Ovarian cancer is the leading cause of mortality from gynecologic cancers in the United States, resulting in approximately 14,500 deaths annually.1 With an incidence of 26,600 cases per year, ovarian carcinoma will develop in one in 55 women in their lifetime. Despite good initial responses to surgery followed by chemotherapy, 75% of women ultimately die of complications associated with disease progression.2,3 During the past several decades, the practice of surgical staging and tumor debulking has had a significant effect on the survival of these patients.2,4 Furthermore, platinum-based regimens have played an essential role in advancing the treatment of ovarian malignancies.2
Conventional survival estimates are often outdated and overly pessimistic because they refer to cohorts of patients diagnosed many years ago.5,6 In fact, some report survival rates of only 5–17% for advanced stage disease.2 However, results from clinical trials often claim higher survival rates of up to 48% and may not translate to the general population due to potential patient selection and surveillance biases.7 To address this, we performed a large population-based study to estimate change in survival across time for women with ovarian cancer. More specifically, we propose to identify the demographic and clinicopathologic factors associated with treatment advances and uncover those that require further investigation.
PARTICIPANTS AND METHODS
Demographic, clinicopathologic, treatment, and survival information of women diagnosed with ovarian cancer during the period from 1988 through 2001 were extracted with permission from the Surveillance, Epidemiology and End Results (SEER) program of the United States National Cancer Institute. These data are representative of approximately 14% of the U.S. population and are reported from 12 population-based registries, including San Francisco–Oakland, Connecticut, metropolitan Detroit, Hawaii, Iowa, New Mexico, Seattle (Puget Sound), Utah, metropolitan Atlanta, Alaska, San Jose–Monterey, and Los Angeles.8
A total of 30,260 women were diagnosed with ovarian carcinoma after primary surgery. Fourteen patients with ovarian choriocarcinoma were excluded because of the small group size. The remaining 30,246 women were divided into three time intervals: 1988–1992, 1993–1997, and 1998–2001. These three intervals were chosen to provide three groups with a comparable number of patients and a similar span of years in each group. Factors including age at diagnosis, race, marital status, stage, tumor histology, grade of disease, type of surgery, and disease-specific survival were extracted.
Age was classified into younger than 50 years and 50 years or older as an arbitrary estimate of menopausal status. Race was classified into four groups, including whites, African Americans, Asians, and others. Asians included women of Chinese, Japanese, Filipino, Korean, and Vietnamese descent. Surgery was classified as surgery without hysterectomy and standard surgery. Standard surgery included women who underwent hysterectomy, debulking surgery, or exenteration. The public-use SEER data does not include information on surgeon specialty, extent of residual disease, chemotherapy, prior hysterectomy, or surgery beyond the primary procedure. Histologic cell types were categorized into non–clear cell epithelial (n=26,753), clear cell tumors (n=1,411), sarcomas (n=818), germ cell (n=778), and sex-cord stromal tumors (n=486).
Pearson’s χ2 tests were performed to analyze distributions in the study cohort across the three periods.9 Because follow-up data from the third interval, 1998–2001, have not yet matured to yield 5-year survival estimates, Kaplan-Meier analyses for 5-year survival rates were performed on the 1988–1992 and 1993–1997 intervals and compared using log rank tests.10,11 The outcome of interest was death from ovarian cancer as determined by the underlying cause of death on the death certificate. Thus, time to death was censored in women who died from causes other than ovarian cancer and who were alive at last follow-up. Cox proportional hazards were used for multivariable analyses.12 Two-tailed tests at P values less than .05 were considered significant. All data were analyzed using Intercooled STATA 8.0 (StataCorp LP, College Station, TX) and SAS 6.12 (SAS Institute Inc., Cary, NC).
A total of 30,246 women were diagnosed with ovarian carcinoma from 1988 to 2001. Demographic, clinical, and pathologic characteristics are summarized in Table 1 across three intervals, 1988–1992, 1993–1997, and 1998–2001. The proportion of whites diagnosed with ovarian cancer decreased from 88.5% to 85.7% to 84.6%, whereas the proportion of Asians increased from 4.0% to 5.8% to 6.3% (P<.001).
