Assessing the necessity of including a crossover period with dual reporting when changing total prostate-specific antigen methods

OBJECTIVES: Options are limited for the laboratory when changing analytical methods for tumor markers. Standardization between methods should alleviate this burden. Our objective was to assess the necessity of dual reporting for total prostate-specific antigen (tPSA) when changing methods that have been calibrated against the World Health Organization PSA reference material. DESIGN AND METHODS: Following an initial validation where 40 patient samples were divided and analyzed on the Roche E-modular (current method) and Abbott ARCHITECT (new method) for tPSA, a crossover period ensued with dual reporting of tPSA over 54days (n=1110 samples with both tPSA results reported). Passing & Bablok regression was performed, with agreement during the crossover period also determined by two methods. Method#1: if the percent difference between the result pair was ≤0.33×the intraindividual biological variation (0.33×CVi=6.0%) for tPSA; and Method#2: if the Roche tPSA result was within an acceptable analytical range based on Abbott's maximum allowable imprecision for the tPSA assay (CV≤8%) to determine if there was a range of optimal agreement that might not require dual reporting. RESULTS: During the crossover period, Passing & Bablok regression yielded the following relationship: Abbott tPSA=1.04(95% CI: 1.03 to 1.04)×(Roche tPSA)-0.02(95% CI: -0.03 to -0.02). Only 52% (95%CI: 49 to55) of the results were in agreement by Method#1 whereas 95% (95% CI: 92 to 97) of the results between 3.3 and 19μg/L were found to have acceptable agreement based on Method#2. CONCLUSIONS: These findings suggest that dual reporting of tPSA is required when changing methods. However, the extent of dual reporting may be limited to specific concentration ranges with future studies required to validate selective dual reporting when changing tPSA assays.