Oophorectomy-Corrected Ovarian Cancer Incidence, Survival, and Mortality by Subtype, Race, Ethnicity.

Nicolas Wentzensen: conceptualization (lead), formal analysis (equal), investigation (equal), methodology (equal), supervision (equal), writing – original draft (equal), writing – review and editing (equal). Camryn Cohen: data curation (equal), formal analysis (equal), investigation (equal), methodology (equal), software (equal), writing – review and editing (equal). Summer Harvey: formal analysis (equal), investigation (equal), methodology (equal), software (equal), writing – review and editing (equal). Britt K. Erickson: formal analysis (equal), investigation (equal), methodology (equal), writing – review and editing (equal). Megan A. Clarke: conceptualization (equal), data curation (equal), formal analysis (equal), methodology (equal), supervision (equal), writing – review and editing (equal).

This work was supported by the Intramural Research Program of the National Cancer Institute (Z01 CP010126).

From the SEER22 database ( seer.cancer.gov ), including 22 population‐based cancer registries, representing 48% of the U.S., we obtained incidence data for microscopically confirmed first malignant primary ovarian tumors diagnosed between 2000 and 2019 (excluding 2020 cases because of COVID). In addition to primary ovarian tumors (C56.9), we conducted analyses combining ovarian, fallopian tube (C57.0), and peritoneal carcinomas (C48.2). We excluded cases among women aged < 15 years and those diagnosed by autopsy or death certificate only. For analyses of incidence‐based mortality and 5‐year relative survival, cases were identified in the SEER17 database, representing 27% of the U.S. population, using the same selection criteria; at the time of analysis, these data were not available in all SEER22 cancer registries. We classified histologic subtypes (histotypes) according to the third edition of the International Classification of Diseases for Oncology as serous, endometrioid, mucinous, clear cell, other epithelial, and other non‐epithelial cancers (Table  S1 ). Data on grade was highly incomplete and unreliable among cases with grade information and therefore could not be used for subtype classification. We classified stage at diagnosis using Summary Stage 2000 as localized, regional, distant, or unstaged/unknown. Stage data were available in SEERStat as of 2004, therefore analyses by stage exclude cases diagnosed in 2000–2003. Race and ethnicity data were abstracted from medical records and grouped by SEER into race and origin categories. Based on available groupings, we reported data separately for Non‐Hispanic White, Non‐Hispanic Black, Hispanic, American Indian/Alaskan Natives, and Asian/Pacific Islanders. This study was exempt from institutional review board approval. The National Health and Nutrition Examination Survey (NHANES) is a nationally representative cross‐sectional survey that assesses the health and nutritional status of U.S. noninstitutionalized civilians [ 18 ]. Survey‐weighted, modeled oophorectomy prevalence estimates for each 2‐year cycle were calculated for adult women over 18 years of age, adjusted for age, year, and race and ethnicity. Oophorectomy was defined as having both ovaries removed, self‐reported in the Reproductive Health Questionnaire survey section. Prevalence estimates were stratified by age grouping (18–39, 40–49, 50–59, 60–69, 70–79, and 80+) and race and ethnicity (Non‐Hispanic White, Non‐Hispanic Black, and Hispanics). Given that oophorectomy prevalence did not change meaningfully during the time period, for corrections by race and age group, oophorectomy estimates were collapsed into two groups, combining data over multiple years (up to 2006 and after 2006). Following NHANES recommendations, Hispanic‐specific estimates were not calculated prior to 2007 [ 19 ]. Oophorectomy rates were not separately reported for other racial/ethnic groups. Age‐adjusted incidence rates, uncorrected and corrected for oophorectomy prevalence, were calculated using SEER*Stat software (version 8.4.4) for ovarian cancer overall and by histologic subtype, race and ethnicity, stratified by six age groups (15–39, 40–49, 50–59, 60–69, 70–79, and 80+). Rates were age‐adjusted to the 2000 US standard population and expressed per 100,000 woman‐years. Trends in ovarian cancer incidence were estimated using the NCI Joinpoint regression software (version 5.0.2), which estimates annual percentage changes (APCs) with 95% CIs and t ‐tests to determine whether APCs are statistically significantly different from zero. An average APC (AAPC) for the whole time period was reported for all comparisons. Multiple segments with different statistically significant APCs were additionally reported. From the SEER17 database, we evaluated 5‐year relative survival rates among women diagnosed with ovarian cancer by histology (serous, endometroid, mucinous, clear cell, other) from 2000 to 2018 with follow‐up through 2019. We estimated 5‐year relative survival as the ratio of the observed to the expected survival rate of patients using the actuarial method; expected survival was estimated using the Ederer II method [ 16 , 20 ]. Cause‐specific mortality was estimated for ovarian cancer deaths between 2000 and 2019 using death certificate data from the National Center for Health Statistics (NCHS; www.cdc.gov/nchs ) provided to the SEER program. However, these data do not contain detailed information on tumor‐level characteristics such as the histotype of the cancer cases. Incidence‐based mortality analyses in SEER allow linking incidence data, including tumor characteristics, with mortality data from NCHS. However, since these data require linking incidence data from a more limited source than the mortality database, the number of mortality outcomes is typically lower compared to an overall mortality analysis. Despite the underestimation of mortality overall, incidence‐based mortality is critical to compare mortality estimates by tumor characteristics. To explore mortality rates by histotype, we evaluated incidence‐based mortality using data from the SEER17 database, which links ovarian cancer cases to deaths [ 21 ]. To allow for adequate follow‐up time after case diagnosis and avoid the underestimation of mortality rates in the years immediately following cancer diagnosis, the analysis was restricted to deaths diagnosed from 2010 to 2019.

