Ovarian Cancer Statistics, 2018

Abstract

In 2018, there will be approximately 22,240 new cases of ovarian cancer diagnosed and 14,070 ovarian cancer deaths in the US. Herein, the American Cancer Society provides an overview of ovarian cancer occurrence based on incidence data from nationwide population-based cancer registries and mortality data from the National Center for Health Statistics. The status of early detection strategies is also reviewed. In the US, overall ovarian cancer incidence declined by 29% from 1985 (16.6 per 100,000) to 2014 (11.8 per 100,000), while mortality declined 33% from 1976 (10.0 per 100,000) to 2015 (6.7 per 100,000). Ovarian cancer encompasses a heterogenous group of malignancies that vary in etiology, molecular biology, and numerous other characteristics. Ninety percent of ovarian cancers are epithelial, the most common being serous carcinoma, for which incidence is highest in non-Hispanic whites (NHWs; 5.2 per 100,000) and lowest in non-Hispanic blacks (NHBs; 3.4 per 100,000) and Asian/Pacific Islanders (APIs; 3.4 per 100,000). Notably, however, APIs have the highest incidence of endometrioid and clear cell carcinoma, which occur at younger ages and help explain comparable epithelial cancer incidence for APIs and NHWs younger than 55 years. Most serous carcinomas are diagnosed at stage III (51%) or IV (29%), for which the 5-year cause-specific survival is 42% and 26%, respectively. For all stages of epithelial cancers combined, five-year survival is highest in APIs (57%) and lowest in NHBs (35%), who have the lowest survival for almost every stage of diagnosis across cancer subtype. Epithelial ovarian cancer survival has plateaued in NHBs for decades, whereas it continues to increase in NHWs, from 40% for cases diagnosed during 1992–1994 to 47% during 2007–2013. Although there has been progress in reducing ovarian cancer incidence and mortality overall, further research is needed to determine the source of racial disparities and identify effective mechanisms for early ovarian cancer detection and prevention.

Introduction

In 2018, there will be approximately 22,240 new cases of ovarian cancer diagnosed and 14,070 ovarian cancer deaths in the US.¹ Ovarian cancer accounts for 2.5% of all malignancies among females, but for 5% of all cancer deaths due to the disease’s relatively high fatality rate, as 4 out of 5 patients are diagnosed with advanced disease.² Improving early detection and prevention is a research priority because local-stage disease has a 5-year relative survival rate of 93%.² Although advancing knowledge about ovarian cancer has previously been hindered by substantial disease heterogeneity and uncertainties about tumor tissues of origin, insight has evolved rapidly in recent years, especially for epithelial tumors, the most common type. This article provides an overview of ovarian cancer occurrence in the US, including incidence, mortality, and survival rates and trends (by subtype when possible), as well as a summary of recent research on early detection strategies.

Materials and Methods

Population-based cancer incidence data in the United States are collected by the National Cancer Institute’s (NCI’s) Surveillance, Epidemiology, and End Results (SEER) program and the Centers for Disease Control and Prevention’s National Program of Cancer Registries (NPCR). The North American Association of Central Cancer Registries (NAACCR) compiles and reports incidence data from 1995 onward for registries that participate in the SEER program and/or the NPCR. These data approach 100% coverage of the US population in the most recent time period and were the source for the projected new cancer cases in 2018 and cross-sectional incidence rates by age, histology, and race/ethnicity.^{3, 4} Some of the incidence data presented herein were previously published in volumes 1 and 2 of Cancer in North America: 2010–2014.^{4, 5}

All invasive ovarian cancer cases were classified according to the International Classification of Diseases for Oncology, version three (C56.9).^{6, 7} Causes of death were classified according to the International Classification of Diseases.⁸ Incident cases were further grouped by histologic subtype, while this information is not available in mortality data.

Long-term incidence trends (1975–2014) were based on data from the 9 oldest SEER databases (Connecticut, Hawaii, Iowa, New Mexico, Utah, and the metropolitan areas of Atlanta, Detroit, San Francisco-Oakland, and Seattle-Puget Sound), representing approximately 9% of the US population.⁹ Beginning in 1992, data became available for Asians/Pacific Islanders (APIs), American Indians/Alaska Natives, and by Hispanic ethnicity from the SEER 13 registries (SEER 9 plus Los Angeles, San Jose-Monterey, rural Georgia and the Alaska Native Tumor Registry), representing 13% of the US population, and were used in analyses of incidence and survival trends by race/ethnicity.¹⁰ Trends by histology from 1995 to 2014 are based on NAACCR data from 25 registries with complete data for the period covering 66% of the US population to maximize the number of cases in rarer subtypes.³ The lifetime probability of developing cancer and contemporary stage distribution and cause-specific survival statistics were based on data from all 18 SEER registries (SEER 13 plus Greater California, Greater Georgia, Kentucky, Louisiana, and New Jersey), covering 28% of the US population.¹¹

