A report in the Journal of the National Cancer Institute (Vol. 94, No. 981-990) said that while PSA screening detects most prostate cancer, as a screening test, PSA may be a bit too good at what it does.

The problem, according to a team of researchers from the Fred Hutchinson Cancer Research Center in Seattle and elsewhere, is that PSA also detects cancer that may not require treatment. In other words, widespread PSA screening leads to overdiagnosis of prostate cancer, they said.

For many people, the term overdiagnosis makes no sense when it is applied to cancer. All cancer, they reason, should be diagnosed and treated. As Cleveland Clinic urologist Eric A. Klein, MD, puts it "most Americans aren't comfortable with the idea of living with cancer."

But cancer experts say it isn't that cut-and-dried.

Ruth Etzioni, PhD, of the Fred Hutchinson Center and one of the authors of the study, explained that the risk for prostate cancer increases significantly with age, but it is unlikely that a prostate cancer that is "detected in a 75-year-old is going to progress to cause clinical symptoms during his lifetime."

In fact, many prostate cancers in older men actually cause no symptoms and don't decrease survival time. Without PSA screening, these slow growing cancers are often only detected during autopsy.

But if one of these non-aggressive cancers is detected by PSA screening, that detection can trigger a number of events, including unnecessary and possibly disabling treatment, said Etzioni.

Side effects from prostate cancer treatment include incontinence, erectile dysfunction, and discomfort from scarring. Thus, PSA screening can "over diagnose" this type of prostate cancer.

As the prostate enlarges, the level of PSA found in the blood increases. Because prostate glands enlarge as a man ages, PSA levels also increase with age. A "normal" level for a 45-year-old may be 1 ng/ml, but a healthy 75-year-old could have a normal PSA of 6.5 ng/ml.

Generally, urologists consider a PSA level above 4.0 ng/ml requires additional testing such as a biopsy.

Etzioni and her colleagues used a computer model to determine the likelihood that PSA screening will result in overdiagnosis of prostate cancer.

"We defined overdiagnosis as detection of a cancer that would not otherwise have been diagnosed during a patient's lifetime," she explained.

The model used an imagined population of 2 million men, aged 60 to 84 in 1988. The authors asked the computer to estimate how many of these men would be diagnosed with prostate cancer by PSA screening. It compared that estimate to a computer-generated estimate of the number of prostate cancer cases that would be diagnosed without the use of PSA screening.

To check the accuracy of this method, the researchers checked the computer-generated findings against actual prostate cancer data collected by the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) registry.

"We found that somewhere between 20% and 40% of prostate cancers are over-diagnosed," said Etzioni.

This is actually lower than the "overdiagnosis" rate reported by some other researchers, but Etzioni said the real take-home message is that "PSA is actually a very good test. Even using our highest estimate for overdiagnosis, this means that PSA diagnosed 60% of cancers in a timely way that led to real clinical benefit," she said.

Nevena Damjanov, MD, a medical oncologist at Fox Chase-Temple Cancer Center in Philadelphia, said most cancer experts know the limits of PSA screening, but she admits that this can often be confusing for patients.

"This is a great deal like the mammography debate: when should we stop doing mammograms? Does it make sense to do mammography or PSA screening on people in their 70s so that we can discover a disease that may never affect their daily lives?" she asked.

"In younger people certainly there is a survival benefit that comes with early detection, but this is not so clear when the patient is elderly," she said.

Klein, who heads the section of urologic oncology at the Cleveland Clinic Foundation's Urological Institute, said, "My take on this whole issue is that we probably are overdiagnosing or over-detecting some clinically insignificant prostate cancers.

"But the simple fact is that there is no available biologic test, no X-ray, no test at all that can tell us is any cancer is aggressive or not aggressive," he said. "We just don't know what cancers can safely go untreated."

Klein said the Cleveland Clinic "has no plans to change our current policy on PSA screening: we recommend annual testing beginning at age 50. High-risk men, meaning those with a family history or African-American men, should consider annual PSA screening beginning in their 40s."

The American Cancer Society estimates that more than 189,000 new cases of prostate cancer are diagnosed annually and about 30,000 men will die in 2002 from prostate cancer.