Among 158,725 cases reported between 2000 and 2019, serous ovarian carcinoma was the most common histotype (67.8%), followed by endometrioid (9.4%), non‐epithelial (6.8%), mucinous (6.1%), clear cell (5.6%), and other epithelial cancers (4.3%) (Table  1 ). Histotype distributions were similar across racial/ethnic groups with notable exceptions: Compared to Non‐Hispanic White women, Non‐Hispanic Black women had a lower proportion of endometrioid (6.1%) and clear cell carcinomas (3.1%), and a higher proportion of non‐epithelial cancers (13.7%). Hispanic women had a lower proportion of serous carcinomas (63.0%) and a higher proportion of non‐epithelial cancers (11.9%). Asian/Pacific Islander women had the lowest proportion of serous carcinomas (55.1%) and the highest proportion of endometrioid (11.7%) and clear cell carcinomas (12.7%), respectively. Ovarian cancer incidence by race/ethnicity and histotype based on SEER22 from 2000 to 2019. Increase is calculated based on the corrected/uncorrected estimates. Age‐adjusted prevalence of oophorectomy overall changed minimally from 11.3% between 1999 and 2006 to 10.6% between 2007 and 2020 (Table  S2 ). No substantial changes in oophorectomy prevalence were observed in age groups, but estimates had wide confidence intervals reflecting small sample sizes within many strata (Table  S3 ). Oophorectomy prevalence ranged from 1.3% at age 18–39 to 27% at age 70–79. In most age strata, oophorectomy rates were highest for Non‐Hispanic White women. Overall, the uncorrected, age‐adjusted ovarian cancer incidence rate between 2000 and 2019 was 12.4 (95% CI: 12.3–12.5) per 100,000 which increased to 15.2 (95% CI: 15.1–15.4) after correcting for oophorectomy (Table  1 , Figure  1A ). Corrected incidence rates were highest for serous ovarian cancer (10.4, 95% CI: 10.3–10.5), followed by endometrioid (1.4, 95% CI: 1.4–1.4), clear cell (0.8, 95% CI: 0.8–0.8), mucinous (0.9, 95% CI: 0.9–0.9), other epithelial histotypes (0.7, 95% CI: 0.6–0.7), and non‐epithelial cancers (1.0, 95% CI: 1.0–1.0). Corrected rates of ovarian cancer were highest in Non‐Hispanic White women (16.6, 95% CI: 16.5–16.7) followed by Hispanic (12.9, 95% CI: 12.8–13.2) and Non‐Hispanic Black women (11.8, 95% CI: 11.7–12.1). Rates were highest in Non‐Hispanic White women for all epithelial histotypes except non‐epithelial ovarian cancer rates which were highest among Non‐Hispanic Black women. Uncorrected and oophorectomy‐corrected ovarian cancer incidence overall (A) and oophorectomy‐corrected incidence by race and ethnicity (B). *Indicates that the Average Annual Percent Change (AAPC) is significantly different from zero at the alpha = 0.05 level. The oophorectomy‐corrected ovarian cancer incidence was 17.6/100,000 in 2000 and 12.3 in 2019, decreasing at an average APC of 2.1% per year (−2.1 to −1.9) between 2000 and 2019. Rates decreased significantly by 1.5% per year (−1.8 to −0.2) between 2000 and 2008 and by 3.6% (−4.7 to −2.2) between 2008 and 2011, followed by stable trends between 2011 and 2015, and a sharp decline of 4.7% (−4.7%, −6.2 to −3.7) between 2015 and 2019. While uncorrected incidence trends were lower, the slopes of the two curves were similar. Oophorectomy‐corrected ovarian cancer incidence trends varied across racial/ethnic groups (Figure  1B ): Among Non‐Hispanic White women, a significant decrease in ovarian cancer incidence was observed between 2000 and 2019 (−2.3%, −2.6 to −2.0), with a modest decrease between 2000 and 2015 (−1.8%, −2.0 to −1.3) and a stronger decrease between 2015 and 2019 (−3.9%, −7.3 to −2.4). Among Non‐Hispanic Black women, there was a slight decrease between 2000 and 2019 (−1.1%, −1.5 to −0.7) with stable incidence rates between 2000 and 2016 and a strong decrease between 2016 and 2019 (−5.1%, −9.6 to −2.1). Among Hispanic women, there was a modest decrease between 2000 and 2019 (−1.1%, −1.5 to −0.6). Similar trends were observed for serous carcinoma rates (Figure  S1 ), with the strongest decreases observed for Non‐Hispanic White women (−2.6%, −3.1 to −2.3), while trends for non‐serous carcinomas showed a slower decline between 2000 and 2019 (−1.3%, −1.7 to −1.0). Overall, 34,932 of 158,725 cases (22%) were diagnosed below 50 years of age. Incidence trends among women 50 years and older mirrored those of the overall trends, declining from 39.3 in 2000 to 26.3 in 2019 (−2.3%, −2.5 to −2.1) (Figure  2 ). In women under 50 years, the decline in oophorectomy‐corrected incidence was small, from 5.8 in 2000 to 4.7 in 2019 (−0.9%, −1.2 to −0.7). Oophorectomy‐corrected ovarian cancer incidence trends by age. *Indicates that the Average Annual Percent Change (AAPC) is significantly different from zero at the alpha = 0.05 level. Oophorectomy‐corrected fallopian tube cancer incidence increased substantially from 2000 (0.6/100,000) to 2019 (1.7/100,000) (average APC 7.0%, 6.1–8.2) with a particularly strong increase starting in 2016. Oophorectomy‐corrected peritoneal cancer incidence rates remained stable at around 0.8/100,000 (Figure  3 ). Oophorectomy‐corrected incidence of primary ovarian, fallopian tube, and peritoneal carcinomas combined was 19.1/100,000 in 2000 and 14.9/100,000 in 2019, with an annual decrease of 1.3% (−1.5 to −1.0) (Figure  4A ). These rates declined between 2000 and 2019 by 1.5% (−1.6 to −1.3) in Non‐Hispanic White women and by 0.6% (−1.0 to −0.1) in Hispanic women. Among Non‐Hispanic Black women, rates were stable until 2016, followed by an annual decrease of 3.4% (−7.7 to −0.7) (Figure  4B ). Oophorectomy‐corrected fallopian tube and peritoneal cancers incidence trends. *Indicates that the Average Annual Percent Change (AAPC) is significantly different from zero at the alpha = 0.05 level. Incidence trends of uncorrected and oophorectomy‐corrected ovarian, fallopian tube and peritoneal cancers combined overall (A) and by race/ethnicity (B). *Indicates that the Average Annual Percent Change (AAPC) is significantly different from zero at the alpha = 0.05 level. Relative 5‐year survival was 43% (95% CI: 42.7–43.5) among all women diagnosed with ovarian cancer with substantial variation by race/ethnicity and histotype (Table  2 ). Survival was highest for Asian/Pacific Islander women (48%, 95% CI: 46.1–49.3), and lowest for Non‐Hispanic Black women (32%, 95% CI: 30.8–33.7). Relative 5‐year survival was highest for endometrioid (77%, 95% CI: 75.4–78.8) and clear cell (67%, 95% CI: 64.4–69.6) histotypes, followed by mucinous (57%, 95% CI: 54.9–58.9), and other epithelial (58%, 95% CI: 55.9–60.7) histotypes with lowest survival observed for serous histotypes (35%, 95% CI: 34.8–35.7). Non‐Hispanic Black women and American Indian/Alaska Native women had the lowest relative 5‐year survival for all histotypes, with a particularly low 5‐year survival observed for serous carcinoma in Non‐Hispanic Black women (25%, 95% CI: 23.5–26.5). In contrast, Non‐Hispanic White women and Asian/Pacific Islander women had the highest survival for all histotypes. In analyses stratified by stage, endometrioid ovarian carcinomas had better survival within all stages compared to all other histotypes. Remarkably, while for all non‐serous histotypes, most cancers were diagnosed at local or regional stage, 79% of serous carcinomas were diagnosed at distant stage (Table  S4 ). Non‐Hispanic Black women had the worst outcomes in all stages. Ovarian cancer survival by subtype and race/ethnicity based on SEER17 from 2000 to 2019. Oophorectomy‐corrected ovarian cancer mortality rates declined from 14.0/100,000 to 9.9/100,000 between 2000 and 2019 (−1.8%, −2.1 to −1.6) trailing decreases in incidence by approximately 1 year (Figure  5 ). Oophorectomy‐corrected mortality was 13.5/100,000 for Non‐Hispanic White women, 10.4/100,000 for Non‐Hispanic Black women, and 8.4/100,000 for Hispanic women (Table  S5 ). Uncorrected and oophorectomy‐corrected ovarian cancer mortality trends. *Indicates that the Average Annual Percent Change (AAPC) is significantly different from zero at the alpha = 0.05 level. To evaluate mortality by ovarian histotype, we conducted an incidence‐based mortality analysis (Table  S6 ). While serous histotypes contributed to 68% of incident cases, they were responsible for 83% of ovarian cancer deaths. The oophorectomy‐corrected incidence‐based mortality rate was 7.5/100,000 overall, and 6.3/100,000 for serous carcinomas (Table  S6 ).

