Clinical features, diagnosis, staging, and treatment of uterine carcinosarcoma

INTRODUCTION — Uterine carcinosarcomas (previously called malignant mixed Müllerian tumors) are rare and aggressive cancers that are dedifferentiated (metaplastic) carcinomas comprised of carcinomatous and sarcomatous elements arising from a single malignant epithelial clone. They may represent a stable disruption of the epithelial mesenchymal transition [1-3]. They are considered a high-risk variant of endometrial adenocarcinoma because carcinosarcomas share similarities in epidemiology, risk factors, and clinical behavior more closely with endometrial carcinoma as opposed to uterine sarcomas.

The clinical features, diagnosis, staging, and treatment of uterine carcinosarcoma will be discussed in this topic review. Other types of endometrial cancer and uterine sarcomas are reviewed separately. (See "Uterine sarcoma: Classification, epidemiology, clinical manifestations, and diagnosis" and "Overview of resectable endometrial carcinoma" and "Endometrial stromal sarcomas, related tumors, and uterine adenosarcoma".)

EPIDEMIOLOGY AND RISK FACTORS — Uterine carcinosarcomas are rare tumors that account for less than 5 percent of all uterine malignancies [4]. As an example, in the United States, the incidence of carcinosarcoma is approximately 1 to 4 per 100,000 women [5]. Carcinosarcomas occur in older women; the median age at diagnosis ranges from 62 to 67 years [6]. Black women have a twofold higher incidence of uterine carcinosarcoma compared with non-Hispanic White women [7,8].

Uterine carcinosarcomas share similar risk factors with endometrial carcinomas. Both neoplasms are associated with obesity, nulliparity, and use of exogenous estrogen and tamoxifen [9-17]. Progestin-containing contraceptives are protective against both types of neoplasms. A history of exposure to pelvic radiation is also associated with an increased risk of developing uterine carcinosarcoma [18-20].

CLINICAL FEATURES — It is not possible to distinguish carcinosarcoma from endometrial carcinoma or uterine sarcoma based on clinical features. The diagnosis requires histologic evaluation. The typical features of carcinosarcoma are reviewed below. (See "Endometrial carcinoma: Clinical features, diagnosis, prognosis, and screening", section on 'Clinical presentation'.)

Presentation — Women with uterine carcinosarcomas may present with a classical clinical triad of pain, bleeding, and a rapidly enlarging uterus. Of these, vaginal bleeding is the most common presenting sign for women with carcinosarcoma [21,22]. In the largest report involving 300 patients, presenting signs included postmenopausal bleeding (82 percent), pelvic pain (13 percent), and vaginal discharge (10 percent) [22].

Over 10 percent of patients with carcinosarcoma will present with metastatic disease, and 60 percent will have extrauterine disease on staging scans [23,24].

On exam, a pelvic mass may be palpated or seen protruding through the cervical os. Up to 15 percent of patients have involvement of the cervix identified through cervical biopsy, endocervical curettage, or both [22]. (See "Endometrial carcinoma: Epidemiology, risk factors, and prevention" and "Endometrial carcinoma: Serous and clear cell histologies".)

Imaging

Ultrasound — For women with suspected uterine cancer, pelvic sonography is often the first-line imaging study. Carcinosarcomas are hyperechoic compared with the myometrium on ultrasound. In addition, expansion of the endometrial canal has been described [25]. However, findings on ultrasound cannot distinguish a carcinosarcoma from an endometrial adenocarcinoma. (See "Endometrial carcinoma: Clinical features, diagnosis, prognosis, and screening", section on 'Ultrasound'.)

MRI — For patients presenting with suspicion of a uterine malignancy, gadolinium-enhanced magnetic resonance imaging (MRI) is occasionally performed to evaluate for local extent of disease. On MRI, carcinosarcoma presents as a heterogenous bulky polypoid mass that often prolapses into the endocervical canal and demonstrates prolonged intense enhancement [26-28]. In one report of 17 cases evaluated by MRI, 88 percent had higher signal on T2 weighted imaging compared with the myometrium [28]. Carcinosarcomas tend to show a lower incidence of hemorrhage compared with leiomyosarcoma [27].

