Epithelial ovarian cancer is usually diagnosed in postmenopausal women, but, in the United States, 7% of cases occur in women younger than 40 years old.1–4 The possibility of fertility preservation is of great importance to many young women diagnosed with cancer, and infertility resulting from surgical excision of the reproductive organs is associated with significant psychosocial distress.5,6 Fertility-sparing treatment for gynecologic cancer has become increasingly relevant in developed countries, where women frequently choose to delay childbearing.7 Although nearly 75% of women with epithelial ovarian cancer are diagnosed with advanced-stage disease, younger patients are more likely to present with early-stage disease and to have better prognoses than older women.3,8–10
Despite the absence of randomized trials, observational studies have driven cautious acceptance of fertility-sparing surgery for stage I ovarian cancer.1,3,5,9,11–13 However, little population-level data are available regarding long-term survival outcomes after fertility-sparing surgery. Studies published to date have often been underpowered or have failed to control for important treatment variables such as administration of chemotherapy. Furthermore, there is disagreement regarding the appropriateness of fertility-sparing surgery for women with certain high-risk features including ovarian surface involvement, clear cell histology, and high pathologic grade.14,15 In this study, we used a national cancer registry to compare all-cause survival between reproductive-aged women who underwent fertility-sparing surgery and those who underwent conventional surgery for stage I epithelial ovarian cancer.
MATERIALS AND METHODS
We conducted a cohort study using prospectively collected data from the National Cancer Database, a national cancer registry that includes information on approximately 70% of all incident cancers diagnosed in the United States. The National Cancer Database, a joint project of the American College of Surgeons and the American Cancer Society, serves as a surveillance mechanism for more than 1,400 hospitals participating in the American College of Surgeons' Commission on Cancer.16 Data are collected by trained hospital cancer registrars and include demographics, clinical data, histopathology, primary treatment, and survival. Because all data are deidentified and publicly available, this study was deemed exempt from internal review board oversight by the Partners Healthcare Research Committee.
We identified all women diagnosed with American Joint Committee on Cancer pathologic stage IA and IC invasive (excluding borderline) epithelial ovarian cancer between January 1, 2004, and December 31, 2012.17 We defined epithelial histology using International Classification of Diseases for Oncology codes corresponding to serous, mucinous, clear cell, endometrioid, or other adenocarcinoma (Appendix 1, available online at http://links.lww.com/AOG/A962).18 We excluded patients who received no treatment at the reporting facility, lacked microscopic confirmation of ovarian cancer, received systemic treatment before surgery, did not undergo surgery, had bilateral ovarian tumors, or in whom it was not possible determine fertility conservation status (Fig. 1).
The exposure of interest was fertility-sparing surgery. We categorized patients who underwent surgery in which one ovary and the uterus were conserved as having had fertility-sparing surgery and those who underwent bilateral oophorectomy or hysterectomy (with unilateral or bilateral oophorectomy) as having undergone conventional surgical management (women in the control group). The primary outcome was overall survival, defined as months from diagnosis to death or the date of last contact, as reported by the cancer registrar.
Control variables included age at diagnosis, race–ethnicity (Asian, black, Hispanic, white, other or unknown), and year of diagnosis. Geographic region, based on the treating facility, was categorized by Census region as Midwest, Northeast, South, or West. As a proxy for patient income, we used median household income in the patient's residential zip code as estimated by the 2012 American Community Survey, categorized in quartiles (less than $38,000, $38,000–47,999, $48,000–62,999, and $63,000 or greater). We categorized insurance status as uninsured, privately insured, insured by Medicare, or insured by Medicaid. The degree of educational attainment in each patient's area of residence was estimated using the 2012 American Community Survey and categorized by quartile of the proportion of adults in the patient's zip code who did not graduate from high school. The treating facility was categorized as an academic cancer program, community cancer program, or integrated network cancer program.