We conducted the first comprehensive evaluation of ovarian cancer incidence and mortality accounting for bilateral oophorectomy. We demonstrate that oophorectomy correction increased ovarian cancer incidence by 23% and ovarian cancer mortality by 27%, with strongest increases in Non‐Hispanic White women. Since oophorectomy rates changed minimally over time, trends were not substantially affected by oophorectomy correction. From 2000 to 2019, we observed a small decrease in oophorectomy‐corrected ovarian cancer incidence, with a stronger reduction since 2016. This decline was strongest for Non‐Hispanic White women. There is increasing evidence that many serous cancers may originate in the fallopian tubes. We assessed trends in fallopian tube cancers and observed increasing rates, with the most pronounced rise since 2015 [ 22 ]. This recent increase of fallopian tube cancers parallels the recent decrease of ovarian cancer incidence. High‐grade serous ovarian cancers frequently involve many sites in the pelvic area, including the ovaries, fallopian tubes, adnexae, the peritoneum and sometimes the endometrium. Over the last decade, many pathology laboratories have introduced more detailed pathology protocols of gynecologic pathology specimens, including the Sectioning and Extensive Examination of the Fimbria (SEE‐Fim) protocol which can increase detection of malignancies in the fallopian tubes [ 23 ]. It is likely that some cancers that previously would have been classified as ovarian cancers are now classified as primary fallopian tube cancers. Notably, this observation was specific to fallopian tube cancers, while the incidence of peritoneal cancers did not change between 2000 and 2020. When combining ovarian, fallopian tube, and peritoneal cancers, the decrease in ovarian cancer incidence starting around 2015 was attenuated to a 1.3% reduction per year. Our data demonstrate that it is important to consider all three sites in descriptive analyses of ovarian cancer, as suggested by the 2020 WHO recommendation on classification of female genital tumors [ 9 ]. Increasingly, in epidemiological studies, ovarian, fallopian, and peritoneal cancers are grouped into a single category, reflecting their likely joint origin [ 24 , 25 ]. Even after accounting for oophorectomy and potential reclassification of cancers by site, a 1.3% annual reduction of ovarian cancer incidence remained since 2002. It has been proposed that the strong decrease in MHT use after the announcement of results from the Women's Health Initiative trial of menopausal hormone therapy (MHT) use in 2002 could explain a subsequent decrease in ovarian cancer incidence [ 10 ]. However, a sharp drop in MHT use observed between 2002 (29.5%) and 2008 (9.8%) was followed by a more constant use level in the population [ 26 ]. These substantial changes in use patterns do not seem to align with the steady reduction in ovarian cancer incidence since 2002. Further, we also observed a decrease in incidence among women under 50 years, where MHT use is extremely rare. With the recent revisiting of MHT use recommendations based on more favorable risk–benefit analyses, it will be important to closely monitor ovarian cancer incidence rates when MHT use increases [ 27 ]. Use of contraceptive approaches has changed substantially over the last two decades, with changed OC formulations and increasing use of intrauterine devices that have no established relationship with ovarian cancer risk [ 28 ]. However, a reduction of OC use, a protective factor for ovarian cancer, would not explain the decreasing incidence rates. Regular aspirin use is associated with reduced risk of ovarian cancer, but aspirin use patterns in the time window of this analysis have been stable [ 29 ]. For other drugs that have shown increased use over the last two decades, like metformin or statins, there is currently no established association between their use and ovarian cancer risk [ 30 ]. While increased genetic testing and increasing uptake of prophylactic surgery for high penetrance mutations may lead to a reduction of ovarian cancer incidence in the population, the prevalence of these mutations and their attribution to ovarian cancer in the general population is low [ 2 ]. Other factors that may affect ovarian cancer incidence trends during the last two decades include changing use of personal hygiene products, changes in oral contraceptive use and formulations, and increasing use of long‐acting reversible contraceptives, among other factors [ 25 , 31 , 32 , 33 ]. Our analysis emphasizes the importance of serous ovarian carcinoma as the most common and most fatal ovarian cancer subtype. Serous histotypes account for 68% of ovarian cancer incidence, but 83% of ovarian cancer deaths, due to late‐stage detection and poor survival. Thus, focus of prevention and early detection efforts should be on serous ovarian cancers to impact ovarian cancer mortality. A striking finding was the stage distribution of serous ovarian cancer, with almost 80% of cases diagnosed with distant metastases. This distribution differs remarkably from other ovarian cancer histotypes and from most cancers in general. We observed