Gynecologic Oncology 82, 504 –509 (2001)

doi:10.1006/gyno.2001.6316, available online at http://www.idealibrary.com on

A Comparison of Ovarian Metastasis between Squamous Cell

Carcinoma and Adenocarcinoma of the Uterine Cervix
Toru Nakanishi, M.D., 1 Kenji Wakai, M.D.,* Hisatake Ishikawa, M.D., Akihiro Nawa, M.D., Yuka Suzuki, M.D.,
Shigeo Nakamura, M.D.,† and Kazuo Kuzuya, M.D.
Department of Gynecology, †Department of Pathology and Clinical Laboratory, Aichi Cancer Center Hospital, and
*Department of Preventive Medicine, Nagoya University School of Medicine, Nagoya 464-8681, Japan

Received February 12, 2001

of squamous cell carcinoma. Recently, however, some studies

Objective. The purpose of our study was to investigate a possible
difference in ovarian metastasis between squamous cell carcinoma showed that ovarian metastases were rare in early invasive
and adenocarcinoma of the uterine cervix and to confirm clinico- adenocarcinoma of the uterine cervix (1.7–2.5%) [4 –7]. The
pathological variables associated with the metastases. reported incidence was almost equal to that of early-stage
Methods. Clinical and pathological variables of 1064 patients squamous cell carcinoma of the uterine cervix, which rarely
with invasive squamous cell carcinoma and 240 with adenocarci- metastasizes to the ovary (0 – 0.71%) [4 –9]. When compared in
noma were studied. clinical stage Ib disease [4, 5], the incidence of ovarian me-
Results. Ovarian metastasis was found in 14 patients (1.3%) tastasis of adenocarcinoma was slightly higher than that of
with squamous cell carcinoma and 15 (6.3%) with adenocarci- squamous cell carcinoma. Since the differences in the charac-
noma. The mean age of patients with ovarian metastasis of squa- teristics of ovarian metastasis from squamous cell carcinoma
mous cell carcinoma was 57.4 years, compared to 50.2 years for
and adenocarcinoma have not been sufficiently studied, the
adenocarcinoma. Ovarian metastasis of adenocarcinoma was
clinicopathological variables associated with the metastases
more likely to be visible (40.0%) and present in both ovaries
(66.7%), while these two characteristics occurred in only 21.4 and remain unclear.
36.7% of patients with squamous cell carcinoma. A logistic regres- This is a retrospective study of ovarian metastasis from
sion analysis with clinical variables indicated that clinical stage cervical cancer over the past 27 years at Aichi Cancer Center.
beyond IIb was a significant variable of squamous cell carcinoma, We attempted to determine possible differences in ovarian
and more than 30-mm tumor size was significant in adenocarci- metastases from squamous cell carcinoma and adenocarcinoma
noma. of the uterine cervix and to confirm their clinical characteristics
Conclusion. The incidence of ovarian metastasis of adenocarci- and any variables associated with the metastases.
noma of the uterine cervix was significantly higher than that of
squamous cell carcinoma. The incidence of adenocarcinoma was
associated more closely with tumor size than clinical stage, PATIENTS AND METHODS
whereas it was more associated with clinical stage in squamous cell
carcinoma. The results thus suggested that the differences in The clinical and pathological records of all patients diag-
ovarian metastases were caused by the different characteristics of nosed with invasive cervical cancer at Aichi Cancer Center
the two types of carcinoma. © 2001 Academic Press
between 1974 and 2000 were reviewed. All patients who
Key Words: cervical cancer; ovarian metastasis.
underwent hysterectomy, pelvic lymphadenectomy, and oo-
phorectomy of one or both sides as primary treatment at the
INTRODUCTION Aichi Cancer Center during these years, and whose primary
cervical lesion was diagnosed as squamous cell carcinoma or
Since it was reported that ovarian metastases were more adenocarcinoma, were included in this study. Patients whose
commonly found in association with adenocarcinoma than with clinical or pathological records were not available, and those
squamous cell carcinoma and the incidence was 5.5–12.5% with adenocarcinoma in which a primary site could not be
[1– 4], the incidence of ovarian metastasis of adenocarcinoma differentiated from a site of endometrial origin, were excluded
of the uterine cervix has been considered much higher than that from this study. A total of 1304 patients formed the basis of the
study. Of these patients, 1134 underwent bilateral oophorec-
1
To whom correspondence should be addressed at Department of Gynecol- tomy, 62 oophorectomy on one side and wedge resection on
ogy, Aichi Cancer Center Hospital, 1-1 Kanokoden, Chikusa-ku, Nagoya the other, and 108 oophorectomy on only one side. Because of
464-8681, Japan. Fax: 001-81-52-763-5233. benign ovarian disease or ectopic pregnancy, one ovary had

