Skip to content
National Cancer Institute National Cancer Institute U.S. National Institutes of Health www.cancer.gov
Division of Cancer Prevention logo
Home Site Map Contact DCP
Programs & Resources

NCI DCP News Note

Prostate Cancer Screening Controversies: A summary of the presentation of Barry Kramer MD, MPH at the June 2012 American Society of Clinical Oncology Meeting

Date:   July 30, 2012

A basic tenet of cancer screening is that it is extremely difficult to make healthy people better off than they already are. An important corollary to that is: before putting large numbers of healthy people in harm's way with any medical intervention, including a screening test, we need strong evidence that it has a clear net benefit. So why is there a controversy in prostate cancer screening? I believe a core issue is the difficulty health professionals have in dealing with the frequent uncertainties associated with medical evidence. Sometimes the answer to a clinical question is not an absolute yes or no, but is far more nuanced. Too frequently, that nuance is lost in the area of cancer screening, since the most-often heard messages do not go beyond oversimplified sound bites.

In 1996, Daniel Wolpaw wrote that in the field of cancer screening "It is particularly important to be able to sort out what is known versus what makes sense…." It makes sense that if we find a small cancer, the patient must benefit. But the reality is that is sometimes true, but too frequently not.

It is easy to take mental shortcuts, so researchers encourage the use of a formal analytic framework to assess the benefits and harms of screening (see figure). The goal of prostate cancer screening is to improve actual health outcomes such as less prostate cancer mortality, less overall mortality, or better quality of life. And the most direct (and reliable) means to assess the health outcomes, both good and bad, is a randomized controlled trial.

Intermediate outcomes shown in the figure, such as detection of more cancers at an early stage, are notoriously misleading. For example, screening can detect very slow growing, or even static, tumors that are not lethal (a phenomenon known as "over-diagnosis"); but it can also miss the fastest growing, life-threatening cancers. This could artifactually produce better survival and "cure" rates without decreasing the actual risk of dying from the cancer.

This flowchart shows the Analytic Framework for Cancer Screening. An arrow depicting a direct pathway shows healthy persons at risk progressing from screening to ultimate health outcomes. The pathway progression starts with screening and its possible adverse effects. The next step is early cancer detection. The third step represents possible intermediate outcomes, such as surgery for cure and decreased late stage disease, and the possible adverse effects of treatment. The fourth and final point on the pathway is health outcomes related to cancer mortality, overall mortality, and quality of life.


There is no doubt that the burgeoning use of the prostate specific antigen (PSA) test has increased the number of men diagnosed with prostate cancer. Beginning around 1988, incidence for prostate cancer jumped rapidly for men in the U.S., due directly to increased use of the PSA test. But in the same time period, prostate cancer death rates were more stable. Most of those additional prostate cancer cancers identified did not seem to change how often men were dying of the disease. (The relatively smaller decreases in mortality may have been due to the unquestionable improvements in therapy over the same era.)

And thus we turn to the only two randomized controlled trials of prostate cancer screening: the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) and the European Randomized Study of Prostate Cancer (ERSPC).

In the PLCO men in the study arm were screened for 6 years and followed for an additional 7 years. Men in the control arm received the usual care in their community, which could include PSA. After 13 years, the cumulative death rate from prostate cancer between these two groups of men is essentially identical.

In the ERSPC trial, multiple countries created screening programs from which data were pooled. The designs among the screening programs differed in randomization methods, screening intervals, biopsy thresholds, screening tests, and start dates. Not all screening centers participating in the study have reported their results yet. However, unlike the PLCO trial, ERSPC has shown a reduction in prostate cancer mortality in the screening group after a median follow-up of 11 years.

So at this point, the benefits of prostate cancer screening are uncertain or certainly modest. However, there are some very real harms. The US Preventive Health Services Task Force summarized the benefits and harms of PSA screening this way: While perhaps 0 to 1 man in 1,000 will not die from prostate cancer because he got PSA screening regularly over a decade, about 270 men will have some kind of harm, ranging from needing a biopsy (100-200 false positive tests leading to biopsy), cardiovascular events (2), sexual impotence (29), urinary incontinence (18), and even death due to the prostate cancer treatment (<1).

Sir Muir Gray summed up the issue of medical screening this way, "All screening programmes do harm; some do good as well, and of these, some do more good than harm at reasonable cost." At this point, the evidence does not clearly show that PSA screening does more good than harm.


Back to top