Share this article on:

Uterine Carcinosarcomas and Grade 3 Endometrioid Cancers: Evidence for Distinct Tumor Behavior

Bansal, Nisha MD1,4; Herzog, Thomas J. MD1,4; Seshan, Venkatraman E.2,4; Schiff, Peter B. MD, PhD3,4; Burke, William M. MD1,4; Cohen, Carmel J. MD1,4; Wright, Jason D. MD1,4

doi: 10.1097/AOG.0b013e318176157c
Original Research

OBJECTIVE: To compare the clinical behavior and outcome of uterine carcinosarcomas and grade 3 endometrioid carcinomas.

METHODS: Data on patients with grade 3 endometrioid adenocarcinomas and uterine carcinosarcomas, from 1988 to 2004, was obtained from the Surveillance, Epidemiology, and End Results database. Mortality was analyzed using Cox proportional hazards models. Survival analysis was performed with the Kaplan-Meier method and log rank test.

RESULTS: The cohort included 8,986 women with 5,024 (56%) grade 3 endometrioid carcinomas and 3,962 (44%) uterine carcinosarcomas. Women with uterine carcinosarcomas were older (aged 70 years compared with 66 years; P<.001) and more often nonwhite (23% compared with 15%; P<.001). These women presented with more advanced disease (stage III/IV 41% compared with 31%; P<.001). Multivariable analysis demonstrated that uterine carcinosarcoma histology, advanced age, nonwhite race, and advanced stage were independent predictors of poor survival. Cancer-specific mortality was 45% lower in women with grade 3 endometrioid carcinomas (hazard ratio 0.55; 95% confidence interval [CI] 0.5–0.6). The 5-year cancer-specific survival was lower for women with uterine carcinosarcoma for each disease stage. Survival for stage IC was 38% (95% CI 33–45%) for uterine carcinosarcoma compared with 68% (95% CI 63–73%) for grade 3 endometrioid carcinoma. For stage III, survival was 22% (95% CI 19–26%) for uterine carcinosarcoma compared with 45% (95% CI 41–49%) for grade 3 endometrioid carcinoma.

CONCLUSION: Carcinosarcomas present at more advanced stage and have worse survival than grade 3 endometrioid carcinomas. Carcinosarcomas may represent a distinct biologic entity.


Uterine carcinosarcomas have a distinct clinical behavior and worse outcome compared with grade 3 endometrioid endometrial cancers.

From the 1Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, 2Department of Biostatistics, and 3Department of Radiation Oncology, Columbia University College of Physicians and Surgeons; and 4the Herbert Irving Comprehensive Cancer Center, New York, New York.

Corresponding author: Jason D. Wright, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, 161 Fort Washington Avenue, 8th Floor, New York, NY 10032; e-mail:

Financial Disclosure The authors have no potential conflicts of interest to disclose.

Uterine carcinosarcomas comprise fewer than 5% of all uterine neoplasms, with an annual incidence of less than 2 per 100,000.1 Carcinosarcomas are unique in that these tumors contain both malignant carcinomatous and sarcomatous components. These are highly aggressive tumors that often have spread beyond the uterus at the time of surgical assessment.1 The rarity of uterine carcinosarcomas has prevented large epidemiologic studies from investigating these neoplasms. A small series of clinical, pathologic, and biologic studies have attempted to elucidate the behavior of uterine carcinosarcomas. Whether carcinosarcomas are more akin to high-grade epithelial carcinomas and therefore should be regarded as such in future studies or they should retain their designation as a type of uterine sarcoma remains an area of active debate.

Recent data have suggested that uterine carcinosarcomas are monoclonal in origin, with the sarcomatous component representing dedifferentiation of the carcinomatous portion.2 The carcinomatous element is thought to be the driving force. Therefore, the argument has been made that uterine carcinosarcomas actually behave as aggressive histologic subtypes of epithelial endometrial carcinomas. Histologically, tumor emboli within lymphovascular channels and metastases primarily consist of the carcinomatous component.3 Further evidence that uterine carcinosarcomas arise from a single stem cell comes from studies examining immunohistochemical expression of p53 protein, which have found concordance of p53 protein expression in both the carcinomatous and sarcomatous portions.4,5 Uterine carcinosarcomas also share similar risk factors with endometrial carcinomas. Both neoplasms are associated with obesity, nulliparity, and exogenous estrogen use.6 In addition, oral contraceptives are protective against both types of neoplasms. These studies argue for a similar pathogenesis of carcinosarcomas and epithelial endometrial neoplasms.

