April 2008 Clinical Laboratory News: Beyond PSA

CLN Banner Logo

April 2008: Volume 34, Number 4

Beyond PSA
Do Novel Markers of Prostate Cancer Offer Better Specificity?
By Deborah Levenson

Since the 1980s, when it first became available as a monitoring test for prostate cancer, the serum prostate specific antigen (PSA) test has become so common that the majority of American men older than 50 now receive it. Widespread PSA screening has increased the number of prostate cancer biopsies to almost one million annually, according to the Urological Sciences Research Foundation. But PSA’s specificity, anywhere from 20% to 40%, leaves much to be desired. The majority of men who get biopsies based on an elevated PSA levels receive a negative result, while at least 15% of men with PSA levels less than 4.0 ng/mL, considered by many to be the upper limit of the normal range, harbor prostate cancer. Meanwhile, most men diagnosed with prostate cancer have forms that aren’t aggressive at that time and don’t need immediate treatment, making a marker that determines whether cancer is likely to spread beyond the prostate the ultimate goal of research on novel biomarkers.

Advances in molecular diagnostics, more knowledge of the human genome, and proteomic tools that allow quicker identification of genes have allowed researchers to propose a growing number of novel prostate cancer biomarkers that could be more specific than PSA, and in some cases, add prognostic information. But at this point, it’s difficult to predict which will ultimately become useful tools for patient care. “In the search for biomarkers of prostate and other cancers, some fall down, and others rise up,” commented Hans Lilja, MD, PhD, who chairs a National Academy of Clinical Biochemistry (NACB) subcommittee now completing guidelines on using biomarkers for prostate cancer. He is Attending Research Clinical Chemist at the Memorial Sloan Kettering Cancer Center in New York, N.Y., Professor of Laboratory Medicine at Lund University in Sweden, and patent holder of other prostate cancer assays employing free PSA and human kallikrein 2. NACB guidelines on biomarkers for several other cancers are also forthcoming.

Which prostate cancer markers stand the best chance of helping physicians determine which men with high PSA scores need first or repeat biopsies or inform men with a family history of disease about their risk? The answer depends on whom you ask. A group of prominent prostate cancer researchers disagree about several markers, but do concur that work on a very few of the many new markers is most likely to yield useful clinical tools. These include: prostate cancer gene 3 (PCA3), early prostate cancer antigen-2 (EPCA-2), and certain prostate cancer risk alleles that would help determine if a man with a family history is at risk.

Which Emerging Prostate Cancer Markers Look Promising?
Prominent prostate cancer researchers weighed in on the leading markers.



Risk Alleles


Eleftherios Diamandis, MD, PhD
Chair, NACB Guidelines Committee
University of Toronto

Interesting, but may miss some cancers. Exam is very unpleasant for patients.


Questions validity of published research.

Hans Lilja, MD, PhD
Chair, NACB Guidelines Prostate Cancer Subcommittee
Memorial Sloan Kettering Cancer Institute

Interesting, but preliminary data suggest it may miss some cancers. Exam is unpleasant for patients. Sample work-up requires special equipment and expertise. Changes in copy number of PSA-mRNA influences performance.

Predictive accuracy of a single routine PSA measurement (AUC 0.76-0.79) outperforms by far the predictive accuracy of five risk alleles (AUC 0.57), according to published research. * It’s unknown whether combination of PSA plus the 5 risk alleles and other new risk alleles predicts prostate cancer better than PSA alone.

Interesting, but validation of target biomarker (identification of structural EPCA) is lacking and data is discordant from those shown for EPCA. Reported EPCA-2 data were based on study cohort with selection biases.

William Catalona, MD
Northwestern University

Adequate sample collection is key. Apt to result in false negatives.

Very promising. Some studies show certain alleles may predict aggressive disease.

Promising, but published results need extensive validation.

Leonard Marks, MD
UCLA School of Medicine

Sensitivity and specificity far exceed PSA’s, but will not replace PSA.


Only Johns Hopkins offers the test, so data are limited.

Robert Getzenberg, MD, PhD
Johns Hopkins University

Requires special equipment and expertise.


Has published a proof of principlestudy that shows ability to identify aggressive disease.

