Abstract
Aim: The aim of the present study was to evaluate the characteristics of borderline ovarian tumors (BOTs). Patients and Methods: Data of 151 patients with BOTs were retrospectively evaluated. Results: A total of 151 cases with BOTs were diagnosed. Histopathological evaluation identified 82.8% with serous, 10.6% with mucinous and 5.3% with mixed histology. Overall, 67.5% had International Federation of Gynecology and Obstetrics (FIGO) stage I, 10.6% FIGO stage II, 14.6% FIGO stage III and 4% FIGO stage IV. A total of 21.9% had peritoneal implants; of which 2.7% were invasive, 17.2% non-invasive and 2% both invasive and non-invasive. Microinvasion was observed in 5.3% and a micropapillary pattern in 12.6%. A total of 12.6% of patients presented second neoplasms. During a median follow-up period of 86 (range=0.1-432) months, there were relapses in 16.8%, of which 52.6% had invasive implants. Overall, 6.2% died of their disease, 28.5% with invasive implants. The median time-to-progression was 48 (range=8-120) months. Conclusion: Patients with BOTs have an excellent prognosis. Long-term follow-up is recommended, since recurrence occurs.
Borderline tumors of the ovary (BOTs) represent a specific group of epithelial ovarian malignancies with a good prognosis. Carl Abel described in 1901 papillary ovarian cystadenomas “on the border line between benign and malignant growths” (1). Taylor characterized them in 1929 as “semimalignant tumors” (2) and the International Federation of Gynecology and Obstetrics (FIGO) defined them in 1971 as cystadenomas with epithelial proliferation with the absence of stromal invasion omitting any word which included the term malignant in order to set a clear differentiation from invasive lesions of the ovary (3). BOTs account for approximately 8-15% of all ovarian neoplasms (4), having an incidence of 1.8-4.8/100,000 per year (5-7). They are more frequently seen in pre-menopausal women, a fact that emphasizes the importance of fertility-sparing surgery in patients who want to preserve their childbearing potential (3, 4, 8).
BOTs are associated with a significantly more favorable prognosis than epithelial ovarian cancer with excellent 5-, 10-, 15- and 20-year survival rates as high as: 97%, 95%, 92% and 89%, respectively (9). Nevertheless, they cannot be exempted from a long-term risk of a borderline or even less frequently an invasive relapse (10).
The aim of the present retrospective study was to evaluate and access clinical characteristics and modalities of treatment of BOTs and to evaluate long-term follow-up.
Patients and Methods
A hundred and fifty-one consecutive patients with primary BOTs, identified between 1973 and 2008 at the Department of Gynecology and Obstetrics (Virchow Campus) (50 patients) and the Department of Gynecology and Gynecologic Oncology (Benjamin Franklin Campus) (101 patients) of the Charité University Hospital, Berlin, were retrospectively reviewed for clinical characteristics, surgical procedures, histopathological subtypes, prognosis and follow-up. In one case, the BOT was discovered at the autopsy of a patient who died of sepsis after an operation for a total endoprothesis of a knee. Survival data of the patients were last updated 01/2010 based on patient's files, interviews and/or responses from their physicians. Surgical data were extracted from operations records and histopathological reports. Follow-up information was available for 113 cases. The remaining 38 patients were lost to follow-up.
Tumors were graded, staged and classified by the Department of Pathology according to WHO and FIGO criteria (11, 12). The diagnosis of borderline malignancy was based on an examination of the primary tumor. The histological criteria used to define a BOT were those described by Serov et al. (13): ovarian tumors histologically distinguished from ovarian carcinoma by the absence of destructive infiltrative growth or stromal invasion.
The following parameters were registered for each patient: age at primary diagnosis, menopause state, gravity and parity status, histology (histological subtypes, microinvasion, micropapillary architecture and implants), stage by FIGO, surgical procedure performed (laparoscopy vs. laparotomy, hysterectomy, unilateral or bilateral adnexectomy, omentectomy, appendectomy, lymphadenectomy, fresh frozen examination) and information about adjuvant therapy.
Eighteen cases with secondary malignancies and five with invasive ovarian cancer in the contralateral or the same ovary with the BOT were excluded from the statistical evaluation.
Statistical analysis. Categorical variables were evaluated using the chi-square test. Survival curves were constructed according to the Kaplan–Meier method and statistical differences between the curves were calculated with the log-rank test. Multivariate survival analysis was performed with the Cox proportional hazard model. A p-value of less than 0.05 was considered statistically significant. All statistical analyses were performed using the Statistical Package for Social Sciences software, version 11.5 (SPSS, Chicago, IL, USA).
