| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1
Department of Clinical Immunology, Albany Medical Center, 43 New Scotland Ave. MC-22, Albany, NY 12208
2
Department of Clinical Chemistry, The Johns Hopkins Medical Institutions, 600 N. Wolfe St., Baltimore, MD 21287-7065
3
Department of Clinical Chemistry, Brigham and Women’s Hospital, 75 Francis St., Boston, MA 02115
4
Physicians Laboratory Services, Inc., 4840 F St., Omaha, NE 68117
5
Immunodiagnostics Development, Beckman Coulter, Inc., PO Box 269006, San Diego, CA 92196-9006
a author for
correspondence: fax 858-621-4750,
kgloveland{at}Beckman.com
Previously reported clinical trials have shown that measurement of the different forms of prostate-specific antigen (PSA) (1) is useful in the differentiation of prostate cancer from benign prostatic conditions. Men with prostate cancer tend to have lower percent free PSA (%FPSA) values than men with benign disease (2)(3)(4); %FPSA is the ratio of free PSA to total PSA times 100%.
Beckman Coulter, Inc. (Fullerton, CA) has developed the Access Hybritech PSA and Hybritech PSA assays for use on the automated Access Immunoassay System. The assays are two-site immunoenzymatic ("sandwich") assays that use mouse monoclonal antibody in alkaline phosphatase conjugate and paramagnetic particles coated with a second mouse monoclonal antibody. After unbound particles are removed by washing, a chemiluminescent substrate, Lumi-Phos 530,1 is added to produce light directly proportional to the amount of analyte in the sample as determined from a stored calibration curve. Calibrators are assigned up to 150 µg/L PSA and 20 µg/L FPSA.
Analytical performance of the new assays was evaluated at four clinical laboratories located in the United States and at the manufacturer’s laboratories. Proficiency of the Access instruments and operators was demonstrated by assaying an identical panel of 18 human samples. Serum aliquots were stored at -70 °C, shipped frozen, and tested in duplicate over 2 days on both the Access and the reference Tandem methods. Proficiency panel results from each laboratory were analyzed by Deming regression and compared with the values assigned by the manufacturer. Each laboratory achieved results within 10% of the target values before proceeding. For the remainder of the study, Lyphochek Immunoassay Plus Control2 and Access QC control with ranges assigned for the Access instrument for each assay were used daily to verify calibration of individual instruments and accurate sample recovery from stored calibration curves. Controls spanning the range of each assay recovered without trending and within the expected 2 SD range at the beginning and end of the 28-day stored assay calibration interval on each instrument. Consistent recovery of controls verified stable systems.
As recommended in NCCLS guideline EP5 (5), we determined within-run, between-run, and total imprecision for four lots of reagent (two PSA and two FPSA), using commercial controls on one instrument in each laboratory. Among the four laboratories, the overall CV (the root mean square of the median variances) across the range of the PSA assay was 3.9% (range, 1.8–6.7%); for the FPSA assay, it was 3.8% (range, 2.5–6.0%).
The detection limit was calculated as the concentration
corresponding to the mean response (relative light units) plus 2 SD of
10 replicates of the zero calibrator when read from the stored
calibration curve derived from the same reagent lot (target detection
limits: 
0.008 µg/L PSA and 0.005 µg/L FPSA). Across
laboratories using two reagent lots, the detection limit for PSA was
0.005 µg/L (range, 0.003–0.008 µg/L) and for FPSA was 0.002 µg/L
(range, 0.001–0.005 µg/L). Secondly, at three laboratories, human
sera (PSA <2.5 µg/L and FPSA <0.25 µg/L) were diluted in PSA
sample diluent, and the within-run CVs of duplicate analyses of each
dilution were calculated. For these three laboratories, the mean lowest
concentration contributing to a plotted straight line of the dilutions
and with CV <20% for the replicates (n = 2) was 0.007
µg/L for PSA and 0.005 µg/L for FPSA.
To verify that an assay does not give falsely low results in the presence of the very high concentrations of PSA or FPSA that may be present in human samples (antigen excess or hook effect), aliquots of the zero calibrator containing PSA concentrations of 19–625 000 µg/L and FPSA concentrations of 5–160 000 µg/L were prepared by the manufacturer. The instrument signal was plotted vs concentration for three lots of each assay to determine the point where excess antigen falsely read within the calibration range. The Access Hybritech PSA assay did not demonstrate antigen excess up to 50 000 µg/L PSA, and FPSA did not hook until >20 000 µg/L.
To demonstrate linearity, 15 samples with increased PSA (65–128 µg/L) and 7 samples with increased FPSA (6–12 µg/L) were diluted and assayed, using two reagent lots for each assay. The expected results vs observed results were analyzed by Deming regression. The target dilution linearity was a regression slope of 0.90–1.10. The slope of the Deming regression for observed PSA concentrations vs expected concentrations on the 15 samples ranged from 0.99 to 1.02 (r >0.99). The slope of the Deming regression for observed FPSA concentrations vs expected concentrations of the seven samples varied from 0.98 to 1.01 (r >0.99).