*NEJM 2008; 358: 910-919

PCA3: Now Available

PCA3 is the only novel prostate cancer marker with both a commercial assay and coverage by Medicare and some private insurers. First discovered in the early 1990s in the lab of William B. Isaacs, PhD, at Johns Hopkins, PCA3 was first known as DD3 and used by Canada-based DiagnoCure to develop a qualitative assay called uPM3. In November 2003, Gen-Probe acquired exclusive worldwide diagnostic rights to this prostate cancer gene from DiagnoCure. Gen-Probe currently offers PCA3 analyte specific reagents to five U.S. reference labs: Bostwick Laboratories (Glen Allen, Va.); Caris/MPI (Phoenix, Ariz.); LabCorp (Burlington, N.C.); Mosaic Diagnostics (Dallas, Texas); and OURLab (Nashville, Tenn). A commercial PCA3 test, PROGENSA PCA3, is approved in Europe as a CE-marked assay.

Getting a PCA3 result can be an involved process for both the clinician and the lab. Following a digital rectal exam to massage the prostate to ensure that enough prostate cells get into a urine sample, that sample is collected and mixed with a lysis buffer and RNA stabilizer. Using a dual assay, the test then separately quantifies both PCA3 and PSA mRNAs from prostatic cells in the urine, explained Junqi Qian, MD, Director of Molecular Diagnostics for Bostwick Laboratories. He added that the dual assay relies on Gen-Probe’s Target Capture to purify target RNA, Transcription-Mediated Amplification to increase the number of copies of the mRNAs, and a hybridization protection assay to quantify the mRNA copy number. The reference labs charge between $350 and $450 for the test, Qian noted.

More Specific than PSA

Results from a recent trial show that PCA3 has sensitivity and specificity that well exceed PSA’s. Researchers from the Urological Sciences Research Foundation (USRF) in Culver City, Calif. evaluated the PCA3 assay by performing a receiver-operating-curve (ROC) analysis on results from 233 men with serum PSA levels persistently 2.5 ng/mL or greater and who had at least one previous negative biopsy. For the PCA3 ratio, the AUC was 0.678. The research team calculated a PCA3 using a ratio of PCA3 mRNA levels over PSA mRNA levels (PCA3 mRNA/PSA mRNA x 1000) and determined that a PCA3 ratio of 35 corresponded to the greatest diagnostic accuracy, with a sensitivity of 58% and specificity of 72% (Urology 2007;69: 532–535). The study’s lead author is Leonard Marks, MD, who is USRF’s medical director, Clinical Professor of Urology at UCLA School of Medicine, a consultant and study investigator for both Gen-Probe and Beckman Coulter, and consultant to Onconome. The companies are all developing prostate cancer biomarkers.

Other researchers expressed cautious optimism about PCA3. Eleftherios Diamandis, MD, PhD, Chair of the committee overseeing development of the NACB guidelines, and Head of Biochemistry at Mount Sinai Hospital and Toronto Medical Laboratories and University of Toronto, noted two potential problems: the test may miss some cancers and the necessary digital rectal exam may be sufficiently unpleasant that some patients might reject the test.

In addition, accurate results could depend on the quality of that exam. “Some cancer may be located such that it’s difficult to squeeze those cells into the urethra. You can get false negatives,” noted William Catalona, MD, Professor of Urology at Northwestern University and Director of the Clinical Prostate Program at Northwestern’s Robert H. Lurie Comprehensive Cancer Center. He receives research support and honoraria from Beckman Coulter, which markets a serum PSA assay, and is the author of many papers on SNPs being evaluated as risk markers for prostate cancer. Catalona also worries about the cutoff for the PCA3 test. “There could be problems with some men if they have ratios close to 35. I see some of these men, and I wonder if some cancerous cells didn’t get into the urethra,” he said. Still, Catalona sees the test as valuable—when the score is high—in making decisions regarding second biopsies. “But the test leaves doubt when it’s negative,” he added.

Lilja considers PCA3 “interesting,” but noted that the expertise of lab personnel running the test and changes in the copy number of PSA-mRNA could also influence test performance. He added, “It will be important to understand whether cancers not detected by the test are significant or insignificant (indolent). Also, at present it is unknown whether PCA3 outperforms or provides further enhancement to that of percent free PSA.”

