Semi-quantitative immunochromatographic test for prostate specific antigen in whole blood: tossing the coin to predict prostate cancer?

A strip test for the semi-quantitative determination of prostate specific antigen (PSA) based on whole blood was evaluated by the Working Group on Laboratory Diagnostics of the German Urological Association.

In 301 men (156 volunteers and 145 referred patients, mean ages 57 years and 64 years) handling and capability of the strip test to predict PSA serum concentrations (Access Hybritech, Beckman Coulter, Inc., Fullerton, CA, USA) relative to a cutoff of 4 micro g/l were investigated. For patients, the strip test was performed by investigators, volunteers accomplished it according to leaflet instructions. Each test result was interpreted by both, the patient/volunteer and an investigator.

24% of the volunteers required assistance with test performance. The overall agreement between Access PSA (range 0.1-29.7) and strip test was 76% (read by participants) and 79% (read by investigators). Within the PSA ranges 0.1-3, 3.1-5 and 5.1-10, participants yielded agreement rates of 89%, 40% and 56%, respectively. Studying the 2.1-10 PSA range most relevant for prostate cancer screening, agreement rate (53%), sensitivity (50%) and specificity (67%) were barely superior to random chance. Volunteers alone (PSA range 0.1-11.3 micro g/l) yielded a specificity of 92%, however, all PSA values above the cutoff (n=8) went unrecognized and all positive strip tests (n=12) were incorrect (sensitivity 0%, positive predictive value 0%).

Test handling and interpretation was difficult. The rate of false strip test results is disappointing even for PSA ranges remote from the cutoff. In its present form the strip test fails to meet the clinical requirements and its launch to the market was therefore discontinued.