Prostate-specific antigen-based screening for prostate cancer in the third millennium: Useful or hype?

Abstract

Prostate cancer is the most prevalent malignancy in men and the third leading cause of cancer deaths worldwide. Although the wide-spread introduction of total prostate-specific antigen (tPSA) testing has revolutionized the approach to the managed care of this disease, there are some biological, analytical, clinical, and economical issues that argue against the cost-effectiveness of tPSA-based population screening for early identification of cancer. The on-going standardization/harmonization efforts, along with the outcomes of recent epidemiological investigations, demonstrate that the current tPSA thresholds might be revised and possibly recalculated according to several demographical variables, such as age, ethnicity, genotype, family history, and body mass index. A major shortcoming of tPSA screening is the lack of reliable evidences of reduction in prostate cancer-associated mortality, due to the large lead-time because of the indolent growth rate, the impossibility to differentiate high-grade from indolent cancers, and the treatment-associated morbidity. Since no single tPSA cut-off was proven able to efficiently identify men at higher risk of death, the jeopardy of over-diagnosis and over-treatment is also tangible. The large expenditure is an additional source of concern. Finally, a wide-spread population screening also carries several ethical, social, and psychological implications, which might overwhelm the potential benefits.

• Prostate
• prostate cancer
• prostate-specific antigen
• screening

Epidemiology of prostate cancer

Prostate cancer is the most prevalent malignancy in men worldwide and the third leading cause of cancer deaths [1], [2]. Nearly 190,000 new cases are diagnosed every year both in Europe [3], [4] and in the US [5], and about 27,360 deaths a year are estimated in the US from this cancer [5]. Prostate cancer is particularly prevalent in developed countries such as the United States and the Scandinavian countries, with about a 6-fold difference between high-incidence and low-incidence countries [4]. Remarkable racial or ethnic differences have been described in incidence and mortality rates; black males are 1.6 times more likely to be diagnosed and 2.4 times more likely to die with prostate cancer than whites [6–8]. Socio-economic factors such as accessibility to prostate screening and appropriate medical treatment, along with racial differences in cancer biology, are likely contributors to this disparity [9]. A remarkable difference in the incidence of prostate cancer has also been reported between men in developed countries as compared with Asian men, such a difference being attributable to life-styles [10]. The risk of prostate cancer appears strongly related to age, since the prevalence rises sharply after age 55 years and peaks at age 70–74, declining slightly thereafter [11]. By the age of 80, approximately 60%–70% of men have evidence of histologic carcinoma at autopsy [11–13].

Key messages

• The wide-spread introduction of total prostate-specific antigen (tPSA) testing has revolutionized the approach to the managed care of prostate cancer, but there are some biological, analytical, clinical, and economical issues that argue against the cost-effectiveness of tPSA-based population screening.

• Several lines of evidence demonstrate that the current tPSA thresholds might be revised and possibly recalculated according to several demographical variables.

• A major shortcoming of PSA screening is the lack of reliable evidences of reduction in prostate cancer-associated mortality.

• Since no single tPSA cut-off was proven able to efficiently identify men at higher risk of death, the jeopardy of over-diagnosis and over-treatment is also tangible.

A well known risk factor is a positive family history of the disease [14], as reflected by studies in twins showing that 42%–57% of the risk of prostate cancer is the result of inheritable factors [15], [16]. Linkage and genome-wide scan studies have shown that prostate cancer susceptibility genes are mostly located on chromosomes 1, 20, 17, X [17–22], and 15q [23]. A major role of steroid hormones in the aetiology of prostate cancer has also been demonstrated, since these hormones, especially testosterone, are essential in the growth and maintenance of the prostate gland [24]. Putnam et al. observed that alcohol consumption increases the risk of prostate cancer and that there is also a positive association between body mass and significant prostate cancer [25].

Biochemical aspects of prostate-specific antigen

The original discovery of prostate specific antigen (PSA) resulted from the combined efforts of numerous scientists searching for antigens in prostate and semen [26]. PSA, also known also as kallikrein-related peptidase 3 (KLK3), was isolated from the prostate tissue in 1979 by Wang et al. [27]. In 1980, Papsidero et al. developed an immunoassay and determined PSA concentrations in serum [28], but the clinical use as a tumour marker was only suggested in 1987 [29].

