“Kissing ovaries”: a sonographic sign of moderate to

severe endometriosis
Fabio Ghezzi, M.D.,a Luigi Raio, M.D.,b Antonella Cromi, M.D.,a Daniela Günter Duwe, M.D.,b
Paolo Beretta, M.D.,a Marco Buttarelli, M.D.,a and Michael D. Mueller, M.D.b
a
Department of Obstetrics and Gynecology, University of Insubria, Del Ponte Hospital, Varese, Italy; and b Endometriosis
Center, Department of Obstetrics and Gynecology, University of Berne, Berne, Switzerland

Objective: To evaluate whether the presence of kissing ovaries at ultrasound is a marker for endometriosis and
whether it correlates with the severity of the disease.
Design: Prospective observational study.
Setting: Gynecologic departments of two university hospitals.
Patient(s): A total of 722 consecutive premenopausal women who had laparoscopic surgery for an adnexal mass
or suspected pelvic endometriosis.
Intervention(s): Preoperative ultrasound evaluation and laparoscopic surgery.
Main Outcome Measure(s): Diagnostic and predictive value of ultrasound identification of kissing ovaries in the
detection of endometriosis.
Result(s): Kissing ovaries were diagnosed at ultrasound and confirmed laparoscopically in 32 patients. Of these,
27 had moderate to severe endometriosis and five others had benign adnexal masses. Bowel (18.5% vs. 2.5%) and
fallopian tube (92.6% vs. 33%) endometriosis were significantly more frequent in patients with kissing ovaries
than in patients without kissing ovaries. In infertile patients (n ⫽ 145), kissing ovaries were associated with a
higher proportion of women with fallopian tube obstruction (80% vs. 8.6%). Considering patients with moderate
to severe endometriosis (n ⫽ 189), the median (range) revised American Fertility Society score (74 [32–148] vs.
35 [16 –146]) and the operative time (115 minutes [65–245 minutes] vs. 50 [15–180 minutes]) were significantly
higher in patients with than in those without kissing ovaries.
Conclusion(s): The detection of kissing ovaries at ultrasound is strongly associated with the presence of
endometriosis and is a marker of the most severe form of this disease. (Fertil Steril威 2005;83:143–7. ©2005 by
American Society for Reproductive Medicine.)
Key Words: Endometriosis, kissing ovaries, ultrasound, laparoscopy, bowel endometriosis

A number of scores have been proposed to interpret the
gray-scale ultrasound appearance of pelvic masses to differ-
entiate benign from malignant tumors (1, 2). The accuracy of
transvaginal ultrasound in evaluating pelvic cysts suspected
to be endometriomas has been widely evaluated (4, 5).
Moreover, the possibility of implementing the conventional
ultrasound assessment of the female pelvis with the use of
color Doppler and power Doppler imaging techniques has
further refined the ability to correctly diagnose endometrio-
mas (6 – 8).
However, even when endometriosis is suspected preoper-
atively, the difficulty in evaluating its extension remains a
major limitation. Previously published studies propose some
tools for preoperative counseling on the basis of the ultra-
sound appearance of the ovarian cysts and their relation to
the anatomical location of the lesions (9 –11). It has been
reported that the involvement of the left side of the pelvis is
associated with a higher recurrence risk and with a worse
pregnancy outcome than endometriosis limited to the right
Received January 26, 2004; revised and accepted May 14, 2004.
Reprint requests: Fabio Ghezzi, M.D., Department of Obstetrics and
Gynecology, University of Insubria, Piazza Biroldi 1, 21100 Varese, Italy
(FAX: 39-0332-299-307; E-mail: fabio.ghezzi@uninsubria.it).
pelvic side (9, 10). Recently, Dessole et al. described a new
procedure, which is based on transvaginal ultrasound com-
bined with the introduction of saline solution in the vagina to
enhance the identification of rectovaginal endometriosis
(11).
An improved preoperative assessment of endometriosis,
particularly the identification of severe forms of the disease,
would allow for better planning of surgery. In these cases, it
is common to observe that the ovaries, although sonographi-
cally normal, are entirely or partly joined together and sta-
bilized behind the uterus in the pouch of Douglas (so-called
kissing ovaries) at the time of laparoscopy. Therefore, we
sought to assess the accuracy of preoperative transvaginal
ultrasound in correctly identifying the presence of kissing
ovaries and to establish whether this finding correlates with
the severity of endometriosis.
