients with apparently low-risk disease actually harbor unfavorable disease owing to the inaccuracies of currently used repeat biopsy protocols [9]. The current problems inherent in the prostate biopsy procedure advocate for the development of noninvasive tools capable of predicting disease progression more accurately that are suitable for repeat measurements over time.

Refinements to the PCa diagnostic pathway focusing on the detection of only clinically significant PCa (csPCa) are needed to make the diagnostic pathway less burdensome to patients. In addition, the diagnostic pathway should be cost-effective and acceptable to the general population and health care providers [10]. Currently, many international guidelines recommend the use of risk stratification tools such as novel biomarkers, risk calculators (RCs), and magnetic resonance imaging (MRI) for the prediction of a positive prostate biopsy as reflex tests after an elevated PSA level [111213141516]. This may support the process of shared, informed decision-making (Fig. 1) [17]; reduce the number of unnecessary biopsies by better identification of those men at risk for PCa; and better differentiate aggressive cancers from nonaggressive cancers.

Fig. 1

Flowchart of men with elevated PSA and/or abnormal results on a DRE with a combination of risk stratification tools. PCa, prostate cancer; PSA, prostate-specific antigen; DRE, digital rectal examination; MRI, magnetic resonance imaging. Adapted from Osses DF, et al. Int J Mol Sci 2019;20:1637 [17].

Several biomarkers that can provide a higher degree of predictive accuracy for PCa and csPCa than currently available clinical tools are now actively investigated as novel tools to improve the detection of PCa before prostate biopsy and reduce unnecessary prostate biopsies. Table 1 summarizes the results of studies investigating the performance of serum and urinary biomarkers for the detection of PCa and reduction of unnecessary prostate biopsies. This review highlights the role of existing and promising serum and urinary biomarkers for the detection of PCa and csPCa before prostate biopsy.

Table 1

Studies investigating the performance of currently available serum and urinary biomarkers

