Conservative treatment of borderline ovarian tumors: a retrospective study.

Borderline ovarian tumors (BOTs) must be distinguished from ovarian carcinomas because they present differently and can be treated differently. Looking at the degree of aggressiveness between epithelial ovarian carcinomas and benign ovarian neoplasms, such as cystadenomas, borderline neoplasms are closer to benign ovarian neoplasms in terms of their clinical behavior [ 1 ]. Unlike ovarian neoplasms, BOTs are often associated with infertility, have a lower incidence, a lower association with breast cancer (BRCA) genes, are diagnosed at early stages and have a good survival rate even when they involve a peritoneal lesion. They are more common in younger patients, so one of the goals in these patients is fertility preservation [ 2 , 3 ]. The treatment of these tumors is highly controversial and continues to be debated. Complete staging with conservative and radical surgical treatment is very important to choose the right treatment and have a good prognosis. To preserve fertility, BOTs in young patients can be treated with conservative surgery [ 4 ]. In most cases, natural fertility is preserved, and spontaneous conception may occur after conservative treatment. Unfortunately, some patients become infertile even after such conservative surgery. The management of such cases is an important dilemma. Some infertile BOT patients undergo infertility treatment to achieve improvement. In this context, the role of infertility treatment in the occurrence of ovarian cancer is controversial [ 5 ]. Aim The aim of this study was to conduct an analysis of the literature and scientific data from the medical community on the diagnosis, conservative management, and treatment of infertility in BOTs.

BOT is suspected when a patient with a serous or mucinous ovarian cyst is found to have a change in the ultrasound (US) image of the cyst on successive examinations, such as an increase in size or the appearance of septa or intracystic vegetations. Although the tentative diagnosis of BOT is made analytically, macroscopically, paraclinically or by US, magnetic resonance imaging (MRI) and positron emission tomography (PET), we cannot distinguish BOT from other ovarian tumors [ 14 , 15 , 16 ]. Since it is a histological diagnosis, BOTs are usually diagnosed based on extemporaneous histopathological examination or paraffin examination [ 17 , 18 ]. The determination of tumor markers, such as cancer antigen 125 (CA125), is not effective for diagnosis. In about 50% of patients, it is in the normal range and rarely exceeds 100 U/mL [ 19 ]. The final diagnosis is made histologically: proliferation of epithelial cells and stratified epithelium, papillary protrusions, mitotic activity, nuclear atypia and cellular pleomorphism. In addition, there must be no stromal invasion, which distinguishes it from invasive carcinoma [ 3 , 6 , 10 ]. In about 10% of cases of BOT, areas of microinvasion occur, in which cells with features of BOT are found. They occur as foci less than 5 mm in diameter or invade the stroma up to an area of 10 mm2. Stromal microinvasion is more common in serous BOTs. It is associated with the presence of micropapillary patterns and peritoneal implants. This is an independent prognostic factor and a sign of progression to an invasive form. Peritoneal spread of BOT in the form of implants, if epithelial, is considered a non-invasive form (85%). In this case, only the peritoneal surface is affected, in contrast to invasive implants, which also spread to the adjacent structures, the intestinal wall or the omentum. For a complete histological diagnosis, it is necessary to perform IHC tests that monitor the expression of the CK7, CK20, Bcl-2, ERα, PR, p53 and Ki67 markers. CK7 is a marker for malignant epithelial tumors of the ovaries, present in borderline tumors of the serous subtype but not in the mucinous and 11 endometrioid subtypes. Wilms tumor 1 (WT1) is a core marker for malignant serous epithelial tumors, present in 80% of cases. P53 and Ki67 are markers of tumor aggressiveness [ 20 ]. The association between BOTs and BRCA1 and BRCA2 gene mutations is uncertain. However, it appears that BRCA gene mutations occur in a lower percentage of borderline tumors than invasive ovarian cancers [ 20 ].