National mortality data were obtained from the SEER program’s SEER*Stat database, as provided by the National Center for Health Statistics.¹² Data for APIs, American Indians/Alaska Natives, and by Hispanic ethnicity are available from 1990. Some of the statistical information presented herein was adapted from data previously published in the SEER Cancer Statistics Review 1975–2014.²

All incidence and death rates were age-standardized to the 2000 US standard population and expressed per 100,000 population, as calculated by NCI’s SEER*Stat software (version 8.3.4).¹³ Whenever possible, cancer incidence rates were adjusted for delays in reporting, which have the greatest impact on the most recent data years and occur because of a lag in case capture or data corrections. The annual percent change in rates was quantified using NCI’s Joinpoint Regression Program (version 4.5.0.1).¹⁴ The probability of developing cancer was calculated using NCI’s DevCan software (version 6.7.5).¹⁵

Selected Findings

Ovarian Cancer Occurrence Overall

This section presents overall incidence and mortality statistics for ovarian cancer. While it is generally more appropriate to provide information separately by histologic subtype, combined incidence allows comparison with other major cancer sites and with ovarian cancer mortality, which is unavailable by subtype, and provides consistency for historic statistics and trends due to changes in tumor classification over time. Information for specific cancer subtypes is presented in the subsequent section.

Contemporary incidence and mortality

The average lifetime risk of developing ovarian cancer is 1.3%, the equivalent of 1 in 78 women (Table 1). The overall ovarian cancer incidence rate in the US was 11.5 per 100,000 women during 2010–2014.³ Incidence rates in NHW women (12.0 per 100,000 women), who have the highest rates, are 30% higher than those in NHB (9.4) and API women (9.2), who have the lowest rates (Figure 1). Racial/ethnic differences in ovarian cancer risk at the population level are partially explained by the prevalence of risk factors. For instance, higher parity, use of oral contraceptives, tubal ligation, and oophorectomy reduce risk, while menopausal hormone use increases risk.^16–19 However, the source of most of the variation remains unknown.²⁰ As with incidence, mortality rates are highest in NHW women (7.9 deaths per 100,000 women) and lowest in API women (4.4; Figure 1). NHB women have the second-highest mortality rates (6.6 deaths per 100,000 women), despite relatively low incidence rates, likely due in part to a lower likelihood of receiving optimal treatment and more comorbidities compared to other women.^{21, 22}

Table 1.

Age-specific Probability of Developing Ovarian Cancer for US Women^*

Age	10-year probability:	or 1 in:
40	0.1%	870
50	0.2%	474
60	0.3%	327
70	0.4%	265
80	0.4%	283
Lifetime risk	1.3%	78

^*Among those who are cancer-free. Based on cases diagnosed 2012–2014. Percentages and “1 in” numbers may not be numerically equivalent due to rounding.

Source: DevCan: Probability of Developing or Dying of Cancer Software, Version 6.7.5

Surveillance Research Program, Statistical Methodology and Applications, National Cancer Institute, 2012. http://surveillance.cancer.gov/devcan/

Figure 1. Ovarian Cancer Incidence and Mortality Rates* by Race and Ethnicity†, US, 2010–2014.

*Per 100,000, age adjusted to the 2000 US standard population.

†Persons of Hispanic origin may be of any race; blacks, American Indians/Alaska Natives, and Asian/Pacific Islanders include those of Hispanic and non-Hispanic origin.

Sources

Incidence: NAACCR, 2017.

Mortality: US mortality data, National Center for Health Statistics, Centers for Disease Control and Prevention, 2017. Data for American Indians/Alaska Natives are based on Contract Health Service Delivery Area (CHSDA) counties.