We compared distributions of demographic, clinical, and treatment characteristics using the Pearson χ2 and Wilcoxon rank-sum tests. Time trends in use of fertility-sparing surgery were evaluated using linear weighted least-squares regression, implemented in the Joinpoint Regression Program, with log-transformed proportion of patients undergoing fertility-sparing surgery as the dependent variable.19
We used propensity score matching to assemble a cohort in which women who underwent fertility-sparing surgery and conventional surgical management were balanced on observed covariates that might confound the effect of treatment approach on survival.20 We fit a logistic regression model to predict each patient's log-odds of receiving fertility-sparing surgery with age group, year of diagnosis, race–ethnicity, geographic region, insurance status, income level, education level, rural or urban status, treating facility type, substage, histology, grade, lymph node dissection, and adjuvant chemotherapy status included as independent variables (Appendix 2, available online at http://links.lww.com/AOG/A962). Missing data were coded with dummy variables. We matched each woman who received fertility-sparing surgery to another who underwent conventional surgery but had a comparable propensity to undergo fertility-sparing surgery using a 1:1 nearest-neighbor algorithm with a caliper width set to 0.2 times the standard deviation of the propensity score.21 To evaluate success of matching, we calculated absolute standardized differences in observed covariates and considered covariates to be balanced if the differences were less than 10% (Appendix 3, available online at http://links.lww.com/AOG/A962).22
We plotted survival curves for women who underwent fertility-sparing surgery and propensity-matched women in the control group using the Kaplan–Meier method. We estimated the probability of survival 5 and 10 years after diagnosis as well as the associated 95% confidence intervals (CIs) and compared the survival functions using the log-rank test. A Cox proportional hazards model was used to estimate the hazard ratio (HR) and 95% CI for all-cause mortality associated with fertility-sparing surgery.
To evaluate whether the findings in our survival analysis persisted among women with high-risk stage I epithelial ovarian cancer, we repeated the main analysis exclusively in patients with one or more high-risk features such as stage IC, grade 3, or clear cell histology.23 Because grade was unknown in 22% of women in the propensity-matched cohort, we tested the robustness of the aforementioned analysis by including women with unknown grade in the high-risk group. Additionally, we calculated subgroup-specific relative hazard of all-cause mortality according to tumor grade, substage, histologic type, lymphadenectomy status, and use of adjuvant chemotherapy. As a result of the frequency of unknown grade (22%), we undertook sensitivity analyses in which all patients with unknown grade were categorized as grade 1, grade 2, or grade 3 to determine whether these missing data were likely to bias subgroup-specific results.
Because propensity score methods can adjust only for confounders that are measured, we assessed the sensitivity of our findings to potential confounding by an unmeasured variable.24,25 We assumed that surgeons were less likely to offer fertility-sparing surgery to women with CA-125 elevation, a finding associated with increased risk of all-cause mortality in stage I ovarian cancer.26 We adjusted the HR estimating the association between fertility-sparing treatment and all-cause mortality under a range of possible differences between groups in the prevalence of elevated CA-125.25
We identified 2,070 women younger than 40 years who were treated for stage IA and IC epithelial ovarian cancer between 2004 and 2012, of whom 1,726 met study inclusion criteria (Fig. 1). Of these women, 825 (47.8%) underwent fertility-sparing surgery. The proportion of women who underwent fertility-sparing treatment increased by an average of 4.4% per year (P<.001; Appendix 3, http://links.lww.com/AOG/A962), from 40.3% in 2004 to 53.7% in 2012.
There were significant demographic and clinical differences between women who received fertility-sparing surgery and those who received conventional surgery (Table 1). Fertility-sparing surgery was utilized more frequently in younger women (85.2% in women younger than 25 years compared with 27.5% in those 35 years and older), nonwhite women (54.5% compared with 42.3%), those treated in the West and Northeast (56.0% compared with 42.7%), and urban dwellers (50.0% compared with 31.3%). Fertility-sparing surgery was also more common among women with serous and mucinous histology (55.7% compared with 40.3%), those who did not undergo lymphadenectomy (68.6% compared with 40.3%), and those who received no adjuvant chemotherapy (53.4% compared with 40.4%).
Based on propensity score, we matched 452 women who underwent fertility-sparing surgery to 452 women who received conventional surgical management. Differences in distributions of observed covariates between women who underwent fertility-sparing and conventional treatments were no longer evident after propensity score matching (Table 1). Furthermore, in the propensity-matched sample, absolute standardized differences were less than 10% for all covariates, signifying excellent balance with respect to these variables (Appendix 4, available online at http://links.lww.com/AOG/A962).20
In matched participants, the median follow-up was 63 months and did not differ between women who underwent fertility-sparing or conventional surgery (P=.72). During the study period, there were 30 deaths among women who received fertility-sparing surgery and 37 deaths among those who received conventional surgery. Survival curves did not differ between women who underwent fertility-sparing surgery and propensity-matched women in the control group (Fig. 2A; log-rank P=.36). In comparison with propensity-matched patients who underwent conventional surgery, women who underwent fertility-sparing treatment exhibited no difference in relative hazard of all-cause mortality (HR 0.80, 95% CI 0.49–1.29). The probability of surviving 5 years after diagnosis was 93.5% (95% CI 90.5–95.6) and 90.5% (95% CI 86.9–93.2) among women who underwent fertility-sparing and conventional surgery, respectively. Ten-year survival was 88.5% (95% CI 82.4–92.6) among those who underwent fertility-sparing surgery and 88.9% (95% CI 84.9–92.0) in women in the control group. A post hoc power calculation demonstrated that at a confidence level of 0.05, this study had 80% power to detect a HR of 2.2 for fertility-sparing surgery, corresponding to an 11.8% decline in 10-year survival.
We performed a subgroup analysis to verify that the overall findings were applicable to patients with high-risk histopathology findings such as stage IC, grade 3, or clear cell histology. Although these 412 women represented 46% of the propensity-matched cohort, 70% of deaths (47/67) occurred in this group. Prognostic factors remained balanced between women who received fertility-sparing and conventional surgery in this subgroup of the propensity-matched cohort (Appendix 5, available online at http://links.lww.com/AOG/A962). The survival function for high-risk patients who underwent fertility-sparing surgery did not differ from that of high-risk women in the control group (Fig. 2B; log-rank P=.61). We observed no significant association between hazard of death and fertility-sparing status among patients with high-risk features (HR 0.86, 95% CI 0.49–1.53). Five-year survival was 89.9% (95% CI 84.1–93.7) and 85.6% (95% CI 78.9–90.3) and 10-year survival was 80.5% (95% CI 68.5–88.3) and 83.4% (95% CI 76.0–88.7) in women who received fertility-sparing and conventional surgery, respectively. Inclusion of women with unknown grade in the high-risk subgroup did not appreciably alter these findings (HR 0.93, 95% CI 0.56–1.69).
Figure 3 depicts relative hazard of death and 95% CIs for fertility-sparing surgery compared with conventional treatment for subgroups of propensity-matched patients defined by stage, grade, histologic type, lymphadenectomy status, and receipt of adjuvant chemotherapy. Fertility-sparing surgery was not associated with all-cause mortality hazard in any subgroup (P>.05 for all). In sensitivity analyses, categorizing women with unknown grade as having grade 1, grade 2, or grade 3 tumors did not appreciably alter grade-specific HRs for fertility-sparing surgery (P>.05 for all).
Estimates of the effect of potential confounding by a difference in the prevalence of elevated CA-125 are tabulated in Appendix 6, available online at http://links.lww.com/AOG/A962. Under all proposed scenarios, differences in the prevalence of CA-125 elevation between groups did not alter our main finding that fertility-sparing surgery was associated with no difference in hazard of death in comparison with conventional treatment.
In this large national study of the use and effectiveness of fertility-sparing surgery in young women with ovarian cancer, we observed no difference in all-cause mortality between young women treated with fertility-sparing surgery and similar women who underwent conventional surgery. These findings persisted among women with high-risk features such as stage IC disease, clear cell histology, or high pathologic grade. In addition, our analyses revealed that fertility conservation has become increasingly common in the United States between 2004 and 2012. Because logistic and ethical issues are significant barriers to a randomized trial of fertility-sparing surgery for epithelial ovarian cancer, well-designed observational studies offer the best available evidence to guide clinical practice.
Although some prior studies have also found that fertility-sparing surgery is safe in ovarian cancer, the large sample size, availability of many demographic and clinical variables, and use of propensity score matching in the present investigation lends additional support for this conclusion.1,3,9,11–13,27 In a retrospective multiinstitutional study performed in Japan, Kajiyama et al27 compared clinical outcomes between 74 patients who underwent fertility-sparing surgery and 498 patients who underwent conventional surgery and observed no difference in disease-free and overall survival after a median of 66 months. Similarly, in a multicenter Italian retrospective study, Fruscio et al11 compared 242 women who received fertility-sparing surgery with 789 patients who had conventional surgery and found that survival was similar between groups after 11.9 years of follow-up. Wright et al1 used the Surveillance, Epidemiology, and End Results database to show that neither ovarian nor uterine preservation was associated with overall or disease-specific survival in stage IA and IC ovarian cancer. However, Wright et al's study was limited by the absence of information about chemotherapy, raising the possibility of confounding. The present study demonstrates that fertility-sparing surgery is not associated with all-cause mortality in a propensity-matched cohort that was balanced on chemotherapy use.
Despite an accumulation of evidence for the safety of fertility-sparing surgery in young women with stage I ovarian cancer, controversy persists regarding the appropriateness of this approach for women with high-risk features.14,15,28–30 A retrospective study performed in France by Morice et al reported a 100% recurrence rate in three patients with stage IC ovarian cancer who underwent fertility-sparing surgery, and a Japanese multiinstitutional study by Satoh et al found a 50% recurrence rate among six patients with stage I grade 3 tumors.14,15 However, the ability of these studies to guide practice is limited by the small number of patients with high-risk features. In contrast to these studies, in this larger study, we observed no evidence that the fertility-sparing approach is inferior to conventional surgery among patients with high-risk features.
This study has several significant limitations. First, because propensity score matching can only control for confounding by measured covariates, we cannot exclude the possibility that the findings of this nonrandomized study were biased by treatment selection. Reassuringly in this regard, our results were insensitive to confounding by an unmeasured covariate with a moderate effect on mortality. Another important limitation is that data reported by hospital tumor registries do not include many variables of interest such as fertility outcomes, recurrence, cause of death, or presence of familial cancer syndrome. Furthermore, we have no information regarding whether or when completion surgery was performed in the patients included in this study. Participants' level of income and educational attainment were estimated from residential zip codes. Although the present study includes approximately 70% of all young women diagnosed with stage I epithelial ovarian cancer in the United States over 9 years, it is nonetheless underpowered to detect an effect smaller than a HR of 2.2. Although smaller effects may be clinically relevant, conducting a study with greater power poses a significant practical challenge and would likely require pooling multiple national databases. This limitation is especially significant with respect to conclusions about the safety of fertility-sparing surgery among small subgroups of patients, like those with clear cell cancer.
This study can reassure clinicians that fertility-sparing surgery does not impart a large risk of death for patients with stage I epithelial ovarian cancer who desire future fertility. Furthermore, there was no evidence of excess mortality among patients with stage IC or high-risk histology undergoing fertility-sparing surgery. Nonetheless, given the limited number of patients with clear cell and other high-grade histology included in the present study, preoperative counseling should convey that the safety of fertility-sparing surgery is less certain for women with clear cell tumors and those with other high-grade histology.
1. Wright JD, Shah M, Mathew L, Burke WM, Culhane J, Goldman N, et al. Fertility preservation in young women with epithelial ovarian cancer. Cancer 2009;115:4118–26.
2. Rauh-Hain JA, Foley O, Winograd D, Andrade C, Clark RM, Vargas RJ, et al. Clinical characteristics and outcomes of patients with stage I epithelial ovarian cancer compared to fallopian tube cancer. Am J Obstet Gynecol 2015;212:600.e1–8.
3. Kashima K, Yahata T, Fujita K, Tanaka K. Outcomes of fertility-sparing surgery for women of reproductive age with FIGO stage IC epithelial ovarian cancer. Int J Gynaecol Obstet 2013;121:53–5.
4. Centers for Disease Control and Prevention, CDC WONDER. United States and Puerto Rico cancer statistics, 1999–2013 incidence incidence request.Available at: http://wonder.cdc.gov/cancer-v2013.html
. Retrived December 1, 2016.
5. Lee SJ, Schover LR, Partridge AH, Patrizio P, Wallace WH, Hagerty K, et al. American Society of Clinical Oncology recommendations on fertility preservation in cancer patients. J Clin Oncol 2006;24:2917–31.
6. Jeruss JS, Woodruff TK. Preservation of fertility in patients with cancer. N Engl J Med 2009;360:902–11.
7. Matthews TJ, Hamilton BE. Delayed childbearing: more women are having their first child later in life. NCHS Data Brief 2009:1–8.
8. Melamed A, Keating NL, Clemmer JT, Bregar AJ, Wright JD, Boruta DM, et al. Laparoscopic staging for apparent stage I epithelial ovarian cancer. Am J Obstet Gynecol 2017;216:50.e1–50.e12.
9. Ghezzi F, Cromi A, Fanfani F, Malzoni M, Ditto A, De Iaco P, et al. Laparoscopic fertility-sparing surgery for early ovarian epithelial cancer: a multi-institutional experience. Gynecol Oncol 2016;141:461–5.
10. Smedley H, Sikora K. Age as a prognostic factor in epithelial ovarian carcinoma. Br J Obstet Gynaecol 2016;92:839–42.
11. Fruscio R, Ceppi L, Corso S, Galli F, Dell'Anna T, Dell'Orto F, et al. Long-term results of fertility-sparing treatment compared with standard radical surgery for early-stage epithelial ovarian cancer. Br J Cancer 2016;115:641–8.
12. Bentivegna E, Gouy S, Maulard A, Pautier P, Leary A, Colombo N, et al. Fertility-sparing surgery in epithelial ovarian cancer: a systematic review of oncological issues. Ann Oncol 2016;27:1994–2004.
13. Ditto A, Martinelli F, Bogani G, Lorusso D, Carcangiu M, Chiappa V, et al. Long-term safety of fertility-sparing surgery in early stage ovarian cancer: comparison to standard radical surgical procedures. Gynecol Oncol 2015;138:78–82.
14. Morice P, Leblanc E, Rey A, Baron M, Querleu D, Blanchot J, et al. Conservative treatment in epithelial ovarian cancer: results of a multicentre study of the GCCLCC (Groupe des Chirurgiens de Centre de Lutte Contre le Cancer) and SFOG (Société Francaise d'Oncologie Gynécologique). Hum Reprod 2005;20:1379–85.
15. Satoh T, Hatae M, Watanabe Y, Yaegashi N, Ishiko O, Kodama S, et al. Outcomes of fertility-sparing surgery for stage I epithelial ovarian cancer: a proposal for patient selection. J Clin Oncol 2010;28:1727–32.
16. Raval MV, Bilimoria KY, Stewart AK, Ko CY. Using the NCDB for cancer care improvement: an introduction to available quality assessment tools. J Surg Oncol 2009;99:488–90.
17. Edge S, DR Burd, Compton CC, Fritz AG, Greene FL, Trotti A, editors. AJCC cancer staging manual. 7th ed. New York (NY): Springer; 2009.
18. Fritz A, Percy C, Jack A, Shanmugaratnam K, Sobin L, Parkin DM, et al, editors. International classification of diseases for oncology. 3rd ed. Geneva (Switzerland): World Health Organization; 2000.
19. Kim HJ, Fay MP, Feuer EJ, Midthune DN. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 2000;19:335–51.
20. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41–55.
21. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat 2001;10:150–61.
22. Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 2009;28:3083–107.
23. Aebi S, Castiglione M; ESMO Guidelines Working Group. Newly and relapsed epithelial ovarian carcinoma: ESMO clinical recommendations for diagnosis, treatment and follow-up. Ann Oncol 2009;20(suppl 4):21–3.
24. Rosenbaum PR, Rubin DB. Assessing sensitivity to an unobserved binary covariate in an observational study with binary outcomes. J R Stat Soc Series B Methodol 1983;45:212–218.
25. Lin DY, Psaty BM, Kronmal RA. Assessing the sensitivity of regression results to unmeasured confounders in observational studies. Biometrics 1998;54:948–63.
26. Paramasivam S, Tripcony L, Crandon A, Quinn M, Hammond I, Marsden D, et al. Prognostic importance of preoperative CA-125 in International Federation of Gynecology and Obstetrics stage I epithelial ovarian cancer: an Australian multicenter study. J Clin Oncol 2005;23:5938–42.
27. Kajiyama H, Shibata K, Mizuno M, Umezu T, Suzuki S, Nawa A, et al. Long-term survival of young women receiving fertility-sparing surgery for ovarian cancer in comparison with those undergoing radical surgery. Br J Cancer 2011;105:1288–94.
28. Bentivegna E, Morice P, Uzan C, Gouy S. Fertility-sparing surgery in epithelial ovarian cancer. Future Oncol 2016;12:389–98.
29. Park JY, Suh DS, Kim JH, Kim YM, Kim YT, Nam JH. Outcomes of fertility-sparing surgery among young women with FIGO stage I clear cell carcinoma of the ovary. Int J Gynaecol Obstet 2016;134:49–52.
30. Kajiyama H, Shibata K, Mizuno M, Hosono S, Kawai M, Nagasaka T, et al. Fertility-sparing surgery in patients with clear-cell carcinoma of the ovary: is it possible? Hum Reprod 2011;26:3297–302.