Prostate Cancer Screening with Prostate-Specific Antigen Testing: More Answers or More Confusion?

Prostate cancer is a leading cause of morbidity and mortality among middle-aged and older men. Of the solid tumors prostate cancer is rather unique in that it presents in 2 distinct forms, a latent form, which grows slowly and poses no threat to the patient’s life, and an aggressive form, which metastasizes quickly and kills the patient. The discovery of prostate-specific antigen (PSA)2 and the demonstration of its utility for early diagnosis and monitoring of prostatic carcinoma have raised hopes that this simple serological test could be invaluable in screening asymptomatic individuals for early prostate cancer diagnosis. The premise is that such early diagnosis may then lead to early therapeutic interventions, which should improve the overall survival of prostate cancer patients. However, PSA screening of asymptomatic individuals has remained controversial during the last 15 years owing to the lack of evidence for improved patient survival. Recently, the results of 2 major randomized clinical trials on the effectiveness of PSA as a screening tool, from both the US and Europe, have been published. These results are not clear cut. For this reason, the controversy surrounding prostate cancer screening will likely continue for years. Below, we examine this issue with 4 authorities in the field. Patrick Walsh3 : Undisputed evidence for benefit! There is no debate that PSA testing has made it possible to diagnose prostate cancer at an earlier, curable stage. Because prostate cancer produces no symptoms until it is far advanced, before PSA testing most patients presented with incurable disease—either locally advanced or metastatic. However, following the advent of PSA testing, suddenly it became possible to diagnose prostate cancer at an earlier curable stage. According to data from the American Cancer Society, in 1990 only 68% of men presented with localized disease and 20.6% had metastatic disease. In 2009, 91% of patients presented with localized prostate cancer, only 4% had metastatic disease, and deaths from prostate cancer between 1996 and 2006 fell from 41 400 to 27 350. To answer the question more succinctly, PSA testing has given men a choice that they did not have before it was available: men can either undergo testing and if they have cancer choose treatment or observation or they can do nothing and run the risk of a diagnosis when it’s too late to cure. Klaus Jung4 : PSA has become the most popular tumor marker ever because of its relationship with tumor stage and because of the increasing risk of developing prostate cancer with increasing PSA, as demonstrated on the basis of PSA results obtained many years before diagnosis. None of the biomarkers for other cancers has been shown to be comparably effective in detecting a carcinoma at a stage early enough for curable treatment that would reduce the prevalence of primary metastatic carcinoma. Despite all problems resulting from noncancer-related biological factors affecting PSA concentrations (e.g., intra- and interindividual biological variability of PSA, age, prostate volume, possible inflammation) as well as analytical issues like insufficient interchangeability of PSA obtained by different commercial PSA assays, there is currently no alternative to PSA and for the urologist this marker will remain the early biochemical signal of prostate cancer in the near future. William Catalona5 : The short answer is, “Because it works.” Patients and physicians have relied upon the test because it gives them valuable information that they want to have, and death rates have continued to fall during the PSA era. Clinicians have learned from their own experience that, although the PSA test is not perfect, it is effective in identifying men at high risk for prostate cancer and for detecting it early. Approximately 10% of men over 50 have a PSA >4 μg/L, and approximately 20% have a PSA >2.5 μg/L. These men are at far greater risk for prostate cancer than the approximately 80% with lower PSA levels. Moreover, there is also a strong correlation between PSA and the aggressive form of the disease. Neil Fleshner6 : PSA screening continues despite conflicting evidence for its survival benefit. There are a variety of factors driving this. First, it is important to recall the landscape with respect to prostate cancer detection before the PSA era. In the late seventies and early eighties, a high proportion of patients with prostate cancer detected by digital rectal examination alone had a high probability of node-positive disease in addition to having a greater chance of extraprostatic extension. The second reason for ubiquitous PSA screening relates to the favorable early and midterm outcomes reported among patients who have had PSA-screened tumors. The rate of organ-confined disease as well as improved functional outcomes due to the nature of nerve-sparing surgery, particularly among patients with early stage disease, has improved outcome. Patrick Walsh: The introduction of any new form of treatment for localized prostate cancer requires at least 10 years before a major effect on mortality is seen. The dramatic decline in mortality since 1992 is the result of 2 events—earlier diagnosis at a curable stage and effective therapy. In 1982, only 7% of men with localized prostate cancer underwent surgery and radiation therapy was too underpowered to cure. Almost no one received treatment with curative intent. However, a decade later, when more men with curable disease were identified and the side effects of radical prostatectomy were reduced, more than 100 000 men underwent surgery. If one applies the findings of the Scandinavian Prostate