Keywords
non-epithelial ovarian tumors - ovarian cancer - biomarkers - germ cell tumors - sex
cord-stromal tumors
Palavras-chave
tumores ovarianos não epiteliais - câncer de ovário - biomarcadores - tumores de células
germinativas - tumores do cordão sexual-estroma
Introduction
Adnexal masses are commonly found on gynecological imaging in women of all ages.[1] It is estimated that 5 to 10% of women will be submitted to surgery to investigate
an adnexal mass in their lifetime.[2] Despite the majority of them being benign, malignant tumors must be promptly diagnosed
and treated.[3] Non-epithelial ovarian cancers are rare, accounting for approximately only 5% of
ovarian malignancies encompassing germ cell tumors (3%) and sex cord-stromal tumors
(2%).[4]
According to the last classification of World Health Organization (WHO),[5] germ cell tumors of ovary comprise dysgerminoma, yolk sac tumor, embryonal carcinoma,
non-gestational choriocarcinoma, mature teratoma, immature teratoma, mixed germ cell
tumor, monodermal teratomas and tumors with malignant transformation arising from
a dermoid cyst. Mature teratoma, is the most common benign ovarian neoplasia, most
occur during reproductive years, with a peak incidence between 20 and 40 years of
age.[6] Benign mature teratomas comprise 95% of all germ cell tumors, and only 5% of germ
cell tumors are malignant.[6]
The WHO last classification for ovarian sex cord-stromal tumors shows that these tumors
have been reclassified into the following clinicopathologic entities: pure stromal
tumors, pure sex cord tumors, and mixed sex cord-stromal tumors. Ovarian fibroma,
thecoma, Leydig cell tumor are pure stromal tumors. Adult granulosa cell tumor, Juvenile
granulosa cell tumor and Sertoli cell tumor are pure sex cord tumors, while Sertoli-Leydig
cell tumor is mixed sex cord-stromal tumor.[5]
[7]
In clinical practice, women undergo surgery to diagnose and treat an adnexal mass
according to findings on clinical exam, transvaginal ultrasound and tumor biomarker.
More specific biomarkers such as β-subunit of human chorionic gonadotropin (β-hCG),
α fetoprotein, lactic dehydrogenase (LDH) may be useful for diagnosing malignant non-epithelial
tumors.[8] On the other hand, there is scarce information in the current literature about the
role of routinely measured biomarkers cancer antigen 125 (CA125) and carcinoembryonic
antigen (CEA); and also more recently discovered biomarker Human Epididymis Protein
4 (HE4) in the preoperative diagnosis of non-epithelial ovarian tumors.
The objective of our study was to evaluate, the role of clinical features and preoperative
determination of CA125, HE4 and CEA serum levels in the differentiation of benign
from malignant non-epithelial ovarian tumors.
Methods
Patients
This is a cross-sectional study that was conducted at Hospital da Mulher Prof. Dr.
José Aristodemo Pinotti at Universidade Estadual de Campinas (Unicamp), from February
2010 to December 2015. The study and was approved by the research ethics committee
of the institution (Protocol 1092/2009). Women referred to the pelvic oncology clinic,
due to adnexal masses detected in ultrasound or other imaging exam, were invited to
participate. Patients were consecutively included after signing a consent form and
were submitted to the study protocol. On the first visit, patients were submitted
to physical exam, and blood was collected to measure biomarker levels. In addition,
transvaginal ultrasound was scheduled. When indicated, women underwent surgery for
disease diagnosis and treatment. The indication of surgery was based on clinical exam,
preoperative biomarkers, and ultrasound scan. The medical files of the patients obtained,
which were from the hospital's digital filing system, were reviewed to obtain information
about treatment, disease recurrence, and patient status. Women were considered postmenopausal
when they had > 1 year of amenorrhea or were > 50 years old in case of previous hysterectomy.
In premenopausal women, tumorectomy or unilateral adnexectomy with contralateral ovarian
preservation without hysterectomy was considered fertility-sparing surgery.
