Human Reproduction Vol.18, No.4 pp. 858±862, 2003
DOI: 10.1093/humrep/deg191
Undescended ovary and unicornuate uterus:
Simpli®ed diagnosis by the use of clomiphene citrate
ovarian stimulation and magnetic resonance imaging (MRI)
W.Ombelet1,4, M.Grieten2, Ph. DeNeubourg1, G.Verswijvel2, L.Buekenhout3, P.Hinoul1 and
E.deJonge1
1
Department of Obstetrics and Gynecology and 2Department of Radiology, Ziekenhuizen Oost-Limburg, Schiepse Bos 6, 3600 Genk
and 3Department of Obstetrics and Gynecology, St Franciscusziekenhuis, P. Perquaylaan 129, 3350 Heusden-Zolder, Belgium
4
To whom correspondence should be addressed. E-mail: [email protected].
BACKGROUND: Although the association between unicornuate uterus and undescended ovaries is well established,
little information is available on this entity, suggesting the possibility that many cases are not recognized at all.
Consequently, important clinical information is missed in many patients worldwide. METHODS: During a period
of 5 years, eight cases of unicornuate uterus were observed in our infertility clinic. The ®rst three patients received
magnetic resonance imaging (MRI) after mild clomiphene citrate (CC) stimulation due to a discrepancy between
estradiol levels and follicular growth and because a suspected ectopic ovary could not be visualized on ultrasound.
Based on this experience, ®ve consecutive patients were offered MRI after CC stimulation as part of this study.
RESULTS: In ®ve out of eight patients with unicornuate uterus (62.5%) an undescended ovary could be visualized
in the upper abdomen. Abdominal ultrasound revealed the ectopic ovary in only two cases. The use of CC stimulation undoubtedly facilitated the diagnosis of the undescended ovary on MRI. CONCLUSION: It is suggested that
MRI after CC stimulation is an excellent non-invasive method to diagnose undescended ovaries in women with a
unicornate uterus.
Key words: ectopic ovary/infertility/MRI/undescended ovary/unicornuate uterus
Introduction
Undescended or ectopic ovaries are characterized by the
attachment of the upper pole to an area above the level of the
common iliac vessels. Although ovarian maldescent also
occurs in patients with a normal uterus, the incidence is
reported to be 20% when the uterus is absent
(Rokitansky±KuÈstner±Hauser syndrome) and as high as 42%
in cases of unicornuate uterus. Bilaterality occurs more often in
women with congenital absence of the uterus (Dabirashra®
et al., 1994). Although the term ectopic ovary is commonly
used for the condition described here, undescended ovary is
probably a more accurate re¯ection of the underlying
pathophysiology.
The incidence of congenital uterine malformations varies
between 0.1±5.0% and MuÈllerian anomalies are found more
often in subfertile patients compared with fertile and sterile
controls (AcieÂn, 1997; Raga et al., 1997; Grimbizis et al.,
2001). Of all MuÈllerian defects, unicornuate uterus is found
in 3±13% of cases (AcieÂn, 1997; Raga et al., 1997;
Grimbizis et al., 2001). Unicornuate uterus is caused by a
failure of one MuÈllerian duct to develop (unicornuate
uterus without rudimentary horn) or to migrate to its proper
location. It is linked to an increase in obstetrical
858
complications such as early miscarriages, ectopic pregnancy,
abnormal fetal presentation, intrauterine growth retardation
and premature labour (Andrews and Jones, 1982; Heinonen
et al., 1982; Donderwinkel et al., 1992; Moutos et al., 1992;;
Heinonen, 1997; Raga et al., 1997). Despite the well-known
association of ectopic ovaries and unicornuate uterus, ectopic
ovaries are reported only sporadically, suggesting the possibility that many cases go unrecognized (Verkauf and Bernhisel,
1996).
This is the ®rst report on the use of magnetic resonance
imaging (MRI) after ovarian stimulation in the detection of
undescended ovaries. We describe four cases of unicornuate
uterus in which the diagnosis of ectopic ovary during a routine
infertility exploration including hysterosalpingography (HSG),
hysteroscopy and laparocopy was missed.
