Clinical Chemistry 43: 491-497, 1997;
(Clinical Chemistry. 1997;43:491-497.)
© 1997 American Association for Clinical Chemistry, Inc.
Clinical evaluation of the Byk LIA-mat CA125 II assay: discussion of a reference value
Johannes M. G. Bonfrer1,a,
Catharina M. Korse1,
Rob A. Verstraeten2,
Gerard J. van Kamp3,
Guus A. M. Hart4 and
Peter Kenemans2
1
Departments of Clinical Chemistry and
4
Statistics, The Netherlands Cancer Institute (Antoni van Leeuwenhoek Huis), Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands.
2
Departments of Obstetrics and Gynecology and
3
Clinical Chemistry, Academic Hospital Free University,
De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands.
a Author for correspondence. Fax +31 20 6172625.
 |
Abstract
|
|---|
The Byk LIA-mat CA125 II assay was compared with the Centocor IRMA CA125
II. Serum samples studied (n = 1012) were obtained from 652
apparently healthy females, 61 pregnant women, and 299 patients with
benign and malignant gynecological tumors. The CA125 II assay value at
the 95th percentile of the total healthy group was 29 kU/L for the
LIA-mat and 32 kU/L for the Centocor assay. For the LIA-mat assay the
95th percentile was 31 kU/L (Centocor 36 kU/L) for the group <45 years
and 21 kU/L (Centocor 25 kU/L) for women >55 years of age. By using
ROC curves we found the optimal pretreatment Byk LIA-mat CA125 II value
differentiating between benign and malignant ovarian tumors to be 95
kU/L. Pretreatment CA125 values >1000 kU/L were detected in serum
samples of patients with advanced epithelial ovarian cancer.
Key Words: indexing terms: tumor marker • immunoassay • gynecological neoplasms
|
Introduction
|
|---|
After the introduction 12 years ago of the CA125 tumor marker
assay, its use has become widespread in clinical practice
(1). Although the initial publications mainly focused on
the monitoring of patients with epithelial ovarian carcinoma
(2), other applications were soon described. For example,
Niloff et al. proposed to use it as an accessory to second-look
laparoscopy (3), and various authors introduced CA125 in
the management of endometriosis (4). Vasilev et al.
included the serum assay values of this marker in the preoperative
evaluation of pelvic masses (5). However, as a diagnostic
tool the CA125 antigen has a serious drawback as increased assay values
have been found in a multitude of conditions (i.e., congestive heart
failure (6), liver cirrhosis (7), and benign
ovarian diseases (8)(9)).
The OC 125 monoclonal antibody (mAb) was first described in 1981
(10).1
It was obtained by immunizing BALB/c mice with
the OVCA 433 cell line isolated from the ascites of a patient with
serous papillary cystadenocarcinoma. Previously, CA125 II assays were
introduced in which the CA125 capture antibody is murine mAb M11, with
a higher avidity for epitopes of the CA125 molecule (11).
An extensive evaluation of technical aspects of the Byk LIA-mat CA125
II has been published recently (12). To assess the impact
of this newly developed sandwich assay for clinical practice, we
carried out a comparative study with second-generation CA125 II assay
with clinical data, in particular with the use of reference values.
|
Materials and Methods
|
|---|
assay procedures
The LIA-mat CA125 II (Byk-Sangtec, Dietzenbach, Germany) is a
heterologous assay with M11 as the solid-phase (coated tube)-bound
capture mAb and OC 125 as tracer mAb, covalently labeled with an
isoluminol derivative. The light signal is proportional to the amount
of CA125 antigen in the sample.
The LIA-mat assays were processed on a LIA-mat 300 System
(13).
The Centocor® CA125 II (Centocor, Malvern, PA) is a
one-step "sandwich" IRMA involving polystyrene beads coated with
the M11 mAb as solid phase. The tracer is the
125I-labeled murine mAb OC 125. Day-to-day CV for the
Byk LIA-mat assay was 10.4% at 215 kU/L and 11.2% at 75 kU/L. The
Centocor CA125 II day-to-day precision was 10.0% at 180 kU/L and
12.2% at 66 kU/L. All tests were performed in duplicate according to
the manufacturer's instructions, and reference sera were included with
each batch. Blood, not processed immediately, was centrifuged within
1 h and the serum was stored at -20 °C the same day.
