Prostate cancer is the most common noncutaneous malignancy in American men. In 2016, approximately 180,890 men were diagnosed with prostate cancer and 26,120 men died from the disease in the United States.1 Unfortunately, prostate cancer encompasses a complex and heterogeneous disease spectrum. Physicians and patients are faced with the ambiguity of who should be screened, biopsied, rebiopsied, treated, or provided with adjuvant therapy.
Advances in urology have led to an array of blood-, urine-, and genetic-based tests marketed toward providing personalized outcomes. These outcomes and treatments are especially important in prostate cancer given the diverse nature of the disease, plethora of treatment options, risks of morbidity, and quality of life.2,3
This article reviews two blood-, one urine-, and five genetic-based tests that can help facilitate the physician-patient decision process regarding prostate cancer management. An emphasis is placed on describing the indication for each test, what specimen is needed, how results are reported, and a brief review of the development of each test. Tests are grouped by function and who to biopsy, rebiopsy, treat, and who may benefit from adjuvant therapy following prostatectomy.
The 4Kscore® (OPKO Health, Miami, FL) is designed to help identify patients who may benefit from a prostate biopsy or rebiopsy after an abnormal prostate-specific antigen (PSA) score or a clinical finding suspicious for prostate cancer. The 4Kscore measures the plasma levels of four different prostate-derived kallikrein proteins: total PSA, free PSA, intact PSA, and human kallikrein-related peptidase 2. Additionally, it incorporates these values, as well as age, digital rectal examination (DRE) results (nodule/no nodule), and previous biopsy status (previous biopsy/no previous biopsy) to assess a patient’s risk of having a Gleason score ≥ 7 on subsequent prostate biopsy. The results provide patients with a personal positive predicative value of having aggressive prostate cancer.4-8
The development of the 4Kscore included early retrospective studies using data from European subjects. Area under the curve (AUC) measurements for prostate cancer detection improved in three populations when the 4-kallikrein panel was incorporated with clinical models including age, PSA level, and DRE results (improvements in AUC with 4-kallikrein panel, 0.72-0.84, 0.70-0.78, and 0.63-0.78, respectively).4-6 Applying the clinical model to these populations and recommending biopsy when the probability of prostate cancer was ≥ 20% would have led to advising against biopsies in approximately 50% of men with an elevated PSA level.5,6 In those men in whom biopsy was advised against, 13% would harbor prostate cancer, of whom 82%, 15%, and 3% would be Gleason grade ≤ 6, 7, and ≥ 8, respectively.5 A prospective US-based validation study found that the 4Kscore had superior ability to detect Gleason ≥ 7 prostate cancer compared with a modified Prostate Cancer Prevention Trial Risk Calculator with AUC 0.82 versus 0.74 (P < .0001). Depending on the 4Kscore threshold, applied biopsy reduction ranged from 30% to 58%, with delay in diagnosis in 1.3% to 4.7% of Gleason grade ≥ 7 prostate cancers.7 The clinical utility of the 4Kscore was shown to be influential in 89% of the physician-patient decisions regarding whether or not to have a prostate biopsy performed, and decreased the rate of biopsies by 65% in one study.8
The 4Kscore is a validated clinical decision aid, not US Food and Drug Administration (FDA)-approved, that has been proven to reduce the number of biopsies in men with elevated PSA levels. Therefore, incorporation of the 4Kscore may improve clinical decision making while reducing potential harms associated with unnecessary prostate biopsies.
The Prostate Health Index (phi) (Beckman Coulter, Brea, CA) can be used in men aged ≥ 50 years with a PSA level between 4 and 10 ng/mL with no suspicion of prostate cancer upon DRE to help determine the need for biopsy. The phi score includes PSA, free PSA, and proenzyme PSA (proPSA) measurements. proPSA is an enzymatically inactive precursor of PSA that is associated with prostate cancer, which is found in both the peripheral zone of the prostate and the serum. The results of the phi score are translated into probability of cancer on biopsy.9
A prospective study, including patients with PSA values between 2 and 10 ng/mL and a negative DRE result, found phi to be highly specific for prostate cancer detection. Additionally, an increased phi score correlated with an increased risk for prostate cancer and Gleason grade ≥ 7.9 The phi score has greater predictive accuracy of Gleason grade 7 disease than PSA, free PSA, or proPSA measurements, individually.10 Clinical utility studies confirmed the ability of phi to predict Gleason grade ≥ 7 disease in biopsy-naive men and potentially reduce biopsy rates. For example, when using a phi score cutoff of 24, 41% of prostate biopsies could possibly be prevented while still maintaining high sensitivity and specificity for significant prostatic disease.11 Incorporating the phi score may help to determine the need for a prostate biopsy in biopsy-naive or previously biopsied men.