MRI techniques and methods
Because the cases presented here stretch over a period of 5
years, the imaging technique differed to some degree. All
studies were performed on a 1 Tesla MRI unit (Siemens
Magnetom Expert, Erlangen, Germany). In the ®rst two cases,
axial and coronal turbo spin echo (TSE) T2 weighted images
ã European Society of Human Reproduction and Embryology
MRI for the detection of a unicornuate uterus
Figure 1. (a) An hysterosalpingogram showing a unicornuate uterus with a patent tube on the right side. (b) Axial T2 weighted image (TR =
4600/TE = 99 ms). A unicornuate uterus is demonstrated on the patient's right (arrow). (c) Axial T2-weighted image (TR = 4600/TE = 99
ms). The right ovary is located a few cm above the unicornuate uterus shown in Fig 1b (thin arrow). On the patient's left side a small
`nodular' structure is demonstrated anterior to the external iliac vessels (large short arrow), which is believed to be a MuÈllerian remnant. (d)
Coronal STIR image (TR = 6000/TE = 60 ms). The left ovary is demonstrated at the level of the pelvic brim, containing several follicles that
are organised in a `cord-like' structure.
were obtained (TE = 99 ms/TR = 4600 ms) followed by short
tau inversion recovery (STIR) sequences in the coronal, sagittal
and axial plane (TE = 60 msec/TR = 6000 ms). Since the
stimulated ovary contains numerous ¯uid-containing follicles,
they are easily detected on this sequence. One series of T1weighted images was obtained in the axial plane using a
gradient echo sequence [Turbo Fast Low Angle Shot
(TFLASH) with TE = 4.2 ms/TR = 11 ms] focused on the
ectopic ovary. This sequence is used to study the signal of the
content of the follicles in the latter. Due to technological
improvement over the years we performed axial, coronal and
sagittal T2-weighted images using a non-breath hold halfFourier acquisition single-shot turbo spin-echo (HASTE)
technique (TR 1400 ms/TE 60 ms) in the last two cases. Also
late-echo T2-weighted HASTE images were obtained in the
same imaging planes (TR = 1250 ms/TE = 360 ms). The latter
has the same imaging capabilities as the STIR images where
¯uid has very high contrast compared with the surrounding
tissue. All patients were examined in the supine position with a
four element phased-array torso coil to optimise the signal-tonoise ratio. Slice thickness was 5 mm in all sequences. The
centre of the coil was positioned at the umbilicus of the patient.
The ®eld of view encompassed the upper border of the kidneys
to the bladder base. Neither sedation or i.v.-contrast were
applied. Image quality was adequate to detect the ectopic
location of the ovary contralateral to the missing uterine horn in
all cases. Detection was most adequate on the T2-weighted
images, especially on the heavily weighted HASTE series and
STIR images.
Case 1
A couple with primary infertility for 3 years was referred to our
infertility centre for donor-insemination. Investigation of the
male partner revealed azoospermia. Genetic testing proved to
859
W.Ombelet et al.
be normal. On testicular biopsy the diagnosis of Sertoli-cell
only syndrome was made. An infertility work-up of the
female, including an endocrinological and immunological
evaluation, was normal. An HSG showed an unicornuate uterus
with one patent salpinx on the right side (Figure1a) and this
was con®rmed laparoscopically. A urological examination
including i.v. pyelography (IVP) could not detect any
abnormality.
The couple was entered in our donor-insemination programme. Clomiphene citrate (CC) (50 mg for 5 days, days 5±9)
for mild ovarian stimulation was used. In the ®rst treatment
cycle, we observed a normal bifollicular response with a serum
estradiol level of 544 pg/ml on the day of hCG administration.
This attempt was unsuccessful. In a consecutive cycle, serum
estradiol reached a level of >600 pg/ml, but no ovarian
response could be seen on ultrasound. A spontaneous LH surge
was noted. In this cycle, we performed no insemination.
