Human Reproduction vol.14 no.10 pp.
2644–2650, 1999
Evaluation of ectopic pregnancy by magnetic resonance
imaging
Milliam L.Kataoka1, Kaori Togashi2,4,
Hisataka Kobayashi2, Takuya Inoue3, Shingo Fujii3,
and Junji Konishi1
1Department
of Radiology and Nuclear Medicine, Faculty of
Medicine, Kyoto University, 54 Shogoin Kawahara-cho, Sakyo-ku,
Kyoto-shi, Kyoto-fu, 2Hitachi Medical Corporation Chair of
Department of Diagnostic and Interventional Imageology, Faculty
of Medicine, Kyoto University, 54 Shogoin Kawahara-cho,
Sakyo-ku, Kyoto-shi, Kyoto-fu, 3Department of Obstetrics and
Gynecology, Faculty of Medicine, Kyoto University,
54 Shogoin Kawahara-cho, Sakyo-ku, Kyoto-shi, Kyoto-fu,
606-8507, Japan.
4To
whom correspondence should be addressed
Patients (n J 37) suspected of ectopic pregnancy were
prospectively evaluated with magnetic resonance (MR)
imaging to assess the capability of MR imaging in the
diagnosis of ectopic pregnancy. Five levels of confidence
were defined: diagnostic, suspicious, equivocal, question-
able, and negative. Tubal wall enhancement and presence
of tubal haematoma or gestational sac-like structure were
considered diagnostic findings. There were 21 diagnostic,
two suspicious, eight equivocal, and six negative findings.
MR findings were compared with the surgical findings in
18 patients. Surgical confirmation was obtained in 12
diagnostic, two suspicious, and four equivocal studies.
Using the MR diagnostic criteria for tubal pregnancy,
MR had 12 true positive, three true negative, three false
negative, and no false positive results for the diagnosis
of tubal pregnancy. Retrospective analysis of the signal
intensity of haematoma and ascites was performed for
these 18 surgically confirmed cases. The predominant signal
intensity of tubal haematoma was an intermediate signal
on T1-weighted image (WI) and a low signal on T2WI.
Ascites showed signal intensity higher than that of urine
on T1WI in 100% of 13 cases. In conclusion, MR imaging
with use of intravenous contrast material allows a specific
diagnosis of tubal pregnancy, recognizing tubal wall
enhancement and fresh tubal haematoma.
Key words: ectopic pregnancy/Fallopian tube/MRI/tubal
pregnancy
Introduction
Recent advances in qualitative serum assays of the β subunit
of human chorionic gonadotrophin (β-HCG) have allowed
detection of earlier and smaller ectopic pregnancies. An early
detection of ectopic pregnancy has led the trend toward more
conservative treatment. Current therapeutic options consist of
2644
expectant management (close observation), medical treatment,
in which methotrexate (MTX) is the agent most often used,
and surgery. MTX injection is a promising non-traumatic
technique for resolving early tubal pregnancy. An accurate
non-invasive diagnosis is especially important to introduce
this new technique. Transvaginal sonography is the imaging
modality of choice; however, due to intrinsic limitations in
tissue characterization, it cannot always produce a definitive
diagnosis of tubal localization of the pregnancy. Magnetic
resonance (MR) imaging has an excellent tissue inherent
contrast, especially in characterizing blood. Cellular structural
differences allow in-vivo visualization of the myometrial zonal
anatomy (a subendometrial layer or junctional zone and an
outer myometrial layer) by T2-weighted MR imaging (Brosens
et al., 1998). Previous reports demonstrated the value of MR
imaging in identifying a haemorrhagic adnexal mass suggestive
of ectopic pregnancy (Hricak et al., 1985; Ha et al., 1993;
Togashi et al., 1994; Kuhl et al., 1995; Momma et al., 1995;
Hamada et al., 1997). However, to our knowledge, no prior
English article has described the criteria for the diagnosis of
tubal pregnancy on MR imaging. We carried out this prospec-
tive study to evaluate the diagnostic accuracy of contrast
enhanced MR imaging for detection of early tubal pregnancy.
Materials and methods
A prospective study was undertaken of all 37 consecutive women
clinically suspected to have an ectopic pregnancy and who underwent
MR image evaluation at our institution between January 1990 and
March 1997. One patient with findings of cervical pregnancy was
excluded. One patient was suspected to have an ectopic pregnancy
on two different occasions, which made a total of 37 cases studied.
All the patients were clinically stable and asymptomatic or minimally
symptomatic. Symptoms observed were minimal vaginal bleeding
(n 5 17) and slight abdominal pain (n 5 13). Ectopic pregnancy was
suspected by clinical history (amenorrhoea, post-curettage, etc.),
positive pregnancy test, and absence of intrauterine pregnancy on
sonography despite their gestational (menstrual) age or high HCG
concentration. All cases studied had already undergone or were
scheduled subsequently to undergo dilatation and curettage (D & C)
(n 5 29 and n 5 8 respectively). Sonographic studies were done
prior to MR imaging in all patients. An exact HCG value (last
measurement before the MR exam) was obtained in 33 cases, (,2–
33 804 mIU/ml, mean 5159 mIU/ml), of which 16 (48%) showed
HCG concentrations ø1000 mIU/ml. Note that an intrauterine gesta-
tional sac seen by transvaginal sonography should coincide with a
HCG value ù1000 mIU/ml (International Reference Preparation)
(Cacciatore et al., 1990). Ten (27%) patients studied were pregnant
with an assisted fertilization, and four (11%) patients studied had a
history of a previous ectopic pregnancy.
The gold standard for the final diagnosis of ectopic pregnancy was
© European Society of Human Reproduction and Embryology