Biomarker	Reference	Year	No. of Result	AUC	Result
Serum biomarkers
PHI	Lazzeri et al. [27]	2012	222	0.67	At a phi cutoff of 28.8, 116 biopsies (52.25%) could be avoided, while 6 patients (8.4%) with PCa would have been missed. However, no patient with high-grade PCa would have been missed.
	Lazzeri et al. [28]	2013	158	0.73	At a phi cutoff of 25.5, 27 biopsies (17.2%) could be avoided, while 3 patients (4.2%) with PCa would have been missed and 2 patients (3.8%) with high-grade PCa would have been missed.
	de la Calle et al. [34]	2015	561	0.78	At a phi cutoff of 25, 40% of unnecessary biopsies could be avoided, and 25% of low-grade PCa would be reduced at the cost of missing 5% csPCa.
	Park et al. [30]	2018	246	0.76	At a phi cutoff of 22.9, 33 biopsies (21.3%) could be avoided, while 2 patients (1.3%) with PCa would have been missed. However, no patient with high-grade PCa would have been missed.
4K-score	Vickers et al. [39]	2008	740	0.83	Using a 20% risk of PCa as the threshold for biopsy, 424 (57%) biopsies could be avoided, while 31 (20.4%) of low-grade PCa and 3 (7.5%) of high-grade PCa would have been missed, respectively.
	Parekh et al. [43]	2015	1,012	0.82	Using a cutoff of 6% risk of csPCa, 30% of biopsies could be avoided, delaying diagnosis for 1.3% of patients with high-grade PCa.
	Braun et al. [44]	2016	749	0.78	Using a cutoff of 6% risk of csPCa, 17% of biopsies could be avoided, delaying diagnosis for 3.8% of patients with high-grade PCa.
Urine biomarkers
PCA3	Marks et al. [51]	2007	226	0.68	Using a PCA3 cutoff of 35, the PCA3 assay had a sensitivity of 58%, a specificity of 72%, and an OR of 3.6. At a PCA3 cutoff of less than 20, the NPV was 0.88 in men undergoing repeat prostate biopsy.
	Seisen et al. [54]	2015	138		Phi outperformed PCA3 for detecting csPCa (AUC, 0.80 vs. 0.55; p=0.03) in a comparative study.
	Cantiello et al. [55]	2015	156		Both phi and PCA3 significantly improved the predictive accuracy for the endpoint of extracapsular tumor extension. However, only phi provided significant incremental predictive accuracy for the prediction of tumor volume >0.5 mL, pathologic GS ≥7, seminal vesicle invasion, and composite endpoint of csPCa.
TMPRSS2:ERG	Tomlins et al. [58]	2011	606		TMPRSS2:ERG score is associated with indicators of csPCa at biopsy and prostatectomy, including tumor volume, high pathologic GS, and Gleason upgrading.
	Salami et al. [61]	2013	45		The combination of PCA3 and TMPRSS2:ERG can improve their ability to predict results of prostate biopsy (AUC=0.88; specificity=90% at 80% sensitivity).
	Leyten et al. [62]	2014	443		The AUC of ERSPC-RC increased from 0.799 to 0.842 when PCA3 and TMPRSS2:ERG scores were added. In multivariable logistic regression analyses, only TMPRSS2:ERG added significant predictive value to ERSPC-RC for predicting biopsy GS (OR, 7.16; p<0.001) and clinical tumor stage (OR, 2.60; p=0.023), whereas PCA3 did not.
	Tomlins et al. [63]	2016	1,244		The AUC of PCPT-RC increased from 0.639 to 0.762 after adding both PCA3 and TMPRSS2:ERG scores for predicting PCa. The AUC was also increased to 0.779 for predicting high-risk PCa, when both markers were added.
EPI test	McKiernan et al. [67]	2016	519	0.73	At a predetermined cutoff of 15.6, EPI test yielded an NPV of 91% and a sensitivity of 92%, with 27% of patients having an EPI score below the cutoff. Applying a cutoff from the training cohort to serve as a threshold for biopsy in the validation cohort decreased unnecessary biopsies by 27% of patients, while missing only 8% of high-grade cancers.
	McKiernan et al. [68]	2018	503	0.70	A validated cutoff of 15.6 would avoid 26% of unnecessary prostate biopsies and 20% of total biopsies, with NPV of 89% and missing 7% of high-grade PCa. An alternative cutoff of 20 would avoid 40% of unnecessary biopsies and 31% of total biopsies, with NPV of 89% and missing 11% of high-grade PCa.
SelectMDx	Van Neste et al. [71]	2016	905	0.76	When this gene expression was combined with PSA, PSAD, DRE, previous negative prostate biopsies, age, and family history in a multimodal model, the overall AUC was 0.90 in the training set and 0.86 in the validation set. A total reduction of biopsies by 42% and a decrease of unnecessary biopsies by 53% were observed in this model, with an NPV of 98% for high-grade PCa.
	Hendriks et al. [72]	2017	172	0.83a	SelectMDx scores are significantly higher in patients with suspicious lesions in mpMRI (p<0.01), with an AUC of 0.83 for the prediction of mpMRI outcome.

AUC, area under the curve; PHI, Prostate Health Index; PCa, prostate cancer; csPCa, clinically significant prostate cancer; OR, odds ratio; NPV, negative predictive value; GS, Gleason score; ERSPC-RC, European Randomized Study of Screening for Prostate Cancer - risk calculator; PCPT-RC, Prostate Cancer Prevention Trial - risk calculator; EPI, ExoDx Prostate Intelliscore; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; DRE, digital rectal examination; mpMRI, multiparametric magnetic resonance imaging.

a:An AUC for the prediction of mpMRI outcome.

PROSTATE HEALTH INDEX (PHI)

The phi is a simple blood test and a mathematical algorithm incorporating three different isoforms of PSA (total PSA [tPSA], free PSA [fPSA], and [−2]proPSA, i.e., p2PSA) combined in a mathematical formula: phi=(p2PSA/fPSA)×(tPSA)1/2. fPSA has three subforms: benign prostatic hyperplasia-associated PSA (BPSA), inactive PSA (iPSA), and proPSA (Fig. 2) [1819]. BPSA and iPSA are associated with benign tissues, whereas proPSA is associated with cancer [20]. It is possible to detect three truncated forms of proPSA (i.e., [−2], [−4], and [−5,−7]proPSA) in serum (Fig. 2) [1921]. Of these three truncated forms of proPSA, p2PSA is the most stable form and is specific for PCa [2022]. Development of the Access 2 immunoassay system (Beckman Coulter, Brea, CA, USA) by Beckman Coulter opened a new field of study for detecting PCa. Beckman Coulter developed the phi in partnership with the National Cancer Institute (NCI)'s Early Detection Research Network (EDRN), and the phi was approved by the U.S. Food and Drug Administration (FDA) in 2012.

