Tangir, Jacob MD; Zelterman, Daniel PhD; Ma, Wenging; Schwartz, Peter E. MD
Ovarian germ cell tumors account for 20–25% of all ovarian neoplasms, and only about 3% of these are malignant.1 During the mid-1980s, fertility-preserving surgery (removing the affected ovary and preserving the contralateral ovary and the uterus) followed by combination chemotherapy became the standard of care for early stages and selected advanced malignant germ cell tumors of the ovary.2 Several studies have reported that this conservative approach is equally effective, in terms of survival, when compared with more radical surgery.3–5 One of the main advantages of this therapy is that patients with malignant germ cell tumors of the ovary could conserve their reproductive function after effective treatment.
To date, however, only a few studies have evaluated the reproductive outcomes of patients treated with this protocol. Furthermore, there are no data regarding the effect of the chemotherapy combinations commonly used on the physical and neurological development of children born to women previously exposed to these agents.
This report will review one of the largest series of these rare cancers. The objectives of this analysis were 1) to evaluate the fertility outcomes of women diagnosed with malignant germ cell tumors of the ovary and treated conservatively, 2) to study any relationship between fertility and disease status, mode of therapy, and type of tumor, 3) to describe the demographic characteristics of all patients with malignant germ cell tumors of the ovary seen at our institution, including those not treated conservatively, and 4) to describe preliminary data on the impact of chemotherapy on the offspring of patients treated for malignant germ cell tumors of the ovary.
MATERIALS AND METHODS
A case series analysis was performed on all patients with malignant germ cell tumors of the ovary seen or consulted on at the Yale Gynecologic Oncology Center from 1975 to 1995. Even though the main outcome to be measured is reproductive function after conservative surgery, patients not treated conservatively were also included for completeness. Only patients with complete information on diagnosis, stage, and surgical treatment were included. Initial information regarding surgery, pathology, and chemotherapy was obtained from hospital medical records for the patients seen at Yale. For patients consulted on, consulting physicians completed an initial detailed questionnaire in addition to submitting operative, pathology, and chemotherapy records. Follow-up information for patients seen at Yale was obtained by a mailed questionnaire. The consulting physician completed a similar questionnaire to provide follow-up information for patients not seen at Yale.
The follow-up questionnaire included the following categories:
1. Medical and surgical history: Previous medical problems and newly developed medical problems since the diagnosis was made; surgical history and any new surgeries in addition to the initial diagnostic or therapeutic procedure for the malignant germ cell tumor of the ovary; long-term side effects from chemotherapy.
2. Reproductive function: The questionnaire requested information regarding menstrual history before and after the treatment (for those in whom the uterus was conserved), changes in the menstrual pattern, and delays in resuming menses. It also asked about attempts to conceive, length of time trying, and whether any fertility drugs had been used, number of pregnancies, and outcome of each one (full term, premature, abortion, or ectopic pregnancy).
3. Offspring: The questionnaire also requested information regarding children born after treatment for malignant germ cell tumors of the ovary, including congenital or developmental defects, other medical problems, and performance in school.
In cases where patients did not respond to the initial questionnaire, one additional questionnaire was mailed. If neither of the two mailed questionnaires was answered, a telephone interview was attempted for patients who had been seen at our center. For patients not seen at our center, a questionnaire was mailed to the consulting physician requesting follow-up information as described above.
The statistical software packages SAS 8.0 (SAS Institute, Cary, NC) and Splus 1.0 (MathSoft, Seattle, WA) were used to calculate:
* Fisher exact test to analyze the differences in fertility outcome among different treatment modalities and menstrual changes after different combinations of chemotherapeutic agents
* logistic regression analysis to compare fertility outcomes among International Federation of Gynecology and Obstetrics (FIGO) stages and among different chemotherapeutic regimes
* binomial distribution to calculate exact confidence intervals (CIs) for selected rates
A total of 106 patients with the diagnosis of malignant germ cell tumor of the ovary who met the criteria outlined above were identified. Eleven patients died of disease, one patient committed suicide (a postmortem examination revealed no evidence of disease), and eight patients were lost to follow-up after the initial diagnostic surgery. These 20 patients were excluded from the analysis. Median follow-up for the remaining 86 patients was 122 months (range 24–384). Table 1 lists the patient characteristics of the entire group of patients. Only one patient reported the diagnosis of another malignancy, a breast cancer, occurring 20 years after the treatment of her stage I immature teratoma. She underwent a lumpectomy and radiation therapy and is currently free of disease.
Sixty-four patients underwent fertility-preserving surgery. Two additional patients underwent a unilateral salpingo-oophorectomy but had previously undergone bilateral tubal ligations. Among the 64 patients with the potential for conceiving, 26 patients had not attempted to conceive at follow-up. Thirty-eight patients attempted to conceive, 29 patients achieved at least one pregnancy spontaneously (76%; exact CI 0.62, 0.89) (Table 2). There was no significant difference between those patients treated with surgery only and those who received surgery and chemotherapy regarding fertility outcomes.
The FIGO stages and fertility outcomes are depicted in Table 3. The difference among the different stages in number of pregnancies achieved was not significant. Of note, ten patients with advanced disease were treated conservatively. Of these, eight were able to conceive (Table 4).
The primary chemotherapy regimens for the patients who attempted to conceive are listed in Table 5. Among the women who conceived, 16 received vincristine, actinomycin D, and cyclophosphamide. Nine received a platinum-based chemotherapy: Three received cisplatin, vinblastine, and bleomycin; three received bleomycin, etoposide, and cisplatin; one received etoposide and cisplatin; and two were treated with other platinum-based combinations. Four of these patients did not receive any chemotherapy. Among the nine women who could not conceive, four received vincristine, actinomycin D, and cyclophosphamide; three received etoposide and cisplatin; one received cisplatin, vinblastine, and bleomycin; and one received no chemotherapy.
There was no difference in fertility outcomes among patients treated with vincristine, actinomycin D, and cyclophosphamide versus those treated with platinum-based combinations.
We performed a logistic regression on the ability to conceive or not as a function of chemotherapy type received, age at time of diagnosis, stage of tumor, and duration of follow-up. None of these variables or their various interactions achieved statistical significance at levels more extreme than .1 in their explanatory value for ability to conceive.
Follow-up information regarding menstrual history was obtained in 40 of 64 patients (62.5%). Four patients were premenarchal at the time of diagnosis (10%). All of them went into menarche normally and have been having regular menses. Two of them were treated with vincristine, actinomycin D, and cyclophosphamide for six cycles; one was treated with cisplatin, vinblastine, and bleomycin for four cycles; and the fourth one was treated with cisplatin, vinblastine, and bleomycin and adriamycin for four cycles. This last patient is the only one thus far to have attempted to conceive, and was able to complete a full-term pregnancy.
Twenty-eight additional patients (69%) reported a return to regular menstruation during or after completion of chemotherapy, with delays that ranged from 0 to 6 months. Eight patients (20%) reported abnormal menstrual patterns after treatment (Table 6). Two of these eight patients went into premature menopause. One of them was diagnosed with a stage IC endodermal sinus tumor at age 29 and then received 18 cycles of vincristine, actinomycin D, and cyclophosphamide. She stopped menstruating during the chemotherapy treatment and went into menopuase. The second patient was diagnosed with an immature teratoma at age 17 and was treated with six cycles of vincristine, actinomycin D, and cyclophosphamide. She had regular periods for several years while on oral contraceptives (OCs). After the pill was stopped her menstruation became lighter and stopped permanently at age 35. Elevated follicle-stimulating hormone levels confirmed menopause.
Three of the remaining six patients who reported irregularities with their menstruation reported persistent irregular bleeding. Among the other three, one was amenorrheic for 1 year after completion of six cycles of cisplatin, vinblastine, and bleomycin, and the remaining two received five and six cycles of etoposide and cisplatin, respectively. One of them reported polymenorrhea after chemotherapy, and the other patient reported increased dysmenorrhea after the treatment.