a consensus between surgery, serial HCG concentration, D & C, and
sonographic findings. Medical records were reviewed in all patients.
Among these 37 cases, 18 (aged 20–40 years; mean 29) underwent
surgery and were confirmed to have ectopic pregnancy. Appropriate
high HCG concentrations at initial presentation accompanied by
falling concentrations after MTX administration or expectant follow-
up were considered as cases of true ectopic pregnancy (n 5 14).
Four cases who were followed expectantly, had low concentrations
of HCG at initial presentation, but who showed adnexal mass on
initial sonography, were also considered as cases of true ectopic
pregnancy. One patient who was followed expectantly, had low
concentrations of HCG at initial presentation and no abnormal
findings on sonography, was excluded from the study because these
observations could represent either an aborted pregnancy, or a topic
or ectopic one. Consequently, 36 cases (aged 20–41 years; mean 29)
were considered to have ectopic pregnancy and comprised this study.
Eighteen patients underwent surgery, and the MR findings of these
patients were correlated to the surgical findings to assess the accuracy
of MR imaging.
MR imaging was performed with a use of 1.5 T superconductive
system (Sigma, General Electric Medical Systems, Milwaukee,
Wisconsin, USA), with a body coil. Sagittal T1 and T2-weighted
images (WI) and contrast enhanced T1WI with gadopentetate dimeglu-
mine (GD-DTPA) (Magnevist; Schering, Osaka, Japan) (0.1 mmol/kg)
were obtained by conventional spin-echo (SE) technique. The para-
meters for T1WI were 600/20 ms repetition time/echo time (TR/TE)
and for T2WI were 2000/70–90 ms (TR/TE). Complementary axial
images were obtained in six cases. Axial and coronal images were
not obtained in all patients due to time constraints in our service.
The slice thickness was 5 mm, with a 2.5 mm gap, 2563192 matrix,
the number of signals averaged two with a 32 cm field of view.
Prospective reading
MR scans were interpreted knowing the ultrasound and laboratory
results. MR imaging findings were prospectively assessed for the
presence or absence of dilatation and contrast enhancement of the
Fallopian tube, haematoma, gestational sac (GS)-like structure, and
ascites. An enhancing tubular structure close to the ovary was
diagnosed as Fallopian tube. GS-like structure was defined as a sac-
like cystic structure surrounded by a thick enhancing ring, similar to
that seen in sonography (Atri et al., 1992).
Each haematoma and GS-like structure was localized as tubal,
adnexal or abdominal. Tubal localization of haematoma or GS-like
structure was established when they were observed within the
enhancing tubular structure (dilated Fallopian tube) or in the uterine
cornus (interstitial).
MR imaging was scored to indicate whether a tubal pregnancy
was demonstrated, and with what confidence, as follows: (i) diagnostic
(definite tubal pregnancy: presence of wall enhancement of dilated
Fallopian tube filled with haematoma or GS-like structure), (ii) sus-
picious [dilated enhancing tube filled with only a non-specific fluid
or only presence of haematosalpinx (cylindrical bloody structure
without enhancement)], (iii) equivocal (haematoma or GS-like struc-
ture without identifiable tubal findings), (iv) questionable (only
ascites), (v) negative.
An addendum to this scoring system was the presence of haematoma
or GS-like structure in the uterine cornus. According to previously
published MR data (Ha et al., 1993; Bassil et al., 1995; Yamashita