Because an ectopic ovary was suspected, vaginal and abdominal ultrasound was carried out but could not con®rm the
diagnosis of an undescended ovary. During the next cycle and
following CC stimulation, an MRI examination (Siemens
I.O.T. magneton) was carried out on day 12 of the cycle. This
con®rmed the presence of a unicornuate uterus and revealed a
left-sided undescended ovary at the level of the pelvic brim,
anterior to the psoas muscle, containing several follicles
(Figure 1b±d). A thin cord-like structure continued from this
ovary in the direction of the internal inguinal canal. The latter
contained a small (15 mm diameter) `nodular' structure at the
level of the external iliac vascular axis. This was considered to
be a MuÈllerian remnant or a remnant of the lower gonadal cord.
Considering these ®ndings, the diagnosis of an ectopic ovary
was made. The right ovary was entirely normal.
Figure 2. Sagittal STIR image (TR = 6000/TE = 60 ms). The leftsided ovary (arrow) is located anterior to the psoas and iliac
muscles at the level of the pelvic brim.
Case 2
A 26-year-old woman with primary infertility was referred for
fertility treatment. An HSG, hysteroscopy and laparoscopy
performed in the referring centre had con®rmed the diagnosis
of a unicornuate uterus with one tube and one normal ovary on
the right side. On vaginal ultrasound, only one ovary could be
seen. An infertility work-up of the male revealed moderate
teratozoospermia with 7% normal forms using strict criteria of
sperm morphology (Ombelet et al., 1997a). An IVP showed no
abnormalities.
The couple was advised to start with intrauterine insemination (IUI) after CC ovarian stimulation. On day 15 of the ®rst
IUI cycle, serum estradiol reached 272 pg/ml with only one
follicle of 12 mm diameter visible on vaginal ultrasound.
Because of the discrepancy between follicular size and
estradiol level an MRI was performed con®rming the presence
of a left-sided ectopic ovary in the upper part of the pelvis,
anterior to the intersection of the psoas and the iliac muscle. It
contained numerous follicles (Figure 2).
Case 3
A couple with primary infertility for 2 years was referred for
IVF after two unsuccessful IUI attempts with homologous
semen. A previous investigation revealed a normogonadotrophe anovulation, a unicornuate uterus and only one ovary
860
Figure 3. T2-weighted image in the sagittal plane demonstrating
follicles in the left ovary. The latter has a linear morphology,
reaching the level of L2 (not shown) and the follicles are clustered
in a grape-like fashion (arrow).
and tube on the right side. Unilateral tubal patency was
documented by HSG and laparoscopy.
A male factor was found with unexplained oligoasthenoteratozoospermia. After the washing procedure, >13106 motile
spermatozoa could be recovered and therefore IUI treatment
with CC ovarian hyperstimulation was continued (Ombelet
et al., 1997b). During the next IUI treatment cycle a
discrepancy between serum estradiol levels and ovarian
ultrasound images (follicular growth) was observed. An
ectopic ovary was suspected and MRI was carried out
after mild ovarian CC stimulation. This con®rmed the
presence of a unicornuate uterus and a normal right-sided
ovary containing numerous follicles. A left retroperitoneal
ectopic ovary was seen as a cord-like structure that contained
several follicles. Its cranial border reached the left side of the
second lumbar vertebra, the caudal part of the cord-like ovary
stretched to the internal inguinal canal (Figure 3). The left
kidney was absent.
MRI for the detection of a unicornuate uterus
Figure 4. Axial STIR image (TR = 6000/TE = 60 ms). The right
ovary was found anterior to the iliac muscle, containing several
follicles (arrows).
Case 4
A couple with primary infertility for 6 years as a result of nonobstructive azoospermia was referred after 12 unsuccessful
trials with IUI using donor semen. An HSG and laparoscopy
showed a unicornuate uterus with one normal tube and ovary
on the left side. Abdominal ultrasound was performed but
could not detect an ovary on the right side. Agenesis of the
right kidney was found on IVP.