notable differences in ovarian cancer incidence, survival, and mortality by race and ethnicity. Incidence and mortality were highest among Non‐Hispanic White women who also had the highest incidence of serous ovarian carcinoma. Non‐epithelial subtypes were most common among Non‐Hispanic Black women, while endometrioid and clear cell subtypes were most common among Asian or Pacific Islander women. Global studies have noted higher incidence of clear cell and endometrioid histotypes in Asia, but it is not clear whether genetic or environmental factors may play a role [ 34 , 35 ]. Endometriosis is an established risk factor for endometrioid and clear cell subtypes, and some studies suggest that Asian women may have a higher prevalence of endometriosis [ 36 ]. Ovarian cancer survival was lowest among Non‐Hispanic Black women overall and after accounting for subtype and stage. This finding is in line with reports from many other cancer sites showing low survival among Non‐Hispanic Black individuals after accounting for cancer subtype and stage, pointing to disparities likely related to access to care, access to clinical trials, among others [ 1 , 16 , 20 ]. Serous ovarian and endometrial cancers have similar molecular characteristics, including p53 mutations, chromosomal instability, and poor outcomes. The decreasing serous ovarian cancer incidence and mortality are in stark contrast to rapidly rising serous endometrial cancer incidence and mortality, particularly among Non‐Hispanic Black women [ 16 , 17 , 21 ]. This is puzzling, since we would expect similar exposures driving serous carcinomas that manifest in the ovaries and the uterus. This suggests that either cells of origin, exposures, and carcinogenic pathways are different between the two sites, or that other factors are at play that influence the progression of initially transformed cell clones differently at both cancer sites. Our analysis has several strengths, including the use of national cancer registry data that cover almost half of the U.S. population, and nationally representative oophorectomy data from the NHANES survey. The incidence‐based mortality approach allowed to estimate ovarian cancer mortality by subtype, which is otherwise not possible with mortality data. We note some important limitations: While nationally representative, the number of individuals included in NHANES with oophorectomy information was small, limiting the oophorectomy correction to Non‐Hispanic White, Non‐Hispanic Black, and Hispanic women. Further, information on salpingectomy without oophorectomy was not available from nationally representative surveys, but prevalence is likely low. Race and ethnicity are self‐reported both in SEER and NHANES, and were grouped in different categories that change over time. The lack of disaggregated data for Asian American and Pacific Islander populations in SEER may mask important differences between these populations [ 37 ]. Further, the reporting of histologic grade information in SEER22 was limited, with over 30% missing data, a lower than expected proportion of low‐grade serous tumors among those with grade information (only 5%), and a lower than expected 5‐year survival among cases without grade information, suggesting that grade was not missing at random. Therefore, we could not use grade for subtype classification, which particularly affects the distinction between low‐grade and high‐grade serous carcinomas, which have different histology and clinical behavior. Also, the histologic data represent local practice and are not formally harmonized or undergo a centralized rereview to ensure reliable, comparable histology coding. In summary, we show a decrease of ovarian cancer incidence and mortality that was also observed in individuals below 50 years. Non‐Hispanic Black women had the lowest survival within strata of subtype and stage, suggesting disparities along the ovarian cancer continuum of care. We see evidence of reclassification of serous carcinomas from primary ovarian to primary fallopian tube sites. As opposed to uterine cancer, we did not observe an increase in serous ovarian carcinomas, challenging the notion that serous cancers at both sites have similar cells of origin and biology. The reasons for the continued decrease of ovarian cancer incidence observed since 2002 are not understood. The fact that almost 80% of serous carcinomas are diagnosed at advanced stages explains the poor outcomes and points to the potential for early detection to reduce ovarian cancer mortality.

