Incidence and treatment outcomes of ovarian carcinosarcoma from the national cancer registry of Korea

Abstract

Objective

To investigate the incidence and survival outcomes of ovarian carcinosarcoma in Korea between 1999 and 2018.

Methods

Patients diagnosed with ovarian carcinosarcoma between 1999 and 2018 were identified from the Korea Central Cancer Registry (KCCR) and their information was collected. Age-standardized incidence rates (ASRs), annual percent changes (APC), and relative survival rates of ovarian carcinosarcoma were calculated and compared to those of epithelial ovarian cancer.

Results

According to the KCCR, 458 cases of ovarian carcinosarcoma were detected, and accounted for 1.5% (458/30,679) of all epithelial ovarian cancers in Korea between 1999 and 2018. The ASR of ovarian carcinosarcoma between 1999 and 2018 was 0.064 per 100,000 women. The incidence rate of ovarian carcinosarcoma increased during the study period, with an ASR of 0.029 per 100,000 in 1999 and 0.073 per 100,000 in 2018. The APC of ovarian carcinosarcoma during 1999–2018 was 5.86 (p<0.001). The median overall survival (OS) of patients with ovarian carcinosarcoma was 39 months, and the 5-year OS rate was 42.5%. Among ovarian carcinosarcomas, patients with localized stages showed better clinical outcomes than those with regional or distant stages (5-year OS, 60.8%, 57.9%, and 32.8%, respectively; p<0.001). In addition, younger (<50 years) patients showed better OS than older (≥50 years) patients (5-year OS, 52.6% vs. 40.2%; p<0.001).

Conclusion

Our nationwide registry-based study demonstrated that the incidence of ovarian carcinosarcoma increased from 1999 to 2018 in Korea. Patients with advanced-stage disease and older age (≥50 years) had poorer survival outcomes.

Synopsis

We investigated the incidence and survival outcomes of ovarian carcinosarcoma, by using the Korea Central Cancer Registry (1999–2018). The age-standardized incidence rates of ovarian carcinosarcoma were 0.064/100,000. The median overall survival of patients with ovarian carcinosarcoma was 39 months.

INTRODUCTION

Ovarian cancer is the second most common gynecologic cancer and the leading cause of death from gynecologic malignancies in the United States [1]. Epithelial ovarian cancer accounts for the majority of ovarian cancers and has several histological subtypes. Ovarian carcinosarcoma, previously known as a malignant mixed Müllerian tumor, is a rare subtype of epithelial ovarian cancer and accounts for only 1%–4% of all ovarian cancers [2]. Carcinosarcomas are of epithelial and monoclonal cell origin, and must be distinguished from pure ovarian sarcomas [3]. Carcinosarcomas are biphasic malignant neoplasms composed of high-grade carcinomatous and sarcomatous elements, and are categorized into homologous and heterologous subtypes based on these sarcomatous elements. The homologous subtype comprises native ovarian tissue and consists of endometrial stromal sarcoma, fibrosarcoma, or leiomyosarcoma. Heterologous elements are tissues that are non-native to the ovary, including cartilage, bone, adipose tissue, and smooth or striated muscles [4]. These tumors are often at a high stage at the time of diagnosis, and typically occur in postmenopausal women aged up to 60 years, with a very poor overall prognosis [5].

Accurate information regarding ovarian carcinosarcoma is needed to improve its clinical outcomes and treatment. However, because it is an uncommon disease, most studies have been case reports, case series, or original articles with a small number of patients.

Therefore, this study aimed to report the incidence of carcinosarcoma using a nationwide registration database. In addition, we report the clinical outcomes of carcinosarcoma.

MATERIALS AND METHODS

This study used data from patients enrolled in the population-based Korea Central Cancer Registry (KCCR), which has collected information on approximately 98% [6] of all cancer cases in Korea since 1999. All patients diagnosed with ovarian carcinosarcoma between 1999 and 2018 were included and followed-up until December 31, 2019. In addition, patients diagnosed with epithelial ovarian cancer in the same period were examined. To confirm the validity of the individual vital statuses used in the survival calculation, the Korea National Cancer Incidence Database was linked to both mortality and population resident registration data, which were obtained from the Ministry of the Interior and Safety.

