European Urology

1. Introduction

Screening and early diagnosis of prostate cancer are among the most important and, at the same time, controversial issues of medicine, with enormous health and economic relevance. Prostate-specific antigen (PSA) can be considered a simple and safe test, within appropriate boundaries, and although robust cost-effectiveness studies are still lacking, current evidence suggests that PSA screening may lie within acceptable limits [1].

Theoretically, the chance to diagnosis prostate cancer earlier would allow clinicians to treat a higher percentage of clinically localized disease, which might guarantee patients higher chance of definitive healing. More practically, however, several issues related to the available diagnostic tools (PSA and prostate biopsy), the risk of overdiagnosis (eg, lead time), inability to identify properly clinically insignificant cancers, and complications of all the curative treatments offer several points for discussion. Above all, the most critical issue is the lack of any demonstration that a screening program significantly reduces cancer-related mortality, which is the most relevant outcome of any screening program.

Despite the fact that the results of two randomized controlled studies (RCTs), the Prostate, Lung, Colorectal and Ovary (PCLO) trial [2] and the European Randomized Screening for Prostate Cancer (ERSPC) trial [3], are expected in the coming years, many scientific societies have recommended early detection policies for prostate cancer, based mainly on PSA sampling.

Current guidelines for the use of PSA for prostate cancer screening differ substantially between the United States and Europe. In the United States, the US Preventive Services Task Force (USPSTF) concluded that the evidence is insufficient to recommend for or against PSA screening. Both the American Urological Association (AUA) [4] and the American Cancer Society (ACS) [5], however, recommended that screening be offered to men ≥50 yr. More recently, the National Comprehensive Cancer Network (NCCN) practice guidelines considered a more aggressive protocol offering baseline digital rectal examination (DRE) and PSA at age 40 [6]. In contrast, the Advisory Committee on Cancer Prevention in Europe stated that “screening for prostate cancer is not recommended as health-care policy and also the most recent view of the World Health Organization (WHO) is that mass screening should not be supported prior to the availability of large-scale, randomised data.” Similarly, the European Association of Urology (EAU) guidelines concluded that there was a current lack of evidence to support mass screening but affirmed, indeed, that the use of PSA in combination with DRE as an aid to early diagnosis in well-informed patients was less controversial and was widely used in clinical practice [7].

Two questions should be considered. First, can a patient truly be informed if medical professionals and researchers are still investigating what the best course of action is? Second, is the conceptual difference between mass and opportunistic screening practically relevant, considering the high penetration of PSA testing in communities of men? The impact of expert opinion and mass media messages inviting the population >50 yr of age to do a simple and safe blood test significantly cut the theoretical difference between mass and opportunistic screening. For these reasons, it is likely that, given the widespread practice and acceptance in communities of men, even null findings from the two ongoing RCTs probably will not reduce future PSA testing rates [8].

2. Prostate-specific antigen for screening and early detection of prostate cancer

The role of PSA in early diagnosis of prostate cancer is debatable. The definition of normal total PSA (tPSA) values and the nonoptimal sensitivity and specificity of the test, mainly for patients with low serum level, are the most critical issues. In the past decade, several studies demonstrated the opportunity to lower the PSA cut-off point used to trigger the indication for biopsy from 4 ng/ml to 3 ng/ml and, subsequently, to 2.5 ng/ml, with the purpose of diagnosing those 20–25% of patients harboring cancer with PSA values <4 ng/ml. Although the lack of sensitivity could be ameliorated by decreasing the tPSA threshold value, this would lead to a decrease in specificity and positive predictive value (PPV), resulting in a substantial burden of biopsies arising from false-positive tests.