Cancer Group’s randomized trial of surgery vs watchful waiting to these men, 7500 to 15 000 fewer men per year should be dying of the disease or suffering from painful metastasis. This, coupled with improvements in radiotherapy, are the reasons why fewer men are dying of the disease. Klaus Jung: This is too complex a question to be answered with a single reason. Following the introduction of PSA screening at the end of the eighties, there has been a continuous decline of the prostate cancer death rate in ensuing years. The influence of PSA screening on the decreased mortality rate seems to be more than likely, but it has not been proven. There are arguments that the decrease of prostate cancer mortality was observed too early after the introduction of PSA screening for prostate cancer, a slow-growing carcinoma with a long biological history. Thus, the mortality decline is probably a combined result of the more frequent detection of prostate cancer at an early, curable stage by PSA screening and the improved treatment modalities such as surgery, radiation, and hormonal therapies during the later years. For example, for patients who have a higher risk of recurrence, there is an improved postoperative strategy with early adjuvant radiation and prolonged hormonal therapy. Radiation therapy now uses a higher dose (approximately 80 Gy) to cure prostate cancer in comparison to earlier years. In addition, successful options for advanced stage patients with asymptomatic metastases (zoledronic acid, chemotherapy with docetaxel) improve survival. However, it also should be mentioned that death misclassification bias and the use of statins have been discussed as contributing to the decline of the prostate cancer mortality rate. William Catalona: It is mainly because of PSA screening. Of course, other factors come into play, such as improved treatments and their earlier application. Nevertheless, it has been estimated that 40% to75% of the reduction is due to PSA screening. One cannot completely dissociate the effects of screening from treatment. Early detection would be useless without effective treatment. And curative treatments are effective mainly in patients with early disease. However, the most effective curative treatment, radical prostatectomy, was available before the PSA era. Therefore, PSA screening is the most important factor responsible for the falling mortality rate. Neil Fleshner: It is of course impossible to know for sure the impact of PSA screening on mortality rates. However, if one looks at the early European Randomized Study of Screening for Prostate Cancer (ERSPC) trial results there appears to be a benefit in mortality in prostate cancer noted as early as 8–10 years after instituting a screening period. Furthermore, we know from the Swedish randomized trial of prostatectomy vs watchful waiting, that mortality rates start to split similarly at approximately 7–9 years. I think it is therefore quite plausible that screening or early detection has been in part responsible for the falling US mortality rates. I also believe that PSA as a marker in detecting early recurrence, and in particular with the use of early hormone therapy that arises from PSA testing, may have also improved prostate cancer mortality. Patrick Walsh: This question is easy to answer. One of the trials, the ERSPC, was well done and informative. The other trial, Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO), was poorly done and uninformative. The ERSPC was a landmark study carried out in 7 countries that enrolled 162 000 men with follow-up out to 14 years. In patients who were actually screened, there was a 27% decrease in prostate cancer deaths. The PLCO, which was half the size of the European Trial and had complete follow-up half as long (7 years), concluded that screening had no effect on prostate cancer mortality. This study contained numerous flaws. (a) The follow-up was too short. Screening and aggressive treatment are typically reserved for individuals with at least a 10-year life expectancy and any patient who dies within 7 years of diagnosis had incurable disease at the time of diagnosis and would have not benefited from screening. In the positive ERSPC trial there was also no improvement in survival before 10 years. (b) The trial did not test screening vs no screening but rather more screening vs less—85% of patients in the screened arm underwent PSA testing compared to 52% in the controls. (c) Forty-four percent of the men who entered the trial already had 1 or more PSA tests. Consequently, these men were not only less likely to have cancer, but also less likely to have life-threatening disease. This is why so few men died from cancer and why there was no decrease in advanced disease in the men who underwent screening. (d) Only 30% of the screened men who developed a PSA >4 ng/mL while in the trial actually underwent a biopsy. Indeed, if the authors had set out to design a study to discredit PSA testing it would have been difficult to do a better job. Klaus Jung: The different study designs are clearly responsible for the contradictory results (e.g., the PLCO trial with 73 000 men between 55–74 years used annual screening with a PSA decision threshold of 4 μg/L; the ERSPC trial with 162 000 men between 55–69 years applied a PSA cutpoint of 3 μg/L and mean screening interval of 4 years, but using different recruitment and randomization procedures between the participating countries). Both studies had limitations. The PLCO trial manifested distinct weaknesses of the study design, (a) with PSA prescreening activities in about 40% of the study participants before randomization, leading to the probable elimination of men with prostate cancer from the study groups; (b) the