Evidence that uterine carcinosarcomas are biologically distinct from high-grade endometrial cancer stems from the worse prognosis of carcinosarcomas. Several small retrospective reports have documented a worse survival for women with uterine carcinosarcomas as compared with high-grade endometrioid cancers as well as serous and clear cell tumors. These reports have therefore argued that carcinosarcomas should not be included in studies of patients with endometrioid cancers.7–8 The poor prognosis for uterine carcinosarcomas underscores the necessity for improved therapeutic strategies. Understanding the epidemiology and outcome of uterine carcinosarcomas as distinct from endometrioid carcinomas allows reliable exclusion of these cancers from studies examining endometrioid carcinomas. This is particularly important because these are rare neoplasms and randomized trials examining treatment options are difficult. The purpose of this study is to compare the clinical behavior, natural history, and outcome of a large cohort of patients with uterine carcinosarcomas compared with women with grade 3 endometrioid endometrial cancers.

Back to Top | Article Outline


This study was approved by the Columbia University Institutional Review Board. Data on patients with grade 3 endometrioid adenocarcinomas and carcinosarcomas of the uterus was obtained through review of the Surveillance, Epidemiology, and End Results (SEER) database from the National Cancer Institute. This is a population-based cancer registry that includes approximately 26% of the United States population.9 The SEER database is composed of a number of geographically distinct tumor registries. The demographic characteristics of the SEER registries are representative of the general population within the United States.10 Data from SEER 17 registries was used.

Women with invasive uterine carcinosarcomas and grade 3 endometrioid endometrial adenocarcinomas treated between 1988 and 2004 were analyzed. Clinical and pathologic data, including age at diagnosis, race (white, African American, or other), number of lymph nodes removed, and marital status were collected. Year of diagnosis was classified as 1988–1993, 1994–1998, or 1999–2004 for analysis. Patients were categorized based on the geographic area of residence at the time of diagnosis: central (Detroit, Iowa, Kentucky, Louisiana), eastern (Connecticut, New Jersey, Atlanta, rural Georgia) and western (Alaska, California, Hawaii, Los Angeles, New Mexico, San Francisco, San Jose, Seattle, Utah) United States. Staging information was derived from the American Joint Cancer Committee staging information and recorded extent of disease codes.10 Patients were stratified based on histology (carcinosarcoma compared with grade 3 endometrioid cancer).

Each patient’s vital status was recorded. Survival was calculated as the number of months from cancer diagnosis to date of death. Overall and cancer-specific survival were calculated for each patient. Patients alive at last follow-up were censored.

Frequency distributions between categorical variables were compared using the χ2 test. Cox proportional hazards models were developed to examine survival. In Cox proportional hazards analyses, we modeled the cancer-specific mortality hazard ratios comparing patients with grade 3 endometrioid cancers to patients with carcinosarcomas, controlling for the other predictor variables (age, race, year of diagnosis, area of residence, marital status, performance of lymphadenectomy, receipt of radiation, and stage). Additionally, separate Cox models were generated for patients with grade 3 tumors as well as for patients with uterine carcinosarcomas. Kaplan-Meier curves were generated to examine overall and cancer-specific survival. All hypothesis tests were two-tailed. Statistical analysis was conducted using R.

Back to Top | Article Outline


A total of 8,986 patients were identified. The cohort included 5,024 (56%) women with grade 3 endometrioid carcinomas and 3,962 (44%) with uterine carcinosarcomas. The clinical and demographic characteristics of the cohort are displayed in Table 1. Carcinosarcoma patients were older; 64% were older than 60 years, with a median age of 70 years, compared with grade 3 endometrioid cancers, of whom 51% were older than 60 years, with a median age of 66 years (P<.001) (Table 1). Carcinosarcomas were more common in African-American women; 18% of patients with carcinosarcomas were African American compared with 8% of those with grade 3 endometrioid tumors (P<.001). Carcinosarcoma patients more often presented with advanced stage disease. Stage III/IV tumors were noted in 40.6% of carcinosarcoma patients compared with 30.5% of grade 3 endometrioid adenocarcinoma patients (P<.001). Lymphadenectomy was undertaken in 61% of women with carcinosarcomas compared with 67% of those with grade 3 neoplasms (P<.001). Patients with carcinosarcoma were less likely to receive adjuvant radiotherapy; 37% of women with carcinosarcomas underwent radiation compared with 43% of those with grade 3 cancers (P<.001).