Among the eight patients who reported menstruation irregularities, six have attempted to conceive and only two have been successful. In contrast, among the 32 patients who reported normalization of menstruation, 19 have attempted to conceive and 16 have been successful. This difference was statistically significant (exact P = .032)
The 29 patients who conceived had a total of 47 pregnancies. There were 38 children born (including three sets of twins), ten elective terminations, and two spontaneous abortions. Eleven patients had one child, eight had two children, two had three children, and one had five children. Follow-up was available for 16 of these children. One child has “speech problems.” Another child, whose mother was treated with chemotherapy during the third trimester of pregnancy, had juvenile arthritis. This, however, is a prevalent disease in this family. One of a patient's twins has problems processing high-order thinking and is currently undergoing further workup. The other 13 children were reported as completely normal at the time of follow-up.
Fertility-preserving treatment for malignant germ cell tumors of the ovary, even in advanced stages, allowed these young women to conceive and have children. Twenty-nine patients of 38 who attempted to conceive were successful (76%). This is one of the largest series reported for this relatively rare disease. These results are comparable to previously published data.
Low et al6 reported on 74 patients with malignant germ cell tumors of the ovary who underwent conservative surgery, 47 of whom (64%) received adjuvant chemotherapy. Of these, 20 attempted conception and 19 were successful (95%). The reproductive outcomes of patients treated with surgery who did not receive chemotherapy, however, were not reported. It is not clear how many of these 20 patients were of advanced stage. Zanetta et al7 reported on 81 patients who were treated conservatively and received adjuvant chemotherapy. Twenty patients attempted to conceive and 16 were successful (80%), compared with 12 of 12 in the group not treated with chemotherapy. There were only six patients with advanced disease in this series. It is not reported, however, how many of the advanced-stage patients were able to conceive. Gershenson8 reported a series of 40 patients with malignant germ cell tumors of the ovary, 16 of whom had attempted to conceive. Nine conceived naturally, and three more did after infertility treatment (75%). Ten patients in this series had stage III disease. It was not stated how many of these ten advanced-stage patients attempted or were able to conceive.
Brewer et al9 reported a series of 16 patients with dysgerminoma who underwent conservative treatment. Three of the 16 who attempted to conceive were successful. All of them had been treated with bleomycin, etoposide, and cisplatin.
An important novel aspect of our study is the finding that eight of ten women with stage III disease successfully treated with fertility-preserving surgery were able to conceive (Table 5).
About the untoward side effect of chemotherapeutic agents, alkylating agents such as cyclophosphamide have been associated with gonadal failure in women.10 In addition, the cumulative dose of chemotherapy administered is another important factor. In the earlier stages of our series, most patients were treated with the vincristine, actinomycin D, and cyclophosphamide combination that includes cyclophosphamide.
Among the 29 patients who conceived, 16 received vincristine, actinomycin D, and cyclophosphamide (55%); four required no chemotherapy; and the remaining nine received other combinations without cyclophosphamide. Among the nine patients who could not achieve pregnancy, four received vincristine, actinomycin D, and cyclophosphamide (44%); one required no other therapy; and the remaining four received other combinations lacking cyclophosphamide.
These results suggest that cyclophosphamide did not have a significant impact in the fertility of these patients and that even with the use of vincristine, actinomycin D, and cyclophosphamide a substantial proportion of women were able to conceive. The use of this effective combination chemotherapy did not decrease the chances of conception in those who had fertility-preserving surgery.
Premature menopause has also been associated with chemotherapy in young women, especially with alkylating agents. Byrne et al11 showed that treatment with alkylating agents carried a relative risk of premature menopause of 9.3. Results in the literature, however, have not been consistent. In one of the largest series recently reported, Chiarelli et al12 studied the risk of early menopause and infertility after treatment for childhood cancer in Ontario. In a series of 719 patients they found no increased risk for either early menopause or infertility in women treated with combinations of several alkylating chemotherapeutic agents including cyclophosphamide. Unfortunately, in this study they also included patients who received carboplatin and cisplatin, described as “alkylating agents.” These two drugs are not alkylating agents, and they are not known to affect gonadal function.13 The grouping of drugs from different categories might have masked the potential adverse effect of alkylating agents like cyclophosphamide.
In our series, however, only two of the 64 patients available for analysis reported premature menopause (3%). Both were treated with vincristine, actinomycin D, and cyclophosphamide. One received 18 cycles and the other received six. One limitation of our analysis and of other similar reports is that most of our patients are still at a young age and long-term reproductive outcome cannot be completely evaluated.
Interestingly, in the study by Chiarelli as well as in others, it was found that women treated before puberty have a reduced risk of developing gonadal damage relative to those treated after menarche.12,14 In one report from The M.D. Anderson Cancer Center, patients who developed menstrual dysfunction after completion of chemotherapy were significantly older at diagnosis than those who had normal menses (mean ages 17.7 and 13.6, respectively).5 Our results are consistent with these findings. Four of the patients in our series were premenarchal when they were treated, and all of them reported regular menses at the time of follow-up. Furthermore, the only one who has attempted pregnancy has conceived without complications. Premenarchal ovaries may be more resistant to the toxicity from the chemotherapy than postmenarchal ovaries because of the relatively larger amount of oocytes in reserve.13 This information is extremely important to parents of young girls diagnosed with malignant germ cell tumors of the ovary who are concerned about their daughters' future reproductive function.
Menstrual history was obtained in 40 of the 64 patients who experienced fertility-preserving surgery (63%). The majority of the patients treated with surgery and chemotherapy resumed normal menses within 6 months of the completion of the treatment. Eighty percent of the patients for whom information was available reported return to normal menses. From this group, 53% received alkylating agents. Interestingly, more patients who received platinum-based chemotherapy developed menstrual irregularities than those who received alkylating agents (15% versus 37.5%). In a study describing gonadal dysfunction due to cisplatin, 57% of the patients reported amenorrhea or irregular menstruation after treatment with platinum-based chemotherapy.15 In a more recent report of patients with dysgerminoma treated with bleomycin, etoposide, and cisplatin, 93% of the patients returned to their prechemotherapy menstrual pattern. These conflicting data regarding the effect of platinum-based chemotherapy in menstrual function warrant further investigation.
The majority of the patients were placed on OCs during chemotherapy treatment. The authors routinely recommend this practice to avoid any possible pregnancy during treatment with cytotoxic agents. Unfortunately, data were not available regarding how many patients remained on OCs after treatment, and how this may have been correlated with reproductive function and menstrual changes.
Another legitimate concern of these young women, in addition to the future reproductive function after the treatment, is what, if any, effect may occur in their offspring after they were treated with toxic agents at a young age. Green et al16 studied the children of 306 men and women who had been treated with chemotherapy during childhood. That report concluded that treatment with mutagenic chemotherapeutic agents did not significantly increase the frequency of congenital anomalies in the offspring. However, there was a small increase of structural congenital heart disease in the children of women treated with actinomycin D. Our study suggests that chemotherapy was not responsible for any major birth defects in this population.
This report has limitations. Because of the rare nature of these tumors, it is difficult to study large numbers of patients. The number of patients in our study is relatively small. Therefore, achieving significant differences in statistical analyses is unlikely. It is reassuring, however, that our results are consistent with those of previous reports for fertility after chemotherapy.
Another limitation is the retrospective nature of the analysis. The patients included were seen over a 20-year period, and treatment and outcomes tend to change and improve over time. This may have influenced our results. In the case of treatment for malignant germ cell tumors of the ovary, two major changes have occurred since the 1970s that have favorably impacted the outcomes: the shift from radiation therapy to chemotherapy-based treatment, and later the change from cyclophosphamide-based chemotherapy to platinum-based chemotherapy. Our series includes only patients treated with chemotherapy, and the shift from cyclophosphamide-based to platinum-based chemotherapy has been analyzed and found not to affect the main outcome measured in this study, which is reproductive function.