Trends in incidence and mortality

Ovarian cancer incidence has been decreasing since the mid-1980s, largely driven by declines in whites that accelerated during the past decade. Overall, the incidence rate dropped 29%, from 16.6 (per 100,000) in 1985 to 11.8 in 2014.² However, trends differ by age. Among whites and blacks 65 years of age and older, incidence rates increased from 1975 until around 1990 before beginning to decline (Figure 2).⁹ This increase may be related to the birth rate decline during the early- to mid-20^th century; invasive epithelial ovarian cancer risk is reduced by about 20% with the first childbirth and by about 10% with each additional birth.¹⁹ The more rapid decline among white women in recent years may be partly due to decreased use of menopausal hormones following publication of a landmark report in 2002 linking them to increased breast cancer risk.²³ Women who have ever used menopausal hormones (estrogen alone or estrogen combined with progesterone) have a 20% higher risk of developing ovarian cancer compared to never-users, with a stronger risk among recent users; current users and those who have stopped within 5 years have an excess risk of about 40%.¹⁸ Risk is increased even with short duration of menopausal hormone use and remains elevated for at least 10 years after discontinuation.

Figure 2. Age-adjusted Ovarian Cancer Delay-adjusted Incidence and Mortality Rates* by Age Group and Race/ethnicity, 1975–2015.

*Per 100,000, age adjusted to the 2000 US standard population. Incidence rates are three-year moving averages. Note: American Indians/Alaska Natives not pictured due to <25 deaths in some years.

Sources

Incidence: SEER 9 (white, black; 1975–2014), SEER 13 (Hispanic, API; 1992–2014). SEER program, 2017.

Mortality: US mortality data, National Center for Health Statistics, Centers for Disease Control and Prevention, 2017. Data for Hispanics excludes Louisiana, New Hampshire, and Oklahoma.

In contrast, incidence among women <65 years has generally declined at a continuous rate since at least 1975 (Figure 2). This is likely due to uptake of oral contraceptives, which confer a substantial risk reduction and also likely contributed to the recent declines in older women.²⁴ Among women who use oral contraceptives for five to nine years total, risk is reduced by about 35%, with the protective effect persisting with diminishing strength for at least 30 years after discontinuation.¹⁷

The mortality rate for ovarian cancer has declined 33% from 1976 (10.0 per 100,000) to 2015 (6.7 per 100,000) due to reductions in incidence and improvements in treatment.^25–28 Age-specific mortality trends closely mirror those of incidence because of generally low survival (Figure 2). Death rates declined during the most recent 10 years of data (2006 to 2015) in all racial/ethnic groups except American Indians/Alaska Natives, among whom rates were stable (Table 2).²⁵ Mortality rates decreased most rapidly among Hispanics and NHWs, by about 2% annually, compared to about 1% annually among NHBs and APIs (Table 2).

Table 2.

Trends in Ovarian Cancer Mortality Rates by Age and Race/ethnicity, United States, 1992 to 2015

	Trend 1			Trend 2			Trend 3			2006–2015
	Years	APC	95% CI	Years	APC	95% CI	Years	APC	95% CI	AAPC
All races
All ages	1992–1998	−1.2*	(−1.8 to −0.6)	1998–2003	0.6	(−0.5 to 1.7)	2003–2015	−2.3*	(−2.5 to −2.1)	−2.3*
<65 years	1992–1998	−2.3*	(−3.2 to −1.3)	1998–2002	0.5	(−2.1 to 3.3)	2002–2015	−2.7*	(−3.0 to −2.4)	−2.7*
≥65 years	1992–2005	0.0	(−0.2 to 0.3)	2005–2015	−2.2*	(−2.6 to −1.9)				−2.2*
Non-Hispanic white
All ages	1992–2005	−0.2	(−0.5 to 0.0)	2005–2015	−2.5*	(−2.8 to −2.1)				−2.5*
<65 years	1992–2005	−1.1*	(−1.4 to −0.7)	2005–2015	−3.1*	(−3.6 to −2.5)				−3.1*
≥65 years	1992–2005	0.3*	(0.0 to 0.5)	2005–2015	−2.2*	(−2.5 to −1.8)				−2.2*
Non-Hispanic black
All ages	1992–2015	−1.0*	(−1.3 to −0.8)							−1.0*
<65 years	1992–2015	−1.3*	(−1.6 to −1.0)							−1.3*
≥65 years	1992–2015	−0.9*	(−1.1 to −0.6)							−0.9*
Asian/Pacific Islander
All ages	1992–2015	−0.6*	(−1.0 to −0.2)							−0.6*
<65 years	1992–2015	−1.0*	(−1.6 to −0.5)							−1.0*
≥65 years	1992–2006	1.9*	(0.6 to 3.1)	2006–2015	−2.9*	(−4.5 to −1.3)				−2.9*
American Indian/Alaska Native
All ages	1992–2015	−0.0	(−1.1 to 1.0)							−0.0
<65 years	1992–2015	−1.3	(−3.0 to 0.4)							−1.3
≥65 years	1992–2015	0.5	(−0.9 to 1.9)							0.5
Hispanic
All ages	1992–2006	−0.1	(−0.7 to 0.6)	2006–2015	−1.8*	(−2.9 to −0.8)				−1.8*
<65 years	1992–2015	−0.9*	(−1.2 to −0.6)							−0.9*
≥65 years	1992–2015	−0.7*	(−1.2 to −0.3)							−0.7*

AAPC indicates average annual percent change over the most recent data years; APC, annual percent change; CI, confidence interval. Based on incidence rates age adjusted to the 2000 US standard population.