According to the cancer registration guidelines, the first date of diagnosis, diagnosis route, primary cancer site, histological diagnosis, Surveillance, Epidemiology, and End Results (SEER) summary stage, laterality, differentiation, diagnostic method, and initial treatment (within 4 months after diagnosis) can be identified in the registry. Ovarian carcinosarcoma and epithelial ovarian cancer (C56.9) were defined based on the International Classification of Diseases for Oncology, 3rd edition [7]. Patients with pure ovarian sarcoma histology, such as leiomyosarcoma, liposarcoma, or fibrosarcoma, were excluded. Patients were grouped according to their age at disease diagnosis (<30, 30–39, 40–49, 50–59, 60–69, 70–79, and ≥80 years), treatment modality, and SEER summary stage. SEER stages at diagnosis comprised localized (invasive cancer limited to the organ of origin), regional (tumor extension beyond the limits of the organ of origin), distant disease (spread to distant areas from the primary tumor), and unknown stage.

The age-standardized incidence rates (ASRs)/100,000 women using Segi’s world standard population [8], and the annual percent changes (APCs) to estimate trends in the incidence rates were calculated for ovarian carcinosarcoma and epithelial ovarian cancer.

Differences in the characteristics of patients with ovarian carcinosarcoma and epithelial ovarian cancer were evaluated using Student’s t-test for continuous variables and Pearson’s χ² test for categorical variables. Survival curves were generated using the Kaplan-Meier method [9] and compared using the log-rank test according to age, SEER stage, and histological type.

All statistical tests were 2-tailed, and a p<0.05 was considered statistically significant. Incidence and survival analyses were performed using SAS ver. 9.4 (SAS Institute, Inc., Cary, NC, USA) and STATA ver. 16 (StataCorp LLC, College Station, TX, USA), respectively.

The study protocol was reviewed and approved by the Institutional Review Board of the National Cancer Center (NCC2021-0075); the requirement for informed consent was waived as this was a secondary analysis of the de-identified data.

RESULTS

According to the KCCR, 458 ovarian carcinosarcoma cases were detected in Korea between 1999 and 2018. The ASR of ovarian carcinosarcoma between 1999 and 2018 was 0.064 per 100,000 women. The incidence rate of ovarian carcinosarcoma increased during this period, with slight fluctuations. Seven cases were registered, and the ASR was 0.029 per 100,000 in 1999; 33 cases were registered, and the ASR was 0.073 per 100,000 in 2018. The APC of ovarian carcinosarcoma between 1999 and 2018 was 5.86 (p<0.001).

A total of 30,639 cases of epithelial ovarian cancer were detected, and ovarian carcinosarcoma accounted for 1.5% (458/30,639) of the epithelial ovarian cancers. The ASR of epithelial ovarian cancer between 1999 and 2018 was 4.504 per 100,000 women, and the APC was 2.74 (p<0.001). The ASRs and number of cases for each year are listed in Table 1.

Table 1. The incidence of ovarian carcinosarcoma from the Korea Central Cancer Registry, 1999–2018.

Year	Ovarian carcinosarcoma		Epithelial ovarian cancer		ASR ratio
	ASR per 100,000 women	Cases	ASR per 100,000 women	Cases
1999	0.029	7	3.550	931	0.008
2000	0.037	10	3.511	952	0.010
2001	0.030	8	3.457	961	0.009
2002	0.035	10	3.559	1,018	0.010
2003	0.048	14	3.780	1,095	0.013
2004	0.045	13	3.776	1,137	0.012
2005	0.031	9	3.937	1,216	0.008
2006	0.058	18	4.040	1,286	0.014
2007	0.067	23	4.463	1,455	0.015
2008	0.078	29	4.226	1,433	0.019
2009	0.044	17	4.154	1,450	0.011
2010	0.066	26	4.578	1,602	0.014
2011	0.084	32	4.540	1,634	0.018
2012	0.070	27	4.853	1,798	0.014
2013	0.077	33	4.766	1,822	0.016
2014	0.056	25	5.168	1,988	0.011
2015	0.082	36	5.046	1,995	0.016
2016	0.100	43	5.544	2,236	0.018
2017	0.093	45	5.515	2,250	0.017
2018	0.073	33	5.696	2,380	0.013
1999–2018	0.064	458	4.504	30,639	0.014
APC (%)	5.86 (p<0.001)		2.74 (p<0.001)

APC, annual percent change; ASR, age-standardized incidence rate.