The most important feature from a histological point of view is the absence of stromal invasion. In a small number of cases, BOTs may show implants on the peritoneal surface even without stromal invasion. BOTs show stratification of the epithelial lining of the papillae with cell proliferation, nuclear atypia, and mitotic activity. Several histological subtypes are described, many of which are serous tumors (approximately 53–65%). Mucinous BOTs account for 32–42% of all BOTs. The remaining BOTs are endometrial tumors, clear cell tumors, burner tumors and other histologies. Of all borderline tumors, about 70% are serous. They are characterized by a micropapillary pattern, invasive implants, and a less favorable prognosis than the other subtypes. A significant percentage of BOTs (about 50%) are bilateral [ 10 ]. Serous BOTs are divided into two subtypes: the most common pattern (90%) is a unilocular cyst with thin septa inside. The micropapillary pattern (10%) has important histological features (micropapillary appearance with a length of more than 5 mm or with a tumor area of more than 10%). The latter have a poor prognosis, as most of them have a higher recurrence rate for the invasive form, a higher percentage of bilateral forms and the association of implants with a tendency to overstay at a later evaluation [ 11 , 12 ]. Recent articles suggest that serous BOT, which have a micropapillary pattern but no implants (stage I), or which have non-invasive implants (stages II and III), have a similar evolution to serous BOT without a micropapillary pattern. It can be concluded that the presence of implants and their invasiveness are suspicious factors for malignant disease. Mucinous BOTs account for approximately 30–50% of all BOT and are less commonly bilateral (7%) [ 11 , 12 ]. They are unilateral tumors often associated with a peritoneal pseudomyxoma (arising from the appendix). Two histological subtypes are distinguished: intestinal subtype (85–90%); most of these cases are unilateral; if bilateral, primary colorectal carcinoma must be excluded. The intestinal type is typical of older age, occurs in older women and has multilocular cysts. This type is associated with pseudomyxoma peritonei and has a good evolution; endocervical or Müllerian cysts (10–15%), they are bilateral in at least 40% of cases. They are associated with pelvic endometriosis or ipsilateral endometriomas in 20–30% of cases and BOT with mixed histology (seromucinous). The endocervical type occurs in younger women and is more common bilaterally (20–30%), has a unilocular cystic tumor, is more advanced in stage, correlates with implants or lymph node metastases and its mortality rate can be as high as 50% depending on stage [ 13 ]. The other histological subtypes – endometrioid tumors with clear or transitional cells (Brenner) – are a rarity. We present five aspects of the histological appearance of BOTs from our casuistry (Figures 1 , 2 , 3 , 4 , 5 ). In the first case, we encountered mucinous BOT – intestinal mucinous epithelium and epithelial proliferation with stratification and villus formation (Figure 1 ). In the second case, we identified mucinous BOT with complex architecture, low-grade dysplasia, and stromal invasion (Figure 2 ). In the third case, we identified mucinous BOT with the complex architecture of the intestinal type of epithelium with cystic transformed areas, papillary growths, villi, and areas of low-grade dysplasia at the level of the epithelial islands invading the stroma. In the fourth case, we identified mucinous BOT – epithelial proliferation with disturbed architecture and areas of cellular atypia at the level of the epithelium (Figure 4 ), and in the last case we identified the presence of low-grade borderline serous ovarian tumors. When immunostaining with the anti-CK7 antibody, we observed moderately positive intracytoplasmic and membranous immunostaining (Figure 5A ). For immunostaining with the anti-CK20 antibody, we observed negative membranous and cytoplasmic immunostaining (Figure 5B ). In the case of immunolabeling with the anti-Ki67 antibody, we obtained nuclear positive focal immunostaining of around 10% for nuclei (Figure 5C ). In the case of immunolabeling with the anti-p53 antibody, we obtained moderately positive nuclear immunostaining (Figure 5D ). In the immunoreaction performed with the anti-Bcl-2 antibody, we obtained moderately positive intracytoplasmic immunostaining (Figure 5E ), and for the immunolabeling for ERs, we obtained moderately positive nuclear immunostaining (Figure 5F ), and for PRs, we obtained intensely positive nuclear immunostaining (Figure 5G ). Mucinous BOT: intestinal mucinous epithelium and epithelial proliferation with stratification and villus formation. HE staining, ×200. BOT: Borderline ovarian tumors; HE: Hematoxylin–Eosin Mucinous BOT: complex architecture, low-grade dysplasia, and stromal invasion. HE staining, ×400 Mucinous BOT: the complex architecture of the intestinal type of epithelium with cystic transformed areas, papillary growths, villi, and areas of low-grade dysplasia at the level of the epithelial islands invading the stroma is observed. HE staining, ×200 Mucinous BOT: epithelial proliferation with disturbed architecture and areas of cellular atypia at the level of the epithelium. HE staining, ×200 Microscopic features of low-grade BOTs: (A) Moderately positive intracytoplasmic and membranous immunostaining for anti-CK7 antibody (×200); (B) Negative membranous and cytoplasmic immunostaining for anti-CK20 antibody (×100); (C) Nuclear positive focal immunostaining of around 10% for anti-Ki6 antibody (×200); (D) Moderately positive nuclear immunostaining for anti-p53 antibody (×100); (E) Moderately positive intracytoplasmic immunostaining for anti-Bcl-2 antibody (×200); (F) Moderately positive nuclear immunostaining for anti-ER antibody (×100); (G) Intensely positive nuclear immunostaining for anti-PR antibody (×200). Bcl-2: B-cell lymphoma-2; CK: Cytokeratin; ER: Estrogen receptor; Ki67: Cell proliferation factor; p53: Tumor protein 53; PR: Progesterone receptor

