Clin Chem Lab Med 2012;50(12):21892198 2012 by Walter de Gruyter Berlin Boston. DOI 10.

1515/cclm-2011-0757
Human epididymis protein 4 as a serum marker for diagnosis of
endometrial carcinoma and prediction of clinical outcome
Laura Zanotti
1
, Eliana Bignotti
1
, Stefano Calza
2
,
Elisabetta Bandiera
1
, Giuseppina Ruggeri
3
, Claudio
Galli
4
, Germana Tognon
5
, Monica Ragnoli
5
, Chiara
Romani
1
, Renata A. Tassi
1
, Luigi Caimi
3
, Franco E.
Odicino
5
, Enrico Sartori
5
, Sergio Pecorelli
1, 5
and
Antonella Ravaggi
1,
*

1
Angelo Nocivelli Institute of Molecular Medicine ,
Division of Gynaecologic Oncology, University of Brescia,
Brescia , Italy

2
Unit of Medical Statistics , Department of Biomedical
Sciences and Biotechnology, University of Brescia, Brescia ,
Italy

3
Department of Laboratory Medicine , A.O. Spedali Civili,
Brescia , Italy

4
Scientic Affairs , Abbott Diagnostics, Rome , Italy

5
Department of Obstetrics and Gynaecology , University of
Brescia, Brescia , Italy
Abstract
Background: The purpose of this study was to assess the
diagnostic and prognostic impact of preoperative serum
determination of human epididymis protein 4 (sHE4), and
to investigate its potential correlation with clinicopathologi-
cal features and survival endpoints in endometrial cancer
patients.
Methods: Preoperative serum samples from 193 endometrial
cancer patients and 125 women with normal endometrium
were measured for sHE4 and serum CA125 (sCA125) con-
centrations by quantitative chemiluminescent microparticle
immunoassays on the automated Architect instrument.
Results: sHE4 concentrations were signicantly higher in
endometrial cancer patients regardless of tumour stage and
grade compared with normal controls. Setting the speci-
city at 95 % , the sensitivities in detecting endometrial cancer
patients were 66 % for HE4, 33 % for CA125 and 64 % for the
combination of the two markers. High concentrations of both
HE4 and CA125 signicantly correlated with all clinicopatho-
logical features characterising a more aggressive tumour phe-
notype. In multivariate analysis, only high preoperative sHE4
concentrations, but not sCA125, were independent prognostic
factors for shorter Overall Survival, Disease-Free Survival
and Progression-Free Survival.
Conclusions: HE4 is more sensitive and specic than CA125
in distinguishing endometrial cancer patients from women
with normal endometrium, regardless of tumour stage and
grade. sHE4 appears to be associated with a more aggres-
sive tumour variant and it could be clinically useful, in iden-
tifying high-risk endometrial cancer patients, for a tailored
surgical and postoperative therapy. HE4 signicant correla-
tion with decreased Overall Survival, Disease Free Survival
and Progression Free Survival suggests its potential role as a
novel prognostic marker for endometrial cancer.
Keywords: biomarkers; CA125; diagnosis; endometrial
cancer; human epididymis protein 4 (HE4); prognosis.
Introduction
Endometrial cancer (EC) is the most commonly diagnosed
gynaecologic tumour in Europe and US and the fourth most
common cancer following breast, lung and colorectal cancer
in women (1) . In the US, 43,470 new cases and 7950 deaths
were estimated in 2010 (2) .
Although usually detected in early, more favourable stages,
the tumours identied at late stage are associated with high lev-
els of morbidity and mortality: 5-year survival rates for loca-
lised, regional, and metastatic disease are 96 % , 67 % and 17 % ,
respectively (2) . The early appearance of symptoms, such as
postmenopausal bleeding, allows that about 70 % of cases are
diagnosed at rst stage (2) . Current diagnostic approaches rely
on EC sampling of macroscopic disease that is driven by the
presence of symptoms (3) , even if numerous studies have shown
that ultrasound is at least as sensitive as endometrial biopsy (4) .
The need for tumour markers to detect early-stage EC is
critical for asymptomatic patients in particular of high-risk
categories, such as women who have a familial association
with Lynch syndrome, hereditary non-polyposis colo rectal
cancer, obesity, diabetes, for breast cancer patients on tamox-
ifen treatment (3) or for patients submitted to other medical
therapies that can promote the endometrial thickening.
Serum analysis is a low cost, non-invasive technique and it
is not subjected to operator variability.
At present, there are no serum markers routinely used in
clinical practice for early detection of EC. CA15.3, YKL-40,
CA125 and hPRL have already been investigated and have
been found elevated in EC patients serum. However, these
markers have not reached suitable levels of sensitivity and
specicity to be used for EC diagnosis (5 8) .
*Corresponding author: Dr. Antonella Ravaggi, Istituto di Medicina
Molecolare, A. Nocivelli , Piazzale Spedali Civili 1, 25123
Brescia, Italia
Phone: + 39 0303996284, Fax: + 39 0303996059 ,
E-mail: ravaggi@med.unibs.it
Received October 18, 2011; accepted March 8, 2012;
previously published online April 4, 2012
2190 Zanotti et al.: Serum HE4 detection in endometrial carcinoma
Even if the majority of EC patients have a favourable
outcome, a subset of endometrial tumours exhibits a bio-
logically aggressive phenotype, deep myometrial invasion,
grade 3 disease, lymph node metastasis or high-risk cellular
histological types, such as serous or clear cell, characterised
by poor prognosis (9) . For this reason, it is of the utmost
importance to discover new potential prognostic biomark-
ers that may allow stratication of patients into meaningful
prognostic categories, with the ultimate goal of improving
treatment strategies. Moreover, patients characterised by
poor prognosis are at increased risk of recurrence and/or
metastasis (3, 10) . To date, most EC recurrences are detected
by imaging techniques or clinical evaluation, because there
are no reliable serum biomarkers used in clinical practice
for regular monitoring of patients during follow-up. Serum
CA125 (sCA125) has limited utility in monitoring response
to chemotherapy and may not predict recurrence in absence
of other clinical signs, because its increase can be observed
in only 50 % 63 % of patients with recurrent/progressive dis-
ease (11) . Therefore, sensitive tumour markers for monitor-
ing response to therapy and for early detection of recurrence
are still needed.
Among newly considered biomarkers, one of the most
promising is human epididymis protein 4 (HE4). HE4, also
known as WFDC2, because it contains two whey acidic pro-
tein and four disulde core, was rst described by Kirchoff
et al. (12) in human epididymal tissue and was subsequently
found expressed in many other normal tissues, particularly of
the reproductive tracts and of the central respiratory airways,
as well as in ovarian carcinoma biopsies (13, 14) .
Helstrom in 2003 (15) and then many other groups have
demonstrated the efcacy of HE4 as a serum marker for ova-
rian cancer and its superiority over CA125 especially in terms
of specicity.