At the same time, the reduction of PSA cut-off points might also be followed by a significant rise in the number of diagnoses of clinically insignificant cancers, which, finally, might cause considerable overtreatment for many patients. Moreover, Thompson et al evaluated the prevalence of prostate cancer in those patients in the control arm of the Prostate Cancer Prevention Trial (PCPT) who underwent protocol biopsy despite tPSA value <4 ng/ml. In this subgroup of patients, no PSA value was able to exclude the presence of prostate cancer, with 7% of patients with PSA <0.5 ng/ml diagnosed with cancer [9]. In the other words, the risks of a positive prostate biopsy and of high-grade disease (Gleason score ≥7) rise with increasing PSA level, but there is not a clear cut-off up to PSA of 4 ng/ml that separates a population with prostate cancer from those without it.

3. Derivatives of prostate-specific antigen

In the past decade, several studies have been performed with the aim of identifying PSA derivatives or new markers that are able to increase sensitivity and specificity of tPSA, especially for those patients with PSA <10 ng/ml [10] for whom the PSA produced by benign prostatic hyperplasia (BPH) may act as a significant confounding factor, especially in patients >50 yr.

Some authors supported the finding that serial PSAs starting at age 40 will allow practicing clinicians to determine which patients are at higher risk for developing prostate cancer and, specifically, will allow clinicians to calculate and to follow the PSA velocity (PSAV) at a time when BPH is less prevalent and PSA is more predictive of cancer [11]. Recent data have determined median PSA levels to be 0.6–0.7 ng/ml for men aged 40–49 yr, 0.9 ng/ml for men aged 50–59 yr, 1.2 ng/ml for men aged 60–60 yr, and 1.5 ng/ml for men >70 yr [12], and [13]. Moreover, Loeb et al showed that men in their 40 s and 50 s with a PSA between their age-specific median value and 2.5 ng/ml have a 14.6-fold and a 7.6-fold increased risk, respectively, of developing prostate cancer [12]. According to the previous data, the question is what to do with these young patients at increased risk who fall below the biopsy threshold of 2.5 ng/ml. A possible solution might be represented by the application of PSAV. Obviously, the historical PSAV of 0.75 ng/ml should be considered as a cut point for prostate biopsy in patients with PSA between 4 and 10 ng/ml or the PSA is not correctly applicable in patients with PSA <4 ng/ml. Data from the literature highlight that PSAV of 0.4 ng/ml per year should be considered as a significant cancer predictor in young men with low PSA value [11]. In agreement with Schroder et al, however, we do believe that the early application of PSA testing starting at age 40 and serial early PSA determinations is still an experimental concept that should be considered with caution [14].

The discovery of various molecular forms of PSA in the early 1990 s facilitated the development of assays for selective immunodetection of PSA, which is not bound to plasma proteins. The chance of dosing free PSA (fPSA) and calculating a ratio of free PSA to total PSA (%fPSA) has allowed a major improvement in patient selection for biopsy, reducing the number of biopsies in those patients with PSA ranging from 2.5 to 10 ng/ml [15]. Recently, Walz et al demonstrated that %fPSA can accurately predict the prostate cancer in a subgroup of patients with PSA ≤2.5 ng/ml [16]. The NCCN guidelines affirm that fPSA is not generally used in deciding whether or not to perform an initial biopsy. In selected cases, however, biopsy may be considered for patients with %fPSA <25% [6]. The EAU guidelines have also selected %fPSA <20% in men with elevated PSA levels as a valid parameter to trigger the indication to perform a prostate biopsy [7]. In contrast, conflicting results were reported with another PSA derivative, complex PSA (cPSA) [17].

Another way to increase specificity of tPSA seems to be the use of molecular subforms of PSA. Specifically, the structural composition of fPSA in serum differs measurably depending on whether it is released from hyperplastic or from cancerous tissue. In the present issue of European Urology, Jansen et al reviewed the role of PSA isoforms and human kallikreins in the diagnosis of prostate cancer [18]. The data of that nonsystematic review of the literature suggest that in patients with PSA ranging from 2 to 10 ng/ml, the level of precursor isoform of PSA (pPSA) and the ratio of pPSA to fPSA (%pPSA) might increase the diagnostic accuracy of tPSA and %fPSA. Data concerning the performance of %pPSA for tPSA levels of <2 ng/ml are currently insufficient to draw any conclusions. On the contrary, the studies evaluating other PSA isoforms [(−2)pPSA, (−4)pPSA, (−5)pPSA, (−7)pPSA, benign PSA] failed to show any adjunct diagnostic value compared with fPSA and tPSA [18].