Abbreviations

Table

ACT	alpha-1-antichymotrypsin
AMACR	alpha-methylacyl-CoA racemase
BPH	benign prostate hypertrophy
BPSA	benign PSA
CRP	C-reactive protein
EPCA	early prostate cancer antigen
FPSA	free PSA
ITS	intention-to-screen
KLK2	kallikrein-related peptidase 2
KLK3	kallikrein-related peptidase 3
LUTS	lower urinary tract symptoms
M	molecular mass
Mr	relative molecular mass
PSA	prostate-specific antigen
PSM	prostate-specific membrane antigen
PSAD	PSA density
PPSA	PSA precursor
SNPs	single nucleotide polymorphisms
TPSA	total PSA
WHO	World Health Organization

PSA is a serine protease belonging to the human kallikrein family, composed of 237 amino acids, with a molecular mass (M) of 26,079 for the peptide moiety of the molecule [30]. There are, however, at least five isoforms of PSA in plasma whose structure differ in the di-, tri-, and tetra-antennary carbohydrate composition, but the predominant molecular species detected by ion spray mass spectrometry was at relative molecular mass (Mr) of 28,430, thus indicating that PSA contains a carbohydrate residue of Mr 2,351, for a total percentage of carbohydrate of 8.3% [30]. Like other serine proteases, PSA is produced under normal conditions from an inactive precursor (pPSA) by the secretory cells in the prostate glands and secreted into the lumen, where a seven-amino-acid propeptide is cleaved by kallikrein 2 [31], [32]. While seminal PSA concentration is usually ~0.35 mg/mL, its concentration in the blood of healthy individuals is very low, usually around 0.6 ng/mL [33]. PSA that enters the circulation intact is rapidly bound by protease inhibitors (‘active PSA’), whereas a fraction which represents only 5%–35% of the total PSA (tPSA) is inactivated in the lumen by internal cleavages at residues 85 to 86, 145 to 146, or 182 to 183, does not form complexes with protease inhibitors or other proteins, and circulates as ‘inactive’ or free PSA (fPSA) [34], [35]. Approximately 70%–90% of tPSA is bound to alpha-1-antichymotrypsin (ACT), whereas a small amount is complexed with alpha-2-macroglobulin, alpha-1-antitrypsin, and protein C (complexed PSA) [34], [35]. The ‘inactive’, free form of PSA is rather heterogeneous in structure, comprising at least three additional distinct variants which include the propeptide (pPSA), an internally cleaved or degraded form also referred to as ‘benign PSA’ (BPSA), and an intact, denatured form that is similar to the native protein, except for changes in structure or conformation that make the molecule enzymatically inactive [36]. Interestingly, several molecular subspecies are also included under the term pPSA, each of these differing in length according to the number of the seven amino acids left after cleavage of the proleader peptide (the most represented variants, however, contain proleader peptides of four and two amino acids, [-4]pPSA and [-2]pPSA, respectively). These truncated forms appear to be more resistant to activation than the intact pPSA containing the full seven-amino-acid proleader peptide [36].

PSA is continuously released into the blood-stream from the normal prostate by diffusion [37]. Under particular circumstances, coinciding with disruption of the basement membrane and normal lumen architecture, the tumour loses contact with the prostatic ducts and greater amounts of both ‘active’ protein and pPSA escape proteolytic degradation, and are released into the blood-stream [37], so that these subfractions are those most frequently encountered in the blood of prostate cancer patients. The molecular basis for the elevation of these truncated isoforms in prostate cancer is uncertain, but it likely reflects decreased cleavage of proPSA by kallikrein 2 in cancer tissue [37]. This pathway of pathological release into the circulation supports the evidence that the amount of ‘inactive’ fPSA is usually lower in the serum of men with prostate cancer as compared with those who have a normal prostate or benign prostate hypertrophy (BPH) [34], thereby justifying the use of the fPSA-to-tPSA ratio in establishing the cause of an increased tPSA [38]. Contrary to the common belief that prostate cancer is accompanied by enhanced production of PSA, protein expression by neoplastic cells is usually reduced as compared with healthy or benign hyperplastic tissues, so that the increased serum concentration observed in patients with cancer is mainly attributable to increased cell numbers, disruption of prostatic architecture, and diffusion of higher amounts of protein into the blood-stream [39].