relatively small numbers of participants; (c) the high PSA contamination rate in the control arm, leading to a comparison of 2 different “screening ” cohorts with 85% and 52% PSA tests performed; and (d) the limited duration of follow-up of 7 years. Thus, it was not really surprising that a decrease of prostate cancer mortality as the benefit of screening could not been seen. Even if the patient number in the ERSPC study is higher but still relatively small, the first results show a benefit; with adjustment for nonattendance in the screening arm (defined as failure to attend the initial screening round) and PSA contamination in the control arm (defined as carrying out at least 1 PSA test after randomization) the reduction in mortality is already 31%. If these results would have been published before the PLCO data, the negative discussion concerning the effectiveness of PSA screening probably would have not taken William Catalona: The European trial (ERSPC) evidence that PSA screening can the US trial is on this Moreover, it is likely that the mortality benefit in ERSPC is an because of the relatively short the relatively long screening the of digital rectal as a screening test after the first and contamination in the control In it is that PLCO results will with follow-up because the study was from the The most than of participants were many men with prostate cancer from the study men had a lower prostate cancer mortality and the to a mortality than 52% of were screened during the This contamination the in prostate cancer mortality rates and also the of the to 1 PLCO had no for men with screening results to undergo and only approximately 40% of men with screening results underwent within 1 early detection and treatment. It patients a of years after biochemical following radical prostatectomy to of prostate However, the follow-up for men with cancer in PLCO of years was insufficient to mortality PLCO men to years who are less likely to have a mortality benefit from screening. In ERSPC 1 evidence that PSA screening prostate cancer mortality by at least 20% by after for with screening in the screening arm and contamination in the control with the authors about detecting tumors that the In PLCO is about the of screening using PSA μg/L PSA PSA for a for men who to findings and undergo treatment. Neil Fleshner: The 2 randomized trials of prostate cancer screening published in the of a not published in the data from a PLCO have shown that of 40% of patients with an PSA in the study a biopsy. to this the number of patients who received a PSA test in the control In this has to a study within the PLCO is the reason, in that this to a Therefore, the European study is more of the effect of PSA screening at an 8–10 years time Patrick Walsh: design, the ERSPC was to data when the threshold for was has been made of the that to 1 prostate cancer death at 10 years, men would to be screened and an men would to be However, because screening in a decrease in the number of men with incurable disease, with follow-up the number to will In the PLCO trial the threshold for results and there is that the trial was published at this time to the by the European Because screening in this study did not improve clinical with follow-up the results will not Klaus Jung: the of prostate cancer and its as well as with the of years as the from PSA screening detection until clinical diagnosis of cancer, it can be that the PLCO study with a follow-up of 7 years and the ERSPC study with a follow-up of years presented data results can only the effectiveness of PSA screening. in the PLCO study is for at least years for all However, the PLCO trial will ever show a with the 2 cohorts is because of the effects of prescreening before randomization and of PSA contamination in the control the ERSPC study will continue to patients and will in addition to the primary of like the PLCO trial, the effect of PSA screening on important of of life and results of the ERSPC have more to show an increasing PSA benefit. Because patients who with metastatic disease can years and it is difficult to an 4 years after of should be William Catalona: In the PLCO data were One question why this study was to in the as the European study that the first evidence that PSA screening The of PLCO the impact of the of PSA screening by ERSPC and among patients and physicians who did not the and of the 2 there will be from both ERSPC will show a greater mortality benefit and lower numbers to and to to 1 it is possible that PLCO show a mortality benefit for because of the in study design, they will not the of screening. Neil Fleshner: one the of prostate cancer I believe that the results were Prostate cancer is a disease to an important risk to patient mortality for between 10 and years after this in I believe that a of these studies will a answer about the of screening. I do not know when these will I do believe that data a mortality benefit particularly in the European study there has been contamination of the control with PSA I do believe the other is overall mortality from increasing patient and it is of course impossible to that will impact the survival Nevertheless, it is that the of screening will become and likely more with Patrick Walsh: is I patients that if they are the of who a to the or their for a and about dying from prostate cancer, they should not undergo PSA the other if they are a 55–69 who not want to from prostate cancer, the European trial evidence that PSA testing can their Klaus Jung: The data the in to PSA testing but only to patients. Thus, the before the should be to patients about the and of PSA screening and its possible with to possible treatment side effects and treatment This decision is with the reduction of