Prostate cancer (PCa) represents a significant burden to men's health. It is the most common cause of cancer in men and the second leading cause of cancer death. In 2008, approximately 900 000 men were diagnosed with PCa worldwide, with the highest rates in Europe, North America, South America and Oceania. An estimated one in six men in the USA will be diagnosed with PCa in their lifetime; however, the risk of dying from PCa is only 1 in 35.1
PCa lends itself to early detection as it has a slow natural course if untreated, allowing for the cancer to be diagnosed before it progresses to an advanced and incurable stage. In addition, most tumours have a long latency period and remain contained within one organ, allowing for potentially life-saving therapeutic interventions.2 Moreover, screening or early detection can be performed with a simple test to determine the levels of prostate-specific antigen (PSA) in the blood.
The benefits and disadvantages of PSA-based screening have become the subject of intense investigations and debate. First and foremost, the high prevalence of PCa makes the disease a socioeconomic, as well as personal, problem. Autopsy series have, for example, shown PCa in a high proportion of men dying from other causes, rising from 27% in men aged 41–50 years to 65% in men aged 71–80 years.3 Second, treatment of PCa can be associated with serious and life-long side-effects such as incontinence and erectile dysfunction;4 therefore, screening should identify only men with clinically significant PCa in its early stages. In recent years, there has been debate concerning the utility and potential harms of PSA-based PCa screening. This review summarizes the latest evidence on this subject and gives an overview of future candidate biomarkers.
A non-systematic literature search using the MEDLINE/PubMed database was conducted to identify original articles, review articles and editorials regarding PSA and novel biomarkers in PCa diagnosis and screening. Searches were limited to the English language, humans and adults. All abstracts were reviewed and the corresponding full-length articles for those that were most relevant were analysed.
Prostate-specific antigen screening
Prostate-specific antigen is a glycoprotein secreted by prostate epithelial cells. It was first studied in the late 1970s and replaced prostatic acid phosphatase as a PCa biomarker in the 1980s.5 In conjunction with a digital rectal examination (DRE), PSA has become a widely used tool for PCa screening.6,7 However, an elevated blood level of PSA does not automatically indicate PCa as the PSA level can be elevated in men with benign prostatic hyperplasia or inflammation.
There is a large body of retrospective or case–control prospective studies on PSA-based PCa screening. However, owing to various methodological limitations, conclusions based on these studies are limited.8 Recently, three prospective randomized trials assessed the impact of PSA-based screening on PCa mortality. Each of these trials is discussed separately below and the results are shown in Table 1.
|Number of participants||77 000 men||182 000 men||20 000 men|
|Age range (years)||55–74||50–74||50–64|
|Ratio of screening to no screening||1:1||1:1||1:1|
|Median follow-up (years)||7||9||14|
|Main results||Incidence of death per 10 000 person-years: 2.0 in the screening group; 1.7 in the control group||
20% reduction in relative risk of death from PCa in screened group compared with control group
To prevent one death: number of men that need to be screened: 1410; number of men that need to be treated: 48
44% reduction in relative risk of death from PCa in screened group compared with control group
To prevent one death: number of men that need to be screened: 293; number of men that need to be treated: 12
|Conclusion||Rate of death from PCa very low and not significantly different between the screened group and control group||
PSA-based screening reduced the rate of death from PCa by 20%
PSA-based screening associated with high risk of overdiagnosis
PCa mortality reduced almost by half in the screened group
Benefit of PCa screening compares favourably to other cancer screening programmes
Prostate, Lung, Colorectal and Ovarian cancer screening trial
The Prostate, Lung, Colorectal and Ovarian (PLCO) trial accrued close to 77 000 men aged 55–74 years from 1993 to 2001 at 10 different centres in the USA.9 The trial reported information on PCa incidence, cancer-specific mortality (CSM), all-cause mortality and cancer staging. Participants were randomly assigned to 6 years of annual PSA tests and 4 years of DRE. The PSA cut-off was 4 ng/ml, and patients with a PSA above the cut-off, or suspicious DRE results, were advised to seek further diagnostic evaluation. Compliance rates were 85% for PSA testing and 86% for DRE. The incidence of PCa per 10 000 person-years after 7 years was 116 in the screened group compared with 95 in the control group. The incidence of death per 10 000 person-years was 2.0 in the screened group and 1.7 in the control group. The authors concluded that, after 7–10 years of follow-up, the rate of death from PCa was very low and did not differ significantly between the two groups.