^*The APC or AAPC is significantly different from zero (p < 0.05)

Source: NCHS, 2017. Data for American Indians/Alaska Natives are based on Contract Health Service Delivery Area (CHSDA) counties. Data for Hispanics excludes Oklahoma and New Hampshire.

Trends were analyzed using the Joinpoint Regression Program, version 4.5.0.1, allowing up to 4 joinpoints.

Ovarian Cancer Subtypes

Overview

Ovarian cancer encompasses a heterogenous group of malignancies differentiated by cell/site of origin, pathological grade, risk factors, prognosis, and treatment.^{19, 29–31} Epithelial cancers are most common among women of all racial/ethnic groups, accounting for 90% of all cases (Figure 3). Epithelial cancers are classified by tumor cell histology as serous (52%), endometrioid (10%), mucinous (6%), or clear cell (6%), with one-quarter being more rare subtypes or unspecified.³ Epithelial malignancies are further grouped as type I or type II based on clinicopathologic factors, with the primary distinguishing molecular factor being marked genetic instability in type II versus type I.²⁹ Type I ovarian cancers are generally large, unilateral, cystic tumors at diagnosis with indolent behavior. They are thought to usually develop from extraovarian benign lesions that embed in the ovary and subsequently undergo a series of mutations resulting in malignant transformation. In this way, low-grade serous carcinomas are thought to originate from benign deposits of fallopian tube epithelium in the ovaries (endosalpingiosis); endometrioid and clear cell carcinomas from benign foci of endometrial tissue in the ovaries (endometriosis); and most low grade mucinous carcinomas from benign foci of transitional epithelium from the tuboperitoneal junction.²⁹ Type I ovarian cancers are considered low grade, with the exception of clear cell carcinomas, and account for only a small fraction of ovarian cancer deaths.²⁹

Figure 3. Distribution (%) of Ovarian Cancer Cases* by Major Subtype and Race/ethnicity, 2010–2014.

*Data are based on microscopically confirmed cases. Persons of Hispanic origin may be of any race; Asian/Pacific Islanders and American Indians/Alaska Natives include those of Hispanic and non-Hispanic origin. † Data for American Indians/Alaska Natives are based on Contract Health Service Delivery Area (CHSDA) counties.

Source: NAACCR, 2017.

Type II epithelial cancers are high grade and characterized by involvement of both ovaries, aggressive behavior, late stage at diagnosis, and low survival.²⁹ They are thought to originate as fallopian tube fimbriae carcinomas that spread to the ovaries and/or peritoneum.^{29, 32, 33} Women with these cancers often present with extensive extraovarian disease and ascites. Type II cancers are primarily high-grade serous carcinomas, the most common epithelial subtype, but also include carcinosarcomas and undifferentiated carcinomas.²⁹ It is notable that while tumor grade is important in clinical practice, it is not a robust independent prognostic indicator. Moreover, grade is not a reliable metric for population-based cancer epidemiology research because of the high proportion of inaccurately recorded and unknown grade, including up to one-third of epithelial cancers, in cancer registry data (Supplementary Figure 1).³⁴

Non-epithelial cancers are typically less aggressive than epithelial malignancies. Germ cell and sex cord-stromal tumors make up the majority of non-epithelial cancers, but account for only 3% and 2%, respectively, of all ovarian cancers (Figure 3). Sex cord-stromal tumors arise from various connective tissue cell types including granulosa, Sertoli, and/or Leydig cells. Other non-epithelial ovarian cancers include small cell carcinoma (hypercalcemic and non-hypercalcemic types) and ovarian sarcoma.

Epithelial Ovarian Cancer Occurrence

Incidence

Epithelial ovarian cancer incidence varies by age and race/ethnicity. The age distribution for serous carcinoma is older than that of other epithelial subtypes, peaking in the seventh versus the fifth decade of life (Figure 4). NHW women have the highest rates for serous (5.2 per 100,000) and mucinous carcinoma (0.7 per 100,000; Table 3). NHW women also have the highest rate of endometrioid carcinoma (alongside APIs; 1.1 per 100,000). Clear cell carcinoma incidence rates are highest in APIs (1.0 per 100,000), about twice that of women in other racial/ethnic groups. Conversely, APIs have the lowest rates for serous carcinoma (alongside NHBs; 3.4 per 100,000), the most common subtype. Clear cell carcinoma incidence in API women peaks at a younger age (Figure 4). Higher rates of clear cell and endometrioid carcinoma in Asian women have been documented in the United States and Eastern Asia, although reasons remain unknown.^35–38 Endometriosis is most strongly associated with these subtypes;^{19, 30} however, it is unclear whether the relationship is causal or a result of shared risk factors,^39–42 and whether prevalence of this condition is higher among Asian women.⁴³ Other genetic and reproductive/hormonal factors may also contribute.^{19, 44} NHBs have the lowest incidence rates for all epithelial subtypes (Table 3). Reasons for the lower rates among NHBs are unknown, even when accounting for a higher prevalence of known ovarian cancer protective factors such as fallopian tube ligation,²⁰ which is associated with a roughly 30% reduced risk of ovarian cancer.¹⁶

Figure 4. Epithelial Ovarian Cancer Incidence Rates* by Age, Race/ethnicity, and Histology, US, 2010–2014.

*Per 100,000, age adjusted to the 2000 US standard population. Rates based on <6 cases are excluded.

†Persons of Hispanic origin may be of any race; blacks, American Indians/Alaska Natives, and Asian/Pacific Islanders include those of Hispanic and non-Hispanic origin.

Source: NAACCR, 2017.

Table 3.

Ovarian Cancer Incidence Rates^* by Race/ethnicity and Histology, US, 2010–2014

	Epithelial					Sex cord-stromal	Germ cell
	All epithelial subtypes	Serous	Endometrioid	Mucinous	Clear cell
All races	9.4	4.9	1.0	0.6	0.6	0.3	0.4
Non-Hispanic white	10.0	5.2	1.1	0.7	0.6	0.2	0.3
Non-Hispanic black	6.9	3.4	0.5	0.4	0.2	0.5	0.4
American Indian/Alaska Native	8.3	4.3	0.9	0.5	0.4	0.3	0.2
Asian/Pacific Islander	7.8	3.4	1.1	0.6	1.0	0.1	0.4
Hispanic	8.1	4.0	0.8	0.5	0.4	0.2	0.5

^*Per 100,000, age adjusted to the 2000 US standard population.

^†Persons of Hispanic origin may be of any race; blacks, American Indians/Alaska Natives, and Asians/Pacific Islanders include those of Hispanic and non-Hispanic origin. Data for American Indians/Alaska Natives are based on CHSDA counties.

Source: NAACCR, 2017.

Declines in endometrioid and serous carcinoma incidence rates of 3.7% and 1.0% per year, respectively, since 2000 (Table 4) may be linked to reductions in the use menopausal hormone therapy, which is associated specifically with these subtypes.^{18, 19, 23} Mucinous carcinoma incidence decreased by more than 5% annually from 1995 to 2009, likely due to improvements in laboratory methods to recognize metastases to the ovary, which can mimic primary mucinous carcinoma,^{23, 45} and perhaps declines in smoking, which only increases risk for this subtype.⁴⁶ However, incidence of mucinous carcinoma has been stable since 2009, as has clear cell carcinoma since at least 1995. Rates for other epithelial cancers combined have declined rapidly in recent years due to ongoing efforts to improve classification, which may have attenuated decreasing trends for some subtypes. Future trends will be similarly impacted by advances in the understanding of ovarian cancer biology. For instance, because many epithelial cancers are now recognized to originate in the fallopian tubes, high-grade serous carcinoma is increasingly being classified as cancer of the fallopian tube rather than of the ovary,^{47, 48} although the accuracy of this designation requires further research.

Table 4.

