AN ANALYTICAL COMPARISON OF THE 3 MOST COMMONLY USED PROSTATE-SPECIFIC ANTIGEN ASSAYS - TANDEM-R, TANDEM-E, AND IMX

Citation
Je. Oesterling et al., AN ANALYTICAL COMPARISON OF THE 3 MOST COMMONLY USED PROSTATE-SPECIFIC ANTIGEN ASSAYS - TANDEM-R, TANDEM-E, AND IMX, Urology, 46(4), 1995, pp. 524-532
Citations number
25
Categorie Soggetti
Urology & Nephrology
Journal title
ISSN journal
00904295
Volume
46
Issue
4
Year of publication
1995
Pages
524 - 532
Database
ISI
SICI code
0090-4295(1995)46:4<524:AACOT3>2.0.ZU;2-I
Abstract
Objectives. To compare the similarity of individual prostate-specific antigen (PSA) results when using the Tandem-R PSA assay, the Tandem-E PSA assay, or the IMx PSA assay; to assess the lot-to-lot variation wi thin (intra-assay interlot) and between (interassay interlot) the Tand em-R, Tandem-E, and IMx PSA assays; and to evaluate the individual and overall potential lot-to-lot bias of the Tandem-R, Tandem-E, and IMx PSA assays. Methods. Forty-nine serum samples (PSA values from 0 to 85 ng/mL) were each tested by three separate lots (manufacturer's reagen t materials) of Tandem-R, Tandem-E, and IMx PSA assays, Therefore, a t otal of nine different lots were utilized per patient sample in this i nvestigation. Analyses primarily focused on three ranges: 0 to 10 ng/m L (low), 10 to 85 ng/mL (high), and 0 to 85 ng/mL (overall). Results, In the 0 to 10 ng/mL range, 93% of the assay comparisons yielded an ac tual difference of less than 1 ng/mL. All three assays demonstrated ex cellent correlation within and between their three respective lots, wi thin all three ranges. The 95% confidence intervals around the percent coefficient of variation (CV) demonstrated similar results for each a ssay (CV range, 3.2% to 6.0%). All lots demonstrated an average actual PSA bias of less than +/- 1 ng/mL. The average percent PSA bias was a lso similar between all three assay systems, All lots demonstrated a l ess than +/- 4% bias. Conclusions. Overall, the Tandem-R and IMx PSA a ssays yielded slightly lower results than the Tandem-E PSA assay. Howe ver, these differences were not statistically significant. In addition , the overall lot-to-lot variation (intra-assay and interassay) was no t statistically significant, and the actual or percent PSA bias was mi nimal with these three assays. Therefore, a clinician can feel confide nt that a patient's serum sample should yield a similar and interchang eable result, whether it is determined by the Tandem-R, Tandem-E, or I Mx PSA assay system.