Considering our experience with the previous cases of
unicornuate uterus, MRI after CC ovarian stimulation was
performed. During that cycle one dominant follicle was
observed with a diameter of 16 mm in the left ovary. Serum
estradiol levels reached 610 pg/ml, again showing a discrepancy between ultrasound ®ndings and serum hormonal level.
MRI con®rmed the presence of an undescended ovary at the
right pelvic brim, anterior to the psoas muscle, containing
several follicles (Figure 4).
An IVF cycle was started and on day 14 an oocyte aspiration
was performed. Seven oocytes were aspirated from the left
ovary. The ectopic ovary could not be observed on abdominal
ultrasound. Six out of seven oocytes fertilized and 2 days after
oocyte retrieval three fertilized oocytes were transferred
laparocopically into the left tube. During this procedure the
ectopic ovary with remnants of a tube (1 cm) and ®mbriae were
seen high up in the abdomen, close to the parietal peritoneum.
Cases 5±8
The next four patients were found to have a unicornuate uterus
on routine fertility exploration in our centre. Because of the
previous experience, MRI after CC stimulation was conducted
in all cases resulting in one extra case of undescended ovary. In
this patient, the undescended ovary could also be visualized on
abdominal ultrasound. In three other patients, two normally
localized ovaries were observed without the presence of an
ectopic ovary. In two patients a rudimentary horn was
diagnosed on MRI.
Discussion
A variety of terms have been used to describe an undescended
or ectopic ovary. Supernumerary ovaries include cases in
which a third ovary is entirely separated from the eutopic ovary
and may be located in the omentum or retroperitoneally. The
term `accessory ovary' refers to cases in which excess ovarian
tissue is situated near and connected to the normally placed
ovary (Wharton, 1959; Printz et al., 1973). Ovarian implant
syndrome covers those cases that develop after pelvic surgery
or after a history of pelvic in¯ammatory disease. It has been
well established that ovarian tissues may continue to function
after transplantation from their original site (Payan and Gilbert,
1987). Because of this confusion in terminology, Lachman and
Berman (1991) proposed a modi®cation of the terminology for
classi®cation of ectopic ovaries. They suggested elimination of
the terms supernumerary and accessory ovary. Instead, they
introduced usage of the term ectopic ovary, which can be
divided in three categories: (i) post-surgical implant, (ii) postin¯ammatory implant and (iii) true embryologic.
The embryological mechanism underlying undescended
ovaries is uncertain, but could be explained by a lack of
caudal descent of the gonads into the true pelvis (Parmley,
1993) or by a retarded differential growth of that portion of the
urogenital ridge giving rise to both the gonad and the Fallopian
tube (Rock et al., 1986). Ectopic ovaries may be unilateral or
bilateral and can be associated with abnormalities of the
MuÈllerian ducts such as unicornuate uterus. The association of
unicornuate uterus and urinary tract anomalies including
ectopic kidney, renal agenesis, double renal pelvis and
horseshoe kidneys is well known (Fedele et al., 1996).
In this study we describe four cases of true embryologic
undescended ovaries, with attachment of the upper pole to an
area above the level of the common iliac vessels. All four cases
were missed by routine fertility exploration. In three cases
(reports 1, 2 and 3) MRI was performed after mild ovarian
stimulation with CC because a discrepancy between serum
estradiol levels and ovarian ultrasound images was found after
ovulation induction. The multifollicular growth in the ectopic
ovary greatly enhanced MRI diagnosis.
State-of-the-art MRI has proven its value in the detection
and characterization of a wide variety of disorders of the
female reproductive organs. Ultra-fast cross-sectional T1- and
T2-weighted images can be used to assess the normal and
pathological ovaries. Usually, MRI is used as an adjunct to
ultrasound and/or computed tomography in selected cases.
However, while reports documenting the value of MRI in
detecting anomalies of the uterus and ovaries are profound in
the literature, reports concerning the detection of ectopic
ovaries are lacking.
As described in these case reports, we used MRI as the
primary imaging tool to detect the location and the appearance
of the contralateral ovary in those patients with an unicornuate
uterus. MRI was chosen because this technique has a superior
soft tissue contrast, has multiplanar imaging capabilities and
lacks ionising radiation. MRI was shown to be more sensitive
in the detection of an undescended ovary compared with
ultrasound.