Tumors were staged according to the International Federation of
Gynecology and Obstetrics (FIGO) classification for gynecological
cancer (14).
sample collection
CA125 was determined retrospectively in a total of 1012 serum
samples:
Group A (n = 652).
Apparently healthy women, age
40 years, who volunteered to take part in a screening program for
early detection of ovarian cancer were included when no abnormalities
were found after they underwent pelvic examination, including an
ultrasound scan. Mean age was 49 with a maximum of 76 years.
Group B (n = 61).
Blood from women with an
uneventful pregnancy was collected in a maternity clinic. The available
sample was used when informed consent was given. Twenty of the women
were in the first trimester, 21 in the second trimester, and 20 in the
third trimester of their pregnancy. Ages were 24–44 years (mean 33
years).
Group C (n = 134).
Patients with benign pelvic
diseases who were operated on in the Departments of Obstetrics and
Gynecology of the Academic Hospital of the Free University were
included when they gave informed consent.The following benign ovarian
tumors were pathologically established: (a) serous
cystadenoma (n = 35), (b) mucinous cystadenoma (n
= 18), (c) fibroadenoma (n = 8), (d) dermoid
cyst (n = 10), and other benign pelvic tumors (n = 63). Mean
age was 44 years, with a range of 14–84 years.
Group D (n = 165).
Sera available from patients
diagnosed with primary malignant gynecological tumors from 1985 to
1993: (a) cervix (n = 23): stage I/II (n = 12)
(adenocarcinoma n = 5, squamous cell n = 6), stage III/IV
(n = 11) (adenocarcinoma n = 7, squamous cell n = 7);
(b) endometrium (n = 57): stage I/II (n = 45),
stage III/IV (n = 8); (c) ovary (n = 85): stage
I/II (n = 26) (epithelial n = 21, nonepithelial n = 4),
stage III/IV (n = 59) (epithelial n = 58, nonepithelial
n = 1).
statistics
Because of the marked skewed distribution and the large overall
range of the CA125 values, these were logarithmically transformed
before any analysis, resulting in reasonable normality of residuals and
constancy of variance. The relation between CA125 values and age in
healthy women was studied by using restricted cubic spline regression
analysis (15). To compare the two assay methods, the
difference of the result was plotted against the corre- sponding
mean after logarithmic transformation. To identify outliers, residual
analyses were performed. Variances were compared with Bartlett's test.
Normality was checked by inspecting normal probability plots with the
Shapiro–Francia test, or occasionally the Shapiro–Wilk test, and
quantified by Royston's V (16). To assess the test
performance characteristics of the assays, the sensitivity and
specificity at maximal overall test accuracy were calculated. Maximal
overall test accuracy, defined as the shortest distance from the upper
left corner of a ROC curve, was derived from actual ROC curves. ROC
curves were constructed according to the NCCLS guidelines
(17) and were evaluated by calculating and comparing areas
under the curves (AUCs) according to DeLong et al. (18).
The SAS 6.10 (Windows) package was used for the main analyses.
|
Results
|
|---|
assay comparison
A comparison between the results of the Byk CA125 II and the
Centocor CA125 II assay was made by using the difference plot (Fig. 1
) (19). Differences of logarithmically transformed
values of 1012 samples were plotted against the average value of both
assays. The Byk assay presented lower results in the lower calibrator
range and somewhat higher values at concentrations >90 kU/L (ln 4.5).
View larger version (25K):
[in this window]
[in a new window]
|
Figure 1. Bland–Altman plot of difference between Byk CA125 II and
Centocor CA125 II assays.
The x-axis shows the mean of logarithmic concentrations of
the Byk and Centocor assays (absolute figure). On the y-axis
the difference of Byk vs Centocor is given.
|
|
distribution of ca125 values
Healthy controls (n = 652).
