The prostate-specific antigen test

Abstract

Before the 1980s, prostate cancer (PC) was considered a deadly disease. The mortality–incidence ratio showed that 1 out of each 2 – 3 PC patients died of this disease. On the other hand, autopsy studies have shown that latent PC is common in middle-aged men. The prostate-specific antigen (PSA), a glycoprotein produced by the epithelial cells of the prostate gland, received FDA's approval in 1986 for monitoring treatment response, and in 1994 as a screening aid for the diagnosis of PC. After the publication of two randomized trials on PC screening using the PSA test, it is generally accepted that systematic PSA-based screening, as compared to a clinical situation with virtually no screening, can reduce suffering from metastatic disease and PC mortality. However, what is also shown is that PSA-based screening coincides with a considerable amount of unnecessary testing and overdiagnosis. Should we abandon the use of the PSA test for the diagnosis of PC, or should we encourage PSA testing and make it freely available for all men at any time? Both the answers should be “No.” What we must do is use the test as wisely as is currently possible and inform men, who want to be tested, in a balanced way about harms and potential benefits.

Keywords::

• biopsy
• prostate
• prostate cancer
• PSA
• screening
• sensitivity
• specificity

Prostate-specific antigen (PSA) is a glycoprotein, with a molecular mass of 34-kD, discovered by Hara et al. in 1971 [1]. PSA is encoded by the KLK3 gene that is almost exclusively produced by the epithelial cells of the prostate gland, and is in fact a more specific organ marker rather than a tumor marker [2]. The PSA test was and is still an outstanding test for monitoring patients who have been actively treated for prostate cancer. Hence, the PSA test received FDA's approval in 1986 for monitoring treatment response and disease recurrence, and in 1994 the FDA approved the use of the PSA test in conjunction with a digital rectal exam (DRE) to test asymptomatic men for prostate cancer.

Before the 1980s, prostate cancer was considered a deadly disease. Reports on series of locally confined prostate cancer showed rapid development of metastases and death. One of the earliest reports of Hanash et al. [3] reported 10 years' survival rates of 52 and 4% for stage A and B disease, respectively. Nowadays, many of these patients would have been very likely classified as M+. In the early 1980s with the digital rectal examination as the only method of diagnosis, 30 – 35% of men had bone metastases, and 45 – 50% had nodal disease. The mortality–incidence ratio in the pre-PSA era showed that on average one out of each two to three prostate cancer patients died of this disease [4]. On the other hand, autopsy studies have shown that latent prostate cancer exists. Depending on a man's age, 29% (men aged 40 – 50 years) to 67% (men aged 70 – 80 years) of men have these latent prostate cancers [5]. In 1994, Catalona et al. showed that early detection of DRE and PSA increased the rate of organ confined disease from 70 – 85% as compared to 30% when detection was driven by DRE only [6]. This landmark study raised the question whether with the use of PSA as a screening tool instead of a monitoring tool the burden of prostate cancer (metastatic disease and mortality) could be reduced.

Screening differs from the clinical use of tests in several important ways. In the clinical situation, patients consult their physician about complaints, that is, the patients request help and the risk and expense of the tests are usually deemed acceptable by the patient. Screening however engages apparently healthy individuals who do not seek medical help. Consequently, the medical and ethical standards justifying a screening program must be of a higher level than those related to a diagnostic process. Potential population-based screening programs involve considering three main issues, namely feasibility, effectiveness and costs. Feasibility is all about organization, facilities and resources, but it is also the acceptability of the screening test used. Effectiveness is evaluated by the extent to which the program affects the main outcome. Screened individuals must have a realistic chance of a better outcome because screening always causes some initial mental and physical damage. Especially in the case of prostate cancer, this is relevant because there is a considerable chance of overdiagnosis and subsequent overtreatment.

In the early 1990s, two large randomized trials on prostate cancer screening using the PSA test were initiated. Although both trials showed contradictory results, it is generally accepted that systematic PSA-based screening as compared to a clinical situation with virtually no screening can reduce suffering from metastatic disease and prostate cancer mortality [7-10]. However, what all these studies also ubiquitously show is that PSA-based screening coincides with a considerable amount of unnecessary testing and overdiagnosis [11]. Especially, the latter is a concern because most of these overdiagnosed cancers (i.e., latent cancers that without screening would never have been detected, lest caused complaints/death) are being treated actively. Active treatment in prostate cancer mostly refers to radical prostatectomy and radiotherapy, both coinciding with frequent and severe side effects.

These drawbacks of PSA-based screening are caused by the simple fact that the PSA test is not prostate cancer specific. This implies that an elevated PSA level in the blood (historically ≥ 4.0 ng/ml today usually ≥ 2.5 ng/ml or 3.0 ng/ml), generally accepted as an indication for further examinations, can also be caused by non-cancerous conditions such as benign enlargement of the prostate (BPH) and prostatitis. In addition, there are prostate cancers, also potentially lethal prostate cancers that do not cause an elevated PSA level. This so-called PSA conundrum results in low positive predictive values at prostate biopsy and a constantly lowering of the PSA cut-off level to avoid a possibly missed diagnosis of prostate cancer. In a purely PSA-based prostate biopsy algorithm (i.e., biopsy if PSA ≥ 4.0 ng/ml), the PPV is around 25 – 35% depending on the population studied [12]. This means that approximately 75% of the prostate biopsies could have been avoided at that point in time. In addition, data from the Prostate Cancer Prevention Trial showed that with applying such a cut-off 75% of the biopsy detectable prostate cancers would have been missed [13]. Perhaps even more important is the fact that with PPVs of around 25 – 30% almost half of the prostate cancers detected are most likely overdiagnosed [11]. So apart from the fact that PSA cannot distinguish between cancer and no cancer, it is also not capable of selectively identifying potentially aggressive prostate cancer.