0090-8258/01 $35.00 504

Copyright © 2001 by Academic Press
All rights of reproduction in any form reserved.
 OVARIAN METASTASIS OF CERVICAL CANCER 505

been already removed in 55 of 108 patients undergoing oopho- TABLE 1

rectomy on one side. One ovary of 115 patients, who were 45 Clinical Presentation of Ovarian Metastasis of Cervical Cancer
years old or less and diagnosed as having early invasive car-
Squamous cell
cinoma, was conserved. Besides these, all patients received carcinoma Adenocarcinoma
hysterectomy and pelvic lymphadenectomy; 1151 underwent
radical hysterectomy, 134 simple hysterectomy, and 19 pelvic Total 1064 240
exenteration. Informed consent was obtained from each patient Mean age (Range) 50.5 (24.2 ⫺ 79.8) 49.5 (25.3 ⫺ 78.2)
regarding the use of clinicopathological variables. Ovarian metastasis 14 15
Mean age (Range) 57.4 (36.8 ⫺ 73.5) 50.2 (30.0 ⫺ 70.1)
The age of patients at the initial evaluation, their menopausal Unilateral 9 5
state, presence or absence of obesity, and year of treatment Bilateral 5 10
were obtained from clinical records. Obesity was defined ac- Enlarged ovary 2 4
cording to the criteria of the Japan Society of Obesity, i.e., a Visible 1 2
body mass index of more than 26.4. Primary disease was Microscopic 11 9
staged at the initial evaluation according to the staging system
of the International Federation of Gynecology and Obstetrics
(FIGO) and was adopted as the clinical stage of disease. cancer. Of 1064 patients with squamous cell carcinoma, 14
Sixteen patients whose disease had not invaded the pelvic wall (1.3%) had metastases and 15 of 240 (6.3%) with adenocarci-
but had caused hydronephrosis were diagnosed as clinical stage noma had metastases (Table 1). Histologically, 2 were kera-
IIIb, and 1 whose disease had invaded the lower third of the tinizing type squamous cell carcinoma, 12 nonkeratinizing
vaginal wall was diagnosed as clinical stage IIIa. Tumor size type, 11 endocervical type adenocarcinoma, 1 endometrial
was determined from pathological slides. For patients with type, 2 clear cell, and 1 adenosquamous carcinoma. While 1 of
gross tumors, the maximum diameter was noted. All available 14 patients with squamous cell carcinoma had no other extra-
pathological slides were reviewed for the diagnosis of histo- cervical spread pathologically, all patients with adenocarci-
logical subtype, presence or absence of pathological parame- noma had spread. The mean age of all patients with squamous
trial invasion, vaginal invasion, lymph node metastasis, endo- cell carcinoma was not different from that of patients with
metrial invasion, and ovarian metastasis. Histological adenocarcinoma (P ⫽ 0.194). Patients with metastasis of
specimens of each ovary were made by cutting it vertically, squamous cell carcinoma had a higher mean age than those
and at least two sections were made for examination. Ovarian with adenocarcinoma, but the difference was not statistically
metastasis was considered present if viable tumor cells were significant (P ⫽ 0.072). The mean age of patients with
observable in ovarian tissue or vessels. Histological diagnosis metastasis was significantly older than that of all patients with
of squamous cell carcinoma included both the keratinizing and squamous cell carcinoma (P ⫽ 0.029), whereas the mean age
the nonkeratinizing types, and adenocarcinoma included both was not different in adenocarcinoma (P ⫽ 0.819). Although
the endocervical and the endometrioid types as well as adeno- 10 of 15 (66.7%) adenocarcinoma cases involved both ovaries,
squamous carcinoma and clear cell adenocarcinoma. 5 of 14 (35.7%) metastases of squamous cell carcinoma were
In comparing the age of patients, Student’s t test analysis found in both ovaries. Ovaries of 4 (26.7%) patients with
was used. Pearson’s ␹ 2 test was used to examine the signifi- adenocarcinoma were enlarged due to metastases and 2
cance of variables in relation to histology and to the ovarian (13.3%) had a visible metastatic tumor of adenocarcinoma on
metastasis. Logistic regression analysis was carried out with normal-size ovaries, while 11 of 14 (78.6%) of the squamous
ovarian metastasis as the dependent variable and with the other cell carcinomas had microscopic metastases. While 4 of 6
variables listed in Table 2 as the independent ones for patho- (66.7%) visible metastases of adenocarcinoma involved both
logical analysis. The reliability of the analysis was validated ovaries, 1 of 3 (33.3%) metastases of squamous cell carcinoma
using a likelihood ratio step-forward and step-backward fitting involved both. Other ovarian diseases were found in 55 pa-
procedure. All variables were assessed for analyses according tients, 1 with an endocervical type adenocarcinoma of the
to the categories listed in Table 2. All tests were two tailed; a uterine cervix coexisting with ovarian cancer of stage IIIc
P value of ⬍0.05 was considered statistically significant. The (endometrioid type) and the others with benign adenomas,
statistical analysis was performed using the SPSS software mature cystic teratomas, or endometrial cysts.
version 6.1J. The findings from our comparison of clinicopathological
variables in squamous cell carcinoma and adenocarcinoma are
summarized in Table 2. While the mean age of all patients did
RESULTS not differ between squamous cell carcinoma and adenocarci-
noma patients, the incidence in adenocarcinoma patients
All patients in this study were Japanese with a mean age at younger than 50 years of age was significantly higher than for
presentation of 50.3 years (age range, 24.2 to 79.8 years). In squamous cell carcinoma patients. The incidence of adenocar-
total, 29 (2.0%) patients had ovarian metastases from cervical cinoma patients in the premenopausal state was also signifi-
506 NAKANISHI ET AL.

