Photo by Todd McNaught
A new study published in the May 27 issue of the New England Journal of Medicine finds that 15 percent of older men with PSA test scores considered to be in the normal range have prostate cancer. Researchers detected tumors in biopsies of healthy men who had PSA test values equal to or less than 4 nanograms per milliliter (ng/ml) for seven consecutive years.
The vast majority of the tumors appeared to be low-grade, or nonaggressive cancers. About 15 percent of the tumors had features commonly associated with high-grade cancers, which have a poorer prognosis than low-grade cancers.
The researchers, including several investigators in the Public Health Sciences Division, do not yet know whether any of the tumors would have developed into harmful cancers that required treatment. Still, the results are expected to add to the discussion over whether the PSA test, which was originally developed to monitor a patient's response to therapy, is a good screening tool for prostate cancer. The controversy stems from a lack of definitive research about whether PSA screening improves survival rates for prostate cancer as well as evidence that the test picks up many harmless cancers for which men receive unnecessary, sometimes debilitating, treatment.
Despite these issues, said Dr. Donna Pauler, one of the study's statisticians and a PHS investigator, "the PSA test remains the best marker available for diagnosing prostate cancer."
The analysis was conducted by the Southwest Oncology Group (SWOG), a national research consortium whose statistical center is jointly housed within PHS and Cancer Research and Biostatistics, a Seattle nonprofit research organization. Dr. Ian Thompson of the University of Texas Health Sciences Center at San Antonio led the study, which was funded by the National Cancer Institute.
Data for the study — the first comprehensive evaluation of prostate-cancer prevalence in men with "normal" PSA levels — was obtained from the Prostate Cancer Prevention Trial (PCPT), a nationwide study published last year that assessed the effects of a drug called finasteride (Proscar) in preventing prostate cancer among men 55 and older. Men who took part in PCPT, half of whom received a placebo, agreed to undergo a yearly PSA test and a prostate biopsy upon the study's conclusion.
"One very unique aspect of the current study was that it involved a population of healthy men who would normally have no reason to undergo a prostate biopsy but had agreed to do so as part of the PCPT," said Phyllis Goodman, lead statistician for the analysis. "That gave us an opportunity to examine the prevalence of cancer among men whose PSA levels were below 4, which is the generally accepted cutoff value for detecting cancer."
The PSA test, developed in 1979, measures the level of a protein called the prostate-specific antigen in a man's blood serum. The protein is produced at low levels by the healthy prostate and is produced in excess by prostate tumors as well as certain benign conditions. Men diagnosed with cancer, regardless of their initial treatment, undergo follow-up PSA testing to monitor for evidence of disease progression or recurrence. The test also is commonly used to screen healthy men for prostate cancer, although researchers haven't demonstrated whether such surveillance affects a patient's survival nor have they conclusively proven the optimal upper limit for a normal test result.
The current study involved 2,950 men aged 62 to 91 who were enrolled in the placebo arm of the PCPT trial and whose PSA levels were below 4 ng/ml for each of the seven years of the trial. Prostate biopsies conducted on each of the men revealed 449 cancers, the equivalent of 15.2 percent. Of these cancers, nearly 15 percent had Gleason scores of 7 or higher, an indication that the cancer was high-grade.
"The prevalence of cancer was substantially higher than what was expected," Pauler said. "Also surprising was the prevalence of high-grade cancers."
The prevalence of prostate cancer was 6.6 percent among men with a PSA level of up to 0.5 ng/ml; 10.1 percent among those with PSA levels of 0.6 to 1.0 ng/ml; 17.0 percent among those with values of 1.1 to 2.0 ng/ml; 23.9 percent among those with values of 2.1 to 3.0 ng/ml; and 26.9 percent among those with values of 3.1 to 4.0 ng/ml.
"That tells us that the PSA test has predictive value even within what is considered the normal range," Pauler said.
Despite the number of cancers found among men with so-called "normal" PSA scores, it is not known whether these cancers actually were dangerous and would require treatment, Goodman said.
"We don't know if these are slow-growing cancers that would never progress beyond their current stage or whether they were early-stage cancers that would have in fact been detected by PSA in the next year or few years," she said.
Multiple studies, including those led by PHS investigator Dr. Ruth Etzioni, have found that the PSA test diagnoses more men with prostate cancer than may ever experience its symptoms. In 2002, she and colleagues determined an "overdiagnosis" rate of about 30 percent among men whose cancers were detected by the PSA test. An earlier study led by Etzioni found that approximately 36 percent of men would develop prostate cancer within their lifetimes, but only one-quarter of these would be diagnosed with the disease in the absence of the PSA test.
"Since we have had widespread PSA testing in this country, a man's lifetime probability of developing prostate cancer has doubled," Etzioni said. "I think that means that the PSA test is already detecting the vast majority of bad tumors."
Many argue that lowering the threshold PSA value for cancer diagnosis could pose more harm than good.
Dr. Leslie Ford, associate director for clinical research in NCI's Division of Cancer Prevention and a coauthor of the study, said that better methods are needed to distinguish harmless, slow-growing cancers from the more aggressive ones.
"If more biopsies are performed at lower PSA levels, more cancers will be found and treated," Ford said. "But some men will undergo treatment, and the risks associated with it, for tumors that would never have been clinically significant."
More definitive results about the benefits of PSA testing are expected to emerge from the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial, a large-scale national study that will assess the impact of screening on survival. Data from the trial is expected to be available in 2008.
Goodman said that the SWOG group is designing a follow-up study to follow the participants who were diagnosed with prostate cancer to determine the long-term outcomes of the participants, such as whether they survived or if their cancers had spread. "We'd like to find out how aggressive these cancers really are," she said.
In addition, the team plans to analyze the PSA test results of all men who took part in the PCPT study to come up with concrete values for how sensitive and specific the PSA test really is.
"The PCPT study has provided us with a fascinating collection of data from which we can expect to learn many more interesting things about prostate cancer," Pauler said.
Other contributors to the study were PHS members Dr. Catherine Tangen, principal investigator of the PCPT; Dr. John Crowley principal investigator of the SWOG Stastistical Center and chief executive officer of Cancer Research and Biostatistics; M. Scott Lucia and E. David Crawford of the University of Colorado Health Science Center; Drs. Howard Parnes, Lori Minasian and Leslie Ford of the National Cancer Institute; Dr. Scott Lippman, M.D. Anderson Cancer Center; and Dr. Charles Coltman, SWOG.