Ultrasound (US)
The following ultrasound parameters were used to decide which women should undergo
surgical treatment: largest diameter of the lesion; maximum diameter of the largest
solid part; if unilocular or multilocular; presence of wall irregularity; ascites;
acoustic shadows; number of papillary projections; color Doppler blood flow.[9] When surgery was not indicated, women were scheduled for clinical follow-up. From
869 women enrolled in the study, we excluded 361 who were not operated, 237 women
with epithelial ovarian tumors, 128 with non-neoplastic and non-ovarian tumors, and
24 with ovarian metastases. One hundred and nineteen consecutive women with benign
and malignant non-epithelial ovarian tumors were included in the study. Women had
their surgery indicated according to their clinical exam, ultrasound results (simple
rules by international ovarian tumor analysis [IOTA]),[9] and serum biomarkers. In the present study, 72 women had their ultrasound analyzed
with IOTA simple rules. Among them, 9 had malignant tumors as per histology diagnosis,
but the IOTA simple rules were malignant in 8 cases and benign in 1 case of granulosa
cell tumor. Among the remaining 63 cases that had a benign tumor histology diagnosis,
the IOTA simple rules were benign in 54 women, indeterminate in 1 case (ovarian fibroma),
and malignant in 8 women (2 with fibroma, 1 with thecoma, 1 with ovarian leiomyoma,
1 with struma ovarii, and 3 with teratomas).
Histopathology
Histopathological diagnosis was the gold standard parameter, performed by pathologists
specialized in gynecologic pathology, following the last “WHO classification of tumors
of female reproductive organs.”[5] Women with bilateral tumors with one of them presenting epithelial histology were
excluded from the study. Our sample comprises 71 women with germ cell tumors (64 benign
and 7 malignant), 46 with sex cord-stromal tumors (32 benign and 14 malignant), and
2 with ovarian leiomyomas.
CA125 and CEA
Serum levels of CA125 and CEA were determined by the CA125 II and CEA tests, respectively,
both biomarkers through the chemiluminescence technic in the automatic analyzer Cobas
e411 (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions,
with CA125 expressed in U/ml and CEA in ng/ml.
Human Epididymis Protein 4 (HE4)
The concentrations of HE4 were measured according to manufacturer's instructions using
the ARCHITECT HE4 assay (Abbott Diagnostics, Abbott Park, IL, USA), with HE4 expressed
in pmol/L.
Statistical Analysis
Data were analyzed using the R Environment for Statistical Computing software (R Foundation
for Statistical Computing, Vienna, Austria).[10] According to the histopathological diagnoses, tumors were classified as benign or
malignant germ cell or sex cord-stromal tumors. Two patients with ovarian leiomyoma
had their tumors grouped with sex cord-stromal tumors for statistical purposes. Women's
clinical features, serum biomarker levels and surgical treatment were compared using
the Chi-square test for categorical variables and the Kruskal-Wallis test for the
continuous variables. Statistical calculations were performed using 95% confidence
intervals (95% CIs) considering p < 0.05 as significant. For missing data reason, we excluded 19 cases without CA125,
18 cases without HE4, and 4 without CEA for the biomarkers' analysis only. Follow-up
time (in months) was considered from the date of diagnostic surgery to last hospital
visit or date of death in the case of one patient.
Results
In [Table 1], among 119 women, 98 (82.4%) presented benign tumors and 21 (17.6%) malignant. Seventy-two
(60.5%) were premenopausal and 47 (39.5%) were postmenopausal. Germ cell tumors were
significantly more frequent in premenopausal women, and those younger than 50 years
old, compared with women with sex cord-stromal tumors. Neither body mass index nor
history of close relatives with breast or ovarian cancer were related to tumor malignancy.
Benign ovarian tumors were bilateral in 16 (16.3% of benign tumors) women. Almost
all women with malignant germ cell and sex cord-stromal tumors presented at the initial
stage of the disease: 20 (95.2%) with stage I, of note, 15 (71.4%) were stage Ia.
Among benign germ cell tumors, 63 (98.4%) were mature teratomas, and, among malignant,
only 4 (57.1%) were immature teratomas. Fibromas were the most common tumors among
the patients in the benign sex cord-stromal group, and in the malignant counterpart,
the most common was granulosa cell tumor.