Is PSA there for useless in the detection of prostate cancer? Far from that, PSA is a very useful marker for initial risk stratification. It has been shown that very low PSA values in middle-aged men can be highly predictive for the future risk of being diagnosed with (aggressive) prostate cancer. The absolute risk of advanced cancer within 15 years in men aged 44 – 50 at time of blood drawing varied from 1% for men with a PSA at the 25th centile of 0.4 ng/ml, to 6% for men at the 90th centile (∼ 1.5 ng/ml), and > 10% for men with a PSA of 3 – 4 ng/ml. This could be translated to different re-screening intervals according to baseline PSA, and hence reduce unnecessary testing and diagnoses [14,15].

In addition, the PSA test together with other relevant clinical information can be combined into the so-called risk prediction tools. These tools are commonly available and can aid in reducing unnecessary side effects and selective diagnosis of potential aggressive prostate cancer [14]. As an example, the multiple validated European Randomized Study of Screening for Prostate Cancer (ERSPC) risk calculator is capable of reducing unnecessary biopsies with approximately 33% [16]. Obviously, there is always room for improvement and currently numerous studies are ongoing exploring the potential value of new biomarkers and combinations [14].

Expert opinion

Prostate cancer is a heterogeneous disease. It can be present while its carrier does not have any complaints; in fact, he can become very old without even knowing that he has prostate cancer. On the other hand, prostate cancer can be lethal within a very short period even at relatively young age.

Data coming from the pre-PSA era clearly show that detection on the basis of a DRE and/or complaints is inadequate. Prostate cancer in that time was considered a lethal disease where in the majority of cases detection came too late and cure was no longer possible. PSA-based screening has shown to be able to reduce the burden of prostate cancer by both reducing suffering from M+ disease and prostate cancer mortality. On the other hand, the incidence of prostate cancer has increased considerably, mainly caused by the diagnosis of potentially indolent prostate cancers which, however, most of the times are being treated invasively. These drawbacks of PSA-based screening, that is, unnecessary testing, overdiagnosis and overtreatment, make screening for prostate cancer highly controversial.

Does this mean that we should stop what we have started? That is, should we abandon the use of the PSA test for the diagnosis of prostate cancer? Definitely not, this will bring us back to the times that prostate cancer often was incurable and caused a lot of suffering.

Should we then encourage PSA testing and make it freely available for all men at any time? Again definitely not, this will bring us to a situation where we will do more harm than good.

What should we do then? It is clear that PSA testing cannot be stopped, so what we must do is use the test as wisely as currently is possible and inform men who want to be tested about harms and potential benefits in a balanced way, as described in the guidelines of many urological associations such as the EAU (Europe), AUA (US) (Table 1) and the SIU (worldwide; http://www.siu-urology.org/psa-aid.aspx). In addition, we must apply (multivariate) risk stratification. There are ample data available that, for example, repeatedly testing men with low PSA values (< 1.0 ng/ml), older men with comorbidities, men with persistently elevated PSA levels due to proven BPH only leads to more harm. In addition, research for new, more specific prostate cancer biomarkers is of high priority; and to reduce over treatment, we should consider more often active surveillance as an initial treatment. Combining all these approaches will undoubtedly result in an acceptable, but effective prostate cancer screening algorithm.

Table 1. The AUA and EAU guidelines/recommendations on PSA screening.

AUA Guideline Statements ( http://www.auanet.org/education/guidelines/prostate-cancer-detection.cfm)	EAU Guidelines on Screening http://www.uroweb.org/?id=23&act=showfull&aid=654 and A. Heidenreich et al. Eur Urol 2012;59:61-71
Shared decision-making for men aged 55 – 69 years that are considering PSA screening, and proceeding based on a man's values and preferences	No evidence for introducing widespread population-based screening programs for early PCa detection in all men. Rather, early detection (opportunistic screening) should be offered to the well-informed man. The decision to undergo early PSA testing should be a shared decision between the patient and his physician based on information balancing its advantages and disadvantages
No PSA screening in men under 40 years of age No routine screening in men between the ages of 40 – 54 years at average risk	A baseline PSA determination at 40 years of age is suggested on which the subsequent screening interval may then be based
A routine screening interval of two years or more, individualized by a baseline PSA level	A screening interval of 8 years might be enough in men with initial PSA levels < 1 ng/ml
No routine PSA screening in men aged 70+ years or any man with less than a 10 – 15-year life expectancy	Further PSA testing is not necessary for men aged > 75 year and a baseline PSA < 3 ng/ml because of their very low risk of dying from PCa

Declaration of interest

The author states no conflict of interest and has received no payment in preparation of this manuscript.