Characterization and Management of Borderline Ovarian Tumors – Results of a Retrospective, Single-center Study of Patients Treated at the Department of Gynecology and Obstetrics of the University Medicine Greifswald

Discussion

Various studies reported the occurrence of BOT as between 28 and 62 years and therefore, at an average of 10 years earlier compared to the ovarian cancer (14-21). Our data are comparable, with an average age of 54.6 (range=23-83) years. In line with our study, where 69% of 54 patients were found to be postmenopausal at the time of diagnosis, in previous investigations, 60%, 70% and 65% of women were postmenopausal at time of diagnosis (18, 22, 23). In contrast, studies of Desfeux et al. (20), Ji et al. (24) and Pirimoglu et al. (25) revealed that only 37%, 41% and 28% of women with BOT were postmenopausal. In our collective, the average age of menopause was 48.0 years (range=33-38). Four patients (7.4%) had undergone a hysterectomy when premenopausal. Furthermore, patients with BOT had an average age of menarche of 13.5 (range=11-19) years, which is comparable to several authors (13.0 to 14.6 years) (15, 19, 20, 22, 26). Overall, no study showed any influence of the time point of menarche or menopause for the risk of subsequent BOT.

Riman et al. identified a risk-reductive effect of childbirth for both serous and mucinous BOT when comparing 193 patients to a control cohort (19). On average, patients with BOT had 1.8 births and 0.2 abortions. On average 17.1% of women were nulliparous. Other studies also showed that patients with BOT had fewer childbirths. as Ji et al. reported 1.9 (range=0-8) pregnancies and 1.6 (range=0-7) childbirths (20), and Ayhan et al. reported 2.1 (range=0-13) pregnancies and 1.9 (range=0-11) childbirths (22). In this study, 31% of patients with BOT were nulliparous. Moreover, Nikrui et al. (23) and Eltabbakh et al. (27) showed that 39% and 22% of patients with BOT were nulligravida. A meta-analysis of nine studies by Harris et al. (28) showed that parity was inversely proportional to the risk of developing BOT and that an increased number of childbirths was associated with a further risk reduction. Interestingly, infertile women had an increased risk for BOT development compared to controls without infertility in medical history. Concerning the number of pregnancies and childbirths, the patient collective in the present study did not substantially differ from those mentioned above and therefore our date strongly support the results of previous studies. Further research has been carried out in terms of risk reduction of BOT due to breast feeding and intake of oral contracepives. Studies by Harris et al. (28), Bernard et al. (29) and Riman et al. (19) showed that breast feeding reduces the risk of BOT development, whereas a study of Jordan et al. (30) detected no risk reduction. Very heterogenous results are associated with the intake of oral contraceptives. Harris et al. did not find any correlation between oral contraceptive use and the risk of BOT development (28), whereas Bernard et al., interestingly, detected a risk reduction (29) and Jordan et al. (30) showed a risk reduction of 50%. In contrast, estrogen substitution for hormone replacement therapy increased the risk of BOT development in the study of Riman et al. (19).

This study aimed to determine a possible relation between tumor diseases in family history and the risk of developing a BOT. Surprisingly, 28 out of 54 patients (51.9%) reported tumor events in their family history, 3.7% had events solely of breast cancer, whereas 9.3% had cases of breast cancer and another tumor type in their family; 1.9% of patients had solely colon carcinoma in their family history, 9.3% reported colon carcinoma and another tumor type. In a study by Nikrui et al., 13% of patients had a positive tumor anamnesis in their family (23). Riman et al. determined 8% of patients had cases of breast cancer in their family (19). Furthermore, Gotlieb et al. showed that 4.3% of their collective of 117 patients with BOT were positive for BRCA1/2 mutation compared to 24.2% among patients with invasive ovarian cancer (31). Overall, 8.2% of patients with BOT had cases of breast cancer in their family history. Various other studies did not identify a correlation between positive tumor family history and the risk of BOT development (27, 30, 32).

Interestingly, this study found that 22.2% of the patients with BOT were diagnosed with other tumor diseases, among these the most frequent secondary tumor was breast cancer (11.1%). Nikrui et al. also found that 16.2% of patients developed a secondary tumor when diagnosed with BOT (23). A study by Bouchardy et al. found that 12% of patients with BOT developed a secondary tumor, most frequently including breast cancer and colorectal cancer (14). The authors postulated a 3.97-fold increased risk for colorectal cancer development. The risk for breast cancer was assumed to be twice as high compared to patients without BOT.

In this study, 32.7% of patients with BOT were asymptomatic. When suffering from symptoms, pain of the upper and lower abdomen (40.7%), alterations of body weight (7%) and girth (7%) were the most frequent events. These findings are in line with other studies, as Ji et al. found 69.5% of patients with BOT were symptomatic (20). Nikrui et al. (23) and Cusido et al. (33) found 14.0% and 30.4% of patients, respectively, were asymptomatic. However, abdominal pain and alterations of body weight and girth were the leading symptoms.

