COMPARABILITY OF THE TANDEM-R AND IMX ASSAYS FOR THE MEASUREMENT OF SERUM PROSTATE-SPECIFIC ANTIGEN

Objectives. To assess the comparability of the Tandem-R and IMx serum prostate-specific antigen (PSA) assays across levels of the ratio of f ree-to-total serum PSA found in a community-based population of health y men. Methods. Banked serum samples from the baseline component of th e Olmsted County Study of Urinary Symptoms and Health Status Among Men were thawed and analyzed using the Tandem-R and IMx PSA assays. Serum levels also were determined for the free, noncomplexed form of PSA, P SA complexed to alpha-1 antichymotrypsin, and total PSA with a researc h-based immunofluorometric assay. Results. The results of the Tandem-R and IMx assays were strongly correlated at all levels of the ratio of free-to-total serum PSA. The Spearman correlation coefficients ranged from 0.87 to 0.98 (all p < 0.001). The relationship between the Tande m-R and IMx assays, however, differed at low levels of free-to-total s erum PSA compared with high levels. In the lowest 10th percentile of t he ratio of free-to-total serum PSA (0.04 to 0.18), the IMx assay read lower than the Tandem-R (slope +/- standard error = 0.92+/-0.04, inte rcept +/- standard error = 0.21+/-0.14); whereas in the upper 10th per centile of free-to-total ratio (0.46 to 0.65) the IMx assay yielded va lues higher than the Tandem-R assay (slope = 1.21+/-0.07, intercept = 0.14+/-0.05). In the middle 90%, the slope did not statistically diffe r from 1.0. Conclusions. For the majority of men, results of the Tande m-R and IMx PSA assays were virtually identical. The small differences found would not be of clinical significance for most men but should b e considered when comparing results of different assays in sequential determinations for a specific man.