Decisional thresholds

As usual in laboratory diagnostics, the selection of optimal decisional thresholds is a crucial issue in PSA-based screening for prostate cancer. Most, if not all, laboratory parameters are continuous variables, so that the choice of the optimal cut-offs for identifying the disease is the result of a delicate balance between sensitivity and specificity, negative and positive predictive value. The case of PSA is paradigmatic, because there are no reliable markers or measures of tumour invasivity in prostate cancer patients so far (the Gleason score might be helpful, but it is not the panacea), and thereby a definitive bench-mark for comparing diagnostic performances is lacking. Understandably, a low cut-off would increase both sensitivity and positive predictive value of the assay, but it will contextually increase the chance of over-diagnosis and over-treatment of indolent cancers, which are basically organ-confined, with no primary or secondary Gleason pattern 4 or 5, and a volume less than 0.5 cm³[40]. Conversely, enhancement of the decisional threshold would decrease the chance of identifying cancers with favourable characteristics, so that missing these cancers may not produce adverse clinical outcomes, but will shirk the inherent purpose underlying a population screening, that is identification of disease in the early stage.

In the late 1980s, without preliminary validation studies to ascertain test performance, a cut-off level of 4.0 ng/mL was widely adopted in clinical practice, so that for nearly two decades prostate cancer was thought to be virtually non-existent at tPSA levels below such a threshold. Moreover, since the vast majority of patients with levels lower than 4.0 ng/mL did not undergo biopsy and were thereby excluded when evaluating the diagnostic performances of the test, the results of several earlier reports on the test's sensitivity and specificity were upwardly biased [41].

When evaluating the diagnostic performances of a screening strategy, the two sides of the coin should be considered, i.e. sensitivity and specificity. In a seminal article in 2004, Thompson et al. highlighted that the use of a 4 ng/mL cut-off was unsuitable to detect all prostate cancers, as reflected by the rather different prevalence of malignancy according to the tPSA value, being 17%, 24%, and 27% among men with tPSA of 1.1–2.0 ng/mL, 2.1–3.0 ng/mL, and 3.1–4.0 ng/mL, respectively [42]. Loeb et al. also showed that base-line tPSA levels in ranges of <0.7, 0.7–2.5, 2.6–4.0, 4.0–10.0, and >10.0 ng/mL were associated with prostate cancer rates of 0.4%, 4.1%, 18.9%, 29.1%, and 50.0%, respectively [43]. In the late 1990s, two separate studies suggested that the 4 ng/mL cut-off might also have a suboptimal specificity, since the risk of prostate cancer at biopsy was similar at tPSA levels of either 2.5–4.0 ng/mL or 4–10 ng/mL [44], [45]. Moreover, one-third of prostate cancers detected at a PSA level of 4.0 ng/mL or greater had already spread to the prostate capsule or beyond [40].

These results confirm that even within the 0–4.0 ng/mL range, the tPSA level is a continuously increasing marker of prostate cancer risk, with no boundary below which no prostate cancer can be detected or ruled out, so that Thompson and Ankerst provocatively suggested abandoning the terms ‘normal’ and ‘elevated’ as descriptors of PSA results [41]. Thus, a great confusion has occurred in the choice of the optimal cut-off, so that a variety of thresholds are now used beside that conventionally established at 4 ng/mL, including 2.5 ng/mL (sensitivity of 80% for detection of Gleason 8 or higher disease) [42], [45], 3 ng/mL [46], or even 10 ng/mL [47].

