July 23, 2003
Common screening test may be missing majority of prostate cancers
Study finds 82 percent of prostate cancers may be missed with PSA test
BOSTON - Through a specialized analysis of existing prostate cancer data, researchers from Brigham and Women's Hospital (BWH), Dana-Farber Cancer Institute (DFCI) and Harvard School of Public Health (HSPH) have shed new light on the debate surrounding the efficacy of a popular cancer screening test. The re-evaluation of prostate-specific antigen (PSA) - a blood test used to detect prostate cancer - suggests that the screening method may have appeared more effective than it really is, potentially missing a high number of cancers. These findings are outlined in the July 24 issue of the New England Journal of Medicine.
"Some past studies have shown PSA testing to be almost perfect, however clinicians suspect the test produces a significant number of false positives and false negatives," said Rinaa Punglia, MD, MPH, a radiation oncologist at BWH and DFCI. "With this in mind, we approached the data in a new way and looked at how research biases may have been influencing the appearance of the test, making it seem to perform better than it really does."
The team found that characteristics of the PSA test might be based on skewed data. Their reanalysis illustrated that current thresholds (meaning the PSA value clinicians use to differentiate abnormal from normal readings to ultimately recommend biopsy) are likely insufficient and can be improved.
"We were interested in finding out how well this test performs," said Punglia. "When adjusting the numbers to account for bias, we discovered that physicians who use a PSA of greater than four (nanograms per milliliter) to recommend biopsy - the current standard - may be missing 82 percent of cancers in men under 60 years of age and 65 percent in men over 60." The team went on to demonstrate that lowering the threshold to 2.6 in younger men may double the cancer-detection rate, without significantly altering the number of false positives.
Prostate cancer is the second most common cause of cancer in American men and it is estimated that almost 29,000 will die of the disease this year. It is one of the only cancers with a screening option, yet endorsement of routine PSA testing is not uniform. Punglia said the debate largely stems from the fact that no study has yet shown PSA screening reduces mortality. Her study may point to a possible reason why.
The study, drawn from data of 6,691 men who underwent PSA screening at Washington University School of Medicine and 705 who subsequently underwent biopsy, is based on computer models. First, the data was adjusted for a condition called verification bias and then plotted on a graph that measures the diagnostic capabilities of the test. After comparing the unadjusted numbers to the adjusted ones, the researchers found that selection bias in earlier data had led some researchers to incorrectly assume that PSA screening had near perfect discriminatory power.
"The medical community has probably overestimated the sensitivity of this test, and it may be reasonable to lower our threshold from four to 2.6," said Punglia. "Although we won't know if PSA testing decreases mortality until the results of randomized clinical trials come in, based on this study, men who undergo screening, along with their physicians, may want to think differently about a PSA of three."
Dana-Farber Cancer Institute is a principal teaching affiliate of the Harvard Medical School and is among the leading cancer research and care centers in the United States. It is a founding member of the Dana-Farber/Harvard Cancer Center (DF/HCC), designated a comprehensive cancer center by the National Cancer Institute.
BWH is a 725-bed nonprofit teaching affiliate of Harvard Medical School and a founding member of Partners HealthCare System, an integrated health care delivery network. Internationally recognized as a leading academic health care institution, BWH is committed to excellence in patient care, medical research, and the training and education of health care professionals. The hospital's preeminence in all aspects of clinical care is coupled with its strength in medical research. A leading recipient of research grants from the National Institutes of Health, BWH conducts internationally acclaimed clinical, basic and epidemiological studies.