861
W.Ombelet et al.
A review of the literature revealed one case of a supernumerary ovary in a patient with persistent elevated plasma
estrogen levels despite bilateral oophorectomy. The ectopic
ovary in that particular patient was diagnosed by stimulation
with hCG (Kosasa et al., 1976).
Although the incidence of unicornuate uterus is low, a
correct diagnosis is mandatory not only to be aware of the
existence of an ectopic ovary but also in considering the
reproductive performance when uterine malformations are
involved. Early and late abortions as well as preterm delivery
are found much more often if unicornuate uterus is involved
(Andrews and Jones, 1982; Donderwinkel et al., 1992; Moutos
et al., 1992; Heinonen, 1997; Raga et al., 1997). Concerning
assisted reproduction, signi®cantly lower implantation rates are
described after IVF±embryo transfer in cases of unicornuate
uterus (Heinonen et al., 2000).
The knowledge of the existence of an ectopic ovary is also of
importance and relevant in the following clinical situations:
follicle aspiration techniques in women with unicornuate
uterus, exploration of unexplained cyclic abdominal pain due
to folliculogenesis, ovulation and cyst formation in the
undescended ovary and medical conditions where surgical
castration is indicated.
There are many reasons to promote MRI after mild ovarian
stimulation in all cases of unicornuate uterus. Stimulation with
CC improves the accuracy of the MRI examination. Therefore
we suggest performing routine MRI after CC stimulation in all
unicornuate uterus cases. Whether MRI in a non-stimulated
cycle is as sensitive as MRI after CC stimulation is doubtful,
but has to be investigated.
Continuous re®nement of the MRI technique and further
improvement of the spatial resolution and increased availability of MRI will progressively increase its application in the
detection of undescended ovaries and other MuÈllerian fusion
defects. Since both intra- and extra-peritoneal locations of an
undescended ovary may occur, the authors suggest that MRI
can precede or replace laparoscopic evaluation in this speci®c
patient population (Doyle, 1992).
There are enough reasons to prefer MRI above laparoscopy
in the diagnosis of undescended ovaries: (i) the costs (for
society) of a laparoscopy are much higher compared with MRI;
(ii) in contrast with laparoscopy, MRI is a non-invasive
procedure; and (iii) renal abnormalities and extra-peritoneal
ovaries (which are probably the most important to diagnose)
cannot be evaluated during laparoscopy.
To conclude, MRI after CC ovarian stimulation in all
unicornuate uterus cases may avoid otherwise invasive diagnostic techniques and could detect undescended ovaries more
frequently than has been appreciated to date. We described a
novel and sensitive tool in the diagnosis of undescended
ovaries, an entity which has probably gone unrecognized until
now in many patients at risk.
Acknowledgements
We are indebted to Godfried Benijts, Wilfried Gyselaers, Val
Huysmans, Greet Mestdagh, Martine Nijs, Peter Sieprath and Jos
Vlasselaer for their helpful comments and suggestions. We gratefully
862
acknowledge Patricia Jarkowski for her technical support in preparing
this manuscript.
References
AcieÂn, P. (1997) Incidence of MuÈllerian defects in fertile and infertile women.
Hum. Reprod., 12, 1372±1376.
Andrews, M.C. and Jones, H.W. Jr (1982) Impaired reproductive
performance of the unicornuate uterus: intrauterine growth retardation,
infertility, and recurrent abortion in ®ve cases. Am. J. Obstet. Gynecol., 144,
173±176.
Dabirashra®, H., Mohammad, K. and Moghadami-Tabrizi, N. (1994) Ovarian
malposition in women with uterine anomalies. Obstet. Gynecol., 83,
293±294.
Donderwinkel, P.F., Dorr, J.P. and Willemsen, W.N. (1992) The unicornuate
uterus: clinical implications. Eur. J. Obstet. Gynecol. Reprod. Biol., 47,
135±139.