Fig. 2

Molecular forms of PSA. PSA, prostate specific antigen; BPSA, benign prostatic hyperplasia-associated PSA; iPSA, inactive PSA; cPSA, complexed PSA; ACT, alpha I-antichymotrypsin; hK2, kallikrein-related peptidase 2. Adapted from Filella X, et al. Pharmgenomics Pers Med 2018;11:83–94 [19].

Numerous studies [23242526272829] have reported that %p2PSA (i.e., the percentage of p2PSA to fPSA) and phi provide significantly better clinical performance for predicting PCa than tPSA or %fPSA (Table 2). Phi has demonstrated a higher accuracy than tPSA and %fPSA for predicting the presence of PCa at biopsy, showing area under the curves (AUCs) from 0.709 to 0.76. Furthermore, phi is associated with the aggressiveness of the tumor, showing higher levels in patients with aggressive PCa [272829]. Similar results were shown by a recent prospective, multi-institutional study evaluating phi in Korean men [30], reporting AUCs of 0.76, 0.74, 0.69, and 0.56 for phi, %p2PSA, %fPSA, and tPSA, respectively, in patients with PSA in the gray zone (Table 2). There was also a significant association between phi and the Gleason score (GS) on biopsy (Spearman's rho=0.757; p<0.001). When phi was divided into quartiles, the proportion of high-grade PCa significantly and markedly increased as phi increased (Fig. 3). At the highest phi interval (phi>80), the chance of a positive biopsy result with high-grade PCa (GS≥7) was 100.0% (Fig. 3).

Table 2

Summarized results of predictive accuracy for prostate cancer of tPSA, %fPSA, %p2PSA, and phi

Reference	AUC tPSA (95% CI)	AUC %fPSA (95% CI)	AUC %p2PSA (95% CI)	AUC PHI (95% CI)
Park et al., 2018 [30] (n=246)
PSA≥3.5 (total)	0.683 (0.620–0.740)	0.68 (0.618–0.738)	0.761 (0.702–0.812)	0.797 (0.741–0.845)
3.5≤PSA<10 (subgroup)	0.556 (0.474–0.636)	0.685 (0.606–0.757)	0.74 (0.664–0.807)	0.763 (0.689–0.828)
Jansen et al., 2010 [24]
Rotterdam (n=405)	0.585 (0.535–0.634)	0.675 (0.627–0.721)	0.716 (0.669–0.759)	0.750 (0.704–0.791)
Innsbruck (n=351)	0.543 (0.473–0.594)	0.576 (0.523–0.629)	0.695 (0.644–0.743)	0.709 (0.658–0.756)
Sokoll et al., 2010 [25] (n=556)	0.58 (0.53–0.64)	0.66 (0.61–0.71)	0.70 (0.65–0.75)	0.76 (0.72–0.81)
Guazzoni et al., 2011 [26] (n=268)	0.53 (0.47–0.59)	0.58 (0.52–0.64)	0.76 (0.71–0.81)	0.76 (0.70–0.81)
Lazzeri et al., 2012a [27] (n=222)	0.52 (0.45–0.59)	0.60 (0.53–0.67)	0.72 (0.66–0.78)	0.67 (0.61–0.73)
Lazzeri et al., 2013b [28] (n=158)	0.55 (0.47–0.63)	0.60 (0.52–0.68)	0.73 (0.66–0.80)	0.73 (0.66–0.80)
Stephan et al., 2013 [29] (n=1,362)	0.56 (0.53–0.59)	0.61 (0.59–0.64)	0.72 (0.70–0.75)	0.74 (0.71–0.76)

tPSA, total prostate-specific antigen; %fPSA, percentage of free PSA to tPSA; %p2PSA, percentage of p2PSA to free PSA; PHI, Prostate Health Index; AUC, area under the curve; CI, confidence interval.

a:An observational prospective study of a clinical cohort of men with previous negative prostate biopsies.

b:A nested case-control study from multicenter European cohort, the PROMEtheuS database.