CT scan — On computed tomography (CT), uterine carcinosarcoma is described as a heterogenous, ill defined, hypodense mass with concomitant dilatation of the endometrial canal [25]. CT scan can also detect myometrial invasion, nodal involvement, and the presence of metastatic disease. This is important because up to 30 percent of patients have node involvement at presentation, and over 17 percent present with metastatic disease (most commonly involving the lung, liver, and bones) [22,26,29].

FDG PET/CT — Positron emission tomography (PET) using (18)-Fluorodeoxyglucose (FDG) with or without computed tomography (CT) can be used for diagnostic and initial staging purposes. Carcinosarcoma is generally positive on PET scan; the mean standardized uptake value (SUV) in one study of 19 patients was 10 ± 5.5 (range, 3 to 22) [30].

Laboratory findings — There are no typical laboratory abnormalities associated with the diagnosis of carcinosarcoma, although anemia is present in up to 10 percent as a result of vaginal bleeding [22].

CA 125 — Cancer antigen (CA) 125 is a glycoprotein encoded by the mucin 16 (MUC16) gene. It is most commonly used in the follow-up of patients with ovarian cancer, but it can be elevated in carcinosarcoma. If elevated, it appears to correlate with metastases or tumor bulk. (See "Adnexal mass: Role of serum biomarkers in diagnosing epithelial carcinoma of the ovary, fallopian tube, or peritoneum", section on 'Cancer antigen 125'.)

Preoperative elevation of the tumor marker CA 125 may signify more advanced disease [31]. In a retrospective study of 54 patients, CA 125 value ≥30 U/mL was significantly associated with higher disease stage (62 versus 38 percent with a CA 125 <30 U/mL), presence of a serous component (75 versus 25 percent), and myometrial invasion ≥50 percent (61 versus 39 percent).

Pathology — Uterine carcinosarcoma is "biphasic" and characterized by both carcinomatous (epithelial) and sarcomatous (connective tissue) elements. Based on the sarcomatous elements, carcinosarcoma is often described as homologous, if the sarcomatous element is native to uterine tissue, or heterologous, if the sarcoma is not native to uterine tissue. The most common carcinoma/sarcoma combination is a mixed homologous tumor consisting of a high-grade serous carcinoma and an endometrial stromal sarcoma. However, different histologies are possible for each of the two components. These include the following histologies [6,32]:

●Carcinoma – Endometrioid, serous, clear cell, or undifferentiated; adenocarcinoma with squamous differentiation may also be described.

●Homologous sarcoma – Endometrial stromal sarcoma, leiomyosarcoma, or undifferentiated endometrial sarcoma.

●Heterologous sarcoma – Rhabdomyosarcoma, chondrosarcoma, osteosarcoma, liposarcoma, or fibrosarcoma.

While both components must be present in the uterine specimen to make this diagnosis, metastatic sites are generally composed of the carcinomatous (epithelial) portion [33-36]. Uterine carcinosarcoma is considered one of the best examples of a stable epithelial-mesenchymal transition (EMT) characterized by transdifferentiation, expression of mesenchymal markers (altered cadherins and expression of the transcription factor, Snail), and a greater migratory potential [37].

EVALUATION OF SUSPECTED DISEASE — Uterine carcinosarcoma is a histologic diagnosis. It is not possible to distinguish carcinosarcoma from endometrial carcinoma or uterine sarcoma based upon symptoms, laboratory evaluation, or imaging findings. For women presenting with a clinical suspicion of a uterine malignancy, endometrial sampling is usually performed prior to definitive surgery [38,39]. However, if the endometrial biopsy is negative but clinical suspicion remains high, further work-up is required. (See "Overview of the evaluation of the endometrium for malignant or premalignant disease".)

Because carcinosarcoma arises from the endometrial lining, it is theoretically more amenable to diagnosis by endometrial sampling than pure uterine sarcomas. However, one study highlights that sampling may not be an accurate test for the diagnosis of uterine carcinosarcoma [38]. In a study involving 51 women with carcinosarcoma, preoperative sampling was diagnostic in 59 percent (30 of 51) [38]. In 16 additional cases, the presence of malignancy was established by endometrial sampling, and in the majority, it was due to the detection of the carcinomatous component.

For patients in whom endometrial sampling is not informative, further evaluation is required. This usually involves pelvic sonography with or without directed biopsy. The evaluation is similar to that in women with suspected endometrial cancer. (See "Endometrial carcinoma: Clinical features, diagnosis, prognosis, and screening", section on 'Approach to patients with nondiagnostic histopathology'.)

DIAGNOSIS — The diagnosis of carcinosarcoma is based on histologic evidence of a dual population of carcinomatous and sarcomatous cells with invasion of the stroma. This is typically based upon pathology evaluation following hysterectomy, but in some cases, the diagnosis is made with endometrial biopsy.

Postdiagnostic evaluation — Patients with a biopsy-proven diagnosis of carcinosarcoma (and those suspected of having a carcinosarcoma) should undergo imaging to rule out metastatic disease because almost 60 percent of patients will have metastatic disease at the time of initial presentation. (See 'Imaging' above.)

There are no studies to suggest a preferred modality. In our practice, we suggest at least a computed tomography (CT) of the chest, abdomen, and pelvis for an extent of disease evaluation. This information is necessary to inform the appropriate treatment plan. While positron emission tomography (PET)/CT may be used instead of CT, there are only limited data to suggest that PET/CT yields more information [40]. In one study of 19 patients, PET/CT had a clinical impact on the management of seven patients [30]. PET excluded bone metastases (suggested on magnetic resonance imaging [MRI]) in two patients, enabling them to undergo surgical treatment. PET findings resulted in the upstaging of five patients, and four of these were deemed inoperable based on the PET results. However, the use of PET was not associated with an improvement in overall survival compared with historic controls. (See 'FDG PET/CT' above.)

DIFFERENTIAL DIAGNOSIS — The differential diagnosis is the same as that for endometrial adenocarcinoma.

TREATMENT APPROACH — Most patients are diagnosed by endometrial biopsy. A smaller number of patients are initially diagnosed after hysterectomy for treatment of presumed fibroids or pelvic pain. The approach to treatment depends on whether the diagnosis was made preoperatively or following a hysterectomy.

Diagnosis based on biopsy — For women with a biopsy-proven diagnosis of carcinosarcoma, the approach is stratified on whether or not advanced disease is present or suspected based on imaging.

●For patients with no evidence of metastatic disease, we recommend surgery and often adjuvant chemotherapy. (See 'Surgery' below and 'Adjuvant therapy' below.)

●For women with extrauterine disease limited to the peritoneum (a subset of stage IVB), we suggest surgical cytoreduction. (See 'Cytoreduction' below.)

●For patients with extra-abdominal metastatic disease, treatment goals are palliative. The role of surgery and/or chemotherapy should be individualized to address disease-related symptoms with consideration of the patient’s personal goals of care and clinical status. (See 'Treatment of metastatic disease' below.)

Diagnosis posthysterectomy — Given the risk of metastatic spread and nodal involvement, patients should undergo postoperative imaging to rule out metastatic disease. (See 'Postdiagnostic evaluation' above.)

●For women with no evidence of extra-abdominal metastatic disease, we generally proceed with a second surgery either done through an exploratory laparotomy or a minimally invasive approach. In addition to careful exploration, we also perform lymphadenectomy. (See 'Cytoreduction' below and 'Lymphadenectomy' below.)

●Women with peritoneal carcinomatosis have stage IVA disease, and those with visceral or extra-abdominal metastases (ie, lung or liver involvement) have stage IVB disease. For these patients, treatment is palliative. The role of surgery and/or chemotherapy should be individualized to address disease-related symptoms with consideration of the patient’s clinical status and goals for treatment. However, we suggest a biopsy of a representative lesion to confirm metastasis and characterize the histologic pattern in the metastatic lesion. (See 'Treatment of metastatic disease' below.)

