Current Controversies in Oncology

Introduction

By Jacek Jassem, MD, PhD
Head, Department of Oncology and Radiotherapy
Medical University of Gdansk, Poland

Prostate cancer is one of the most frequent causes of mortality in men. There is clearly an urgent need for effective screening strategies in this malignancy. The prostate-specific antigen (PSA) test was introduced into clinical practice in the mid-1980s and approved by the U.S. Food and Drug Administration (FDA) in 1992. Since then, the PSA blood test has been widely used as a screening tool. At first glance, detecting the tumor at an early, asymptomatic, and curable stage of development seems to be of high relevance. Additionally, the PSA test is a simple, widely available, harmless, and inexpensive method.

In spite of all this, the value of PSA screening has become one of the most controversial issues in contemporary oncology. The reason is obvious: so far, there is no direct evidence that PSA screening affects disease-specific survival. Indeed, two major randomized studies initiated on both sides of the Atlantic in the early 1990s are still ongoing. Screening advocates claim that the increased discovery rate of early curable cancer, a fall in the incidence of metastatic disease, and a reduction in related mortality after widespread PSA screening began to provide enough evidence to recommend the procedure. On the other hand, prostate cancer is a peculiar malignancy, with a large proportion of tumors that are indolent and pose no life threat. Consequently, screening may lead to overdetection and overtreatment of insignificant prostate cancers with the resulting costs, side effects, complications, and increased patient anxiety.

Additionally, the test is neither highly specific nor sensitive: most patients with an elevated PSA level do not have cancer, whereas cancer may be present despite normal PSA values. False positive results cause anxiety and further increase the rate of unnecessary diagnostic procedures. As a result, the likelihood of harm may outweigh the prospect of benefit. The currently used PSA threshold (= 4 ng/mL) is also debated: lowering it will improve sensitivity, but only at the expense of increased anxiety and additional unnecessary biopsies. Last but not least, the cost-effectiveness of PSA screening has been difficult to calculate owing to the lack of data on screening effectiveness.

The issue of prostate cancer screening was debated in the July 2007 issue of ASCO News & Forum. However, the previous debate presented the perspective of U.S.-based physicians, which may not be representative of other geographical regions. For example, despite the lack of consensus on the efficacy of PSA screening among major American societies, the PSA blood test has become a routine practice in the United States and Medicare has approved payment of one PSA screening test per year. In Europe, the PSA test is available and paid for in some countries, but it is only accepted as “opportunistic screening” in men who, in spite of controversy and uncertainty, choose to be tested.

In this column, the problem of PSA screening is debated by two European experts in the field: Monique Roobol, MD (in favor) and Fritz Schröder, MD (against). I hope you will find this debate interesting and inspiring.


“Current Controversies in Oncology” is a forum for the exchange of views on topical issues in the field of oncology. The views and opinions expressed therein are those of the authors alone. They do not necessarily reflect the views or positions of the editor or of the American Society of Clinical Oncology.

Men Should be Screened for Prostate Cancer

By Monique J. Roobol, MD
Department of Urology,
University Medical Centre
Rotterdam, The Netherlands

The World Health Organization website (www.who.int) addresses the subject of early detection of cancer as follows: “Early detection of cancer greatly increases the chances for successful treatment,” a sentence that leaves little doubt about the value of early detection. Yet prostate cancer, the second highest cause of cancer death in men in the Western world, is in the midst of a controversy. This is mainly due to the lack of evidence that early detection of prostate cancer will indeed reduce disease-specific mortality. This evidence will come from two randomized controlled trials^1,2 which have been ongoing for more than a decade, and from which results can be expected within a few years at the earliest.

But why should a man wait for a possible mortality reduction of 20% or 25% (on a population level) when a screening test is already available that might detect his prostate cancer while it is still curable? Prostate cancer develops asymptomatically into an incurable stage. The only way to avoid this outcome is through active screening.

Screening can provide reassurance.
Although the PSA test has limitations, it is unique in the field of early detection of cancer—not because it is able to selectively identify men who have prostate cancer but because it can identify men who do not need further testing at all, or for at least several years. The result of a serum PSA test can alleviate a man’s worries about the uncertainty of having prostate cancer.

This application of the PSA test is confirmed by several studies. Within the European Randomised study of Screening for Prostate Cancer (ERSPC)³, two different centers identified men at a very low risk of developing prostate cancer. The Swedish center concluded that a man age 50 to 65 years with a PSA level < 1.0 ng/mL can safely forego screening for three years.⁴ The Dutch center, which used longer follow-up data, estimated this period to be as long as eight years.⁵ Recently, data from the Malmö Health Care Project, which has a very long follow-up time (i.e., 25 years), concluded that men age 44 to 50 years with a baseline PSA = 0.5 ng/mL are at a very low risk of developing cancer.⁶ Knowing the PSA distribution of men age 55 to 65 years, where 42% still have a PSA = 1.0 ng/mL5, it is evident that a simple, relatively cheap, non-invasive test can reassure many, if not most, of the men who worry about prostate cancer. This possibility of risk stratification considerably increases the effectiveness of early detection of prostate cancer and can reduce the frequency of PSA and additional testing (with the associated costs and stress).