A study now being planned by Gen-Probe aims to show how much enhancement PCA3 can add to PSA. Intended to yield data for FDA submission, the trial will involve approximately 1,000 patients at 10 U.S. sites to evaluate the predictive value of the test in patients with a previous negative biopsy and an elevated serum PSA level, according to Harry Rittenhouse, PhD, Senior Director of Gen-Probe’s Cancer Program. Rittenhouse says the company hopes to begin the trial within the next 2 or 3 years, and is now adapting a PCA3 assay to a second generation platform called Panther. “This platform will be automated and easy to run,” he noted. Gen-Probe is also planning a study that would determine the extent to which PCA3 and certain gene fusions present in about 50% of prostate cancer patients offer complementary diag-nostic and prognostic information, Ritten-house added. The company is now testing approximately 4,000 patients in 37 countries who are enrolled in the GSK REDUCE clinical study, which aims to determine if the drug Avodart (dutasteride) can reduce incidence of prostate cancer in men at increased risk of the disease.

Risk Alleles: Up and Coming

Risk alleles may be another promising tool for determining whether a man is at risk for developing the disease and would benefit from aggressive screening. The idea that risk is determined by a group of genes versus a single gene is relatively new. “We used to think that one bad gene segregated down through families. Now we have about a dozen risk alleles located in a few regions of the genome,” Catalona explained.

A recent study shows that SNPs in five chromosomal regions— three at 8q24 and one each at 17q12 and 17q24.3— plus family history have a cumulative and significant association with prostate cancer. A multinational group of researchers led by Henrik Grönberg, MD, PhD of Karolinka Institutet in Stockholm, Sweden evaluated these SNPs in 2,983 Swedish subjects with prostate cancer and 1,781 controls. After researchers selected the most significant SNP from each of the five regions and included it in a multivariate analysis, each remained significant after adjustment for other SNPs and family history. Together, the five SNPs and family history were estimated to account for 46% of the cases of prostate cancer in the study population. In men who had any five or more of these factors associated with prostate cancer, the odds ratio was 9.46, as compared with men without any of the factors (NEJM 2008; 358: 910–919).

“Predictive value of a single routine PSA measurement outperforms by far the predictive accuracy of the combination of all risk alleles,” said Lilja, who noted that the PSA measurement provides an AUC for prediction of 0.76– 0.79, compared to 0.57 for the combination of all risk alleles. “At present, we do not know whether the combination of routine PSA plus all five risk alleles predicts prostate cancer better than PSA alone,” he noted. But he added, “we don’t yet have the data telling us if combining genetic analysis with other methods improves the accuracy of the prediction models.”

Catalona is excited about where the findings might lead. “Researchers are just getting started on this. It’s like a jigsaw puzzle. After you’ve worked on it a while, fitting the pieces gets easier because it’s more evident where they should go.” He added that some research on these genes has also yielded evidence that they can help predict risk of not just cancer, but of aggressive disease.

A new test based upon the five gene variants is forthcoming from Proactive Genomics (Winston-Salem, N.C.). Founded by Wake Forest University researchers who were part of the team whose paper was published in NEJM, the start-up company is developing a commercial assay called the Focus5 Prostate Cancer Risk Test. According to Aubrey Turner, MS, Proactive Genomics’s Genetic Counselor, the company expects to launch the test sometime this year at a cost of less than $300. At this point, the test won’t be used to predict aggressive disease because the research team didn’t find that the five genes did so among the study’s Swedish subjects who are less likely to be diagnosed with aggressive disease than Americans because they get less regular screening.

In February, deCODE Genetics launch-ed the deCODE PRCa assay, a reference laboratory test that detects a total of 8 SNPs, including 2 on chromosomes X and 2. A recent paper describes these 2 variants’ association with risk of prostate cancer (Nature Genetics 2008; doi:10.1038/ng.89).

