Method Comparison Studies for Prostate Specific Antigen and Unconjugated Estriol Immunoassays

Abstract

OBJECTIVE: Method comparison studies were performed in order to move a semi-automated prostate specific antigen (PSA) immunoassay and a manual unconjugated estriol (uE₃) immunoassay to an automated chemistry immunoassay analyzer. The results of the two method comparison studies are compared.

DESIGN: Serum samples collected on patients with physician orders for PSA or uE₃ were assayed by both methods. PSA samples were assayed on a Hybritech Tandem Photon ERA and on two Beckman Coulter Access instruments. UE₃ samples were assayed by RIA and on two Beckman Coulter Access instruments. Linear regression analysis was performed on both sets of data and within-run precision and dilution studies were performed on the PSA Access method.

SETTING: Clinical chemistry laboratory, West Virginia University Hospitals Inc, Morgantown WV.

RESULTS: PSA linear regression analysis for the two methods (ERA and Access 1) were y = 1.0008x + 0.0393, r = 0.9976, SE = 0.1319, n = 37 and (ERA and Access 2), y = 1.0019x + 0.0486, r = 0.9964, SE = 0.1632, n = 37. Within-run precision studies for both Access instruments produced acceptable coefficient variations and dilution study results were in PSA reportable range. uE₃ linear regression analysis for the two methods (RIA and Access 1) were y = 1.4105x - 0.3741, r = 0.8696, SE = 0.8330, n = 33 and (RIA and Access 2) were y = 1.315x - 0.2292, r = 0.8643, SE = 0.7964, n = 33.

CONCLUSION: The results of the method comparison studies for PSA were acceptable and the automated PSA immunoassay method was adopted. The results of the uE₃ comparison studies did not show good correlation; the automated method was not adopted.

ABBREVIATIONS: CV = coefficient of variation; MoM = multiple of the median; PSA = prostate specific antigen; r = correlation coefficient; RIA = radioimmunoassay; SE = standard error; uE₃ = unconjugated estriol.