TABLE 2
Comparison of Clinicopathological Variables between Squamous Cell Carcinoma and Adenocarcinoma of the Uterine Cervix

Squamous cell carcinoma Adenocarcinoma

n % n % P

Clinical variables
Age (years) ⫺40 147 13.8 52 21.7 0.003
41–50 367 34.5 86 35.8
51⫹ 550 51.7 102 42.5
Postmenopausal state No 556 52.3 146 60.8 0.016
Yes 508 47.7 94 39.2
Obesity No 965 90.7 213 88.8 0.357
Yes 99 9.3 27 11.3
FIGO clinical stage Ia–IIa 862 81.0 204 85.0 0.149
IIb⫹ 202 19.0 36 15.0
Treatment year ⫺1985 620 58.3 83 34.6 ⬍0.001
1986⫹ 444 41.7 157 65.4
Tumor size (mm) ⫺30 654 61.5 151 62.9 0.435
31–50 324 30.5 65 27.1
51⫹ 86 8.1 24 10.0
Pathological variables
Vaginal invasion No 821 77.2 201 83.8 0.002
Yes 243 22.8 39 16.3
Parametrial invasion No 817 76.8 190 79.2 0.427
Yes 247 23.2 50 20.8
Endometrial invasion No 947 89.0 201 83.8 0.023
Yes 117 11.0 39 16.3
Pelvic lymph node metastasis No 789 74.2 178 74.2 0.997
Yes 275 25.8 62 25.8
Depth of invasion (mm) ⫺10 555 52.2 132 55.0 0.571
10–15 240 22.6 55 22.9
15⫹ 269 25.3 53 22.1

