PSA Test Losing Diagnostic Value, Study Says
Monday, March 10, 2008; 12:00 AM
MONDAY, March 10 (HealthDay News) -- Changes in medical practice have undermined the value of the prostate-specific antigen (PSA) test as a diagnostic tool for prostate cancer, a new study asserts.
However, that assertion was immediately challenged by other experts in the field.
The American Cancer Society recommends annual screening for prostate cancer for most men after age 50. The screening is done in two steps: a digital rectal exam, followed by a blood test for PSA, a protein produced by the prostate gland.
"In men with an abnormal digital rectal examination, it [the PSA test] still is very predictive," said study author Dr. Douglas S. Scherr, clinical director of urological oncology at Weill Cornell Medical Center in New York City. "For men with a normal digital rectal examination, it has lost the predictive accuracy it had in the 1990s."
For those men, exact PSA readings matter less, because doctors are doing more biopsies now and are using improved biopsy techniques, Scherr explained. His report was expected to be published in the April 15 issue ofCancer. In a retrospective analysis, his team looked at 1,607 prostate biopsies performed at Weill Cornell between 1993 and 2005, and uncovered several trends.
In the 1990s, biopsies generally were limited to men with PSA readings higher than 4 nanograms per milliliter of blood. Later, they were often done when the PSA levels were between 2.5 and 4, Scherr said.
"In the early and mid-90s, we used six needles when we did a biopsy," he said. "Now, we stick in 14. Taking more cores and doing biopsies at lower PSA levels has allowed us to keep the prostate cancer detection level steady. If we kept the PSA level at 4 rather than at 2.5, the cancer detection rate would have fallen rapidly."
Increased awareness of prostate cancer has also changed the screening picture, Scherr said. "You almost never have an abnormal rectal examination these days," he said.
There is a need to develop new biomarkers for prostate cancer beyond PSA, Scherr said. "Should a man when he turns 50 have a biopsy?" he said. "It is certainly an important question that needs being looked at."
But the study merely "makes a statistical point about an unusual population," said Andrew Vickers, a statistician whose title at Memorial Sloan-Kettering Cancer Center in New York is attending research methodologist.
"It apples to men who have regular screening," Vickers said. "It doesn't apply to men who haven't been screened. For those men, the PSA test might be more accurate."
And the study is misleading because of information it doesn't contain, said Dr. Ian M. Thompson, chairman of urology at the University of Texas Health Science Center at San Antonio.
The study includes only men on whom biopsies were done, Thompson said. "If you don't biopsy something, you don't know if you have cancer or not," he said. "The study is informative only of men who had biopsies. There is nothing in it about the criteria they used to recommend biopsies."
Risk factors for prostate cancer, such as ethnicity, age and family history, are not mentioned in the study, Thompson added. "You don't know anything about the people in their practice who never had a biopsy," he noted.
The PSA test has flaws, Thompson acknowledged. Most notably, it is unable to distinguish between prostate cancers that will grow so slowly that they do not endanger life and those that are aggressive enough to kill.
"But of all the tools we currently have, the PSA test is valuable," he said. "People dealing with other kinds of cancer wish they had a blood test like PSA."
You can learn about the diagnosis and treatment of prostate cancer from the National Cancer Institute.
SOURCES: Douglas S. Scherr, M.D., clinical director, urological oncology, Weill Cornell Medical Center, New York City; Andrew Vickers, Ph.D., statistician, Memorial Sloan- Kettering Cancer Center, New York City; Ian M. Thompson, chairman, urology, University of Texas Health Science Center, San Antonio; April 15, 2008,Cancer