ConfirmMDx® (MDx Health, Irvine, CA) is indicated for men who have a documented history of a previous negative prostate biopsy result for cancer or cellular atypia suspicious for cancer. ConfirmMDx uses the methylation status of three biomarkers (GSTP1, RASSF1, and APC) from prostatic tissue to help determine a patient’s chance for having prostate cancer on a subsequent biopsy. Two studies used to develop and validate the test reported a positive test result upon the methylation of at least one gene in a single biopsy sample.12,13
In a retrospective design, the Methylation Events to Locate Occult Cancer (MATLOC) and the subsequent multicenter validation Detection of Cancer Using Methylated Events in Negative Tissue (DOCUMENT) study assessed the methylation status of three genes, GSTP1, APC, and RASSF1, in relation to ACTB.12,13 These genes are involved in DNA damage repair, tumor cell invasion/metastasis, and cell cycle control, and have been reported to be present in upward of 90%, 80%, and 54% of prostate cancer tissues, respectively. The results of both the MATLOC and DOCUMENT studies reported that the methylation patterns of these three genes had a negative predictive value of 88% to 90% for prostate cancer and could potentially decrease rebiopsy rates by 64%.12,13 To assess clinical utility, a field observation study of 138 men from 5 practices, all of whom had negative ConfirmMDx results, were pooled and showed that less than 5% of these men had repeat biopsies.14 A larger prospective study assessing clinical utility, the Prostate Assay Specific Clinical Utility at Launch (PASCUAL) study, recently completed enrollment and results should be forthcoming.
ConfirmMDx is a validated test that can predict the chance of a repeat negative prostate biopsy result. Clinically it can help a physician and patient determine the need for a repeat biopsy.
Progensa® (Gen-Probe, San Diego, CA) prostate cancer antigen-3 (PCA3) urine assay is another test helpful in determining management after a negative prostate biopsy result. Following a DRE, PCA3 score is calculated from a patient’s urine. The PCA3 score is a continuum with a cutoff score of 35 used to estimate a patient’s risk for prostate cancer on a subsequent prostate biopsy. A higher score translates into higher risk. The amount of PCA3 present is normalized to the amount of PSA in the urine.15
The PCA3 gene is specific to prostatic tissue and has been found to be increased in 95% of prostate cancer specimens. Two independent studies completed in the United States and Europe reported that a PCA3 cutoff score of 35 had a specificity of 72% for detecting prostate cancer.16,17 Stated differently, a patient with a PCA3 score less than 35 has a 28% chance of having prostate cancer on a subsequent biopsy. The PCA3 test is minimally invasive and can provide additional information to patients and physicians when they are considering a repeat prostate biopsy.
Prolaris® (Myriad Genetic Laboratories, Salt Lake City, UT) was the first commercially available gene-based assay to evaluate tumor aggressiveness. The Prolaris ScoreTM is a quantitative measure of the RNA expression levels based on a 46-gene expression signature. It incorporates the expression of 31 genes involved only in cell cycle progression (CCP), which are normalized against 15 housekeeping genes, to quantify prostate cancer cellular proliferation. This assay was determined from 126 cell cycle genes, and did not focus on other biologic pathways. The test’s endpoint is based on correlation with prostate tumor cell proliferation, low gene expression associated with a low risk of disease progression, and high gene expression associated with disease progression.
The results of the Prolaris Score report an individual’s cancer aggressiveness and risk of mortality from prostate cancer. It allows the patient to compare his results within well-defined risk groups (very low, low, and intermediate risk). Results indicate that for each 1-unit increase in the Prolaris Score, patients had approximately twice the risk of dying from prostate cancer over 10 years.
The clinical development of the Prolaris Score first reported results of 761 men from the United Kingdom who had prostate cancer diagnosed with needle biopsy from 1990 to 2004 and who were “conservatively managed” with watchful waiting (WW). Tumor tissue from this prospective cohort was later retrospectively analyzed with the Prolaris test. The investigational and validation study found that the CCP score predicted biochemical reoccurrence (BCR) and that the hazard ratio for each unit of increase in CCP was 1.89 and 2.10.18,19 Using prostate biopsy samples, investigators observed a similar hazard ratio of 1.60 and 2.55 for BCR per unit increase in CCP in men who later had a prostatectomy and external beam radiation therapy, respectively.20,21 When using the CCP as a predictor of mortality from prostate cancer, a hazard ratio of 2.02 and 2.08 per unit increase in CCP have been reported.22,23
A recent systemic review and meta-analysis of five studies found pooled hazard ratios for disease-specific mortality to be 2.42 and 2.08 for a 1-unit increase in CCP scores in univariate and multivariate models, respectively. The pooled hazard ratio for 1-unit increases in CCP scores for BCR were 1.88 and 1.63 in univariate and multivariate models, respectively.24
In order to provide clinical utility, the CCP score was combined with the Cancer of the Prostate Risk Assessment Score (CAPRA). The combination of CCP and CAPRA results in a combined clinical cell cycle risk (CCR) and can be used to predict likelihood of mortality from prostate cancer. The CCR was found to stratify patients’ 10-year prostate cancer mortality risk according to their CAPRA risk category.23
The Open Registry Measuring Impact of Genomic Testing on Treatment Decision After Biopsy in Newly Diagnosed Prostate Cancer Patients (PROCEDE-1000) study evaluated the ability of CCP to impact the physician-patient consensus. Results showed that incorporating the CCP into a shared decision-making process changed 48% of treatment plans, 72% of which were reductions in treatment intensity. The change in therapeutic modality was observed across all American Urological Association (AUA) risk groups.25 In a similar study, it was reported that the majority of treatment plan changes occurred in men with high CCP scores and an AUA low-risk category, as well as men with an AUA intermediate-risk category and low CCP score.26
The Prolaris Score has Medicare approval for men who have very low- to low-risk disease as defined by Epstein criteria and AUA risk strata found on prostate biopsy. In 2014, Prolaris was included into the National Comprehensive Cancer Network (NCCN) guidelines for the management of prostate cancer. The Prolaris Score can help differentiate patients within similar risk profiles (per D’Amico risk classification), as well as better assess the risk of prostate cancer-specific mortality in 10 years. Therefore, Prolaris may help both patients and physicians accurately determine the malignant potential of an individual’s cancer and perhaps allow men with indolent cancer to avoid potential morbidity associated with treatment.
Oncotype Dx® (Genomic Health, Redwood City, CA) and the genomic prostate score (GPS) can be used in men with very low-, low-, and intermediate-risk disease as defined by NCCN guidelines.27 Oncotype DX uses a quantitative 17-gene reverse transcription-polymerase chain reaction assay on manually microdissected tumor from as little as 1 mm of prostate tissue from biopsy specimens. The assay consists of 12 cancer-related genes and 5 reference genes. The 12 cancer-related genes represent 4 distinct biologic pathways (androgen signaling, stromal response, cellular organization, and proliferation) that contribute to the predictive value of the assay. The results are reported as a GPS, which is reported on a scale of 0 to 100. A lower GPS predicts a greater likelihood of favorable pathology, which is defined as freedom from high-grade disease (primary pattern 4 or any pattern 5 Gleason score and/or non–organ-confined disease). Adverse pathology was defined as high-grade, presence of primary Gleason pattern 4 or any pattern 5, and/or non–organ-confined (pT3) disease.28,29
The development of the GPS used 727 candidate genes from prostatectomy tissue that were evaluated in 441 patients with cT1/T2 disease and Gleason scores from 6 to 9 from a cohort of 2641 patients. The most promising 81 genes were then assessed in biopsy specimens and narrowed to 12 genes from 4 pathways. In the validation study, biopsy samples were compared with surgical pathology samples and the GPS was found to predict adverse pathology, with an odds ratio of 1.9 per 20-unit increase in GPS with men in low or intermediate risk groups.29 In an independent validation study and meta-analysis, the GPS was also found to substratify patients within CAPRA, NCCN, and AUA/European Association of Urology risk groups, and was found to be a significant predictor of adverse pathology and likelihood of favorable pathology.28,30 For example, it was found that for men with a CAPRA score of 2, the average probability of favorable pathology was 67%; incorporating GPS further refined this population, with probabilities ranging from 37% to 85% for favorable pathology.28,29
In a multicenter prospective trial, including 211 patients from 10 sites across the United States, assessing the impact of GPS on treatment selection, GPS resulted in a treatment recommendation change in 26% of all patients. An increase in the use of active surveillance (AS), along with increased physician confidence in treatment recommendations, was also reported. All changes were directionally consistent with GPS. Post-GPS treatment recommendations changed from surgery to AS in 30% of patients. Post-GPS, the greatest change (37%) was seen in the NCCN low-risk group.31 This suggests that the GPS, by providing an individual biologic assessment of tumor aggressiveness, can influence treatment recommendations, particularly in men with NCCN low-risk disease.
As a follow-up to the first clinical utility study, a retrospective chart review study in community-based urology practices was performed to assess the impact of Oncotype DX GPS on treatment recommendations and treatments implemented. In this study, there was a 22% relative increase in AS recommendations. Although the baseline (pre-GPS) rates of conservative management with AS/WW were high at 43%, the incorporation of GPS in the work-up of men with low-risk clinical features resulted in a 24% absolute and 56% relative increase in implementation of AS. Within each clinical risk group, the incorporation of GPS into decision making resulted in sizeable increases in the implementation of AS/WW. Largest absolute increases in AS were observed in NCCN very low- and low-risk patients. When GPS was integrated, AS was recommended more frequently and physicians saw a dramatic increase in patients accepting AS.32
Oncotype Dx GPS is a validated test, covered by Medicare, included in the NCCN guidelines, that can be used to help predict whether an individual’s cancer has favorable pathology. Furthermore, GPS provides substantial utility and increased confidence in treatment recommendations, which may lead to increased acceptance of the urologist’s treatment recommendation for AS, and patient confidence with treatment decisions.
ProMark® (Metamark Genetics, Cambridge, MA) is intended for use in men with Gleason 3+3 or 3+4 disease to help determine future management or treatment options. Promark is a protein-based proteomic prognostic test that incorporates eight biomarkers from a prostate biopsy sample to predict an individual’s risk of favorable or nonfavorable/aggressive prostate cancer. Favorable results are defined as surgical Gleason ≤ 3+4, ≤ T2, and nonfavorable/aggressive disease is defined as surgical Gleason score ≥ 4+3, or T3a, T3b, +N, or +M disease.33 The theory behind proteomic-based tests is that they may more directly reflect signaling pathways and oncogenic phenotypes, as they represent post\translational modifications.34 A Promark score is reported in a range of 0 to 100, with higher scores representing an increased risk for aggressive disease. The score report compares individual results with biopsy pathology and incorporates a patient’s NCCN risk category.
In the development of Promark, automated integrated multiplex immunofluorescence was used to identify eight biomarkers in prostatectomy tissue. These eight biomarkers were identified from 381 patient biopsies with matched prostatectomy specimens to predict prostate pathology aggressiveness and lethal outcome.33,35
In a clinical validation study, Promark® was able to delineate between favorable and nonfavorable disease from 276 blinded biopsy samples using a risk score range of 0 to 1. A score of ≤ 0.33 had a 90% sensitivity, or 10% false-negative rate for favorable pathology. In those with a Promark score of ≤ 0.33, the positive predictive values for favorable disease in NCCN risk groups were 95%, 81.5%, and 75% for patients in very low-, low-, and intermediate-risk groups, respectively. A risk score of 0.80 was found to have a 95% specificity for nonfavorable pathology, which translates into a 5% chance of having favorable pathology with a risk score above 0.80. The predictive value of nonfavorable pathology with a Promark score < 0.80 was 76.9%, and increased to 100% with a score . 0.90.33 Promark’s protein expression-based test is a novel instrument that can stratify an individual’s risk for prostate cancer and potentially aid in physician-patient discussion about treatment or AS.
Decipher® (GenomeDx Biosciences, Vancouver, British Columbia, Canada) genomic classifier (GC) score can be used in the postprostatectomy period for men with pT3 disease, positive margins, or BCR to determine risk of metastasis. The GC score uses prostatectomy tissue to assess coding and noncoding RNA via microarray; 22 markers related to cell proliferation, differentiation, androgen signaling, motility, and immune modulation are incorporated into the GC score, which ranges from 0 to 1. Higher scores translate into increased risk for metastasis.
A validation study reported that GC scores when used as an independent predictor of metastasis have a hazard ratio of 1.5 for every 10% increase in GC score in a multivariable model. In a 5-year cumulative incidence of metastasis, patients with a GC < 0.4 (low risk) had a 2.4% chance of metastasis compared with 6.0% for a GC score between 0.4 and 0.6 (intermediate risk). There was also a 22.5% chance of metastasis in patients with a GC score > 0.6 (high risk).36 In a separate patient population, the mean GC for patients with rapid metastasis was 0.58, and every 10% increase in GC increased the odds of rapid metastasis by 1.48.37 The GC was also able to predict 8-year cumulative incidence for biochemical failure and distant metastasis in men receiving adjuvant radiation therapy following prostatectomy.38
In 188 patients who received postoperative radiotherapy for pT3 or margin-positive disease, Decipher found the cumulative incidence of metastasis at 5 years to be 0%, 9%, and 29% for low, average, and high GC scores, respectively. For those with GC scores < 0.4, or low risk, there were no differences in 5-year metastasis in men who received adjuvant or salvage radiotherapy (P = .79). However, for men with GC scores ≥ 0.4 the cumulative incidence of 5-year metastasis was 6% and 23% for adjuvant and salvage radiotherapy, respectively (P < .01).39 In a retrospective, multicenter study including 170 men receiving salvage radiotherapy for extraprostatic extension, seminal vesicle invasion, and positive margins, the 5-year cumulative incidence of metastasis following salvage radiotherapy was 2.7%, 8.4%, and 33.1% for men with low-, intermediate-, and high-risk GC scores, respectively. The hazard ratio for metastasis was 1.58 per 0.1 change in GC score.40
Decipher tends to redirect patients to the appropriate treatment after prostatectomy. In patients who are postprostatectomy with high-risk adverse pathologic features, Decipher aids in determining if they can safely delay adjuvant therapy. With Decipher, approximately 60% of patients are reclassified as low risk. Decipher low risk patients with a postoperative CAPRA-S score of ≤ 5 had a negative predictive value of > 98% for 5-year metastatic-free survival despite having postprostatectomy-adverse pathologic features.39 Two decision-impact studies looked at postprostatectomy patients who had pT3 disease or positive surgical margins. When Decipher was not available, 49% of physicians chose adjuvant radiotherapy. When Decipher was available, 77% of low-risk patients moved into an observation arm, whereas 84% of high-risk patients were treated with adjuvant radiotherapy.41,42
Decipher’s genomic classifier is currently the only test available that helps determine the need for additional treatment following radical prostatectomy, and has been shown to change physician treatment recommendations in both low- and high-risk groups.41 Recent advancements include investigating if the GC can be used to predict metastasis from prostate biopsy43 and the Genomic Resource Information Database, an RNA expression database for research use.
Today’s practicing urologists have a multitude of blood, urine, and genomic-based tests at their disposal, generating discussion about their appropriate use. This article focuses on two blood, one urine, and five genomic-based tests that, when used in the appropriate clinical setting, may aid in the patient-physician decision-making process. These tests help patients and physicians decide who should have an initial prostate biopsy, repeat biopsy, or begin treatment for diagnosed prostate cancer, and who may benefit from adjuvant therapy following prostatectomy. To our knowledge, no prospective study has assessed the ability of any of the tests reviewed to decrease mortality. Furthermore, no head-to-head comparisons among tests with the same function have been completed. The lack of such studies may be secondary to the recent development of a majority of these tests.
The authors wish to thank Mihir Thaker, DO, and Michael Wilson, DO, for their editing assistance.