MATERIALS AND METHODS
Consecutive premenopausal women with an adnexal mass or
with clinical signs suggestive of pelvic endometriosis and
scheduled for laparoscopic surgery between January 1, 2000,
and November 30, 2003, were enrolled at the participating
institutions. Exclusion criteria were [1] previous surgical

0015-0282/05/$30.00 Fertility and Sterility姞 Vol. 83, No. 1, January 2005 143
doi:10.1016/j.fertnstert.2004.05.094 Copyright ©2005 American Society for Reproductive Medicine, Published by Elsevier Inc.
 FIGURE 1
Transvaginal ultrasound image of “kissing ovaries”
in the presence of a unilateral endometrioma.
Ghezzi. Kissing ovaries at ultrasound and endometriosis. Fertil Steril 2005.
intervention on the adnexae or the uterus, [2] history of
breast, gastrointestinal tract, or genitourinary tract malig-
nancy, [3] history of infertility without symptoms or signs of
endometriosis, and/or [4] clinical or ultrasound suspicion of
malignancy. The study was approved by the Ethical Re-
search Committee of the participating institutions.
Ultrasound evaluation of the pelvis was performed with
commercially available ultrasound units (Sequoia 512, Acu-
son, Mountain View, CA; Aloka 1700 Dyna view II, Aloka,
Tokyo) equipped with a 5-/7.5-MHz transvaginal probe.
Using B-mode, an ovarian endometrioma was suspected in
the presence of round-shaped cysts with thick walls, regular
margins, and homogeneous low echogenicity. The diagnosis
of “kissing ovaries” was made when both ovaries were
joined together behind the uterus in the cul-de-sac (Fig. 1)
and were not separable by pushing the transvaginal probe
and by moving the uterus transabdominally. The presence of
definite endometriomas was not a prerequisite for the diag-
nosis of kissing ovaries (Fig. 2).
All ultrasound examinations were performed within 1
week before surgery by three of the authors. Each case was
classified (kissing ovaries or not), electronically stored, and
sent to the other two authors for review. In case of a
discordant opinion, an agreement was reached after a colle-
gial discussion of the case. If isolated peritoneal endometri-
otic foci were found at surgery in case of a nonendometriotic
adnexal mass, the patient was classified in accordance with
adnexal mass histology.
Bladder endometriosis was defined as the presence of
lesions infiltrating the muscle layer of the bladder wall with
spillage of urine after the removal of the lesion. Ureteral
144 Ghezzi et al. Kissing ovaries at ultrasound and endometriosis
FIGURE 2
Transvaginal ultrasound image of “kissing ovaries”
in the absence of definite endometriomas.
Ghezzi. Kissing ovaries at ultrasound and endometriosis. Fertil Steril 2005.
endometriosis was considered to be present when endometri-
otic tissue was responsible for a ureteral stenosis requiring
ureterolysis. Bowel endometriosis was considered to be
present when one or more nodules were touching the muscle
layer of the intestinal wall and the bowel mucosa was visible
or cut after the removal of the nodules. Fallopian tube
involvement was considered to be present in the presence of
dense adhesions distorting and dislocating at least one fal-
lopian tube.
Operative laparoscopy was performed by video control
through an umbilical incision and two or three lower abdom-
inal incisions. Instrumentation included 5-mm scissors and
graspers. Irrigation was performed with lactated Ringer’s
solution, and hemostasis was achieved with bipolar coagu-
lation. Lysis of adhesions was performed with the use of
sharp dissection to fully mobilize the ovaries. In the presence
of endometriosis, the surgical goal was the complete exci-
sion of all lesions. In the presence of ovarian endometrio-
mas, one of the following techniques was used: [1] stripping
of the inner lining of the cyst from the normal ovarian tissue,
[2] unilateral salpingo-oophorectomy, and [3] bilateral sal-
pingo-oophorectomy. In infertile patients, laparoscopic tubal
patency was assessed by chromopertubation. All surgical
specimens were sent for histological examination. Patho-
logic descriptions consistent with endometrial glands and
stroma were considered endometriosis. Descriptions of he-
mosiderin-laden macrophages alone, although suspected of
being endometriosis, were not considered clear evidence of
the disease.