cantly higher. The incidence of adenocarcinoma has increased in 15 patients with clinical stage Ia and in 11 with stage IIa
in recent years, as described previously [10, 11]. In the present adenocarcinoma. When classified by tumor size, the meta-
study, the incidence of pathological vaginal invasion of squa- static incidence of adenocarcinoma of 0% (0/102) in tumors
mous cell carcinoma was significantly higher than that of smaller than 20 mm and of 2.0% (1/49) in 21- to 30-mm
adenocarcinoma, while the incidence of endometrial invasion tumors increased immediately to 10.8% in 31- to 50-mm
of squamous cell carcinoma was significantly lower. The inci- tumors and 29.2% in tumors larger than 51 mm (Table 3).
dence of pathological parametrial invasion of squamous cell The incidence of squamous cell carcinoma was 0.2% (1/
carcinoma was slightly higher than that of adenocarcinoma, but 444) in tumors smaller than 20 mm, 1.9% (11/583) in 21- to
the difference was not statistically significant. The incidences 60-mm tumors, and 5.4% (2/37) in tumors larger than 61
of large tumor maximum diameter greater than 30 mm, pelvic mm in diameter.
lymph node metastasis, and deep stromal invasion were not The logistic regression analysis with clinicopathological
significantly different between squamous cell carcinoma and variables revealed that the presence of pathological endo-
adenocarcinoma. metrial invasion and lymph node metastasis were the sig-
The associations of ovarian metastases with clinicopatho- nificant variables associated with ovarian metastasis of
logical variables are summarized in Table 3. While the inci- squamous cell carcinoma of the uterine cervix (Table 4). In
dences of the metastases of clinical stage Ia, Ib, and IIa the analysis of adenocarcinoma, pathological endometrial
squamous cell carcinoma were 0.8% (1/132), 0.5% (3/614), invasion, lymph node metastasis, and pathological parame-
and 0.9% (1/116), respectively, the incidence increased imme- trial invasion were significant. When analyzed with clinical
diately to 4.0% (7/175) in clinical stage IIb and to 7.4% (2/27) variables only, clinical stage beyond IIb was a significant
in stage III–IV. The incidence of stage Ib adenocarcinoma was variable of squamous cell carcinoma, and a tumor size of
4.0% (7/178), compared to 14.8% (4/27) in stage IIb and more than 30 mm was a significant variable in adenocarci-
44.4% (4/9) in stage III–IV. No ovarian metastasis was found noma (Table 4).
 OVARIAN METASTASIS OF CERVICAL CANCER 507

TABLE 3
Comparison of Clinicopathological Variables and Ovarian Metastasis between Squamous Cell Carcinoma
and Adenocarcinoma of the Uterine Cervix

Squamous cell carcinoma Adenocarcinoma

Ovarian metastasis Ovarian metastasis

Total Total
n n % n n % P

Total 1064 14 1.3 240 15 6.3 ⬍0.001

Age (years)
⫺40 147 1 0.7 52 1 1.9 0.440
41–50 367 2 0.5 86 9 10.5 ⬍0.001
51⫹ 550 11 2.0 102 5 4.9 0.082
Postmenopausal state
No 556 4 0.7 146 10 6.8 ⬍0.001
Yes 498 10 2.0 94 5 5.3 0.003
FIGO clinical stage
Ia–IIa 862 5 0.6 204 7 3.6 ⬍0.001
IIb⫹ 202 9 4.5 36 8 23.5 ⬍0.001
Treatment year
⫺1985 620 7 1.1 83 6 7.2 ⬍0.001
1986⫹ 444 7 1.6 157 9 5.7 0.005
Tumor size (mm)
⫺30 654 6 0.9 151 1 0.7 0.895
31–50 324 5 1.5 65 7 10.8 ⬍0.001
51⫹ 86 3 3.5 24 7 29.2 ⬍0.001
Pathological vaginal invasion
No 821 7 0.9 201 5 2.5 0.054
Yes 243 7 2.9 39 10 25.6 ⬍0.001
Pathological parametrial invasion
No 817 6 0.7 190 3 1.6 0.265
Yes 247 8 3.2 50 12 24.0 ⬍0.001
Pathological endometrial invasion
No 947 7 0.7 201 5 2.5 0.027
Yes 117 7 6.0 39 10 25.6 ⬍0.001
Pelvic lymph node metastasis
No 789 2 0.3 175 3 1.7 0.016
Yes 275 12 4.4 62 12 19.4 ⬍0.001