There is no single hypothesis, but it is more likely that an overlap of mechanisms involved in each hypothesis fully explains the reproductive effects. The ovulation hypothesis assumes that the occurrence of a malignant tumor in the ovaries is due to repeated microtraumatization of the ovarian surface epithelium (OSE) at ovulation [ 7 ]. Pregnancy, breastfeeding, and hormonal contraception inhibit ovulation and have a protective effect. A 2–4-fold increased risk of BOTs has been observed in patients taking ovarian stimulation drugs prior to in vitro fertilization (IVF) [ 8 , 9 ], and the hypothesis of malignant transformation due to exposure to hormonal stimulation is confirmed by the registration of an increased number of borderline serous ovarian tumors after ovarian stimulation and multiple ovaries [ 9 ]. The risk of borderline serous ovarian tumors is increased when estrogens are administered in excess without compensation by progesterone, and in obesity when conversion occurs in the periphery. Progesterone can lead to the “elimination” of cells on OSE that have sublethal deoxyribonucleic acid (DNA) damage by inducing apoptosis. Androgens have also been shown to cause tumor cell growth, which has been observed in polycystic ovary syndrome (PCOS) [ 9 ].

The prognosis of BOTs is excellent, except that 11% of these tumors recur and 20–30% undergo malignant transformation [ 10 ]. It has been noted that there is no predictability in BOTs regarding the evolution of prognosis in recurrences. Previously, five characteristics were associated with poor prognosis: cell type, stage, implant type (for serous borderline tumors), presence of micropapillary architecture (for serous borderline tumors) and microinvasion [ 21 ].

The standard treatment for BOTs is radical and based on hysterectomy, bilateral salpingo-oophorectomy, and peritoneal staging. The recurrence rate increases to 15–35% but has no impact on survival. The spontaneous pregnancy rate is almost 50% [ 22 , 23 , 24 ]. Fertility-preserving treatments, which preserve the uterus and at least part of an ovary, have evolved greatly in recent years, with 33% of BOTs diagnosed in patients less than 40 years of age. Conservative treatment, preserving at least part of the ovary and uterus, may be recommended to preserve fertility, at least in early-stage patients BOT. In serous BOT, conservative surgical treatment depends on several factors, the most important of which are the presence of healthy ovarian tissue, which can be detected by US and MRI, the number of antral follicles and the age of the patient. Determination of serum levels of anti-Müllerian hormone is only required for bilateral BOTs. As for conservative surgical treatment, four surgical techniques are described: unilateral adnexectomy, unilateral adnexectomy with contralateral cystectomy, unilateral cystectomy, and bilateral cystectomy. Peritoneal cytology and biopsy of the contralateral ovary were systematically performed as part of conservative surgery. These surgical procedures are also associated with multiple peritoneal biopsies, omentectomy and pelvic or para-aortic lymphadenectomy. The recurrence rate after conservative treatment is 0–25% compared to 0–5% after radical surgery [ 25 ]. Complete staging is not required for unilateral BOTs unless macroscopic peritoneal micropapillary implantation patterns are found. Extraovarian extension is present in about 15% of unilateral tumors, so thorough intraoperative exploration is required. In bilateral BOTs, the percentage of association with extraovarian extension is about 56%. For this reason, complete staging, and removal of as many implants as possible is required, with the omentum being the preferred site [ 26 ]. In the case of BOT with only one ovary, cystectomy is the practiced surgical technique. An exception to this rule applies to very young patients, in whom as much functional ovarian tissue as possible should be preserved to detect possible recurrences with the help of transvaginal US. Later surgical intervention to preserve fertility is possible because US can detect even very small recurrences in the normal ovarian parenchyma, so that functional ovarian tissue can be preserved [ 27 ]. The incidence of postoperative recurrence is 30–31%. The most important factors for detecting recurrence are: intrafocal ovarian tumor, tumor resection margins and intraoperative rupture of the BOT [ 28 , 29 ]. Even in unilateral BOTs, mucinous ovarian tumors are often associated with overtly invasive lesions. In this case, conservative surgical treatment is not recommended [ 30 ]. In patients with large bilateral BOTs, bilateral salpingo-ovariectomy is performed, in which the ovarian tissue cannot be preserved. Oocytes retrieved before bilateral salpingo-ovariectomy can also be preserved [ 31 , 32 ]. Pregnancy rates in patients who have undergone conservative surgical treatment vary widely, ranging from 30–80% [ 33 , 34 , 35 ]. When cystectomy was performed, fertility outcomes were better if the patient’s age was less than 40 years, and it was histologically a non-serous BOT type. In 10–35% of cases, patients with BOT have a history of infertility, especially if BOT is serous, bilateral or has a micropapillary pattern [ 36 , 37 ]. In BOT, surgical procedures are a cause of infertility as they determine the formation of adhesions and the impairment of ovarian function and oocyte reserve [ 38 ]. If conservative treatment is aimed at preserving fertility, the question arises as to how to treat subsequent infertility, especially by ovulation induction. Laparoscopy is increasingly used in patients treated conservatively. Laparoscopic treatment of BOT is associated with rupture of cysts and incomplete staging. In a review by du Bois et al., a high recurrence rate was observed with laparoscopic conservative treatment compared to laparotomy (14.9 vs. 7.7%) [ 39 ].