Fig. 3

Proportion of prostate cancer with Gleason score (GS)≥7 in relation to Prostate Health Index (phi) intervals.

Phi is the least expensive ($80 in the USA) of the currently available commercial multiplex biomarkers and is suggested in the initial and repeat biopsy setting [313233]. On average, using phi with a cutoff of ≥25 to biopsy could avoid 40% of biopsies and reduce 25% of GS 6 diagnoses at the cost of missing 5% of csPCa [34].

There has been a paradigm shift in PCa decision-making from a one-size-fits-all approach using tPSA toward multivariable risk assessment that takes the characteristics of individual patients into account. Given the substantial international evidence showing the superiority of phi over PSA, several tools have been created by combining phi with other clinical risk factors to aid in prostate biopsy decisions. For example, Lughezzani et al. [35] reported that the addition of phi to a multivariable model with age, prostate volume, digital rectal examination (DRE), and prostate biopsy history can lead to a statistically significant gain of 7% in predictive accuracy. They created a nomogram combining five variables that had an AUC of 0.80. Foley et al. [36] also created a multivariable phi-based nomogram including age, family history, DRE, previous negative biopsy, and either PSA or phi to aid in prostate biopsy decisions. The model using phi had an AUC of 0.77 for overall PCa and of 0.79 for highgrade disease. In a subset of men undergoing repeat prostate biopsy, the phi-based multivariable model had an AUC of 0.85 for any PCa and of 0.88 for high-grade disease. More recently, Loeb et al. [37] also reported that the addition of phi to the Prostate Cancer Prevention Trial risk calculator (PCPT-RC) and the ERSPC risk calculator (ERSPC-RC) can significantly improve the prediction of aggressive PCa (from AUC 0.58 to AUC 0.70 and from AUC 0.65 to AUC 0.71, respectively). These studies confirm that phi is a useful addition to multivariable nomograms for initial or repeat biopsy to improve the accuracy of risk stratification. ERSPC-RC including phi is easily accessible through the internet (www.prostatecancer-riskcalculator.com) or mobile applications (Fig. 4) [38] for easier use at the point of care.

Fig. 4

Screenshots of a smartphone app of the European Randomized Study of Screening for prostate cancer risk calculator (ERSPC-RC), which includes the Prostate Health Index (phi). Adapted from Zhang K, et al. Asian J Urol 2017;4:86–95 [38].

4K SCORE

Several studies [39404142] have indicated that a panel of four-kallikrein markers (tPSA, fPSA, iPSA, and kallikrein-related peptidase 2 [hK2]) can be used to improve the predictive accuracy of biopsy outcome and reduce unnecessary biopsies. Using data from the Swedish section of the ERSPC (n=740), Vickers et al. [39] reported that a panel of four-kallikrein markers shows significantly better predictive accuracy of biopsy outcome in previously unscreened men with elevated PSA than does PSA alone (AUC from 0.68 to 0.83 for those without DRE and from 0.72 to 0.84 for those with DRE). They estimated that using a 20% risk of PCa as the threshold for biopsy could reduce the number of biopsies by 424 (57%) while missing only a small number of cancers (31 of 152 low-grade cancers and 3 of 40 high-grade cancers).

The four-kallikrein panel is conceptually similar to phi by using a combination of PSA-based markers (tPSA, fPSA, iPSA, and hK2). This panel has been commercialized by Opko Diagnostics with the name of the 4K score. The test combines the four-kallikrein panel together with patient age, DRE, and history of prior biopsy in an algorithm that calculates an individual patient's risk of csPCa on biopsy. The 4K score is currently not available in Europe. It costs around $500 in the USA.

A multi-institutional prospective trial in the USA [43] has shown that the 4K score could discriminate between patients with and without GS ≥7 with an AUC of 0.821 in 1,012 men undergoing a prostate biopsy. The authors of that study indicated that 30% of biopsies could be avoided using a cutoff value of 6%, delaying diagnosis for 1.3% of patients with high-grade PCa. Using a cohort of 749 men, Braun et al. [44] reported similar results, indicating that 17% of biopsies could be avoided using a threshold of 6% risk of high-grade PCa, whereas diagnosis would be delayed for 3.8% of patients with high-grade PCa. The AUC is 0.784 for a model including age, the four-kallikrein panel, and DRE [44].

Lin et al. [45] recently evaluated the ability of the 4K score to predict the presence of high-grade PCa in 718 men enrolled for AS in nine centers. They showed that a clinical model including the 4K score significantly improved the accuracy of predicting reclassification compared with a clinical model including PSA at the first repeat biopsy, showing AUCs of 0.783 and 0.740, respectively (p=0.043). However, when the prediction of reclassification at subsequent biopsies was studied, there was no significant difference between the two models (AUC, 0.754 vs. 0.755).

The 4K score is currently recommended for patients undergoing initial and repeat biopsy. A systematic review to evaluate the performance of the 4K score in a prebiopsy setting showed a pooled AUC of above 0.80 for the discrimination of csPCa, which was highly consistent across 11 studies involving over 10,000 subjects [46]. On average, using the 4K score with a cutoff risk of 9% csPCa to indicate systematic biopsy could avoid 43% of biopsies at the cost of missing 2.4% of csPCa [394041424344].

In a comparative study including 531 Swedish men undergoing first-time biopsy, the phi and 4K score showed similar discriminative abilities for the prediction of all PCa (AUC, 0.70 vs. 0.69) and high-grade PCa (AUC, 0.71 vs. 0.72) [47]. Both are serum-based simple blood tests that can reduce the number of unnecessary biopsies compared with screening with tPSA, representing important new options for reducing harm.

PCA3

Prostate cancer antigen 3 (PCA3) is a prostate-specific noncoding messenger RNA (mRNA) that is highly overexpressed in specific PCa cell lines and prostatic tumors [4849]. PCA3 mRNA levels can be measured by using quantitative real-time polymerase chain reaction in a urine sample obtained after a prostate massage to obtain the maximum amount of prostatic cells. Measuring PSA mRNA allows for standardization of the number of PCA3 RNA copies by calculating the ratio of PCA3 to PSA (called the PCA3 score). The currently commercialized Progensa PCA3 test (Hologic Inc., Marlborough, MA, USA) is a robust urine test. It costs around $300 in the USA. According to the results of a meta-analysis [50], its overall sensitivity, specificity, and AUC values were 0.63, 0.88, and 0.82 in case-control studies and 0.65, 0.73, and 0.75 in prospective studies, respectively. That meta-analysis included 46 different studies, underlying that a cutoff of 35 was used in 26 institutions. However, the choice of the most appropriate cutoff for the PCA3 score remains controversial.

Marks et al. [51] demonstrated the superiority of the PCA3 over PSA by using a PCA3 assay in men (n=226) undergoing repeat prostate biopsy (AUC, 0.68 vs. 0.52; p=0.008). With the use of 35 as the most balanced PCA3 cutoff score, the PCA3 assay had a sensitivity of 58%, a specificity of 72%, and an odds ratio (OR) of 3.6. In particular, the authors reported that the negative predictive value (NPV) was 0.88 at a PCA3 cutoff of less than 20. Thus, the PCA3 score has remarkable utility in men who have one or more previous negative prostate biopsy results. The FDA approved this test as an aid for repeat prostate biopsy decisions in 2012.

Several studies [515253] have investigated correlations of the PCA3 score with features of PCa aggressiveness such as tumor volume, GS, pT stage, and percentage of positive biopsy cores. However, data about the association of PCA3 score with csPCa are conflicting. PCA3 is not associated with locally advanced disease [53]. Thus, its utility is limited in the prediction of aggressive cancer.

Comparative studies have demonstrated that the phi outperforms PCA3 for the prediction of csPCa on biopsy [5455]. Seisen et al. [54] reported that the phi outperforms PCA3 for detecting csPCa (AUC, 0.80 vs. 0.55; p=0.03). Similarly, Cantiello et al. [55] compared the performances of the phi and PCA3 to predict adverse pathologic features at radical prostatectomy. In a multivariable model, both the phi and PCA3 significantly improved the predictive accuracy for the endpoint of extracapsular tumor extension. However, only the phi provided significant incremental predictive accuracy for the prediction of tumor volume >0.5 mL, pathologic GS ≥7, seminal vesicle invasion, and the composite endpoint of csPCa [55]. Thus, many current guidelines recommend PCA3 with a cutoff of 35 in men with moderately elevated PSA for whom repeat biopsy is being considered.

TMPRSS2:ERG GENE FUSION

The TMPRSS2:ERG gene fusion is a genetic rearrangement of the TMPRSS2 gene (an androgen-regulated transcriptional promoter) and the ERG oncogene, which commonly occurs in PCa [56]. Similar to PCA3, a TMPRSS2:ERG rearrangement can be detected in urine after DRE [57]. It can also be normalized to the amount of PSA mRNA to generate a TMPRSS2:ERG score. Hessels et al. [57] reported that the TMPRSS2:ERG fusion gene has greater diagnostic accuracy than tPSA, with a high specificity of 93% and a PPV of 94% for detection of PCa. Unlike PCA3, a TMPRSS2:ERG score is associated with csPCa [5859]. Tomlins et al. [58] reported that the TMPRSS2:ERG score was associated with indicators of clinically significant cancer at biopsy and prostatectomy, including tumor volume, high pathologic GS, and Gleason upgrading in a large-scale multicenter study. Another population-based study found that TMPRSS2:ERG gene fusion was associated with an increased cumulative incidence ratio of 2.7 for developing metastases and PCa-specific mortality [59]. However, the low sensitivity of this biomarker reduces its value as a standalone test. Combining PCA3 with TMPRSS2:ERG can improve the prediction of PCa [575860616263]. Robert et al. [60] have shown that the combination of these two markers can lead to a higher sensitivity than TMPRSS2:ERG alone (sensitivity: 93.6% vs. 45.8%) while preserving its high specificity (98.8% vs. 97.5%) at the prostatic tissue level. Salami et al. [61] also confirmed that the combination of these two markers can improve their ability to predict the results of prostatic biopsies (AUC=0.88).

In a prospective multicenter study (n=443) [62], both PCA3 and TMPRSS2:ERG had independent additional predictive values over ERSPC-RC parameters for predicting PCa in multivariate analyses (OR, 3.64; p<0.001 and OR, 3.28; p=0.002, respectively). The AUC of ERSPC-RC increased from 0.799 to 0.842 when PCA3 and TMPRSS2:ERG scores were added. Interestingly, in multivariate logistic regression analyses, only TMPRSS2:ERG added significant predictive value to ERSPC-RC for predicting biopsy GS (OR, 7.16; p<0.001) and clinical tumor stage (OR, 2.60; p=0.023), whereas PCA3 did not. A prospective study by Tomlins et al. [63] has also shown the value of PCA3 and TMPRSS2:ERG scores when they are added to PCPT-RC. AUC was increased from 0.639 for PCPT-RC to 0.762 after adding both PCA3 and TMPRSS2:ERG scores. PCA3 and TMPRSS2:ERG scores were also evaluated to predict high-risk PCa, with an AUC of 0.779 when both markers were added.

As a commercial test, the MiProstate Score (MiPS; University of Michigan, Ann Arbor, MI, USA) incorporates PSA, PCA3, and TMPRSS2:ERG to predict the risk of PCa and csPCa. MiPS costs around $700 in the USA. It is a promising test following PSA screening. However, it has not yet been validated in prospective studies or directly compared with other biomarkers.

