Tudor Borza, MD, Ramdev Konijeti, MD and Adam S. Kibel, MD∗akibel@partners.org,     Department of Surgery, Division of Urology, Brigham and Women’s Hospital, 45 Francis Street, Boston, MA 02115, USA

∗Corresponding author.

Prostate cancer diagnosis and treatment rates have increased significantly since the introduction of prostate-specific antigen (PSA) screening. Although it was initially thought that most prostate cancers would lead to death or significant morbidity, recent randomized trials have demonstrated that many patients with screening-detected cancer will not die of their disease. Modifications to PSA screening, screening guideline statements, and novel screening markers have been developed to minimize the risk and morbidity associated with overdiagnosis and overtreatment. Less aggressive management strategies such as active surveillance may lead to lower treatment rates in men who are unlikely to benefit while maintaining cure rates.

Keywords

Prostate cancer • PSA screening • Early detection • Overdiagnosis • Active surveillance

Key points

Introduction

Prostate cancer is the most commonly diagnosed cancer in men with an expected 238 600 new cases and 28 000 deaths in 2013 accounting for the second leading cause of cancer death in US men.1 Over the past 20 years, a clinical focus has been on early detection using prostate-specific antigen (PSA) screening to find cancer before it becomes incurable. The result has been a dramatic increase in diagnosis accompanied by a decrease in mortality of almost 3.7% per year since 1994.1 Recently, however, the utility of PSA screening has come into question because of an increase in diagnosis of cancers that are unlikely to cause significant harm to patients during their lifetime. Although controversy exists as to the effect of PSA screening on mortality, what is unquestioned is that a relatively large number of patients must be screened in order to prevent one cancer-related death.2,3 As a result, efforts are focused on screening men most likely to benefit, development of new tests designed to aid in screening, and limiting treatment in men at low risk for dying of cancer. Herein, the authors review recent data on PSA screening, the rationale for new recommendations, novel screening tests, and the role of active surveillance in the management of this disease.

PSA screening

PSA was introduced in the late 1980s as a screening tool for prostate cancer.4 The initial recommendations were that all men older than 50 years undergo yearly screening with PSA and digital rectal examination (DRE), and screening should begin at 40 years of age for men at a higher risk. This screening paradigm led to a dramatic increase in the number of cases from about 100 per 100 000 US men in the pre-PSA era to approximately 160 per 100 000 US men subsequently.1 In addition, a dramatic stage migration has been noted. Currently, 85% of prostate cancers diagnosed in the United States are clinically localized. Before PSA testing, as many as two-thirds of patients with apparently clinically localized disease were found to have pathologically advanced disease at the time of prostatectomy5 and approximately one-third were found to have lymph node metastatic disease.6 The number of patients presenting with locally advanced disease (stage T3 or T4) has decreased from 19.2% in the pre-PSA era to only 4.4% a decade later.7 The rate of metastatic disease at presentation has also been greatly affected by PSA screening. Scosyrev and colleagues8 performed an analysis of Surveillance Epidemiology and End Results (SEER) data and demonstrated a 3-fold decrease in the age-specific and race-specific observed annual incidence rates of metastatic prostate cancer at diagnosis in 2008 compared with the expected rates based on the incidence between 1983 and 1985.

Although initially it was thought that most of the tumors diagnosed by PSA screening had the potential to cause harm, there has been a realization over time that many of these tumors were similar to those detected in autopsy series and thought to be clinically insignificant. Incidental tumors were first appreciated in autopsy studies that demonstrated histologic evidence of prostate cancer in 1 in 3 men older than 50 years, with up to 80% of these tumors being clinically insignificant because of their limited size and grade.9–12 These studies illustrated the high latent disease prevalence, thus, allowing for potential overdiagnosis by screening tests.

Selecting a threshold for normal PSA has also proven difficult because significant cancer has been shown to occur even with low PSA values. The Prostate Cancer Prevention Trial (PCPT) randomized men with PSA of less than 4 ng/mL to receiving finasteride or placebo to investigate whether finasteride prevented prostate cancer.13 The study incorporated an end-of-study biopsy for all patients, including men on placebo, which allowed a more full understanding of prostate cancer risk in men with a normal PSA. Cancer was identified at the end-of-study biopsy in 15.2% of men with normal annual PSA and DRE over a 7-year interval. These tumors were not all clinically indolent because Gleason 7 or higher cancer was found in 2.3% of these patients. In men with a PSA between 2.1 and 4.0 ng/mL, 24.7% had prostate cancer and 5.2% had Gleason 7 or higher cancer. Even a PSA cutoff of 1.1 ng/mL would miss cancers because 25% of all cancers and 18% of Gleason 7 to 10 cancers were found in patients with PSA of less than this value.14 It is clear from this data that there is no level of PSA that can be thought of as truly normal or abnormal.

Clinical trial data on the utility of PSA screening also have mixed results. Despite the decreases in prostate cancer mortality and rates of advanced disease at presentation1,5–8 during the era of PSA screening and aggressive therapy, data emerged demonstrating that patients with localized prostate cancer at diagnosis had high non–prostate cancer–specific mortality if left untreated, raising the possibility that treatment in these patients would not have provided a benefit.15 These seemingly contradictory findings highlighted the clear need for prospective, randomized controlled trials investigating the utility of PSA screening.

Randomized controlled trials of PSA screening

Although the debate on PSA screening has continued for a decade or more, the recent publication of randomized trials has brought this discussion into the public eye. The Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial enrolled 76 693 men between 55 and 74 years of age and randomized them to annual PSA screening for 6 years and DRE for 4 years or usual care as the control group.3 The study demonstrated no significant difference in prostate cancer mortality between the intervention cohort (screened) and the control cohort (usual care) at the 13-year follow-up. There were 4250 cancers detected it the screening group compared with 3815 cancers in the control group, resulting in a relative increase of 12% (relative risk [RR] 1.12, 95% confidence interval [CI] 1.07–1.17). A total of 158 prostate cancer deaths occurred in the screened arm and 145 deaths in the control arm, corresponding to a statistically insignificant difference in cumulative prostate cancer mortality rates of 3.7 and 3.4 deaths per 10 000 person-years, respectively (RR 1.09, 95% CI 0.87–1.36). There were some flaws in the PLCO trial. There was significant screening before randomization, with 45% of the total control cohort having a PSA before enrollment. Contamination was also a significant problem, with 52% in the control group undergoing at least one PSA test during the study. In addition, 15% of men in the screening group did not undergo PSA testing during the trial. As a result, the study lost power. Lastly, the study was designed to test PSA screening in the community. PSA was obtained; the results were sent to the primary care physician; and recommendations were made. However, neither biopsy nor treatment was mandated. As a...