However, few studies have detailed HE4 serum (sHE4)
concentrations in endometrial cancer. Moore and colleagues
(16) have examined multiple markers among which HE4
showed the highest sensitivity in detecting EC regardless of
stage; a similar superiority of HE4 compared to CA125 in
discriminating EC patients from healthy controls has been
found by other research groups (17, 18) . Moreover, a recent
study from our group (19) demonstrated that HE4 gene and
protein were signicantly upregulated in EC compared to
healthy controls with normal endometrium (NE) and that
high sHE4 concentrations were signicantly associated with
worse outcome only in poorly differentiated EC patients.
The correlation between a more aggressive EC phenotype
and this serum marker has been described also by another
group (20) .
To date, the most of the studies about HE4 in EC patients
have been carried out with a conventional enzyme immuno-
assay (EIA) (16 20) , but recently a fully automated chemilu-
minescent immunoassay (CMIA) has been made available. In
our recent paper (21) we have evaluated this new HE4 assay
in comparison with the conventional EIA in ovarian can-
cer and we have demonstrated a good correlation and simi-
lar clinical value of the two assays, with a greater precision
for CMIA.
In this report, with the aim to complete our previous work
(19) , taking advantage of this new automated method for
HE4 detection, we have evaluated sHE4 concentrations in a
larger cohort of EC patients and controls. We have compared
serum concentrations of CA125 and HE4 in patients with
EC and in women with normal endometrium (NE), to assess
their sensitivity and specicity, alone and in combination.
Moreover, we have correlated sHE4 and sCA125 concen-
trations with patients clinicopathologic characteristics and
survival endpoints to determine their prognostic potential.
Materials and methods
Patients and samples
This retrospective study included 193 patients with EC referred to
the Gynaecologic Oncology Department of the University of Brescia
from 2003 to September 2010. The study also recruited 125 healthy
volunteers, without any benign or malignant gynaecologic patho logy,
as healthy controls (NE). The absence of gynaecological disease was
assessed by anamnesis, gynaecological examination, transvaginal
pelvic ultrasound and pap smear. All subjects with a past or con-
comitant history of malignancy were excluded from the study. All
enrolled women provided informed consent prior to the collection
of any blood samples; the study was approved by the Institutional
Review Board. Blood samples were collected from patients the day
before surgery; sera were separated by centrifugation within 30 min
and then stored at 80 C until analysis.
For healthy controls, age and menopausal status were recorded.
EC patients charts were reviewed to note age, menopausal status,
stage, grade, treatment and follow-up information. Endometrial can-
cer patients were staged in accordance with International Federation
of Gynecologists and Obstetricians (FIGO) guidelines issued in
1998. None of the patients had received preoperative chemotherapy
or radiation.
Patients and controls were fairly balanced as regards menopausal
status and age.
EC patients clinicopathologic characteristics are summarised in
Table 1 . The median age of EC patients was 66 years (range 28 92).
The median age of NE controls was 61 years (range 35 76): 24
(19.2 % ) premenopausal and 101 (80.8 % ) postmenopausal women.
Patients were followed from the date of surgery until death or the
last observation in November 2010.
HE4 and CA125 serum concentrations measurements
Preoperative sHE4 concentrations were measured using the al-
ready described (21) CMIA assay (Abbott Diagnostics Division,
Wiesbaden, Germany) on the fully automated Architect instrument
(Abbott Diagnostics Division). The dynamic range of HE4 detection
goes from 20 to 1500 p M with an automated 1:10 dilution protocol
that extends the linear range up to 15,000 p M . The intra-assay and
total imprecision (CV % ) of the CMIA HE4 assay ranged from 2.11
to 2.93 and from 3.13 to 3.70 depending on the concentrations of the
assays positive controls (21) .
Preoperative sCA125 values were also determined by CMIA
assay (Abbott Diagnostics Division) on Architect. The assay
is linear up to 1000 U/mL and has a normality threshold at 35
U/mL (22) . Testing was carried out at the III Laboratory Service,
Spedali Civili di Brescia, Italy, following the manufacturer s
recommendations.
Zanotti et al.: Serum HE4 detection in endometrial carcinoma 2191
Table 1 sCA125 and sHE4 preoperative levels in EC patients divided according to clinicopathological features.
Variable n, % CA125, U/mL HE4, p M
Median Range 95th perc p
a
-Value Median Range 95th perc p
a
-Value
Age at diagnosis
< 65 82 (42.5) 15.9 3.3 212.3 82.4 60 6.5 541.9 158.8
65 111 (57.5) 18.2 3.9 961.0 109.4 0.134 93 19.4 1348.4 431.7 < 0.01
Menopause
Negative 18 (9.3) 22.1 6.8 96.7 81.2 66.6 28.0 541.9 218.6
Positive 175 (90.7) 17.2 3.3 961.0 93.0 0.025 81.6 6.5 1348.4 301.8 < 0.01
WHO grading
G1 35 (18.1) 16.7 5.2 54.7 42.7 63.8 33.0 304.3 160.1
G2 89 (46.1) 17.8 3.3 497.4 88.6 88.4 6.5 1348.4 439.3
G3 69 (35.8) 17.1 3.9 961.0 137.8 0.45 81.6 31.7 541.9 283.0 < 0.01
FIGO stage
I 106 (54.9) 16.2 3.3 138.6 44.3 65.6 6.5 572.4 191.0
II 34 (17.6) 15 6.7 120.2 54.7 93 34.5 1348.4 304.3
III 31 (16.1) 33.3 3.9 497.4 203.8 97.2 50.5 639.4 400.5
IV 9 (4.7) 82.8 22.9 961.0 615.3 < 0.01 143.6 50.5 541.9 537.9 < 0.01
Unknown 13 (6.7)
Histology
Endometrioid 152 (78.7) 16.9 3.3 497.4 74.8 78.8 6.5 1348.4 321.4
Mixed 16 (8.3) 22.8 6.0 961.0 399.5 95.2 40.8 532.0 361.2
Indifferentiated 9 (4.7) 38.1 6.7 165.2 128.5 81.6 45.9 154.2 137.2
Clear cell 9 (4.7) 12.7 8.1 82.8 60.1 61.5 51.0 184.3 145.0
Serous 6 (3.1) 25.3 4.7 78.5 67.6 59.2 31.7 168.0 154.1
Mucinous 1 (0.5) 54.7 0.195 118.4 0.638
Myometrial invasion
M0 15 (7.8) 16.9 4.7 71.2 59.7 57.3 6.5 118.4 105.0
M1 81 (42.0) 14.5 3.3 195.2 42.7 66.2 31.7 1348.4 220.4
M2 97 (50.2) 20.5 3.7 961.0 143.9 < 0.01 93 19.4 639.4 506.6 < 0.01
Lymph node status
Negative 136 (70.5) 16.2 3.3 195.2 44.0 70.5 6.5 537.2 221.2
Positive 26 (13.5) 38.1 3.9 25.7 22.9 < 0.01 95.1 50.5 532.0 450.4 < 0.01
Unknown 31 (16.0)
Cervical involvement
Absent 124 (64.3) 16.9 3.3 138.6 71.6 69.1 6.5 572.4 194.6
Glandular involvement 26 (13.5) 16.1 4.6 120.2 62.3 90.5 36.6 300.7 222.7
Stromal involvement 41 (21.2) 19.5 7.1 961.0 196.1 0.026 110.6 34.5 1348.4 355.5 < 0.01
Unknown 2 (1.0)
Adnexal involvement
Negative 170 (88.1) 16.7 3.3 194.5 71.5 76.4 6.5 1348.4 294.0
Positive 22 (11.4) 47.6 9.4 961.0 483.1 < 0.01 95.1 50.5 541.9 519.3 < 0.01
Unknown 1 (0.5)
Peritoneal cytology
Negative 173 (89.6) 16.9 3.3 497.4 72.4 76.7 6.5 1348.4 291.8
Positive 14 (7.3) 47.7 7.1 961.0 443.7 < 0.01 97.7 50.5 541.9 535.5 0.011
Unknown 6 (3.1)
Lymphovascular invasion
Absent 83 (43.0) 13.8 3.3 194.5 62.8 66.2 6.5 500.3 178.8
Present 102 (52.9) 22.1 3.9 961.0 96.7 < 0.01 90.5 19.4 1348.4 522.7 < 0.01
Unknown 8 (4.1)

a
p-Values were calculated using non-parametric Wilcoxon-Mann-Whitney or non-parametric ANOVA. Unknown levels were not considered
in the test procedures.
Statistical analysis
Differences in sHE4 and sCA125 concentrations between the groups
were calculated using parametric ANOVA. Spearman rank correla-
tion was used to evaluate association between sHE4 and sCA125 val-
ues. Combination of markers was investigated comparing receiver
operating characteristic curves (ROC) based on the method pro-
posed by DeLong et al. 1988 (23) . Different ROC are built based on
predicted values for each sample derived from logistic models ac-
counting for different combinations of markers. In order to evaluate
generalisability of the model, predicted values are computed using
leave-one-out cross validation. Briey, a logistic model is tted
2192 Zanotti et al.: Serum HE4 detection in endometrial carcinoma
using a combination of markers on every sample but one. The tted
model is then used to predict the probability of the event on the ex-
cluded sample. This is performed on every sample and the resulting
predicted probability is used as marker value.
The association between sHE4 or sCA125 concentrations
and clinicopathologic parameters was investigated using either
Wilcoxon-Mann-Whitney test or non-parametric ANOVA followed
by post-hoc Nemenyi-Damico-Wolfe-Dunn tests. For survival analy-
sis, three endpoints (cancer relapse, cancer progression and cancer
related death) were used to calculate Disease-Free Survival (DFS),
Progression-Free Survival (PFS) and Overall Survival (OS), respec-
tively. DFS was dened as the time interval between the date of sur-
gery and the date of identication of disease recurrence; PFS was
dened as the time interval between the date of surgery and the date
of identication of progressive disease (disease not treatable with
curative intent) and OS was dened as the time interval between the
date of surgery and the date of death. For all three endpoints the last
date of follow-up was used for censored subjects. Survival models
were tted using the Cox proportional hazard models, while survival
curves were drawn based on the Kaplan-Meier methods. All the mod-
els assumptions were tested. The effect of both sHE4 and sCA125
concentrations on prognosis was evaluated categorising them on the
basis of the tertiles, computed on the whole cohort (low; medium;
high). In all analyses, a p-value < 0.05 was considered signicant and
the test was two-sided. All the statistical analyses were performed
using R (R Development Core Team, 2010).
Results
HE4 and CA125 serum concentrations
HE4 serum concentration was signicantly higher in EC
patients (median = 79.9 p M , range 6.5 1348.4), regardless of
tumour stage and differentiation grade, compared with NE
(median = 39.0 p M , range 21.6 84.5, p < 0.0001) (Figure 1 A, B).
CA125 serum concentrations were signicantly higher in
EC patients (median = 17.2 U/mL, range 3.3 961.0), regardless
of tumour stage and differentiation grade, compared with NE
(median = 10.8 U/mL, range 2.7 53.3, p < 0.001) (Figure 1C, D).
ROC curves for sHE4 and sCA125 in endometrial
cancer patients
The areas under the ROC curves (AUC) were determined for
HE4, CA125 and in combination, in differentiating EC from
3.0
2.7
A B
C D
2.4
2.1
1.8
1.5
L
o
g

H
E
4
,

p
M
L
o
g

C
A
1
2
5
,

U
/
m
l
L
o
g

C
A
1
2
5
,

U
/
m
l
L
o
g

H
E
4
,

p
M
1.2
0.9
0.6
0.3
0
NE I II III IV
Stage EC
G1 NE
G2 G3
EC
Median
Mean
DS
39
41.4
11.5
65.6
81.9
65.4
93
117.9
76.1
143.6
219.8
193.1
97.2
146.9
129.2
NE I II III IV
Stage EC
Median
Mean
DS
10.8
12.1
7.1
16.2
20.3
17.4
15.0
19.2
11.4
82.8
163.3
300.6
33.3
68.5
99.1
Median
Mean
DS
39
41.4
11.5
63.8
80.4
52.9
88.4
132.6
174.4
81.6
112.7
95.9
G1 NE
G2 G3
EC
Median
Mean
DS
10.8
12.1
7.1
16.7
20.6
12.9
17.8
33.6
58.3
17.1
45.3
118.9
3.0
2.7
2.4
2.1
1.8
1.5
1.2
0.9
0.6
0.3
0
3.0
2.7
2.4
2.1
1.8
1.5
1.2
0.9
0.6
0.3
0
4.0
3.6
3.2
2.8
2.4
2.0
1.6
1.2
0.8
0.4
0
Figure 1 Box plots showing sHE4 (AB) and sCA125 (CD) concentrations in NE controls and EC patients, represented according to FIGO
stage (AC) and differentiation grade (BD).
Zanotti et al.: Serum HE4 detection in endometrial carcinoma 2193
NE. The AUC comparison for individual markers revealed
that sHE4 had a signicantly higher AUC relative to sCA125
in all EC (p < 0.0001) and in each subset of EC analysed (all
p < 0.05) (Figure 2 ).
Using a cut-off value of 51 p M , sensitivity and speci-
city of sHE4 in discriminating EC from NE were 78.8 % and
84.8 % , respectively, with an accuracy of 81.1 % . Regarding
sCA125, setting the cut-off value at 12.7 U/mL, the sensi-
tivity and specicity were 71.5 % and 65.6 % , respectively,
with an accuracy of 69.2 % . In both cases, the cut-off was
chosen as the value of the curve that minimises the dis-
tance to the upper left corner. Even considering only the
subpopulation of stage I EC, the diagnostic performance of
HE4 remains higher when compared with that of CA125.
Using the cut-off of 51 p M , HE4 was able to detect 67.9 %
of stage I patients (accuracy of 77.6 % ), while the sensiti vity
of CA125 was 66.1 % (accuracy of 65.8 % ) at the cut-off
value of 12.7 U/mL.
The correlation between sCA125 and sHE4 concentrations
was rather small (Spearman rs = 0.38, p < 0.01), therefore a
combination of the two markers was analysed. AUC deriving
from the combination of HE4 and CA125 was signicantly
increased when compared to CA125 (p < 0.01), but not to HE4
(p = 0.2533), regardless of tumour stage and differentiation
grade.
Then we compared the sensitivity of each tumour marker
for the discrimination of EC from NE at specicity of 90 % ,
95 % and 99 % (Table 2 ). HE4 showed a notably higher sen-
sitivity in detecting EC compared to CA125. For all EC
stages the combination of CA125 and HE4 showed 40 % ,
29 % , 36 % improvement in sensitivity over that of CA125
alone, and 4 % , 2 % , 11 % improvement over that of HE4
alone.
HE4 and CA125 serum concentrations and
clinicopathologic variables
As represented in Table 1, both high sHE4 and sCA125 con-
centrations were signicantly correlated with menopausal
status, higher stage (I vs. > I), deeper myometrial invasion
(M0/M1 vs. M2), lymph node metastasis, presence of cervical
invasion, adnexal involvement, positive peritoneal cytology
and lymphovascular invasion. Moreover, high sHE4 concen-
trations signicantly correlated also with age 65 and higher
grade (G1 vs. G2/G3). Neither sHE4 nor sCA125 correlated
with tumour histotype.
HE4 and CA125 serum concentrations and clinical
outcome
For survival analysis, only 190 EC patients were considered,
because three patients were lost to follow-up. Patients were
followed from the date of surgery until death or the last obser-
vation (median follow-up, 31 months, range 2 86 months). At
the time of the last follow-up, 156 patients (82.1 % ) were alive
without evidence of disease, seven (3.7 % ) were alive with
disease, 24 (12.6 % ) were dead of disease and three (1.6 % )
died for other causes.
In the univariate analysis for OS, DFS and PFS, high pre-
operative sHE4 concentrations (high vs. low and medium
tertiles) were signicantly associated to worse prognosis
as well as the other traditional prognostic factors, such as
FIGO stage, histology, myometrial invasion, lymphovascu-
lar invasion, cervical and adnexal involvement, lymph node
metastasis, positive peritoneal cytology (Table 3 ). High pre-
operative sCA125 concentrations (high vs. low and medium
tertiles) were signicantly associated to shorter OS and PFS
(Table 3).
In the multivariate analysis, only high preoperative sHE4
concentrations and not high preoperative sCA125 concentra-
tions were demonstrated to be independent prognostic factors
for poor OS, shorter DFS and PFS (Table 3 and Figure 3 ).
Regarding conventional prognostic factors, only FIGO stage
showed an independent association with OS (III-IV vs. I-II;
p = 0.034) and PFS (p = 0.050), but not with DFS (p = 0.471).
Differentiation grade and histological type were not found
signicantly correlated with anyone of the three survival
endpoints.
1.0
0.8
0.6
S
e
n
s
i
t
i
v
i
t
y
0.4
0.2
0.0
0.0 0.2 0.4 0.6
1-Specificity
EC vs. NE
EC stage I vs. NE
EC stage>I vs. NE
0.8
Marker AUC
CA125
HE4
CA125+HE4
0.7247
0.8834
0.8952
Marker AUC
CA125
HE4
CA125+HE4
0.6774
0.8227
0.8409
Marker AUC
CA125
HE4
CA125+HE4
0.7837
0.9495
0.9528
1.0
1.0
0.8
0.6
S
e
n
s
i
t
i
v
i
t
y
0.4
0.2
0.0
0.0 0.2 0.4 0.6
1-Specificity
0.8 1.0
1.0
0.8
0.6
S
e
n
s
i
t
i
v
i
t
y
0.4
0.2
0.0
0.0 0.2 0.4 0.6
1-Specificity
0.8 1.0
Figure 2 Receiver operating characteristics (ROC) curves for HE4, CA125 and CA125 + HE4 combination, for differentiation of normal
endometrium controls (NE) and all endometrial cancer patients (EC) (A), stage I EC (B) and stage > I EC (C).
AUC, area under the curve.
2194 Zanotti et al.: Serum HE4 detection in endometrial carcinoma
T
a
b
l
e

2

S
e
n
s
i
t
i
v
i
t
i
e
s

(
%
)

f
o
r

i
n
d
i
v
i
d
u
a
l

a
n
d

c
o
m
b
i
n
e
d

b
i
o
m
a
r
k
e
r
s

a
t

v
a
r
i
o
u
s

s
p
e
c
i

c
i
t
y

l
e
v
e
l
s
:

N
E

v
s
.

E
C
,

N
E

v
s
.

s
u
b
s
e
t
s

o
f

E
C
.

N
E

v
s
.

a
l
l

E
C
S
p
e
c
i

c
i
t
y
,

%

N
E

v
s
.

s
t
a
g
e

I

E
C
S
p
e
c
i

c
i
t
y
,

%

N
E

v
s
.

G
1

E
C
S
p
e
c
i

c
i
t
y
,

%

N
E

v
s
.

s
t
a
g
e
>
I

E
C
S
p
e
c
i

c
i
t
y
,

%

N
E

v
s
.

>
G
1

E
C
S
p
e
c
i

c
i
t
y
,

%

9
0
9
5
9
9
9
0
9
5
9
9
9
0
9
5
9
9
9
0
9
5
9
9
9
0
9
5
9
9
C
A
1
2
5
3
9
3
5
2
1
3
1
2
6
1
1
4
0
3
1

9
5
0
4
6
2
8
3
9
3
3
1
9

C
u
t
-
o
f
f
,

U
/
m
L
2
0
.
6
2
4
.
8
3
7
.
1
2
0
.
3
2
4
.
8
3
7
.
1
2
1
.
1
2
5
.
9
4
2
.
6
2
0
.
3
2
5
.
7
4
2
.
6
2
0
.
6
2
5
.
7
4
2
.
7
H
E
4
7
4
6
6
4
6
5
7
5
3
3
2
6
0
5
1
2
9
9
1
8
2
6
5
7
7
6
9
5
0

C
u
t
-
o
f
f
,

m
M

5
5
.
7
6
3
.
5
8
5
.
0
5
7
.
3
6
3
.
5
8
5
.
0
5
7
.
3
6
3
.
8
8
4
.
5
5
6
.
9
6
3
.
0
8
4
.
5
5
5
.
7
6
3
.
0
8
5
.
0
C
A
1
2
5
+
H
E
4
7
9
6
4
5
7
6
8
5
2
4
2
6
0
4
6
4
0
9
3
7
7
6
9
8
2
6
9
6
1

C
u
t
-
o
f
f
,

U
/
m
L
;

m
M

1
1
.
7
;

5
5
.
7
3
3
.
6
;

5
0
.
5
1
5
.
9
;

7
6
.
8
1
1
.
7
;

5
5
.
7
1
2
.
3
;

6
8
.
8
4
2
.
7
;

5
3
.
1
2
1
.
8
;

4
7
.
8
2
1
.
1
;

6
2
.
7

5
.
2
;

9
3
.
1
1
5
.
4
;

5
4
.
0
2
5
.
7
;

5
9
.
5

7
.
1
;

8
8
.
5
1
2
.
3
;

5
5
.
7
3
3
.
6
;

5
0
.
5
1
5
.
9
;

7
6
.
8
Discussion
Endometrial cancer is the most common malignant neoplasm
of the female genital tract (2, 24) . Although most EC patients
often exhibit symptoms that lead to an early diagnosis, both
annual incidence and death rate are rising (24) . The need to
identify diagnostic biomarkers for this neoplasm is essential
especially for patients at high risk (3) .
Up to now, the main diagnostic approach is endometrial
sampling after the presentation of symptoms (3, 25) and at
present, there are no early detection serum biomarkers for EC
to be used in the diagnosis of preclinical disease or in the
screening of women who are at high EC risk (11) .
Several serum tumour markers have been investigated
for their potential diagnostic use, both individually and in
combination, but they have not reached acceptable diagnos-
tic performances (5 7, 26, 27) . The most commonly used
tumour marker in the clinical management of EC is CA125,
even if it has been reported to be elevated in only 11 % 34 %
of patients and particularly to exhibit a low sensitivity in
detection of early stage (11) . Accordingly, several studies
have shown that sCA125 concentrations correlate only with
advanced stages and with presence of extrauterine disease
(28, 29) .
HE4 has been recently identied as a new candidate marker
for ovarian cancer. Elevated HE4 concentrations were found
in sera of patients with ovarian tumours and HE4 showed
a similar sensitivity to CA125, but an increased specicity
in discriminating malignancies from benign masses (15, 17,
18) . Galgano et al. reported an increased tissue expression of
HE4 in various types of human carcinomas (30) and Bignotti
et al. conrmed a signicantly stronger HE4 immunohis-
tochemical staining in EC samples compared with NE tissues
(19) . Only few recent studies have proposed HE4 as a serum
marker for endometrial malignancies (16, 17, 19, 20) .
In the present study, as a continuation of our previous work
(19) , HE4 and CA125 serum secretions were analysed in
patients with EC and NE. We used the new automated HE4
chemiluminescent immunoassay to investigate the diagnostic
and prognostic performance of this new biomarker compared
to CA125.
Our data were consistent with a previous paper (16) indi-
cating that sHE4 has a signicantly higher AUC compared
to sCA125 in differentiating EC patients from NE, regard-
less of tumour stage and grade. At 95 % specicity, the sen-
sitivities for discrimination between controls and all stages
of EC were 66 % for HE4 and 35 % for CA125, similarly to
values obtained with a smaller series of patients and with
the manual EIA detection kit (19) . If the evaluation was lim-
ited in stage I EC, HE4 retained a sensitivity of 53 % , while
CA125 only of 26 % . In fact, in agreement with Powell et al.,
CA125 was elevated only in advanced stages ( III) and high
grade EC (28) .
The best combinations of sensitivity and specicity for
HE4 and CA125 were obtained with the cut-off of 51 p M
and 12.7 U/mL, respectively. The specicity of HE4 for dis-
criminating EC from NE was 84.8 % and the sensitivity for
detecting all EC or only stage I EC patients was 78.8 % and
Zanotti et al.: Serum HE4 detection in endometrial carcinoma 2195
Table 3 Univariate and multivariate survival analyses in relation to HE4 and clinical parameters.
Variables OS DFS PFS
HR 95 % CI p-Value HR 95 % CI p-Value HR 95 % CI p-Value
Univariate analysis
Preoperative HE4 serum level
High vs. medium and low 3.43 1.50 7.84 0.003 2.45 1.15 5.21 0.020 3.31 1.43 7.66 0.005
Age
65 vs. < 65 1.17 0.51 2.67 0.715 1.70 0.74 3.88 0.209 1.14 0.50 2.64 0.754
Menopause
Yes vs. no 1.20 0.28 5.12 0.802 1.42 0.34 6.01 0.630 1.14 0.27 4.86 0.860
FIGO stage
III and IV vs. I and II 13.70 5.64 33 28 0.000 4.80 2.22 10.39 0.000 13.18 5.37 32.37 0.000
WHO grading
G3 and G2 vs. G1 1606.96 0 5 10
17
0.642 1665.66 0 3 10
16
0.609 1603.65 0 1 10
18
0.650
Histology
Non-endometrioid vs.
endometrioid
2.68 1.19 6.03 0.018 2.38 1.09 5.21 0.030 2.45 1.06 5.66 0.036
Myometrial invasion
M2 vs. M0 and M1 6.89 2.06 23.08 0.002 3.86 1.56 9.57 0.004 6.86 2.04 23.09 0.002
Lymphovascular invasion
Yes vs. no 10.12 2.37 43.22 0.002 5.06 1.72 14.90 0.003 9.32 2.18 39.89 0.003
Cervical involvement
Stromal vs. no 4.29 1.74 10.56 0.002 5.09 2.19 11.80 0.000 4.56 1.80 11.59 0.001
Adnexal involvement
Yes vs. no 11.91 5.30 26.77 0.000 10.35 4.65 23.03 0.000 13.32 5.76 30.77 0.000
Lymph node metastasis
Yes vs. no 8.74 3.43 22.25 0.000 2.59 0.94 7.14 0.066 8.39 3.21 21.91 0.000
Peritoneal cytology
Positive vs. negative 13.99 5.93 32.99 0.000 10.54 4.18 26.60 0.000 14.52 6.01 35.09 0.000
Preoperative CA125 serum
level, U/mL
High vs. medium and low 2.64 1.18 5.88 0.018 1.78 0.82 3.83 0.143 2.49 1.10 5.63 0.029
Multivariate analysis
a

Preoperative HE4 serum level
High vs. medium and low 2.78 1.16 6.63 0.022 2.49 1.13 5.49 0.023 2.66 1.10 6.45 0.031
Preoperative CA125 serum
level, U/mL
High vs. medium and low 0.98 0.35 2.76 0.970 1.72 0.50 3.22 0.610 0.98 0.35 2.72 0.970
CI, condence interval of the estimated HR; HR, hazard ratio.
a
Multivariate models were adjusted for stage of disease, grade and adnexal
involvement. Multiple imputation data.
67.9 % , respectively. Regarding CA125, the specicity was
65.6 % and the sensitivity for detecting all EC or only stage I
EC patients was 71.5 % and 66.1 % , respectively.
Therefore, HE4 performed better than CA125 in both
early and advanced stage EC and it is worth noting that our
results showed an improved diagnostic performance than that
obtained by Moore et al. (16) . Regardless of tumour stage and
differentiation grade, the combination of sHE4 with sCA125
raised considerably the sensitivity in EC detection compared
to CA125 alone, but did not yield any statistical signicant
improvement to the performance of HE4. Our data were con-
sistent with other studies (16, 19) even if Moore et al. showed
that the combination of HE4 and CA125 exhibited a little
but signicant improvement compared to HE4 alone in stage
II-IV EC (16) .
Our ndings underline the important role of sHE4 as a new
potential biomarker for EC detection that could be evaluated
as a clinical diagnostic test for EC women who do not have
common EC-associated symptoms.
In agreement with our previous results (19) , in this report
we demonstrated, on a larger number of patients, that HE4
is signicantly correlated with all pathological variables that
dene a more aggressive tumour phenotype (deep myometrial
invasion, lymph nodes metastasis, adnexal involvement, vas-
cular or lymphatic invasion, positive peritoneal cytology, cer-
vical invasion, high grade and high stage). Conversely, Moore
et al. (20) recently reported that HE4 is correlated only with
deep myometrial invasion and grade: this could be explained
by the smaller number of patients enrolled in that study.
Accordingly with the literature (28, 29) , we found a sig-
nicant correlation between CA125 and the principal prog-
nostic indicators except tumour grade. Since HE4 correlates
also with tumour grade, it is more powerful than CA125 in
identifying adverse ECs factors.
2196 Zanotti et al.: Serum HE4 detection in endometrial carcinoma
0.0
0 10 20 30 40
Months
50 60 70 80
0.2
0.4
Low/medium
High
Low/medium
High
Low/medium
High
Low/medium
High
Low/medium
High
Low/medium
High
p=0.030
p=0.970 p=0.610
p=0.970
p=0.024
DFS
HE4
CA125
PFS OS
DFS
PFS OS
p=0.020
S
u
r
v
i
v
a
l

p
r
o
b
a
b
i
l
i
t
y
0.6
0.8
1.0
0.0
0 10 20 30 40
Months
50 60 70 80
0.2
0.4
S
u
r
v
i
v
a
l

p
r
o
b
a
b
i
l
i
t
y
0.6
0.8
1.0
0.0
0 10 20 30 40
Months
50 60 70 80
0.2
0.4
S
u
r
v
i
v
a
l

p
r
o
b
a
b
i
l
i
t
y
0.6
0.8
1.0
0.0
0 10 20 30 40
Months
50 60 70 80
0.2
0.4
S
u
r
v
i
v
a
l

p
r
o
b
a
b
i
l
i
t
y
0.6
0.8
1.0
0.0
0 10 20 30 40
Months
50
0.2
0.4
S
u
r
v
i
v
a
l

p
r
o
b
a
b
i
l
i
t
y
0.6
0.8
1.0
0.0
0 10 20 30 40
Months
50
0.2
0.4
S
u
r
v
i
v
a
l

p
r
o
b
a
b
i
l
i
t
y
0.6
0.8
1.0
A B C
D E F
Figure 3 Kaplan-Meier survival curves for EC patients on the entire patients cohort according to HE4 concentrations (AC) and to CA125
concentrations (DF).
The analyses were adjusted for stage, grade, histology and adnexal involvement. sHE4 concentrations were categorised on the basis of
tertiles
In this study we also analysed the impact of both serum
biomarkers and of clinical parameters on survival endpoints.
In univariate analysis, as expected, FIGO stage, ovarian or
lymph node metastasis and myometrial invasion showed a
statistically signicant association with OS, DFS and PFS,
proving the validity of the patients cohort recruited in this
study.
Similarly, both high sCA125 (high vs. low and medium ter-
tiles) and sHE4 concentrations (high vs. low and medium ter-
tiles) were signicantly more often observed in patients with
poor OS and PFS. In univariate analysis, the correlation of
CA125 with survival endpoints is conrmed by all published
studies (11) .
In the multivariate analysis, only HE4 resulted in an inde-
pendent prognostic factor for shortened OS, PFS and DFS, in
the whole cohort of EC patients. Regarding the conventional
prognostic factors, only the FIGO stage was independently
associated with OS and PFS. Interestingly, HE4 appear to
be better prognostic factor than the FIGO stage: these data
suggest that HE4 could reect intrinsic tumour aggressive-
ness that established prognostic factors were not able to
identify in our cohort of EC patients. Before this study, only
Bignotti et al. (19) assessed the impact on the prognosis of
HE4, however, in this previous paper, HE4 was found to
be an independent prognostic factor only in poorly differ-
entiated EC patients. Otherwise, the present paper demon-
strated the relevant prognostic impact of HE4 on the whole
EC cohort, regardless of tumour grade, and this guarantees a
higher robustness of the results. Indeed, also moderately dif-
ferentiated EC can be aggressive, showing 34 % positivity in
pelvic nodes at diagnosis (31) and 15 % of recurrences even in
stage I (32) .
The reports on the independent value of preoperative
sCA125 are conicting (33, 34) . Our data are in accordance
with Lundstr m et al. (33) and Hoskins et al. (34) , showing
that sCA125 is not predictive of clinical outcome.
In conclusion, this study is among the rst ones to demon-
strate that HE4 is a serum marker with a better sensitivity than
CA125 for EC detection. Its good sensitivity in early stage
disease suggests that HE4 may potentially be also a more
effective marker for detecting early recurrences or for moni-
toring response to treatment.
Notably, to our knowledge, this is the rst study that
demon strates the importance of HE4 as an independent prog-
nostic factor for EC, irrespective of FIGO stage, histological
grade, histotype and adnexal involvement. The correlation
between HE4 and a more aggressive EC phenotype can
allow the identication of high-risk subsets of patients with
worse clinical outcome that could benet from a more radical
sur gical treatment and a tailored adjuvant therapy.
The current availability of a highly sensitive and specic
assay for automatic measurement of sHE4 will facilitate fur-
ther multicentre studies to validate its clinical usefulness in
the management of EC patients.
Acknowledgments
We wish to thank all the nurses working in the Department of
Obstetrics and Gynaecology of the University of Brescia and
Zanotti et al.: Serum HE4 detection in endometrial carcinoma 2197
especially Ms. Margherita Franzoni, for the assistance in collect-
ing blood samples. We are also grateful to Ms. Michela Faustini for
her essential technical contribution in performing the Architect as-
says and to Ms. Ilary Toresini for informatic support. This inves-
tigation was supported by the grant from the Istituto Superiore di
Sanit (Programma Italia-USA Farmacogenomica Oncologica ,
convenzione 527/B4/4), Rome, Italy.
Conict of interest statement
Authors conict of interest disclosure: The authors stated that
there are no conicts of interest regarding the publication of this
article. Research funding and employment played no role in the
study design; in the collection, analysis, and interpretation of data;
in the writing of the report; or in the decision to submit the report
for publication.
Research funding: Dr. Antonella Ravaggi was invited for a presen-
tation as an expert testimony to a workshop sponsored by Abbott
Diagnostics at the ISOBM congress in October 2011.
Employment or leadership: Dr. Claudio Galli is currently employed
by Abbott Diagnostics as the Scientic Affairs Manager, Italy.
Honorarium: None declared.
References
1. Amant F, Moerman P, Neven P, Timmerman D, Van Limbergen
E, Vergote I. Endometrial cancer. Lancet 2005;366:491 505.
2. Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA
Cancer J Clin 2010;60:277 300.
3. Sorosky JI. Endometrial cancer. Obstet Gynecol 2008;111:
436 47.
4. Doubilet PM. Diagnosis of abnormal uterine bleeding with imag-
ing. Menopause 2011;18:421 4.
5. Scambia G, Benedetti Panici P, Baiocchi G, Perrone L, Greggi S,
Mancuso S. CA 15-3 as a tumor marker in gynecological malig-
nancies. Gynecol Oncol 1988;30:265 73.
6. Diefenbach CS, Shah Z, Iasonos A, Barakat RR, Levine DA,
Aghajanian C, et al. Preoperative serum YKL-40 is a marker for
detection and prognosis of endometrial cancer. Gynecol Oncol
2007;104:435 42.
7. Ginath S, Menczer J, Fintsi Y, Ben-Shem E, Glezerman M,
Avinoach I. Tissue and serum CA125 expression in endometrial
cancer. Int J Gynecol Cancer 2002;12:372 5.
8. Kanat-Pektas M, Yenicesu O, Gungor T, Bilge U. Predictive
power of sexual hormones and tumor markers in endometrial
cancer. Arch Gynecol Obstet 2010;281:709 15.
9. Sohaib SA, Houghton SL, Meroni R, Rockall AG, Blake P,
Reznek RH. Recurrent endometrial cancer: patterns of recur-
rent disease and assessment of prognosis. Clin Radiol 2007;62:
28 36.
10. Hacker NF. Uterine cancer. In: Berek JS, Hacker NF, editors.
Practical serum cancer antigen CA125 gynecologic onco-
logy. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins,
2005:397 442.
11. Gadducci A, Cosio S, Carpi A, Nicolini A, Genazzani AR. Serum
tumor markers in the management of ovarian, endometrial and
cervical cancer. Biomed Pharmacother 2004;58:24 38.
12. Kirchhoff C, Habben I, Ivell R, Krull N. A major human
epididymis-specic cDNA encodes a protein with sequence
homology to extracellular proteinase inhibitors. Biol Reprod
1991;45:350 7.
13. Bingle L, Singleton V, Bingle CD. The putative ovarian tumour
marker gene HE4 (WFDC2) is expressed in normal tissues and
undergoes complex alternative splicing to yield multiple protein
isoforms. Oncogene 2002;21:2768 73.
14. Drapkin R, von Horsten HH, Lin Y, Mok SC, Crum CP, Welch
WR, et al. Human epididymis protein 4 (HE4) is a secreted gly-
coprotein that is overexpressed by serous and endometrioid ova-
rian carcinomas. Cancer Res 2005;65:2162 9.
15. Hellstr m I, Raycraft J, Hayden-Ledbetter M, Ledbetter JA,
Schummer M, McIntosh M, et al. The HE4 (WFDC2) pro-
tein is a biomarker for ovarian carcinoma. Cancer Res
2003;63:3695 700.
16. Moore RG, Brown AK, Miller MC, Badgwell D, Lu Z, Allard WJ,
et al. Utility of a novel serum tumor biomarker HE4 in patients
with endometrioid adenocarcinoma of the uterus. Gynecol Oncol
2008;110:196 201.
17. Huhtinen K, Suvitie P, Hiissa J, Junnila J, Huvila J, Kujari H,
et al. Serum HE4 concentration differentiates malignant ovarian
tumours from ovarian endometriotic cysts. Br J Cancer 2009;
100:1315 9.
18. Montagnana M, Lippi G, Ruzzenente O, Bresciani V, Danese
E, Scevarolli S, et al. The utility of serum human epididymis
protein 4 (HE4) in patients with a pelvic mass. J Clin Lab Anal
2009;23:331 5.
19. Bignotti E, Ragnoli M, Zanotti L, Calza S, Falchetti M, Lonardi
S, et al. Diagnostic and prognostic impact of serum HE4 detec-
tion in endometrial carcinoma patients. Br J Cancer 2011;104:
1418 25.
20. Moore RG, Miller CM, Brown AK, Robison K, Steinhoff M,
Lambert-Messerlian G. Utility of tumor marker HE4 to predict
depth of myometrial invasion in endometrioid adenocarcinoma
of the uterus. Int J Gynecol Cancer 2011;21:1185 90.
21. Ruggeri G, Bandiera E, Zanotti L, Belloli S, Ravaggi A, Romani
C, et al. HE4 and epithelial ovarian cancer: comparison and
clinical evaluation of two immunoassays and a combination
algorithm. Clin Chim Acta 2011;412:1447 53.
22. Mongia SK, Rawlins ML, Owen WE, Roberts WL. Performance
characteristics of seven automated CA 125 assays. Am J Clin
Pathol 2006;125:921 7.
23. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing
the areas under two or more correlated receiver operating
characteristic curves: a non-parametric approach. Biometrics
1988;44:837 45.
24. Sonoda Y, Barakat RR. Screening and the prevention of gyne-
cologic cancer: endometrial cancer. Best Pract Res Clin Obstet
Gynaecol 2006;20:363 77.
25. Farias-Eisner G, Su F, Robbins T, Kotlerman J, Reddy S, Farias-
Eisner R. Validation of serum biomarkers for detection of
early- and late-stage endometrial cancer. Am J Obstet Gynecol
2010;202:73.e1 e5.
26. Cherchi PL, Dessole S, Ruiu GA, Ambrosini G, Farina M,
Capobianco G, et al. The value of serum CA 125 and associa-
tion CA 125/CA 19-9 in endometrial carcinoma. Eur J Gynaecol
Oncol 1999;20:315 7.
27. Hareyama H, Sakuragi N, Makinoda S, Fujimoto S. Serum and
tissue measurements of CA72-4 in patients with endometrial car-
cinoma. J Clin Pathol 1996;49:967 70.
28. Powell JL, Hill KA, Shiro BC, Diehl SJ, Gajewski WH.
Preoperative serum CA-125 levels in treating endometrial can-
cer. J Reprod Med 2005;50:585 90.
29. Jhang H, Chuang L, Visintainer P, Ramaswamy G. CA 125 levels
in the preoperative assessment of advanced-stage uterine cancer.
Am J Obstet Gynecol 2003;188:1195 7.
2198 Zanotti et al.: Serum HE4 detection in endometrial carcinoma
30. Galgano MT, Hampton GM, Frierson HF Jr. Comprehensive
analysis of HE4 expression in normal and malignant human tis-
sues. Mod Pathol 2006;19:847 53.
31. Creasman WT, Odicino F, Maisonneuve P, Quinn MA,
Beller U, Benedet JL, et al. Carcinoma of the corpus uteri.
FIGO 26th Annual Report on the Results of Treatment in
Gynecological Cancer. Int J Gynaecol Obstet 2006;95(Suppl 1):
S105 43.
32. DiSaia PJ, Creasman WT, editors. Clinical gynecologic onco-
logy. 7th ed. Philadelphia, PA: Mosby Elsevier, 2007:159 61.
33. Lundstr m MS, H gdall CK, Nielsen AL, Nyholm HC. Serum
tetranectin and CA125 in endometrial adenocarcinoma. Anti-
cancer Res 2000;20:3903 6.
34. Hoskins PJ, Le N, Correa R. CA 125 normalization with che-
motherapy is independently predictive of survival in advanced
endometrial cancer. Gynecol Oncol 2011;120:52 5.