A positive likelihood ratio of 10 is usually considered to “rule in disease”, but with PSA testing, lower ratios are seen: 4.5, 5.5, and 6.4 with cutoff values of 3, 4, and 5 ng/mL, respectively. Conversely, a negative likelihood ratio of 0.1 is typically considered to “rule out disease”, but the ratios with PSA testing were higher: 0.47, 0.61, and 0.70 corresponding to the above cutoff values.
The authors did, however, find that a PSA level below 1.0 ng/mL virtually rules out prostate cancer.
“It seems clear that further biomarkers and diagnostic tools are needed before population based screening of healthy men can be recommended,” study co-author Dr. Mattias Johansson told Reuters Health. “In particular, tools that help distinguish rapidly growing and potentially lethal tumors from slow growing tumors are warranted in order to minimize overdiagnosis and overtreatment.”
Dr. Johansson, from the International Agency for Research on Cancer, Lyon, France, and his colleagues analyzed data from a case-control study nested in a longitudinal cohort. Included were 540 case subjects and 1034 controls matched by age and date of blood draw.
On average, blood samples were taken 7.1 years before prostate cancer diagnosis, the report indicates.
The area under the curve for PSA was 0.84. At cut-off values of 3, 4, and 5 ng/mL, the sensitivity of PSA testing for prostate cancer was 59%, 44%, and 33%, respectively, and specificity was 87%, 92%, and 95%, respectively.
In an accompanying analysis, Dr. Jennifer Stark, from Harvard School of Public Health, Boston, and colleagues examined the benefits and harms of PSA screening and concluded that at present there is simply not enough data to support population based screening. Moreover, they note, further studies with more precise measures are needed to gauge the financial and psychological toll of false positive PSA results, overdiagnosis, and overtreatment of prostate cancer.
Before men undergo PSA screening, they should be fully informed of the benefits, harms, and uncertainty associated with the test, Dr. Stark and colleagues emphasize.
Editorialists Dr. Dragan Ilic and Dr. Sally Green, from Monash University, Clayton, Australia are of the same opinion and add that “clinicians should consider using likelihood ratios together with a patient’s individual risk factors for the disease to explain the potential benefits and harms of the PSA test, allowing patients to contribute to decisions.”
BMJ Online First 2009.