Molecular diagnostics of prostate cancer: impact of molecular tests

Abstract

Prostate cancer (PCa) is the second leading cause of cancer-related death among men. Prostate-specific antigen (PSA) testing is used in screening programs for early detection with a consequent reduction of PCa-specific mortality at the cost of overdiagnosis and overtreatment of the nonaggressive PCa. Recently, several assays have been commercially developed to implement PCa diagnosis, but they have not been included in both screening and diagnosis of PCa. This review aims to describe the actual and novel commercially available molecular biomarkers that can be used in PCa management to implement and tailor the screening and diagnosis of PCa.

INTRODUCTION

Prostate cancer (PCa) is the second leading cause of cancer-related death among men, with subclinical PCa being quite common in men over 50 years.1 Age, ethnicity, and familiar history of PCa are relevant risk factors.1 Prostate-specific antigen (PSA) testing has provided a method for early detection of PCa, and its use in screening programs has reduced PCa-specific mortality by 25%; however, overdiagnosis and overtreatment of the nonaggressive PCa have been the consequences of its application.1,2 The diagnosis of PCa is based on clinical and imaging investigations and microscopic features, as summarized in Figure 1. Fine-needle biopsies are performed for suspicious lesions, establishing PCa diagnosis on microscopic criteria.3 The correlation of morphological features with additional immunohistochemical stains may also be helpful in some cases for a correct diagnosis.3

Figure 1.

Diagnostic workflow of prostate cancer (PCa). PSA: prostate-specific antigen; MRI: magnetic resonance imaging; PCa: prostate cancer.

PCa is a highly heterogeneous tumor, clinically, morphologically, and molecularly. Multifocality of primary PCa is frequently observed and has been supported molecularly by the presence of clonally distinct lesions within a patient.4 Multifocality and multiclonality are challenging for diagnosing PCa in fine-needle biopsies and are the main factors for the underperformance of molecular biomarkers.4 In reflex to the abovementioned features and difficulties related to PCa, the conventional portfolio of biomarkers is limited in clinics. Except PSA and the recent introduction of genomic testing in metastatic castration-resistant PCa (mCRPC), no other biomarkers have been implemented in clinical practice. Nonetheless, in the era of personalized medicine, there is a need to improve PSA screening to avoid unnecessary biopsies, overdiagnosis, and overtreatment. Several analytical assays were commercialized for urine and serum analysis to support this need, but they have yet to be included in official guidelines for the management and/or screening of PCa.

Taken this scenario, the present review aims to describe the molecular biomarkers used in PCa management and focus on possible implementations for tailoring the screening and the diagnosis. To this aim, the biomarkers will be divided into liquid and tissue biomarkers for potential applications.

LIQUID BIOMARKERS

Screening procedures for PCa are based on PSA detection by blood tests. In case of a positive level over 2.5 ng ml⁻¹, the diagnostic workflow is described in Figure 1. In many cases, biopsies are negative despite PSA positivity over the normal value. The possibility of avoiding unnecessary prostate biopsies is strictly linked to the robustness of liquid biomarkers. Although, at present, they are not validated and universally used, several commercially available tests have proven their capability in stratifying PCa patients with mildly elevated PSA, typically between 2.5 ng ml⁻¹ and 10 ng ml⁻¹. The main characteristic of those tests is to improve the poor specificity of PSA and avoid the risks associated with unnecessary biopsies, including the risk of overdiagnosis of Grade Group 1 (GG1) PCa in patients with a low probability of harboring GG2-positive disease.5 Those newer biomarkers range from serum to urine analysis.

