A preponderance of clinical evidence supports a significant public health benefit for prostate-specific antigen (PSA)–based screening and early detection of prostate cancer in appropriately counseled and selected men. Population-based screening with PSA decreases prostate cancer mortality; however, because of relatively poor specificity, PSA-based screening may also increase the detection of clinically insignificant cancers that would otherwise never require treatment. Use of newer biomarkers that increase the specificity for prostate cancer detection may aid in risk stratification and the appropriate identification of men for prostate biopsy. The authors review the 4-kallikrein panel and 4K probability score.
Key points
- •
Overdetection of prostate cancer is in part attributable to the relatively low specificity and positive predictive value of prostate-specific antigen (PSA)-based screening.
- •
Biomarkers that enhance the performance characteristics of PSA and differentiate between indolent and lethal prostate cancers are needed.
- •
Tests with increased specificity for clinically significant cancer may decrease the number of men undergoing prostate biopsy and diminish the detection of indolent, clinically insignificant disease.
- •
As observed in multiple studies, the serum 4-kallikrein (4K) panel accurately predicts the risk of biopsy-detectable high-grade cancer (Gleason score ≥7) in men who have never undergone prostate biopsy or in those with a prior negative biopsy.
- •
The 4Kscore provides a probability score for the risk of biopsy-detectable cancer and may inform care in men who have never undergone prostate biopsy and in those with a prior negative biopsy.
Introduction
Since the early 1990s, prostate cancer mortality has decreased 45%, a trend largely attributed to early detection through widespread use of prostate-specific antigen (PSA)–based prostate cancer screening. Screen-detected prostate cancers constitute a spectrum of disease ranging from the indolent to the highly aggressive. The goal of screening is to maximize the early diagnosis of potentially aggressive but curable disease while minimizing both the detection of indolent disease and the number of invasive confirmatory tests. However, the relatively poor specificity of PSA has contributed to the overdetection of indolent disease. Efforts have, therefore, focused on identifying new robust tools, such as novel biomarkers, that increase specificity for prostate cancer detection. A newer biomarker, the serum 4-kallikrein (4K) panel, may inform initial screening in men who have never undergone biopsy as well as in those with a prior negative biopsy. The authors review the 4K panel and its potential ability to improve prostate cancer screening.
Introduction
Since the early 1990s, prostate cancer mortality has decreased 45%, a trend largely attributed to early detection through widespread use of prostate-specific antigen (PSA)–based prostate cancer screening. Screen-detected prostate cancers constitute a spectrum of disease ranging from the indolent to the highly aggressive. The goal of screening is to maximize the early diagnosis of potentially aggressive but curable disease while minimizing both the detection of indolent disease and the number of invasive confirmatory tests. However, the relatively poor specificity of PSA has contributed to the overdetection of indolent disease. Efforts have, therefore, focused on identifying new robust tools, such as novel biomarkers, that increase specificity for prostate cancer detection. A newer biomarker, the serum 4-kallikrein (4K) panel, may inform initial screening in men who have never undergone biopsy as well as in those with a prior negative biopsy. The authors review the 4K panel and its potential ability to improve prostate cancer screening.
Limitations in prostate-specific antigen–based screening for prostate cancer
The National Comprehensive Cancer Network supports the continued use of PSA testing for the early detection of prostate cancer based on randomized trials confirming its efficacy to diminish prostate cancer mortality. However, despite the survival benefits associated with PSA-based screening, the concern regarding the overdetection of insignificant prostate cancers remains. Aggressive PSA-based screening decreases mortality by maximizing the early detection of both indolent and aggressive cancers. Overdetection is the diagnosis of screen-detected indolent prostate cancer that, left untreated, would otherwise not diminish overall or prostate cancer–specific survival. The rate of overdetection within the European Randomized Study of Screening for Prostate Cancer (ERSPC) was estimated to be approximately 50%.
Overtreatment of screen-detected indolent cancers with surgery and radiation may unnecessarily expose patients to substantial risks that diminish health-related quality of life without affecting long-term oncologic outcomes. Overtreatment of prostate cancer in the United States is a momentous concern as 90% of US men diagnosed undergo treatment, and approximately 66% of those treated will be confirmed to have indolent disease.
The lack of specificity and low positive predictive value of PSA contribute to the overdetection and, thus, overtreatment of prostate cancer. PSA is an organ-specific rather than prostate cancer–specific biomarker. No single PSA cutoff threshold has particularly good test performance characteristics in prostate cancer screening, with a particularly low predictive accuracy in the gray zone of elevated PSA (3 ng/mL to 10 ng/mL). Although higher PSA is associated with a higher stage and Gleason sum among men with cancer, a prostate biopsy is positive for cancer in only around 25% of men with a PSA between 2 and 10 ng/mL and no prior diagnosis. Moreover, some men with a low PSA may have clinically significant cancer.
It is clear that blanket cessation of all screening for prostate cancer, as recommended by the US Preventative Services Task Force, fails to acknowledge the biological heterogeneity of prostate cancer. Without screening, the 20% to 30% of men diagnosed with aggressive prostate cancer at presentation would potentially lose an opportunity for cure. One potential solution is to focus on the detection of clinically significant prostate cancer by using tests with enhanced specificity.