The distribution of
CA125 values in the healthy control population is nongaussian and skews
positively. The relation between Byk LIA-mat CA125 values after
back-transformation to the linear scale and age is depicted in Fig. 2
. The predicted values (central line) and 95% prediction limits
(outer lines), according to the restricted spline method, are
indicated. There is evi- dence for nonlinearity
(P = 0.0004): relatively constant values up to age 45,
then a decrease until age 55. After age 55, the CA125 values seem to
stabilize again. The variation around the straight line of the normal
plot is higher (as could be expected from chance alone), the main
problem being a larger-than-expected group of females at the lower end
(studentized residual around -3). Results of the analyses of the Byk
LIA-mat assay were equivalent to the Centocor assay. On the basis of
the aforementioned results, we have partitioned the healthy women in
age groups of <45, 45–55, and >55 years. A description of the CA125
values is given in Table 1
. The 95th percentile of the group <45 years is 31 kU/L for the
Byk LIA-mat and 36 kU/L for the Centocor test assay, decreasing to 21
kU/L and 25 kU/L for the group >55 years (P <0.001),
respectively. The 99th percentile of the same age groups showed a
difference of 8 kU/L (P <0.001): 33 kU/L for the Byk
LIA-mat assay and 25 kU/L for the Centocor. The upper limit of normal
(95%) for the whole population should have been 29 kU/L for the Byk
assay and 32 kU/L for the Centocor assay.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 2. Scatter diagram of Byk LIA-mat CA125 II results in a
healthy population (n = 657) related to age.
The central line indicates the predicted line according to
the restricted spline method. The upper and lower
lines show the 95th percentile prediction limits.
|
|
View this table:
[in this window]
[in a new window]
|
Table 1. Comparison of Centocor CA125 II (C) and Byk LIA-mat CA125
II (B) assay results in apparently healthy women stratified by
age.
|
|
Patients.
1) Pregnancy samples (n = 61). A median
serum CA125 value of 19 kU/L was found with a highest value of 80 kU/L
with the Byk LIA-mat (87 kU/L for Centocor). Concentrations >35 kU/L
were present in 13 cases (21%). However, 16 females (26%) had a
concentration >31 kU/L, the 95th percentile for the appropriate age
group. The percentage decreased to 16% (10 of 61) when using the
corresponding 32 kU/L found with the Centocor assay.
2) Benign pelvic diseases (n = 134). For patients with benign
pelvic disease, the Byk assay values are on average lower than the
Centocor results. This group had a median Byk LIA-mat CA125 II
concentration of 22 kU/L with values from 3 to 483 kU/L. Twenty-six
percent of the patients had a marker concentration >35 kU/L. Applying
29 kU/L as a cutoff concentration, 35% of the tested samples were
found to have increased values. There was some evidence of a difference
between the five histological groups of benign ovarian tumors
(P = 0.026). However, there was a rather large
variation observed for the calibrated residual SD. Results of the
nonparametric Kruskal–Wallis test showed no significant difference
(P = 0.07). The median Centocor CA125 II value was 27
kU/L with a range of 5–475 kU/L. Twenty-seven percent of the Centocor
CA125 II values were >35 kU/L, and 37% were above the 95th percentile
value of 32 kU/L.
3) Gynecological malignancies (n = 449). CA125 assay values in
serum from patients with different gynecological malignancies and
different clinical stages of ovarian carcinoma were determined (Table 2
). Very high concentrations of the CA125 antigen mainly occurred
in cases of ovarian carcinomas. Measured with the LIA-mat CA125 II
assay, the median values in the latter group were 116 kU/L for FIGO
stage I/II and 1290 kU/L for stage III/IV, significantly differing from
each other (P <0.001). This difference is still valid after
correction for variations in histology in the different stage groups.
These results hold for the Centocor CA125 II test results as well. A
scatterplot of the CA125 assay values measured in serum from patients
with different gynecological malignancies as determined with the Byk
LIA-mat shows that in nonovarian cancer, concentrations >100 kU/L are
less common and serum assay values >1000 kU/L are rare (Fig. 3
).
View larger version (22K):
[in this window]
[in a new window]
|
Figure 3. Scatter diagram of Byk LIA-mat CA125 II concentration in
serum from patients with benign pelvic disease and gynecological
cancer.
St. 1/2, ovarian cancer stages I and II; St. 3/4, ovarian cancer stages
III and IV.
|
|
roc curves
ROC curves are an objective way to assess the test performance
characteristics of a certain test in a specific population. A
constructed ROC curve from values of healthy controls (group A) vs
values of patients with benign pelvic diseases (group C) is shown in
Fig. 4
. The observed difference in AUC is in favor of the Centocor
assay, 0.789 [95% confidence interval (CI) ± 0.045] vs 0.754 (95%
CI ± 0.049) for the Byk LIA-mat (P <0.01).
Discrimination between patients with benign pelvic disease and patients
with ovarian carcinoma yielded a slightly larger AUC for the Byk method
[0.912 (95% CI ± 0.041) vs 0.904 (95% CI ± 0.045)],
although this difference did not reach significance (Fig. 5
). From the ROC curves, the point at the line having the
shortest distance to the upper left corner (100% sensitivity, 100%
specificity) was calculated. At this concentration the test reaches its
optimal accuracy in terms of discriminating the evaluated patient
groups. In Table 3
the CA125 values with corresponding sensitivity and specificity
are given. Both assays show a highly similar sensitivity and
specificity spectrum. Optimal test accuracy to discriminate between
healthy females and a group of patients with benign pelvic disease was
found at concentrations of 15 and 20 kU/L for the Byk and Centocor
assay, respectively. A specificity rate of 90% is reached at
concentrations of 24 and 27 kU/L, close to the reported reference
concentrations calculated for the healthy controls. The corresponding
sensitivity to find the disease is, however, only ~50%. The ROC
curve constructed on the basis of patients with benign pelvic disease
and those patients with ovarian carcinoma displays a similar pattern
for both assays (Fig. 5
). Optimal test accuracy to discriminate
malignant disease from benign tumors is found at 95 kU/L. This is
virtually corresponding to the 90% specificity point. The two CA125
assays reach a sensitivity of little over 60% at this specificity
concentration. The sensitivity to detect malignancy in this patient
group is increased to 90% at concentrations of 28 and 31 kU/L for the
Byk and Centocor assays, respectively. This will result in almost 40%
of the results being falsely considered positive.
View larger version (18K):
[in this window]
[in a new window]
|
Figure 4. ROC curves for Byk LIA-mat CA125 II (AUC 0.754 ±
0.025 SE) and Centocor CA125 II assays (AUC 0.789 ± 0.023 SE)
from healthy females (n = 652) vs patients with benign pelvic
disease (n = 134).
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Figure 5. ROC curves for Byk LIA-mat CA125 II (AUC 0.912 ±
0.021 SE) and Centocor CA125 II assays (AUC 0.904 ± 0.023 SE)
from patients with ovarian cancer (pretreatment sera, n = 85) vs
patients with benign pelvic disease (n = 134).
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. CA125 concentration at 90% specificity, 90% sensitivity,
and optimal test accuracy, derived from actual ROC
curve.
|
|
|
Discussion
|
|---|
The introduction of a new generation of CA125 assays has led to
several investigations comparing the original homologous two-site
"sandwich" assay based on a single mAb (OC 125) with the previously
developed assays that incorporate the OC 125 and the M11 mAbs
(20)(21). The Bland–Altman difference plot
indicates that the Byk assay reveals a negative deviation at the lower
end of the calibrator range. This bias has implications for the
distribution of values in the various groups studied. In earlier
publications we demonstrated a good overall correlation between all the
CA125 tests studied, but also noted that differences in the lower range
could occur (12)(20). It is customary to use a
95th percentile interval as a reference range (22); this
interval is usually established between the 2.5th and 97.5th
percentiles. However, for tumor markers there are no lower limits for a
healthy population, and therefore the upper reference value is set at
the 95th percentile. In this study the 95th percentile of CA125 serum
values in a group of 652 healthy women used as a control group was 29
kU/L for the Byk LIA-mat and 32 kU/L for the Centocor assay,
concentrations clearly lower than the 35 kU/L given in the package
inserts. The reference value currently in use was introduced by Bast et
al. (2), arbitrarily set from the 99th percentile of test
results for a group of 888 blood donors, including 537 males. A
previous overview of studies reporting reference values showed that
2.5% of >10 000 apparently healthy women had concentrations >35
kU/L (23). In this study we demonstrated nonlinearity in
the distribution of serum CA125 values and a significant relation
between age and CA125 serum concentrations. The 95% prediction limits
of the mean value after logarithmic transformation prompted us to
divide the population of healthy women into three age groups. In the
group of women over age 55 years, none had a serum CA125 concentration
>35 kU/L. This is a strong argument for not applying a common
reference range to the whole female population, as is also suggested by
Bon et al. (24). We find a reference range of 25 kU/L for
the Centocor assay more appropriate for this age group. From our
results one may conclude that each laboratory may consider using the
reference value related to the kit in use and the patient group under
investigation. However, the clinical benefit of such an adjustment
remains doubtful for the main use of CA125 testing, monitoring
treatment. In addition, the reference value of 35 kU/L is not validated
for the CA125 test.
Premenopausal values of CA125 may exceed 100 kU/L. This is in agreement
with publications indicating that increased concentrations of this
antigen may occur in the follicular phase of the cycle, particularly
during menstruation (25). The measurement of CA125
concentrations has been considered of limited value in screening
procedures, the low predictive value of the test being the main
obstacle for its applicability (26). However, most studies
find that the sensitivity and specificity are better in postmenopausal
women as compared with premenopausal women. A careful application of
age-adjusted cutoff concentrations may further improve the
identification of subgroups with a high risk for epithelial ovarian
carcinoma.
The test results of both assays for the group of patients with benign
ovarian tumors are strikingly identical: The Byk assay showed 26% and
the Centocor assay showed 27% with serum concentrations >35 kU/L.
Di-Xia et al. (8) measured CA125 concentrations >35 kU/L
in 35% of their group of patients with benign ovarian tumors,
comparable with our findings when related to the adjusted cutoff
concentrations. However, positivity rates of 5% (2 of 41) and 10% (3
of 31) were reported by Einhorn et al. (27) and Vasilev et
al. (5), respectively. Several reasons can explain this
difference: The various assays used in these studies do not correlate
exactly and the size and composition of the patient groups are
different. Di-Xia et al. described a group of patients with a majority
having cystadenomas, whereas the Einhorn et al. population was not
further characterized. The fact that no significant difference in serum
CA125 concentrations was found between the patients with pelvic disease
and the group with benign ovarian tumors may point to the plausibility
that the increase of CA125 concentrations is not derived from the
release of the antigen from the benign tumor but from the shedding of
CA125 from surrounding irritated peritoneal tissue.
It is evident that further information on the age distribution of
patients and control groups might help to explain the observed
differences. Application of ROC curves could further diminish the
impact of possible disparities in the CA125 assays used. The CA125
assay cannot achieve discrimination between healthy females and the
group with benign pelvic diseases, as the predictive value reaches only
50% at optimal test accuracy concentrations.
The pretreatment serum CA125 assay values in our group of patients with
gynecological malignancies are in agreement with earlier reports. The
serum CA125 assay values of patients with endometrial carcinoma are
increased in 25% of the cases >35 kU/L. The same percentage is
reported by Kenemans et al. (23); moreover, ~30% of
patients with cervical carcinoma were reported to have an increased
serum CA125 concentration. In this study we found increased values in
39% of the pretreatment sera obtained from patients with cervical
cancer, probably due to a higher tumor load in patients classified in
higher FIGO stages. The sensitivity for detecting epithelial ovarian
carcinoma is comparable for both assays if the cutoff concentration is
adjusted to the legitimate 95th percentile. A study to assess the
effect of an age-related reference range to possibly improve the
sensitivity for low-stage disease is required. We calculated optimal
CA125 assay values by ROC curve analysis and correlated optimal
sensitivity and specificity in a given clinical situation. The obtained
results indicate that each assay has a typical cutoff value for
distinguishing specific subgroups of patients. Because CA125 is
frequently used as one of the parameters for diagnosing pelvic masses,
the choice of the cutoff value is important
(8)(28).
In conclusion, the new LIA-mat CA125 II assay has a high
correlation with the Centocor CA125 II IRMA, generally regarded as the
reference assay for the measurement of CA125 in serum. There are,
however, slight differences, and results are not simply transferable.
Therefore, to acquire insight into the performance of the assay in use,
we recommend participation in an external quality-control assessment
scheme.
|
Acknowledgments
|
|---|
We thank D. Linders, J. van Bezu, and H.M. de Feij-de Graaf for
their technical assistance and A. Lansdorp for preparing the
manuscript.
|
Footnotes
|
|---|
1 Nonstandard abbreviations: mAb, monoclonal antibody; FIGO, International Federation of Gynecology and Obstetrics; AUC, area under the curve; and CI, confidence interval. 
|
References
|
|---|
-
Bates SE. Clinical applications of serum tumor markers. Ann Intern Med 1991;115:623-638.
-
Bast RC, Jr, Klug TL, St John E, Jenison E, Niloff JM, Lazarus H, et al. A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. N Engl J Med 1983;309:883-887.
[Abstract]
-
Niloff JM, Bast RC, Schaetzl EM, Knapp RC. Predictive value of CA125 antigen levels in second look procedures for ovarian cancer. Am J Obstet Gynecol 1985;151:981-986.
[Medline]
[Order article via Infotrieve]
-
Barbieri RL, Niloff JM, Bast RC, Schaetzl EM. Elevated serum concentrations of CA125 in patients with advanced endometriosis. Fertil Steril 1986;45:630-634.
[ISI][Medline]
[Order article via Infotrieve]
-
Vasilev SA, Schlaerth JB, Campeau J, Morrow CP. Serum CA125 levels in preoperative evaluation of pelvic masses. Obstet Gynecol 1988;71:751-756.
[Abstract/Free Full Text]
-
Hopman EH, Helmerhorst ThJM, Bonfrer JMG, Ten Bokkel Huinink WW. Highly elevated serum CA125 levels in patients with cardiac failure. Eur J Obstet Gynecol Reprod Biol 1993;48:71-73.
[Medline]
[Order article via Infotrieve]
-
Pavesi F, Lotzniker M, Marbello L, Garbagnolo P, Franchi M, Moratti R. CA125 in benign and malignant diseases involving coelomic epithelium. J Tumor Marker Oncol 1988;3:49-57.
-
Di-Xia C, Schwartz P, Xinguo L, Zhan Y. Evaluation of CA125 levels in differentiating malignant from benign tumors in patients with pelvic masses. Obstet Gynecol 1988;72:23-27.
[Abstract/Free Full Text]
-
Kaesemann H, Caffier H, Hoffmann FJ, Crombach G, Würz H, Kreienberg R, et al. Monoclonale Antikorper in Diagnostik und Verlaufskontrolle des Ovarialkarzinoms. CA125 als Tumormarker. Klin Wochenschr 1986;64:781-785.
[Medline]
[Order article via Infotrieve]
-
Bast RC, Jr, Feeney M, Lazarus H, Nadler LM, Colvin RB, Knapp RC. Reactivity of a monoclonal antibody with human ovarian carcinoma. J Clin Invest 1981;68:1331-1337.
-
Hardardottir H, Parmley TH, Quirk JG, Sanders MM, Miller FC, O'Brien TJ. Distribution of CA125 in embryonic tissues and adult derivatives of the fetal periderm. Am J Obstet Gynecol 1990;163:1925-1931.
[ISI][Medline]
[Order article via Infotrieve]
-
Bonfrer JMG, Baan AW, Jansen E, Lentfer D, Kenemans P. Technical evaluation of three second generation CA125 assays. Eur J Clin Chem Clin Biochem 1994;32:201-207.
[ISI][Medline]
[Order article via Infotrieve]
-
Wood WG. Automation of luminescence immunoassays with
coated tubes using the LIA-mat 300 system as example. In: Klapdor R,
ed. Tumor associated antigens, oncogenes, receptors, cytokines in tumor
diagnosis and therapy at the beginning of the nineties. Cancer of the
breast—state and trends in diagnosis and therapy. München:
Zuckschwerdt, 1992:348–50..
-
Shepard JH. Revised FIGO staging for gynaecological cancer. Br J Obstet Gynaecol 1989;96:889-892.
[ISI][Medline]
[Order article via Infotrieve]
-
Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med 1989;8:551-562.
[ISI][Medline]
[Order article via Infotrieve]
-
Royston P. Estimating departure from normality. Stat Med 1991;10:1283-1294.
[ISI][Medline]
[Order article via Infotrieve]
-
Zweig MH, Beck JR, Collinsworth WL, Donlon JA, King JS, Krieg AF, et al. Assessment of clinical sensitivity and specificity of laboratory tests; proposed guideline 1987:143-175 NCCLS publication GP10-P Villanova, PA. .
-
DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-846.
[ISI][Medline]
[Order article via Infotrieve]
-
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307-310.
-
Kenemans P, Van Kamp GJ, Oehr P, Verstraeten RA. Heterologous double-determinant immunoradiometric assay CA125 II: reliable second-generation immunoassay for determining CA125 in serum. Clin Chem 1993;39:2509-2513.
[Abstract]
-
Oehr P, Liu Q, Girardin P. Comparison of three CA125 II tumour marker assays. Tumourdiagn Ther 1993;14:39-40.
-
Solberg HE. Establishment and use of reference values. Burtis CA Ashwood ER eds. Tietz textbook of clinical chemistry 1989:454-484 WB Saunders Philadelphia. .
-
Kenemans P, Yedema CA, Bon GG, Von Mensdorff-Pouilly S. CA125 in gynecological pathology—a review. Eur J Obstet Gynecol Reprod Biol 1993;49:115-124.
[ISI][Medline]
[Order article via Infotrieve]
-
Bon GG, Kenemans P, Verstreaten R, van Kamp GJ, Hilgers J. Serum tumor marker immunoassays in gynecologic oncology: establishment of reference values. Am J Obstet Gynecol 1996;174:107-114.
[ISI][Medline]
[Order article via Infotrieve]
-
Jäger W, Meier C, Wildt L, Sauerbrei W, Lang N. CA125 serum concentrations during menstruation cycle. Fertil Steril 1988;50:223-227.
[ISI][Medline]
[Order article via Infotrieve]
-
Helzlsouer KJ, Bush TL, Alberg AJ, Miller Bass K, Zacur H, Comstock GW. Prospective study of serum CA125 levels as markers of ovarian cancer. JAMA 1993;269:1123-1126.
[Abstract]
-
Einhorn N, Bast RC, Jr, Knapp RC, Tjernberg B, Zurawski VR. Preoperative evaluation of serum CA125 levels in patients with primary epithelial ovarian cancer. Obstet Gynecol 1986;67:414-416.
[ISI][Medline]
[Order article via Infotrieve]
-
Schutter EMJ, Kenemans P, Sohn C, Kristen P, Crombach G, Westermann R, et al. Diagnostic value of pelvic examination, ultrasound, and serum CA125 in postmenopausal women with pelvic mass. An international multicenter study. Cancer 1994;74:1398-1406.
[ISI][Medline]
[Order article via Infotrieve]
The following articles in journals at HighWire Press have cited this article:
|
|
|
|
 
A. Santillan, R. Garg, M. L. Zahurak, G. J. Gardner, R. L. Giuntoli II, D. K. Armstrong, and R. E. Bristow
Risk of Epithelial Ovarian Cancer Recurrence in Patients With Rising Serum CA-125 Levels Within the Normal Range
J. Clin. Oncol.,
December 20, 2005;
23(36):
9338 - 9343.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