Should the dosage of pPSA be introduced in clinical practice? According to Jansen et al, the retrospective nature of many available studies, the use of different assay systems and antibodies to measure the concentration of PSA subforms in the available studies, and the small number of cases analyzed in most of the studies discourage recommendation of the use of pPSA in everyday clinical practice. Prospective studies with appropriate methodological design should reconfirm similar findings first. Moreover, according to the five phases of biomarker development for the early detection of cancer proposed by the US National Cancer Institute (NCI) [19], only the first two steps were completed for pPSA (discovery and predictive analysis), with the last three (assay refinement, blinded characterization, and retrospective/longitudinal evaluation) still to be done [20].

4. Putative novel markers for prostate cancer detection

Considering the limitations of PSA as a biomarker for prostate cancer screening, new markers that reduce the number of unnecessary biopsies and that differentiate indolent from aggressive cancers are needed to decrease the risk of overtreatment. Prostate stem cell antigen, α-methyl coenzyme-A racemase, prostate cancer gene 3 (PCA3), early prostate cancer antigen, human kallikrein 2 (hK2), and hepsin are promising markers that are currently undergoing validation studies. In their review of the literature, Jansen et al analyzed the potential diagnostic role of other members of the tissue kallikrein family different from human kallikrein 3 (hK3). The review supported a possible role of human kallikrein 2 (hK2) and the ratio of hK2 to fPSA (%hK2) as potential additional markers to select those patients with tPSA values ranging from 4 to 10 ng/ml who should undergo prostate biopsy. For tPSA levels <4 ng/ml, however, the role of hK2 appears to be limited [18]. Similarly to pPSA, this promising biomarker completed the first two phases (discovery and predictive analysis) of the biomarker development process for the early detection of cancer [20]. Regarding the other members of the tissue kallikrein family, future research should clarify whether there is a possible role in prostate cancer diagnosis.

To date, tPSA is still the most commonly used marker for diagnosis of prostate cancer. Currently, the situation is probably not significantly different from 2004, when the last WHO international consultation on new markers took place. The consultation concluded that total serum PSA was still the best marker, although no specific cut-off point for normal PSA could be defined [21].

Two attractive future perspectives are represented by the use of nomograms to calculate the risk of prostate cancer on biopsy, including a panel of kallikrein markers, and by the implementation of urinary markers. Recently, Vickers et al demonstrated that the addition of fPSA and intact PSA and hK2 improved the prognostic accuracy models including only age and PSA (“laboratory” model) and including age, tPSA, and DRE (“clinical” model) [22]. In contrast, measurement of PCA3 mRNA concentration in the urine after DRE seems to be superior to %fPSA for predicting repeat prostate biopsy outcome and may be indicative of clinical stage and significance of prostate cancer [23].

5. Conclusions

At a time when definitive evidence on reduction of cancer-specific mortality following screening or early detection programs for prostate cancer are lacking, the scientific community should be more cautious in proposing aggressive measures such as lower PSA and lower tPSA cut-off points or early PSA testing during the fourth decade of life. More aggressive PSA screening in younger men will detect more cancers but will not necessarily result in improvements in death rates, cancer-specific survival, or overall outcomes for those men. Moreover, the option of an active surveillance program cannot be regarded as a valid shield against the risk of overtreatment, considering that most of the urologists (and the patients) are more prone to propose (and to accept) radical surgical or radiotherapy treatments, regardless of long-term complications. Although promising, the role of PSA isoforms and of human kallikrein is, to date, not significantly helpful in this dispute.
Conflicts of interest: The authors have nothing to disclose.