Trends in Ovarian Cancer Incidence Rates by Histology, United States, 1995 to 2014

	Trend 1			Trend 2			Trend 3			2005–2014
	Years	APC	95% CI	Years	APC	95% CI	Years	APC	95% CI	AAPC
Epithelial	1995–2000	−1.2*	(−1.8 to −0.6)	2000–2014	−2.1*	(−2.2 to −1.9)				−2.1*
Serous	1995–2000	0.9	(−0.2 to 2.0)	2000–2014	−1.0*	(−1.2 to −0.8)				−1.0*
Endometrioid	1995–2000	−1.5	(−3.1 to 0.2)	2000–2014	−3.7*	(−4.0 to −3.3)				−3.7*
Mucinous	1995–2009	−5.3*	(−5.6 to −5.0)	2009–2014	−0.0	(−1.9 to 1.9)				−2.4*
Clear cell	1995–2014	−0.1	(−0.4 to 0.2)							−0.1
Other	1995–2005	−3.1*	(−3.5 to −2.7)	2005–2008	−0.6	(−5.5 to 4.5)	2008–2014	−5.5*	(−6.4 to −4.7)	−3.9*
Germ cell	1995–2014	0.4	(−0.1 to 0.9)							0.4
Sex cord-stromal	1995–2004	−1.9	(−4.0 to 0.3)	2004–2014	2.3*	(0.5 to 4.0)				2.3*

AAPC indicates average annual percent change over the most recent data years; APC, annual percent change; CI, confidence interval. Based on incidence rates age adjusted to the 2000 US standard population.

^*The APC or AAPC is significantly different from zero (p < 0.05)

Source: NAACCR, 2017.

Trends were analyzed using the Joinpoint Regression Program, version 4.5.0.1, allowing up to 3 joinpoints.

Stage at diagnosis and survival

Stage at diagnosis varies substantially by epithelial subtype (Figure 5). Most serous carcinomas are diagnosed at stage III (51%) or IV (29%), reflecting the aggressive nature of predominant high-grade serous carcinomas. In contrast, the majority (58%−64%) of endometroid, mucinous, and clear cell carcinomas are diagnosed at stage I, similar to non-epithelial tumors. Consequently, the 5-year cause-specific survival for serous carcinoma is 43%, compared with 82%, 71%, and 66% for endometrioid, mucinous, and clear cell carcinoma, respectively (Table 5).

Figure 5. AJCC* Stage Distribution (%) for Ovarian Cancer by Histology, US, 2007–2013.

*6^th edition

Source: SEER 18 Registries, National Cancer Institute, 2017.

Table 5.

Five-year Cause-specific Survival (%) for Ovarian Cancer by Race/ethnicity, AJCC Stage^*, and Histology, US, 2007–2013

	Epithelial					Sex cord-stromal	Germ cell
	All epithelial subtypes	Serous	Endometrioid	Mucinous	Clear cell
All races
All stages	47	43	82	71	66	88	94
Stage I	89	86	95	92	85	98	99
Stage II	71	71	84	69	71	84	93
Stage III	41	42	59	30	35	61	90
Stage IV	20	26	29	13	16	41	69
Non-Hispanic white
All stages	47	47	82	72	67	88	94
Stage I	89	86	94	94	86	100	99
Stage II	72	73	82	69	68	78	84
Stage III	40	42	58	32	36	61	93
Stage IV	20	25	27	10	22	^	^
Non-Hispanic black
All stages	35	36	77	50	39	85	88
Stage I	81	78	93	80	53	95	97
Stage II	58	57	91	^	^	^	^
Stage III	36	40	55	^	^	^	^
Stage IV	13	16	^	^	^	^	^
Asian/Pacific Islander
All stages	57	47	84	79	70	94	95
Stage I	89	80	95	93	87	87	100
Stage II	76	76	89	^	83	83	^
Stage III	45	47	77	^	24	24	83
Stage IV	25	30	^	^	15	15	^
Hispanic
All stages	52	48	82	71	62	90	95
Stage I	91	90	99	90	84	98	99
Stage II	68	65	87	^	^	^	100
Stage III	45	48	49	41	38	^	93
Stage IV	24	31	40	23	0	^	^

^*6th edition

^{^}Statistic not shown due to fewer than 25 cases. American Indian/Alaska Natives not shown due to sparse data.

Source: SEER 18 Registries, National Cancer Institute, 2017.

Survival also varies by race/ethnicity (Table 5). For example, 5-year cause-specific survival for serous carcinoma is 47%−48% in NHW, API, and Hispanic women compared to 36% in NHB women (Table 5). Indeed, serous carcinoma survival is lowest for NHB women for every stage of diagnosis, likely due in part to lower receipt of guideline-adherent treatment.^{22, 49, 50} Further, NHBs have among the lowest stage-specific survival across epithelial cancer subtypes. The relatively high survival in APIs for epithelial cancers overall reflects their low incidence of serous carcinoma, as well as high survival across subtypes (Table 5). Reasons for this survival advantage are unknown, but may include better response to treatment.^{38, 49} Overall five-year cause-specific epithelial cancer survival in NHW women has improved from 40% for cases diagnosed in 1992–1994 to 47% for cases diagnosed in 2007–2013.¹⁰ In contrast, five-year survival in NHB women remained around 35% over the same period. Slower dissemination of treatment advances, including less access to optimal debulking surgery and intraperitoneal chemotherapy,^51–53 may have contributed to the stagnation.