Table 1

Table 1

In a Cox proportional hazards model of women with carcinosarcoma advanced age, African-American race, later year of diagnosis, treatment in the western United States, and advanced stage were associated with a worse prognosis. In contrast, age younger than 40 years, single status, and receipt of radiation were associated with an improved survival (Table 2). African-American women with carcinosarcomas were 19% more likely to die from their tumors than white women (hazard ratio [HR] 1.19; 95% confidence interval [CI] 1.07–1.36). Likewise, women aged older than 60 years had a 61% increase in mortality compared with patients aged 40–60 years with carcinosarcomas (HR 1.61; 95% CI 1.47–1.76).

Table 2

Table 2

The effect of histology on survival was examined for the entire cohort of patients with grade 3 tumors and carcinosarcomas. Even after adjusting for all other prognostic factors, patients with carcinosarcomas had a decreased survival. The risk of death was 45% lower in patients with grade 3 tumors than in women with carcinosarcomas (HR 0.55; 95% CI 0.51–0.59) (Table 2). Age, date of diagnosis, race, marital status, radiotherapy, lymphadenectomy, and stage all remained as prognostic variables.

Patients were then stratified by stage to analyze 5-year survival (Table 3). Carcinosarcomas had a worse survival for every stage examined. Among women with stage IA tumors the 5-year cancer-specific survival was 59% (95% CI 54–66%) for carcinosarcomas compared with 78% (95% CI 73–84%) for grade 3 endometrioid tumors (P<.001). Similar findings were noted for patients with advanced stage disease. Only 22% (95% CI 19–26%) of patients with stage III carcinosarcomas were alive at 5 years compared with 45% (95% CI 41–49%) of those with grade 3 endometrial tumors (P<.001). In Kaplan-Meier analysis, cancer-specific survival was worse in patients with carcinosarcomas (P<.001). The results were similar for both early stage (stage I and II; P<.001) and advanced stage tumors (stage III and IV; P<.001) (Fig. 1).

Table 3

Table 3



Back to Top | Article Outline


Given that carcinosarcomas are rare uterine neoplasms, large epidemiologic studies to describe clinical behavior and outcome are limited. In our study, carcinosarcomas had a distinctly worse outcome than matched grade 3 endometrial tumors. Patients with grade 3 endometrial cancers were 45% less likely to die from their tumors than matched patients with carcinosarcomas. The decreased survival in women with carcinosarcomas was noted for both early- and late-stage disease.

Prior work has examined the survival of patients with uterine carcinosarcomas. In general, these studies have shown that survival is poor. The reported overall survival for women with uterine carcinosarcomas ranges from 20 to 35%.1,11–15 In a series of 300 patients with clinical stage I to III uterine carcinosarcoma, the 5-year overall survival was 31%. This report includes patients treated over the span of several decades.11

A number of studies have attempted to compare the clinical behavior of carcinosarcomas to other high-grade endometrial tumors. The majority of these reports are limited by small sample size and inclusion of a variety of histologic subtypes. Amant and colleagues16 analyzed 137 patients, including 50 with grade 3 endometrioid carcinomas, 54 with serous or clear cell cancers, and 33 uterine carcinosarcomas. Median survival was 23 months for the carcinosarcomas compared with 60 months for grade 3 endometrioid cancers and 40 months for serous and clear cell carcinomas. In a subgroup analysis of patients with stage I–II disease, just 44% of the carcinosarcoma patients survived, compared with 86% of grade 3 endometrioid cancer patients. Similarly, in a study of 71 patients with high-grade uterine neoplasms, the 5-year overall survival was significantly lower for women with carcinosarcomas than for those with grade 3 endometrial carcinomas (25% compared with 64%).7 Finally, a recent case–control study of 45 patients with uterine carcinosarcomas reported similar findings, with a median survival of 18 months for carcinosarcoma patients compared with 36 months for patients with grade 3 endometrioid, serous, and clear cell cancers.8 In each of these studies the authors concluded that uterine carcinosarcomas are biologically distinct, with a different prognosis and worse outcomes. Our findings are in accord with these data. For every stage of disease, the presence of sarcomatous elements conveyed a worse prognosis. Even for women with stage IA neoplasms confined to the endometrium, the 5-year survival was 19% lower for those women with carcinosarcomas.

