© The Author 2007. Published by Oxford University Press.
EDITORIALS |
Prostate-Specific Antigen Levels, Prostate-Specific Antigen Kinetics, and Prostate Cancer Prognosis: A Tocsin Calling for Prospective Studies
Affiliation of authors: Department of Urology, The University of Texas Health Science Center at San Antonio, San Antonio, TX
Correspondence to: Dipen J. Parekh, MD, Department of Urology, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr, San Antonio, TX 78229 (e-mail: parekhd{at}uthscsa.edu).
In 1985, a US man had an 8.7% lifetime risk of receiving a diagnosis of prostate cancer and a 2.5% risk in his life of death from the disease; these two risks today are 17% and 3%, respectively (1,2). Prostate-specific antigen (PSA) screening is largely responsible for this increase in diagnosis; its contribution to the 31% fall in mortality between 1990 and 2003 remains a subject of heated debate (2). Along with the substantial increase in detection of the disease has been a concomitant stage migration from tumors of high volume and grade to tumors of lower volume and grade—and presumably of lower biologic potential.
Recognizing the substantial impact of all prostate cancer treatment options on a man's quality of life, including sexual, urinary, and bowel function, interest in active surveillance as a management option for prostate cancer has increased substantially. Active surveillance is a management strategy that acknowledges that many tumors are slow growing and may not cause morbidity or mortality during a patient's lifetime. Active surveillance commonly includes regular PSA tests and rectal examinations as well as periodic prostate biopsy to detect any increase in tumor volume and grade. The premise with this management is that should evidence of increased tumor volume or grade be detected, potentially curative treatment can be initiated to control the disease. This approach differs from that of Fall et al. (3), who followed patients expectantly, without attempting early intervention for cure.
Although a strategy of active surveillance makes intuitive sense, in practice there are few data to guide physicians and their patients with the question of when to treat. Because we are a risk-averse society, most recommendations favor early treatment to avoid the potential loss of the "window of curability" of these tumors. Current strategies often focus on absolute PSA levels, PSA kinetics, and evidence of increase in tumor grade or volume (4). In practice, the most common reason for treatment is a "worrisome" PSA level or an increase in PSA velocity.
In this context, Fall et al. show the challenges with using PSA measures to identify tumors that meet the ultimate definition of "biologic aggressiveness," causing the death of the patient. It must be recognized, however, that the patients in this series were substantially different from the majority of patients with localized prostate cancer today in countries where PSA screening is highly prevalent (such as the United States): PSA levels were above 10 ng/mL in 48% of the patients in that population, and more than 60% of men had tumors that were palpable on rectal examination. At diagnosis, these were high-volume tumors, tumors more likely to exhibit stronger PSA kinetics driving more favorable PSA operating characteristics than would be expected from populations containing higher fractions of screen-detected cancers (5). As a result, we would expect less optimistic performance of these PSA measures in a typical US population.
Nevertheless, the receiver operating characteristic curves in Fig. 3 in Fall et al. (3) continue the sobering saga concerning PSA as a tool for detecting or measuring prognosis of prostate cancer. We show the implications of these data when applied to clinical decision making (Table 1). In the ideal circumstances, a man on active surveillance for prostate cancer expects that his PSA data will tell his physician one of two things: 1) he has a potentially deadly tumor that must now be treated or 2) he has an indolent tumor that can be safely watched, sparing him the side effects of treatment. If we simply use a PSA doubling time of 5 years as a guide, in this group of patients approximately 36% of deadly tumors would be missed and 40% of men with indolent tumors would be treated unnecessarily. If our goal is to find all lethal cancers, the very best trigger for treatment would be a PSA value of 7 ng/mL; the trade-off for this threshold is that approximately 80% of men with nonlethal tumors would then be treated.
|
The implications from these data are troublesome. Without becoming despondent and simply treating all men with prostate cancer—accepting the substantial risk of overtreatment and its consequences—how can we better identify the patient with low-risk disease in whom active surveillance is a reasonable option for management? The first step must be to acknowledge that, although PSA level and its kinetics are clearly associated with the behavior of individual prostate cancers, they are not sensitive or specific for the tumor that will ultimately cause harm to a patient. We have found the same results with PSA as a measure of the risk of prostate cancer. Specifically, in our analysis of the Prostate Cancer Prevention Trial (6), we found that other measures (rectal examination findings, family history, ethnicity, age, and prior prostate biopsy results) all provided independent predictive value of prostate cancer risk. Currently, the same methodology of incorporating multiple prognostic factors is used for the evaluation of a man with newly diagnosed prostate cancer. Variables that physicians currently assess include measures of tumor aggressiveness (Gleason grade), measures of tumor volume (PSA, local stage, number of involved cores, percent or linear amount of tumor per core), and measures related to the time an individual man will be at risk from his tumor (age and comorbidities). Adding to the complexity of this assessment is that these variables can change over time while the man is on active surveillance as PSA testing, rectal examination, and prostate biopsy are periodically repeated and as his health status also changes with time. A further complexity is the need to incorporate treatment selected by the patient. Because treatment is perhaps the greatest confounder in the assessment of tumor risk (i.e., a life-threatening tumor may never be recognized if treated and cured), future assessments must control for this treatment.
Thirty years ago, Thomas Chalmers noted that the moment you recognize a clinical challenge, "randomize the first patient" (7). These data demand that clinical trials commence now to examine surveillance strategies to help patients and their physicians identify and treat tumors that will otherwise be life threatening and to carefully follow those that will not. Our limited health care resources and the quality of life of an enormous number of men will benefit from this for decades to come.
NOTES
Supported by National Cancer Institute grant #5UO1CA86402 from the Early Detection Research Network.
D. J. Parekh is supported by The University of Texas Health Science Center at San Antonio Institute for Integration of Medicine and Science Mentored Career Development Award.
REFERENCES
(1) Seidman H, Mushinski MH, Gelb SK, Silverberg E. Probabilities of eventually developing or dying of cancer—United States, 1985. CA Cancer J Clin (1985) 35:36–56.
(2) Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, 2007. CA Cancer J Clin (2007) 57:43–66.
(3) Fall K, Garmo H, Andrén O, Axelson A, Adolfsson J, Adami H-O, et al. Prostate-specific antigen levels as a predictor of lethal prostate cancer. J Natl Cancer Inst (2007) 99:526–32.
(4) Zhang L, Loblaw A, Klotz L. Modeling prostate specific antigen kinetics in patients on active surveillance. J Urol (2006) 176((Pt 1)):1392–7. discussion 1397–8.[CrossRef][Web of Science][Medline]
(5) Thompson IM, Ankerst DP, Chi C, Lucia MS, Goodman PJ, Crowley JJ, et al. Operating characteristics of prostate-specific antigen in men with an initial PSA level of 3.0 ng/ml or lower. JAMA (2005) 294:66–70.
(6) Thompson IM, Ankerst DP, Chi C, Goodman PJ, Tangen CM, Lucia MS, et al. Assessing prostate cancer risk: results from the Prostate Cancer Prevention Trial. J Natl Cancer Inst (2006) 98:529–34.
(7) Chalmers TC. Randomization of the first patient. Med Clin North Am (1975) 59:1035–8.[Web of Science][Medline]
Related Articles in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2007 99: 526-532.
J Natl Cancer Inst 2007 99: 493.
J Natl Cancer Inst 2007 99: 493.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||