et al., 1995; Hamada et al., 1997), this finding was considered
diagnostic for interstitial pregnancy, and was included in the diagnostic
group (i) for tubal pregnancy, even in the absence of identification
of Fallopian tube.
The therapeutic options in our institution were expectant manage-
Evaluation of ectopic pregnancy by MRI
ment (close observation), medical treatment with MTX, and surgery.
The treatment decision followed criteria amply discussed (Ylostalo
et al., 1992; Stovall and Ling, 1993; Falk, 1994; Leventhal, 1994;
Atri et al., 1996).
Retrospective reading
In the second portion of this study, MR images of all 18 surgically
confirmed cases were retrospectively reviewed by two radiologists
(KT and MLK), in conference, to evaluate signal intensity of
haematoma (tubal, adnexal, abdominal) and ascites. Signal intensity
of the haematoma was classified as high, intermediate, or low. On
T1WI, high signal intensity meant signal similar to or higher than
the bone marrow, intermediate signal intensity meant signal similar
to the myometrium, and low signal intensity meant signal similar to
simple fluid. On T2WI, high signal intensity meant signal intense of
simple fluid, intermediate signal intensity meant signal similar to
outer myometrium, and low signal intensity meant signal similar to
signal void or skeletal muscle.
Signal intensity of the ascites was classified as intermediate [if
comparable with water (urine)], high (if signal intensity was similar
to myometrium) or prominently high (if signal intensity was higher
than signal of the myometrium) on T1WI. On T2WI, signal intensity
was compared with the signal of urine, and classified as high,
intermediate or low.
Results
Prospective reading
Based on the gold standard consensus specified above, ectopic
pregnancy was considered as the final diagnosis in 36 cases.
MR findings of all cases are summarized in Table I. There
were 21 (58%) diagnostic (Figure 1 and 2), two (5%) suspicious
(Figure 3), eight (22%) equivocal (Figure 4), and five (14%)
negative findings. Fallopian tube was identified as an enhancing
tubular structure with tubal haematoma in 18 (Figure 1 and 2)
and without tubal haematoma or GS-like structure in two.
There were three cases of interstitial pregnancy (two GS-like
structure and one haematoma in the uterine cornus) classified
as diagnostic for tubal pregnancy. Thus, in total, there were
21 diagnostic findings. Haematomas were observed in 25
(69%) cases, including 19 tubal (including one interstitial), 10
adnexal and nine abdominal. Six (17%) cases had GS-like
structure, including three tubal (two being interstitial), two
adnexal and one in rudimentary horn. One case had both tubal
haematoma and GS-like structure. Ascites was observed in 20
(55%) cases.
Table II relates MR findings to treatment results. Eighteen
(50%) cases underwent surgery, including five cases operated
after initiating MTX treatment because of increasing symptoms
or free fluid or decreasing haemoglobin concentration. High
HCG concentrations (17 169 mIU/ml and 1950 mIU/ml)
justified MTX administration in two cases which were negative
on MR imaging. Three cases negative on MR imaging,
followed expectantly, had low concentrations of HCG (,2–
572 mIU/ml) at initial presentation.
MR findings in 18 surgically confirmed cases were compared
with surgical findings in Table III. Enhancing tubular structures
were identified in 14 cases. Out of 12 diagnostic, two suspi-
cious, and four equivocal studies (Table II), all diagnostic and
2645
M.L.Kataoka et al.

Table I. MR findings in patients with proven ectopic pregnancy (n 5 36)

MR findings Diagnostic Suspicious Equivocal Questionable Negative Total

(n 5 21) (n 5 2) (n 5 8) (n 5 0) (n 5 5) n (%)

Tubal dilatation and wall enhancement 18 2 0 0 0 20 (55)

Tubal haematoma 19a 0 0 0 0 19 (53)
Adnexal (not tubal) haematoma 7 1 2 0 0 10 (28)
Abdominal (not adnexal) haematoma 6 0 3 0 0 9 (25)
Gs-like structure 3b 0 3 0 0 6 (17)
Ascites 15 1 4 0 0 20 (55)
None 0 0 0 0 5 5 (14)

GS 5 gestational sac.
Numbers in parentheses are percentages of totals.
aIncluding one interstitial haematoma.
bIncluding two interstitial GS.
Diagnostic group includes three cases of interstitial pregnancy.

Figure 1. Magnetic resonance (MR) imaging from a 29-year-old woman at 7 weeks 1 day gestation, post-dilatation and curettage (D & C)
(chorionic tissue absent), HCG 5 1000 mIU/ml. An example of a diagnostic study-case of tubal pregnancy. (a) Sagittal T1-weighted spin-
echo (SE) MR image (TR/TE 5 600/20) reveals a hyperintense area (white arrow) with surrounding ascites of intermediate signal, posterior
to the uterus (U). (b) Sagittal T2-weighted SE MR image (TR/TE 5 2000/70): haematoma (black arrow) is obvious within ascites.
Haematoma is elliptical in shape, and exhibits distinct hypointensity mixed with hyperintensity. The distinct hypointensity areas indicate
fresh blood. (c) Post-contrast T1-weighted SE MR image (TR/TE 5 600/20) shows well-enhancing tubal wall (arrowheads) identified
around the haematoma. Surgery confirmed unruptured left tubal pregnancy and bloody ascites.

suspicious studies proved to have tubal pregnancy. The four
equivocal cases proved to have tubal pregnancy (n 5 1),
abdominal pregnancy (n 5 2), and pregnancy in rudimentary
horn (n 5 1). In 15 cases of tubal pregnancy, 13 were localized
in the ampullary portion and two in the isthmic portion. Using
the diagnostic criteria for tubal pregnancy (tubal enhancement
and haematoma), MR gave 12 true positive, three true negative,
and three false negative results, for the diagnosis of tubal
pregnancy, distinguishing tubal pregnancies from other types
of ectopic pregnancy. No false positive result was obtained.
The three false negatives obtained included one patient at
10 weeks gestation and HCG .4000 mIU/ml on admission
date. MR examination was obtained at gestational age of 11
weeks and 6 days, and rising HCG (16 000 mIU/ml), revealing
findings suspicious of tubal pregnancy. Two days after adminis-
tration of MTX, this patient was urgently operated and a
432 cm haematoma was found in ampullary portion of the
tube. Another patient was at 2 weeks and 3 days gestation and
HCG 778 mIU/ml on initial presentation. MR exam was
performed at 4 weeks and 2 days gestation and rising HCG
2646
(22 500 mIU/ml). MTX was administered following MR
examination and HCG decreased to 3320 mIU/ml. Seven days
later, an emergency salpingectomy was performed and an
ampullary portion dilated by a 635 cm haematoma was
identified. The other false negative patient was seen at 2 weeks
and 4 days gestation and HCG 1280 mIU/ml on admission,
and an MR examination was performed on the same day. Two
days later HCG concentration fell to 960 mIU/ml, and on the
following day an emergency laparotomy was performed. A
332 cm haematoma was observed near rupture in the ampullary
portion of the tube. These false negatives could not be
diagnosed retrospectively on MR imaging.
Of 10 patients pregnant after assisted fertilization, three
underwent surgery, five MTX treatment and two, expectant
management.
Retrospective reading
The results of the evaluation of signal intensity of the
haematoma are summarized in Table IV.
The presence of haematoma was confirmed in all cases.
 Evaluation of ectopic pregnancy by MRI

Figure 2. A 27-year-old woman at 8 weeks gestation suspected of ectopic pregnancy, post-D & C (chorionic tissue absent), HCG 5 256
mIU/ml. An example of a diagnostic study-case of tubal pregnancy. (a) Endovaginal sonography in transverse plane through right adnexa
shows ill-defined heterogeneous mass (calipers), which is a non-specific finding for the diagnosis of ectopic pregnancy. C. UT 5 corpus
uterine. (b) Sagittal T1-weighted SE MR image (TR/TE 5 600/20) shows tiny hyperintense area (black arrow). (c) Sagittal T2-weighted SE
MR image (TR/TE 5 2000/70) shows haematoma (white arrow) of 1 cm in diameter which has distinct hypointensity with central high
intensity, in proximity to the hyperintensity on T1WI (black arrow). (d) On post-contrast T1-weighted SE MR image (TR/TE 5 600/20) a
well enhancing tubular structure (arrowheads) is displayed around and behind the haematoma (white arrow). Note that the parallel lines are
twisted.

The predominant signal intensity of tubal haematoma was
intermediate on T1WI (n 5 10), and distinctly low on T2WI
(n 5 7) (Figure 1).
Ascites showed signal intensity higher than that of urine on
T1WI in 100% of 13 cases (high intensity in seven cases and
predominantly high intensity in six cases). The signal intensity
results on T2WI were four intermediate, six high and three
low signal intensity.
Discussion
An ectopic pregnancy is a clinical condition in which implanta-
tion and growth of the fertilized ovum occur in an area other
than the uterine cavity. The most common site (about 98%) is
the Fallopian tube, especially in the ampullary portion (Chow
and Lindahl, 1979). The prevalence of ectopic pregnancy has
increased (Atri et al, 1996) due to an increase in risk factors
(Pellerito et al., 1992). Because many ectopic pregnancies are
now being detected earlier and in an unruptured condition, the
majority of the patients present with no or minimal symptoms,
which can be evaluated and managed in a timely manner.
Surgery is still the standard treatment for ectopic pregnancy
(Leventhal, 1994); however, there is a recent trend towards
treating ectopic pregnancy either medically or not at all
(expectant management) (Ylostalo et al., 1992; Stovall and
Ling, 1993). These medical regimens most often use MTX
administered systemically or locally into the ectopic sac or
haematosalpinx. Early institution of medical therapy with MTX
diminishes the potential risk of rupture, decreases patient
morbidity and health care costs, avoiding surgery and hospital-
2647
M.L.Kataoka et al.

Figure 3. A 31-year-old woman at 4 weeks 2 days gestation, post-D & C (chorionic tissue absent), HCG 5 22 500 mIU/ml; example of a
suspicious study-case. (a) Sagittal T1-weighted SE MR image (TR/TE 5 600/20); and (b) sagittal T2-weighted SE MR image (TR/TE 5
2000/70). Behind the ovary (O), a cystic structure (white arrow) and small haematoma (black arrow) is observed. The haematoma exhibits
intermediate signal mixed with hyperintensity on T1WI (a), and distinct hypointense signal mixed with hyperintense area on T2WI (b).
(c) On post-contrast T1-weighted SE MR image (TR/TE 5 600/20), thick enhancing wall (arrowheads) is identified around the cystic
structure. Although the Fallopian tube is surrounded by haematoma, its content is non-specific fluid. After MTX injection, surgery was
performed due to abdominal pain, and confirmed left tubal pregnancy.

Figure 4. A 33-year-old woman at 3 weeks 1 day gestation, pre D & C, HCG 5 967 mIU/ml; an example of an equivocal study-case.
(a) Sagittal T1-weighted SE MR image (TR/TE 5 600/20). (b) Sagittal T2-weighted SE MR image (TR/TE 5 2000/70). (c) Post-contrast
T1-weighted SE MR image (TR/TE 5 600/20). Between the uterus and the bladder, a haematoma (arrows) is observed and shows
intermediate signal on T1WI (a) and distinct low intensity on T2WI (b). However, there is no enhancing tubular structure within or around
the haematoma. Presence of acute-phase haematoma is strongly indicative of ectopic pregnancy; however, its localization is unclear because
of an absence of enhancing tubular structure. Laparotomy revealed an abdominal pregnancy.

ization (Emerson et al., 1992; Falk, 1994). This approach
requires a highly accurate non-invasive test for the evaluation
of patients with suspected ectopic pregnancy.
Endovaginal sonography is the current alternative to diag-
nostic laparoscopy for the diagnosis of ectopic pregnancy.
It has 84.4% sensitivity, 98.9% specificity, 96.3% positive
predictive value, and 94.8% negative predictive value for
accuracy in the diagnosis of ectopic pregnancy on the basis
of observing any adnexal mass except for a simple cyst or an
intra-ovarian lesion (Brown and Doubilet, 1994). The most
highly specific finding is the identification of an extra-uterine
GS that contains a yolk sac or an embryo (with or without
cardiac activity). However, the most common adnexal finding
2648
is an echogenic homogeneous or heterogeneous, rounded, or
elongated solid structure, located outside but in proximity to
the ovary, indicating haematosalpinx (Atri et al., 1992). In
tubal pregnancy, the Fallopian tube is enlarged by products of
conception, blood, and blood clots (Rottem et al., 1990; Atri
et al., 1992). Sonography identifies this haematosalpinx based
on echogenic and morphological characteristics. Unfortunately
this diagnosis is not specific since sonography is not able to
characterize blood. On the other hand, MR imaging is capable
not only of identifying blood but also of determining the age
of blood products as acute (intermediate signal intensity on
T1WI and marked low signal intensity on T2WI), subacute
(peripheral high signal intensity with a distinct central area of
 Evaluation of ectopic pregnancy by MRI

Among six patients who underwent surgery and had been

Table II. Relationship between MR findings and treatment results (n 5 36) classified as suspicious or equivocal by MR imaging, tubal
MR finding Treatment pregnancy was found in three and other types of ectopic
pregnancy were confirmed in the other three. Other clinical
Surgery MTX Observation conditions that might be included in the differential diagnosis
Diagnostic (n 5 21) 12 8 1 are salpingitis or pyosalpinx for suspicious cases, and other
Suspicious (n 5 2) 2 0 0 haemorrhagic disease entities (such as corpus luteum,
Equivocal (n 5 8) 4 3 1 haematoma, or endometriosis) for the equivocal group. How-
Questionable (n 5 0) 0 0 0
Negative (n 5 5) 0 2 3 ever, these conditions can be excluded by clinical history.
Total 18 13 5 Appropriate high HCG concentrations justified MTX
administration in two cases negative on MR imaging. These
MTX 5 methotrexate.
The three cases diagnosed as interstitial pregnancies on MR were treated cases show that, although highly specific in the diagnosis of
with MTX. tubal pregnancy, MR imaging can produce false negative
results.
The predominant signal intensity of the tubal haematomas
Table III. MR findings compared with surgical findings (n 5 18) was intermediate on T1WI, and distinct hypointensive on
MR findings Surgical findings Total T2WI, which means an acute phased haematoma (Rubin
et al., 1987). Identification of fresh blood seems to be highly
Tubal pregnancy EP important in the diagnosis of early tubal pregnancy, and MR
(n 5 15) (n 5 3)
imaging is unique as the only modality to enable identification
Tubal dilatation and wall enhancement 14 0 14 of the stage of a haematoma.
Tubal haematoma 12 0 12 It is not our purpose to replace sonography as the standard
Adnexal (not tubal) haematoma 7 0 7
Abdominal (not adnexal) haematoma 4 2 6 imaging study for the diagnosis of tubal pregnancy. MR study
GS-like structure 1 1 2 is proposed to provide additional information for a limited
Ascites 12 1 13 number of patients who need precise diagnosis. For example,
EP 5 ectopic pregnancy. in patients with pre-existing damage in the contralateral tube,
and desiring future pregnancy, where the preservation of
patency in the remaining tube is very important. In our study,
Table IV. Signal intensity of the haematoma in patients treated surgically 10 patients were undergoing treatment for infertility, and
(n 5 18) MR examination allowed a specific early diagnosis of tubal
Spin-echo Signal intensity Tubal Adnexal Abdominal pregnancy, contributing to the decision for determining early
non-surgical therapy.
T1WI predominantly intermediate 10 4 1 In conclusion, MR imaging with contrast enhancement is a
predominantly high 2 3 5
T2WI predominantly low 7 3 2 promising modality for diagnosis of tubal pregnancy, enabling
predominantly high 5 3 3 the recognition of tubal wall enhancement and fresh tubal
other pattern 0 1 1 haematoma.
Total 12 7 6

low signal intensity on T1WI and T2WI) or chronic (entirely
high signal intense on T1WI and T2WI) (Gomori et al., 1985;
Rubin et al., 1987).
Our MR imaging criteria for diagnosis of tubal pregnancy
are based on two findings: one is haematosalpinx and the other
is wall enhancement of dilated Fallopian tube. These criteria
are supported by a report that no lumen was identifiable within
normal Fallopian tubes on specimen studies (Outwater et al.,
1996). In 1998, Outwater et al described the correct identifica-
tion of dilated Fallopian tubes in 31 of 41 study patients and
correct exclusion of dilated Fallopian tubes in a mean 34 of 38
control subjects (Outwater et al., 1998). Increased vascularity of
the tubal wall observed following implantation can explain
wall enhancement (Togashi et al., 1994; Momma et al., 1995).
In our surgically confirmed cases, MR gave 12 true positive,
three true negative, three false negative, and no false positive
results for the diagnosis of tubal pregnancy, distinguishing
them from other rare types of ectopic pregnancy, with the use
of diagnostic criteria (definitive tubal pregnancy indicated by
simultaneous presence of tubal enhancement and haematoma).
References
Atri, M., de Stempel, J., Bret, P.M. et al. (1992) Accuracy of transvaginal
ultrasonography for detection of hematosalpinx in ectopic pregnancy. J. Clin.
Ultrasound, 20, 255–261.
Atri, M., Leduc, C., Gillett, P. et al. (1996) Role of endovaginal sonography
in the diagnosis and management of ectopic pregnancy. Radiographics, 16,
755–774.
Bassil, S., Gordts, S., Nisolle, M. et al. (1995) A magnetic resonance imaging
approach for the diagnosis of a triplet cornual pregnancy. Fertil. Steril., 64,
1029–1031.
Brown, D.L. and Doubilet, P.M. (1994) Transvaginal sonography for
diagnosing ectopic pregnancy: positivity criteria and performance
characteristics. J. Ultrasound Med, 13, 259–266.
Brosens, J.J., Baker, F.G. and de Souza, N.M. (1998) Myometrial zonal
differentiation and uterine junctional zone hyperplasia in the non-pregnant
uterus. Hum. Reprod. Update, 4, 496–502.
Cacciatore, B., Stenman, U.H. and Ylostalo, P. (1990) Diagnosis of ectopic
pregnancy by vaginal ultrasonography in combination with a discriminatory
serum hCG level of 1000 IU/l (IRP). Br. J. Obstet. Gynaecol., 97, 904–908.
Chow, T.T. and Lindahl, S. (1979) Ectopic cervical pregnancy. J. Clin.
Ultrasound, 7, 217–218.
Emerson, D.S., Cartier, M.S., Altieri, L.A. et al. (1992) Diagnostic efficacy
of endovaginal color Doppler flow imaging in an ectopic pregnancy
screening program. Radiology, 183, 413–420.
Falk, R.J. (1994) Methotrexate treatment of tubal pregnancy. Int. J. Fertil.
Menopausal Stud., 39, 197–201.

2649
M.L.Kataoka et al.

Gomori, J.M., Grossman, R.I., Goldberg, H.I. et al. (1985) Intracranial

hematomas: imaging by high-field MR. Radiology, 157, 87–93.
Ha, H.K., Jung, J.K., Kang, S.J. et al. (1993) MR imaging in the diagnosis
of rare forms of ectopic pregnancy. Am. J. Roentgenol., 160, 1229–1232.
Hamada, S., Naka, O., Moride, N. et al. (1997) Ultrasonography and magnetic
resonance imaging findings in a patient with an unruptured interstitial
pregnancy. Eur. J. Obstet. Gynecol. Reprod. Biol., 73, 197–201.
Hricak, H., Lacey, C., Schriock, E. et al. (1985) Gynecologic masses: value
of magnetic resonance imaging. Am. J. Obstet. Gynecol., 153, 31–37.
Kuhl, C.K., Heuck, A., Kreft, B.P. et al. (1995) Combined intrauterine and
ovarian pregnancy mimicking ovarian malignant tumor: imaging findings.
Am. J. Roentgenol., 165, 369–370.
Leventhal, J.M. (1994) Surgical management of tubal ectopic pregnancy. Int.
J. Fertil. Menopausal Stud., 39, 197–201.
Momma, C., Kobayashi, F., Fujihara, H. et al. (1995) Characteristic of
magnetic resonance imaging of ectopic pregnancy. Clinical Gynecology
and Obstetrics, 49, 247–251. [in Japanese]
Outwater, E.K., Siegelman, E.S., Chiowanich, P. et al. (1998) Dilated Fallopian
tubes: MR imaging characteristics. Radiology, 208, 463–469.
Outwater, E.K., Talerman, A. and Dunton, C. (1996) Normal adnexa uteri
specimens: anatomic basis of MR imaging features. Radiology, 201,
751–755.
Pellerito, J.S., Taylor, K.J., Quedens-Case, C. et al. (1992) Ectopic pregnancy:
evaluation with endovaginal color flow imaging. Radiology, 183, 407–411.
Rottem, S., Thaler, I., Levron, J. et al. (1990) Criteria for transvaginal
sonographic diagnosis of ectopic pregnancy. J. Clin. Ultrasound, 18,
274–279.
Rubin, J.I., Gomori, J.M., Grossman, R.I. et al. (1987) High-field MR imaging
of extracranial hematomas. AJR Am. J. Roentgenol., 148, 813–817.
Stovall, T.G. and Ling, F.W. (1993) Single-dose methotrexate: an expanded
clinical trial. Am. J. Obstet. Gynecol., 168, 1759–1765.
Togashi, K., Kobayashi, F.and Mori, T. (1994) Obstetrical imaging: MRI 2.
Ectopic pregnancy. Obstetrical and Gynecological Therapy, 68, 225–234.
[in Japanese]
Yamashita, Y., Harada, M., Torashima, M. et al. (1995) Unruptured interstitial
pregnancy: a pitfall of MR imaging. Comput. Med. Imaging Graph., 19,
241–246.
Ylostalo, P., Cacciatore, B., Sjoberg, J. et al. (1992) Expectant management
of ectopic pregnancy. Obstet. Gynecol., 80, 345–348.

Received on April 15, 1999; accepted on July 12, 1999

2650