Introduction

Ovarian carcinoma is the fifth most common carcinoma in women [1], and is the leading cause of death from gynecologic malignancies [2, 3]. Metastatic ovarian tumors are neoplasms of the ovaries which usually originate from distant organs such as the stomach, the breast or the colon and constitute 5% of the total ovarian tumors and 10–30% of the total malignant ovarian tumors [410].

This difference in the incidence of metastatic ovarian neoplasms may be explained by various factors such as geographic distribution (e.g. great incidence of gastric cancer in Japan), the age groups, the experience of the laboratory personnel examining the specimens, and the laboratory methods used in the correct diagnosis [11, 12].

In many cases, metastatic ovarian tumors are detected by the pathology examination before the clinical detection of the primary tumor. It is reported that metastatic ovarian tumors preceded the primary tumor detection in 38% of patients [13].

In certain cases there are difficulties in the differential diagnosis of primary or metastatic ovarian neoplasms by routine histopathological examination. The main diagnostic problems include metastatic ovarian neoplasms from colon carcinomas and primary ovarian mucinous cystadenocarcinomas. Up to 45% of metastatic ovarian neoplasms from colon cancer have been primarily diagnosed as primary ovarian neoplasms [14, 15].

In cases of breast carcinomas, microscopic ovarian metastases have been diagnosed after prophylactic oophorectomy in 50% of the cases and the ovaries in 63.3% of cases present a grossly normal appearance [16].

In this study, we present the clinical and histopathological characteristics of metastatic ovarian neoplasms and the overall 3-year survival of those patients and the immunohistochemical characteristics of the tumors that are helpful in the correct diagnosis

Materials and methods

A retrospective study was conducted in the 2nd Department of Obstetrics and Gynecology, University of Athens, Aretaieion Hospital, Athens, Greece. During the period January 2000 to December 2009, 1,055 ovarian neoplasms were managed in our department from which 520 were primary malignant tumors and 97 were metastatic ovarian neoplasms. All the patients were operated in our department and the histopathological examination was performed in the Histopathology Department of our hospital. The available clinical records of patients with metastasis to the ovary were reviewed for age at diagnosis, origin of the neoplasm, consistency and appearance of the neoplasm, histological type and the 3-year survival. The gross, microscopic and immunohistochemical findings of metastatic ovarian neoplasms were reviewed from the pathological files.

All specimens were formalin fixed, routinely processed and stained by hematoxylin and eosin. The immunohistochemical study was performed by Ventana automated system with the use of following abs: CK7 (Monosan, 1:25), CK20 (DAKO, 1:50), CA125 (Monosan 1:25), CEA (Monosan 1:25), ER (Novocastra 1:50), PgR (Thermoscience 1:50), TTF1 (Cellmarque 1:25).

The survival rates were calculated in months from the time of surgical intervention.

Results

Ovarian metastatic neoplasms comprise 97/1055 (9.19%) of all ovarian neoplasms and 97/520 (18.65%) of the total malignant ovarian neoplasms treated in our Clinic.

The median age of the patients was 55 years (range 26–78 years).

43/97 (44.33%) patients were premenopausal, whereas 54/97 (55.67%) were postmenopausal. 88/97 (90.72%) had a previous cancer history, 21/97 (21.65%) patients reported oral contraceptive use and 9/97 (9.28%) patients reported hormone replacement treatment (Table 1). In relation to localization of the metastatic ovarian neoplasms, 51/97 cases (52.58%) had bilateral development, 24/97 (24.74%) localized in the right and 22/97 (22.68%) in the left ovary.

Table 1 Characteristics of patients with metastatic ovarian tumors

Of the total study group, 61 cases (62.89%) originated from non-genital tract malignancies and 36 cases (37.11%) from genital tract malignancies (Table 2).