The stage shift due to active screening is undeniable.
As mentioned above, it is to be expected that active screening indeed results in the detection of relatively more prostate cancers with favorable characteristics, a so-called stage shift. This is confirmed by reviewing the SEER database over a period that encloses both the pre-PSA and the PSA era. Advanced stage or metastatic tumors, which in 1980 constituted 25% of newly diagnosed and staged cases, have become a rarity; by 2002, only 4% of prostate cancer cases were metastatic at the time of diagnosis.6 Numerous studies comparing tumor characteristics of clinically and screen-detected prostate cancer cases show similar results.^7,8,9,10 In all studies, the screen-detected prostate cancers showed more favorable characteristics and a higher percentage were suitable for treatment with curative intent. A recently published modelling study concluded that PSA screening accounts for approximately 80% of the observed decline in distant-stage incidence of prostate cancer.¹¹

Active treatment is more effective, especially in prostate cancers detected at an early stage.
Because the window of opportunity for cure is relatively short, it is vital to apply active treatment before the cancer cells invade neighboring tissues and organs or metastasize to other sites. A comparison of active treatment (i.e., radical prostatectomy) with a watchful waiting policy in organ-confined disease showed a decreased hazard ratio of developing metastases from 27.3% to 13.3% (63%, p=0.03) for those men who underwent surgery.¹² Studies have also shown that, among men with clinical stage T1c prostate cancer, those with a PSA level of 2.6 to 4.0 ng/mL had the greatest rate of organ-confined disease, the lowest pathologic Gleason grade, and the greatest 10-year progression-free survival rate.¹³ But perhaps even more convincing that treatment is more effective when applied to early diagnosed prostate cancers are the Kattan nomograms on outcome in the management of prostate cancer. Within these nomograms, the most powerful predictors for a decreased chance of recurrence after active treatment are a low PSA value, low clinical stage, and Gleason score, characteristics obviously belonging to early detected prostate cancer cases.^14,15

One of the major reasons for the existing controversy on prostate cancer screening is the fear of detecting potentially indolent prostate cancer cases which may lead to unnecessary invasive treatment. However, it is known that most PSA-detected cancers have the characteristics of clinically significant cancers that require immediate active treatment.¹⁶ This can, if the patient desires, be postponed by applying active surveillance.

Survival of prostate cancer patients in the PSA era has improved.
More effective treatment will subsequently affect disease-specific survival and mortality rates. This trend is already clearly visible in many countries. Relative survival rates of prostate cancer patients have increased from below 80% to 99% in the PSA era.¹⁷ Prostate cancer mortality has declined in both the United States and Europe.

Prostate cancer mortality in the United States declined after 1994 by -4.17% (95% CI -4.34 to -3.99); in the United Kingdom, the rate of decline after 1992 was -1.14% [-1.44 to -0.84] per year. This difference in decline is directly related to the uptake of PSA testing.^18,19

Why wait until the outcome of randomized trials?
It is understandable that health policymakers await the outcome of the randomized trials before starting population-based screening programs that will consume a considerable part of the total health care budget. However, as the available evidence clearly shows that prostate cancer can be identified and treated successfully after early detection, it is also understandable that men want to be tested. Ideally, a screening test or combination of tests should be able to identify men who have lifethreatening prostate cancer. Such a test is not available and the question is, how long it will take to become available, if it ever will? Is this a reason to deny men the opportunity to detect a possible life-threatening prostate cancer while it is still curable?

While awaiting the introduction of population- based screening programs, we should join forces and use all readily available evidence to apply a screening algorithm, which entails careful patient selection and effective treatment.