Table 1
Clinical features by tumor type
|
Characteristics
|
Germ cell tumors (n = 71)
|
Sex cord-stromal tumors (n = 48)
|
|
|
Benign
n = 64 (%)
|
Malignant
n = 7 (%)
|
P1
|
Benign
n = 34 (%)
|
Malignant
n = 14 (%)
|
P2
|
P3
|
|
Age (years)
|
|
|
|
|
|
|
|
|
< 35
|
29 (45.3)
|
6 (85.7)
|
|
4 (11.8)
|
4 (28.6)
|
|
|
|
35–50
|
17 (26.6)
|
0
|
0.110
|
6 (17.6)
|
5 (35.7)
|
0.125
|
0.0001
|
|
> 50
|
18 (28.1)
|
1 (14.3)
|
|
24 (70.6)
|
5 (35.7)
|
|
|
|
Menopausal status
|
|
|
|
|
|
|
|
|
Premenopausal
|
48 (75)
|
6 (85.7)
|
0.528
|
9 (26.5)
|
9 (64.3)
|
0.024
|
0.0001
|
|
Postmenopausal
|
16 (25)
|
1 (14.3)
|
|
25 (73.5)
|
5 (35.7)
|
|
|
|
Body mass index (kg/m2)
|
|
|
|
|
|
|
|
|
< 30
|
44 (68.7)
|
4 (57.1)
|
|
23 (67.7)
|
8 (57.1)
|
|
|
|
30–35
|
14 (21.9)
|
2 (28.6)
|
0.816
|
8 (23.5)
|
4 (28.6)
|
0.842
|
0.963
|
|
> 35
|
6 (9.4)
|
1 (14.3)
|
|
3 (8.8)
|
2 (14.3)
|
|
|
|
Close relatives with cancer (breast/ovary)
|
|
|
|
|
|
|
|
|
No
|
31 (48.4)
|
2 (28.6)
|
0.437
|
16 (47.1)
|
9 (64.3)
|
0.232
|
0.478
|
|
Yes
|
33 (51.6)
|
5 (71.4)
|
|
18 (52.9)
|
5 (35.7)
|
|
|
|
Laterality
|
|
|
|
|
|
|
|
|
Unilateral
|
53 (82.8)
|
6 (85.7)
|
0.845
|
29 (85.3)
|
14 (100)
|
0.117
|
0.301
|
|
Bilateral
|
11(17.2)
|
1 (14.3)
|
|
5 (14.7)
|
0
|
|
|
|
Stage
|
|
|
|
|
|
|
|
|
I
|
|
7 (100)
|
|
|
13 (92.9)
|
|
|
|
II
|
|
|
|
|
|
|
|
|
III
|
|
|
|
|
1 (7.1)
|
|
|
|
IV
|
|
|
|
|
|
|
|
|
Histological subtype
|
|
|
|
|
|
|
|
|
Mature teratoma
|
63 (98.4)
|
|
|
|
|
|
|
|
Struma ovarii
|
1 (1.6)
|
|
|
|
|
|
|
|
Immature teratoma
|
|
4 (57.1)
|
|
|
|
|
|
|
Carcinoid
|
|
1 (14.3)
|
|
|
|
|
|
|
Dysgerminoma
|
|
1 (14.3)
|
|
|
|
|
|
|
Yolk sac tumor
|
|
1 (14.3)
|
|
|
|
|
|
|
Fibroma
|
|
|
|
26 (76.5)
|
|
|
|
|
Thecoma
|
|
|
|
5 (14.7)
|
|
|
|
|
Ovarian leiomyoma
|
|
|
|
2 (5.9)
|
|
|
|
|
Sclerosing stromal tumor
|
|
|
|
1 (2.9)
|
|
|
|
|
Granulosa cell
|
|
|
|
|
10 (71.4)
|
|
|
|
Steroid cell tumor
|
|
|
|
|
1 (7.1)
|
|
|
|
Sertoli-Leydig cell tumor
|
|
|
|
|
2 (14.3)
|
|
|
|
Gynandroblastoma[*]
|
|
|
|
|
1 (7.1)
|
|
|
Abbreviations: P1, including only germ cell tumors; P2, including only sex cord-stromal
tumors; P3, including germ cell and sex cord-stromal tumors.
* currently known as sex cord-stromal tumor, not otherwise specified; which is a subtype
of the mixed sex cord-stromal tumors group.
In [table 2], mean CA125 and HE4 were higher in malignant germ cell tumors, although without
statistical significance. Among them, a postmenopausal patient with carcinoid tumor
presented an HE4 value of 211.2 pmol/l. There was no difference in the expression
of these markers in women with benign or malignant sex cord-stromal tumors, regarding
mean serum levels. The analysis of biomarkers concentration by cutoff points showed
that women with malignant germ cell tumors presented significantly elevated CA125,
HE4, and CEA levels.