Table 2 Origin of the primary tumor

Forty-four cases (%) originated from the gastrointestinal tract (24 cases of gastric carcinoma, 15 cases of colon carcinoma, 3 cases of appendix carcinoma, 2 cases of pancreatic carcinoma) and 15 cases originated from the breast.

Of the primary genital tract malignancies that metastasized to the ovary, endometrial neoplasms were the most common (18/97).

Gross morphology

45/97 of the metastatic ovarian neoplasms were ovoid or kidney-shaped, often with cystic areas occupying <50% of the tumor mass. The external surface was generally smooth and nodular in 43 cases and irregular in 2 cases. In 40 cases extensive necrosis was observed.

Carcinomas metastatic from colon cancers were large and cystic (15/15 cases) with the exception of gastric carcinomas which were large and solid tumors presenting the typical gross and microscopical morphology of Krukenberg tumors in 20/24 cases.

The mean diameter of the metastatic ovarian neoplasms was 7.1 cm with a range of 1–24 cm.

The smallest ovarian neoplasms originated from the breast, whereas the largest ovarian neoplasms originated from the colon with the biggest recorded metastatic neoplasm weighting 1,580 g (Figs. 1, 2, 3).

Fig. 1
figure 1

Bilateral metastasis of a pancreas cancer to the ovary with characteristic kidney-shape tumor (Krukenberg)

Fig. 2
figure 2

Small ovary with a characteristic nodular surface due to breast cancer

Fig. 3
figure 3

Gross section of a large metastatic colonic tumor with multicystic mucinous surface

Histology

All cases of the primary gastrointestinal tumors were adenocarcinomas (100%) and in 20/24 cases of gastric cancer a signet ring cell morphology was observed (Figs. 4, 5).

Fig. 4
figure 4

Signet ring cells of metastatic ovarian tumor from the stomach (Krukenberg tumor)

Fig. 5
figure 5

Microscopic findings of metastatic ovarian tumors (“dirty” necrosis)

The predominant histological type of the metastatic breast tumors was that of infiltrating ductal cell carcinoma (NOS) (13 cases) and in two cases a lobular type was observed.

Metastatic genital tract tumors were endometrioid adenocarcinomas in 12/18 cases (66.67%) serous papillary carcinomas in 6/18 cases (33.33%) squamous cell cervical cancers in 9 cases and in 3 cases a primary fallopian tube cancer was diagnosed.

The cases of metastatic sarcoma, melanoma and lung cancer presented a typical morphology that was consistent with the previously reported primary tumors.

Metastatic GI tract cancers that presented the main diagnostic problem because of their morphological similarity with primary ovarian mucinous tumors had a distinct immunoprofile (CK7−, CK20+, CA125−, TTF1−, ER−). Metastatic breast cancer also presented a distinct immunoprofile (ER+, PGR+, CK7+, CA125−, CK20−, TTF1− (Fig. 6). The rest of the cases presented an histological morphology that in combination with the clinical history permitted the correct diagnosis.

Fig. 6
figure 6

Immunohistochemical findings of metastatic ovarian tumors [CK20 (+), CEA (+), CK7 (−), ER (−), PR (−), CA125 (−)]

The 3-year survival of patients with metastatic ovarian neoplasms was 26.8% after surgical resection of the neoplasms.

The 3-year survival rate of the 63 patients with metastatic ovarian neoplasms from a non-genital tract site were 25.4%, and of the 34 patients with metastatic ovarian neoplasms from a genital tract site the survival rates were 29.4% (Tables 3, 4).

Table 3 Survival time and proportion after surgical removal of ovarian metastatic neoplasms originating outside the genital tract
Table 4 Survival time and proportion after surgical removal of ovarian metastatic neoplasms originating from the genital tract

Discussion

The differential diagnosis between primary and metastatic ovarian neoplasms is important and the primary site of the latter should be diagnosed by the histopathological examination of the ovary. However, according to Yada-Hashimoto et al. [16] not all histopathological examinations lead to a diagnosis of the primary site. About 4.4% of female patients who died from malignant neoplasms were found to have ovarian metastases at autopsy [18], while 25% of patients with breast carcinoma who underwent prophylactic oophorectomy had ovarian metastases [16].