Blood-based tests

Recent assays based on serum biomarkers’ detection are primarily associated with PSA specialization. There are different forms of PSA; the free PSA (fPSA) is usually associated with benign conditions. Nonetheless, the fPSA consists of three other forms: the benign PSA, the intact inactive PSA, and the PSA precursor (proPSA), which is associated with PCa.6 In turn, the proPSA has several molecular isoforms, and the most suitable for serum analysis is the [-2]pro-prostate-specific antigen (p2PSA).6 An automated immunoassay was developed in 2010 for its detection, and an index, namely the prostate health index (PHI), was generated by combining the values of total PSA, fPSA, and p2PSA in a specific formula. In 2012, the USA Food and Drug Administration (FDA) approved PHI for PCa detection in a subset of men with specific characteristics.7 Several reports have evaluated the performance of PHI with variable sensitivities and specificities, which were merged in a recent meta-analysis.6 Overall, PHI seems to be an appropriate tool for PCa7 and a better predictor than fPSA for finding PCa at biopsy.8 Another specialization of the PSA test is the IsoPSA® (Cleveland Dx, Cleveland, OH, USA), which detects the PSA isoform composition in blood by combining a physicochemical partitioning of different structural variants of the PSA with conventional immunoassays.9 Based on the results of a multicentric study, Klein and colleagues suggest omitting MRI and biopsy if the IsoPSA index is ≤6 because of the low risk of high-grade disease. Contrarily, for an IsoPSA Index >6, the suggestion is to perform the biopsy regardless of MRI findings.9 Furthermore, in men with total PSA ≥4 ng ml⁻¹, IsoPSA testing resulted in a 55% net reduction in recommendations for prostate biopsy.10

In line with the abovementioned assays, the 4KScore incorporates a panel of four kallikrein protein biomarkers, namely total PSA, fPSA, intact PSA, and human kallikrein-related peptidase 2, with other clinical information in an algorithm providing a risk for a high-grade (Gleason score ≥7) cancer on biopsy.11 Recently, the 4KScore was tested in a multicentric study including 1111 men to investigate the possibility of using the test combined with a multiparametric MRI to avoid biopsies.12 The reported data are somewhat inconclusive, stressing the need to carefully weigh the risk of an unnecessary biopsy versus the risk of a missed clinically significant PCa. As a possible consequence, the more biopsies are avoided, the more clinically significant prostate cancers are missed.12

Total PSA and fPSA have been further included in another blood-based test, the Proclarix. This, in addition to men’s age, provides for the detection of thrombospondin 1 (THBS1) and cathepsin D (CTSD) to compute a risk score, which has already shown its analytical and clinical performance.13,14 Recently, the risk score has also been correlated with PCa aggressiveness.15

In a more comprehensive asset including single-nucleotide polymorphisms (SNPs) and familial data, the Karolinska Institute developed the Stockholm 3 (STHLM3) algorithm as an alternative to PSA testing to improve the early detection of clinically significant PCa.16 The STHLM3, in addition to PSA, includes the detection of four additional plasma proteins, 101 single-nucleotide polymorphisms, and the patient’s clinical information (including the age, the family history, previous prostate biopsy, and the use of 5-alpha reductase inhibitors).17 The possible introduction of the STHLM3 tests has been estimated to cause a reduction of biopsied men by 53%.18 Thanks to the comprehensive asset of the test, compared to the PSA test as a possible initial screening, the STHLM3 testing showed improved incremental effectiveness in reducing biopsies and detecting clinically significant PCa, but at additional costs.17

In this scenario, there are relevant analytical tools that can implement the management of PCa with the aim of reducing unnecessary biopsies. The blood-based assays described here can be used to substitute classical PSA screening or implement PSA tests in specific cases. Their use in clinical practice should be validated in dedicated studies to define the proper workflow, which needs to be presented in international guidelines for PCa management.

