American Association for Clinical Chemistry
Better health through laboratory medicine
June 2009 Clinical Laboratory News: PSA Testing: Yes, No, Maybe?

CLN Banner Logo

June 2009: Volume 35, Number 6

PSA Testing: Yes, No, Maybe?
Recent Research, Guideline Updates Add to Controversy
By Gina Rollins

Since its introduction in the 1980s, the prostate-specific antigen (PSA) test has been a controversial screening tool for prostate cancer. Because of the varied but typically long natural history of prostate cancer and the test's limited specificity in detecting the disease, there have been persistent doubts about the efficacy of population-based screening. Yet in the U. S., at least, PSA testing is common: an estimated 80% of men older than age 50 have had a least one PSA level taken. Although intense research continues in all aspects of prostate cancer epidemiology, pathogenesis, diagnosis, risk stratification, and treatment, no study to date has unequivocally demonstrated that the benefits of PSA screening outweigh its potential harms. Recently published interim reports from two landmark clinical trials did not end the confusion, and guideline updates of major professional organizations have moved already discordant PSA-related recommendations further apart.

However, despite all the ambiguities, the PSA test still is the mainstay of prostate cancer screening, and experts agree that the marker will continue to have a solid role in the future, even as the race accelerates to find biomarkers that are more specific for prostate cancer. “PSA is not going away. We will need this assay because it is an informative biomarker,” said Hans Lilja, MD, PhD, attending research clinical chemist in the departments of clinical laboratories, surgery, and medicine at Memorial Sloan-Kettering Cancer Center. “But we also need to use PSA testing a bit more wisely. In and of itself, it is not sufficient to warrant a decision on biopsy, but we will have it as a trigger to do additional testing.” Lilja chaired the prostate cancer writing group for the National Academy of Clinical Biochemistry Laboratory Management Practice Guideline (NACB LMPG) on the use of tumor markers in testicular, prostate, colorectal, breast and ovarian cancers.

An Unacceptable Number

Even as the role of PSA testing remains controversial, there has been a well-described stage migration towards earlier stage of disease at the time of diagnosis. Today only about 5% of cases are metastasized at the time of initial diagnosis, in comparison to about 50% before the PSA test was available. Likewise, there has been an approximate 30% decline in prostate cancer mortality, which also roughly corresponds to introduction of the PSA test.

If PSA testing has not been definitively linked to these findings, it has without doubt lead to over-detection and over-treatment of prostate cancer. One study estimated that over-detection increases with age, rising from about 27% in men age 55 to about 56% at age 75 (J Natl Cancer Inst 2003; 95:868-78). An interim report from the European Randomized Study of Screening for Prostate Cancer (ERSPC), released in the same issue of the New England Journal of Medicine as an interim report from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO), bears out the over-detection and over-treatment problem (N Engl J Med 2009;360:1320Ð8). The study, which involved 182,000 subjects in seven European countries, found the rate of death from prostate cancer was 20% lower in the screening versus control arm, but 1,410 men would need to be screened and 48 cases of prostate cancer would need to be treated to prevent one death from the disease. “It has been known that there is a considerable amount of over-detection, but it was brought back into perspective with solid numbers for the first time in our article, with the ratio of 48 cases detected to prevent one death at nine years of follow-up,” said lead author Fritz Schröder, MD, PhD, professor of urology at Erasmus University in Rotterdam, The Netherlands. “That is an unacceptable number.”

The amount of over-detection would not be such a concern if it did not also lead to over-treatment with the consequent costs and psychologically devastating side-effects like erectile dysfunction and, more rarely, urinary or fecal incontinence. Despite the fact that active surveillance is an option recommended by various guidelines, upwards of 90% of men in the U.S. choose to be treated, even if they have low-grade disease that is unlikely to create problems. “The over-detection carries a stigma psychologically, which often results in aggressive treatment. Urologists, patients, and the medical community are saying ‘you’re a young man, you have a little prostate cancer, and we better do something.' So most men end up receiving aggressive treatment,” explained lead author of the PLCO study, Gerald Andriole, MD, professor and chief of urologic surgery at Washington University in St. Louis and the Siteman Cancer Center at Barnes-Jewish Hospital.