Human kallikreins
Human tissue kallikreins are a family of 15 secreted serine proteases, the regulatory functions of which are linked with the development of malignancy, respiratory disease, neurodegeneration, schizophrenia, and inflammation. Kallikreins have received focused investigation over the past decade after the clinical applicability of human kallikrein 3 (hK3, or PSA) in prostate cancer screening was discovered.
Most of the kallikreins are coexpressed in the prostate in varying amounts, although PSA (hK3) expression is restricted to the prostate and is almost exclusively produced by prostate epithelial cells. PSA is produced under androgen regulation and acts to liquefy semen after ejaculation. PSA is present in several molecular forms within blood and prostatic fluid and is normally found at low concentrations in serum compared with that found in ejaculate. Most of the PSA in blood is complexed to protease inhibitors and is catalytically inactive. The noncomplexed form, free PSA (fPSA), has several molecular forms (nicked, intact, and pro-PSA). In proportion to total PSA (tPSA), fPSA is lower in men with prostate cancer than in men with benign prostate hyperplasia ; the percentage of fPSA has been used in an attempt to improve PSA specificity.
Pro-PSA is an inactive precursor of PSA, which is cleaved by the protease action of human kallikrein 2 (hK2) rendering the mature form of PSA. The hK2 has several structural and functional similarities to PSA, although the concentration of hK2 in serum is lower than PSA. The development of accurate assays for hK2 suggested that it may be useful in prostate cancer detection and was subsequently found to be strongly expressed in high-grade prostate cancer. Intact PSA (iPSA) and nicked PSA have also been found to be useful in discriminating men with benign from malignant prostatic disease. Similar to hK2, iPSA increases as prostate cancer becomes more aggressive.
Kallikrein panel
Combining other kallikreins with PSA in a screening panel enhances its performance characteristics. Multiple studies have observed that a panel of 4 serum kallikrein markers combined (the 4K panel) (tPSA, fPSA, single-chain iPSA, and hK2) improves the prediction of biopsy-detectable cancer compared with PSA alone. Assays for tPSA and fPSA are currently widely available; however, iPSA and hK2 require sophisticated assay calibration, which is not readily available.
Method for Evaluating the 4 Kallikrein Panel
Diagnostic and prognostic models that predict the accuracy of a test may not address the relevance of such a test on actual clinical outcomes. Most studies on the 4K panel have assessed the predictive accuracy of adding kallikreins to a comparison base model of PSA alone and reported the results as the area under the receiver operating characteristic curve ( Table 1 ). High-grade cancer was defined as biopsy Gleason sum 7 or higher.
| Study, Year | Study Population | Study Sample | Outcome Assessed | AUC Increase HG Cancers a | Set Threshold Risk | Clinical Outcome Reduction (%) | HG Cancers Missed |
|---|---|---|---|---|---|---|---|
| Vickers, 2008 | Göteborg ERSPC | 740 Unscreened men | 6 Core bx | 0.04 | 20% Any cancer | 60 bx | 2.5% (1 of 40) |
| Vickers, 2010 | Rotterdam ERSPC | 2914 Unscreened men | 6 Core bx | 0.03 | 20% Any cancer | 51 bx | 12% (12 of 100) |
| Benchikh, 2010 | Tarn, France ERSPC | 262 Unscreened men | 10–12 b Core bx | 0.10 | 20% Any cancer | 49 bx | 6.8% (12 of 175) |
| Vickers, 2010 | Göteborg ERSPC | 1241 Screened men | 6 Core bx | 0.11 | 20% Any cancer | 41 bx | 2.3% (1 of 43) |
| Vickers, 2010 | Rotterdam ERSPC | 1501 Screened men | 6 Core bx | 0.09 | 20% Any cancer | 44 bx | 4.4% (4 of 91) |
| Gupta, 2010 | Rotterdam ERSPC | 925 Men with prior neg bx | 6 Core bx | 0.11 | 15% Any cancer | 71 bx | 16.7% (3 of 18) |
| Carlsson, 2013 | Rotterdam ERSPC | 392 Men after RP | RP pathology | 0.03 | 30% Aggressive c cancer at RP | 33 RP insignificant disease | 3.9% (26 of 666) c |
| Parekh, 2014 | US prospective trial | 1012 Men referred for bx | 10 Core bx | 0.08 d | 9% HG cancer | 43 bx | 10.4% (24 of 231) |
a Area-under-the-curve increase of full kallikrein model compared with the base clinical model.
b Decision to biopsy based on clinical judgment following digital rectal examination or repeat PSA.
c Aggressive radical prostatectomy pathology defined as pT3-T4, extracapsular extension, tumor volume greater than 0.5 cm 3 , or any Gleason grade of 4 or greater.
d An 4Kscore area-under-the-curve increase compared with modified Prostate Cancer Prevention Trial Risk Calculator 2.0.