ExoDx PROSTATE INTELLISCORE (EPI)

The EPI test (Exosome Diagnostics, Boston, MA, USA) evaluates urine-based exosomal RNA expression levels of three genes, utilizing PCA3 and ERG (V-ets erythroblastosis virus E26 oncogene homologs) RNAs from urine normalized to SPDEF (SAM pointed domain-containing Ets transcription factor). Exosomes are small vesicles with doublelipid membranes that are secreted from cells. Exosomes encapsulate a portion of the parent cell cytoplasm. They are shed into various biofluids, including blood and urine (Fig. 5) [64]. They are rich sources of cellular proteins and RNAs. They are promising for profiling RNA expression in tumor cells because they are highly representative of their cells of origin and provide protection for mRNA during sample processing [65].

Fig. 5

Exosomes are extracellular vesicles secreted from cells. Exosomes encapsulate a portion of parent cell cytoplasm and are shed into various biofluids, including blood and urine. MVB, multivesicular body. Adapted from Shurtleff MJ, et al. Elife 2016;5:e19276 [64].

The EPI test is a simple urine test and somewhat unique since it does not require pre-catch DRE or special post-catch handling [66]. This test assigns an individual risk score for patients ranging from zero to 100 to help physicians evaluate a patient's risk for high-grade, potentially more aggressive PCa. Scores above a pre-defined cutoff are associated with an increased likelihood of high-grade PCa on a subsequent biopsy [67]. Thus, this test proposes to discriminate PCa with a GS of ≥7 from PCa with a GS of 6 and benign disease at initial biopsy.

McKiernan et al. [67] evaluated the diagnostic performance of the exosomal gene signature and demonstrated that the ExoDx panel used in combination with PSA, age, race, and family history was able to detect GS ≥7 PCa better than a standard-of-care variable alone (AUC, 0.73 vs. 0.63; p<0.001). A validation study also demonstrated its good assay performance [67]. A predetermined cutoff point (15.6) yielded an NPV of 91% and a sensitivity of 92%, with 27% of patients having an EPI score below the cutoff point. Applying a cutoff point from the training cohort to serve as a threshold for biopsy in the validation cohort decreased unnecessary biopsies by 27% (138 of 519) while missing only 8% (12 of 148) of GS ≥7 cancers. Recently, it was shown that the test can improve the identification of patients with higher grade disease and reduce the total number of unnecessary biopsies in a phase II study [68]. Thus, the ExoDx panel in combination with PSA, age, race, and family history has been successfully validated in over 1,000 patients across two prospective validation trials. However, this test is only available from the manufacturer through a CLIA-approved (Clinical Laboratory Improvement Amendments) laboratory. It has not been approved by the FDA. In current National Comprehensive Cancer Network (NCCN) guidelines, EPI is mentioned as an investigational biomarker only [69]. The NCCN panel states that additional evidence will be reviewed as it becomes available.

SelectMDx®

SelectMDx test (MDx Health, Irvine, CA, USA) is a gene expression assay performed with post-DRE urine that measures the mRNA levels of a two-gene panel (DLX1 and HOXC6) using reverse-transcription PCR. The expression of KLK3 is used as an internal reference. Leyten et al. [70] originally found that a validated urinary three-gene panel (HOXC6, TDRD1, and DLX1) showed higher accuracy for the detection of csPCa than PSA and the Progensa PCA3 test (AUC, 0.77 vs. 0.72 vs. 0.68, respectively). DLX1 and HOXC6 might be involved in the onset of PCa. They are also associated with PCa aggressiveness [70].

Van Neste et al. [71] collected post-DRE urine samples of 905 patients from two independent prospective clinical trials and evaluated the diagnostic value and clinical utility of the two-gene panel in prostate biopsy specimens for clinical validation. The assay was developed with an initial training set of 519 patients. It was then validated in a separate set of 386 patients from these trials. Using the expression of DLX1 and HOXC6 alone resulted in an AUC of 0.76, a sensitivity of 91%, a specificity of 36%, an NPV of 94%, and a PPV of 27% for the prediction of high-grade PCa. When gene expression was combined with PSA, PSA density, DRE, previous negative prostate biopsy results, age, and family history in a multimodal model, the overall AUC was 0.90 in the training set and 0.86 in the validation set. A total reduction of biopsies by 42% and a decrease of unnecessary biopsies by 53% were observed in this model, with an NPV of 98% for high-grade PCa. This means that a low-risk SelectMDx score is correlated with a 90% probability that a man does not have PCa and a 98% probability that he does not have high-grade PCa. Recently, Hendriks et al. [72] found that SelectMDx scores are significantly higher in patients with suspicious lesions in mpMRI (p<0.01), with an AUC of 0.83 for the prediction of mpMRI outcome. In current NCCN guidelines, SelectMDx is also mentioned as an investigational biomarker [69]. The NCCN panel states that additional evidence will be reviewed as it becomes available.