The PLCO trial has been criticized for various reasons, the first of which was the high contamination rate in the control group. Reflecting community practice in the USA, approximately 44% of patients in the control group had at least one PSA test prior to study entry. Furthermore, in the first year of the trial, the rate of PSA testing in the control group was 40% and increased to 52% by the sixth year of study inclusion, suggesting that the control group contained men who were less likely to harbour PCa. In addition, some cancers that are detectable on initial screening may have been already removed from the study population. Second, the biopsy rate was only around 40% in men who had an elevated PSA.12 Third, a PSA level of 4 ng/ml was used as cut-off and a low cut-off level may lead to detection of low-stage cancers that are associated with higher CSM. Fourth, the follow-up time of 7 years was too short to detect a significant effect of PCa screening, given the natural course of untreated PCa.
An analysis of the PLCO data with a focus on comorbidities showed a significant decrease in the risk of CSM in men with no or minimal comorbidities who were randomly assigned to PSA screening versus usual care (22 vs. 38 deaths). It was reported that treatment needed to be given to an additional five men before one PCa death at 10 years could be prevented. These data suggest that selective use of PSA screening for men in good health appears to reduce the risk of PCa CSM with minimal overtreatment.13
European Randomized Study of Prostate Cancer
The European Randomized Study of Prostate Cancer (ERSPC) trial was a European-based multi-institutional trial that was initiated in 1991 for men between the ages of 50 and 74 years. In total, 72 952 men were randomly assigned to PSA screening at 2-year or 4-year intervals in centres in Belgium, Finland, France, Italy, the Netherlands, Spain, Sweden and Switzerland.10 The control group consisted of 89 435 men who did not undergo screening for PCa. Most centres used a PSA cut-off of 3.0 ng/ml as an indication for biopsy, except in Finland, where a PSA of 4.0 ng/ml was used.10 During a median follow-up of 9 years, the cumulative PCa incidence was 8.2% in the screened group compared with 4.8% in the control group. The rate ratio for CSM in the screened group was 0.80, indicating a 20% relative risk reduction in death from PCa in relation to the control group. However, in order to prevent one death from PCa, 1410 men would need to be screened and 48 men would need to be treated. The reduction in mortality due to PSA screening in the ERSPC study was further increased to 30% when adjusting for non-compliance in the screened population and contamination in the control group.14
An analysis of follow-up data regarding the development of metastatic disease from the four centres conducting the ERSPC study was published in 2012. After a median follow-up of 12 years, 256 men in the PSA-screened group and 410 men in the control group were diagnosed with metastatic PCa, resulting in a cumulative PCa incidence of 0.67% and 0.86% per 1000 men, respectively. This finding translates into a 42% reduction for men who were actually screened and an absolute risk reduction of metastatic disease of 3.1 per 1000 randomized men.15
Despite several methodological strengths, the results of the ERSPC trial should be interpreted with caution. The risk of overdiagnosis was estimated to be around 50% and the benefit of screening appears to be restricted to men in the core age group of 55–69 years.16 In addition, the results of the ERSPC trial showed that PCa screening is associated with a high risk of overtreatment. Moreover, a control subject with high-risk PCa was more likely than a subject in the PSA-screened group to receive radiotherapy, expectant management or hormonal treatment instead of radical prostatectomy. However, the trial group (screened or control) had only a minor role in treatment choice compared with other variables, and the differences in treatment between groups are unlikely to play a major role in the interpretation of the results of the ERSPC trial.17
The Göteborg screening trial was originally an independent trial and later joined the ERSPC.11 It accrued 20 000 men aged 50–64 years who were randomized in a 1:1 ratio to screening with PSA every 2 years versus no screening.11 The primary endpoint of the study was CSM. During the median follow-up of 14 years, 12.7% of men in the screening group and 8.2% of men in the control group were diagnosed with PCa. The rate ratio for CSM was 0.56, which translated into a 44% relative risk reduction of death from PCa in the PSA-screened group of the study. In order to prevent one death from PCa, 293 men would need to be screened and 12 men would need to be treated.
Interestingly, the Göteborg study reported better outcomes of screening than the PLCO and the ERSPC trial. This may be explained by differences in trial design and cohort composition such as younger patients, shorter intervals for screening, lower rate of PSA testing prior to study entry, lower rate of contamination in the control group and longer follow-up.16
Current debate on prostate-specific antigen screening
Based on the results of the three major screening trials, various health organizations published recommendations on PSA screening. The United States Preventive Services Task Force (USPSTF) published the most widely debated statement. In a 2012 update of a previous analysis, the USPSTF recommended against PSA-based PCa screening for all men in the general US population, regardless of age.18 The American Urological Association (AUA) clearly opposed this position and published a response to the recommendation of the USPSTF, which can be found online (http://www.auanet.org/content/media/USPSTF_AUA_Response.pdf). The AUA currently promotes a baseline PSA value in all men aged 40 years and screening is discouraged in men with a life expectancy less than 10 years.19 The ongoing debate reflects the current uncertainties and opposing positions regarding PCa screening and emphasizes the necessity for novel biomarkers that could help avoid the overdiagnosis and overtreatment associated with PSA-based PCa screening.