Along with results of PSA and digital rectal examination (DRE), there are, however, other well known predictive variables of prostate cancers, including age, family history of prostate cancer, ethnicity, hereditary and environmental factors and attributes (e.g. diet, body mass index, supplement use), and a prior biopsy with negative results for cancer [41]. Using logistic regression analysis, Thompson et al. found four such variables to be independently and significantly associated with the presence of prostate cancer: PSA level, family history of prostate cancer, an abnormal result from a digital rectal examination, and (inversely) a previous negative biopsy result. They also observed that predictors of high-grade (Gleason score 7) cancer were PSA level, abnormal findings from a digital rectal examination, older age, and African-American ancestry, whereas a previous negative biopsy result was associated with a lower risk of high-grade cancer [48]. In line with these findings, a risk calculator was developed (available at:www.compass.fhcrc.org/edrnnci/bin/calculator/main.asp) that incorporates these variables to provide a more accurate estimate of prostate cancer risk as well as high-grade disease [49].

Guide-lines and recommendations for screening

The American Urological Association and the American Cancer Society recommend annual tPSA testing and digital rectal examination beginning at the age of 50 years to men with a normal risk of prostate cancer and beginning at an earlier age to men at high risk [50], [51]. The National Comprehensive Cancer Network recommends a risk-based screening algorithm, including family history, race, and age [52]. In contrast, the US Preventive Services Task Force recently concluded that the current evidence is insufficient to assess the balance of benefits and harms of prostate cancer screening in men younger than age 75 years, and it recommends against screening for prostate cancer in men aged 75 years or older [1]. In no case, however, do the major clinical practice guide-lines advocate PSA screening to be routinely performed in asymptomatic men older than 75 years or younger than 40 years or with less than a 10-year life expectancy, even if the patient is in a group at high risk for prostate cancer. Regardless of these clear recommendations, the daily practice does not reflect a virtuous behaviour, and the prevalence of inappropriate requests (e.g. tPSA screening in patients older than 75 years or younger than 40 years) is as high as 20% [53], [54], being nearly three times higher from general practitioners than from hospital physicians [54].

Non-cancer-related factors influencing PSA concentration

The intra- and inter-individual biological variabilities of tPSA have been reported to be as high as 18% and 72%, respectively [55], which are much higher than the analytical variability of the automated techniques commonly used for their measurement. The concentration of PSA is influenced by a constellation of non-cancer-related factors, including genotype, age, ethnicity, prostate volume, and the presence of inflammation/prostatitis.

As for other forms of cancers, several sequence variants in the genome have been reported to be associated with prostate cancer risk. In a recent genome-wide association study with plasma tPSA levels, it was also shown that six single nucleotide polymorphisms (SNPs) (located at 7p15, 10q11, 10q26, 17q12, 19q13, and Xp11) had a cumulative effect on PSA levels, the mean PSA levels in men being almost 2-fold increased across increasing quintile of number of PSA associated with these alleles, so that genetic-specific tPSA cut-off values may be used to improve the discriminatory performance of this test for prostate cancer [56]. The level of tPSA steadily increases with age (e.g. median PSA level is 0.7 ng/mL for men aged 40–49 years and 0.9 ng/mL for men aged 50–59) [43], [57–59], so that the reference values are strongly age-dependent, and the upper reference limit should be substantially increased in elderly subjects. There are also racial differences in the concentration of base-line serum tPSA [60]. In particular, for men 70 years and older, tPSA at the 75th percentile were 3.45, 3.40, and 5.40 ng/mL for Mexican American, non-Hispanic white, and non-Hispanic black persons, respectively [61]. Accordingly, approximately 3.6% of Mexican American men, 6.2% of non-Hispanic white men, and 7.8% of non-Hispanic black men had total PSA of 4.0 ng/mL or more [61]. Some clinical investigations also reported slightly lower levels of PSA among obese men [62], [63]. However, tPSA levels are inversely associated with both lean and fat mass (a 2.3 kg difference in lean mass and fat mass is associated, respectively, with a -0.9% and -0.7% difference in tPSA test results) [64]. The results of a recent retrospective epidemiological analysis also show that C-reactive protein (CRP) levels are positively associated with tPSA in a general population of out-patients undergoing tPSA screening, supporting the existence of an intriguing mutual relation between tPSA values and systemic inflammation [65]. Among non-cancer related factors, most benign prostate pathologies can elevate tPSA in blood, including traumas, prostatitis, lower urinary tract symptoms (LUTS), and, especially, benign prostate hypertrophy (BPH). There are several iatrogenic sources that can also produce elevated tPSA values, including biopsy, prostatic massage, transrectal ultrasonography, digital rectal examination, and glandular manipulation [40].