In our Department, the incidence of metastatic ovarian neoplasms was 97/520 malignant ovarian tumors (18.65%).

However, the incidence of such malignancies ranged from 6 to 30% in previous studies [14, 17, 1921]. It should be mentioned that our material did not include cases of prophylactic oophorectomy for breast carcinomas.

The main primary sites of ovarian metastatic neoplasms were as follows in our study: gastrointestinal tract 44/97 (45.36%), breast 15/97 (15.46%) and genital tract 36/97 (37.11%). Our study is consistent with previous studies which showed that the most common origin for metastatic ovarian neoplasms is the gastrointestinal tract [19]. However, the Mediterranean diet in comparison to the Japanese diet is considered responsible for the difference in the incidence of ovarian metastatic gastric carcinoma (24/97—24.74% of the cases in our study) compared to the results of the study of Yada-Hashimoto et al. [16]. Moreover, in our study, Krukenberg tumors were identified in 20/97 cases (20.62%), a finding which is in accordance with previous studies [21]. It should be noticed that we characterized neoplasms as Krukenberg tumors according to the Woodruff and Novak criteria [22] which defined that Krukenberg tumor is a tumor with fibrous stroma infiltrated by characteristic mucin-filled, signet ring adenocarcinoma cells.

The median age of the patients in this study was 55 years ranging from 26 to 78 years. The finding seems to be higher than previous studies; however, the high incidence of metastatic ovarian neoplasms originating from the genital tract could explain it as such malignancies are found in older ages [15]. Moreover, in our study, the younger median age (46 years) of metastatic ovarian neoplasms originating from the stomach should be mentioned.

Localization of the neoplasms tends to be bilateral and is attributed to lymphatic dissemination of breast carcinoma and neoplasms of the upper abdomen [5, 15].

Cystic tumors were found in 48/97 cases (49.48%) in our study and 39/48 of them (81.25%) originated from tumors of the gastrointestinal tract and the finding is in consistence with previous studies [20, 23]. Breast metastases tend to create small nodules in the ovary and usually are inconspicuous on observation.

The diameter of most ovaries in our study did not exceed 7 cm and is in accordance to the reported size of metastatic ovarian neoplasms described in previous studies [24, 25].

The immunohistochemical investigation showed that colorectal cancers metastatic to the ovaries, which present the main differential diagnostic problem from primary mucinous tumor, present a distinct immunoprofile [tumor cells CK20 (+), CEA (+), CK7 (−), ER (−), PR (−), CA125 (−)] that permit the accurate diagnosis [16].

The survival of patients with metastatic ovarian neoplasms is rather poor ranging from 19 to 47% in neoplasms originating from non-genital tract to genital tract organs, respectively [17]. In that study, the median survival time of patients with ovarian metastatic neoplasms with origin genital tract tumors was 48 months, while for those originating from non-genital tract tumors was 12 months. In our study, the overall 3-year survival for surgically managed ovarian metastatic neoplasms was 31/97 (31.96%), while 15/36 (41.67%) and 16/63 (25.4%) were the survival rates in the genital and non-genital tract tumors, respectively.

In conclusion, patients with ovarian metastatic neoplasms are in the fifth to sixth decade and present bilateral lesions of the ovaries. The primary sites are more frequently the gastrointestinal tract which create cystic mucinous tumors, the breast, and the endometrium. Immunohistochemistry is a valuable aid in the correct diagnosis of metastatic tumors especially in cases of neoplasms originating from the colon.

The presence of metastatic ovarian neoplasms should always be included in the differential diagnosis of a pelvic tumor to avoid diagnostic or treatment pitfalls. The prognosis is generally poor since these tumors represented cases in late stages of the disease.