ConfirmMDx®

Sampling errors inherent with the random prostate biopsy procedure can result in a false-negative rate of approximately 25% for standard-of-care histopathology [73]. Previous studies on repeated biopsy procedures have shown that initial prostate biopsy histopathology has a false-negative rate of 20% to 30% [7475]. Repeat biopsies are common in men with previous histopathologically negative findings in an attempt to detect missed PCa.

ConfirmMDx (MDx Health) is a tissue-based multiplex epigenetic assay that aims to evaluate cancer-negative men being considered for repeat prostate biopsy due to still high PSA. The assay is commercially available. It uses multiplex methylation-specific PCR to measure the epigenetic status of PCa-associated genes GSTP1, APC, and RASSF1 in residual cancer-negative prostate biopsy core tissue samples [7677]. By detecting epigenetic abnormalities in a halo around the tumor, which is shown to be associated with oncogenesis, these biomarkers aid in finding evidence of occult PCa unseen by histopathology [767778]. This field effect on adjacent benign-appearing biopsy core tissues is a strong independent predictor for diagnosing PCa in a subsequent biopsy, with a negative epigenetic result providing a higher NPV of approximately 90% than standard histopathology alone [757677]. Test results from this epigenetic assay help guide decision-making for repeat biopsy for patients with a previous negative biopsy result who are still considered to be at risk for PCa.

ConfirmMDx has been validated in two blinded multicenter studies that show the superior NPV of this epigenetic test over standard histopathology for cancer detection in prostate biopsies [7677]. The European MATLOC study [76] blindly tested this assay using archived prostate biopsy needle core tissue samples of 498 subjects with histopathologically negative prostate biopsies followed by positive or negative repeat biopsy within 30 months. The NPV was 90% (95% confidence interval [CI], 87%–93%). In multivariate analysis, ConfirmMDx was a significant independent predictor of patient outcome (OR, 3.17; 95% CI, 1.81–5.53). A similar validation study was performed in the United States using archived cancer-negative prostate biopsy core tissue samples of 350 subjects who had repeat biopsy within 24 months from a total of five urological centers [77]. The NPV was 88% (95% CI, 85%–91%). The test was again found to be the most significant independent predictor of patient outcomes on multivariate analysis (OR, 2.69; 95% CI, 1.60–4.51). These two independent trials showed that the negative findings of this test could be used to decrease concerns about unsampled cancer and effectively avoid unnecessary repeat biopsies [7677]. ConfirmMDx is also available from one CLIA-certified laboratory. It is not FDA-approved. ConfirmMDx can be considered as an option for men contemplating repeat biopsy because the assay may identify individuals at higher risk for PCa on repeat biopsy [69].

CONCLUSIONS

PSA screening has been controversial, leading to an intensive search for alternative PCa biomarkers with better diagnostic and predictive potentials. In particular, there is a quest for biomarkers that can distinguish between aggressive and indolent tumors, thereby leading to better treatment decisions. Several FDA-approved and clinical laboratory-based tests have been developed. These tests show improved sensitivity and specificity over PSA. Nevertheless, serum PSA is still being used in conjunction with other parameters, highlighting the fact that PSA remains an indispensable tool in the clinical management of men with PCa. Emerging alternative biomarkers may continue to supplement or possibly replace PSA over time. These novel biomarkers appear to have better detection and diagnostic values or better prognostic abilities and predictive values. Careful validation of emerging biomarkers may fulfill the so far unmet clinical challenges and guide clinicians to better diagnoses and better treatment options for PCa. In the future, more high-level evidence is needed to answer whether the rational use of the currently available novel biomarkers can effectively decrease biopsy rates and thus decrease costs and morbidity. Such high-level evidence is needed to justify and advocate the broader use of these novel biomarkers.