Results
Patients' characteristics. A total of 151 BOT cases were identified. Patients and disease-related characteristics are presented in Table I. The mean age at primary diagnosis was 51.1 years (range=16-85 years); 35.7% (n=54) of women were in the reproductive age group of 16-45 years, 45% (n=68) were pre-menopausal and 55% (n=83) post-menopausal. Histology was known in 149 cases and revealed serous BOT in the majority of women (n=125). Information about the FIGO stage was available in 131 cases of serous and serous/mucinous tumors and in 15 cases of mucinous tumors. Of the serous and serous/mucinous group, most (64.4%) patients were diagnosed with disease at FIGO stage I, and the 15 patients with mucinous tumors all had FIGO stage I disease. Implants were found in 33 (21.9%) cases. They were localized at the omentum, tubes, peritoneum, parametria, pouch of douglas, uterus, cervix, lymph nodes, skin, sigmoid, rectum and appendix vermiformis.
A pelvic lymphadenectomy was performed in 44 cases and a pelvic and para-aortic lymphadenectomy in 26 cases. Lymph nodes were involved in 14 cases. Half of patients who had a mucinous or a serous/mucinous tumor undercut an appendectomy. The appendix vermiformis was affected by the BOT in one case, in two cases there was endosalpingiosis and in one further case, the appendix was the origin of a pseudomyxoma peritonei with ovarian metastases.
A detailed report about the surgical procedures performed is presented in Table I.
Demographics, pathology and surgery data of patients with borderline ovarian tumor (BOT) (n=151).
Clinical and pathological features of the seven patients with borderline ovarian tumor who died of disease.
Secondary malignancies were seen in 18 cases: 9 with breast cancer, 3 with colorectal cancer, 1 with breast and colorectal cancer, and plasmocytoma, bronchial, endometrial, stomach and cervical cancer each in one case. In 5 cases, a co-existence of invasive epithelial cancer and BOT was present in the same or contralateral ovary. These patients were then excluded from the present analysis.
Follow-up. Follow-up information was available for 113 patients (74.8%). The median follow-up time was 86 months (range=0.1-432 months). A follow-up time of at least five years was available for 72 (63.7%) patients, while 39 (34.5%) patients were followed-up for at least 10 years, and 20 (17.7%) for at least 15 years.
The outcomes were as follows: 81 (71.7%) patients had no evidence of disease, 10 (8.9%) patients were alive with disease, 14 (12.3%) had died of causes not related to BOT, 7 (6.2%) of disease and in one case (0.9%) the BOT was incidentally discovered during the autopsy of a patient who died of sepsis after a total endoprothesis of a knee. The characteristics of the 7 patients who died of disease are shown in Table II.
Survival analysis. In the univariate analysis, higher FIGO stage (>II) (p=0.02) was significantly related to a poorer overall survival rate using Kaplan–Meier survival estimation curves with log-rank comparisons. Histology (serous, mucinous, serous/mucinous) and implants (invasive, non-invasive, invasive and non-invasive) had no influence on the survival (Figure 1). None of the examined parameters, however, retained prognostic significance for survival in the multivariate analysis.
Relapse. Relapse was observed in 19 (16.8%) patients. The median disease-free-survival was 48 months (range=8-120 months). Six (31.6%) patients had a FIGO stage I disease, one (5.25%) stage II, five (26.3%) stage III and six (31.6%) stage IV. The histological subtypes were serous in 16 (84.2%) cases and mucinous in 2 (10.55%) cases. The morphology of the relapse was as follows: invasive implants in eight (42.1%) cases, non-invasive implants in seven (36.8%) cases, invasive and non-invasive implants in two (10.55%) cases and BOT in the contralateral ovary in further two (10.55%) cases. Micropapillary architecture was observed in three (15.8%) cases. Two patients with initial invasive implants and only one patient with initial non-invasive implants experienced relapse. One of these patients (28 years old) also developed an invasive ovarian carcinoma 168 months after the initial diagnosis of a BOT in the contralateral ovary.
The follow-up outcome of those 19 patients was: 10 were alive at last contact, six had died from their disease and three had died of causes unrelated to BOT.
Discussion
BOTs are associated with a favorable prognosis. While in general they initially present as a benign entity, they may later manifest as either a new, more aggressive primary lesion, or undergo malignant transformation (14, 15).