Analytical issues

As previously mentioned, the choice of the assay may influence PSA results and, thereby, the likelihood of prostate biopsy and prostate cancer detection. Traditionally, several analytical variables, including the differences in epitope recognition by the assay antibodies and the calibration against different reference preparations, are responsible for the largest part of the variability among different immunoassays. To overcome the latter issue, an international reference preparation has been devised, where 90% of the PSA is complexed to ACT and the remaining 10% is fPSA [66], [67]. Therefore, several commercial PSA assays are now standardized to this World Health Organization (WHO) standard (WHO 96/670 reference preparation), whereas others remained aligned with the original Hybritech assay. The choice of the calibration is, however, not interchangeable, so that a potential bias might result when using assays calibrated against the two different standards. A recent prospective screening population study showed that median prostate-specific antigen was 17% lower using WHO- versus Hybritech-based assay standardization [68]. In further investigations, the WHO calibration also yielded results approximately 20% to 25% lower when compared with the conventional Hybritech calibration [69], [70], and other studies highlighted that the interchangeability of tPSA, fPSA, and%free PSA values obtained by commercial PSA assays remains elusive [71], [72]. More importantly, it was recently shown that application of the WHO standard for tPSA assays with commonly used thresholds might lead to a significant decrease (~20%) in prostate cancer detection [70].

It is thereby essential at this point in time to acknowledge that metrological traceability of PSA assay calibration to WHO standards is necessary to improve the comparability between tPSA assays, which would enable longitudinal comparison of patient data and major harmonization/comparability of results of multicentre studies, where different assays might be used. However, since the wide-spread introduction of the new WHO standard would produce significantly lower values of PSA, consideration should be given to revising the relationship between the WHO standard and the thresholds used for prostate biopsy, basically lowering the actual decisional tPSA thresholds. Moreover, although the introduction of the WHO standard would indeed contribute to a major uniformity of results among different tPSA assays, it is unlikely that the bias will be completely overcome, so that further efforts should be placed towards harmonization rather than standardization.

Outcome analyses

It is now unquestionable that the wide-spread use of tPSA testing has resulted in a dramatic reduction in the proportion of men who present with metastatic disease and in the age-adjusted incidence and mortality rates of prostate cancer [40], so that the association between PSA testing and clinical outcomes is very unlikely to have occurred as a result of chance. There are, however, further considerations that should be taken into account. First and foremost, it should be clearly acknowledged that PSA should only be considered an early ‘biochemical signal’ of prostate cancer, whereas the diagnostic approach and the clinical decision-making should both consider several other variables, such as age, family history, and results of digital rectal examination [48]. This basic assumption is essential for identifying the ideal setting of PSA testing and establishing its relationship with the natural history of the disease. The time by which tPSA screening advances prostate cancer diagnosis, called the lead-time, has been reported by several studies, but results have varied widely due to the heterogeneous definition and the study population, with mean lead-times ranging from 3 to 12 years. Accordingly, the over-diagnosis was also heterogeneous, ranging from 23% to 66% of all screening-detected cancers [73]. Like other population screenings, the efficiency in early detection of prostate cancer by tPSA testing must be weighed against over-diagnosis, over-treatment, quality of life, and last but not least cost-effectiveness. In the 11-year follow-up of the 1988 Quebec prospective randomized controlled trial, the Cox proportional hazards model of the age at death from prostate cancer showed a 62% reduction (P<0.002) of cause-specific mortality in the screened men [74]. However, a further intention-to-screen (ITS) analysis of the data determined the ‘real’ relative risk of death from prostate cancer to be 1.01 (95% confidence interval (CI) 0.76–1.33) [75]. A 15-year follow-up of a randomized controlled trial in Sweden also showed no significant difference in total or prostate cancer-specific survival between a group of men screened every third year and those who were not [76]. Also in this setting, an ITS analysis at 15 years’ follow-up yielded a non-significant relative risk of death from prostate cancer of 1.04 (95% CI 0.64–1.68) [75]. A pooled analysis of these two ITS-analysed studies produced a cumulative relative risk of death from prostate cancer of 1.01 (95% CI 0.80–1.29) [76]. The recent outcome of two very large randomized prostate cancer screening trials carried out both in the US and Europe was very similar. The Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial on prostate cancer mortality, involving more than 76,693 men, concluded that after 7–10 years of follow-up the rate of death from prostate cancer was very low and did not differ significantly between the two study groups (rate ratio 1.13; 95% CI 0.75–1.70) [77]. Although in the European Randomized Study of Screening for Prostate Cancer different tPSA cut-offs were used, the tPSA-based screening slightly but significantly reduced the rate of death from prostate cancer (rate ratio 0.80; 95% CI 0.65–0.98; adjusted P=0.04) but was associated with a risk of over-diagnosis as high as 50% in the screening group [47], [78]. This is consistent with the hypothesis that the group of missed cancers in these large epidemiological investigations contained fewer tumours with a high stage or a high Gleason score and more tumours with favourable characteristics, so that missing these cancers may not have clinical consequences but would be favourable for limiting over-detection and over-treatment of prostate cancer. It should also be mentioned here that in several epidemiological investigations a large number of prostate cancer cases was diagnosed because of elevated PSA, whereas controls were men with low PSA levels. Due to the postulated genetic regulation of PSA levels, it is therefore conceivable that cases and controls were not accurately matched for the most informative SNPs which are known to influence PSA levels in blood. Since no study assessed the impact of screening upon quality of life, costs, or the potential harms of screening, a recent meta-analysis of available studies concluded that there is insufficient evidence to either support or refute the use of screening for prostate cancer, in agreement with the conclusions of the Cochrane Review [75].