Cases of endometriosis were classified into four groups ac-
cording to the revised American Fertility Society (rAFS) clas-
sification (12). Statistical analysis was performed with Graph-
Pad Prism version 3.00 for Windows (GraphPad Software, San
Diego). The Mann-Whitney U-test and Fisher’s exact test were
Vol. 83, No. 1, January 2005
 TABLE 1
Histological characteristics according to the ultrasound findings.
Absence of kissing
Characteristics ovaries (n ⴝ 678)
Serous cystoadenoma 144 (21.2)
Mucinous cystoadenoma
Dermoid cyst
Luteal cyst
Pelvic endometriosisa 282 (41.6)
Ovarian leiomyoma
Hydrosalpinx
Pelvic abscess
Borderline ovarian tumor
Normal pelvisb
Others
Note: Data are presented as number (%).
a
Histological confirmation obtained for 243 cases.
b
No histology.
Ghezzi. Kissing ovaries at ultrasound and endometriosis. Fertil Steril 2005.
used to compare continuous variables and proportions, respec-
tively. P⬍.05 was considered statistically significant.
RESULTS
A total of 722 women were enrolled in the study. Twelve
cases were excluded for the following reasons: in six cases a
uterine myoma was misdiagnosed as an adnexal mass, in
four cases a malignant ovarian tumor was revealed at frozen
section examination, in one case an appendicular mucocele
was erroneously diagnosed as a sactosalpinx, and in one case
a large lymphocyst was misdiagnosed as an ovarian cyst. A
total of 309 women underwent ultrasound and laparoscopy
because of clinical or ultrasound signs suggestive of endo-
metriosis. In the remaining cases, surgery was performed for
an adnexal mass nonsuggestive of endometriosis.
Among the patients who underwent laparoscopy for sus-
pected endometriosis (chronic pelvic pain, dyspareunia, dys-
menorrhea), a normal pelvis was found at laparoscopy in 21
cases. In nine cases, a benign adnexal mass was misdiag-
nosed as an endometrioma. In 20 cases, mono-/bilateral
endometriomas were misdiagnosed as nonendometriotic be-
nign adnexal tumors.
Kissing ovaries were diagnosed in 32 (4.4%) patients at
preoperative ultrasound. Laparoscopy confirmed the pres-
ence of joined ovaries in the pouch of Douglas, adherent to
surrounding tissue in all cases. Table 1 displays the histo-
logical findings of the study population according to the
ultrasound appearance of the ovaries.
Pelvic endometriosis was present in 309 (43.5%) cases.
The rAFS classification and location of the disease are pre-
sented in Table 2. Among patients with moderate to severe
Fertility and Sterility姞
Presence of kissing
ovaries (n ⴝ 32) P
0 ⬍.01
91 (13.4) 1 (3.1) .07
67 (9.9) 1 (3.1) .35
24 (3.5) 0 .69
27 (84.4) ⬍.0001
4 (0.6) 0 .66
18 (2.6) 1 (3.1) .58
15 (2.2) 2 (6.2) .17
8 (1.2) 0 1.0
21 (3.1) 0 .61
4 (0.6) 0 .66
endometriosis (n ⫽ 189), the median rAFS score was sig-
nificantly higher in those with the ultrasound presence of
kissing ovaries than among patients without kissing ovaries
(median, 74; range, 32–148; vs. median, 35; range 16 –146;
P⬍.001). The median (range) operative time was signifi-
cantly longer in the former than in the latter group of patients
(115 minutes [65–245 minutes] vs. 50 minutes [15–180
minutes]; P⬍.001). Considering only patients with bilateral
ovarian endometriomas (n ⫽ 45), the median (range) rAFS
score (74 [50 –148] vs. 49 [21–146]; P⬍.005) and the me-
dian operative time (135 minutes [90 –245 minutes] vs. 65
minutes [40 –135 minutes]; P⬍.001) were significantly
higher in patients with (n ⫽ 13) than in those without (n ⫽
32) kissing ovaries. Among patients with unilateral endo-
metriomas (n ⫽ 126), those with kissing ovaries at ultra-
sound (n ⫽ 9) had a higher median (range) rAFS score (80
[32–128] vs. 33 [19 –120]; P⬍.005) and a longer median
operative time (110 minutes [65–185 minutes] vs. 65 min-
utes [20 –180 minutes]; P⬍.005) than women without kiss-
ing ovaries (n ⫽ 117). Among patients without definite
endometriomas (n ⫽ 138), those with ultrasound kissing
ovaries (n ⫽ 5) had a higher median (range) rAFS score (75
[32–106] vs. 8 [1–30]; P⬍.005) and a longer median oper-
ative time (80 minutes [65–100 minutes] vs. 35 minutes
[15–70 minutes]; P⬍.005) than women without kissing ova-
ries (n ⫽ 133).