The final diagnosis at BOT is made histologically. There is a potential risk of infertility that must be considered in infertile patients following conservative surgical treatment of BOTs. If infertility remains in patients diagnosed with early-stage BOT, IVF should be considered. In this case, it is necessary to limit the number of IVF cycles. In patients with advanced stage BOT, long-term clinical trials are needed to specify the safety and efficacy of conservative treatment COH-IVF. It is currently suggested that IVF may be considered for infertile patients with early-stage BOT. Patients should be informed of the potential risks associated with COH. After COH-IVF, careful follow-up is required.

There are patients who are infertile despite conservative surgical treatment. For this reason, assisted human reproductive technologies are used. Previous infertility and previous use of drugs to stimulate the ovaries are associated with the risk of BOT. A case-control study draws attention to the fact that infertility per se increases the risk of ovarian cancer [ 40 ]. Other studies reported an increased risk of invasive and BOTs after repeated protocol stimulation with Clomiphene citrate (over 12 cycles) or after treatment with human menopausal gonadotropin [ 41 ]. In early-stage patients BOT the risk of ovarian hyperstimulation is relatively low. The impact on pregnancy achieved after fertility treatment shows that BOTs have no obvious negative impact on history, diagnosis, and treatment. There is no evidence in the literature that the use of IVF needs to be limited in patients with early-stage BOT after conservative treatment [ 42 ]. There are some studies on the safety of controlled ovarian stimulation (COH) with IVF in advanced-stage BOTs after conservative treatment. Other studies show that the presence of invasion in a peritoneal implant is a poor prognostic factor; 16% of them had recurrence with a non-invasive implant and 64% with an invasive implant [ 43 , 44 ]. Since patients with invasive peritoneal implants have a poor prognosis, it is logical to recommend conservative treatment only for BOTs without invasive implants [ 45 , 46 , 47 ]. Previous studies have not shown that pregnancy affects the development of BOT. In advanced stages, we need to be more careful and repeat IVF cycles with stimulation several times. At these stages, accelerated progression from BOT to invasive ovarian cancer has been reported [ 48 , 49 , 50 , 51 , 52 ]. It is possible that the pathogenic mechanisms of tumor progression are related to hormonal influences [ 53 , 54 ]. Recently, expression of ERs has been demonstrated in BOTs, and high serum estradiol (E2) levels occurring in COH with IVF may play a role in tumor development [ 55 ]. It should be emphasized that IVF procedures do not increase the risk of BOT or ovarian cancer [ 56 ]. Other studies suggest that gonadotropins and high doses of E2 in IVF cycles do not induce proliferation of borderline cell cultures [ 57 ]. However, few authors have reported their experience with the dilemma of whether IVF can be performed in the case of BOT with infertility. A literature review of these medical studies shows a median estimate for pregnancy of 80% [95% confidence interval (CI): 68–92%] and for recurrence of 23% (95% CI: 6–39%). This low recurrence rate is because patients who are recommended and undergo IVF have a better prognosis since they are diagnosed at an early-stage BOT. Patients diagnosed with BOT who wish to become pregnant should be assessed by a complex multidisciplinary team of oncologists and fertility experts. This is necessary to make a correct evaluation of the chosen conservative surgical treatment as well as the results of assisted human reproductive technology.

Ovarian tumors with borderline malignancy account for 10–15% of epithelial ovarian malignancies. Compared to patients with ovarian neoplasms, patients with BOT are younger at the time of diagnosis. They are more common in women around 40 years of age, but 27–36% of cases occur at a younger age. Most BOT are diagnosed at an early stage, 70–80% at stage I, compared to 25% of carcinomas. BOT and ovarian cancers do not differ significantly in terms of epidemiological substrate, and numerous studies have confirmed that BOT has similar reproductive effects as ovarian cancers, except for a higher incidence of infertility [ 6 ].