Ovarian cancer is the second most common, and most case‐fatal gynecological cancer in the US [ 1 ]. About 90% of ovarian cancers are of epithelial origin. Ovarian cancer is often detected late when systemic spread has already occurred. While effective prevention efforts based on prophylactic bilateral salpingo‐oophorectomy exist for the few individuals with high genetic or familial risk, there are currently no general population prevention or screening options [ 2 , 3 , 4 ]. Histologic subtypes of ovarian cancer (histotypes) have different molecular [ 5 ], risk factors [ 6 ], and clinical outcomes [ 7 ]. High grade serous carcinoma, the most fatal histotype, accounts for most cases, while remaining epithelial cancers include a mixed group of histotypes [ 8 ]. The profound heterogeneity of ovarian cancers may reflect different cells of origin and different carcinogenic pathways. There is growing evidence that a proportion of high grade serous ovarian carcinomas may originate in the fallopian tubes and some endometrioid and clear cell ovarian carcinomas may originate from endometrial cells or endometriosis lesions. This is reflected in the most recent WHO classification of epithelial ovarin tumors, which includes the ovaries, the fallopian tubes, and the peritoneum as sites of origin and distinguishes five major epithelial tumor types based on morphological assessment and grade (high‐grade serous, low‐grade serous, mucinous, endometrioid, clear cell) and two rare epithelial types (carcinosarcoma, malignant Brenner tumor) [ 9 ]. Evaluating ovarian cancer population trends by histotype improves understanding of the disease burden over time and contextualizes trends with important exposures. Previous studies have shown a decrease in ovarian cancer incidence [ 10 , 11 ]. However, these studies have not corrected incidence rates for oophorectomy prevalence, which can underestimate incidence in the population at risk and could bias comparisons over time and across groups. Reasons for oophorectomy in premenopausal patients include ovarian mass, endometriosis, ovarian torsion, pelvic or other gynecologic malignancy, estrogen dependent malignancy (such as estrogen‐receptor‐positive breast cancer), or cancer risk reduction in patients with BRCA 1/2 or other high risk germline mutations [ 12 ]. Additionally, peri‐ and postmenopausal patients often undergo oophorectomy at the time of benign hysterectomy [ 13 ]. While two prior studies reported oophorectomy‐corrected ovarian cancer incidence, they relied on incomplete data on oophorectomy or were restricted to hospital claims from a single state [ 14 , 15 ]. Strong increases in uterine cancer incidence have been reported recently, including among women under age 50, and particularly for serous uterine cancers [ 16 , 17 ]. Given biologic similarities of serous cancers at both sites, including frequent somatic TP53 mutations, it is important to evaluate whether similar increases are observed for ovarian cancer. We conducted an analysis of oophorectomy‐corrected ovarian cancer incidence and incidence‐based mortality, as well as survival in the Surveillance, Epidemiology, and End Results (SEER) database.

The authors declare no conflicts of interest.

Figure S1A: Oophorectomy‐corrected serous ovarian carcinoma incidence trend. *Indicates that the Average Annual Percent Change (AAPC) is significantly different from zero at the alpha = 0.05 level. Figure S1B: Oophorectomy‐corrected serous ovarian carcinoma incidence trends by race/ethnicity. *Indicates that the Average Annual Percent Change (AAPC) is significantly different from zero at the alpha = 0.05 level. Table S1: Classification of ovarian cancer histotypes according to ICD‐O codes. Table S2: Modeled oophorectomy prevalence by year. Table S3: Modeled oophorectomy prevalence by year and age group. Table S4: Ovarian cancer survival by stage. Table S5: Uncorrected and oophorectomy‐corrected ovarian cancer mortality by race/ethnicity. Table S6: Uncorrected and oophorectomy‐corrected incidence‐based mortality by ovarian histotype.