Urine-based assays

The urine is a promising fluid for the detection of biomarkers for PCa. Biomarkers from PCa cells can be released into prostatic fluids and then into urine for easy detection.19 After digital rectal manipulation of the prostate, an enrichment of PCa biomarkers can be achieved. With various applications, those biomarkers range from PCa cells, DNAs, RNAs, proteins, to other small molecules. Several urine assays have been commercialized to provide help in PCa detection. The first to be approved by the FDA was the prostate cancer antigen 3 (PCA3) test.20 PCA3, or DD3, is a long noncoding RNA, which is overexpressed in PCa. The commercial assay used to detect PCA3 also measures the mRNA level of PSA, which is used for normalization.20 The mean PCA3 scores were significantly lower in men with biopsy Gleason score <7 versus Gleason score ≥7.21 According to the study of van Poppel et al.,21 a threshold of 20 in PCA3 score may be helpful in selecting men with clinically insignificant PCa in whom active surveillance may be appropriate. In contrast, a threshold of 50 may identify men at high risk of having a significant prostate cancer to be submitted to radical prostatectomy.21 On multivariable logistic regression, men with a PCA3 score <25 were 4.56 times as likely to have a negative repeat biopsy as men with a score of 25 or greater.22 In a meta-analysis reporting data from 12 295 patients, the sensitivity of the PCA3 test was 65%, the specificity was 73%, and the area under the curve (AUC) was 0.7482, supporting a good, but not excellent diagnostic performance.23 Combining PCA3 with the detection of the RNA of the fusion gene TMPRSS2:ERG (between the transmembrane protease, serine 2 gene [TMPRSS2] and the erythroblast transformation-specific-related gene [ERG]), which occurs in approximately 50% of prostate cancers and is associated with an aggressive PCa phenotype,24,25 improved specificity for predicting aggressive PCa.26 MyProstateScore (MPS), is a further evolution of PCA3 test, which combines, in addition to TMPRSS2:ERG, serum PSA levels in locked regression models providing values from 0 (very unlikely to detect grade group 2, GG≥2, cancer on prostate biopsy) to 100 (very likely to detect GG≥2 cancer on prostate biopsy).27 An MPS cutoff of 25 would have prevented 39% of biopsies, while missing only one GG≥2 cancer (6%).27 In postdigital rectal exam (DRE) urine, the SelectMDx test quantifies the expression levels of the homeobox C6-gene (HOXC6) and the distal-less homeobox 1-gene (DLX1), two genes that are overexpressed in aggressive PCa. The test also combines clinical variables, including DRE result, age, and PSA density.28 The use of SelectMDx was estimated to have a potential reduction of 40.7% in biopsies.28 The synergistic combination of SelectMDx and Prostate Imaging Reporting and Data System (PI-RADS) into a sum score has shown an increased sensitivity to avoid unnecessary biopsies.29

With a more straightforward approach, two other assays were developed to help PCa diagnosis in urine samples. Those tests are the Exosome Dx Prostate IntelliScore (EPI) and the miR Sentinel. Both are based on analyzing different RNA populations included in urine exosomes. The EPI test quantifies 3 exosome-derived urine biomarkers by polymerase chain reaction (PCR) without the need for other clinical information, included PSA.30 In detail, the test measures the expression levels of PCA3, ERG, and a control gene, the SAM-pointed domain-containing Ets transcription factor (SPEDF).20 In an analysis of over 1000 men from two cohorts of men aged ≥50 years and PSA ranging from 2 ng ml⁻¹ to 10 ng ml⁻¹, scheduled for initial prostate biopsy, the EPI test predicted clinically significant PCa on initial biopsy. Based on an expert consensus panel, an EPI score of >15.6 would discriminate patients as high risk for GG2 PCa on their initial biopsy.31 The follow-up analysis of the abovementioned study demonstrated that men with EPI low-risk scores (<15.6) significantly deferred the time to the first biopsy and remained at a low pathologic risk by 2.5 years after the initial study.32 The EPI test is also available in Europe, where it received the mark for European In Vitro Diagnostics (CE-IVD) and showed similar performance to the previously published EPI laboratory-developed tests (LDT) in a cohort of patients primarily from European clinical sites.33

At the urine exosome level, the miR Sentinel tests interrogate small noncoding RNAs in three independent tests: the Sentinel PCa, CS, and HG. In detail, the Sentinel PCa test classifies patients with and without PCa. For cancer patients, the Sentinel HG test discriminates patients with higher-grade (GG3–5) PCa who can benefit from therapeutic intervention from those with lower-grade disease (GG1 and GG2) who are candidates for active surveillance. Those lower-grade patients in active surveillance can be monitored longitudinally for evidence of molecular progression using the Sentinel HG test or the Sentinel CS test, distinguishing between low-grade disease (GG1) and higher (GG2–5).34,35 Both EPI and miR Sentinel received FDA Grants for Breakthrough Device Designation.