The authors declare that they have no conflict of interests.

This study included a comprehensive review of the literature using PubMed, Scopus, and MEDLINE databases to demonstrate the impact of conservative management of borderline tumors on fertility. The keywords used were borderline ovarian tumor, histology, conservative treatment, and fertility. Databases were searched for reviews, original articles, and meta-analyses. Articles in English or French published in the last 30 years between December 1992 and December 2022 were included in our study. For this review, relevant information was extracted from 57 articles. The macroscopic and microscopic evaluation of the ovary was performed in the Research Center for Microscopic Morphology and Immunology, University of Medicine and Pharmacy of Craiova, Romania. The excised ovarian structures were fixed in 10% neutral buffered formalin and processed for paraffin embedding. A HMB350 microtome equipped with a section transfer to water bath system (STS, microM) performed the serial section of the block at 4–5 μm. The sections were stained classically, using Hematoxylin–Eosin (HE) and immunohistochemically, using anti-cytokeratin (CK)7, anti-CK20, anti-estrogen receptor alpha (ERα), anti-progesterone receptor (PR), anti-tumor protein 53 (p53), anti-cell proliferation factor 67 (Ki67), and anti-B-cell lymphoma-2 (Bcl-2) antibodies (Table 1 ). Immunohistochemical panel of antibodies Antibody Manufacturer Clone Antigenic exposure Secondary antibody Dilution Labeling Anti-CK7 Dako OV-TL 12/30 Citrate, pH 6 Monoclonal mouse anti-human CK7 1:50 Covering epithelium of the reproductive tract Anti-CK20 Dako Ks20.8 Citrate, pH 6 Monoclonal mouse anti-human CK20 1:50 Glandular digestive epithelium Anti-ER Dako 1D5 EDTA, pH 9 Monoclonal mouse anti-human ERα 1:50 ERs Anti-PR Dako PgR 636 EDTA, pH 9 Monoclonal mouse anti-human PR 1:50 PRs Anti-p53 Dako DO-7 EDTA, pH 9 Monoclonal mouse anti-human p53 protein 1:50 p53 Anti-Ki67 Dako MIB-1 EDTA, pH 9 Monoclonal mouse anti-human Ki67 1:50 Cell proliferation factor Anti-Bcl-2 Dako 124 EDTA, pH 9 Monoclonal mouse anti-human Bcl-2 oncoprotein 1:50 Oncoprotein Bcl-2: B-cell lymphoma-2; CK: Cytokeratin; EDTA: Ethylenediaminetetraacetic acid; ER: Estrogen receptor; Ki67: Cell proliferation factor; p53: Tumor protein 53; PR: Progesterone receptor. The slides were deparaffinized, rehydrated with successive alcohol baths with decreasing concentration of 100%, 96%, 90%, 70% (5 minutes each) and then with distilled water (dH2O) for 15 minutes. After the rehydration with dH2O, the nuclei were labeled with Hematoxylin and the cytoplasm of the cells with Eosin. After staining, the slides were fixed with Canada balsam. The immunohistochemical (IHC) procedures involved antigenic exposure which was performed using Ethylenediaminetetraacetic acid (EDTA) pH 9 or with citrate pH 6. This technique was followed by endogenous peroxidase deactivation with the aid of 3% oxygenated water (H2O2) (30 minutes), non-specific endogenous situses blockage with skim milk (30 minutes). Following these procedures, the primary and secondary antibodies were applied. After the primary antibody, the slides were kept at 4°C (for 18 hours). The following day, secondary antibody was applied [mouse/rabbit immunoglobulin G (IgG) antibody, VC002-025, R&D Systems, VisUCyte Horseradish peroxidase (HRP) Polymer] (one hour). The aid of 3,3’-Diaminobenzidine (DAB) (Dako) allowed the development of the slides, and the nuclei were labeled with Hematoxylin solution. As a final procedure, the slides were dehydrated with increasing concentration of alcohol 70%, 90%, 96%, 100% (5 minutes each), clarified in three successful xylene baths (45 minutes). Moreover, a slide was fixed on the tissue using Canada balm. The stained slides on which we could identify borderline tumor structures were scanned with a Motic Easy Scan device (Motic China Group Co., Ltd., Xiamen, China), at a 20× lens, and saved in proprietary format in a database in the Motic Digital Slide Assistant package. They were then exported as *.tiff files at full resolution for further processing.