The characteristics of the patients with ovarian carcinosarcoma are described in Table 2. The mean age at diagnosis of ovarian carcinosarcoma was 60.2 years and the mean follow-up duration from disease registration was 4.1 years. Ovarian carcinosarcoma was most frequently diagnosed at the ages of 50–59 years (30.4%), followed by 60–69 years (24.7%), and 70–79 years (21.8%). In addition, ovarian carcinosarcoma was most frequently diagnosed at the distant stage (61.2%), followed by the regional (20.2%), and local (15.5%) stages. Surgical treatment was performed on most of the patients with ovarian carcinosarcoma (97.6%), and 74.7% of the patients received chemotherapy. The age distribution, stage distribution, and treatment modality of ovarian carcinosarcoma were significantly different from those of epithelial ovarian cancer (Table 2).

Table 2. Clinicopathologic characteristic of patients with ovarian carcinosarcoma and epithelial ovarian sarcoma in Korea, 1999–2018.

Characteristic			Ovarian carcinosarcoma	Epithelial ovarian cancer	p-value
Follow-up from registration (yr)			4.11±4.51	5.78±5.15	<0.001
Age at diagnosis			60.18±12.28	53.57±13.48	<0.001
Total			458	30,639	-
Age group at diagnosis (yr)					<0.001
	<30		8 (1.75)	1,322 (4.31)
	30–39		11 (2.4)	2,803 (9.15)
	40–49		66 (14.41)	7,794 (25.44)
	50–59		139 (30.35)	8,833 (28.83)
	60–69		113 (24.67)	5,811 (18.97)
	70–79		100 (21.83)	3,321 (10.84)
	≥80		21 (4.59)	755 (2.46)
Stage at diagnosis (since 2006)					<0.001
	Local		60 (15.5)	6,075 (26.04)
	Regional		78 (20.16)	4,281 (18.35)
	Distant		237 (61.24)	11,715 (50.22)
	Unknown		12 (3.1)	1,258 (5.39)
Treatment modality
	Surgery				<0.001
		Yes	447 (97.6)	26,765 (87.36)
		No	11 (2.4)	3,874 (12.64)
	Radiation therapy				<0.001
		Yes	15 (3.28)	540 (1.76)
		No	443 (96.72)	30,099 (98.24)
	Chemotherapy				<0.001
		Yes	342 (74.67)	21,737 (70.95)
		No	116 (25.33)	8,902 (29.05)

Values are presented as mean ± standard deviation or number of cases (%).

The median overall survival (OS) of patients with ovarian carcinosarcoma was 39 months, and the 5-year OS rate was 42.5% (Fig. 1). Among ovarian carcinosarcomas, the prognosis was particularly worse in the distant stage and in older patients. The survival curves by stage showed that the 5-year survival rate of the localized stage patients was 60.8%, that of the regional stage was 57.9%, and that of the distant stage was 32.8% (p<0.001). The median OS in the distant stage was 27 months. Data for the localized and regional stages were not available. In the survival curves by age, the 5-year survival rates for patients younger than 50 years were 52.6% and 40.2% for patients older than 50 years (p=0.016). The median OS for patients younger than 50 years was 85 months, and that of patients older than 50 years was 34 months (Fig. 2).

Fig. 1. Comparisons of survival between ovarian carcinosarcoma and epithelial ovarian cancer.

Fig. 2. Clinical outcomes of ovarian carcinosarcoma according to the stage and age. (A) Stage. (B) Age.

mOS, median overall survival; NA, not available.

The clinical outcomes of ovarian carcinosarcoma were worse than those of epithelial ovarian cancer (the median OS of epithelial ovarian cancer was 110 months, and the 5-year survival rate was 60.5%), which was observed across all stages. In localized stage patients, the 5-year survival rates were 60.8% and 88.9% (p<0.001), and in regional stage patients, the 5-year survival rates were 57.9% and 70.8% (p<0.001). In patients with distant stage disease, clinical outcomes worsened, with 5-year survival rates of 32.8% and 41.9%, respectively (p<0.001; Fig. S1). In addition, when comparing the clinical outcomes before and after the age of 50 years, the OS of patients with ovarian carcinosarcoma was worse than that of patients with epithelial ovarian cancer (Fig. S2).

