Millions of men over age 50 rely on the prostate-specific antigen, or PSA, test each year in screenings for prostate cancer. In October Stanford University urologist Thomas A. Stamey made headlines when he declared that the test is not a reliable predictor of cancer.
Back in 1987, Stamey was among the first to suggest that the level of PSA, a protein normally produced by the prostate gland, might be useful in detecting prostate cancer. But based on an analysis of more than 1,300 prostates removed over the past 20 years, Stamey reported in the October issue of the Journal of Urology that the PSA test is currently predictive of cancer in only 2 percent of cases. Because of the increase in screening and detection of prostate cancer over the past two decades, he now says a higher PSA level may most often reflect a harmless age-related increase in prostate size.
When doctors follow up a high PSA level with a biopsy, they often find cancer. But this is only because most men have some degree of prostate cancer. Studies have shown that 80 percent or more of men over age 70 die with—but not from—prostate cancer. As counterintuitive as it seems, detecting prostate cancer is not always in the patient’s best interest. Once cancer is diagnosed, most men opt for treating it either with radiation or removal of the prostate. In many cases, that leads to impotence, urinary incontinence, and other unpleasant side effects.
Although some experts question Stamey’s interpretation of the data, his skepticism about the meaning of PSA levels has significant support. “First of all, it is not known how often PSA testing saves lives,” says Howard Parnes, an oncologist at the National Cancer Institute. Randomized clinical trials have shown, for instance, that screening for breast cancer saves lives, but similar studies for the PSA test haven’t been completed.
It seems likely that PSA testing leads to many more people getting a cancer diagnosis and treatment than is necessary, says Parnes. What would be more useful, he says, is a diagnostic tool—such as a protein marker for prostate cancer—that could determine who among those people actually needs to be treated. For now, some experts advise doctors to notify patients of the low death rate from prostate cancer, explain the risks of treatment, and allow them to make an informed decision about testing. Physicians can also perform yearly rectal exams and practice watchful waiting rather than immediate intervention.
The American Urological Association, however, recommends that the PSA test be used in conjunction with a rectal exam and a full medical history to determine if a biopsy is warranted. And some clinicians argue that a sharp increase in PSA from one year to the next can be a significant clue. Still, without definitive proof of the PSA test’s worth, the National Cancer Institute, the American Cancer Society, and others say only that the test should be “offered”; in contrast, they recommend that women “should be screened” for breast and cervical cancer. “We feel there isn’t enough information to recommend for or against,” says Parnes. “We acknowledge that we don’t know.”