Novel biomarkers for prostate cancer diagnosis
Currently, PSA remains the predominant tool in PCa diagnosis and screening; however, PSA is a poor predictor of disease. Approximately 65–70% of men presenting with an elevated PSA between 4 and 10 ng/ml will have a negative biopsy result.20 In addition, prostate biopsies can be associated with significant morbidity21 and the limitations of PSA in screening and prognostication of PCa facilitated the search for new blood-based or urine-based biomarkers for PCa diagnosis and screening.22,23 A large part of the limitation of serum PSA is the low specificity because it is not a PCa-specific marker. Methods to enhance PSA specificity have assisted clinicians in deciding which patients should undergo biopsy, but have not necessarily improved diagnostic accuracy or facilitated optimal therapeutic decisionmaking (e.g. PSA density, complex PSA, free PSA, etc).23 More accurate tests that can stratify patients according to their risk of developing PCa and identify men who require repeat prostate biopsy and those at risk for aggressive disease are required. Since the introduction of PSA, advances in DNA sequence and RNA transcriptome profiling have broadened our understanding of cancer biology24 and novel assays are able to detect PCa-associated biomarkers in various biological fluids including serum and urine. The three most advanced urinary tests are discussed below.
Prostate cancer antigen 3
Prostate cancer antigen 3 (PCA3) is the most prominent, non-PSA-based biomarker for PCa. PCA3 is a non-coding RNA that has been shown to be elevated in > 90% of PCa tissues, but not in normal or benign prostatic hyperplastic tissue.25,26 Owing to the high sensitivity and specificity of PCA3 in PCa tissue, assays were further developed to detect PCA3 in urine, which led to the currently available transcription-mediated amplification-based urine assay.27 The assay uses PSA transcripts as internal controls for RNA quality and presence of prostate-specific nuclear material. The PCA3 score is generated by the PCA3 to PSA ratio.28 In several studies, PCA3 appeared to be superior to PSA in predicting PCa based on biopsy alone.28–32 A urine PCA3 score of > 35 had an average specificity of 76% and sensitivity of 66% for the prediction of PCa at biopsy, whereas serum PSA had a specificity of only 47% at about the same sensitivity (65%).29 Urine PCA3 measurements provide additional information to the PSA test with area under the curve (AUC) values of 0.66–0.72, compared with 0.54–0.63 for serum PSA alone.30 Combining PCA3 analysis with PSA testing further improves both measures, resulting in an AUC of 0.71–0.75.31 In a study of 495 men, PCA3 had a negative predictive value of 90% and was independently associated with PCa.32
The PCA3 score is not influenced by age, prostatitis, prostate volume or 5α-reductase inhibitors. In men with elevated serum PSA levels and one or two previous negative biopsy results, elevated PCA3 scores were associated with a higher risk of PCa. In 2012, the US Food and Drug Administration (FDA) approved PCA3 as a diagnostic test for PCa in the setting of a previous negative biopsy. PCA3 has become the first basis for the molecular diagnostics in clinical urological practice.
One limitation of the PCA3 test is the cut-off score used to determine the test as positive. As expected, while sensitivity increases, specificity of the assay decreases with lower cut-offs. Furthermore, a low PCA3 score does not exclude cancer and, therefore, not reaching a satisfactory negative predictive value in order to rule out PCa.
Studies on the predictive value of PCA3 for aggressive disease have been contradictory. No significant association between PCA3 score and any prognostic parameter (including stage, Gleason score, tumour volume or extraprostatic extension) was found in independent studies;33,34 however, other studies found an association between high PCA3 scores and a Gleason score of ≥ 7, extracapsular extension and high tumour volume.35,36 Recently, PCA3 score was found to be a valuable predictor of low-volume, insignificant cancer35,36 and, as such, PCA3 could be of use in selecting patients for active surveillance. However, in men with significant PCa, PCA3 may not differentiate the aggressiveness of a tumour; therefore, a biomarker indicative of tumour aggressiveness is still an unmet need for patients suffering from PCa.