Table 2
Mean serum levels of CA125, HE4, and CEA and distribution by cutoff points according
to histopathological diagnosis
|
Biomarkers
|
Germ cell tumor n = 71
|
Sex cord-stromal tumors n = 48
|
|
Benign
n = 64
|
Malignant
n = 7
|
p-value
|
Benign
n = 34
|
Malignant
n = 14
|
p-value
|
|
CA 125 (U/ml)
|
63.8
|
376
|
0.37
|
90.9
|
65.8
|
0.53
|
|
(11–29.3)
|
(89.8–518)
|
|
(13–33.9)
|
(9.2–63.3)
|
|
|
HE4 (pmol/L)
|
46.5
|
114.0
|
0.44
|
64.3
|
53.5
|
0.47
|
|
(33.9–54.6)
|
(67.5–167.5)
|
|
(47.3–78.8)
|
(45.1–59.5)
|
|
|
CEA (ng/ml)
|
4.4
|
3.9
|
0.46
|
2.6
|
1.8
|
0.49
|
|
(1.2–2.9)
|
(1.6–5.8)
|
|
(1.2–3.0)
|
(1.1–2.6)
|
|
|
Biomarkers
|
|
CA 125
|
|
|
|
|
|
|
|
< 35 (U/ml)
|
46
|
2
|
0.0002
|
22
|
7
|
0.1889
|
|
≥ 35 (U/ml)
|
7
|
5
|
|
6
|
5
|
|
|
HE4
|
|
|
|
|
|
|
|
Premenopausal
|
|
|
|
|
|
|
|
< 70 (pmol/L)
|
39
|
2
|
0.00002
|
5
|
8
|
0.3747
|
|
≥ 70 (pmol/L)
|
1
|
4
|
|
2
|
1
|
|
|
Postmenopausal
|
|
|
|
|
|
|
|
< 140 (pmol/L)
|
13
|
0
|
0.0001
|
20
|
5
|
0.6187
|
|
≥ 140 (pmol/L)
|
0
|
1
|
|
1
|
0
|
|
|
CEA
|
|
|
|
|
|
|
|
< 5 (ng/ml)
|
47
|
4
|
0.028
|
28
|
14
|
0.1662
|
|
≥ 5 (ng/ml)
|
6
|
3
|
|
4
|
0
|
|
Abbreviations: CA125, cancer antigen 125; CEA, carcinoembryonic antigen; HE4, human
epididymis protein.
Missing data - CA125 for 19 women, HE4 for 18 women and CEA for 4 women.
In [table 3], there was no statistical difference related to surgical type, performed in women
with benign and malignant tumors. Fertility-sparing surgery was the treatment of choice
in 47 (73.4%) patients in benign germ cell tumor group and in 5 (71.4%) in malignant
group. As related to sex cord-stromal tumors, 12 (35.3%) of the patients in the benign
group and 11 (78.6%) of those in the malignant group were treated with fertility-sparing
surgery. According to data verified on March 22, 2019 in the digital files of patients,
the mean follow-up time of those 21 patients with malignant tumors was 44.2 months.
Besides, one patient with germ cell tumor recurred and died of the disease, and one
patient with sex cord-stromal tumor recurred.
Table 3
Women's treatment and follow-up
|
Germ cell tumor
|
|
Sex cord-stromal tumor
|
|
|
|
Benign n = 64
|
Malignant n = 7
|
P1
|
Benign n = 34
|
Malignant n = 14
|
P2
|
P3
|
|
Surgical type
|
|
|
Laparoscopy
|
28 (43.7)
|
2 (28.6)
|
0.595
|
10 (29.4)
|
3 (21.4)
|
0.4916
|
0.077
|
|
Laparotomy
|
36 (56.3)
|
5 (71.4)
|
|
24 (70.6)
|
11 (78.6)
|
|
|
|
Surgical treatment
|
|
|
Fertility sparing surgery
|
47 (73.4)
|
5 (71.4)
|
0.909
|
12 (35.3)
|
11 (78.6)
|
0.013
|
0.0034
|
|
HT + BSO/staging
|
17 (26.6)
|
2 (28.6)
|
|
22 (64.7)
|
3 (21.4)
|
|
|
|
Chemotherapy
|
|
|
|
|
|
|
|
|
Yes
|
|
5 (71.4)
|
|
|
1 (7.1)
|
|
|
|
No
|
|
2 (28.6)
|
|
|
13 (92.9)
|
|
|
|
Recurrence
|
|
|
|
|
|
|
|
|
Yes
|
|
1 (14.3)
|
|
|
1 (7.1)
|
|
|
|
No
|
|
6 (85.7)
|
|
|
13 (92.9)
|
|
|
Abbreviations: HT + BSO, total hysterectomy and bilateral salpingo-oophorectomy; P1,
including only germ cell tumors; P2, including only sex cord-stromal tumors; P3, including
germ cell and sex cord-stromal tumors.