Non-epithelial Ovarian Cancer Occurrence

Compared with epithelial cancers, non-epithelial malignancies have a younger age distribution, especially germ cell tumors (Figure 6). Incidence of sex cord-stromal tumors is highest among NHBs at every age of diagnosis (Figure 6), with an overall rate (0.5 per 100,000) five-fold higher than that among APIs (0.1 per 100,000), who have the lowest rate (Table 3). In contrast, rates of germ cell tumors are highest in Hispanics (0.5 per 100,000) and lowest in American Indians/Alaska Natives (0.2 per 100,000). Whereas incidence of germ cell tumors has remained stable since at least 1995, rates of sex-cord stromal tumors increased by 2.3% annually from 2004 to 2014 (Table 4). Reasons for this increase are unknown because risk factors for these cancers are poorly understood. The majority of sex-cord stromal (64%) and germ cell (57%) tumors are diagnosed at stage I (Figure 5), for which 5-year cause-specific survival is 98% and 99%, respectively (Table 5). Survival for these tumors remains relatively high even for Stage IV disease, at 41% and 69%, respectively.

Figure 6. Non-epithelial Ovarian Cancer Incidence Rates* by Age, Race/ethnicity, and Histology, US, 2010–2014.

*Per 100,000, age adjusted to the 2000 US standard population. Rates based on <6 cases are excluded.

†Persons of Hispanic origin may be of any race; blacks and Asians/Pacific Islanders include those of Hispanic and non-Hispanic origin.

Familial Risk and Risk Reduction

The strongest risk factor for ovarian cancer is a family history of breast or ovarian cancer.⁴⁴ Risk of developing invasive epithelial ovarian cancer is increased by approximately 50% among women with a first-degree relative with a history of ovarian cancer, and by 10% with a first-degree relative with breast cancer.¹⁹ Approximately 18% of epithelial ovarian cancer cases, particularly high-grade serous carcinomas, are estimated to be due to inherited mutations that confer elevated risk, the majority in BRCA1 and BRCA2.⁵⁴ Mutations in BRCA1 and BRCA2 account for almost 40% of ovarian cancer cases in women with a family history of the disease.⁵⁵ Rare moderate penetrance gene mutations for epithelial ovarian cancer include those in BRIP1, RAD51C, and RAD51D.^{56, 57} There are also more common low penetrance gene variants which potentially result in a substantial number of cancer cases because of their ubiquity. The identification of additional genes associated with increased risk, and their potential utility in a clinical setting for risk prediction, continues to evolve. Non-epithelial cancers are often associated with non-BRCA1/2 gene mutations, including in FOXL2 for adult-type granulosa cell tumors and DICER1 for Sertoli-Leydig tumors.^58–60 Due to this relatively high prevalence of identified genetic mutations, the National Comprehensive Cancer Network recommends genetic testing for all women diagnosed with ovarian cancer to inform their medical and reproductive decisions and those of their relatives.⁶¹

Among women with BRCA1 or BRCA2 mutations, the risk of developing ovarian cancer by age 80 is 44% and 17%, respectively.⁶² While these mutations are rare in the general population (less than 1%), they are more common in certain ethnic or geographically isolated groups, such as those of Ashkenazi (Eastern European) Jewish descent (about 2%).⁶³ The United States Preventive Services Task Force (USPSTF) recommends that primary care providers evaluate the potential risk for BRCA1 or BRCA2 mutations among women with a family history of breast, ovarian, tubal, or peritoneal cancer using one of several available screening tools (e.g. BRCAPRO or BOADICEA). Women with a positive screen should receive genetic counseling and, if indicated, BRCA1/2 genetic testing.⁶⁴