DISCUSSION

This study revealed that the ASR of ovarian carcinosarcoma was 0.064 per 100,000 women in Korea between 1999 and 2018. Moreover, the APC in ovarian carcinosarcoma was 5.86 (p<0.001). Ovarian carcinosarcoma accounted for 1.5% (458/30,639) of epithelial ovarian cancers.

Because of its rarity, many efforts have been made by researchers to study ovarian carcinosarcoma using the cancer registry database. Barnholtz-Sloan et al. [10] reported that 13,643 women were diagnosed with primary invasive ovarian cancer, and 382 (2.8%) of the women had ovarian carcinosarcoma between 1988 and 1997, using data from the SEER Program. In the present study, ovarian carcinosarcoma accounted for 1.5% (458/30,639) of epithelial ovarian cancers, and the incidence in both studies was comparable to that reported in other studies, i.e., 1%–3% [2,11,12]. However, caution is required when interpreting studies using data extracted from the SEER program, since the range of diseases used as the denominator may be different. For example, George et al. [13] described 25,974 (93.6%) patients with serous ovarian cancer and 1,763 (6.4%) with ovarian carcinosarcoma using data extracted from the SEER program. However, we cannot directly compare this study with the present study because the denominator was serous ovarian cancer, rather than epithelial ovarian cancer.

Despite slight fluctuations, the incidence of epithelial ovarian cancer, including carcinosarcoma, steadily increased throughout the study period. In particular, we observed that carcinosarcoma had a higher APC than epithelial ovarian cancer during the same period, which indicates a stiffer increase in ovarian carcinosarcoma than in ovarian epithelial cancer from the KCCR data. To our knowledge, this is the first report of an increasing trend in the incidence of carcinosarcoma.

However, there is no known reason for this rapid increase in carcinosarcoma incidence. Instead, several reasons for the higher APC of ovarian carcinosarcomas have been speculated. First, ovarian carcinosarcoma could be considered as another biphasic histology, such as mixed germ cell tumors, owing to its mimicked histological features. Recently, immunohistochemical markers, including SALL4 and CD10, have been developed as distinguishable markers for carcinosarcoma [14,15]. SALL4 is highly expressed in somatic malignancies, including lung and rhabdoid tumors, whereas other types of ovarian epithelial tumors are negative for this marker [16]. In addition, CD10 expression is associated with Müllerian system-derived neoplastic mesenchymal cells [17]. With the advent of novel immunohistochemical markers, a sarcomatous component of ovarian carcinosarcoma has been confirmed, and is possibly attributed to an increase in the proportion of APC.

Second, a previous study from the National Cancer Institute’s SEER registry suggested that ovarian carcinosarcoma is more commonly diagnosed in unmarried women; therefore, it may be clinically relevant in this data registry [18]. In recent decades, socioeconomic status, such as income, education, and female labor participation, has increased in Korea. These trends have negatively affected the timing of marriage and the APC of ovarian carcinosarcoma [19].

However, these explanations are not warranted, because the database does not contain additional information to support the increase in APC. Further research is needed on environmental or epidemiological correlations attributable to the increased incidence of carcinosarcoma. In addition, it is worth considering whether the increasing feature of APC in ovarian carcinosarcoma is valid in other countries.