In our study, factors independently associated with survival for women with carcinosarcomas included age, race, year of diagnosis, radiation, performance of lymphadenectomy, and stage. These findings are in concordance with prognostic factors reported in the literature. Amant et al16 reported that carcinosarcoma histology and advanced stage were of prognostic significance. The largest study to report on pathologic predictors of survival was conducted by the Gynecologic Oncology Group.17 Prognostic factors related to progression-free interval in multivariate analysis included adnexal spread, lymph node metastases, cell type, and grade of sarcoma. Age is a known risk factor for women with carcinosarcomas. Callister and co-workers11 noted that the relative risk of death was 3.3 in postmenopausal women with carcinosarcomas. Among African-American women in our cohort, the risk of death was 19% higher than in matched white patients. A recent Gynecologic Oncology Group trial also evidenced the importance of race for patients with uterine carcinosarcomas.18

The primary impetus for comparing the outcomes of grade 3 endometrioid tumors and carcinosarcomas stems from the need to refine strategies for adjuvant management. The poor survival for carcinosarcomas highlights the need for improved therapeutic strategies. Although the differential clinical behavior of carcinosarcomas suggests that these neoplasms should be regarded as distinct in clinical trials, the rarity of these neoplasms precludes timely completion of clinical protocols. To date, many studies of high-grade endometrial cancer have included carcinosarcomas. The results of our study would strongly argue against this approach, because patients with carcinosarcoma have significantly worse outcomes. A large randomized, prospective phase III trial by the Gynecologic Oncology Group19 to examine uterine carcinosarcomas independently was recently published. The protocol examined single-agent ifosfamide compared with the combination of ifosfamide and paclitaxel and determined that there was a significant survival advantage with the combination regimen. Although paclitaxel has been demonstrated also to have activity in the first- and second-line treatment of endometrial cancer, other agents that are active for endometrial cancer such as cisplatin and doxorubicin have only moderate activity for the treatment of carcinosarcomas.18,20 The differential drug effect may be due to the two cellular elements of carcinosarcomas.

The primary strength of our study is the large number of patients with a relatively rare tumor that were examined. However, a number of limitations must be acknowledged. Although SEER data provide a representative sample of the U.S. population, as with any registry study, centralized pathology review is not available. This is particularly important in studies of rare tumors such as carcinosarcomas. Although data on radiation use is available, other treatment modalities such as chemotherapy and hormonal therapy are not recorded in the SEER database. In the past 5 years there has been an increasing shift toward the use of chemotherapy for women with advanced stage endometrial cancer. Although it is difficult to determine whether the patterns of care for carcinosarcomas have also shifted, we cannot exclude that our survival data may have been influenced by differential patterns of chemotherapy use. Last, as is true of any retrospective report, we were unable to account for the role of patient and physician preferences in the use of initial and adjuvant therapy.

Uterine carcinosarcomas are rare, aggressive neoplasms that present at a more advanced stage and have a worse prognosis than grade 3 endometrioid endometrial carcinomas. Despite pathologic similarities between these two tumor types, the differential behavior and outcome, even when disease is confined to the uterus, argues for continued distinct classification of these tumors. The poor prognosis for uterine carcinosarcomas underscores the importance of identifying novel therapeutic options for treatment of this disease.