References:
1. Roobol MJ, Schröder FH (guest editors). BJU Int. 2003;92 Suppl 2:117-122.
2. Prorok PC, Andriole GL, Bresalier RS, et al. Control Clin Trials. 2000;21 Suppl 6:273s-309s.
3. Aus G, Damber JE, Khatami A, et al. Arch Intern Med. 2005;165:1857-61.
4. Roobol MJ, Roobol DW, Schröder FH. Urology. 2005;65:343-6.
5. Lilja H, Ulmert D, Björk T, et al. J Clin Oncol. 2007;25:431-6.
6. National Cancer Institute. Surveillance, Epidemiology, and End Results registry. Available at http://seer.cancer.gov/
7. Efstathiou JA, Chen MH, Catalona WJ, et al. Urology. 2006;68:342-7.
8. van der Cruijsen-Koeter IW, Vis AN, Roobol MJ, et al. J Urol. 2005;174:121-5.
9. Aus G, Bergdahl S, Lodding P, et al. Eur Urol. 2007;51:659-64.
10. Pelzer AE, Colleselli D, Bektic J, et al. BJU Int. 2008;[Epub ahead of print].
11. Etzioni R, Gulati R, Falcon S, et al. Med Decis Making. 2008;[Epub ahead of print].
12. Holmberg L, Bill-Axelson A, Helgesen F, et al. N Engl J Med. 2002;347:781-9.
13. Antenor JA, Roehl KA, Eggener SE, et al. Urology. 2005;66:156-60.
14. Stephenson AJ, Scardino PT, Eastham JA, et al. J Natl Cancer Inst. 2006;98:715-7.
15. Zelefsky MJ, Kattan MW, Fearn P, et al. Urology. 2007;70:283-7.
16. Thompson IM, Ankerst DP, Chi C, et al. JAMA. 2005;294:66-70.
17. American Cancer Society. Available at www.cancer.org/downloads/STT/2008CAFFfinalsecured.pdf. Accessed May 2, 2008.
18. Collin SM, Martin RM, Metcalfe C, et al. Lancet Oncol. 2008;[Epub ahead of print].
19. Etzioni R, Tsodikov A, Mariotto A, et al. Cancer Causes Control. 2008;19:175-81. Epub 2007 Nov 20.

The Potential Harm from Prostate Screening

Fritz H. Schröder, MD, PhD
University Medical Centre
Rotterdam, The Netherlands

Screening for prostate cancer has remained a controversial issue in spite of clear demonstration of stage migration toward more favorable disease with the use of PSA-based testing. The main reason for the persisting controversy is that level 1 evidence for effectiveness with respect to decreasing prostate cancer mortality from randomized controlled trials is not yet available. In the present situation of uncertainty, it does not make sense to prevent men, who wish to be screened and who are well informed, from undergoing testing. The issue of introducing screening as a health care policy and the testing of well-informed individuals must be considered separately.

What speaks against screening?
Harm due to screening is known, benefit has yet to be shown. What is the harm which may result from screening? Next to the potential physical and mental side effects of the various steps of the diagnostic workup¹ and the potential complications of the biopsy procedure² in a situation of low specificity of testing (only one in three to five men have a positive biopsy), the main harm must be seen in overtreatment. The only available randomized study of surgical treatment of locally confined, clinically diagnosed prostate cancer shows a 5.8% absolute reduction in prostate cancer mortality at 10 years. This reduction in mortality is achieved at the price of treating 15 men who are not at risk of dying within 10 years of prostate cancer.³ Overdiagnosis from screening, usually leading to overtreatment, has been shown to occur in more than 50% of cases of screen-detected prostate cancer. ⁴ The number of screen-detected prostate cancers that must be treated to prevent one prostate cancer death is presently unknown, but can be assumed to be much higher than the 15:1 ratio in the study referenced above. Next to the trouble and cost of unnecessary treatment, there are considerable side effects of the most common treatment regimens: radical prostatectomy and external beam radiotherapy. The Prostate Cancer Outcome Study (which is not based on centers of excellence) only showed that, with five-year follow-up, erectile dysfunction, incontinence, and bowel problems in the radiotherapy group were highly prevalent.⁵ Similar findings were obtained in the management of screen-detected cancers within the ERSPC.⁶ At present, definitions of overdiagnosis, indolent prostate cancer, and clinically insignificant prostate cancer are controversial and arbitrary. However, even with the application of similar criteria (probably due to differences in patient populations), the proportion of cancers classified as potentially insignificant after radical prostatectomy varies between 5.8% and 49%.⁷

Does stage migration lead to an absolute reduction of aggressive disease?
Stage migration toward more favorable disease occurs with the use of PSA-driven testing. This has been demonstrated in multiple settings and does not require referencing. However, the question of whether stage migration leads to an absolute decrease in aggressive prostate cancer, specifically in metastatic disease, remains controversial. A recent paper from the Swedish section of ERSPC seems to suggest that this is the case.⁸ However, rates of metastatic disease between the screening and control arms of this study are only reported at the time of diagnosis, and these findings have limitations.⁹ Data from the Rotterdam ERSPC section allowed investigators to make an inventory of M1 disease comparing the control group and three rounds of screening, including interval cancers.¹⁰ The difference in incidence of metastatic disease between the two arms amounted to 74%. This figure is almost identical with the relative reduction of metastatic disease seen in the SEER database.¹¹ However, the difference in the numbers of cancers found in the screening and the control arms amounted to 73%. The data show that in this setting, over a period of almost 5.5 years, there is no difference in the proportion of metastatic cases found between the screening and control arms. Stage migration may not translate into a benefit in terms of prostate cancer mortality; the apparently lower proportion of aggressive cases may be the result of diagnosing more cancers if PSA-driven testing is applied.

