Prostate cancer continues to challenge clinicians and epidemiologists alike. Why do some prostate cancers become clinically significant whereas others remain quiescent? Which prostate cancers are truly localized and which are systemic? What is the optimal method of managing this disease? Three decades ago, these questions were academic because most men presented with clinical signs and symptoms of systemic disease and received palliative treatment with androgen deprivation therapy. The introduction of prostate-specific antigen (PSA) testing dramatically changed our understanding of this disease, but it has also raised many questions and concerns.

A review of worldwide cancer statistics demonstrates that the incidence of prostate cancer has increased substantially in most Western countries and this increase coincides with the introduction of PSA testing.1 The largest increase in prostate cancer incidence has occurred in the United States, where prostate cancer rates more than tripled between 1985 and 1990 and then declined after 1993, stabilizing at a value that is nearly twice the rate observed before 1985. The majority of epidemiologists are confident that PSA testing has driven these changes because they have not been observed in countries such as the United Kingdom, where PSA testing is still relatively uncommon when compared with North America.2 Epidemiologists monitoring cancer statistics believe that PSA testing has impacted men residing in countries in which PSA testing is prevalent in 3 primary ways: 1) they are now diagnosed with prostate cancer approximately 5 to 10 years earlier (lead time); 2) they are now primarily diagnosed with localized disease as opposed to disseminated disease (stage shift); and 3) they are much more likely to be diagnosed with latent or indolent disease (overdiagnosis) when compared with the pre-PSA era.

A much more difficult problem to resolve is how many of the additional cancers identified by PSA testing are clinically significant. Pathologists have long recognized that many men harbor small foci of prostate cancer and never experience clinical symptoms.3 The finasteride chemoprevention trial provided dramatic evidence of the large reservoir of indolent disease.4 When this trial was designed, clinicians and statisticians estimated an event rate of approximately 6% during the course of the trial. Researchers powered the study to detect a 50% reduction in incidence rates, anticipating that the men in the treatment arm might achieve a prostate cancer incidence rate of ≤ 3%. After performing biopsies in as many men as possible, researchers were astounded to discover that prostate cancer incidence rates were > 20% in both arms of the trial. A substantial number of the tumors identified in men with essentially normal PSA values were Gleason 3 + 3 tumors.5 This study, along with others, suggests that the natural history of tumors identified by PSA testing is much more protracted than previously anticipated.

Epidemiologists have also observed that prostate cancer mortality rates have declined substantially since 1990.1 This decline also coincides with the introduction of PSA testing. Unfortunately, clinicians and researchers cannot agree on how much of this decline can be directly attributed to PSA testing. Clinicians who strongly support PSA testing usually cite the decline in prostate cancer mortality as proof that PSA screening works. Clinicians who question the efficacy of PSA testing usually argue that the observed decline has occurred much too soon after the introduction of PSA testing to claim a cause-and-effect relationship. Deciphering the true impact of PSA testing and subsequent prostate cancer treatment is a complex exercise that requires information concerning 2 key variables: 1) what is the natural history of screen (PSA)-detected prostate cancer and 2) how do specific treatments such as radical prostatectomy or radiotherapy alter the natural history of this disease. Ideally, information concerning these outcomes should come from randomized trials. Unfortunately, to the best of my knowledge, very little data are available.

In the absence of such data, researchers often turn to computer models to explore the relative impact of screening and treatment. Two decades ago, we published a Markov analysis that explored the relative impact of conservative management, radical prostatectomy, and radiotherapy on 10-year clinical outcomes.6 Based on data available at the time, we concluded that radical interventions had a relatively modest impact on 10-year survival and prostate cancer mortality rates. The primary insight provided by the model was to highlight the paucity of data available concerning the natural history of this disease and the relative impact of treatment. Although there have been significant contributions since this early model, many issues raised in a 2003 editorial remain unresolved.7

During the past decade, researchers working under contract from the National Cancer Institute (NCI) have developed models of prostate cancer progression.8 Three teams, one from the Fred Hutchinson Cancer Institute in Seattle, one from the University of Michigan, and one from Erasmus University in Rotterdam, the Netherlands have independently explored this problem using data collected from the Surveillance, Epidemiology, and End Results(SEER) program of the NCI. Using the computer simulation models they developed, they have explored the changing prostate cancer incidence rates, initial staging, and prostate cancer mortality rates using information derived from multiple sources. Computer simulation models provide researchers with considerable power to explore large-scale epidemiologic trends. Researchers often vary key values such as estimates of disease incidence and mortality rates along with estimates of the relative impact of competing therapies. The purpose of these sensitivity analyses is to understand which variables truly impact survival and which are relatively inconsequential. In our original Markov simulation, for example, we demonstrated that despite the concerns of many prominent clinicians, estimates of prostate cancer mortality associated with treatment had very little impact on observed prostate cancer mortality rates.6

In their article published in this issue of Cancer, Etzioni et al performed several sensitivity analyses and arrived at some remarkable conclusions.9 First, they demonstrated that primary treatment with surgery or radiotherapy explains only approximately 25% to 33% of the decline in prostate cancer mortality noted in the United States over the past 20 years. Second, they have concluded that PSA testing is most likely responsible for some of the decline in prostate cancer mortality, but not by the mechanism assumed by most clinicians. Their model supports the concept that PSA testing appears to identify prostate cancer much earlier in the natural course of this disease, thereby allowing for the treatment of localized disease. However, the decline in mortality appears to result from the combination of both primary treatment and the treatment of recurrent and progressive disease rather than the impact of primary treatment alone. In other words, screening appears to identify at least 2 populations of patients: 1) a subset of patients with indolent disease of no clinical significance and 2) a subset of patients who have clinically significant disease and who are often treated with multiple interventions that collectively provide effective control of prostate cancer in a substantial number of cases. This is a profound insight. It suggests that prostate cancer is a chronic disease that often progresses slowly and relentlessly. In many instances, initial treatments do not cure this disease but rather prolong survival by slowing the progression of clinically significant cancers.

This insight carries significant practical implications. First, it refutes the notion that PSA screening is worthless. Many men do not benefit from PSA testing, but those with sufficient longevity appear to benefit from having treatment initiated early in the course of their disease.10 Second, it provides strong support for current efforts to prevent the overdiagnosis of this disease. Because many men diagnosed with prostate cancer as a result of PSA testing are more likely to experience side effects from surgery and radiotherapy rather than from their disease, it is important to avoid labeling these patients with a diagnosis of prostate cancer. One potential way to resolve this problem is to alter the current diagnostic pathway. Rather than subjecting men with elevated PSA values to transrectal ultrasound-guided biopsies, perhaps we should consider some type of interim testing or imaging first. Contemporary magnetic resonance imaging (MRI) studies or ultrasound/MRI fusion techniques offer the possibility of identifying clinically significant lesions that could lead to targeted biopsies rather than the random techniques currently used.11 Men with negative studies would simply be observed rather than proceed to biopsy. Another possibility could be a downstream confirmatory test such as prostate cancer antigen 3 (PCA3). Although this particular test has not proven to be sufficiently sensitive or specific, another serum marker or genetic test may become available in the near future.12

The computer simulation models developed by Etzioni et al9 suggest that clinically significant prostate cancers most likely require multiple therapies in many cases. Most surgeons have been disappointed to see rising PSA values among patients they believed were cured of their disease. Despite negative surgical margins and negative lymph nodes, many patients still demonstrate evidence of residual disease, suggesting that clinically significant prostate cancer is a systemic disease much earlier than commonly appreciated. These patients often receive radiotherapy accompanied by antiandrogen treatment. Fortunately, many survive several additional years before their PSA begins to rise again. Radiation therapists are also frequently disappointed to learn that PSA levels can rise many years after radiotherapy has been completed, thereby necessitating additional androgen deprivation therapy at a later date. The computer models suggest that prostate cancer mortality is lowered by delaying disease progression for a sufficiently long period of time to allow patients to die of other diseases.

Finally, and most intriguing, the computer models have suggested that additional factors beyond surgery, radiotherapy, and other traditional treatments are having a significant effect. Even under the most optimistic scenarios, prostate cancer treatments explain less than one-half of the observed decline in prostate cancer mortality. What other factors could be impacting this disease? Epidemiologists have long recognized that environmental influences can have a profound impact on disease incidence and progression. Most clinicians and patients now recognize that the dramatic rise and subsequent decline in lung cancer incidence and mortality can be linked to tobacco abuse. Similarly, the dramatic decline in gastric cancer appears to be related to the introduction of refrigeration as a method of food preservation rather than screening and treatment. Who could have guessed that salting and smoking could have had such a deleterious effect? What environmental factors could be impacting prostate cancer progression? What is happening in virtually all Western countries, and especially in the United States, to explain the declining rate of prostate cancer mortality?1 Has the aggressive treatment of heart disease or the control of hypertension had the unanticipated effect of lowering the impact of clinically significant prostate cancer? Could obesity or diabetes have an effect on the development of this disease? Could common drugs such as statins, metformin, or warfarin serve as chemopreventive protective agents? Could vitamin D deficiency play a role? All these theories and more deserve increased scrutiny as a result of the insights provided by the computer models developed by Etzioni et al.9

Fortunately, several randomized trials are currently well underway that promise to provide additional insights into the “prostate cancer conundrum.” These trials provide important data demonstrating cause-and-effect relationships. Long-term follow-up of the European Randomized Study of Screening for Prostate Cancer (ERSPC) trial and the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial should provide important information concerning the natural history of screen-detected disease.13, 14 The Comparison Arm for ProtecT (CAP) trial currently underway in the United Kingdom will also add to this growing pool of data.15 Information from the Prostate Cancer Intervention versus Observation Study (PIVOT) trial conducted in the United States and the ProtecT (Prostate testing for cancer and Treatment) trial conducted in the United Kingdom will provide much need data concerning the relative impact of surgery and radiotherapy.16, 17 These additional data will allow Etzioni et al9 to further refine their computer simulations.

We may never completely resolve the “prostate cancer conundrum,” but the combined efforts of dedicated clinicians, indefatigable trialists, and imaginative epidemiologists should guide us toward a more coherent and appropriate treatment of this prevalent cancer. Prostate cancer is now the most frequently diagnosed cancer in the Western world. Governments and other health funding agencies are struggling with the costs associated with treating patients with this disease. We can no longer afford to use a shotgun approach to treatment. Instead, we need to adopt screening protocols that target those men who are likely to develop clinically significant disease and treatment pathways that achieve the desired goal of lowering prostate cancer mortality at the lowest possible cost both in terms of morbidity and monetary expenditures. Computer models such as those developed by Etzioni et al9 offer a superb method of guiding these research efforts.


  1. Top of page

Supported by the University of Connecticut.


The author made no disclosures.


  1. Top of page
  • 1
    Center MM, Jemal A, Lortet-Tieulent J, et al. International variation in prostate cancer incidence and mortality rates. Eur Urol. 2012; 61: 1079-1092.
  • 2
    Hussain S, Gunnell D, Donovan J, et al. Secular trends in prostate cancer mortality, incidence and treatment: England and Wales, 1975-2004. BJU Int. 2008; 101: 547-555.
  • 3
    Sakr WA, Haas GP, Cassin BF, Pontes JE, Crissman JD. The frequency of carcinoma and intraepithelial neoplasia of the prostate in young male patients. J Urol. 1993; 150( 2 pt 1): 379-385.
  • 4
    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.
  • 5
    Thompson IM, Pauler DK, Goodman PJ, et al. Prevalence of prostate cancer among men with a prostate-specific antigen level < or = 4.0 ng per milliliter. N Engl J Med. 2004; 350: 2239-2246.
  • 6
    Fleming C, Wasson JH, Albertsen PC, Barry MJ, Wennberg JE. A decision analysis of alternative treatment strategies for clinically localized prostate cancer. Prostate Patient Outcomes Research Team. JAMA. 1993; 269: 2650-2658.
  • 7
    Albertsen PC. The prostate cancer conundrum. J Natl Cancer Inst. 2003; 95: 930-931.
  • 8
    National Cancer Institute Cancer Intervention and Surveillance Modeling Network (CISNET). Prostate Cancer Model Profiles. Accessed June 13, 2012.
  • 9
    Etzioni R, Gulati R, Tsodikov A, et al. The prostate cancer conundrum revisited: treatment changes and prostate cancer mortality declines. Cancer. 2012; 118: 5955-5963.
  • 10
    Moyer VA; on behalf of the U.S. Preventive Services Task Force. Screening for Prostate Cancer: U.S. Preventive Services Task Force Recommendation Statement [published online ahead of print May 21, 2012]. Ann Intern Med.
  • 11
    Sciarra A, Panebianco V, Salciccia S, et al. Value of magnetic resonance spectroscopy (MSR) and dynamic contrast-enhanced magnetic resonance (DCEMR) imaging for the characterization of high-grade prostatic intraepithelial neoplasia (HGPIN) foci. Urol Oncol. 2011; 29: 634-640.
  • 12
    Roobol MJ, Schroeder FH, van Leeuwen P, et al. Performance of the prostate cancer antigen 3 (PCA3) gene and prostate-specific antigen in prescreened men: exploring the value of PCA3 for a first-line diagnostic test. Eur Urol. 2010; 58: 475-481.
  • 13
    Schroder FH, Hugosson J, Roobol MJ, et al; ERSPC Investigators. Prostate-cancer mortality at 11 years of follow-up. N Engl J Med. 2012; 366: 981-990.
  • 14
    Andriole GL, Crawford ED, Grubb RL3rd, et al; PLCO Project Team. Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: mortality results after 13 years of follow-up. J Natl Cancer Inst. 2012; 104: 125-132.
  • 15
    Lane JA, Hamdy FC, Martin RM, Turner EL, Neal DE, Donovan JL. Latest results from the UK trials evaluating prostate cancer screening and treatment: the CAP and ProtecT studies. Eur J Cancer. 2010; 46: 3095-3101.
  • 16
    Wilt TJ. The Prostate Cancer Intervention Versus Observation Trial (PIVOT) in NIH State of the Science Conference: Role of Active Surveillance in the Management of Men with Localized Prostate Cancer. National Institutes of Health, December 5-7, 2011;105106.
  • 17
    Donovan J, Hamdy F, Neal D, et al. ProtecT Study Group. Prostate Testing for Cancer and Treatment (ProtecT) feasibility study. Health Technol Assess 2003;7:188.