Gynecological Neoplasms are the Major Cause for Malignant Effusions

Background: Malignant effusions occur frequently in gynecologic neoplasms and may the prognosis for these patients. Therefore a more detailed analysis for the different cavities is necessary to describe the association between the various histological subtypes and the time related occurrence of malignant effusions in gynecological malignancies. Malignant effusion specimens from patients diagnosed Bayreuth Hospital 2018 were reevaluated retrospectively and correlated with the histolgogical subtype of primary tumors and with the clinical follow-up.

Conclusions: Gynecological neoplasms were the major cause of malignant effusions in our study. The high-grade serous papillary subtype of lower tract gynecological is most aggressive with predominant occurrence in the peritoneum and exclusive secondary involvement of the pleural cavity. Therefore, an alone/exclusive/sole involvement of the pleural cavity characterize the invasive breast carcinoma of no special type compared to predominant occurrence of the lobular subtype in malignant ascites. Breast cancer showed a statistical signi cant late occurrence in effusion uids 12 months after primary diagnosis in contrast to early involvement in gynecological-as well as pulmonary-or gastrointestinal malignancies.
The occurrence of malignant effusions complicates the clinical follow-up and worsens the prognosis for the tumor patients. In contrast to lung cancer, breast cancer is characterized by a late metastatic pattern in malignant pleural effusions, particularly in the main groups of hormone receptor-positive or HER2receptor-positive patients [13]. The mean period between diagnosis of a primary tumor and the development of malignant pleural effusion has been reported as 41.5 months, with a mean survival time after malignant effusion of 15.7 months [14]. The mean survival time for patients with malignant ascites is less than 20 weeks, and for those with ovarian tumors is 32 weeks [15].
Cytological examination has become the gold standard for detecting tumor cells in effusion uids in comparison to serological analysis. Ancillary immunostains using the cell block technique may allow a better discrimination between epithelial cells and mesothelial proliferations and helps to detect the phenotype of the tumor cells [16,17].
The present large retrospective single center study, including 435 patients, provides an overview of the distribution pattern for malignant tumors in the different cavities with special emphasis on gynecological neoplasms. We suppose that a more detailed analysis of the histological subtypes of gynecological cancer may better characterize the heterogenous metastatic potential of these tumors and in addition we

Methods
In a retrospective analysis of the period from June 2013 to December 2018, cytological ndings for malignant effusion were reevaluated at the Institute of Pathology in Bayreuth Hospital. We were focussed to nd correlations between the occurrence of malignant effusions to the different histological subtypes of gynecological malignancies in this collective. Histological probes had been analyzed using Hematotoxylin-Eosin stain and in addition we characterised receptor-and Her2-status of breast cancer patients ( Table 1). The results we compared with pulmonary and gastrointestinal neoplasms. For a better overview we summarized a group of other epithelial tumors, which mainly includes urological malignancies and a group of nonepithelial tumors comprising malignant lymphomas, malignant melanomas, malignant mesotheliomas and sarcomas. We correlated our cytological data with previously histological and cytological reports and described a simultane occurrence of malignant effusion from different cavities, if we received the probes within one months. Effusion uids from the peritoneal cavity includes intraoperative lavage probes in 32 female patients with ovarian carcinoma and 6 with endometrium carcinoma. The specimens had been routinely examined using May-Grünwald-Giemsa and Papanicolaou staining. Additional immunostaining were carried out using the cell block technique (Table 1). Approval for the retrospective analysis of malignant ascites was received from the Ethics Committee at Friedrich Alexander University, Erlangen (no. 267_19 Bc).
To examine the relationship between occurence of NST (invasive carcinoma of no special type) or lobular breast cancer and pleura(l) effusion or malignant ascites as well as the time related occurence of malignant effusions, odds ratios (OR) were calculated and Fisher's exact test was performed. The signi cance level was set to 0.05. All statistical analysis was performed using the statistical programming language R V3.6.3 [18].