To evaluate whether the panel would improve clinical decision making, all relevant studies use a decision curve analysis . This analytical method was first described in 2006 and estimates the net benefit of using a prediction model by summing the benefits (true positives) and subtracting the harms (false positives), whereby the latter is weighted by a factor related to the relative harm of a missed disease or event compared with an unnecessary intervention.
When used in 4K analyses, the weighting is derived from the threshold probability of prostate cancer at which a patient would choose to be biopsied, and the net benefit is calculated across a range of probabilities. The model incorporating the 4K panel was deemed to be of clinical value if it had the highest net benefit across the full range of threshold probabilities at which a patient would choose to be biopsied.
Four kallikrein studies in men before biopsy
Unscreened Men
The 4K panel was initially evaluated in 2008 among 740 previously unscreened men who underwent a 6 core biopsy for a PSA of 3 ng/mL or greater in the Göteborg cohort of the ERSPC. The panel was found to increase the predictive accuracy of diagnosing Gleason 7 or greater prostate cancer when added to a base model, including age, digital rectal examination (DRE), and tPSA. The area under the curve (AUC) for detecting high-grade cancer for the base model increased from 0.87 to 0.90 when the full kallikrein panel was added. These investigators concluded that using the full kallikrein panel would reduce biopsy rates by 60% for men with elevated PSA while missing only a small number of cancers (31 of 152 low-grade and 1 of 40 high-grade cancers).
The 4 kallikrein proteins were subsequently evaluated in several additional European cohorts of unscreened men. The full panel, including age, DRE, tPSA, fPSA, intact PSA, and hK2, was assessed among 2914 unscreened men from the Rotterdam arm of the ERSPC. Similar improvement in predictive accuracy for high-grade cancers was observed (AUC improved from 0.81 to 0.84); a reduction of biopsy rates was reported as 513 per 1000 men with elevated PSA and only missing a small number of high-grade cancers (12 per 100 men).
In the Göteborg and Rotterdam ERSPC cohorts, all men with an elevated PSA were referred for biopsy per ERSPC protocol. In usual clinical practice, men with elevated PSA usually undergo clinical work-up before making the decision to proceed with biopsy. This work-up commonly includes an assessment of prior prostatitis, prostate hypertrophy, and family history of prostate cancer. This type of work-up can affect biomarker properties and potentially the properties of a predictive model for prostate cancer. In an independent validation study, Benchikh and colleagues determined whether a statistical model based on patients in the Rotterdam arm of ERPSC, who underwent biopsy secondary to elevated PSA alone, would lose its predictive utility in a cohort of men who underwent a biopsy based on elevated PSA as well as clinical judgment, such as men biopsied in the Tarn, France arm of ERSPC. In the France protocol, the decision to biopsy based on clinical judgment following DRE or repeat PSA test. The kallikrein panel had a significantly higher predictive accuracy for any grade of cancer than the base model, with an AUC increase from 0.63 to 0.78, which was similar for high-grade cancer (0.77–0.87). Using a threshold of 20% or greater risk as indication for biopsy, 492 out of 1000 men would avoid biopsy and 61 men with cancer would be missed, of whom 12 would have high-grade disease. Individual contribution of each kallikrein to the panel was also assessed; fPSA was found to have the greatest contribution, although removing intact PSA and hK2 also led to a reduction in AUC, supporting the use of all 4 kallikreins in the panel. The investigators concluded the 4K panel can predict the result of prostate biopsy in men with elevated PSA and retains its value in men who underwent additional clinical evaluation before biopsy.
Screened Men
The predictive accuracy of PSA is lower in men who have been previously screened. Initial reports of the kallikrein panel were based on the first round of the ERSPC, in which a small proportion had been previously screened with PSA (∼3%). To assess whether the 4K panel retains its value in men with a recent PSA testing, Vickers and colleagues applied the panel to 1241 men with an initial PSA less than 3 who, in subsequent rounds of the Göteborg arm of ERSPC, were found to have PSA greater than 3 and, therefore, underwent biopsy. The previously published model was poorly calibrated and missed many cancers. The investigators observed findings similar to those in unscreened men; a significant number of biopsies could be avoided (413 per 1000 men) and delay the diagnosis of only 1 high-grade cancer per 1000 men with elevated PSA.
The model was subsequently applied to 1501 previously screened men in the Rotterdam section of ERSPC in order to determine if findings could be replicated. The AUC of high-grade cancer in the base model increased from 0.71 to 0.80 when the kallikrein panel was added. When the biopsy threshold was set at 20% risk of any cancer, the decrease in biopsy rate was deemed to be 362 for every 1000 men with elevated PSA and delayed 47 cancers, which would predominately be low grade. The investigators noted that a new model for recently screened men was necessary with the Göteborg cohort but not with the Rotterdam cohort. Newer assays were used to measure iPSA and hK2 for Rotterdam measurements.
Related posts:
Biomarkers in Urologic Cancer
Current Status of Urinary Biomarkers for Detection and Surveillance of Bladder Cancer
Emerging Bladder Cancer Biomarkers and Targets of Therapy
Pharmacogenomics
MicroRNAs in Testicular Cancer Diagnosis and Prognosis
The Emerging Role and Promise of Biomarkers in Penile Cancer
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