TISSUE BIOMARKERS

In the conventional workflow of the PCa diagnosis, the diagnostic confirmation relies on the histological diagnosis of the biopsies. Gleason score of PCa, which quantifies the intratumor heterogeneity of PCa, continues to be the most prognostic indicator of PCa to date.36 Although at the tissue level, PCa diagnosis is mainly based on microscopic features, several immunohistochemical (IHC) markers have proven their clinical utility and are routinely used.37 In particular, the absence of basal cells, a major criterion for diagnosing PCa, can be confirmed by immunohistochemical markers. The so-called prostatic intraepithelial neoplasia (PIN) cocktail combines a double IHC stain with p63 and alpha-methylacyl-CoA racemase (AMACR) for the diagnosis of high-grade PIN and PCa,3,37 as shown in Table 1. High-molecular-weight cytokeratins such as CK5/6/14 can be used as alternatives to p63 in separate confirming reactions.37 Those biomarkers are helpful in fine-needle biopsy analysis.

Table 1.

Immunohistochemical profile for prostate cancer using prostatic intraepithelial neoplasia cocktail

Diagnosis	p63 profile	AMACR profile
Neoplastic prostatic gland	Negative (lack of myoepithelial cells)	Positive glands
High-grade PIN	Positive (surrounding glands)	Positive glands
Normal prostatic gland	Positive (surrounding glands)	Negative glands

PCa: prostate cancer; AMACR: alpha-methylacyl-CoA racemase; PIN: prostatic intraepithelial neoplasia

Although the prognosis of PCa is mainly based on the Gleason score, common IHC biomarkers have proven their possible utility in defining the aggressiveness of PCa. Among them, the well-known proliferation marker Ki67 at the transcriptomic level was related to PCa aggressiveness.38 In IHC tests, when expressed in over 10% or 15% of PCa cells, Ki67 has proven to predict outcomes and to be a prognostic factor in PCa patients.38,39 Biomarkers helping the diagnosis of metastatic PCa are the PSA, the prostate-specific membrane antigen (PSMA), and the NK3 homeobox 1 (NKX3.1). The latter is an androgen-regulated tumor suppressor gene strongly expressed in the nuclei of normal prostatic epithelium and with higher intensity in tumor tissue.40 Those markers are useful in identifying prostatic origin in metastatic carcinomas and in diagnosing bone metastases from PCa.41

Other IHC markers studied in PCa are those related to neuroendocrine differentiation, such as chromogranin A (CgA), synaptophysin, and neuron-specific enolase.39 Neuroendocrine differentiation usually occurs in PCa due to selective pressure from androgen deprivation therapy or androgen receptor antagonists used for treatment. Therefore, it represents a sign of the disease progression in response to treatment. Recently, Ostano et al.42 identified molecular features including protein-coding, long noncoding, and microRNAs, at the time of surgery suggestive of the neuroendocrine (NE) transformation process of PCa. Among neuroendocrine differentiation (NED) markers, CgA expression was reported as an independent adverse prognostic factor for metastatic PCa, both hormone-sensitive and castration-resistant PCa.43 However, CgA expression at the primary core biopsy does not seem to influence the survival of patients with Gleason score 7–10.44 Taken the published observations, the increment of CgA positivity from hormone-sensitive to castration-resistant metastatic PCa could be a negative prognosticator, which could help the clinical evaluation of advanced-stage patients with distant metastases.43

In addition to conventional IHC analyses, molecular methods have proven their utility in detecting occult PCa in apparently negative biopsies. This is the case of ConfirmMDx, a tissue-based molecular test aimed at detecting the methylation status of several genes, namely the glutathione S-transferase pi 1 (GSTP1), the adenomatous polyposis coli (APC), and the Ras-association domain family member 1 (RASSF1).20 A positive methylation result for at least one gene in at least one tissue core produces a positive assay result.45 DNA methylation of GSTP1, APC, and/or RASSF1 in initial, histopathologically cancer-negative biopsies was reported as a significant independent predictor of unsampled PCa.45 The ConfirmMDx assay has proven to help risk stratification in Afro-American men who had an initial negative biopsy.46

Germline and somatic testing for homologous recombination genes has been added to the biomarkers portfolio for PCa, in particular mCRPC to select patients eligible for poly(ADP-ribose) polymerase inhibitor (PARPi) therapy. Therefore, the detection of homologous recombination repair genes alteration in mCRPC, particularly those involving the breast cancer genes, breast cancer 1, early onset (BRCA1) and BRCA2, is important to guide treatment.47 In detail, guidelines and consensus statements recommend genetic testing (germline and/or somatic) for men with metastatic PCa, both hormone sensitive and castrate resistant.47 In addition, major guidelines recommend germline and somatic testing for men with personal or family history of PCa, Lynch syndrome or Ashkenazi Jewish ancestry, and early onset disease.47 There are several additional specifications related to national scientific or oncological societies on the eligibility of genetic/somatic testing. This is the case of the Italian Scientific Societies recommending germline BRCA testing for men with a family history of hereditary breast or ovarian cancer or paternal family with breast or ovarian cancer.48 Approved tests are based on next-generation sequencing (NGS) and can include genes with moderate- to high-risk hereditary cancer susceptibility, namely the homologous recombination repair genes (BRCA1, BRCA2, the checkpoint kinase 2 [CHEK2], the ataxia telangiectasia mutated serine/threonine kinase [ATM], the partner and localizer of BRCA2 [PALB2], and the DNA repair protein RAD51 homolog 4 [RAD51D]), the mismatch repair genes (mutL homolog 1 [MLH1], mutS homolog 2 [MSH2], mutS homolog 6 [MSH6], and PMS1 homolog 2, mismatch repair system component [PMS2]), and the pathogenic variant homeobox B13 (HOXB13), which seems to be the only PCa-specific marker gene.49 In the USA, the FDA approved specific companion diagnostics for PARPi treatment in PCa. The European Society of Molecular Oncology (ESMO) Precision Medicine Working Group recommends tumor multigene panel tests for somatic alterations in PCa. In detail, in countries where PARP inhibitors are accessible for patients with PCa, multigene panel testing on tumor samples assessing the mutational status of, at least, BRCA1/2 is recommended.50 The impact of genetic testing on PCa survival and treatment effectiveness is ongoing. The generation of tumor registry of PCa outcomes, such as the PROMISE, for sure can help in understanding the disease characteristics and treatment responses across the disease spectrum for PCa genetic variants.51

CONCLUSIONS

The development of novel tests for the specialization of PSA screening calls for implementing screening guidelines for PCa with the goal of reducing biopsies and overtreatment. The definition of reflex testing with early diagnosis biomarkers in men with elevated PSA would be a possible and reliable solution. An ideal workflow would be based, as represented in Figure 2, on integrating different biomarkers at different levels of the screening and diagnostic procedure. Taking the analytical possibilities at the level of blood and urine testing, there is the need to validate the proper Reflex testing that could accurately identify high-grade PCa with a consequent improvement of the net benefit of screening.52 Overall, liquid biomarkers described in this review have proven additional clinical utility to the current diagnostic standard of care.53 The choice of the proper test in a reflex testing scenario is not simple and should consider several aspects, including the workflow, the choice of the sample for the analysis (urine or blood), the best analytical method among the available ones, and the overall cost of the test. Maybe in this panorama, if blood and urine biomarkers could have the same performance as negative and positive predictive values, the reflex testing would benefit from using blood samples, because except PSA, no other samples would be requested for the analysis. In any case, the introduction of reflex testing in PCa management could only derive from concerted multicentric and international studies comparing the tests and defining international guidelines to apply.

Figure 2.

Possible workflow of prostate cancer diagnosis with reflex testing. PSA: prostate-specific antigen; DRE: digital rectal exam; PCA3: prostate cancer antigen 3; TMPRSS2:ERG: the fusion gene between the transmembrane protease, serine 2 gene (TMPRSS2) and the erythroblast transformation-specific-related gene (ERG); MPS: MyProstateScore; PHI: prostate health index; IsoPSA: PSA isoform; STHLM3: Stockholm 3.

AUTHOR CONTRIBUTIONS

SB designed the work and drafted the manuscript. EA and SB collected, analyzed, and interpreted the data. EA performed the critical revision of this manuscript. Both authors read and approved the final manuscript.

COMPETING INTERESTS

Both authors declare no competing interests.