Does the decreasing prostate cancer mortality rate in many parts of the world reflect the success of widespread screening?
In the United States, prostate cancer mortality between 1992 and 2008 has decreased by 16.3% in absolute numbers (1992: 34,240 prostate cancer deaths; 2008: 28,660 deaths). If a joint point analysis is applied to prostate cancer mortality, the decrease during the same 10- year period is 41%. Another section of the same paper claims a decrease of 35%.^11,12 In the Netherlands, during the period from 1993 to the end of 2004, prostate cancer mortality has decreased by 22%.¹³ Can we assume that these changes are due to screening? The answer is clearly no. The contribution of radical prostatectomy in the only available randomized study is modest and is likely to be smaller in a screendetected setting. This is also confirmed by a recent comparison of screen-detected and clinically detected prostate cancers that showed seven-year progression-free survival rates differing by only 6% between both groups (77% versus 83%).¹⁴

Other improvements in management and prevention are likely to contribute to the decrease in mortality. Even in a screening setting, about 20% of cases are diagnosed with T3 disease. Major progress in this disease category has been made. Several studies have shown improvements of five-year specific survival in the range of 15% to 25%.¹⁵ Data from the U.S. Male Health Professionals Study have recently shown a strong association between the occurrence of advanced prostate cancer or death from prostate cancer, events which were reduced in long-term statin users by 51%.¹⁶ 25% of American men are long-term statin users. It is safe to conclude that the decrease in prostate cancer mortality seen around the world is not exclusively a result of screening, but that screening might be contributory, as suggested recently by two different modelling experiments.¹⁷

Screening upon request
In the present situation of uncertainty, testing cannot be refused to men who wish to be screened. There is, however, a professional obligation to provide careful information with the available nomograms and accompanying narratives. One example is the ERSPCbased prostate cancer risk calculator, which is available online (www.prostate-riskindicator.com). In addition to providing necessary information, the available nomograms also allow improvements of the specificity of testing (avoiding unnecessary biopsies). Maximization of detection in the present situation seems ill-advised. Recommendations of the national comprehensive cancer center guidelines to lower PSA cut-offs to a level of 2.5 and to indiscriminately apply yearly screening are not in line with a rapidly developing understanding of available screening procedures.¹⁸ In addition, emerging evidence suggests that immediate biopsy may not be necessary in men with low PSA values, in spite of knowledge of the presence of moderately and poorly differentiated cancers in such men.¹⁹


References
1. Essink-Bot ML, De Koning HJ, Nijs HGT, et al. J Natl Cancer Inst. 1998;12:925-931.
2. Raaijmakers R, Kirkels WJ, Roobol MJ, et al. Urology. 2002;60:826-830.
3. Bill-Axelson A, Holmberg L, Ruutu M, et al. New Engl J Med. 2005;352:1977-84.
4. Draisma G, Boer R, Otto SJ, et al. J Natl Cancer Inst. 2003;95:868-78.
5. Potosky AL, Davis WW, Hoffman RM, et al. J Natl Cancer Inst. 2004;96:1348-9.
6. Korfage IJ, de Koning HJ, Habbema JD, et al. BJUI. 2007;99:801-6.
7. Schröder FH. Ann Oncol. 2006;17:201-206, supplement 10.
8. Aus G, Bergdahl S, Lodding P, et al. Eur Urol. 2007;51:659-54.
9. Schröder FH, Habbema DF, Roobol MJ, et al. Eur Urol. 2006;51:588-90.
10. van der Cruijsen-Koeter IW, Vis AN, Roobol MJ, et al. J Urol. 2005;174:121-125.
11. American Cancer Society. Available at: http://www.cancer.org/downloads/STT/2008CAFFfinalsecured.pdf. Accessed
May 4, 2008.
12. Jemal A, Siegel R, Ward E, et al. CA Cancer J Clin. 2008;58:71-96.
13. Karim-Kos HE, de Vries, E, Soerjomataram I, et al. Eur J Cancer. 2008;44:1345-89.
14. Roehl KA, Eggener SE, Loeb S, et al. Urol Oncol. 2006;24:465-71.
15. Bolla M, Collette L, Blanl L, et al. Lancet. 2002;360:103-6.
16. Platz EA, Leitzmann MF, Visvanathan K, et al. J Natl Cancer Inst. 2006;98: 1819-25.
17. Etzioni R, Tsodikov A, Mariotto A, et al. Cancer Causes Control. 2008;19:175-81.
18. Schröder FH, Carter HB, Wolters T, et al. Eur Urol. 2008;53:468-77.
19. Schröder FH, Bangma CH, Roobol MJ. Eur Urol. 2008;53:901-8.