Prostate cancer, the most common cancer in U.S. men and the second-most common cause of cancer deaths, often produces no symptoms until it is too far advanced to cure. That’s why in 1990, before there was PSA testing, only 68 percent of newly diagnosed men had cancer that was localized to the prostate and 21 percent had disease that had metastasized to the bone.
Today, 91 percent of patients are diagnosed with localized disease and only 4 percent have metastases. Because of this ability to diagnose cancer at a curable stage, coupled with improvements in both surgery and radiation therapy, deaths from prostate cancer in the United States have fallen 40 percent in the past 10 years, a decline that is greater than for any other cancer in men or women.
So, why is there a debate on whether men should undergo PSA tests? Critics argue that because the test is not specific for cancer—it measures the amount of a protein produced by the prostate, which may increase in benign conditions such as enlargement—many men undergo needless biopsies and worry.
Let’s compare the accuracy of PSA testing with mammography. If a man has an elevated PSA, his risk of having prostate cancer is 25 percent. If his wife has an abnormal mammogram, her risk of having breast cancer is 10percent to 25 percent. Does anyone question the value of mammography for screening?
Others in the field quote the inconclusive evidence from two large randomized trials published earlier this year that examined how PSA testing affects patient outcomes. One from Europe demonstrated that men who underwent PSA testing had a 27 percent reduction in prostate cancer deaths at 14 years, but another from the United States showed no improvement in mortality at seven years. What’s responsible for this disparity?
The European trial was twice as large, followed patients twice as long, and was carried out in a setting where it was possible to evaluate men who underwent testing versus no testing. In contrast, the U.S. study was far too small and was performed in a setting where half of the control group also underwent PSA testing. In other words, this study did not evaluate testing versus no testing but more testing versus a little less.
Furthermore, only 30 percent of the men in the screening arm who had an elevated PSA during the trial actually underwent a biopsy, and the seven-year follow-up was too short to draw conclusions about mortality.
Many critics believe that screening will diagnose cancer in men who are too old or too ill to live long enough to benefit or who have tumors too small to require aggressive treatment, and that all treatments have too many side effects to be worth it. But I also believe there are ways to overcome these objections. Men who are too old or too ill to live longer than 10 years should never have a PSA test. Patients older than 65 with small, well-differentiated tumors should be followed carefully to determine whether these tumors progress to the point where they require treatment, an approach known as active surveillance that is often used here at Johns Hopkins.
Men diagnosed with prostate cancer that requires treatment need to be well informed about options and seek out qualified physicians who can offer curative treatments with the fewest side effects.
Finally, it is incumbent upon physicians to continue to improve treatments to reduce their morbidity. One day, when it is possible to reduce treatment side effects to a minimum, the debate over PSA testing will end.
The take-home message is that today, men have a choice they didn't have 20 years ago: They can undergo PSA testing and, if they have cancer, choose curative treatment or observation, or they can do nothing and risk a diagnosis when it’s too late to cure.
If you are a person who doesn't wear a seat belt or go regularly to the dentist or the doctor for a checkup and are not worried about dying from prostate cancer, do not undergo PSA testing.
But if you are a healthy man age 55 to 69 who does not want to die from prostate cancer, there’s conclusive evidence that PSA testing could save your life.