Ovarian carcinosarcoma is generally associated with worse clinical outcomes than epithelial ovarian carcinoma. Cantrell and Van Le [12] reported that the clinical prognostic factors associated with poor survival include advanced stage at presentation, suboptimal debulking, and older age. However, these are not specific factors for ovarian carcinosarcoma and are universal factors commonly associated with epithelial ovarian cancer. The exact molecular and biological mechanisms underlying the poor clinical outcomes of ovarian carcinosarcoma are unknown; however, the reason can be indirectly inferred as a response to treatment. Ovarian carcinosarcomas are usually excluded from clinical trials, making it difficult to generate prospective data for this subtype. Therefore, treatment recommendations were based on retrospective studies with small patient populations. Recent treatment guidelines for ovarian cancer from the National Comprehensive Cancer Network [20] recommend surgical staging and intravenous paclitaxel/carboplatin adjuvant chemotherapy as the primary treatment, which is analogous to the treatment of epithelial ovarian cancer. Similarly, the Gynecologic Cancer InterGroup states that the mainstay of treatment remains cytoreductive surgery, even in the early stage, followed by platinum-based chemotherapy [11]. However, patients with ovarian carcinosarcoma respond poorly to platinum-based chemotherapies. Rauh-Hain et al. [21] performed a case-control study that included 50 patients with advanced-stage ovarian carcinosarcoma who underwent cytoreductive surgery and first-line platinum/taxane-based chemotherapy. Each patient was matched to 2 women with serous epithelial ovarian cancer, and the response rates to chemotherapy were 62% and 83%, respectively (p=0.03). Brown et al. [22] conducted a retrospective study and reported an inferior response rate to platinum-based chemotherapy in 37 patients with ovarian carcinosarcoma compared to 545 with serous ovarian cancer (25% vs. 60%; p=0.02). Because platinum sensitivity is a well-known prognostic factor for ovarian cancer, poorer response rates to platinum-based treatment could be an indicator of worse clinical outcomes in ovarian carcinosarcoma. In fact, there are no effective anticancer agents for both sarcomatous and carcinomatous components, and toxicity should be considered when several anticancer agents are used in combination. Therefore, the clinical outcomes of ovarian carcinosarcoma are worse than those of epithelial ovarian cancer. However, systematic studies are needed to elucidate the exact molecular mechanisms underlying carcinosarcoma survival.

In ovarian carcinosarcoma, the carcinomatous component is most often high-grade serous carcinoma [23], while other histological types may be seen. Tumors with an epithelial component of high-grade serous carcinoma may exhibit BRCA1 or BRCA2 gene mutations, with potential therapeutic implications [3]. Sonoda et al. [24] reported that the BRCA2 pathogenic variant was found in ovarian carcinosarcoma in both carcinoma and sarcomatous elements. The National Comprehensive Cancer Network Guideline also asks ovarian carcinosarcoma to use PARP inhibitors if pathogenic variants of BRCA1 or BRCA2 are found. The study of the frequency of BRCA1 or BRCA2 pathogenic variants in ovarian carcinosarcoma and the response to PARP inhibitors should be implemented in the future.

Ovarian carcinosarcoma is usually diagnosed at an older age than epithelial ovarian cancer is. In addition, the age at diagnosis is higher in patients with pure ovarian sarcoma or epithelial ovarian cancer. According to a previous study published by our research group, the mean age at diagnosis of pure sarcoma was 54.2 years, while that of carcinosarcoma in the present study was 60.2 years [25]. The exact reason for the late onset of carcinosarcoma remains unknown. However, we can deduce that the pathogeneses of pure ovarian sarcoma and carcinosarcoma are different. The tumors are of epithelial cell origin, and molecular studies support a monoclonal origin by revealing concordant TP52 abnormalities within the carcinoma and sarcoma components. However, further studies are required to elucidate how this developmental process is related to the late onset of carcinosarcoma.

This study has several strengths. First, we analyzed the incidence of ovarian carcinosarcoma using a population-based cancer registry, which contains approximately 98% of the Korean cancer cases. To our knowledge, this is the first nationwide registry-based study of ovarian carcinosarcoma in Korea. In addition, we report, for the first time, the recent trend of an increasing incidence of ovarian carcinosarcoma using APC. However, this study has some limitations. The KCCR database does not include relevant personal information, such as parity, body mass index, the presence of BRCA1 or BRCA2 gene mutations, and personal history of other cancers. The KCCR database only contains the first date of diagnosis, primary cancer site, SEER summary stage, and initial treatment (within 4 months after diagnosis). Moreover, the present study did not describe the treatment history or subtypes of the carcinomas and sarcomas. Therefore, careful attention should be paid when interpreting these results.

In conclusion, according to the KCCR, the incidence of ovarian carcinosarcoma in Korea increased between 1999 and 2018. Advanced stage and older (≥50 years) patients with ovarian carcinosarcoma had poorer survival outcomes than those at the local stage and younger patients.

SUPPLEMENTARY MATERIALS

Fig. S1

Comparisons of survival between ovarian carcinosarcoma and epithelial ovarian cancer by stage. (A) Localized. (B) Regional. (C) Distant.

Fig. S2

Comparisons of survival between ovarian carcinosarcoma and epithelial ovarian cancer by the age of diagnosis. (A) Age at diagnosis <50. (B) Age at diagnosis ≥50.