|Ian Thompson, M.D.|
Professor and Chair of Urology, University of Texas Health Science Center, San Antonio, Texas
|Right now our only option for addressing prostate cancer is to find and treat it early. We have neither a cure nor new advances in managing metastatic disease. And as of yet, we have no evidence that we can prevent prostate cancer, although a long-term study is under way.
Because many men develop slow-growing, well-differentiated tumors, some experts feared PSA screening would lead to needless treatments. But about 95% of tumors diagnosed after a positive PSA test are moderately or poorly differentiated and raise the risk of mortality.
In early tests, men often had PSAs of 15, 20, or 35. A PSA over 10 has a 60%-plus likelihood of being outside the prostate.
Today, most men have had a series of PSA exams, so when they convert from negative to positive PSA, the disease is confined in 95% of cases. Very conservatively, we can say the cancer never recurs in 70% of these men after treatment.
Even without therapy, it takes about a decade for the cancer to metastasize and an average five years more before the man dies of the disease. So even the 30% of patients who fail treatment still have about a 15-year survival. Bottom line: with annual PSA screening, most patients will not die of prostate cancer.
Around 1994, some five years after PSA screening began, we saw a fall in prostate cancer-specific mortality. In Europe and Latin America, where PSA screening is used less, mortality is going up.
Until we get results from the Prostate, Lung, Colorectal, Ovarian (PLCO) study, this is our best evidence supporting screening.
The evidence for PSA testing is at least as strong as that for cervical-cancer screening.
|Peter Albertsen, M.D.|
Professor and Chief of Urology, University of Connecticut Health Center, Farmington, Conn.
|A proven screening test meets five basic criteria. We all agree on the first two: prostate cancer is a significant health hazard, and we have a test to find the disease while it is still localized.
Criterion three, sensitivity and specificity, is problematic. The PSA test is extraordinarily sensitive; its many false positives lead to unnecessary ultrasounds and biopsies. At best, only a third of men with elevated PSA have prostate cancer. And with an estimated 80% specificity, screening will miss one cancer in five men with a normal PSA.
The real problems, however, lie in criteria four and five. Prostate cancer's slow growth makes assessing and comparing treatments difficult. Several studies show that men with low-grade disease are not likely to die from their cancer, so treatment won't significantly reduce mortality.
Men with high-grade tumors have a high risk of dying from cancer, but we don't have good estimates of the relative efficacy of surgery, radiation, or brachytherapy to cure these patients. Many will have a recurrence.
Finally, criterion five demands that screening improves outcomes. The recent decline in U.S. prostate-cancer mortality rates could be ascribed to attribution bias, changes in diet, or environmental factors.
With PSA screening, most men are diagnosed five to seven years earlier than a decade ago, often with localized disease. But because the cancer progresses slowly, we would not expect aggressive screening to reduce mortality for another few years. And to date, some of the largest drops have been in states with the lowest screening and treatment rates.
Ultimately, if mortality rates continue to drop the way cervical-carcinoma mortality rates did after introduction of the Pap smear, the case for PSA screening will have firmer support. Today, however, its value is still uncertain.