The most severe form of pelvic endometriosis (rAFS
ⱖ70) was more frequent among patients than among those
without kissing ovaries (16/32 [50.0%] vs. 27/678 [4.0%];
P⬍.0001). The presence of dense adhesions requiring sur-
gical lysis between the adnexae and the sigmoid was signif-
icantly higher in patients with endometriosis than in those
145
 TABLE 2
Grade and disease location in patients with pelvic endometriosis.

Absence of kissing Presence of kissing

Characteristics ovaries (n ⴝ 282) ovaries (n ⴝ 27) P

Stage of the diseasea

Minimal 48 (17) 0 .01
Mild 72 (25.5) 0 ⬍.001
Moderate 91 (32.3) 3 (11.1) .02
Severe 71 (25.2) 24 (88.9) ⬍.0001
Location of ovarian endometriomas
Unilateral endometrioma 117 (41.5) 9 (33.3) .54
Bilateral endometriomas 32 (11.3) 13 (48.1) ⬍.0001
Absence of endometriomas 133 (47.2) 5 (18.6) ⬍.005
Fallopian tube involvement 93 (33.0) 25 (92.6) ⬍.0001
Completely obliterated cul-de-sac 62 (22.0) 17 (63.0) ⬍.0001
Mono-/bilateral ureteral endometriosis 9 (3.2) 5 (18.5) .05
Uterosacral ligaments endometriosis 59 (20.9) 19 (70.4) ⬍.0001
Bladder endometriosis 9 (3.2) 1 (3.7) .60
Bowel endometriosis 7 (2.5) 5 (18.5) ⬍.01
Diameter of endometriomas (cm)b 5 (2–10) 5 (2–9) .30
Note: Data are presented as number (%) or median (range).
a
According to the revised American Fertility Society classification.
b
Considering the 191 patients with definite endometriomas.
Ghezzi. Kissing ovaries at ultrasound and endometriosis. Fertil Steril 2005.

without kissing ovaries on ultrasound (25/27 [92.6%] vs.
99/282 [35.1%]; P⬍.0001).
The laparoscopic tubal patency test was performed in 145
infertile patients. Of these, five cases had kissing ovaries,
107 patients had pelvic endometriosis; and the remaining 33
patients had nonendometriotic adnexal masses. A higher
proportion of cases with complete obstruction of both fallo-
pian tubes was found among patients with kissing ovaries
than among those without kissing ovaries (4/5 [80%] vs.
12/140 [8.6%]; P⬍.001). This figure did not change when
the analysis was restricted only to patients with endometri-
osis (4/5 [80%] vs. 11/107 [10.3%]; P ⫽ .001).
In all five cases diagnosed to have kissing ovaries and not
affected by endometriosis, dense adhesions were present
between the ovaries, requiring lysis at the time of operation.
DISCUSSION
The present study shows that the detection of kissing ovaries
at ultrasound is strongly associated with the presence of
endometriosis and in particular is a marker of the most
severe form of the disease. Indeed, in all cases an important
adhesion formation between the ovaries and the surrounding
pelvic organs and structures (e.g., with the sacrouterine
ligaments, rectum, posterior uterine wall) was present.
The development of kissing ovaries is probably mimick-
ing the processes induced in pelvic inflammatory disease.
Although the etiology of endometriosis still remains an
enigma, most investigators agree that the persistence and
progression of ectopic endometrial tissue is intimately con-
nected with a local pelvic inflammatory process (13, 14).
Thus, it has been demonstrated that women with endometri-
osis have higher concentrations of proinflammatory cyto-
kines in their peritoneal fluid than women without endome-
triosis (15–17). The peritoneal fluid accumulates in the
dependent portions of the pelvis (18), especially in the pouch
of Douglas, and the inflammation triggered by the endome-
triosis induces adhesion formation between the adjacent
peritoneal surfaces and organs (19), displacing the ovaries
medially and caudally into the Douglas.
