This Article Full Text (PDF) eLetters: Submit a response to this article Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services E-mail this article link to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Barry, M. J. Search for Related Content PubMed Articles by Barry, M. J.

Genitourinary Cancer

Commentary: How Serious Is Getting a Diagnosis of Prostate Cancer?

Medical Practices Evaluation Center, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA

Correspondence: Michael J. Barry, M.D., General Medicine Unit, Massachusetts General Hospital, 50 Staniford Street—9th Floor, Boston, Massachusetts 02114, USA. Telephone: 617-726-4106; Fax: 617-726-4120; e-mail: mbarry{at}partners.org

Received January 14, 2008; accepted for publication January 28, 2008.

Disclosure: No potential conflicts of interest were reported by the author, planners, reviewers, or staff managers of this article.

In this issue of The Oncologist, Drs. Loeb and Catalona propose a more aggressive strategy of screening for prostate cancer with the prostate-specific antigen (PSA) test than has generally been used in the U.S. [1]. The most controversial of these recommendations include beginning screening at the age of 40 (the "stopping age" is not addressed), routinely performing prostate biopsy on men with a total PSA level >2.5 ng/ml (regardless of age), and performing biopsies on men with lower total PSA levels who have a PSA velocity >0.4 ng/ml per year.

Implicit in such recommendations is the growing concern about the sensitivity of the PSA test raised, for example, by data from the Prostate Cancer Prevention Trial (PCPT). The PCPT demonstrated that about 15% of men with a PSA level less than the traditional biopsy threshold of 4.0 ng/ml have prostate cancer at biopsy, with about 15% of those cancers being Gleason grade 7 or higher [2]. In truth, thanks to the PCPT, we now know that the PSA test is quite limited in terms of its ability to separate men with prostate cancer at biopsy from men without prostate cancer, with an area under the receiver operating characteristic (ROC) curve of 0.67, which means that just 67% of the time, a man with prostate cancer can be expected to have a higher PSA level than a man without prostate cancer (for a worthless test, the ROC curve area would be 0.5). Compare that discriminating ability with, for example, the ability of an atrial natriuretic peptide measurement to discriminate patients with heart failure from "all comers" with shortness of breath; 94% of the time, a patient with heart failure will have a higher level than a patient with another cause of dyspnea [3].

To compensate for the poor discriminating ability of total PSA, it is indeed tempting to lower biopsy thresholds and add new PSA derivative biopsy criteria to find as many prostate cancers that may be destined to cause future morbidity and mortality as early as possible. However, maximizing sensitivity comes at a price, and in this case, a double price.

The first price is obvious, and is attributable to the deterioration in specificity that results from any attempt to increase sensitivity. This price will be paid by the additional men without prostate cancer who will undergo biopsy as a result of the more aggressive recommendations. In doing so, they will face the small but finite risks of complications a set of 10–12 biopsies entails, and more importantly, subsequently live with the uncertainty and anxiety about whether their biopsy was a false negative, and the increased surveillance and subsequent testing that inevitably follows [4]. Welch and colleagues have estimated that the number of men in the U.S. aged 40 years and older requiring biopsy would increase from about 3 million with a biopsy threshold of 4.0 ng/ml to about 6 million with a biopsy threshold of 2.5 ng/ml [5]. And adding a relatively low PSA velocity threshold for biopsy would increase that number far further.

The second and steeper price, however, would be paid by the additional men not previously destined ever to present with prostate cancer who would be diagnosed as a result of the more aggressive screening recommendations. The reality is that regular PSA testing substantially increases a man's risk of finding out he has prostate cancer; and the more aggressive the screening strategy, the greater the likelihood a screenee will get prostate cancer. Is this statement just hyperbole? What actually happens is that increasingly aggressive screening increases the likelihood of diagnosing more and more cancers from the underlying prevalence pool, with inevitably diminishing returns. Technically, screening simply converts "unknown" to "known" cancers. But from a patient's perspective, suddenly knowing he has cancer when he didn't before is equivalent to "getting it"; he must face the diagnosis, decide on treatment, and risk the side effects of his treatment.

The influence of PSA screening on the incidence of prostate cancer became obvious early in the "PSA era" in the U.S., when the prostate cancer incidence roughly doubled over just 4 years. Current estimates of the lifetime risk for prostate cancer for men in the U.S. have risen from about 10% in the pre-PSA era to about 17% now [6], with widespread but not universal testing. It is likely, then, that regular PSA testing with a traditional testing strategy would roughly double a man's chances of eventually facing a prostate cancer diagnosis, and that risk would be even higher with the more aggressive screening strategy proposed by Drs. Loeb and Catalona. After all, in the PCPT, about 25% of the average-risk men in the placebo group were diagnosed with prostate cancer in just 7 years (let alone a lifetime) with a very aggressive surveillance strategy including an end-of-study biopsy [7].

Drs. Loeb and Catalona argue that data showing better treatment outcomes for men diagnosed with lower PSA levels justifies paying these prices. However, in the absence of untreated controls, the superior prognosis for men with low PSA levels at the time of diagnosis could just as well mean that population simply includes more men who don't need to be treated in the first place, that is, who were never destined to present with cancer in the absence of screening. But the authors are correct that "overdiagnosis" is really a population concept. At the individual level, one can never be certain which men with prostate cancer are destined to have future morbidity and mortality without treatment. A Gleason 6 cancer in part of one biopsy core in a 65-year-old man may have clinical consequences if he is destined to live another 25 years, while a Gleason 8 cancer in multiple biopsy cores in a 55-year-old man may be inconsequential if he is destined to die of a myocardial infarction in the next year.