Beside the mostly unspecific symptoms of BOT, serum tumor markers such as CA125 and CA19-9 play an important role in BOT diagnostics. Among the patient collective in this study, the tumor markers CA125 and CA19-9 were preoperatively determined in 75.9% and 59.3% of cases, with an average of 79.4 U/ml and 1,690.7 U/ml, respectively. CA125 was increased in 34.2%, whereas CA19-9 was greater than the reference level in 31.3% of cases. Our results are in line with those of Wong et al. who found CA125 to be increased in 39% of his collective. Among 234 patients included in a study by Ren et al., CA125 and CA19-9 were increased in 62% and 11% of cases, respectively (17). CA125 thereby correlated with advanced-stage BOT, however, it was not an independent factor for the probability of tumor relapse. Pirimoglu et al. found an increased CA125 level in 57.8% (25), whereas Engelen et al. reported that 24% of patients with BOT had a CA125 level above the reference value and an increased CA19-9 level in 46% (34). The authors suggested CA19-9 to be the more reliable tumor marker and recommended its determination during the follow-up of Bo's. Overall, however, the determination of serum tumor markers seems to increase early detection of tumor recurrence in only a minority of cases.

Concerning the histological differentiation of BOT, there were no major differences between this study and the literature. Most recent studies report rates of between 55% and 66% for serous tumors, and 34% and 45% for mucinous tumors (14, 15, 19, 35, 36), both being the most frequent histological differentiation of BOT. Some studies with other patient collectives found mucinous BOT to be more abundant (27, 34, 37). Ji et al. (20) and Donnez et al. (38) showed a nearly balanced distribution of serous (48% and 49%) and mucinous (47% and 48%) BOT. In this study, 25.8% of serous tumors and 8.7% of mucinous tumors were found to be bilateral. Except for the studies of Hung-Yaw et al. (27) and Ren et al. (17), this is in line with most of the recent research in this field (20, 22, 34, 35, 37, 39). Furthermore, serous tumors appeared to be greater in size compared to mucinous tumors (20, 27, 36, 40-42), which is validated by this study.

Concerning the actual German guidelines for diagnostics and therapy of malignant tumors of the ovary (13), the performance of a frozen section is not imperative. Among the collective of this study, an intraoperative frozen section was performed in 66.7% of patients. Of these, in 88.9%, the diagnosis of the frozen section was concordant with the definitive histological result. This success rate is outstanding compared to the results of other studies in which it ranged between 60% and 78.4% (22, 27, 40, 43, 44). However, convincing studies on the definitive benefit of the performance of an intraoperative frozen section are still lacking.

Apart from this, the performance of peritoneal lavage and sample taking for cytology is recommended by international therapy guidelines and is part of adequate surgical staging (2, 7, 13, 45). In this study, peritoneal lavage followed by cytology was carried out in 72.2% of cases; among these, tumor cells were found in 7.4%, and in 3.2%, tumor cells could not be excluded. In spite of the fact that the performance of cytology is a simple and an effective method to diagnose peritoneal extension of BOT, only few studies describe the application of this technique (15, 46). In literature, the rate of cytology following peritoneal lavage was described as being between 21.1% and 72.2% (16, 27). In total, 4.4% to 80.0% of these cytology samples tested positively for tumor cells.

There is no evidence that adjuvant chemo- or radiotherapy for BOT is associated with increased overall survival (13). Therefore, the only recommended therapy of BOT is radical surgery, including careful abdominal exploration, total hysterectomy, bilateral salpingo-oophorectomy, omentectomy, excision of all conspicuous lesions, performance of various peritoneal biopsies, and peritoneal lavage with subsequent cytology (12-13). According to actual therapy guidelines, conservative fertility-sparing surgery may be discussed with patients under 40 years of age and with incomplete childbearing (12, 13). After informed consent about the fourfold increased risk of tumor recurrence, this includes unilateral salpingo-oophorectomy or unilateral cystectomy, abdominal exploration, peritoneal lavage and cytology, omentectomy, excision of suspicious lesions and peritoneal biopsies (2, 13, 47). In fact, in this study, only eight patients (14.8%) primarily underwent surgical treatment according to therapy guidelines. Five more patients (9.3%) completed the treatment according to therapy guidelines after a secondary surgery. Retrospectively, only one (1.9%) out of nine (16.7%) patients treated by fertility-sparing surgery had a serious wish for childbearing. Furthermore, eight patients (14.8%) received adjuvant chemotherapy. The reasons why these treatment modes, divergent from actual therapy guidelines, were chosen between 1990 and 2010 is retrospectively difficult to comprehend.

During a mean follow-up of 70.3 months (median=40.5, range=0-231 months), 14 patients (25.9%) died, with a mean time to death of 225 months. None of these deaths was tumor associated. In this study, two patients (3.7%) experienced recurrence of disease 9 and 29 months after tumor resection. This is a very low percentage compared to literature, with stated rates of disease recurrence of up to 11.5% (16, 33-35). Among these cases, tumor recurrence was observed within the first 10 years after initial surgery. Nevertheless, the mean duration of follow-up in literature was only 21 to 105.6 months (14, 16-18, 20, 24, 33-35, 37).

In conclusion, even though only eight out of 54 patients received surgical treatment of BOT according to actual therapy guidelines between 1990 and 2010. To date only, two cases of tumor recurrence were detected.

Interestingly, the number of patients who were primarily treated according to therapy guidelines has increased since 2007. However, further studies are necessary to investigate recent therapy procedures at the Greifswald University Hospital with its associated peripheral academic teaching departments Wolgast, Demmin, Anklam and Ueckermünde.