Previous studies (1–3) that explored the predictive abil-
ity of ultrasound in identifying women with endometriosis
showed an optimal agreement between ultrasound and
intraoperative diagnosis only for ovarian endometriomas.
The lack of a noninvasive tool to predict anatomic loca-
tions and extension of the disease in the absence of
definite endometriomas has deeply limited the preopera-
tive assessment. Our findings are in keeping with those of
Exacoustos et al. (20), who found a high correlation
between ultrasound and surgical findings. In particular,
these investigators reported an 82% agreement between
ultrasound and intraoperative evaluation in cases of se-
vere endometriosis. According to our results, when ultra-
sound reveals the presence of kissing ovaries, the surgeon

146 Ghezzi et al. Kissing ovaries at ultrasound and endometriosis Vol. 83, No. 1, January 2005
should expect to find a tangled pelvis with subverted
anatomy. Of interest, the proportion of bowel involvement
in patients with kissing ovaries was statistically higher
than in other women with endometriosis. This observation
has important clinical value for the preoperative counsel-
ing and when deciding on appropriate surgical manage-
ment. Chapron et al. (21), in their retrospective series of
patients with deep infiltrating endometriosis, showed that
laparotomy was still the procedure of choice in 23 out of
241 cases, particularly those with intestinal involvement.
Redwine and Wright (22) reported that in the presence of
complete cul-de-sac obliteration, aggressive laparoscopic
excision of endometriosis is often required, including
major intestinal surgery. Therefore, it is of the utmost
importance to inform patients preoperatively of the ultra-
sound diagnosis of kissing ovaries and of the fact that
some degree of intestinal surgery and eventually laparot-
omy could be required to achieve complete removal of
endometriosis.
Another interesting observation of the present study is that
the preoperative detection of kissing ovaries seems to be a
negative prognostic marker for fertility. Although the high
proportion of patients with nonpatent fallopian tubes and
kissing ovaries found in the present study has to be taken
with caution because of the small number of cases consid-
ered, it is very likely that, in the presence of kissing ovaries,
the fallopian tubes are deeply involved in dense adhesions
and are nonpatent.
Moreover, infertile patients with kissing ovaries are likely
to have the most severe form of endometriosis, which, in
turn, is associated with lower pregnancy rates. Several re-
ports suggest that the peritoneal microenvironment of pa-
tients with severe endometriosis could negatively affect the
response to ovarian stimulation, fertilization, and implanta-
tion rates, impairing even assisted reproductive technology
outcome (23, 24).
In conclusion, the finding of kissing ovaries at ultrasound is
suggestive of severe pelvic endometriosis, a condition often
affecting the menstrual cycle and sexual activity, compromising
quality of life and fertility. The preoperative identification of
kissing ovaries at ultrasound should be considered an indication
for referring the patient to a qualified endometriosis center to
optimize the surgical treatment in terms of radicality, fertility
preservation, and risk of recurrence.
REFERENCES
1. Timor-Tritsch LE, Lerner JP, Monteagudo A, Santos R. Transvaginal
ultrasonographic characterization of ovarian masses by means of color
flow-directed Doppler measurements and a morphologic scoring sys-
tem. Am J Obstet Gynecol 1993;168:909 –13.
2. Lerner JP, Timor-Tritsch LE, Federman A, Abramovich G. Transvag-
inal ultrasonographic characterization of ovarian masses with an im-
proved, weighted scoring system. Am J Obstet Gynecol 1994;170:81–5.
3. Eskenazi B, Warner M, Bonsignore L, Olive D, Samuels S, Vercellini
Fertility and Sterility姞
P. Validation study of nonsurgical diagnosis of endometriosis. Fertil
Steril 2001;76:929 –35.
4. Mais V, Guerriero S, Ajossa S, Angiolucci M, Paoletti AM, Melis GB.
The efficiency of transvaginal ultrasonography in the diagnosis of
endometrioma. Fertil Steril 1993;60:776 – 80.