Doyle, M.B. (1992) Magnetic resonance imaging in MuÈllerian fusion defects.
J. Reprod. Med., 37, 33±38.
Fedele, L., Bianchi, S., Agnoli, B., Tozzi, L. and Vignali, M. (1996) Urinary
tract anomalies associated with unicornuate uterus. J. Urol., 155, 847±848.
Grimbizis, G.F., Camus, M., Tarlatzis, B.C., Bontis, J.N. and Devroey, P.
(2001) Clinical implications of uterine malformations and hysteroscopic
treatment results. Hum. Reprod. Update, 7, 161±174.
Heinonen, P. (1997) Unicornuate uterus and rudimentary horn. Fertil. Steril.,
68, 224±230.
Heinonen, P.K., Saarikoski, S. and Pystynen, P. (1982) Reproductive
performance of women with uterine anomalies. An evaluation of 182
cases. Acta Obstet. Gynecol. Scand., 61, 157±162.
Heinonen, P.K., Kuismanen, K. and Ashorn, R. (2000) Assisted reproduction
in women with uterine anomalies. Eur. J. Obstet. Gynecol. Reprod. Biol.,
89, 181±184.
Kosasa, T.S., Grif®ths, C.T., Shane, J.M., Leventhal, J.M. and Naftolin, F.
(1976) Diagnosis of a supernumerary ovary with human chorionic
gonadotropin. Obstet. Gynecol., 47, 236±237.
Lachman, M.F. and Berman, M.M. (1991) The ectopic ovary. A case report
and review of the literature. Arch. Pathol. Lab. Med., 115, 233±235.
Moutos, D.M., Damewood, M.D., Schlaff, W.D. and Rock, J.A. (1992) A
comparison of the reproductive outcome between women with a
unicornuate uterus and women with a didelphic uterus. Fertil. Steril., 58,
88±93.
Ombelet, W., Bosmans, E., Janssen, M., Cox, A., Vlasselaer, J., Gyselaers, W.,
Vandeput, H., Gielen, J., Pollet, H., Maes, M. et al. (1997a) Semen
parameters in a fertile versus subfertile population: a need for change in the
interpretation of semen testing. Hum. Reprod., 12, 987±993.
Ombelet, W., Vandeput, H., Van de Putte, G., Cox, A., Janssen, M., Jacobs, P.,
Bosmans, E., Steeno, O. and Kruger, T. (1997b) Intrauterine insemination
after ovarian stimulation with clomiphene citrate: predictive potential of
inseminating motile count and sperm morphology. Hum. Reprod., 12,
1458±1463.
Parmley, T. (1993) The embryology of the female genital tract. In Verkauf,
B.S. (ed) Congenital malformations of the female reproductive tract and
their treatment. Appleton and Lang, Norwalk, USA.
Payan, H.M. and Gilbert, E.F. (1987) Mesenteric cyst-ovarian implant
syndrome. Arch. Pathol. Lab. Med., 111, 282±284.
Printz, J.L., Choate, J.W., Townes, P.L. and Harper, R.C. (1973)
The embryology of supernumerary ovaries. Obstet. Gynecol., 41,
1246±1252.
Raga, F., Bauset, C., Remohi, J., Bonilla-Musoles, F., Simon, C. and Pellicer,
A. (1997) Reproductive impact of congenital MuÈllerian anomalies. Hum.
Reprod., 12, 2277±2281.
Rock, J.A., Parmley, T., Murphy, A.A. and Jones, H.W. Jr. (1986) Malposition
of the ovary associated with uterine anomalies. Fertil. Steril., 45, 561±563.
Verkauf, B.S. and Bernhisel, M.A. (1996) Ovarian maldescent. Fertil. Steril.,
65, 189±192.
Wharton, L.R. (1959) Two cases of supernumerary ovary and one case of
accessory ovary with an analysis of previously reported cases. Am. J.
Obstet. Gynecol., 78, 1101±1109.
Submitted on August 19, 2002; resubmitted on November 6, 2002; accepted on
January 9, 2003