Perspectives

In addition to tPSA, the use of per cent fPSA (free/total PSA), complexed PSA, PSA velocity, and PSA density with ultrasound of the prostate have been proposed as additional tests to enhance the specificity of tPSA [40]. PSA velocity, which basically reflects the rate of PSA increase over time (e.g. annual rate of increase of 0.5–0.75 ng/mL per year), has been associated with prostate cancer risk and has also been considered a marker of cancer aggressiveness and mortality [1]. The assessment of PSA density (PSAD), which is the relationship of the PSA level to the prostate mass (volume) measured by transrectal ultrasonography, was earlier proposed as a helpful tool for characterizing the significance of elevated PSA levels in men with very large prostates, but it is now generally discouraged [1]. Per cent fPSA is a simple estimation of non-protein-bound PSA as a percentage of the tPSA level. Although this parameter might potentially help differentiate between cancer and benign prostate disease (the percentage of free PSA generally is lower in men with prostate cancer) in men with border-line tPSA levels (4–10 ng/mL), there is no universal consensus regarding the optimal cut-off due to the trade-offs in sensitivity and specificity [1].

Although tPSA and eventually%fPSA remain the best available tools for discriminating between localized prostate cancer and benign prostatic hyperplasia, the quantitative assessment of PSA mRNA, but not that of prostate-specific membrane antigen (PSM) mRNA, in blood might be helpful in the biochemical grading of prostate cancer [38], [79], [80].