STAGING — Uterine carcinosarcoma is surgically staged according to the 2017 International Federation of Gynecology and Obstetrics (FIGO)/Tumor, Node, Metastasis (TNM) classification system, the same system as for endometrial carcinoma (table 1). Findings of disease outside of the abdomen constitute stage IVB disease, in which case, surgical staging is of limited value.

SURGERY — The primary management of carcinosarcoma limited to the abdomen is surgery for both staging and initial treatment [41]. Complete surgical staging includes total hysterectomy, bilateral salpingo-oophorectomy, pelvic and para-aortic lymph node dissection, cytology of peritoneal washings, omentectomy, and biopsies of peritoneal surfaces. Minimally invasive approaches for apparent stage I and II disease appear appropriate for these patients. A posthoc analysis of the Gynecologic Oncology Group LAP2 trial evaluated the 72 patients with carcinosarcoma combined with the other "type II" cases for a total of 403 cases and concluded that patterns of recurrence and survival were not impacted by surgical approach [42].

Lymphadenectomy — All patients with carcinosarcoma should undergo a lymphadenectomy. In addition to informing the appropriate stage of patients, lymphadenectomy may be associated with an improvement in overall survival (OS) [43,44]. In one of the largest reports (n = 855 women with stage I to III carcinosarcoma), median survival with and without lymphadenectomy was 54 and 25 months, respectively [43]. A multicenter, prospective, cohort study that evaluated sentinel lymph node mapping in women with endometrial cancer included 13 women with carcinosarcoma and did not definitively address the efficacy of this approach in this patient population [45].

Cytoreduction — Although there are low-quality data available, surgical cytoreduction is a reasonable option for women who present with advanced uterine carcinosarcoma (ie, stage III to IVA), particularly if complete cytoreduction to no gross residual disease can be achieved. (See "Endometrial carcinoma: Staging and surgical treatment".)

One study retrospectively profiled the outcomes following surgical cytoreduction of 44 highly selected patients with advanced carcinosarcoma (14 with stage IIIC, 30 with stage IVB) [46]. Complete surgical cytoreduction was associated with a significant survival advantage compared with patients with gross residual disease postoperatively (median 52 versus 9 months, p <0.0001). On multivariate analysis, complete cytoreduction was independently associated with OS.

ADJUVANT THERAPY — Our approach to adjuvant treatment following surgical resection for carcinosarcoma is based on stage at diagnosis, which is consistent with the National Comprehensive Cancer Network (NCCN) guidelines [41]. In the future it is likely that biomarkers such as abnormal p53 will be used to select patients for adjuvant chemotherapy per PORTEC-3 [47,48].

Stage IA — For women with carcinosarcoma that invades less than half of the myometrium (stage IA), some experts prefer observation rather than adjuvant treatment because women with stage IA carcinosarcoma have a better prognosis compared with those with IB or later-stage disease [49]. In addition, it is unclear if their survival is improved with any form of adjuvant treatment [50-52]. However, given that European Organisation for Research and Treatment of Cancer (EORTC) 55874 showed that pelvic RT improved local control (but not progression-free survival [PFS] or overall survival [OS]) for carcinosarcoma [50], and that Gynecologic Oncology Group (GOG) 258 has demonstrated that chemotherapy has a greater impact on systemic control of disease [53], other UpToDate experts favor adjuvant chemotherapy over observation for high-risk tumors. Moreover, some women with stage IA disease may prefer to be treated. For those who desire treatment and accept the toxicities of therapy, adjuvant treatment is reasonable, provided these patients are aware that there is no evidence base for an impact on OS. In such cases, chemotherapy is administered using a treatment algorithm similar to one used for serous carcinoma of the uterus [54]. (See "Adjuvant treatment of high-risk endometrial cancers", section on 'Noninvasive stage IA disease'.)

No trials have evaluated the value of adjuvant chemotherapy specifically in women with stage IA disease. However, trials in women with early-stage disease (stage I and II) have consistently reported improvement in recurrence rates and PFS, but not in OS. Such patients were studied in GOG 261, discussed below. (See 'Chemotherapy plus radiation therapy' below.)

The outcomes of women with early-stage carcinosarcoma are informed by the following data:

●A multi-institutional retrospective study that included 111 women with early-stage uterine carcinosarcoma (85 percent with stage I disease) evaluated the outcomes following surgical staging in women who underwent observation (n = 44), radiation therapy (RT) (n = 23), chemotherapy plus RT (n = 15), or chemotherapy alone (n = 29) [52]. Stage, receipt of chemotherapy, and the presence of lymphovascular space invasion (LVSI) were significantly associated with PFS. However, none of these factors were associated with OS.

●In GOG 20, 156 women with stage I or II uterine sarcomas were randomly assigned to postoperative surveillance or adjuvant doxorubicin [51]. Although adjuvant doxorubicin resulted in a lower recurrence rate (41 versus 53 percent, respectively), the result was not statistically significant. In addition, there was no difference in PFS or OS. Toxicities included lymphopenia (33 percent) and six cases of cardiac toxicity, though, as reported, none of them were deemed life threatening.

●In a trial performed by the European Organization for the Research and Treatment of Cancer (EORTC 55874), over 200 women with stage I and II uterine sarcoma (including 91 with carcinosarcoma) were randomly assigned to postoperative observation or adjuvant pelvic RT [50]. Although RT resulted in a significantly lower rate of any locoregional recurrence compared with observation (24 versus 47 percent, respectively), there was no improvement in PFS or OS. As reported, 24 percent of those patients receiving RT experienced "transient side effects," which were not clarified further. However, three cases of delayed gastrointestinal issues requiring surgical intervention were reported.

Stage IB to IV — For women with surgically staged IB to IV carcinosarcoma, we suggest combination platinum-based chemotherapy, rather than RT, single-agent chemotherapy, or observation [55], although we note that supporting evidence is limited. Additionally, in the absence of prospective data, the administration of multi-modality treatment (ie, chemotherapy plus RT) is investigational and should not be used outside of a clinical trial, as discussed below. (See 'Chemotherapy plus radiation therapy' below.)

The benefit of adjuvant chemotherapy over RT was shown in GOG 150, which included 206 women with uterine carcinosarcoma (70 percent of whom had disease outside of the uterus) who were randomly assigned to adjuvant treatment with whole-abdominal RT or cisplatin (20 mg/m²/day for four days), ifosfamide (1.5 g/m²/day for four days), plus mesna (120 mg/m² IV as a continuous infusion for four days) administered every three weeks for three cycles [55]. With a median duration of follow-up of five years, the main results were:

●The risk of recurrence was numerically lower among those treated with chemotherapy rather than whole-abdominal RT (hazard ratio [HR] 0.79, 95% CI 0.53-1.18); although this difference was not statistically significant, it would be clinically meaningful, if true.

●There was also a nonsignificant trend toward lower risk of death following treatment with chemotherapy rather than whole-abdominal RT (HR 0.71, 95% CI 0.48-1.05).

●There appeared to be a difference in the toxicity and recurrence patterns based on treatment:

•Compared with whole-abdominal RT, chemotherapy was associated with a higher incidence of vaginal recurrence (10 versus 4 patients) and serious (grade 3/4) neuropathy (9 versus 0 patients).

•Whole-abdominal RT was associated with a higher incidence of abdominal relapse (29 versus 19 patients), metastases to other distant sites (13 versus 10 patients), and grade 2 to 4 gastrointestinal events (10 versus 0).

When using chemotherapy, we suggest combination rather than single-agent treatment. This recommendation is supported by a meta-analysis of two trials (n = 373) comparing combination chemotherapy with single-agent ifosfamide in patients with advanced or recurrent disease [56]. Combination chemotherapy resulted in:

●A lower risk of death (HR 0.75, 95% CI 0.60-0.94) compared with single-agent treatment

●A lower risk of disease progression (HR 0.72, 95% CI 0.58-0.90)

●A higher risk of severe nausea and vomiting (relative risk [RR] 3.53, 95% CI 1.33-9.37), diarrhea (RR 1.51, 95% CI 0.31-7.52), and neuropathy (RR 1.59, 95% CI 0.99-2.55)

Choice of regimen — For patients with newly diagnosed uterine carcinosarcoma who have been surgically resected, we suggest carboplatin and paclitaxel. The data to support this recommendation are discussed below. (See 'Treatment of metastatic disease' below.)

Radiotherapy — There is no clear role for RT in the adjuvant setting [22,57]. In the largest report of 300 patients (53 percent treated with surgery and radiotherapy), adjuvant RT resulted in a lower rate of pelvic recurrence compared with surgery alone (28 versus 48 percent, respectively, p = 0.0002). However, there was no difference in five-year OS (27 versus 36 percent, p = 0.10) with adjuvant radiation [22].

Some experts prefer to administer vaginal brachytherapy (VBT) following surgery to reduce the risk of a local recurrence. This practice is based on the role of adjuvant VBT in the treatment of high-risk endometrial adenocarcinoma. (See "Adjuvant treatment of high-risk endometrial cancers", section on 'Treatment approach'.)

Chemotherapy plus radiation therapy — There are low-quality data to support the administration of combined chemotherapy plus RT (ie, combined-modality treatment, most commonly given as an RT sandwich or sequentially after chemotherapy) following surgery for carcinosarcoma [36,58-60]. Despite this, some experts prefer to combine chemotherapy plus VBT in order to reduce the risks of both local and distant metastases, which is similar to the approach for patients with advanced serous endometrial adenocarcinoma [61]. (See "Adjuvant treatment of high-risk endometrial cancers", section on 'Timing of pelvic radiation with chemotherapy'.)

Investigational approaches — Because patients with carcinosarcoma have a poor prognosis regardless of stage, we actively encourage patients to enroll on an appropriately designed clinical trial. Among these, agents that target the phosphatidylinositol-4,5-bisphosphate 3-kinase, protein kinase B, and mammalian target of rapamycin (PI3K/AKT/mTOR) pathway are in trials because of the finding that mutations involving this pathway are present in carcinosarcoma [62]. Other agents of interest include anti-angiogenic agents, histone deacetylase inhibitors, and Hedgehog inhibitors.

PROGNOSIS — Stage has consistently been identified as the most important prognostic factor for patients with uterine carcinosarcoma [43,63]. In a series of 121 patients with uterine carcinosarcoma who were treated with surgery followed by observation (29 percent), radiation therapy (RT) (38 percent), or chemotherapy (33 percent), regardless of treatment, five-year disease-specific survival rates were 59, 22, and 9 percent for women with stage I/II, III, or IV disease, respectively [36].

Other factors that have been associated with prognosis include depth of myometrial invasion, lymphovascular space invasion (LVSI), age, late-onset menopause, race (eg, Black women have a poorer prognosis than White women), single marital status, and the presence of gross residual disease [64-67].

Finally, for patients who had a preoperative elevation of CA 125, one report suggests that an elevated CA 125 postoperatively predicts a worse prognosis. In this study of 54 patients, an elevated postoperative CA 125 was also associated with a greater risk of death (hazard ratio 5.73, 95% CI 1.56-21.1) [31].

POST-TREATMENT SURVEILLANCE — The recommendations for posttreatment surveillance are identical to those for women treated for endometrial adenocarcinoma. (See "Overview of resectable endometrial carcinoma", section on 'Post-treatment considerations'.)

Preoperative elevation of cancer antigen (CA) 125 has been noted in some patients with carcinosarcoma, particularly in the context of advanced-stage disease (ie, stage III to IV) [31,68,69]. For these patients, monitoring CA 125 every three months following completion of primary treatment is a reasonable option. However, we prefer not to use changes in CA 125 for decisions regarding future treatment, given the lack of a survival benefit to support CA 125 surveillance in a large trial conducted in women with ovarian cancer. (See 'CA 125' above and "First-line chemotherapy for advanced (stage III or IV) epithelial ovarian, fallopian tube, and peritoneal cancer", section on 'Post-treatment surveillance'.)

TREATMENT OF METASTATIC DISEASE — The treatment of women with metastatic uterine carcinosarcoma is palliative. Therefore, we base therapeutic decisions on the presence of symptoms, pattern of recurrence, and whether or not prior treatment was administered.

Isolated vaginal recurrence — For women who experience an isolated vaginal recurrence, we suggest local therapy. The approach to treatment is similar to that of women with recurrent endometrial cancer. (See "Management of locoregional recurrence of endometrial cancer".)

Metastatic disease, no prior chemotherapy — For women with previously untreated metastatic carcinosarcoma, we suggest carboplatin and paclitaxel rather than an ifosfamide-based regimen. Carboplatin and paclitaxel has activity in carcinosarcoma and causes less toxicity than ifosfamide-based regimens.

The evidence to support this recommendation is discussed below.

Carboplatin plus paclitaxel — In Gynecologic Oncology Group (GOG) 261, carboplatin plus paclitaxel was compared with ifosfamide and paclitaxel in 637 women with newly diagnosed stage I to IV carcinosarcoma or recurrent chemotherapy-naϊve disease (536 with uterine and 101 patients with ovarian carcinosarcoma) [70]. The study met its primary objective, demonstrating that carboplatin and paclitaxel was not inferior to ifosfamide and paclitaxel in regards to overall survival (OS, 37 versus 29 months; hazard ratio [HR] 0.87, 90% CI 0.70-1.075) and resulted in a longer progression-free survival (PFS, 16 versus 12 months; HR 0.73). Although grade ≥3 toxicities occurred in 90 percent receiving carboplatin and paclitaxel versus 65 percent receiving ifosfamide and paclitaxel, most of the increased toxicity was hematologic, which was to be expected given that given that growth factor support was required in the ifosfamide and paclitaxel arm and only rarely used (n = 2) in the carboplatin plus paclitaxel arm. Confusion and genitourinary hemorrhage were worse among those receiving ifosfamide and paclitaxel.

The efficacy of carboplatin and paclitaxel was also previously demonstrated in GOG 232B [71], in which the response rate among women with chemotherapy-naϊve, stage III or IV, persistent, or recurrent carcinosarcoma was 54 percent, and the median PFS and OS were 8 and 15 months, respectively. [68,72].

Ifosfamide-based regimens — Two randomized trials have demonstrated improved response rates when ifosfamide is administered in combination with paclitaxel or cisplatin rather than as a single agent [73,74]. Although not compared directly, ifosfamide plus paclitaxel appears to be more effective and better tolerated compared with ifosfamide and cisplatin. Notably, decisions about optimal therapy have been driven by unacceptable rates of chemotherapy toxicity while the two have not been compared head to head, but serially to single-agent iphosphamide. These trials are reviewed below.

●Ifosfamide plus paclitaxel – A phase III trial conducted by the GOG (GOG 161) enrolled 179 women with previously untreated stage III to IV, recurrent or advanced disease and randomly assigned them to treatment using 21-day cycles of single-agent ifosfamide (2 g/m² daily for three days) or ifosfamide (1.6 g/m² daily for three days) plus paclitaxel (135 mg/m² on day 1) for up to eight cycles [73]. Compared with single-agent ifosfamide, ifosfamide plus paclitaxel resulted in a higher response rate (45 versus 29 percent) and a significant reduction in the risk of death (median OS, 14 versus 6 months; HR 0.69, 95% CI 0.49-0.97). Sensory neuropathy (grade 1 to 4) was significantly worse with combination therapy (30 versus 8 percent). Serious (grade 3/4) central nervous system toxicity was reported in three patients treated with single-agent ifosfamide and five patients treated with the combination.

●Ifosfamide plus cisplatin – A separate phase III trial conducted by the GOG (GOG 108) compared ifosfamide with or without cisplatin [74]. Compared with ifosfamide alone, ifosfamide plus cisplatin resulted in a higher response rate (54 versus 36 percent) and a significant improvement in PFS (median six versus four months, relative risk [RR] 0.73, p = 0.02). However, there was no improvement in OS (median, 9 versus 8 months, RR 0.80, p = 0.07). Combination therapy resulted in more serious (grade 3/4) events compared with single-agent ifosfamide, including anemia (17 versus 8 percent) and peripheral neuropathy (12 versus 1 percent). Both treatments were associated with serious central nervous system toxicity (14 versus 19 percent).

Metastatic disease, prior chemotherapy — For women who develop disease progression following adjuvant combination chemotherapy, we suggest single-agent chemotherapy. The options include agents with activity in the treatment of metastatic endometrial cancer. (See "Initial treatment of metastatic endometrial cancer".)

A choice between agents should be individualized, taking into account the prior toxicities the patient has experienced and the toxicities associated with an alternative agent. However, in light of the limited evidence of a survival benefit with treatment, hospice care is also reasonable, particularly for patients at risk for significant treatment-related toxicity, and should be considered early (See "Hospice: Philosophy of care and appropriate utilization in the United States".)

All patients with metastatic carcinosarcoma should be offered palliative care. (See "Benefits, services, and models of subspecialty palliative care".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Uterine cancer".)

SUMMARY AND RECOMMENDATIONS

●Uterine carcinosarcoma is classified as a metaplastic carcinoma with elements of carcinoma and sarcoma. (See 'Introduction' above.)

●Endometrial sampling should be performed in any woman in whom there is a high clinical suspicion for uterine sarcoma (ie, presenting with a large uterine mass, bleeding, and/or pain). However, a negative sampling does not exclude uterine carcinosarcoma, and further evaluation should be performed. (See 'Presentation' above and 'Evaluation of suspected disease' above.)

●Uterine carcinosarcoma is surgically staged according to the 2017 International Federation of Gynecology and Obstetrics (FIGO)/Tumor, Node, Metastasis (TNM) classification system, the same system as for endometrial carcinoma (table 1). (See 'Staging' above.)

●For women with advanced carcinosarcoma without extra-abdominal disease, we suggest surgical cytoreduction (Grade 2C). (See 'Surgery' above.)

●There are low-quality data to inform the management of women with surgically staged IA carcinosarcoma. Some of our experts suggest observation rather than chemotherapy, while others suggest adjuvant chemotherapy. (See 'Adjuvant therapy' above.)

●For women with stage IB to IVA carcinosarcoma, we suggest adjuvant chemotherapy rather than adjuvant RT (Grade 2C), although we acknowledge limitations in supporting evidence. Additionally, some experts also administer vaginal brachytherapy to these patients to reduce the risk of a local recurrence. When administering chemotherapy, we suggest combination platinum-based chemotherapy rather than single-agent treatment (Grade 2B). (See 'Adjuvant therapy' above.)

●We suggest carboplatin and paclitaxel in the adjuvant setting rather than an ifosfamide-based combination (Grade 2C). (See 'Choice of regimen' above.)  

●For women with an isolated vaginal recurrence, we suggest local therapy (Grade 2C). The approach is similar to the treatment of locally recurrent endometrial cancer. (See "Management of locoregional recurrence of endometrial cancer".)

●For women with previously untreated metastatic (stage IVB) or recurrent carcinosarcoma, we suggest carboplatin and paclitaxel rather than an ifosfamide-based regimen (Grade 2C). (See 'Metastatic disease, no prior chemotherapy' above.)

●For women who experience disease progression following receipt of adjuvant or first-line chemotherapy, we suggest single-agent chemotherapy (Grade 2B). A choice between agents should be individualized, taking into account the prior toxicities the patient has experienced and the toxicities associated with an alternative agent. However, in light of the limited evidence of a survival benefit with treatment, hospice care is also reasonable, particularly for patients at risk for significant treatment-related toxicity. (See 'Metastatic disease, prior chemotherapy' above.)