Limitations of both biopsy and PSA testing also contribute to over-treatment. “The problem with surveillance is that 10% to 15% of biopsies are underestimating disease, so that drives a lot of people to have treatment just in case their disease is worse than indicated by biopsy or because they have to have repeat biopsies,” said John W. Davis, assistant professor of urology at the University of Texas M.D. Anderson Cancer Center in Houston. At the same time, research has shown that a prostate cancer risk exists even at low PSA levels. For example, one study found that 6.6% of men with PSA levels <0.5µg/L had prostate cancer (N Engl J Med 2004; 350:2239Ð46). That biopsy and PSA levels are imperfect diagnostic and screening tools reflects the variable natural history of prostate cancer, according to Davis. “Two men can be identical in prostate size, shape and PSA levels and have totally different clinical scenarios,” he noted.

Facts on Prostate Cancer and PSA

Prostate cancer is the most common tumor in men and the second leading cause of cancer mortality in men. In the U.S. each year there are nearly 220,000 new cases of the disease and approximately 28,000 deaths. Despite its prevalence, prostate cancer has a relatively low mortality rate: while an estimated 16% of men will be diagnosed with the disease at some point, the lifetime risk of dying from it is only about 3%. Autopsy studies indicate that at least one-third of men older than age 50 have histologic evidence of prostate cancer, but up to 80% of these tumors are not clinically significant, meaning they are of small size and grade.

Serum elevation of PSA, which is a glycoprotein produced by the epithelial cells of the prostate, is a highly sensitive marker for prostate disease, but not necessarily prostate cancer. Other conditions that cause PSA levels to rise include prostatitis, benign prostatic hyperplasia, and trauma to the urethra or prostate. At lower PSA levels, the assay’s specificity for cancer is as low as 20%, but it increases to 40% to 50% when levels rise to ≥10µg/L. Population-based median levels are <0.6µg/L for men ≤ age 50, but rise with age. Levels also are higher in African American men.

The PSA test initially was cleared by the FDA in 1986 as a marker for monitoring treatment in men who already had been diagnosed with prostate cancer. In 1994, formal approval was extended for use in detecting prostate cancer in men age 50 and older. However, by that time the test already had entered the mainstream, as evidenced by data recently published from the ongoing Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO). Subjects in the control arm were to receive usual care, which could include PSA testing, and even in the first year of the study—1993—40% had been screened (N Engl J Med 2009;360:1310–9).

What's the Evidence?

The cloud of controversy surrounding over-detection and over-treatment exists at least partially because there are no evidence-based, standardized thresholds for what constitutes a clinically actionable PSA score. For instance, a cut-off of 4 µg/L has been commonly adopted, but it was introduced based on a single report. Many clinicians now use a level of 2.5µg/L as a trigger for biopsy and 1.5µg/L in younger men with a family history of prostate cancer. However, these thresholds have not been formally endorsed by professional organizations. In fact, the NACB LMPG recommends not using age-specific reference intervals and notes that the reported benefits from lowering the clinical decision limit below 4µg/L are too uncertain to make a blanket recommendation. (DOI:10.1373/clinchem.2008.105601)

Analytical challenges represent still more wrinkles in the PSA picture. Although clinicians and patients generally assume that any PSA test is like another, there are two reference standards, and depending on which one an assay is calibrated to, mean PSA levels can vary by at least 20%, although one study found as much as a 38% difference (J Urol 2008; 180:1959Ð63). The Hybritech Tandem-R PSA assay, which was the first widely adopted FDA-cleared commercial assay, was standardized using an absorptivity of PSA of 1.42mL/mg/cm, but subsequent analysis showed the correct absorptivity to be 1.84mL/mg/cm. The WHO First International Standard for PSA (IRR96/670) adopted the correct absorptivity, and many manufacturers have calibrated their assays to this standard, but assay-to-assay discrepancies still remain. An estimated biological variation of up to 20% also can skew interpretation of any test results.

The NACB LMPG and other guidelines recommend that labs report and that clinicians and patients know the particular assay used for each PSA reading. This is especially noteworthy given that it is common for men to seek out second opinions. “Men being monitored for prostate cancer or who have prostate cancer and are on active surveillance hop around a lot in terms of both doctors and labs,” observed Andriole.

MD Anderson Cancer Center has gone so far as to distribute cards with a grid for men to record the date of each PSA test, the test level, and the assay it was run on. “All of that information is typically available on the lab report, and people move around a lot,” said Davis. “As someone who is on the receiving end of patients showing up with a diagnosis of prostate cancer, but with very limited information about prior test results, that type of documentation can be very useful.”

The Take-Home Message

Clinicians and researchers had hoped data from the PLCO and ERSPC trials would provide definitive answers regarding the role of PSA screening, but that was not the case, at least from the interim reports. In contrast to ERSPC, the PLCO study, which involved nearly 77,000 men enrolled at 10 U.S. study centers, found no significant reduction in prostate cancer mortality associated with screening (N Engl J Med 2009;360:1310-9). However, the report reflected a median 7 years of follow-up versus 9 years for ERSPC. The trials also used different PSA thresholds to trigger diagnostic evaluation (4ng/mL for PLCO versus 3ng/mL for ERSPC). News coverage played up other differences between and potential weaknesses of the studies, but Andriole and Schršder are united in the theme of over-detection and over-treatment. “As a medical community, we need to sort out from the prostate cancers discovered, which are truly destined to be problems and should be treated aggressively,” said Andriole. “We have to find a way not to treat the lion's share of these tumors.”

Interestingly, many men took in a different message from mainstream media reports of the studies, according to Davis. “The main conclusion reached was not that PSA testing was not effective, but that treatment was not effective,” he noted.

Schröder expects that subsequent follow-up from the ERSPC study will enable more definitive conclusions. “Within 2 to 4 years there will be enough strength of evidence that population-based screening will be introduced in most European studies,” he predicted. “Crucial to that will be decreasing over-detection and over-treatment.” Still, he believes the results thus far are note worthy. “The message has changed dramatically. In the past there was a sort of Russian roulette—we didn't know if there would be any benefit for screening but we did know for sure there would be harm. Now we can relate to men who consider PSA testing that there is benefit in terms of decreasing the chance of dying of prostate cancer, but we still know there is possibly harm,” he explained.

Discordant Recommendations

Even as the medical community was coming to terms with the PLCO and ERSPC reports, the American Urological Association (AUA) broke ranks with other professional organizations in April by issuing updated clinical guidelines that called for the PSA test to be offered to well-informed men age 40 or older who have a life-span of at least 10 years. While the AUA expanded its recommended screening parameters, other guideline groups, at least thus far, are not making changes, or are even contracting prior recommendations. For example, the NACB LMPG, which was updated in December 2008, concluded that there was not enough evidence to support or refute routine population-based PSA screening and that outcomes from trials such as PLCO and ERSPC would be needed for a definitive recommendation. Meanwhile, in 2008 the influential U.S. Preventive Services Task Force modified its existing recommendation that there was insufficient evidence to endorse screening to advise that screening men older than age 75 was no longer recommended. For now, the American Cancer Society (ACS) is holding steady to its prior recommendation that routine or mass screening is not warranted, but that it should be a shared decision between physicians and patients age 50 or older who have a life expectancy of 10 years.

With all the discordant advice swirling about, informed choices will be critical, according to Schröder. “If someone understands the risk of possibly being diagnosed with cancer and treated for cancer that he would never have had trouble from, and still wants to be screened, then he can't be refused,” he said. “But the accent has to be on well-informed.” In a written commentary, Otis Brawley, MD, chief medical officer of the ACS, suggested that ”men know the facts about screening and talk to a doctor who knows the facts. You will have to know enough to judge whether the physician knows what he or she is talking about. Make your own choice about screening. Do not be overly influenced by what others decide. Avoid mass screening where informed decision-making is not possible.”

The Search is On

A flurry of research is aimed at identifying and validating biomarkers that will better detect which prostate cancers are likely to advance. Listed here are selected biomarkers under investigation.

Type of Biomarker Candidate

Proposed Use/Comments

Circulating biomarkers

PSA subfractions

May help discriminate between malignancy and benign conditions

Human kallikrein 2

More sensitive than PSA at detecting extracapsular extension

Molecular urine/tissue markers

Prostate cancer antigen 3

Highly expressed in prostate cancer compared to other genitourinary tissues and non-neoplastic prostate tissues

Alpha-methylacyl-CoA racemase (AMACR)

Over-expressed in prostate cancer

Cell/gene tests

Circulating prostate cells RT-PCR gene targets PSA, hK2, and PSMA mRNAs

Measurements of the frequency in the shedding of circulating prostate/tumor cells in blood using RT-PCR assays to define invasive disease stage


Tumor suppressor; gene somatically deleted or mutated in some prostate cancers; protein can be detected by IHC; decreased levels associated with higher grade and stage.

Excerpted from NACB LMPG for Use of Tumor Markers in Testicular, Prostate, Colorectal, Breast and Ovarian Cancer, DOI:10.1373/clinchem.2008.105601.

The Quest for New Markers

The plethora of recommendations and conflicting evidence underscore the necessity of addressing the over-detection and over-treatment issue, either by improving on the PSA test itself or using it in combination with other biomarkers. Numerous refinements have been proposed to improve the specificity of total PSA concentrations, including PSA density, PSA velocity, PSA doubling time, percentage of free PSA, percentage proPSA to freePSA, and complexed PSA. Patient age, family history, ethnicity, prior biopsy history, and comorbidities should also be taken into consideration. Nomograms have been developed to help estimate a man's risk of prostate cancer given various PSA levels, and the Dutch Prostate Cancer Research Foundation has created an online risk calculator based on PSA levels, biopsy history and results, digital rectal exam results, and other factors.

The NACB LMPG recommends using percentage of free PSA to help distinguish prostate cancer from benign prostatic hypertrophy when total PSA levels are between 4µg/L and 10µg/L and digital rectal exam is negative. Complexed PSA has comparable cancer detection to total PSA, but yields a ”somewhat better specificity in a narrow concentration range,” according to the guidelines. Still, complexed PSA levels alone do not produce a specificity similar to percentage of free PSA, the guidelines state. The updated AUA guidelines observe that there is no consensus on optimal strategies for using the various modifications to total PSA.

With no easy answers from the various PSA refinements, clinicians and researchers are looking to the promise of novel biomarkers that would have more robust sensitivity and specificity, not only for screening but also risk assessment and treatment monitoring. “If we had another marker that would enable you to monitor patients without biopsy and make it a urine or blood-based proposition—and accurate—then the acceptance of active surveillance would go up tremendously,” Davis predicted.

Many researchers are pursuing alternative biomarkers. The NACB LMPG identified 14 promising markers, while a review article published in the December 2008 issue of Clinical Chemistry highlighted 21 (2008;54:1951Ð60). The lead author of that analysis, Girish Sardana, PhD, predicts that a combination of several biomarkers will be needed to change the over-detection, over-treatment paradigm. “Because of the heterogeneity of prostate cancer, finding an all-in-one marker that can diagnose most cases is unrealistic. The trick will be determining which patient population can benefit from a particular marker and understanding how it works in the grand scheme of a panel of markers,” he said. Sardana is a biomarker research and development associate at the Ontario Cancer Biomarker Network in Toronto. While a plethora of potential biomarkers have been proposed, Sardana and his co-authors “highlighted the most promising ones that have been validated the most. So many markers are out there but many have only been studied in vitro or had one or two papers published about them,” he said.

The National Cancer Institute is supporting research involving numerous detection methods, including micro-RNA assays, gene methylation, gene fusion, metablomics, and proteo-imaging. “The field is wide-open,” said Sudhir Srivastava, PhD, MPH, chief of the Cancer Biomarkers Research Group at NCI and principal investigator in the NCI's Early Detection Research Network. Several promising lines of NCI-sponsored investigation together represent a “comprehensive analysis of prostate cancer biology that hopefully will lead to a panel of biomarkers that can be detected in urine,” he said. “When creating a panel, we have to look at each biomarker's contribution to diagnostic performance. The goal is to increase specificity, but if we just add biomarkers there will be a point beyond which there isn't much additional benefit.”

As the medical community awaits these research developments, PSA will remain an indispensible, if controversial, tool in prostate cancer detection and treatment.

Dr. Lilja holds a patent for a free PSA immunoassay.