The measurement of urinary biomarkers is a relatively novel diagnostic approach to diagnosis of prostate cancer, mainly because of their non-invasivity. Several urinary markers for prostate cancer have been suggested over the past decades (i.e. prostatic inhibin-like peptide, thymosin β15, calgranulin B or myeload related protein-14 MRP-14, mini-chromosome maintenance-5, transferrin), but most studies have used small cohorts, making sensitivity and specificity rather heterogeneous, so that it is difficult to assess their potential clinical impact [81]. Sarcosine, an N-methyl derivative of the amino acid glycine that can be non-invasively measured in urine specimens, was recently identified as a differential metabolite highly expressed in invasive prostate cancer cell lines relative to benign prostate epithelial cells. Sarcosine levels were also increased during prostate cancer progression to metastasis [82]. While fPSA molecular isoforms and human kallikrein-related peptidase 2 (KLK2) hold the promise for detection, staging, prognosis, and monitoring of prostate cancer, evidence from large prospective clinical trials remains to be reported [83]. A promising approach to the diagnosis of prostate cancer and to distinguishing between indolent and aggressive tumours has recently been proposed by Vanaja et al. through detection of epigenetic biomarkers. The investigators determined the methylation status of eight genes, including FLNC, EFS, ECRG4, RARB2, PITX2, GSTP1, PDLIM4, and KCNMA1, in non-recurrent, recurrent primary prostate tumours, and benign prostate tissues. Specific hypermethylation of RARB2 and GSTP1 CpG sites were proven useful for diagnosis of prostate cancer, whereas dinucleotide cytosine-guanine (CpG) site hypermethylation of genes FLNC, EFS, ECRG4, PITX2, PDLIM4, and KCNMA1 were associated with prediction of biochemical, local, and systemic recurrence of prostate cancer [84]. Some additional emerging markers at various stages of development show some promise for prostate cancer diagnostics, including KLK2, kallikrein 11, early prostate cancer antigen (EPCA), PCA3, hepsin, prostate stem cell antigen, and alpha-methylacyl-CoA racemase (AMACR) [85–87], but only the future will tell whether appropriate study designs and clinical data analyses will confirm preliminary findings on their clinical usefulness. Moreover, on the assumption that pathologic states within the prostate may be reflected by changes in serum proteomic patterns, Petricoin et al. first suggested that proteomics may be of value in deciding whether to perform a biopsy on a man with a PSA level in the range between 4 and 10 ng/mL [88]. Unfortunately, no further reports confirmed either the clinical usefulness or the favourable cost-to-benefit ratio of using multiple biomarkers for detecting prostate cancer, so that the advantages of using this approach have remained mostly circumstantial, if not elusive, so far. It is, however, predictable that further advances in the application of metabolomics to prostate cancer would help identify additional metabolites, representing end-points of the molecular pathways that are perturbed by other ‘omes’ such as the genome, transcriptome, and proteome. Translation of this basic research into clinical practice would expectantly help develop non-invasive screening procedures to be used for effective cancer diagnosis and prognosis [89].

Conclusions

Although the wide-spread introduction of tPSA testing has revolutionized the approach to identification and clinical management of men with prostate cancer, being circumstantially associated with a decreased mortality, there are still several biological, analytical, clinical, and economical issues that argue against its cost-effectiveness as a screening tool for detecting prostate cancer. The on-going standardization/harmonization efforts, including the introduction of a reference WHO standard, along with the outcomes of recent large epidemiological investigations, demonstrate that the current tPSA thresholds might be revised to increase the diagnostic efficiency, and possibly recalculated according to several demographical variables such as age, ethnicity, family history, and body mass index. In this perspective, the inclusion of SNP genotyping in a nomogram approach might be profitable to improve the positive predictive value of the PSA test, but further studies are needed [90]. The major shortcoming of tPSA screening so far is the lack of any reliable evidence of reduction in prostate cancer-associated mortality, which is due to the large lead-time because of the indolent growth rate, the impossibility to differentiate high-grade, high-volume from indolent cancers, and the treatment-associated morbidity. Since no single tPSA cut-off was proven able so far to efficiently identify men at higher risk of death, the risk of over-diagnosis and over-treatment is tangible and might offset the advantages of early detection [83]. Remarkably, men diagnosed with metastatic prostate cancer during the follow-up after treatment for clinically localized disease report a better quality of life than those who were metastatic at the time of diagnosis. This apparent paradox is justified by the earlier identification of metastases and the more rigorous follow-up in the course of disease, rather than by benefits conferred by the primary treatment [91].

Although the economic costs of early-stage prostate cancer are significant and will increase further in an ageing world, especially in industrialized countries, the cost-effectiveness of PSA-based screening for prostate cancer cannot be easily assessed. In the US, total costs have been estimated to range from US$1.72 billion to US$4.75 billion annually (1990 costs), whereas cost-effectiveness models of population-based prostate cancer screening indicate that such screening could result in as much as US$27.9 billion (1988 values) in charges to the US health care system, which might produce significant wasted expense [92]. Last but not least, it should also be mentioned here that wide-spread screening also carries several ethical, social, and psychological implications, which might ultimately overwhelm the potential benefits. Taken together, all these issues would suggest that there is as yet insufficient clinical and economical evidence to recommend routine population screening with tPSA.

Acknowledgements

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.