5. Moore J, Copley S, Morris J, Lindsell D, Golding S, Kennedy S. A
systematic review of the accuracy of ultrasound in the diagnosis of
endometriosis. Ultrasound Obstet Gynecol 2002;20:630 – 4.
6. Reles A, Wein U, Lichtenegger W. Transvaginal color Doppler sonog-
raphy and conventional sonography in the preoperative assessment of
adnexal masses. Clin Ultrasound 1997;25:217–25.
7. Kurjak A, Kupesic S. Scoring system for prediction of ovarian endo-
metriosis based on transvaginal color and pulsed Doppler sonography.
Fertil Steril 1994;62:81– 8.
8. Alcazar JL, Laparte C, Jurado M, Lopez-Garcia G. The role of trans-
vaginal ultrasonography combined with color velocity imaging and
pulsed Doppler in the diagnosis of endometrioma. Fertil Steril 1997;
67:487–91.
9. Ghezzi F, Beretta P, Franchi M, Parissis M, Bolis PF. Recurrence of
ovarian endometriosis and anatomical location of primary lesion. Fertil
Steril 2001;75:136 – 40.
10. Vercellini P, Aimi G, De Giorgi O, Maddalena S, Carinelli S, Crosignani
PG. Is cystic ovarian endometriosis an asymmetric disease? Br J Obstet
Gynaecol 1998;105:1018 –21.
11. Dessole S, Farina M, Rubattu G, Cosmi E, Ambrosini G, Nardelli GB.
Sonovaginography is a new technique for assessing rectovaginal endo-
metriosis. Fertil Steril 2003;79:1023–7.
12. American Fertility Society. Revised American Fertility Society classi-
fication of endometriosis. Fertil Steril 1985;43:351–2.
13. Lebovic DI, Mueller MD, Taylor RN. Immunology of endometriosis.
Fertil Steril 2001;75:1–10.
14. Mueller MD, Mazzucchelli L, Buri C, Lebovic DI, Dreher E, Taylor
RN. Epithelial neutrophil-activating peptide 78 concentrations are ele-
vated in the peritoneal fluid of women with endometriosis. Fertil Steril
2003;79(3 Suppl 1):815–20.
15. Cheong YC, Shelton JB, Laird SM, Richmond M, Kudesia G, Li TC, et
al. IL-1, IL-6 and TNF-␣ concentrations in the peritoneal fluid of
women with pelvic adhesions. Hum Reprod 2002;17:69 –75.
16. Wu MY, Ho HN. The role of cytokines in endometriosis. Am J Reprod
Immunol 2003;49:285–96.
17. Calhaz-Jorge C, Costa AP, Santos MC, Palma-Carlos ML. Peritoneal
fluid concentrations of interleukin-8 in patients with endometriosis
depend on the severity of the disorder and are higher in the luteal phase.
Hum Reprod 2003;18:593–7.
18. Taylor RN, Lebovic DI, Mueller MD. Angiogenic factors in endome-
triosis. Ann NY Acad Sci 2002;955:89 –100.
19. Redwine DB. Ovarian endometriosis: a marker for more extensive
pelvic and intestinal disease. Fertil Steril 1999;72:310 –5.
20. Exacoustos C, Zupi E, Carusotti C, Rinaldo D, Marconi D, Lanzi G, et
al. Staging of pelvic endometriosis: role of sonographic appearance in
determining extension of disease and modulating surgical approach.
J Am Assoc Gynecol Laparosc 2003;10:378 – 82.
21. Chapron C, Fauconnier A, Vieira M, Barakat H, Dousset B, Pansini V,
et al. Anatomical distribution of deeply infiltrating endometriosis: sur-
gical implications and proposition for a classification. Hum Reprod
2003;18:157– 61.
22. Redwine DB, Wright JT. Laparoscopic treatment of complete obliter-
ation of the cul-de-sac associated with endometriosis: long-term fol-
low-up of en bloc resection. Fertil Steril 2001;76:358 – 65.
23. Arici A, Oral E, Bukulmez O, Duleba A, Olive DL, Jones EE. The
effect of endometriosis on implantation: results from the Yale Univer-
sity in vitro fertilization and embryo transfer program. Fertil Steril
1996;65:603–7.
24. Barnhart K, Dunsmoor-Su R, Coutifaris C. Effect of endometriosis on
in vitro fertilization. Fertil Steril 2002;77:1148 –55.
147