Other limitations inherent to retrospective studies include the potential introduction of selection bias when analyzing the effect of treatment modalities on patients who were not randomized, as well as the potential for recall bias when collecting information via questionnaires. It is highly unlikely, however, that a randomized trial could ever be performed to answer these and other questions in a more rigorous way. Our study results, however, warrant further confirmation pooling larger numbers of patients from multiple centers and comparing them with the fertility outcomes of normal matched controls.
The data presented suggest that fertility-preserving surgery followed by combination chemotherapy may not impede normal reproductive and menstrual function in women diagnosed with malignant germ cell tumors of the ovary. To date, children born to these patients have not experienced an increased risk for congenital malformations or developmental abnormalities. Fertility-preserving surgery followed by appropriate chemotherapy should remain the standard of care even for women with malignant germ cell tumors of the ovary in advanced stages. Patients with this disease and their family members should be reassured about the high chance of retaining the patient's ability to conceive and have normal children after conservative surgery and combination chemotherapy.
1. Disaia P, Creasman W. Germ cell, stromal and other ovarian tumors. In: Disaia P, Creasman W, eds. Clinical gynecologic oncology, St. Louis: Mosby-Year Book, 1997: 351–74.
2. Schwartz P. Combination chemotherapy in the management of ovarian germ cell malignancies. Obstet Gynecol 1984;64:564–72.
3. Kurman R, Norris H. Malignant germ cell tumors of the ovary. Hum Pathol 1977;8:551–64.
4. Schwartz P, Chambers S, Chambers J, Kohorn E, McIntosh S. Ovarian germ cell malignancies: The Yale University experience. Gynecol Oncol 1992;45:26–31.
5. Gershenson D. Management of early ovarian cancer: Germ cell and sex cord-stromal tumors. Gynecol Oncol 1994;55:S62–72.
6. Low JJH, Lewis CP, Crandon AJ, Hacker NF. Conservative surgery to preserve ovarian function in patients with malignant ovarian germ cell tumors. Cancer 2000;89:391–8.
7. Zanetta G, Bonazzi C, Cantu MG, Binidagger S, Locatelli A, Bratina G, et al. Survival and reproductive function after treatment of malignant germ cell ovarian tumors. J Clin Oncol 2001;19:1015–20.
8. Gershenson D. Menstrual and reproductive function after treatment with combination chemotherapy for malignant ovarian germ cell tumors. J Clin Oncol 1988;6:270–5.
9. Brewer M, Gershenson D, Herzog C, Mitchell M, Silva E, Wharton J. Outcome and reproductive function after chemotherapy for ovarian dysgerminoma. J Clin Oncol 1999; 17:2670–5.
10. Meistrich M, Vassilopoulou-Sellin R, Lipshultz L. Gonadal dysfunction. In: Rosemberg S, ed. Cancer, principles and practice of oncology. Philadelphia: Lippincott-Raven, 1997:2758–73.
11. Byrne J, Fears TR, Gail MH, Pee D, Connelly RR, Austin DF, et al. Early menopause in long-term survivors of cancer during adolescence. Am J Obstet Gynecol 1992; 166:788–93.
12. Chiarelli A, Marrett L, Darlington G. Early menopause and infertility in females after treatment for childhood cancer diagnosed in 1964–1988 in Ontario, Canada. Am J Epidemiol 1999;150:245–54.
13. Meistrich M, Vassilopoulou-Sellin R, Lipshultz L. Gonadal dysfunction. In: DeVita V, Hellman S, Rosemberg S, eds. Cancer, principles and practice of oncology. Philadelphia: Lippincott Williams & Wilkins, 2001: 2923–39.
14. Chapman R, Sutcliffe S, Malpas J. Cytotoxic-induced ovarian failure in women with Hodgkin's disease, I, hormone function. JAMA 1979;242:1877–81.
15. Wallace W, Shalet S, Crowne E, Morris-Jones P, Gattamaneni H, Price D. Gonadal dysfunction due to cisplatinum. Med Pediatr Oncol 1989;17:409–13.
16. Green D, Zevon M, Lowrie G, Seigelstein N, Hall B. Congenital anomalies in children of patients who received chemotherapy for cancer in childhood and adolescence. N Engl J Med 1991;325:141–6.