For women who are found to carry a BRCA1 or BRCA2 mutation, risk-reducing bilateral salpingo-oophorectomy decreases the risk of ovarian cancer by about 80%⁶⁵ and is recommended by the American College of Obstetricians and Gynecologists (ACOG) and the Society of Gynecologic Oncology (SGO) once childbearing is complete.^{66, 67} Bilateral salpingectomy alone for pre-menopausal high-risk women is not currently recommended; however, clinical trials are underway.⁶⁶ The National Comprehensive Cancer Network (NCCN) also recommends consideration of bilateral salpingo-oophorectomy for women ages 45–50 years who are carriers of BRIP1, RAD51C, and RAD51D mutations.⁶⁸ The ACOG and SGO suggest that women with BRCA1 or BRCA2 mutations consider oral contraceptives, which reduce risk by about 50% among high risk women who have ever used them, for ovarian cancer prevention.^{66, 69, 70} While there has been concern that this would further increase breast cancer risk, to date no clear significantly increased risk with modern oral contraceptive use has been found in these women.^{66, 67}

Lynch syndrome (hereditary nonpolyposis colon cancer) is a rare hereditary condition associated with an increased risk of colorectal, endometrial, ovarian, and other cancers. Families with Lynch syndrome are mainly characterized by a germline mutation in a DNA mismatch repair gene (e.g. MLH1, MSH2, MSH6 or PMS2). Women with Lynch syndrome have approximately an 8% risk of developing ovarian cancer (usually non-serous epithelial tumors) by age 70,^{71, 72} compared to 0.7% in the general population.¹⁵ The NCCN recommends consideration of hysterectomy along with bilateral salpingo-oophorectomy for women with Lynch syndrome who have completed childbearing.⁶⁸

Early Detection and Screening

Epithelial ovarian cancer is usually diagnosed at an advanced stage, as early stages of the disease have no obvious symptoms and to date, the efficacy of screening has not been demonstrated in prospective randomized controlled trials. The most common sign of advanced disease is swelling of the abdomen caused by ascites.⁷³ However, studies indicate that some women experience persistent nonspecific symptoms in the months prior to diagnosis, including back pain, abdominal distension, pelvic or abdominal pain, difficulty eating or feeling full quickly, vomiting, indigestion, altered bowel habits, or urinary urgency or frequency.^{49, 73–75} Women who experience such symptoms daily for more than a few weeks should seek prompt medical evaluation. Women with non-epithelial tumors often present with more specific early signs, including irregular vaginal bleeding.⁷⁶ Sex cord-stromal tumors often produce sex hormones, which may also affect menstruation and/or cause male physical characteristics, such as a deep voice or body hair.⁷⁷

Currently, there is no recommended screening test for ovarian cancer, although large scale randomized clinical studies to identify effective screening modalities are ongoing. The Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial, which assessed the use of transvaginal ultrasound (TVU) and a fixed cutpoint (≥35 U/mL) in the tumor marker CA125 for early detection, did not observe a reduction in ovarian cancer mortality after up to 19 years of follow-up.⁷⁸ The UK Collaborative Trial of Ovarian Cancer Screening, another large randomized trial based in the United Kingdom, evaluated TVU combined with a risk algorithm incorporating changes in CA125 levels and found reduced mortality in average-risk women after 15 years.⁷⁹ However, the use of secondary analysis to reach these results has been a matter of active debate and further years of follow-up may clarify the potential benefits of this screening method.^{80, 81} Based on these studies, the USPSTF continues to recommend against screening for ovarian cancer in the general population, concluding that there is adequate evidence that annual screening does not reduce ovarian cancer mortality and can lead to important harms, mainly surgical interventions in women without ovarian cancer.⁸⁰ Identifying an effective screening method is complicated by accumulating evidence that ovarian cancer, particularly aggressive high-grade serous carcinoma, begins as a microscopic lesion in the fallopian tube that is undetectable with current strategies.²⁹ For women who are at high risk, a thorough pelvic exam in combination with TVU and a blood test for changes in the level of the tumor marker CA125 may be offered, although this strategy has not proven effective in reducing ovarian cancer mortality.^{82, 83}

Conclusion

Ovarian cancer mortality has declined by more than 30% since the mid-1970s due to reductions in incidence and improvements in treatment in recent decades. Still, fewer than half of women survive beyond five years after diagnosis because of the predominance of aggressive high-grade serous carcinomas and the absence of specific early symptoms and effective early detection strategies. Notably, risk factors for high-grade serous carcinoma remain largely unknown, stymying prevention efforts. Survival rates have improved only slightly over the past three decades among NHWs and remained stagnant among NHBs, likely due to differential access to high-quality treatment. Moreover, NHB women experience the lowest survival for almost every stage of diagnosis across cancer subtypes. Further research is needed to more specifically determine the reasons for this disparity and to advance understanding of the disease in order to identify modifiable risk factors, develop effective early detection methods, and improve treatment.