A 21-year-old patient with yolk sac tumor was submitted to unilateral salpingo-oophorectomy
(tumor stage was Ic) and adjuvant chemotherapy. After 16 months, she presented serum
α fetoprotein elevation and imaging exam showed bladder implant, perigastric, splenic
and mesenteric lymph nodes disease. She was treated with chemotherapy, and since tumor
presented partial response; she was submitted to laparotomy 29 months after the 1st treatment. Biopsies were negative but after 5 months, imaging exam revealed intestinal
disease progression. Once more, chemotherapy was used; however, the patient succumbed
to the disease after 51 months from the diagnostic surgery.
A 53-year-old patient with gynandroblastoma was initially treated with total hysterectomy
with bilateral salpingo-oophorectomy (HT + BSO) and pelvic and paraortic lymphadenectomy.
Tumor stage was Ia. She recurred 19 months after it, and she was submitted to a laparotomy
to resect a pelvic tumor (this time an adult granulosa cell tumor) and remained without
disease until April 2017 when she was discharged from the hospital. Gynandroblastoma
is currently classified as a sex cord-stromal tumor, not otherwise specified; which
is a subtype of mixed sex cord-stromal tumor group.
Discussion
In this single center study, we evaluated 119 benign and malignant non-epithelial
ovarian tumors, as related to women's clinical features, preoperative CA125, HE4,
and CEA serum levels, surgical and chemotherapy treatment, and disease recurrence.
The majority of tumors were benign, mainly of germ cell origin (mature teratomas),
and most of malignant tumors were diagnosed in the initial stage.
Among sex cord-stromal tumors, we detected a higher incidence of fibromas. Ovarian
fibroma tumors account for ∼ 4% of all ovarian tumors. Fibromas can occur at any age,
affecting adolescents and young women, although the mean age of occurrence is in the
late forties. Interestingly, ∼ 10 to 15% of fibromas present with ascites, and less
than 1% appear with both ascites and hydrothorax, known as Meigs syndrome, mimicking
advanced ovarian cancer.[7]
[11] Fibromas usually present as solid adnexal mass on transvaginal ultrasound and, if
CA125 is elevated, suspicion of malignancy increases. Shen et al showed that elevated
serum CA125 level was found in 66 of 580 (11.3%) of patients with ovarian fibroma/fibrothecoma.
Elevated serum CA125 level was significantly correlated with tumor diameter ≥ 10cm,
ascites, and hydrothorax.[11]
Women with benign tumors presented similar mean serum levels of CA125, HE4, and CEA
when compared with women with malignant tumors. However, when we analyzed biomarker
concentration by cutoff points, women with malignant germ cell tumors were significantly
associated to elevated CA125, HE4, and CEA levels. Due to the rarity of non-epithelial
ovarian cancer, there are limited data regarding the role of CA125, HE4, and CEA in
the preoperative diagnosis of these rare tumors. In a previous study from our group,
we found that women with non-epithelial ovarian cancer did not express elevated CA125
and HE4 levels such as women with epithelial ovarian cancer.[12]
In our present study, fertility-sparing surgery was performed in 47 (73.4%) women
with benign and in 5 (71.4%) women with malignant germ cell tumors. Preconized surgical
treatment for young women is conservative (cystectomy or tumorectomy) with maintenance
of ovarian parenchyma. Maintained cortical tissue contains follicles that are able
to supply hormonal function and fertility.[13] In malignant germ cell tumors, conservative surgeries were performed, followed by
adjuvant chemotherapy in 5 (71.4%) cases, with satisfactory response, except for the
patient with yolk sac tumor. Fertility-sparing surgery was associated to chemotherapy
in only one woman with malignant sex cord-stromal tumor, because of the poor response
of this tumor to systemic therapeutics.[14] Recurrence was a relatively rare event for women with malignant non-epithelial ovarian
tumors.
Surgeons should take into account the possibility of a synchronous or asynchronous
bilateral benign ovarian tumor and focus on conservative surgery in young women. Although
uncommon, even benign ovarian tumors can recur, for example, mature teratomas present ∼ 3
to 4% postsurgical recurrence.[15] On the other hand, postmenopausal women are safely treated with HT + BSO.[16]
Conclusion
In conclusion, mature teratomas were the germ cell ovarian tumors more frequently
found in our casuistic. Among sex cord-stromal tumors, fibromas were the most common
in our sample. Malignant cases were diagnosed at initial stages with good prognosis.
Serum determination of CA125, HE4, and CEA levels were not useful for the preoperative
diagnosis of malignancy in women with non-epithelial ovarian tumors.