EPCA-2: Showing Promise

EPCA-2 is also attracting attention because a recent study demonstrates that it is potentially more specific than PSA and might detect aggressive disease. A team from Johns Hopkins School of Medicine led by Robert H. Getzenberg, PhD, Professor of Urology and Director of Research at the James Buchanan Brady Urological Institute measured EPCA-2 in 330 patients separated into several groups: those with PSA levels less than 2.5 ng/mL, PSA levels of 2.5 ng/mL or greater with negative biopsy findings, benign prostatic hyperplasia, organ-confined prostate cancer, non-organ-confined disease, and prostate cancer with PSA levels less than 2.5 ng/mL. Using a cutoff of 30 ng/mL, the EPCA-2 assay had a 92% specificity (95% CI, 85%–96%) for healthy men and men with benign prostatic hyperplasia and 94% sensitivity (95% CI, 93%–99%) for overall prostate cancer. The specificity for PSA in these selected groups of patients was 65% (95% CI, 55%–75%). Additionally, EPCA-2 showed some success in differentiating between disease confined to the prostate and cancer that’s spread beyond it (AUC 0.89, 95% CI, 0.82–0.97), in contrast to PSA (AUC 0.62, 95% CI 0.50–0.75).

Getzenberg, who intends to submit a test to the FDA with data from studies funded by Onconome and whose institution stands to profit from a future EPCA-2 assay, called his research “a proof of principle study.” He described EPCA-2 as a potential adjunct to PSA that’s less expensive and labor-intensive than PCA3. He pointed to EPCA-2’s advantages. “It’s more specific for prostate cancer than PSA and is not elevated in men with benign prostatic hyperplasia or prostatitis. It can differentiate between organ-defined disease at the time of surgery and more aggressive disease that’s spread outside the prostate.” But he added, “Clearly it doesn’t have as much data as PSA, which has been run

in millions of men.” Getzenberg is now studying EPCA’s velocity and doubling time and is examining EPCA-2 levels in men with Gleason score of 8–10 in an effort to relate the marker to the aggressiveness of cancers.

EPCA-2 might initially be useful in identifying men with elevated PSA levels that may need a prostate biopsy, Getzenberg suggested.

The paper on EPCA-2 has drawn enthusiastic coverage from the healthcare and general press. But all the attention has proven to be a double-edged sword, Getzenberg said. “I wish the study drew less attention. Now urologists have to explain to patients that it will take a while to have a clinical tool, and that adds to patients’ frustrations.”

Catalona expressed cautious hope regarding EPCA-2, noting that Getzenberg’s data needs to stand up in a population-based setting. “EPCA-2 could be very good. But I’d like to see it validated in a real-world situation when the samples do not come from the freezer. Often, when you take a new test and use it to measure all-comers, it doesn’t do as well,” Catalona remarked. If the data holds up under careful, lengthy scrutiny, EPCA-2 should have the potential to one day displace PSA, he added.

Calling EPCA-2 “interesting,” Lilja not-ed uncertainty about it. “The reported EPCA-2 data were based on a study cohort with selection biases that should exclude us to draw any conclusions on the test performance. Hence, we need to wait for additional data to make any conclusions on the putative value of the test.”

Diamandis was very critical of the EPCA-2 research and the attention it has received. In a letter to Clinical Biochemistry (2007; 40: 1437–39), he decries the media attention and questions the Urology paper’s characterization of EPCA-2 as a nuclear protein, whether small, localized tumors would yield measurable amounts of EPCA in serum, and the consistency of an ELISA assay used by researchers.

“We’ve done some validation and will continue to validate EPCA-2 to see how well this marker works,” responded Getzenberg, who added that new data from tests on samples from 1,200 men will appear in two papers that are now in progress.

The End of the PSA Era?

If any of these markers truly take off, what will happen to PSA, which despite its imperfections has been credited with making advanced disease upon presentation a relatively rare occurrence? PSA won’t be replaced any time soon, according to Getzenberg and Marks. “PSA has been around for 25 years. Clinicians are comfortable with it, and they should be. It’s changed the course of the disease. Now, men don’t often present with advanced disease. The tumors we see are smaller and we’re seeing more men within the window of time when they’re treatable,” Getzenberg emphasized. Even if new markers someday surpass PSA as a screening and diagnostic tool, the original prostate cancer marker will still have value in monitoring for disease and marking prostate volume, Marks predicted. “If the prostate is removed, the PSA level should be 0. So if it rises, you know there’s a problem. And in men with benign enlargement, PSA correlates extremely well with volume of the prostate.”

Page Access: