A C TA Obstetricia et Gynecologica

AOGS M A I N R E SE A RC H A R TI C LE

Preoperative staging of endometrial cancer using TVS, MRI,

and hysteroscopy
GITTE ØRTOFT1, MARGIT DUEHOLM1, OLE MATHIESEN1, ESTRID S. HANSEN3, ERIK LUNDORF2,
CHARLOTTE MØLLER1, EDVARD MARINOVSKIJ2 & LONE K. PETERSEN1
Departments of 1Gynecology and Obstetrics, 2
Diagnostic Imaging, and 3Histopathology, Aarhus University Hospital
Aarhus, Denmark

Key words Abstract

Endometrial biopsy, endometrial cancer,
hysteroscopy, magnetic resonance imaging,
myometrial invasion, preoperative staging,
tumor grade TVS
Correspondence
Gitte Ørtoft, Department of Gynecology and
Obstetrics, Aarhus University Hospital,
DK-8200 Aarhus N, Denmark.
E-mail: ortoft@dadlnet.dk
Conflict of interest
The authors have stated explicitly that they
have no conflicts of interest in connection
with this article.
Please cite this article as: Ørtoft G, Dueholm
M, Mathiesen O, Hansen ES, Lundorf E,
Møller C, Marinovskij E, Petersen LK.
Preoperative staging of endometrial cancer
using TVS, MRI, and hysteroscopy. Acta
Obstet Gynecol Scand 2013; 92:536–545.
Received: 17 September 2012
Accepted: 21 December 2012
DOI: 10.1111/aogs.12103
Objectives. To evaluate the accuracy of different preoperative modalities for
staging of endometrial cancer to restrict extensive surgery to patients at high
risk of metastatic disease. Setting. Aarhus University Hospital. Population. 156
women referred in 2006–2011 because of atypical endometrial hyperplasia (G0)
or endometrial cancer. Methods. Patients were offered preoperative transvaginal
ultrasonography (TVS), magnetic resonance imaging (MRI), and hysteroscopic-
directed biopsies from the uterine tumor and cervix. Final pathology of the
removed uterus was the reference standard. Patients were divided into low risk
(<50% myometrial invasion, and grades 0, 1, 2, and no cervical invasion) or
high risk (all others). Main outcome measures. Accuracy, sensitivity, specificity,
positive/negative predictive value. Results. Patients were aged 32–88 years, with
a mean body mass index of 29. At final pathology 81% had cancer and 19%
G0 or no residual tumor; 54% were high risk. Hysteroscopy-directed biopsies
had a higher accuracy (92%) than endometrial biopsy (58%) for differentiating
G0 from cancer (p < 0.001); grade 3 tumor identification had similar accuracy
(93 vs. 92%). Deep myometrial invasion was estimated with higher accuracy by
MRI (82%) than TVS (74%) (p < 0.02). For cervical involvement, hystero-
scopy-directed biopsies had higher accuracy (94%) than MRI (84%,) and TVS
(80%) (p < 0.02). Accuracy for identifying high-risk women was highest (83%)
using a combination of MRI and hysteroscopic-directed biopsies, compared
with TVS and endometrial biopsy (72%) (p < 0.05). Conclusion. Preoperative
staging with MRI and hysteroscopy-directed biopsy can identify eight of 10
women with high risk of lymph node metastases and spare eight of 10 low-risk
women extended surgery.

Abbreviations: ASA, American Society of Anesthesiologists; D&C, dilatation and

curettage; FOV, field of view; G0, atypical endometrial hyperplasia; MRI,
magnetic resonance imaging; TE, time to echo; TR, repetition time; TVS,
transvaginal ultrasound.

Introduction
In today’s modern cancer surgery, preoperative or peri-
operative staging is crucial to ensure that the surgery
matches the individual patient’s disease stage (1). In
women with endometrial cancer, hysterectomy and bilat-
eral salpingo-oophorectomy are the main surgical treat-
Key Message
Preoperative staging with MRI and hysteroscopy-
directed biopsy can identify women at high risk of
lymph node metastasis (accuracy 81%). Hystero-
scopy-directed biopsies had the highest accuracy for
identifying cervical invasion.

ª 2013 The Authors

536 Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545

G. Ørtoft et al.
ment modalities, whereas lymph node resection of the
pelvic and perhaps paraaortic vessels is performed mainly
to ensure proper staging for the allocation of women to
adjuvant therapy (2). Some centers also suggest radical
hysterectomy for women with cervical involvement to
ensure local control of the disease (3).
Previously, many centers performed operative staging
with lymph node resection in all women. Growing evi-
dence suggests that low-risk women have a very low risk
of lymph node metastases and also a high five-year over-
all survival (90–91%) as well as an excellent cancer-
specific survival (96–97%) (4,5). Furthermore, in low-risk
women postoperative radiotherapy does not seem to
improve survival (4,6,7), whereas the benefit of ensuring
local control seems well documented (6–8). Although
studies indicate that chemotherapy may increase overall
survival in stages III–IV, the effect of chemotherapy on
stages I–II is so far less documented (9–11). Therefore
women at low risk should probably be offered surgical
treatment with hysterectomy and bilateral salpingo-
oophorectomy only. The search for methods able to pre-
dict tumor grade, myometrial invasion, as well as cervical
invasion is therefore warranted to restrict extended sur-
gery to women at high risk of metastatic disease (1,2,12).
The method must have high accuracy because the lack of
lymph node resection may introduce an increased risk of
recurrence, whereas unnecessary lymph node resection
may increase the risk of surgical complications and also
the risk of life-long lymph edema (2).
Intraoperative methods for the determination of deep
myometrial invasion and tumor grade have been demon-
strated to be highly accurate (13,14). A preoperative
method would, however, have an advantage in operative
planning, in selecting women for minimal invasive sur-
gery, and in ensuring that any extensive surgery is per-
formed by well-trained surgeons. The aim of the present
study was to evaluate the preoperative methods of hyste-
roscopically directed biopsies, transvaginal ultrasound
(TVS) and magnetic resonance imaging (MRI) for the
evaluation of tumor grade, myometrial invasion, and
cervical involvement in women referred due to atypical
hyperplasia or endometrial cancer, with the intent to
restrict extensive surgery to women with a high risk of
metastatic disease.
Material and methods
In 2006, pelvic lymph node resection for high-risk endo-
metrial cancer women and radical hysterectomy for
women with known cervical involvement were introduced
and centralized to five Danish centers. In 2010, women
with atypical hyperplasia or low-risk cancer were further
centralized. From 2006 to 2011, women with endometrial
ª 2013 The Authors
Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545
16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Preoperative staging of endometrial cancer
cancer and atypical endometrial hyperplasia demonstrated
in an endometrial biopsy, dilatation and curettage
(D&C), or hysteroscopy referred directly from a general
practitioner or from one of four county hospitals, were
asked to participate in this prospective study. Inclusion
was slow during the first years and improved in 2010–
2011 due to further centralization. Within two weeks
prior to their surgery, women were offered preoperative
staging with TVS, MRI, and hysteroscopy-directed biop-
sies. Women with known serous histology were immedi-
ately offered extensive surgery and hysteroscopy was not
performed. In all, 176 women agreed to participate; 20
were excluded for reasons given in Figure 1, leaving 156
women for final analysis. A total of 51 (33%) women
were primarily diagnosed by D&C or hysteroscopy upon
referral and were not included in the endometrial biopsy
group. Six percent did not have the results of ultrasound
recorded, 8% did not have MRI due to obesity/claustro-
phobia, and 14% did not have hysteroscopy due to
known serous/clear tumor, women’s choice, prior hyster-
oscopy or high disease stage demonstrated on MRI.
Inclusion was both dependent on the women’s willingness
to be included as well as the surgeons’ willingness to
include their patients. The study was approved by the
local ethics committee and approved by the Danish Data
Protection Board.
Patients were preoperatively divided into low/high risk
according to the following scheme:
Uterine corpus manifestations only:
• Low risk: grades 0, 1, 2, <50% myometrial invasion.
• High risk: grade 3 (including serous, clear, undiffer-
entiated), and/or >50% myometrial invasion.
Uterine corpus and cervix manifestations including
both features of the corpus and the cervix:
• Low risk: grades 0, 1, 2, <50% myometrial invasion
and no cervical invasion.
• High risk: grade 3 (including serous, clear, undiffer-
entiated), and/or >50% myometrial invasion and/or cervi-
cal invasion.
Based on the final pathologic diagnosis, the preopera-
tive risk stratification was compared with the final low/
high risk. Final low-/high risk was accordingly defined as
either apparent low-/high risk including only uterine cor-
pus manifestation (final apparent low risk: grades 0–2,
<50% myometrial invasion; final apparent high risk: all
others] or true final low/high risk [true final low risk:
grades 0–2, <50% myometrial invasion, no cervical inva-
sion and no extra-uterine manifestation, true final high
risk: all others). Prior to surgery, women were informed
of the test results and which operation they were to be
offered.
TVS was performed by one of three blinded ultrasono-
graphic-trained general gynecologists using a transvaginal
537
 16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Preoperative staging of endometrial cancer G. Ørtoft et al.

Eligible patients (pt)

n = 176

20 pt excluded
Benign changes: 9
Decision not to participate: 3
Uterine metastases from cancer of
the ovary and kidney: 2
Comorbidities: 2
Inoperable stage 4: 1
Unknown reason: 3

156 pt included

22 pt no hysteroscopy: 9 pt ultrasound
due to serous or clear 13 pt no MRI due to high
results not
(10) pt choice (9) or high BMI/claustrophobia
recorded in file
disease stage (3) 9 low-risk,
4 low-risk
4 low-risk 4 high-risk
5 high-risk
18 high-risk

134 pt 143 pt MRI 147 pt

Hysteroscopy with transvaginal ultrasound
biopsies 2 pt myometrial
invasion not
evaluated
1 high-risk
Benign hysteroscopy
biopsies (6) 1 low-risk
112 pt
Pt no residual tumor at
final pathology (13). included in
Benign hysteroscopy and Kappa analysis 77 pt 64 pt 81 pt 66 pt
no residual tumor (3) on grade < 50% invasion > 50% invasion < 50% invasion > 50% invasion

Figure 1. Flow chart for patients participating in the study.

probe (GE Voluson scanner, E8, GE-Healthcare, Wauke-
sha, WI, USA). All TVS were included independent of
quality. Tumors were identified and measured in longitu-
dinal as well as transversal planes. Myometrial involve-
ment was defined as maximum myometrial involvement
(distance from deepest tumor invasion to the serosal sur-
face) divided by full wall thickness (myometrial thickness
at a point of no invasion or the least invasion). Cervical
involvement was assessed based on tumor growth or
irregularity on the borders to or in the cervix.
MRI was performed with patients in the supine posi-
tion using cardiac and body matrix coils [1.5 Tesla Gen-
eral Electric Signa (GE Healthcare, Hatfield, UK), Philips
Achieva (Philips Healthcare, Amsterdam, Holland)/Sie-
mens Avanto (Siemens, Malvern, PA, USA) MRI scan-
ners] after administration of an anti-peristaltic agent
(1 mg glucagon intramuscular). The protocol included
sagittal T2-weighted fast spin echo sequence [repetition
time (TR) = 4.140–5.830 ms, time to echo (TE) = 85–
104 ms, field of view (FOV) = 240 mm] with a slice
thickness of 5 mm of the small pelvis, axial oblique
T2-weighted fast spin echo sequence (TR = 3.000 ms,
TE = 85–109 ms, FOV = 240 mm) with a slice thickness
of 3–4 mm perpendicular to the long axis of the uterine
corpus and cervix, axial IV contrast-enhanced [gadoterate
meglumine (Dotarem, Guerbet, Roissy, France)
55.86 mg/kg] T1-weighted fat-saturated sequence
(TR = 400–700 ms, TE = 14 ms, FOV = 240–260 mm;
sequence was acquired 60–90 s after contrast administra-
tion] through the uterine corpus, and axial T2-weighted
fast spin echo fat-saturated sequence (TR = 5.000 ms,
TE = 90 ms, FOV = 340 mm) with a slice thickness of
8 mm from symphysis to diaphragm. The MRI scan was
completed in 30 min (15). One of two blinded MRI radi-
ologists evaluated tumor size, visible or invisible junc-
tional zone, junctional zone interruption or discontinuity,
myometrial wall thickness, myometrial invasion, cervical
involvement, and any extra-uterine spread. Myometrial
involvement was detected when disruption of the junc-
tional zone was present, and measured as stated by ultra-
sound. In women with no visible junctional zone, the
presence of a smooth/irregular endometrial–myometrial
interface was considered as no/presence of myometrial
involvement. Cervical involvement was present when
tumor boarders were present in the cervix or if disruption
of the normal low-signal cervical stroma/or high signal
intensity cervical mass was detected.
Hysteroscopy was performed in the day surgery unit
under short general anesthesia by one of five gynecolo-
gists using a resectoscope (Storch, N. Westport, CT,

ª 2013 The Authors

538 Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545

G. Ørtoft et al.
USA). A subjective assessment of cervical involvement
was noted if tumor borders reached the internal cervical
ostium. Resectoscopic biopsies (3–5) were taken from the
endometrium at the tumor site, the endo-myometrium at
the tumor borders, and from the cervix (3), starting from
the internal cervical border to the surface of the cervix.
Myometrial invasion and stromal cervical invasion were
determined on the removed uterus (reference standard).
For the hysteroscopy-directed biopsies, cervical involve-
ment was determined by stromal invasion in areas of
biopsies also containing cervical glandular tissue. Cervical
biopsies containing only cancer were classified as having
cervical invasion (three patients). Endometrial biopsies
from referral hospitals and all other tissues for pathology
were analyzed by one of two pathologists specializing in
gynecological oncology. All pathologic results were re-
evaluated blindly after study termination, resulting in
only minor alterations (three of 156 evaluations altered
with regard to both deep myometrial invasion and cervi-
cal invasion). Cervical involvement was differentiated
from primary cervical cancer/metastasis disease using
immunohistochemistry. Atypical hyperplasia was defined
as cytologic atypia (nuclear with abnormal chromatin
structure and pleomorphism, loss of polarity, prominent
nuclei) with no myometrial invasion. Histological grade
was determined by the percentage of solid epithelial
growth: G1  5%, G2 6–50%, and G3 > 50% and
upgraded in case of nuclei atypia. Serous, clear cell or
undifferentiated tumors were classified as high grade.
Cancer patients were followed up for 2.8  1.1 years
(1.4–5.7 years). Twenty-six women had recurrent cancer.
Eight women had vaginal recurrences. Nineteen women
died. The five-year overall survival was 85%, 93% for
low-risk (excluding atypical hyperplasia) and 89% for
high-risk stage I. Three low-risk stage I patients died: one
from pancreatic cancer and two from endometrial cancer:
(i) the cancer recurred shortly after a vaginal hysterec-
tomy (body mass index of 48), and (ii) experienced early
recurrence and died from vaginal and retroperitoneal
metastases within the first year).
Statistics
A power assessment demonstrated the need for at least 63
women in a test group (sensitivity: 95%, standard devia-
tion: 10%, alpha: 0.05, power: 0.80). Data were analyzed
using STATA Statistical Software (STATACorp, College
Station, TX, USA). Differences between means were eval-
uated using Student’s t-test. The exact agreement and
Kappa  SD value were calculated for evaluation of
tumor grade (G0–G3). For each method and for combi-
nations of tests, accuracy, sensitivity, specificity, positive
predictive value and negative predictive value were deter-
ª 2013 The Authors
Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545
16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Preoperative staging of endometrial cancer
mined to identify patients at high risk. The McNemar test
was used to evaluate differences between test modalities.
Results
Ninety-three percent of women were postmenopausal and
7% were pre- or perimenopausal. American Society of
Anesthesiologists (ASA) health scores were: ASA1 42%,
ASA2 46%, ASA3 12%. At inclusion, the diagnosis was
atypical hyperplasia in 44% of women and cancer in
56%. The referral diagnosis was based on D&C in 19%,
hysteroscopy  D&C in 14%, and endometrial biopsy in
67%. At final pathology, 11% had no residual tumor, 8%
atypical hyperplasia and 81% cancer; and of these latter
women, 34% were low risk and 66% high risk (Table 1).
Only women graded on both primary and final pathol-
ogy were included for the analysis of tumor grade
(Tables 2 and 3). Non-revised endometrial biopsy (55
patients) determined grade (G0–G3) with an accuracy of
32% (Kappa = 0.17  0.06), revised endometrial biopsy
(105 patients) 43% (Kappa = 0.27  0.05) and hystero-
scopic-directed biopsies 77% (Kappa = 0.60  0.06).
Hysteroscopy failed to identify three women with grade 1
and 2 tumors (identified by MRI) and one patient with a
grade 3 tumor (identified on examination of hysteroscopy
tissue from referral hospital; final pathology showed a
small number of residual carcinosarcoma cells in an ade-
nomyositic focus). With an accuracy of 92%, hysteros-
copy-directed biopsy was significantly better than
endometrial biopsy (58%) in differentiating atypical
endometrial hyperplasia from cancer, whereas no differ-
ences were found when differentiating grade 3 tumor
from grade 1–2 or grade 0–2 tumors (Table 3). Perfora-
tion occurred in 3/134 women treated with antibiotics.
Postoperative bleeding occurred in 2/134 women; one
had to be hospitalized for 24 h due to abdominal pain.
The estimation of deep myometrial invasion (>50%) was
significantly better using MRI (accuracy: 82%) than using
TVS (accuracy: 74%) (Table 4). Visual cervical involvement
was evaluated after dilatation in 122 of 133 hysteroscopies,
with an accuracy of 86%, the sensitivity being only 38%
(specificity: 92%). The estimation of cervical involvement
was significantly more accurate using hysteroscopy-directed
biopsy (accuracy: 95%) than MRI (84%) or TVS (80%),
even though the sensitivity using hysteroscopy-directed
biopsies for identifying women with cervical involvement
was only 73% (Table 4). No complications were seen after
TVS or MRI.
Table 5 demonstrates the results of combining tests for
differentiating apparent/true low risk from apparent/true
high risk. MRI combined with hysteroscopy-directed biop-
sies revealed a significant higher accuracy for differentiating
between women at apparent low risk from those at high risk
539
 16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Preoperative staging of endometrial cancer G. Ørtoft et al.

Table 1. Patient characteristics for all patients and for patients defined as low-risk (grades 0, 1, and 2, <50% myometrial invasion and no
cervical invasion, no extrauterine spread) and high-risk (all others) on final pathology.

All patients Low-risk High-risk

n 156 72 84

Mean  SD Ranges Mean  SD Ranges Mean  SD Ranges

Age 66  9.7 32–88 62  8.2 42–83 68  10 a

32–88
Body weight (kilo) 80  19 46–143 84  22 50–143 77  16a 46–118
Body mass index 29  7 17–53 31  8.2 17–53 28  5.3a 18–42

Final histology n % n % n %

No residual tumor 17 11 17 24 – –
Atypical endometrial hyperplasia 12 8 12 17 – –
Endometroid adenocarcinoma 102 65 43 60 59 70
Clear 10 6 – – 10 12
Serous 2 1 – – 2 2
Undifferentiated carcinoma 3 2 – – 3 4
Carcinosarcoma 10 6 – – 10 12

Final grade of tumor n % n % n %

No residual tumor 17 11 17 24 – –
Grade 0 (atypical hyperplasia) 12 8 12 17 – –
Grade 1 75 48 36 50 39 (46) 46
Grade 2 18 12 7 10 11 (13) 13
Grade 3, serous, clear 34 22 – – 34 (40) 40

Final surgical stage n % n % n %

No residual tumor 17 11 17 24 – –
0 (atypical hyperplasia) 12 8 12 17 – –
I
Low-risk stage I 43 28 43 60 – –
High-risk stage I 44 29 – – 44 52
II 15 10 – – 15 18
III 20 13 – – 20 24
IV 5 3 – – 5 6

Mean  SD (0–100% range) or n and %.

a
p < 0.01 against high-risk patients by Student’s t-test.

(accuracy: 83%) compared with TVS and hysteroscopy-
directed biopsies (77%), MRI and revised endometrial
biopsy (73%), TVS and revised endometrial biopsy (72%),
TVS and the non-revised endometrial biopsy (70%), and
TVS and any primary test (17%). MRI combined with hyst-
eroscopy-directed biopsies also revealed significantly better
accuracy (81%) than all other combinations of tests, except
MRI and revised biopsy, for the differentiation between true
low-risk women and true high-risk women. Thirteen final
low-risk women had lymph node resection and none had
lymph node metastases. Sixty-three of 84 final high-risk
women had lymph node resection, and 12 had lymph node
metastases (19%). Only six of 12 women with positive
lymph nodes were identified by MRI.
Of the 48 women included due to atypical hyperplasia
on endometrial biopsy, 70% had cancer on final pathol-
ogy, and of these, 29% had high-risk features and 6%
(two women) cervical involvement. Altogether, 31% of
women presenting with atypical endometrial hyperplasia
needed extended surgery.
Discussion
Endometrial cancer women at low risk of lymph node
metastasis have excellent survival after hysterectomy and
bilateral salpingo-oophorectomy and should be spared
lymph node resection to avoid increased risk of surgical
complications and risk of lymph edema (2,4,5,12).
Women with a high risk of lymph node metastasis, evalu-
ated on the basis of tumor grade, myometrial invasion,
and perhaps also cervical involvement, may benefit from
a staging operation with resection of pelvic and perhaps

ª 2013 The Authors

540 Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545
 16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G. Ørtoft et al. Preoperative staging of endometrial cancer

Table 2. Accuracy of endometrial (before and after revision) and hysteroscopic-directed biopsy for the determination of tumor grade.

Grade 0 Grade 3 All

(atypia) Grade 1 Grade 2 serous clear No residual tumor n

Non-revised endometrial biopsy 55

Grade 0 (atypia) 4 21 3 1 9 38
Grade 1 0 3 2 0 0 5
Grade 2 0 0 1 1 0 2
Grade 3, serous, clear 1 1 0 6 0 8
Not graded 0 1 0 1 0 2
Revised endometrial biopsy 105
Grade 0 (atypia) 4 29 4 1 10 48
Grade 1 1 17 4 0 1 23
Grade 2 0 1 4 1 0 6
Grade 3, serous, clear 2 2 1 18 1 24
Not graded 0 3 0 1 0 4
Hysteroscopic-directed biopsya 134
Grade 0 (atypia) 6 4 1 0 6 17
Grade 1 3b 57 6 2 5 73
Grade 2 0 4 5 3 1 13
Grade 3, serous, clear 1 0 2 18 1 22
Benign pathologyc 2 1 2 1 3 9

a
p < 0.05 when hysteroscopy was compared with non-revised endometrial biopsy (n = 73) and revised endometrial biopsy (n = 75) using
McNemar test.
b
Three patients with endometroid adenocarcinoma G1 on hysteroscopy and only atypia on pathology were included.
c
Six patients with benign histology on hysteroscopy were excluded in the Kappa analysis and McNemar test. Non-revised endometrial biopsy:
agreement: 31.8, Kappa = 0.17, SD 0.06. Revised endometrial biopsy agreement: 42.5, Kappa = 0.27, SD 0.05, hysteroscopy: agreement: 77%,
Kappa = 0.60, SD 0.06.

Table 3. Accuracy of different diagnostic test for the differentiating of (a) atypical endometrial hyperplasia (grade 0) from cancer; (b) Grades 0,
1, 2, from grade 3 (including unfavorable tumor types); (c) Grades 1, 2, from grade 3.

Negative predictive
Specificity Sensitivity value Positive predictive value Accuracy

95%
% CI n/N % 95% CI n/N % 95% CI n/N % 95% CI n/N % n/N

a. Atypical endometrial hyperplasia from all cancers

Endometrial biopsy
Non-revised 80 28–99 4/5 36 21–53 14/39 14 4–32 4/29 93 68–100 14/15 41 18/44
Revised 57 18–90 4/7 59 47–69 48/82 11 3–25 4/38 94 84–99 48/51 58 52/89
Hysteroscopy 60 26–87 6/10 95a 89–98 97/102 55 23–83 6/11 96 90–99 97/101 92b 103/112
b. Grades 0,1,2, from grade 3
Endometrial biopsy
Non-revised 94 81–99 34/36 75 35–97 6/8 94 81–99 34/36 75 35–97 6/8 91 40/44
Revised 93 84–98 64/69 90 68–99 18/20 97 89–100 64/66 78 56–93 18/23 92 82/89
Hysteroscopy 97 90–99 86/89 78 56–93 18/23 95 88–98 86/91 86 64–97 18/21 93 104/112
c. Grades 1,2 from grade 3
Endometrial biopsy
Non-revised 86 42–100 6/7 86 42–100 6/7 86 42–100 6/7 86 42–100 6/7 86 12/14
Revised 90 73–98 26/29 95 74–100 18/19 96 82–100 26/27 86 64–97 18/21 92 44/48
Hysteroscopy 94 85–98 72/74 90 68–99 18/23 97 91–100 72/77 78 56–93 18/20 93 90/97

McNemar test was used for comparing sensitivity, specificity and accuracy among tests.
a
p < 0.05 compared with non-revised (n = 34) and revised endometrial biopsy (n = 70).
b
p < 0.05 compared with non-revised endometrial biopsy (n = 38) and revised endometrial biopsy (n = 75).

ª 2013 The Authors

Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545 541
 16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Preoperative staging of endometrial cancer G. Ørtoft et al.

Table 4. Accuracy of different diagnostic test for evaluating low- and high-risk features of the uterus (myometrial invasion more or less than
50% and cervical involvement).

Positive predictive
Specificity Sensitivity Negative predictive value value Accuracy

% 95% CI n/N % 95% CI n/N % 95% CI n/N % 95% CI n/N % n/N

Myometrial invasion
TVS 72 61–81 58/81 77 65–87 51/66 79 68–88 58/73 69 57–79 51/74 74 109/147
MRI 83a 73–91 64/77 80 68–89 51/64 83 73–91 64/77 80 68–89 51/64 82b 115/141
Cervical involvement
TVS 89 82–94 105/118 38 20–59 10/26 87 79–92 105/121 43 23–66 10/23 80 115/144
MRI 91 84–95 106/117 54 33–73 14/26 90 83–95 106/118 56 35–76 14/25 84 120/143
Hysteroscopy 97c 93–99 113/116 73 45–92 11/15 97 91–99 113/117 79 49–95 11/14 95d 124/131

McNemar test was used for comparing sensitivity, specificity, and accuracy among tests.
a
p < 0.05 compared with TVS (n = 71).
b
p < 0.05 compared with TVS (n = 132).
c
p < 0.05 compared with TVS (n = 111).
d
p < 0.05 compared with TVS (n = 125) and MRI (n = 118).
TVS, transvaginal ultrasound; MRI, magnetic resonance imaging.

paraaortic lymph nodes with regard to allocation to post-
operative adjuvant therapy (2,3). In the present study, we
applied different preoperative methods to optimize pre-
operative planning of women referred because of endo-
metrial cancer or atypical endometrial hyperplasia in
order to restrict extended surgery to women at high risk.
The combination of MRI and hysteroscopy-directed biop-
sies was found to have the highest efficacy for differenti-
ating between low- and high-risk women. The accuracy
was 81%, with a sensitivity of 83% and a specificity of
79%, which means that if this combination is used to
determine the surgical procedure, then 81% of women
would be operated on according to the Danish guidelines,
whereas 8% would endure unnecessary lymph node resec-
tion and 9% would need reoperation with removal of
lymph nodes or adjuvant therapy because of improper
staging. To evaluate cervical involvement, MRI, TVS, and
hysteroscopy-directed biopsies were evaluated, and the
highest accuracy was obtained using hysteroscopy-direc-
ted biopsy (95%). However, in the 131 women in whom
cervical biopsies were taken, the sensitivity was only 73%,
resulting in only 11 of the 15 women with cervical
involvement being identified preoperatively.
Although myometrial invasion in endometrial cancer
can be evaluated with high accuracy using intraoperative
methods (gross examination: 85–89%, frozen sections: 91
–95%) (13,14,16,17), a preoperative method would have
the advantage of preparing the patient for the extent of
surgery needed, differentiating the level of surgical exper-
tise present in the operating theater, allocating women to
minimal invasive surgery, and preserving the option of
radical hysterectomy for women with cervical involve-
ment. Of preoperative methods, TVS has been evaluated
with regard to determining deep myometrial invasion and
accuracies of 78–87% have been reported, similar to the
results of the present study (17–19). MRI can determine
myometrial invasion with accuracies of 47–86% (17,19–
23). A British national audit on MRI stated that low case-
load seems to decrease accuracy (23). Also found was a
high diagnostic accuracy (90%) for evaluating cervical
involvement but a low sensitivity (42%) similar to that in
the present study (8). Rockall et al. (21) demonstrated a
high sensitivity (84%) of MRI for evaluating deep
myometrial invasion and a sensitivity for cervical involve-
ment of 69% (21). Interestingly, they found that 50% of
women with cervical involvement had lymph node meta-
stases, compared with only 25% of women in the present
study.
An earlier study combined intraoperative gross examina-
tion during surgery with preoperative revision of tumor
grade and found an accuracy of 89% and a sensitivity of
76% for identifying high-risk women (13). Another study
combined revised endometrial biopsy with MRI and found
a sensitivity of 73% (65% in the present study) (22).
The accuracy (80%) of hysteroscopy-directed biopsies
for determining tumor grade including atypical endome-
trial hyperplasia (G0–G3) was superior to that of endo-
metrial biopsy. Cutillo et al. (24) demonstrated that
hysteroscopy-directed biopsy could determine the grade
of tumor with an accuracy of 97% (24). If we excluded
women with atypical hyperplasia (G0), the accuracy for
identifying grades 1–3 was 81% in the present study.
Studies have previously demonstrated that endometrial
biopsy only correctly diagnosed grade of tumor in 58%
of cases, rising to 74% for initial grade 1 tumors (12,25).
However, in the present study, the accuracy for determin-
ing grade 3 tumors with endometrial biopsy (92%) was
similar to that of hysteroscopy-directed biopsies (93%).

ª 2013 The Authors

542 Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545
 16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G. Ørtoft et al. Preoperative staging of endometrial cancer

Table 5. Combination of different diagnostic test for predicting (a) apparent low-risk (grade 1 or 2 and <50% myometrial invasion) and
apparent high-risk (grade 3 and/or >50% myometrial invasion) based on uterine corpus manifestation; (b) true low-risk (grade 1 or 2, less than
50% myometrial invasion, no cervical invasion and no extrauterine manifestation, and true high-risk (all others).

Negative predictive Positive predictive

Specificity Sensitivity value value Accuracy

95% 95% 95% 95%

n % CI n/N % CI n/N % CI n/N % CI n/N % n/N

a. Apparent low/high-risk
TVS & any primary biopsy 147 81 70–89 59/73 62 50–73 46/74 68 57–77 59/87 77 64–87 46/60 71 105/147
TVS & non rev endo. 50 84 67–95 27/32 44 22–69 8/18 73 56–86 27/37 62 32–86 8/13 70 35/50
biopsy
TVS & rev endo. biopsy 95 82 69–91 41/50 60 44–74 27/45 69 56–81 41/59 75 58–88 27/36 72 68/95
MRI & rev endo biopsy 93 81 67–91 38/47 65 50–79 30/46 70 56–82 38/54 77 61–89 30/39 73 68/93
TVS & hysteroscopy 121 74 61–84 45/61 80a 68–89 48/60 79 66–89 45/57 75 63–85 48/64 77 93/121
MRI & hysteroscopyg 112 82 69–91 45/55 84b 72–93 48/57 83 70–92 45/54 83 71–91 48/58 83c 93/112
b. True low/high–risk
TVS & any primary 147 78 66–87 53/68 65 53–75 51/79 65 54–76 53/81 77 65–87 51/66 71 104/147
biopsy
TVS & non rev endo. 50 80 61–92 24/30 50 27–73 10/20 71 53–85 24/34 63 35–85 10/16 68 34/50
biopsy
TVS & rev endo. biopsy 95 76 61–87 35/46 65 50–78 32/49 67 53–80 35/52 74 59–86 32/43 71 67/95
MRI & rev endo biopsy 93 81 67–92 35/43 66 51–79 33/50 67 53–80 35/52 80 65–91 33/41 73 68/93
TVS & hysteroscopy 121 72 59–83 42/58 78d 66–87 49/63 75 61–86 42/56 75 63–85 49/65 75 91/121
MRI & hysteroscopyg 112 79 65–89 41/52 83e 71–92 50/60 80 67–90 41/51 82 70–91 50/61 81f 91/112

McNemar test was used for comparing sensitivity, specificity, and accuracy among tests.
a
p < 0.05 compared with TVS & any primary biopsy (n = 60), TVS & non rev endometrial biopsy (n = 15), TVS & rev endometrial biopsy (n = 39),
MR & rev endometrial biopsy (n = 36).
b
p < 0.05 compared with TVS & any primary biopsy (n = 56), TVS & non rev endometrial biopsy (n = 13), TVS & rev endometrial biopsy (n = 37),
MR & rev endometrial biopsy (n = 36).
c
p < 0.05 compared with TVS & any primary biopsy (n = 108), TVS & non rev endometrial biopsy (n = 35), TVS & rev endometrial biopsy
(n = 74), MR & rev endometrial biopsy (n = 74) and TVS & hysteroscopy (n = 108).
d
p < 0.05 compared with TVS & any primary biopsy (n = 63).
e
p < 0.05 compared with TVS & any primary biopsy (n = 59), TVS & rev endometrial biopsy (n = 40), MR & rev endometrial biopsy (n = 40).
f
p < 0.05 compared with TVS & any primary biopsy (n = 108), TVS & non rev endometrial biopsy (n = 35), TVS & rev endo biopsy (n = 74) and
TVS & hysteroscopy (n = 108).
g
Patients with benign pathology on hysteroscopy excluded.
non rev endo, non-revised endometrial; rev endo, revised endometrial; TVS, transvaginal ultrasound; MRI, magnetic resonance imaging.

Larson et al. (25) found an accuracy of 56% for determi-
nation grade 3 tumors, whereas Traen et al. (13) found
an accuracy of 88% using revised endometrial biopsy.
The strength of the present study is our ability to apply
a range of different preoperative tests in a rather large
number of women. A limitation was that not all women
participated in all of the examinations. However, exclud-
ing women with extreme obesity or women with cervical
stenosis which makes endometrial biopsy impossible
would also introduce bias. The fact that we included
women primarily diagnosed with atypical endometrial
hyperplasia could be a limitation; however, 71% of
women referred because of atypical endometrial hyper-
plasia on endometrial biopsy had endometrial cancer at final
pathology: 29% had high-risk features and two women
had cervical involvement. Suh-Burgmann et al. (26) dem-
onstrated that only 48% of their women primarily diag-
nosed as having atypical endometrial hyperplasia actually
had endometrial cancer on final pathology (26). The dif-
ference may be caused by different referral patterns or
differences in the pathologists’ diagnosis of atypical endo-
metrial hyperplasia. Using hysteroscopy-directed biopsies,
atypical endometrial hyperplasia is still more difficult to
evaluate than tumor grade (29% of atypical hyperplasia
women had cancer at final pathology). These results
underline the importance of treating women with atypical
endometrial hyperplasia as having an unrecognized endo-
metrial cancer.
Previously, hysteroscopy was avoided in women sus-
pected of having endometrial cancer because of the risk of
endometrial cancer cells spreading into the peritoneal cav-
ity. Studies indicate, however, that even though hysteros-

ª 2013 The Authors

Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545 543

Preoperative staging of endometrial cancer
copy may increase the number of women with positive
cytology, this does not seem to worsen their prognoses (27
–29). The FIGO international staging committee has
changed the staging system so that positive peritoneal
cytology does not affect tumor stage (30). In the present
study, eight of 103 women with known cytology had posi-
tive peritoneal cytology after hysteroscopy. All but two had
intra-abdominal spread at primary surgery, and five of
eight women had grade 3 tumors. We monitored our
women for an average of 2.8 years, and the preliminary
survival data seem in line with earlier published data (4,5).
We believe that the increased knowledge from preoper-
ative staging will facilitate the introduction of minimal
invasive surgery for early staged disease and optimize the
planning of extensive surgery.
Conclusion
With an accuracy of 81%, preoperative staging with MRI
and hysteroscopy-directed biopsy can identify women at
high risk of lymph node metastasis and spare 82% of
low-risk women lymph node resection. Hysteroscopy-
directed biopsy had the highest accuracy for evaluating
cervical involvement.
Acknowledgments
We thank Morten Frydenberg for statistical advice, and
secretaries Inge Udsen and Lisbeth Conrad and the sur-
geons Charlotte Søgaard, Jan Blaakjær, Jørgen Præst, and
Thora Christiansen for their help with the project.
Funding
Financial support was gratefully received from the Danish
Cancer Society, M.D. og H. From Haderslevs Fond,
Forsknings Initiative

Arhus Universitet, FIGO Fonden, Sa-
nitorielæge Ellen Pedersens Fond, and Den Classenske
Fideicommis Kirurgiske Fond.
References
1. Seracchioli R, Solfrini S, Mabrouk M, Facchini C, Di
Donato N, Manuzzi L, et al. Controversies in surgical
staging of endometrial cancer. Obstet Gynecol Int.
2010;2010:181963.
2. Chan JK, Kapp DS. Role of complete lymphadenectomy
in endometrioid uterine cancer. Lancet Oncol. 2007;8:
831–41.
3. Holland CM. The role of radical surgery in carcinoma of
the endometrium. Clin Oncol. 2008;20:448–56.
4. Poulsen HK, Jacobsen M, Bertelsen K, Andersen JE,
Ahrons S, Bock JE, et al. Patients with early stages of
544 Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545
16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G. Ørtoft et al.
endometrial cancer should be spared adjuvant
radiotherapy. Danish Endometrial Cancer Group. Int J
Gynecol Cancer. 1996;6:38–43.
5. Bertelsen K, Ortoft G, Hansen ES. Survival of Danish
patients with endometrial cancer in the intermediate-risk
group not given postoperative radiotherapy. Int J Gynecol
Cancer. 2011;21:1191–9.
6. Creutzberg CL, van Putten WL, Koper PC, Lybeert ML,
Jobsen JJ, Warlam-Rodenhuis CC, et al. Surgery and
postoperative radiotherapy versus surgery alone for
patients with stage-1 endometrial carcinoma: Multicentre
randomised trial-PORTEC Study Group. Post operative
radiation therapy in endometrial carcinoma. Lancet.
2000;355:1404–11.
7. ASTEC/EN.5 Study Group, Blake P, Swart AM, Orton J,
Kitchener H, Whelan T, et al. Adjuvant external beam
radiotherapy in the treatment of endometrial cancer (MRC
ASTEC and NCIC CTG EN.5 randomised trials): pooled
trial results, systematic review, and meta-analysis. Lancet
2009;373:137–46.
8. Alders J, Abeler J, Kolstad P, Onsrud M. Postoperative
external irradiation and prognosis parameters in stage I
endometrial carcinoma: clinical and histopathogenic study
of 540 patients. Obstet Gynecol. 1980;56:419–26.
9. Hogberg T, Rosenberg P, Kristensen G, de Oliveira CF, de
Pont Christensen R, Sorbe B et al. A randomized phase-III
study on adjuvant treatment with radiation
(RT)  chemotherapy (CT) in early-stage high-risk
endometrial cancer (NSOG-EC-9501/Eortec 55991)). J Clin
Oncol. 2007;25:5503.
10. Maggi R, Lissoni A, Spina F, Melpignano M, Zola P,
Favalli G, et al. Adjuvant chemotherapy vs radiotherapy in
high-risk endometrial carcinoma: results of a randomised
trial. Br J Cancer. 2006;95:266–71.
11. Susumu N, Sagae S, Udagawa Y, Niwa K, Kuramoto H,
Satoh S, et al. Randomized phase III trial of pelvic
radiotherapy versus cisplatin-based combined
chemotherapy in patients with intermediate- and high-risk
endometrial cancer: a Japanese Gynecologic Oncology
Group study. Gynecol Oncol. 2008;108:226–33.
12. Bernardini MQ, May T, Khalifa MA, Bland AE, Nofech-
Mozes S, Berchuck A, et al. Evaluation of two
management strategies for preoperative grade 1
endometrial cancer. Obstet Gynecol. 2009;114:7–15.
13. Traen K, Hølund B, Mogensen O. Accuracy of
preoperative tumor grade and intraoperative gross
examination of myometrial invasion in patients with
endometrial cancer. Acta Obstet Gynecol Scand.
2007;86:739–41.
14. Franchi M, Ghezzi F, Melpignano M, Cherchi PL,
Scarabelli C, Apolloni C, Zanaboni F. Clinical value of
intraoperative gross examination in endometrial cancer.
Gynecol Oncol. 2010;76:357–61.
15. Dueholm M, Lundorf E, Hansen ES, Ledertoug S, Olesen
F. Accuracy of magnetic resonance imaging and
ª 2013 The Authors

G. Ørtoft et al.
transvaginal ultrasonography in the diagnosis, mapping,
and measurement of uterine myomas. Am J Obstet
Gynecol. 2002;186:409–15.
16. Celik C, Ozdemir S, Esen H, Balci O, Ylmaz O. The
clinical value of preoperative and intraoperative
assessments in the management of endometrial cancer. Int
J Gynecol Cancer. 2010;20:358–62.
17. Ozdemir S, Celik C, Emlik D, Kiresi D, Esen H.
Assessment of myometrial invasion in endometrial cancer
by transvaginal sonography, Doppler ultrasonography,
magnetic resonance imaging and frozen section. Int J
Gynecol Cancer. 2009;19:1085–90.
18. Karlsson B, Norstr€ om A, Granberg S, Wikland M. The use
of endovaginal ultrasound to diagnose invasion of the
endometrial carcinoma. Ultrasound Obstet Gynecol.
1992;2:35–9.
19. Savelli L, Ceccarini M, Ludovisi M, Fruscella E, De Iaco
PA, Salizzoni E, et al. Preoperative local staging of
endometrial cancer: transvaginal sonography vs. magnetic
resonance imaging. Ultrasound Obstet Gynecol.
2008;31:560–6.
20. Rieck GC, Bulman J, Whitaker R, Leeson SC. A
retrospective review of magnetic resonance imaging in
assessing the extent of myometrial infiltration for patients
with endometrial carcinoma. J Obstet Gynecol.
2005;25:765–8.
21. Rockall AG, Meroni R, Sohaib SA, Reynolds K, Alexander-
Sefre F, Shepherd JH, et al. Evaluation of endometrial
carcinoma on magnetic resonance imaging. Int J Gynecol
Cancer. 2007;17:188–96.
22. McComiskey MH, McCluggage WG, Grey A, Harley I,
Dobbs S, Nagar HA. Diagnostic accuracy of magnetic
resonance imaging in endometrial cancer. Int J Gynecol
Cancer. 2012;22:1020–5.
ª 2013 The Authors
Acta Obstetricia et Gynecologica Scandinavica ª 2013 Nordic Federation of Societies of Obstetrics and Gynecology 92 (2013) 536–545
16000412, 2013, 5, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/aogs.12103 by National Health And Medical Research Council, Wiley Online Library on [31/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Preoperative staging of endometrial cancer
23. Duncan KA, Drinkwater KJ, Frost C, Remedios D, Barter
S. Staging cancer of the uterus: a national audit of MRI
accuracy. Clin Radiol. 2012;67:523–30.
24. Cutillo G, Cignini P, Visca P, Vizza E, Sbiroli C.
Endometrial biopsy by means of the hysteroscopic
resectoscope for the evaluation of tumor differentiation in
endometrial cancer: a pilot study. Eur J Surg Oncol.
2007;33:907–10.
25. Larson DM, Johnson KK, Broste SK, Krawisz BR, Kresl JJ.
Comparison of D&C and office endometrial biopsy in
predicting final histopathologic grade in endometrial
cancer. Obstet Gynecol. 1995;86:38–42.
26. Suh-Burgmann E, Hung Y, Armstrong M. Complex
atypical endometrial hyperplasia. The risk of unrecognized
adenocarcinoma and value of preoperative dilation and
cvurettage. Obstet Gynecol. 2009;114:523–9.
27. de la Cuesta SR, Espinosa JA, Crespo E, Granizo JJ, Rivas
F. Does fluid hysteroscopy increase the stage or worsen the
prognosis in patients with endometrial cancer? A
randomized controlled trial. Eur J Obstet Gynecol Reprod
Biol. 2004;10:211–5.
28. Cicinelli E, Tinelli R, Colafiglio G, Fortunato F, Fusco A,
Mastrolia S, et al. Risk of long-term pelvic recurrences
after fluid minihysteroscopy in women with endometrial
carcinoma: a controlled randomized study. Menopause.
2010;17:511–5.
29. Takac I, Zegura B. Office hysteroscopy and the risk of
microscopic extrauterine spread in endometrial cancer.
Gynecol Oncol. 2007;107:94–8.
30. FIGO Committee on gynecologic oncology. Revised
FIGO staging for carcinoma of the vulva, cervix, and
endometrium. Int J Gynaecol Obstet. 2009;105:
103–4.
545