Back to Top | Article Outline


1. Nielsen SN, Podratz KC, Scheithauer BW, O’Brien PC. Clinicopathologic analysis of uterine malignant mixed müllerian tumors. Gynecol Oncol 1989;34:372–8.
2. McCluggage WG. Uterine carcinosarcomas (malignant mixed Mullerian tumors) are metaplastic carcinomas. Int J Gynecol Cancer 2002;12:687–90.
3. Silverberg SG, Major FJ, Blessing JA, Fetter B, Askin FB, Liao SY, et al. Carcinosarcoma (malignant mixed mesodermal tumor) of the uterus. A Gynecologic Oncology Group pathologic study of 203 cases. Int J Gynecol Pathol 1990;9:1–19.
4. Mayall F, Rutty K, Campbell F, Goddard H. p53 immunostaining suggests that uterine carcinosarcomas are monoclonal. Histopathology 1994;24:211–4.
5. Szukala SA, Marks JR, Burchette JL, Elbendary AA, Krigman HR. Co-expression of p53 by epithelial and stromal elements in carcinosarcoma of the female genital tract: an immunohistochemical study of 19 cases. Int J Gynecol Cancer 1999;9:131–6.
6. Zelmanowicz A, Hildesheim A, Sherman ME, Sturgeon SR, Kurman RJ, Barrett RJ, et al. Evidence for a common etiology for endometrial carcinomas and malignant mixed mullerian tumors. Gynecol Oncol 1998;69:253–7.
7. George E, Lillemoe TJ, Twiggs LB, Perrone T. Malignant mixed müllerian tumor versus high-grade endometrial carcinoma and aggressive variants of endometrial carcinoma: a comparative analysis of survival. Int J Gynecol Pathol 1995;14:39–44.
8. Vaidya AP, Horowitz NS, Oliva E, Halpern EF, Duska LR. Uterine malignant mixed mullerian tumors should not be included in studies of endometrial carcinoma. Gynecol Oncol 2006;103:684–7.
9. Surveillance, Epidemiology and End Results (SEER) Program, SEER*Stat Database: Incidence-SEER 9 Regs Limited-Use, Nov 2006 Sub (1973-2004), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2007, based on the November 2006 submission. Available at: Retrieved October, 2007.
10. The Surveillance, Epidemiology, and End Results (SEER) Program, Data Quality. National Cancer Institute, 2007. Available at: Retrieved October, 2007.
11. Callister M, Ramondetta LM, Jhingran A, Burke TW, Eifel PJ. Malignant mixed Müllerian tumors of the uterus: analysis of patterns of failure, prognostic factors, and treatment outcome. Int J Radiat Oncol Biol Phys 2004;58:786–96.
12. Dinh TV, Slavin RE, Bhagavan BS, Hannigan EV, Tiamson EM, Yandell RB. Mixed müllerian tumors of the uterus: a clinicopathologic study. Obstet Gynecol 1989;74:388–92.
13. Podczaski ES, Woomert CA, Stevens CW Jr, Manetta A, Larson JE, Zaino RJ, et al. Management of malignant, mixed mesodermal tumors of the uterus. Gynecol Oncol 1989;32:240–4.
14. Schweizer W, Demopoulos R, Beller U, Dubin N. Prognostic factors for malignant mixed müllerian tumors of the uterus. Int J Gynecol Pathol 1990;9:129–36.
15. Yamada SD, Burger RA, Brewster WR, Anton D, Kohler MF, Monk BJ. Pathologic variables and adjuvant therapy as predictors of recurrence and survival for patients with surgically evaluated carcinosarcoma of the uterus. Cancer 2000;88:2782–6.
16. Amant F, Cadron I, Fuso L, Berteloot P, de Jonge E, Jacomen G, et al. Endometrial carcinosarcomas have a different prognosis and pattern of spread compared to high-risk epithelial endometrial cancer. Gynecol Oncol 2005;98:274–80.
17. Major FJ, Blessing JA, Silverberg SG, Morrow CP, Creasman WT, Currie JL, et al. Prognostic factors in early-stage uterine sarcoma. A Gynecologic Oncology Group study. Cancer 1993;71 suppl:1702–9.
18. Sutton G, Brunetto VL, Kilgore L, Soper JT, McGehee R, Olt G, et al. A phase III trial of ifosfamide with or without cisplatin in carcinosarcoma of the uterus: a Gynecologic Oncology Group study. Gynecol Oncol 2000;79:147–53.
19. Homesley HD, Filiaci V, Markman M, Bitterman P, Eaton L, Kilgore LC, et al. Phase III trial of ifosfamide with or without paclitaxel in advanced uterine carcinosarcoma: a Gynecologic Oncology Group study. J Clin Oncol 2007;25:526–31.
20. Omura GA, Major FJ, Blessing JA, Sedlacek TV, Thigpen JT, Creasman WT, et al. A randomized study of adriamycin with and without dimethyl triazenoimidazole carboxamide in advanced uterine sarcomas. Cancer 1983;52:626–32.
© 2008 The American College of Obstetricians and Gynecologists