Gene rearrangements are known events in cancers such as haematological malignancies and have also been described in PCa. TMPRSS2-ERG gene fusions are one of the most common genetic events in PCa, are present in 50% of all cases and are specific for PCa but can also be detected in prostate intraepithelial neoplasia.24 It was shown that men with extremes of TMPRSS2–ERG gene fusion and PCA3 have different risks of cancer and different aggressiveness of the cancer, as seen in the biopsy. In combination with other clinical and pathological parameters (combined using the multivariate prostate cancer prevention trial risk calculator), urine TMPRSS2–ERG and PCA3 may guide the urgency of biopsy after the detection of an elevated serum PSA.37 The combined measurement of PCA3 and TMPRSS2-ERG in urine outperformed serum PSA for PCa diagnosis and may represent a promising urine test for PCa.37 However, urine expression of PCA3 and TMPRSS2–ERG is related to urine PSA mRNA and thus the tests are not informative if the PSA mRNA level is too low. Furthermore, it has to be kept in mind that both PCA3 and TMPRSS2–ERG are currently helpful only as adjuncts to PSA. There is currently a need for studies determining the performance of the tests in the absence of PSA.
Prostate-specific membrane antigen
Prostate-specific membrane antigen (PSMA) is a transmembrane glycoprotein that is expressed on the surface of prostate epithelial cells. It has been reported that PSMA is upregulated in PCa tissue but not in benign prostatic tissues and no overlap in PSMA expression has been found between benign prostatic hyperplasia and PCa, indicating that PSMA is a very promising diagnostic marker.38 It has been shown that high PSMA expression in PCa cases is correlated with tumour grade, pathological stage, aneuploidy and biochemical recurrence.39,40 In addition, increased PSMA mRNA expression in primary PCas and metastasis has been reported to correlate with PSMA protein overexpression.40 The clinical utility of PSMA as a diagnostic or prognostic marker for PCa has been hindered by the lack of a sensitive immunoassay for this protein.
Reverse transcriptase polymerase chain reaction studies have shown that PSMA, in combination with its splice variant, could be used as a prognostic marker for PCa.41 In the normal prostate, PSMA splice variant expression is higher than that of PSMA, whereas in PCa tissues the PSMA expression is more dominant; therefore, the ratio of PSMA to its splice variant is highly indicative of disease progression. Designing a quantitative PCR analysis that discriminates between the two PSMA forms could yield another application for PSMA in diagnosis and prognosis of PCa.42
In men with serum PSA values between 4 and 10 ng/ml with no previous biopsies, PSMA mRNA measured in postprostate massage urinary sediments was superior to PCA3 in predicting cancer upon biopsy. At 70% specificity, the sensitivity for PSMA and PCA3 was 64% and 46%, respectively.43 Because of its specific expression on prostate epithelial cells and its upregulation in PCa, PSMA has become a target for therapies and the proposed strategies range from targeted toxins and radionuclides to immunotherapeutic agents.44 First-generation products have entered clinical testing.
Prostate-specific antigen testing has dramatically changed the diagnostic approach and clinical presentation of PCa. While PSA is associated with PCa, it has only moderate specificity. The benefits of PSA screening have been established in different trials and include earlier diagnosis of disease, reduction in initial diagnosis of metastasis and improvement in PCa-specific mortality. However, PSA screening is also associated with substantial overdiagnosis and overtreatment; therefore, the treatment decision should be tailored to the risk of metastasis and PCa-specific mortality. New biomarkers are currently adjunctive to PSA and can guide further diagnostic tests after initial negative biopsy. When used in the proper context, future biomarkers may prevent unnecessary biopsies and help identify patients who are likely to benefit from early diagnosis and treatment. Furthermore, they may be helpful in risk stratification that can lead to an individualized treatment decision.
To achieve this goal, there is an urgent need to discover more accurate non-invasive tests for PCa diagnosis and to allow the stratification of patients with life-threatening PCa. Because of the ease of collection, and the fact that prostate cells are directly released into the urethra through prostatic ducts after DRE or prostate massage, urine has become the future of non-invasive biomarker testing. Many studies26,27,29–34 have shown the feasibility of urine for the non-invasive detection of PCa and biomarker research is a focus at many laboratories, where several biomarkers are promising owing to their specificity for the disease in tissue. To our knowledge, only a few of these biomarkers have shown to be useful as urinary marker to date, and two PCa-specific RNA-based biomarkers have been identified (PCA3 and TMPRSS2–ERG gene fusions). The recent FDA approval of PCA3 has led to its introduction in clinical practice and the combination of both markers has been marketed for clinical use as well. Compared with single biomarkers, the combination of several biomarkers considerably improves the prediction of PCa in urine samples and also reflects the biology of the disease, which can be indolent or aggressive.
In the era of individualized therapy, the biomarker combinations are necessary not only to predict PCa at biopsy, but also to predict the aggressiveness of the cancer. Preliminary results show that the PCa-specific TMPRSS2–ERG gene fusion may be indicative of the aggressiveness of cancer upon biopsy, although further studies are warranted. As mentioned above, in PCa biomarker development, the greatest unmet need is a biomarker that stratifies men at risk of aggressive PCa, eventually leading to a reduction of unnecessary interventions.