This Issue: PSA Screening
According to statistics from the American Cancer Society, an estimated 218,890 men will be diagnosed with prostate cancer in 2007, representing the most common noncutaneous cancer in U.S. men. Serum prostate-specific antigen (PSA) testing has proved useful in identifying men with potentially aggressive prostate cancer. However, while deaths have declined since screening was introduced, there is controversy over whether its benefits outweigh the risks of follow-up tests and treatments.
Currently, Medicare provides coverage for an annual PSA test for all men age 50 and older. But recent research has prompted calls for the age to be lowered to 40 to provide a baseline for measuring PSA velocity. For the past decade or so, a velocity of 0.75 ng/ml/year has been used as a threshold for biopsy in men with total PSA levels higher than 4 ng/ml, but the National Comprehensive Cancer Network—responding to recent research—now recommends that biopsy be considered for men with a PSA velocity of >.5 ng/ml/year and that the total PSA threshold be lowered for younger men.
Will more aggressive screening at an earlier age lead to detecting more tumors at an earlier, more treatable stage? Or would it simply lead to overdiagnosis of tumors that may never become life-threatening? In this issue of ACP Observer, three experts in the field take opposing views of the issue.
Screening increases likelihood of curable prostate cancer
By Stacy Loeb, MD and William J. Catalona, MD
In 1994, the U.S. FDA approved the Hybritech PSA test as an aid in the early detection of prostate cancer, using a threshold value of 4 ng/ml. Nevertheless, since that time many studies have shown that a considerable proportion of prostate cancers are found at levels <4 ng/ml. For example, the Prostate Cancer Prevention Trial recently showed that among men with a "normal" PSA <4 ng/ml and negative DRE, 15.2% of men had biopsy-detectable prostate cancer (Thompson, Pauler et al. 2004). Moreover, even at PSA levels <4 ng/ml, the total PSA level correlates strongly with prostate cancer risk.
Dr. Loeb // Dr. Catalona
In addition to missing many prostate cancers, by the time the PSA level reaches 4 ng/ml, up to one-third of prostate cancers have spread to the prostate capsule or beyond. Thus, just as prostate cancer risk correlates directly with the total PSA level, so does prostate cancer aggressiveness. As a result of these findings, in 1995, we lowered the PSA threshold for biopsy in our screening study to 2.5 ng/ml.
Among 2,196 men from the screening study (Loeb, Gonzalez et al., 2006) who were treated with radical prostatectomy, we found a smaller proportion of organ-confined disease and significantly larger tumor volumes as the PSA increased (p<0.0001 for both). Furthermore, the likelihood of "insignificant" prostate cancer (defined as being organ-confined with clear surgical margins and no Gleason pattern 4 or 5 and a tumor volume <0.5cc) in the prostatectomy specimen was less than 10% throughout the study period, even in the final era of the study after the PSA threshold was lowered.
Although reducing the PSA threshold for biopsy could theoretically increase the number of unnecessary biopsies, many of these biopsies would be required later as the PSA level continued to rise, and some unnecessary biopsies could be eliminated through the use of PSA derivatives. While it is true that conditions such as benign prostatic hyperplasia (BPH) can cause PSA elevations, the time course of such PSA increases typically differs from that of prostate cancer. Thus, measurements of PSA kinetics, particularly PSA velocity (PSAV), may be useful to increase the specificity for prostate cancer screening. For example, in a recent randomized trial of saw palmetto for men with moderately to severely symptomatic BPH, the annual rise in PSA among the placebo-treated group was only 0.15 ng/ml/year.(Bent, Kane et al. 2006) Since 1992, the urological community has used a PSA velocity of 0.75 ng/ml/year to help distinguish men with prostate cancer.
However, recent evidence has shown that even a more modest PSAV of 0.35 ng/ml/year or above predicted a significantly greater likelihood of prostate cancer death in patients with more than 10 to 15 years follow-up.(Carter, Ferrucci et al. 2006) Overall, PSA velocity is useful to help identify men with aggressive prostate cancer. As a result, the National Comprehensive Cancer Network now recommends that biopsy be considered for men with a PSA velocity >.5 ng/ml/year. Other PSA-based parameters, such as the percentage of free PSA (not complexed to plasma proteins) and PSA density (serum PSA divided by prostate volume), are also associated with both prostate cancer risk and treatment outcomes.
According to data from the Surveillance, Epidemiology and End Results program, 16% of men presented with metastatic disease at the time of diagnosis from 1985 to 1989. The PSA era has resulted in a dramatic stage migration, such that now fewer that 4% of men currently present with metastatic disease (a 75% reduction). Moreover, there has been a 32.5% decrease in the age-adjusted prostate cancer mortality rate in the U.S. since the introduction of PSA screening.
Because the likelihood of curable prostate cancer correlates with the PSA level at diagnosis, we recommend the expansion of PSA screening to begin at age 40 to establish a baseline to measure PSA velocity, using a lower total PSA threshold, and the adjunctive use of PSA-based parameters such as PSAV to increase the specificity of screening for aggressive prostate cancer.
Stacy Loeb MD is at the Department of Urology, Georgetown University School of Medicine, Washington, D.C. William J. Catalona MD is from the Department of Urology, Northwestern Feinberg School of Medicine, Chicago, Ill.
Lowering PSA thresholds increases overdiagnosis risk
By Richard M. Hoffman, FACP, MPH
Among the controversies in prostate cancer screening is when to refer a patient for prostate biopsy following PSA testing. The American Urological Association screening guidelines recommend that men with a PSA level >4.0 ng/mL be referred for biopsy. The 4.0 ng/mL biopsy threshold has also been used by the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial, the National Cancer Institute's randomized controlled trial evaluating the efficacy of PSA screening in more than 75,000 American men. Even though outcome results from the PLCO are several years away from being reported, some prostate cancer experts as well as the National Comprehensive Cancer Network have already begun recommending that biopsy referral be considered for men with PSA levels between 2.5 to 4.0 ng/mL.
Dr. Hoffman
Arguments to lower the PSA threshold for biopsy are based on studies finding cancers with poor prognostic features (moderately or poorly differentiated histology) in men with PSA levels below 4.0 ng/mL and the assumption that earlier diagnosis and treatment will improve clinical outcomes. The best estimate for the prevalence of cancers in men with low PSA levels comes from an analysis of the Prostate Cancer Prevention Trial (PCPT) data, published March 2004 by the New England Journal of Medicine. During the seven-year study, subjects with either an elevated PSA level (>4.0 ng/mL) or abnormal digital rectal examination were recommended to undergo prostate biopsy. Investigators then performed end-of-study biopsies on 2,950 placebo-treated subjects who had consistently normal PSA levels and digital rectal examinations during the entire follow-up period. Overall, 15% of these subjects with normal PSA levels were found to have prostate cancer, though only 2% had high-grade cancers (Gleason scores >7). Cancer was detected in about 25% of subjects with PSA levels between 2.1 to 4.0 ng/mL and 5% had a high-grade cancer.
However, lowering the PSA threshold to detect more cancers has potentially troubling consequences. In the PCPT trial, 44 men required biopsy to find one high-grade cancer. A study by Welch and colleagues, published in August 2005 by the Journal of the National Cancer Institute, estimated that lowering the PSA threshold to 2.5 ng/mL could lead to an additional 3 million men being classified as having an abnormal test. Although this estimate was based on all men undergoing PSA testing, if even a fraction of this testing occurred the ensuing biopsy referrals would still obligate substantial health care resources to find a relatively limited number of cancers. A greater concern is that many of these cancers would never cause clinical problems during a man's lifetime—and would be detected only as a result of screening.
Overdiagnosing prostate cancers unnecessarily subjects patients to psychological burdens and treatment costs and complications. Already, the risk of overdiagnosis with the current PSA threshold is considerable. In a paper published in July 2002, (Etzioni et al, J Natl Cancer Inst. 2002;94:981-90), the authors estimated overdiagnosis rates of 29% for whites and 44% for African Americans among men with PSA-detected cancers. Draisma and colleagues evaluated overdiagnosis by using data from the European Randomized Study of Screening for Prostate Cancer, where the PSA threshold for biopsy was either 3.0 ng/mL or 4.0 ng/mL depending upon the screening round. The analyses, published in June 2003 by the Journal of the National Cancer Institute, estimated a 56% overdiagnosis rate for cancers detected by annually screening men aged 55 to 75. Overall, this screening strategy would increase the lifetime risk of being diagnosed with prostate cancer by 124%.
Although lowering the PSA threshold would find more cancers, the likely tradeoff for this uncertain benefit would be substantial increases in resource utilization and treatment complications. More importantly, we do not yet know whether earlier diagnosis and treatment will improve clinical outcomes. The American and European randomized controlled trials—the gold standard study design for evaluating the efficacy of screening—are still in progress. In the absence of convincing evidence that screening is effective and safe, providers should not screen outside the conventional guidelines, but rather educate their patients about the potential benefits and harms of current screening practices.
Richard M. Hoffman, FACP, MPH, is a professor of medicine at the University of New Mexico School of Medicine and a staff physician at the New Mexico VA Health Care System.
The information included herein should never be used as a substitute for clinical judgment and does not represent an official position of ACP.
Internist Archives Quick Links
Introducing ACP Summer Session
The most popular learning formats from the Internal Medicine meeting offered at only $35 for ACP members! Combine CME and leisure time in San Francisco, CA or Orlando, FL
The International Medical Graduate's Guide to US Medicine and Residency Training
Providing an overview of residency training and life in America, this is the one-source reference tool international medical graduates have been waiting for.