An early publication from the Göteberg branch of the European Randomised Study of Screening for Prostate Cancer (ERSPC), although not reporting mortality data, begins to give some sense of the tradeoff between the harms and benefits of population-based prostate cancer screening [8]. In this branch of the ERSPC, about 20,000 men aged 50–66 were randomized to a strategy of biennial invitations for PSA screening (no digital rectal exam) to age 70 or a control group. Participation was high. The biopsy threshold was >3.0 ng/ml (>2.54 ng/ml later in the study), though only sextant biopsies were performed. After 10 years of follow-up, the risk for a prostate cancer diagnosis was only 4.4% in the control group and 8.1% in the intervention group. The number needed to harm with screening was 27; that is, 27 men needed to be screened for 10 years to cause one of them to have to face a diagnosis of prostate cancer. On the other hand, the risk of initially presenting with metastatic prostate cancer or a PSA >100 ng/ml was 0.47% in the control group and only 0.24% in the intervention group. The number needed to screen (NNS) was 435; that is, 435 men needed to be screened for 10 years to prevent one man from initially presenting with metastatic disease. The outcome of initial presentation with metastatic disease is still very susceptible to lead-time bias, as men detected earlier with nonmetastatic disease may still develop metastatic disease later. And obviously this NNS would be considerably larger for the more important outcome of preventing a prostate cancer death, if the trial ultimately shows a mortality benefit. Put in other terms, based on the best randomized trial evidence to date, and using the ERSPC screening strategy for men aged 50–66, serial PSA screening would cause about 16 additional men to face a diagnosis of prostate cancer for each single case of an initial presentation with metastatic disease averted over 10 years.

Two other points in the article by Drs. Loeb and Catalona deserve comment. First, in some circumstances, the authors encourage antibiotic treatment based on PSA levels for the possibility of "subclinical prostatitis." This practice, while widespread, is poorly supported by evidence. Which patients, based on their levels and patterns of PSA elevation, would benefit from a short-term repeat PSA test, with or without a course of antibiotics, is an urgent topic for research. Second, the authors report an ecological study by Jemal and colleagues as showing a relationship between higher regional PSA testing rates and lower prostate cancer mortality rates in the U.S. [9]. The figures in that paper actually do not directly address that association. However, the data needed to test the hypothesis are provided in the paper's table 1. There appears to be no statistically significant relationship between the likelihood of a recent PSA test and population-based prostate cancer mortality rates among either whites or blacks in the regions studied (Fig. 1). In that respect, the results of this study are consistent with a previously published cohort study [10] and another ecologic analysis [11] showing no significant relationship between regional PSA screening intensity and declines in prostate cancer mortality in the U.S.

View larger version (14K): [in this window] [in a new window]   Figure 1. Prostate cancer death rates in 1996–2000 and the percentage of men aged 50 and older with a prostate-specific antigen (PSA) test in the last year based on a 2001 survey in populations served by U.S. cancer registries. Data from Jemal A, Ward E, Wu X et al. Geographic patterns of prostate cancer mortality and variations in access to medical care in the United States. Cancer Epidemiol Biomarkers Prev 2005;14:590–595.

	REFERENCES

Top References

 

1. Loeb S, Catalona WJ. What to do with an abnormal PSA test? The Oncologist 2008;13:299–305.[Abstract/Free Full Text]
2. Thompson IM, Ankerst DP, Chi C 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.[Abstract/Free Full Text]
3. Januzzi JL Jr, Camargo CA, Anwaruddin S et al. The N-terminal pro-BNP investigation of dyspnea in the emergency department (PRIDE) study. Am J Cardiol 2005;95:948–954.[CrossRef][Medline]
4. McNaughton-Collins M, Fowler FJ Jr, Caubet JF et al. Psychological effects of a suspicious prostate cancer screening test followed by a benign biopsy result. Am J Med 2004;117:719–725.[CrossRef][Medline]
5. Welch HG, Schwartz LM, Woloshin S. Prostate-specific antigen levels in the United States: Implications of various definitions for abnormal. J Natl Cancer Inst 2005;97:1132–1137.[Abstract/Free Full Text]
6. Ries LAG, Melbert D, Krapcho M et al, eds. SEER Cancer Statistics Review, 1975–2004. Bethesda, MD: National Cancer Institute, Available at http://seer.cancer.gov/csr/1975_2004/, based on November 2006 SEER data submission, posted to the SEER web site, 2007. Accessed July 11, 2007.
7. Thompson IM, Goodman PJ, Tangen CM et al. The influence of finasteride on the development of prostate cancer. N Engl J Med 2003;349:215–224.[Abstract/Free Full Text]
8. Aus G, Bergdahl S, Lodding P et al. Prostate cancer screening decreases the absolute risk of being diagnosed with advanced prostate cancer—results from a prospective, population-based randomized controlled trial. Eur Urol 2007;51:659–664.[CrossRef][Medline]
9. Jemal A, Ward E, Wu X et al. Geographic patterns of prostate cancer mortality and variations in access to medical care in the United States. Cancer Epidemiol Biomarkers Prev 2005;14:590–595.[Abstract/Free Full Text]
10. Lu-Yao G, Albertsen PC, Stanford JL et al. Natural experiment examining the impact of aggressive screening and treatment on prostate cancer mortality in two fixed cohorts from the Seattle area and Connecticut. BMJ 2002;325:740–745.[Abstract/Free Full Text]
11. Shaw PA, Etzioni R, Zeliadt SB et al. An ecologic study of prostate-specific antigen screening and prostate cancer mortality in nine geographic areas of the United States. Am J Epidemiol 2004;160:1059–1069.[Abstract/Free Full Text]

This Article Full Text (PDF) eLetters: Submit a response to this article Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services E-mail this article link to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Barry, M. J. Search for Related Content PubMed Articles by Barry, M. J.