As recommended in NCCLS guideline EP7 (6), we tested serum for potential interference from endogenous substances in blood and 33 drugs frequently used by a population of men >50 years of age, and none interfered in the assay.
PSA and FPSA results from serum samples with the new assays
were compared with results obtained by comparison methods: Tandem-R PSA
at one laboratory, Tandem-E PSA (on the Photon®
ERA) at three laboratories, and Tandem-R free PSA at all
laboratories. Specimens for method comparisons were processed by each
laboratory by routine methods within 3 h after blood was
collected. Excess serum was stored frozen at -20 °C. Serum frozen
for longer than 1 month was stored at -70 °C
(7)(8). All identifying and demographic
information was removed from these leftover sera in accordance with
local Institutional Review Board policies. No results from
investigational reagents were reported to physicians. Method
comparisons were analyzed by Deming regression for results falling
within the range of each assay. Table 1
shows results by assay for each laboratory and for all
laboratories combined. Fig. 1
A shows the overall Access Hybritech PSA vs Tandem equation:
y = 1.074x + 0.19 (slope range,
0.989–1.145; n = 400). Fig. 1B
shows the overall equation of
Access Hybritech free PSA vs Tandem free PSA: y =
1.047x - 0.04 (slope range, 0.967–1.071; n =
391). Because controls on the instruments in each laboratory
consistently recovered within 2 SD of one another throughout the study
and the overall imprecision of the Access system was <4%, differences
in method comparison results were attributed to the differences in
distribution of PSA concentrations in the different patient populations
and variations in each individual laboratory environment.
|
|
Because the data for these specimens left over from routine testing did not include clinical or demographic information, %FPSA could not be interpreted with respect to diagnosis. However, comparison of %FPSA from four sites derived from the Access Hybritech methods and the Tandem reference methods analyzed by Deming method showed an overall regression of Access (y) assays vs Tandem (x) of: y = 1.01x - 0.01 (r = 0.933; n = 391).
To assess the ability of the Access Hybritech PSA assay to
recover each PSA form [FPSA and PSA complexed with
1-antichymotrypsin (PSA-ACT)] equally,
four sets of samples with constant PSA concentrations (4, 7, 15, and 30
µg/L) and increasing proportions of FPSA (0%, 25%, 50%, 75%, and
100%) were prepared with purified, characterized, and fractionated
FPSA and PSA-ACT (9). At each PSA concentration, the slope
of the PSA results was near zero (0.002, 0.010, 0.013, and 0.028),
which indicates that the Access Hybritech PSA assay is unaffected by
the proportions of FPSA in the sample. The average molar response ratio
(100% FPSA/0% FPSA) for the Access Hybritech PSA assay was 1.11.
These data demonstrate that the Access Hybritech PSA assay is equimolar
(nonskewed). For 391 human samples in the multicenter study, the ratio
of the Access Hybritech PSA assay over the Tandem PSA was plotted vs
%FPSA in each sample. Again, the slope of the data was near zero
(y = 0.09x), indicating that the PSA results
from both assays responded the same in regard to the concentration of
FPSA and PSA-ACT in the sample. Overestimation of true PSA recovery by
a "skewed" assay can lead to unnecessary prostate biopsies, whereas
underestimation can lead to delayed diagnosis. This is important
because previous work has shown that the concentrations of FPSA and
total PSA in a given specimen determined with assays from different
manufacturers can vary widely because of differences in assay methods
and reagent specificity (10)(11)(12). Therefore, values obtained
with different manufacturers’ assays cannot be used interchangeably.
In conclusion, these automated chemiluminescent Access Hybritech PSA and free PSA assays supply information to the physician equivalent to that provided by the Hybritech Tandem manual assays. Both systems use the same antibody pairs, and each assay is approved by the Food and Drug Administration for the same clinical indications. This multicenter evaluation shows that the performance of the automated Access Hybritech PSA and free PSA assays is analytically specific, sensitive, linear, accurate, and precise.
Acknowledgments
This study was funded by Beckman Coulter, Inc.
Footnotes
1 Lumi-Phos 530 is a trademark of Lumigen, Inc. 
2 Lyphochek Immunoassay Plus Control is a trademark of Bio-Rad Laboratories.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  F. H. Schroder, J. Hugosson, M. J. Roobol, T. L.J. Tammela, S. Ciatto, V. Nelen, M. Kwiatkowski, M. Lujan, H. Lilja, M. Zappa, et al. Screening and Prostate-Cancer Mortality in a Randomized European Study N. Engl. J. Med., March 26, 2009; 360(13): 1320 - 1328. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Stephan, M. Klaas, C. Muller, D. Schnorr, S. A. Loening, and K. Jung Interchangeability of Measurements of Total and Free Prostate-Specific Antigen in Serum with 5 Frequently Used Assay Combinations: An Update Clin. Chem., January 1, 2006; 52(1): 59 - 64. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