Discussion
Gynecological cancers occurs most frequently in 34% of our patients with malignant effusions including lavage probes for ovarian tumors. These ndings becomes importance, because they results from a single center, which exclude an unproportional in uence of any clinical section on the tumor distribution in comparison to various other studies [2,3,4,7,8,11,15,[19][20][21][22][23]. Further leads the predominance of gynecological tumors to an overbalance of female patients in malignant effusions and required both from clinical as well as from immunohistochemical viewpoint a sex speci c data analysis [1].
Most (65.3%) of all gynecologic malignancies we diagnosed in the peritoneal cavity and in 71.1% of female patients with malignant abdominal cytology we observed gynecological neoplasms. It is widely accepted in the literature that ovarian carcinoma is the prototype of dissiminated peritoneal spread in malignant ascites [5,6,12,[24][25][26][27]. In our previous studies on ovarian cancer we described the highest malignancy rate, at 85.7%, the earliest occurrence, and a predominance of the high-grade papillary subtype in patients with malignant abdominal cytology [10]. The correlation between malignant abdominal cytology and high-grade serous papillary subtype we could extent in the present study on endometrium carcinomas. These ndings underline the discrimination between non aggressive type 1 endometrium cancer with lacking occurrence of malignant ascites in our study from aggressive type 2 endometrium cancer with appearance of malignant ascites in high-grade serous papillary subtype and also MMMT [28,29]. Moreover, we obtained in tumors from the lower genital tract the highest rate of 31.4% for secondary pleural involvement compared to 6.1% for female patients with GI tumors. Therefore is the knowledge of previous ndings for all patients necessary, which exclude in our study an alone pleural involvement for ovarian carcinomas. These nding is remarkable because ovarian carcinoma achieves in various other studies the third or fourths most common epithelial tumor of malignant pleura effusions [12,24,25]. Further it may affect the dayly diagnostic practice, because in both involvement of peritoneal and pleural involvement a gynecological malignancy is more probably, whereas GI tumors showed a predominant peritoneal or an infrequently pleural involvement. The exclusive involvement of peritoneal and pleural cavity by high-grade serous papillary subtype highlights the aggressive potential of this subtype and should be noted at initial diagnosis for continous follow-up controls.
Malignant ascites from gynecological tumors include in 10.5% patients with breast cancer, which showed a lobular subtype in 80% with a characteristic diffuse pattern in effusion uids and is in agreemet with literature data [30,31]. In contrast to other gynecolocigal and gi malignancies appears malignant ascites in breast cancer patients only in 20% within 12 months, compared to 60% in endometrium cancer, 98.7% in ovarian carcinoma and 97% in female patients with gastrointestinal tumors. These late metastatic potential in patients with breast cancer should be correlated with a positive receptor state for this patients, which we also obtained in 90% of our patients [13,[32][33][34][35].
Second most frequent gynecological neoplasms were found in 33.3 % of malignant pleura effusions after patients with lung cancer [12,14]. As mentioned above reduced our cytological clinical correlation the occurrence of malignant pleura effusions in patients with gynecological neoplasms to breast cancer and if we compared only female patients we received a slight overbalance of 49 patients with breast cancer to 48 patients with lung cancer. In contrast to malignant ascites the most often occurring subtype of NST carcinomas predominates in malignant pleura effusion. Similar to malignant ascites they obtained only in 22.4% of the patients within 12 months after primary diagnosis compared to 81.2% for female lung cancer patients and 88.2% in female patients with GI tumors [33,35]. These time associated occurrence of breast cancer in malignant pleural effusion may help at rst in dayly diagnostic practice to discriminate lung from breast carcinoma and could further prove by immunostains.
Only 1.4% of the malignant effusions we could detect in the pericardium with female predominance and similar tumor distribution as in the pleural cavity, which is in agreement with literature data [12,36,37,38]. Also similar clinical symptoms, such as dyspnea, cough and chest pain characterize patients both with pleural and pericardial effusions [19,37]. Patients with malignant pericardial effusion were 7.3 years younger than patients with malignant pleura effusions and the effusion uid is extremely bloody because of cavitary and intralymphatic tumor spread [4]. We diagnosed only 1 patient with breast cancer and 1 patient with cervix carcinoma and both patients showed an involvement of the pericardium within 12 months. These major ndings correlated with literature data, however, the number of patients is low [12,37,38].
Hematological neoplasms were the most common nonepithelial tumor in malignant effusions in our study, occuring in 3.4% of the patients with female predominance in 66.6%. In agreement with the ndings in the literature 80% of them were located in the pleura and 20% in the peritoneum [39][40][41].
Immunohistochemically, B-cell lymphomas were represented in 86.7% and high grade subtype in 66.6% of the patients. The cytological impact of these ndings is the diffuse tumor cell pattern for malignant lymphomas in differential diagnosis to lobular breast carcinoma and all other diffuse enddifferentiated carcinomas, which require applications of further immonstains.
Strengths of the present study are the large number of patients included. Limitation of the study is its retrospective single center design and the fact that the effusion uids from the peritoneal cavity includes intraoperative lavage probes. Availability of data and materials Tumor origin of malignant effusions in Bayreuth Hospital