During the study period, surgery remained as the initial treatment in the majority of women with ovarian cancer. The proportion of all women receiving standard procedures has increased from 76.1% to 78.8% to 87.9% across the three intervals (P<.001). More specifically, the proportion of white women who received standard surgery has increased across the three intervals (64.7% to 65.9% to 68.3%; P<.001). Surgical distribution in African-American and Asian women show similar increases, although differences did not reach statistical significance (51.8% to 56.7% to 56.7% in African Americans, P=.11, and 65.6% to 68.2% to 69.4% in Asians, P=.23). However, the proportion of African Americans receiving standard surgeries remained significantly lower than whites and Asians in the interval 1998–2001 (55.3% compared with 66.3% and 68.1%, P<.001). Although there was no change in the proportion of advanced stage III-IV patients across time, there was a significant increase in the proportion of patients presenting with stage III disease, from 26.7% to 34.9% to 42.9%, with a concomitant decrease in the proportion of stage IV disease over the study period (P<.001).
During the 5-year intervals 1988–1992 and 1993–1997, women diagnosed with ovarian carcinoma had an improvement in 5-year survival rate from 45.4% to 48.6% (P<.001) (Fig. 1). After dividing our study population into early (stage I-II) and advanced (stage III-IV) stage disease, women with early stage cancers did not benefit from an improvement in 5-year survival rate across time (82.7% to 84.0%; P=.51), whereas those with advanced stage cancers had a significantly increased 5-year survival rate, from 25.4% to 29.4%, (P<.001; Table 2). When analyzed by histologic cell type, those with non–clear cell epithelial ovarian tumors had a significantly improved survival rate from 42.5% to 45.8%, (P<.001). Patients with sarcomas showed a nonsignificant increase in disease-specific survival rate (33.5% to 38.8%, P=.07). Patients with germ cell tumors showed a similar although nonsignificant increase in survival rate (91.9% to 94.7%, P=.17). However, patients with clear cell and sex cord stromal tumors of the ovary did not show improvement across time (P=.82 and P =.18, respectively; Table 3).
Age at diagnosis was a significant prognostic factor in survival in our study group. When analyzed across the study periods, women with ovarian cancer and aged younger than 50 years had a 5-year survival rate of 70.5% compared with 40.6% in those aged 50 years or older (P<.001). The significance of age persisted after controlling for stage and cell type. We also analyzed survival rates of the younger and older cohorts across time to estimate survival in each group of patients to see whether each has benefited from the improvement found in our study. Between 1988 and 1997, women aged 50 years or older had an increased 5-year survival rate from 37.5% to 41.5% (P<.001). On the other hand, patients aged younger than 50 years did not show any survival benefit (70.6% compared with 69.9%, P=.16; Table 2, Fig. 2).
When survival of ovarian cancer was stratified by race, whites had a significantly improved survival across the two periods, with 5-year survival rates increasing from 45.0% to 47.8%, (P<.001; Table 2). Improvements in 5-year survival rates of similar magnitude were noted for African Americans and Asians; however, they did not reach statistical significance (P=.36 and P=.12). It should ne noted that when survival was analyzed across the entire study period, African Americans had the lowest overall 5-year survival rate at 46.6%, whereas Asians had the highest survival rate at 58.5% (P<.001). In a subanalysis, we found that a higher proportion of Asians were diagnosed at younger ages (age of diagnosis younger than 50 years; 34.0% compared with 20.9% compared with 27.1%; P<.001) and with earlier stage (stage I-II; 43.6% compared with 31.7% compared with 30.0%; P<.001) compared with whites and African Americans, respectively.
Patients with stage III and IV disease who received standard surgery had significantly higher 5-year disease-specific survival rates compared with those who received surgery without hysterectomy. Of the advanced stage patients, those who underwent debulking surgery had a median survival advantage of 8 months over those who had surgery without hysterectomy (P<.001). The 5-year survival rates of those who received standard surgery and surgery without hysterectomy were 32.5% and 27.7%, respectively (P<.001). Across time, women with advanced stage disease benefited from the standard surgery with an improved 5-year survival rate from 30.1% to 34.3% (P<.001).
In multivariable analysis, younger age (P<.001), earlier stage (stage I compared with II compared with III compared with IV, P<.001), low grade of disease (grade 1, P<.001), favorable histologic cell types (germ cell, P<.001), standard surgery (P<.001), and recent time interval (year of diagnosis from 1993–1997, P=.002) remained as independent prognostic factors for improved survival (Table 4).
Because previous reports on ovarian cancer survival estimates were based on patients diagnosed many years ago, they report dismal 5-year survival rates at only 15–20% for stage III and IV disease.5,6 These outdated and overly pessimistic estimates are in contrast to the more up-to-date and encouraging results from this current study. This population-based study detected a significant increase in 5-year survival in advanced stage disease during a 10-year period. Similarly, studies from Barnholtz-Sloan et al13 and others14,15 reported that women with distant ovarian disease have an improved 5-year survival rate from 17% to 27%. However, many of these reports were limited by the lack of International Federation of Gynecology and Obstetrics staging and histologic information. This current study is one of the largest population-based studies that consist exclusively of patients with detailed histologic and surgical information.
Despite the progress in treatment of advanced ovarian cancers during the last 10 years, we have not improved the survival of young patients. The authors recognize that germ cell tumors of the ovary occur mainly in girls and young women with ages ranging from 6–40 years depending upon histologic cell type.16 Furthermore, the vast majority of these patients will survive after surgical staging and adjuvant chemotherapy.17 As such, it may be difficult to detect a survival improvement in these young patients given that these patients have excellent survivals from their germ cell cancers. However, the lack of survival improvement in young women may only be partially explained by the higher proportion of germ cell tumors compared with the older cohorts. Some studies have found that a number of these young patients with early-stage but poor prognostic ovarian cancers do not undergo adjuvant chemotherapy. Cress et al18 studied 2,150 women with ovarian cancer and found that approximately 20% of patients younger than 55 years with stage IC and II ovarian cancer did not receive chemotherapy. However, the likelihood of receiving chemotherapy was significantly increased if a gynecologic oncologist was involved in the patient’s care.
With respect to the surgical treatment of ovarian cancer, there was an increase in the proportion of standard surgery performed across time. These findings may likely reflect an improved understanding of the ovarian cancer disease process and the importance of tumor debulking surgery in the primary treatment. Bristow et al19 showed that maximal cytoreduction was one of the most powerful determinants of survival among patients with stage III or IV ovarian cancer. These authors also advocated that consistent referral of patients with advanced ovarian cancer to expert centers for primary surgery may improve the overall survival of ovarian cancer patients. Clearly, the improvement in survival across time may also be attributed to the advances in chemotherapy and cancer supportive care.
During the 10-year period, we found an improvement in the survival of women with surgically staged III and IV disease but not in those with early stage disease. On the other hand, Barnholtz-Sloan et al13 showed a relative survival improvement across time in patients with both localized and regional disease. This current report included analysis based on the more commonly used International Federation of Gynecology and Obstetrics staging for ovarian cancer. Moreover, we reported on an updated analysis on disease-specific survival rather than all-cause survival, with complete histology and surgery information. With more updated and detailed information, we found that the most important factor contributing to the improved survival of ovarian cancer patients seems to be associated with the enhanced survival of those with advanced stage cancers.
In our subanalysis of epithelial tumors, we found a lower 5-year survival rate in those with advanced clear cell cancers compared with their epithelial counterparts. Moreover, we were unable to detect any survival improvements in clear cell carcinomas across time despite advances in surgery and adjuvant chemotherapy over the study period. In a literature review of more than 400 cases of clear cell tumors, the 5-year survival rate for stage I disease was only 60% and 12% for all other stages. Previous analyses have also shown that ovarian clear cell carcinoma is a poor prognostic cell type compared with other epithelial malignancies adjusted for stage of disease.20 However, other studies have not found significant differences between clear cell and other epithelial cell types.21 Our study showed that those with advanced stage clear cell cancers have significantly worse 5-year survival rates compared with other epithelial tumors (P =.008). Although we were unable to demonstrate a statistically significant difference in the survival of those with sarcomas, there seems to be clinical improvement in the outcomes over the years (33.5% compared with 38.8%; P =.07). Nevertheless, further investigations are required in these aggressive histologic cell types given the overall poor prognosis. Although younger age, earlier stage, lower grade, standard surgery, and favorable histologic cell types are all independent prognostic factors for improved survival in multivariate analysis, these favorable demographic and clinicopathologic prognostic factors do not entirely explain the improvement in survival across time. These findings most likely reflect the advances in chemotherapy and cancer supportive care over the years.
The strength of this study lies in the large number of patients, which offers the ability to perform detailed, stratified analyses without sacrificing statistical strength. However, these conclusions can potentially be misleading if there is an overreliance on interpreting the P values for each subgroup. Our analytic approach requires that smaller subgroups exhibit a much greater benefit than that experienced by their larger cohorts to attain statistical significance. For example, even though we showed a statistically nonsignificant +5.3% survival rate improvement in the small subgroup of patients with ovarian sarcomas, this finding may still suggest a clinically meaningful improvement because the overall survival of sarcoma patients is so poor. On the other hand, we showed a +3.3% statistical benefit in those with non–clear cell epithelial ovarian cancers. This improvement is relatively small and may not provide a notable clinical benefit given this dismal improvement and the overall poor outcome of advanced epithelial ovarian cancer patients.
Our study was limited by the lack of information on surgeon specialty, extent of residual disease, neoadjuvant and postoperative chemotherapy, subsequent cytoreductive surgeries, second-look surgeries, and lack of central pathology review. Even though we showed an improved survival rate in those who underwent a standard surgery, these findings may be confounded by the selection of potentially healthier patients in the standard surgery group. Because the performance status of patients is not reported in our database, it is difficult to prove definitively that standard surgery alone enhanced the survival of these women. Given that SEER does not provide past medical information on prior hysterectomy or subsequent surgeries, the limitations of these data can potentially confound our results by understating our results on the benefits of standard surgery.
To determine whether there are discrepancies between registry and referral pathologists, Piver et al22 reviewed slides from a large cancer registry and found a 95.3% complete agreement between pathologists on the disease site of origin. Moreover, there was a 61.7% complete histopathologic agreement with only 1% of cases that were considered as having major differences. Most importantly, SEER uses several quality control measures to ensure accuracy; thus, they are able to maintain the highest level of certification of data quality and completeness as reported by the Northern American Association of Central Cancer Registries.23 The Surveillance, Epidemiology and End Results program adheres to strict quality assessment measures by ensuring the accuracy of sample cases by reabstracting data from the medical records annually.8 A recent study by Virnig et al24 found a 98% completeness in each sample case.
The results from this updated analysis reflect the improved survival rates reported in clinical trials. The 3-year estimates for patients diagnosed with ovarian cancer from 1997–2001 are more than 60%, suggesting that the projected 5-year survival estimates for these patients may reach more than 50%. Encouraging results from this study will help to counsel women diagnosed with ovarian cancer and prevent undue discouragement from outdated reports. Despite the overall improvement in ovarian cancer survival, novel treatment strategies are still warranted for patients diagnosed with epithelial and sarcomas of the ovary, given their overall poor prognosis.25
1. Edwards BK, Brown ML, Wingo PA, Howe HL, Ward E, Ries LA, et al. Annual report to the nation on the status of cancer, 1975-2002, featuring population-based trends in cancer treatment. J Natl Cancer Inst 2005;97:1407–27.
2. Hoskins WJ, Perez CA, Young RC. Principles and practice of gynecologic oncology. Philadelphia (PA): Lippincott Williams & Wilkins; 2000.
3. DiSaia PJ, Creasman WT. Clinical gynecologic oncology. St. Louis (MO): Mosby; 2002.
4. Averette HE, Janicek MF, Menck HR. The National Cancer Data Base report on ovarian cancer. American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer 1995;76:1096–103.
5. Yancik R. Ovarian cancer: age contrasts in incidence, histology, disease stage at diagnosis, and mortality. Cancer 1993;71:517–23.
6. Young RC, Walton LA, Ellenberg SS, Homesley HD, Wilbanks GD, Decker DG, et al. Adjuvant therapy in stage I and stage II epithelial ovarian cancer. Results of two prospective randomized trials. N Engl J Med 1990;322:1021–7.
7. Ozols RF, Bundy BN, Greer BE, Fowler JM, Clarke-Pearson D, Burger RA, et al. Phase III trial of carboplatin and paclitaxel compared with cisplatin and paclitaxel in patients with optimally resected stage III ovarian cancer: a Gynecologic Oncology Group study. J Clin Oncol 2003;21:3194–200.
8. National Cancer Institute, Division of Cancer Control and Population Sciences, Surveillance Research Program, Cancer Statistics Branch. Surveillance, Epidemiology, and End Results (SEER) Program, Public-Use Data (1973–2001). Available at: http://seer.cancer.gov/about
. Retrieved November 30, 2004.
9. Pearson K. On the criterion that a given system of deviations from the probable in the case of correlated system of variables is such that it can be reasonably supposed to have arisen from random sampling. Philos Mag 1900;50:157–75.
10. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–81.
11. Bland JM, Altman DG. The logrank test. BMJ 2004;328:1073.
12. Cox DR. Regression models and life-tables. J R Stat Soc Br 1972;34:187–220.
13. Barnholtz-Sloan JS, Schwartz AG, Qureshi F, Jacques S, Malone J, Munkarah AR. Ovarian cancer: changes in patterns at diagnosis and relative survival over the last three decades. Am J Obstet Gynecol 2003;189:1120–7.
14. Engel J, Eckel R, Schubert-Fritschle G, Kerr J, Kuhn W, Diebold J, et al. Moderate progress for ovarian cancer in the last 20 years: prolongation of survival, but no improvement in the cure rate. Eur J Cancer 2002;38:2435–45.
15. Brenner H. Long-term survival rates of cancer patients achieved by the end of the 20th century: a period analysis. Lancet 2002;360:1131–5.
16. Gershenson DM, Copeland LJ, Kavanagh JJ, Cangir A, Del Junco G, Saul PB, et al. Treatment of malignant nondysgerminomatous germ cell tumors of the ovary with vincristine, dactinomycin, and cyclophosphamide. Cancer 1985;56:2756–61.
17. Williams S, Blessing JA, Liao SY, Ball H, Hanjani P. Adjuvant therapy of ovarian germ cell tumors with cisplatin, etoposide, and bleomycin: a trial of the Gynecologic Oncology Group. J Clin Oncol 1994;12:701–6.
18. Cress RD, O’Malley CD, Leiserowitz GS, Campleman SL. Patterns of chemotherapy use for women with ovarian cancer: a population-based study. J Clin Oncol 2003;21:1530–5.
19. Bristow RE, Tomacruz RS, Armstrong DK, Trimble EL, Montz FJ. Survival effect of maximal cytoreductive surgery for advanced ovarian carcinoma during the platinum era: a meta-analysis. J Clin Oncol 2002;20:1248–59.
20. Vergote IB, Kaern J, Abeler VM, Pettersen EO, De Vos LN, Trope CG. Analysis of prognostic factors in stage I epithelial ovarian carcinoma: importance of degree of differentiation and deoxyribonucleic acid ploidy in predicting relapse. Am J Obstet Gynecol 1993;169:40–52.
21. Pettersson F. International Federation of Gynecology and Obstetrics: annual report on the results of treatment in gynecological cancer. Vol 20. Stockholm (Sweden): Panorama Press AB; 1988. p. 110.
22. Piver MS, Tsukada Y, Werness BA, DiCioccio RA, Whittemore AS, Ponder BA. Comparative study of ovarian cancer histopathology by registry pathologists and referral pathologists: a study by the Gilda Radner Familial Ovarian Cancer Registry. Gynecol Oncol 2000;78:166–70.
23. North American Association of Central Cancer Registries. Available at: http://www.naaccr.org/
. Retrieved June 12, 2006.
24. Virnig BA, Warren JL, Cooper GS, Klabunde CN, Schussler N, Freeman J. Studying radiation therapy using SEER-Medicare-linked data. Med Care 2002;40:IV-49–54.
25. Chan JK, Hamilton CA, Cheung MK, Karimi M, Baker J, Gall JM, et al. Enhanced killing of primary ovarian cancer by retargeting autologous cytokine-induced killer cells with bispecific antibodies: a preclinical study. Clin Cancer Res 2006;12:1859–67.