mortality in the European experience has shown that patients with tumors as for in of or treatment due to the of disease The correlation of PSA with prostate cancer and the 80% reduction of patients with metastatic disease after PSA are arguments to PSA screening. William Catalona: I PSA screening to patients. I this because it would them with the of their risk for having prostate cancer and the chance of death from this disease. Neil Fleshner: Because the number to among patients at higher risk of disease is I would be more with patients who are at higher risk of prostate cancer such as men with positive or men of It is also possible that in the a variety of men with may be also better for screening. I would to men before screening that not all men with prostate cancer therapy. is now a for the of men and I would this to patients a I believe that the of into screening may be a to and detection and overall Patrick Walsh: I was I 3 of from prostate I many of PSA testing have actually a with metastases of the they that from to PSA screening and effective treatment have the rate of prostate cancer death in the US by I have a PSA ng/mL and have had PSA since and will do so until I years of Klaus Jung: I did this in the last years in to the The reason is that we do PSA and because there are few men in I have for The the with a complete PSA with different William Catalona: to risk for prostate Neil Fleshner: I have a strong of prostate I first PSA at and have continued with Patrick Walsh: better the time we are not to the and of cancer within the One when this is we will be to patients who are with tumors that are too small to and them more for Klaus Jung: I would 2 possible the and the into the of all factors and their I a of risk to we have used an in to This the possible risk of prostate cancer on the basis of age, prostate volume, and results of the digital rectal with and percent risk are on the more and more who want to use all options before this and new like prostate cancer 3 The screening may improve with new and with only which can be used on a commercial like or have a chance to improve screening for aggressive as in addition to PSA and percent the still have a too and are not available so PSA on an basis without to the with prostate or prostate we could the cancer detection rate. Thus, can be currently with of patients for while can be by of patients with prostate The will be by the of detecting only aggressive of prostate cancer by use of The is to tumors that would and if without treatment. The for better on new the of that all clinical and biochemical data are the The of prostate cancer screening would be most likely a of and and The of the clinical will be the of in and to in and as and In strong with can with their in analytical and issues as well as in data William Catalona: to there has not been a test that is to PSA as a screening There is that the PSA may be or that other test would be that would be more The use of such tests could but completely and Neil Fleshner: I believe biomarkers may a in the that with respect to PSA screening. However, this and data are to earlier, in with screening would be a of of patients who would not therapy for prostate is an for these men. This of PSA and prostate with only when disease become more such as an in or of The of has been well demonstrated in the in a variety of randomized trial is currently is that would be a for the of men with disease. strategy that may be as an alternative to the side effects of radical therapy is therapy. This using modalities such as or to prostate combined with or the this is to the of the prostate prostate cancer without the The is to and side effects while cancer The of these modalities to be Of course the nature of prostate cancer poses about this However, small of prostate cancer may have effect in a patient during their William Catalona: In the prostate cancer is the second leading cause of cancer death cancer in men the US as well as many other It arises and in a curable stage for a time before into an incurable stage. There is no for aggressive prostate cancer and no of it it has an advanced stage. Thus, to reduce suffering and death from prostate cancer, it be detected early, and the most effective of early detection is PSA screening. the diagnosis of cancer is the for treatment be by the of the tumor and of the patient. The ERSPC has that cure is possible when by a 20% reduction in prostate cancer mortality among screened men as reduction after for and this mortality benefit are data a 40% decrease in the US prostate cancer mortality rate and data from the a mortality in countries PSA screening is and or increasing rates it is The of screening is to cancers that could cause suffering and but screening may also cancers that would cause because of limited to between and most cancers are and therapy have as to physicians should be not to out the with the or treatment may be in patients with an or will the and will be to suffering and death from prostate and the detection of cancers that have the prostate and in curable are also important that have received less than the is on physicians to that patients treatment to cure rates and side of of and and of of of PSA, prostate-specific ERSPC, European Randomized Study of Screening for Prostate PLCO, Prostate, Lung, Colorectal and Ovarian Cancer Screening prostate cancer Patrick of Klaus of William of Prostate Cancer Cancer of Neil of of of authors they have to the of this and have the following 3 (a) to the and design, of data, or and of (b) or the for and (c) of the published of of all authors the of of of or None or None and and of The no in the design of choice of enrolled and of data, or or of and for and for on a of would like to for