DISCUSSION Comparison of the clinicopathological variables revealed no

clear differences. Although adenocarcinoma tended to attack
Although some recent studies showed that ovarian metas- young patients under 50 years of age, this could be explained
tases were rare (1.7–2.5%) in early invasive adenocarci- by the decrease in the number of older patients [12–14] and the
noma of the uterine cervix [4 –7], several authors reported increase in adenocarcinoma, as described previously [10, 11].
that the incidence of ovarian metastases (5.5–12.5%) was
Squamous cell carcinoma seemed to invade the vaginal wall
higher in association with adenocarcinoma than with squa-
and adenocarcinoma the endometrium, yet the incidence of
mous cell carcinoma [1– 4]. In the present study, the inci-
ovarian metastasis of adenocarcinoma was still higher than for
dence of ovarian metastasis from adenocarcinoma of 6.3%
squamous cell carcinoma in comparison with the incidence of
was significantly higher than that of squamous cell carci-
noma. When the comparison was limited only to early vaginal invasion, as well as that of endometrial invasion (Table
invasive disease, the incidence of adenocarcinoma was sig- 3). These results suggested that the higher incidence of ovarian
nificantly higher than that of squamous cell carcinoma (Ta- metastasis of adenocarcinoma would be caused by the different
ble 3), suggesting that clinical behavior of an ovarian me- characteristics of the two types of carcinoma.
tastasis from adenocarcinoma might be different from that If the characteristics of ovarian metastasis from adenocarci-
of squamous cell carcinoma. We compared clinicopatholog- noma of the uterine cervix were different from those of squa-
ical variables to determine possible differences in clinical mous cell carcinoma, the variables associated with such me-
behavior between squamous cell carcinoma and adenocar- tastasis would also be different. Thus, we performed a separate
cinoma. logistic regression analysis in squamous cell carcinoma and
508 NAKANISHI ET AL.

TABLE 4
Comparison of Logistic Regression Analysis of Ovarian Metastasis between Squamous Cell Carcinoma
and Adenocarcinoma of the Uterine Cervix

Squamous cell carcinoma Adenocarcinoma

OR 95% CI P OR 95% CI P

Clinicopathological variables
Pelvic lymph node metastasis
No 1.00 0.001 1.00 0.034
Yes 12.66 2.71–59.15 4.81 1.13–20.57
Pathological endometrial invasion
No 1.00 0.011 1.00 0.008
Yes 4.19 1.39–12.69 5.43 1.56–18.97
Parametrial invasion
No NS 1.00 0.025
Yes 5.37 1.23–23.42
Clinical variables
FIGO clinical stage
Ia–IIa 1.00 ⬍0.001 NS
IIb⫹ 7.99 2.65–24.12
Tumor size
⬍30 mm NS 1.00 ⬍0.001
31–50 mm 18.06 2.18–149.72
⬎51 mm 61.62 7.16–530.29

Note. OR, odds ratio; 95% CI, 95% confidence interval.

adenocarcinoma. Although the independent significant vari- metastasis from adenocarcinoma would be different from those
ables were not different in the clinicopathological analysis, the from squamous cell carcinoma.
analysis using clinical variables indicated a difference. While In conclusion, the incidence of the ovarian metastasis of
an independent significant variable in the analysis of squamous adenocarcinoma uterine cervix was 6.3%, compared to 1.3%
cell carcinoma was clinical stage beyond IIb, a tumor size of for squamous cell carcinoma. Logistic regression revealed that
more than 30 mm was the variable in adenocarcinoma. Indeed, the independent significant variable for squamous cell carci-
the incidence of ovarian metastasis was increased in tumors noma was clinical stage beyond IIb, while the variable for
larger than 30 mm. These results suggest that the clinical adenocarcinoma was a tumor size of greater than 30 mm.
variables associated with ovarian metastasis differ between These results suggested that the incidence of ovarian metastasis
squamous cell carcinoma and adenocarcinoma and that the from cervical cancer would be rare in patients with disease
incidence of metastasis of adenocarcinoma associates more earlier than clinical stage IIa squamous cell carcinoma and with
closely with tumor size than clinical stage. small adenocarcinoma less than 30 mm in maximum diameter
The present study also revealed other characteristics of ovar- and that characteristics of ovarian metastasis would differ
ian metastasis of adenocarcinoma that are different from those between squamous cell carcinoma and adenocarcinoma.
of squamous cell carcinoma. One of 14 patients with metasta-
ses of squamous cell carcinoma had no other extracervical
REFERENCES
spread, whereas all with adenocarcinoma had spread. Patients
with metastases of squamous cell carcinoma had a higher mean 1. Mann WJ, Chumas J, Amalfitano T, Westermann C, Patsner B. Ovarian
age than those with adenocarcinoma. Although 66.7% of ovar- metastases from stage IB adenocarcinoma of the cervix, Cancer 1987;60:
ian metastases of adenocarcinoma were observed in both ova- 1123– 6.
ries, only 35.7% of patients with squamous cell carcinomas had 2. Tabata M, Ichinoe K, Sakuragi N, Shiina Y, Yamaguchi T, Mabuchi Y.
metastases in both. Some 40.0% of the metastases of adeno- Incidence of ovarian metastasis in patients with cancer of the uterine
cervix, Gynecol Oncol 1987;28:255– 61.
carcinoma were visible, compared to only 21.4% for squamous
3. Natsume N, Aoki Y, Kase H, Kashima K, Sugaya S, Tanaka K. Ovarian
cell carcinoma. This proportion did not change when analyzing
metastasis in stage Ib and II cervical adenocarcinoma, Gynecol Oncol
patients with metastasis in both ovaries. However, these dif- 1999;74:255– 8.
ferences were not significant because of the small number of 4. Toki N, Tsukamoto N, Kaku T, Toh N, Saito T, Kamura T, Matsukuma K,
patients with ovarian metastasis. Further studies should be Nakano H. Microscopic ovarian metastasis of the uterine cervical cancer,
conducted on a large patient sample with ovarian metastases. Gynecol Oncol 1991;41:46 –51.
The results also suggested that the characteristics of ovarian 5. Sutton GP, Bundy BN, Delgado G, Sevin BU, Creasman WT, Major FJ,
 OVARIAN METASTASIS OF CERVICAL CANCER 509

Zaino R. Ovarian metastases in stage IB carcinoma of the cervix: a 10. Ishikawa H, Nakanishi T, Inoue T, Kuzuya K. Prognostic factors
Gynecologic Oncology Group study [see comments], Am J Obstet Gy- of adenocarcinoma of the uterine cervix, Gynecol Oncol 1999;73:
necol 1992;50 –3. 42– 6.
6. Ellsworth LR, Allen HH, Nisker JA. Ovarian function after radical hys- 11. Nakanishi T, Ishikawa H, Suzuki Y, Inoue T, Nakamura S, Kuzuya K. A
terectomy for stage IB carcinoma of cervix, Am J Obstet Gynecol 1983; comparison of prognosis of pathologically staged pt1b adenocarcinoma
145:185– 8. and squamous cell carcinoma of the uterine cervix, Gynecol Oncol 2000;
7. Wu HS, Yen MS, Lai CR, Ng HT. Ovarian metastasis from cervical 79:289 –93.
carcinoma, Int J Gynaecol Obstet 1997;57:173– 8. 12. Beral V, Booth M. Predictions of cervical cancer incidence and mortality
8. Windbichler GH, Holzner EM, Leck GN, Meisel U, Dapunt O, Marth C. in England and Wales, Lancet 1986;1:495.
Ovarian preservation in the surgical treatment of cervical carcinoma, Am J 13. Devesa SS, Young JJ, Brinton LA, Fraumeni JJ. Recent trends in cervix
Obstet Gynecol 1999;180:963–9. uteri cancer, Cancer 1989;64:2184 –90.
9. Parente JT, Silberblatt W, Stone M. Infrequency of metastasis to ovaries 14. Larsen NS. Invasive cervical cancer rising in young white females [news],
in stage I carcinoma of the cervix, Am J Obstet Gynecol 1964;90:1362. J Natl Cancer Inst 1994;86:6 –7.