For surgical therapy of BOTs, like that for invasive ovarian cancer, adequate staging and maximal surgical effort aiming at optimal cytoreduction are the two most relevant cornerstones (12, 16). In younger patients, who wish to preserve fertility, a fertility-sparing procedure may be considered in cases of early stage and unilateral ovarian involvement. Numerous authors have reported the safety of fertility-preserving surgery in patients with BOT, although in general, the ovarian reserve in these patients may be impaired and so may negatively affect conception potential. However, the patients need to be informed about a higher relapse rate in cases of intraoperative cyst rupture without removal of the affected ovary (17-19). For this reason, fertility-preserving surgery should only be performed after a detailed informing of the patient with the emphasis on undergoing a careful and prolonged follow-up (16).
Many authors have described an increased risk for developing second neoplasms, particularly breast and colorectal cancer, for women with BOTs (20-22). Those may have been diagnosed in the past or after the BOTs. The most common secondary malignancies in our study group were breast (6.6%) and colorectal cancer (2.6%). These incidences are ten-fold those described in the general population. Although Gotlieb et al. describe a lower incidence of Breast Cancer Gene 1 (BRCA1) and BRCA2 mutations in patients with BOTs than in patients with invasive epithelial cancer (23), a common pathway should be considered. The high prevalence of breast cancer in the history of patients with BOTs especially suggests the need for regular screening in this patient group.
Kaplan–Meier survival curves related to stage by FIGO (a), histology (b); presence of implants (c). Statistical differences were calculated by log-rank comparisons.
Microinvasion and micropapillary architecture were not significant risk factors for overall survival in our study group. This was also the case for those with implants, although many authors describe them as the only risk factors for relapse or death from disease (24-26).
The high percentage of FIGO stage I both in serous and mucinous tumors (64.4% and 93.8%) in our study group is in concordance with the literature (27-30). On the other hand, we more frequently observed higher FIGO stages in those with serous tumors than previously described by other authors (stage III: 16.3% stage IV: 4.5% vs. 7.9% and stage IV: 0%, respectively) (31). This could be an effect of the selection bias caused by the treatment at a large University Hospital, reference center for ovarian malignancies. Higher FIGO stage was identified also in our analysis as a negative predictor for overall and disease-free survival, congruent to hitherto reported experience (29, 32-34).
Despite the fact that all our patients with mucinous BOTs were alive during the entire follow-up, the mucinous histological subtype, mostly in association with invasive implants, has been reported to be associated with a worse prognosis in comparison with serous BOTs (8, 21, 27, 35, 36). Moreover, mucinous BOTs may constitute metastasis from a primary mucinous bowel adencarcinoma (37), as was the case in one of our patients, and be falsely misdiagnosed.
The recurrence of BOT in the form of an invasive disease has been reported by numerous authors (21, 22). If a relapse has the morphology of a BOT, the 5-year survival rate will be approximately 81%. However, if the transformation leads to an invasive carcinoma, the 5-year survival rate is reduced to 68% (31). The only patient in our study group who experienced relapse with an invasive carcinoma in the contralateral ovary 168 months after the primary diagnosis of a BOT was alive and was under treatment 32 months after relapse.
In our series lymph nodes were tumor-involved in 14 out of the 44 cases (31.8%), where a lymphadenectomy was performed. These data are in accordance with other studies (8, 38). In a variety of studies, no significant prognostic relevance was identified for affected lymph nodes in borderline disease (26, 38, 39). Therefore systematic lymphadenectomy in the absence of bulky nodes is not indicated for these patients.
We recorded higher relapse rates and a longer disease-free-survival (16.8% in a median time of 48 months) than those described in the literature (approximately 11% in 30 months) (15, 18, 31, 40). Cusidó et al. state that invasive implants and microinvasion are the only independent risk factors for development of relapse (31). This is not the case in our study group, where the multivariate analysis showed no independent negative risk factors for overall survival.
The long follow-up (median=86 months; mean=99 months) and the large number of patients are the strengths of our study and allow us to make important conclusions about this specific tumor entity. All patients were treated in two centers of the same large institution, with common standards and experts, in this way avoiding the disadvantages of a multicenter study, where non-homogenous surgical and pathological/anatomical quality would possibly constitute a negative bias for our observations. However, the retrospective character of the present study has to be recognized as a considerable limitation. Therefore, prospective evaluations with systematic assessment of all disease-related characteristics are warranted for a better understanding and a conclusive evaluation of the biological behavior of those enigmatic tumors.
Conclusion
The results of our study underline the good prognosis of patients with BOT, although relapse and transformation into invasive disease may occur even after a long period of time. The treatment of BOTs should be oriented on individual criteria of the patient and tumor. Prolonged follow-up is required to detect patients at risk of relapse and to select appropriate treatment for them.
- Received March 26, 2014.
- Revision received June 22, 2014.
- Accepted June 23, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved