Abstract
Purpose
To propose a categorization model of uterine carcinosarcoma (UCS) based on tumor cell types (carcinoma and sarcoma) and sarcoma dominance.
Methods
This secondary analysis of a prior multicenter retrospective study examined 889 cases of UCS with available histologic evaluation. Based on survival outcome, cases were clustered into three groups: low-grade carcinoma with nondominant homologous sarcoma [type A, n = 96 (10.8%)], (1) low-grade carcinoma with heterologous sarcoma or any sarcoma dominance and (2) high-grade carcinoma with nondominant homologous sarcoma [type B, n = 412 (46.3%)], and high-grade carcinoma with heterologous sarcoma or any sarcoma dominance [type C, n = 381 (42.9%)]. Tumor characteristics and outcome were examined based on the categorization.
Results
Women in type C category were more likely to be older, obese, and Caucasian, whereas those in type A category were younger, less obese, Asian, and nulligravid (all P < 0.01). Type C tumors were more likely to have metastatic implants, large tumor size, lymphovascular space invasion with sarcoma cells, and higher lymph node ratio, whereas type A tumors were more likely to be early-stage disease and small (all P < 0.05). On multivariate analysis, tumor categorization was independently associated with progression-free survival (5-year rates: 70.1% for type A, 48.3% for type B, and 35.9% for type C, adjusted P < 0.01) and cause-specific survival (5-year rates: 82.8% for type A, 63.0% for type B, and 47.1% for type C, adjusted P < 0.01).
Conclusion
Characteristic differences in clinicopathological factors and outcomes in UCS imply that different underlying etiologies and biological behaviors may be present, supporting a new classification system.
Uterine carcinosarcoma (UCS) is a rare type of high-grade endometrial cancer, however the proportion of UCS has been gradually increasing among endometrial cancer in the USA, exceeding 5% in recent years.1 Histologically, UCS contains both carcinoma and sarcoma components in the uterine tumor site, with the sarcoma element being dedifferentiated from the carcinoma component.2,3 Both the carcinoma and sarcoma components have important prognostic implications pertaining to treatment response and survival in women with UCS.4
Additionally, the dominant tumor component has been shown to be a salient tumor factor in UCS. A recent analysis showed that the dominant pattern in the uterine tumor, either carcinoma or sarcoma, is an independent predictor of survival in UCS.4 In this analysis, sarcoma dominance had the largest impact on survival among the uterine tumor factors in UCS.4 Type of sarcoma dominance (homologous vs heterologous) also impacts survival.5
UCS is understudied due to its rarity and the complexity of the tumor factors (carcinoma, sarcoma, and dominant component), making it difficult to interpret the true effects of histological components on survival. For UCS, it is hypothesized that the carcinoma component is the main driver of tumor progression, and the sarcoma component and sarcoma dominance are secondary factors, altering tumor behavior. However, the association between combination patterns of these three tumor factors and survival of UCS has not been examined.
The objective of this study is to examine the clinicopathological pattern and survival based on carcinoma, sarcoma, and dominant components in UCS, with the goal of proposing a risk-based categorization model of UCS.
MATERIALS AND METHODS
Eligibility Criteria
We utilized the previously organized database from a large multicenter retrospective study to conduct this secondary analysis. Previously, we examined consecutive women with stage I–IV UCS who underwent primary hysterectomy-based surgical treatment in 26 institutions between 1993 and 2013 (906 cases).4–11 Institutional review board approval was obtained at each participating site. By querying this database, we examined cases with information available concerning all three tumor factors (carcinoma and sarcoma types, and dominant component).
Clinical Information
Among eligible cases, we abstracted information for patient demographics, tumor characteristics, surgical performance, postoperative treatment type, and survival. Patient demographics included age, race/ethnicity, country, body mass index (BMI, kg/m2), pregnancy history, history of tamoxifen use, history of pelvic irradiation, and preoperative cancer antigen 125 (CA-125, IU/L) levels. Surgical performance included residual disease at the end of surgery.
Tumor characteristics abstracted were carcinoma type, sarcoma type, sarcoma dominance, cancer stage, tumor size, depth of myometrial tumor invasion, lymphovascular space invasion (LVSI), and pelvic/paraaortic lymph node status [rate of metastasis among staged cases, and lymph node ratio (LNR) among metastatic cases]. Postoperative treatment types included use of postoperative chemotherapy and/or radiotherapy. Survival outcomes were abstracted for progression-free survival (PFS) and cause-specific survival (CSS).
Histopathology Evaluation
In all the cases, archived histopathology slides for hematoxylin–eosin stains and immunohistochemistry stains, if available, were reviewed at each institution. Pathologists were blinded to clinical information. In a comprehensive evaluation, carcinoma type, sarcoma type, sarcoma dominance, and LVSI cell types were assessed as described previously (Supplementary Method).4
Tumor Factor-Based Categorization
Among 906 cases of UCS in the database, there were 889 cases with available information regarding histologic component and sarcoma dominance. We first plotted the crude PFS results based on the combination patterns of carcinoma type, sarcoma type, and sarcoma dominance (Supplementary Fig. S1). Based on the survival outcomes, these eight groups were further clustered into three independent categories (Table 1).
TABLE 1.
Proposed grouping criteria based on tumor factors
| Carcinoma type | Sarcoma type | Dominant component | No. | 2-Year PFS (%) | 5-Year PFS (%) | Proposed categorization |
|---|---|---|---|---|---|---|
| Low-grade | Homologous | Carcinoma | 96 (10.8%) | 77.0 | 70.1 | Type A |
| Low-grade | Homologous | Sarcoma | 65 (7.3%) | 57.4 | 47.4 | Type B |
| Low-grade | Heterologous | Carcinoma | 40 (4.5%) | 54.1 | 54.1 | |
| Low-grade | Heterologous | Sarcoma | 52 (5.8%) | 55.5 | 48.8 | |
| High-grade | Homologous | Carcinoma | 255 (28.7%) | 59.2 | 47.5 | |
| High-grade | Homologous | Sarcoma | 110 (12.4%) | 47.5 | 41.5 | Type C |
| High-grade | Heterologous | Carcinoma | 134 (15.1%) | 45.5 | 32.6 | |
| High-grade | Heterologous | Sarcoma | 137 (15.4%) | 48.2 | 35.3 |
Risk stratification based on three tumor factors: carcinoma component, sarcoma component, and sarcoma dominance
No. number, PFS progression-free survival
In this pilot exploratory study, we termed these three categories as type A, B, and C tumors for the purpose of convenience. Type A tumors are defined as low-grade carcinoma with nondominant homologous sarcoma. Type B tumors are defined as (1) low-grade carcinoma with heterologous sarcoma or any sarcoma dominance or (2) high-grade carcinoma with nondominant homologous sarcoma. Type C tumors are defined as high-grade carcinoma with heterologous sarcoma or any sarcoma dominance (Table 1).
Study Definition
Obesity was defined as BMI ≥ 30 kg/m2. We reclassified cancer stage based on the 2009 International Federation of Gynecology and Obstetrics (FIGO) definition. Low-grade carcinoma (grade 1–2 endometrioid) and high-grade carcinoma (grade 3 endometrioid, serous, clear cell, undifferentiated, and mixed) were defined based on a prior study.4 Similarly, sarcoma types were grouped as homologous (endometrial stromal sarcoma, leiomyosarcoma, fibrosarcoma, and undifferentiated sarcoma) or heterologous (rhabdomyosarcoma, osteosarcoma, chondrosarcoma, and liposarcoma).4
LVSI types were grouped based on presence of sarcoma cells within the lymphatic or vascular capillary, as described previously (carcinoma alone, sarcoma with or without carcinoma, and none).4 LNR was defined as percent proportion of lymph nodes containing tumor cells among resected lymph nodes. PFS was defined as the time interval between hysterectomy and first recurrence/progression of disease or death due to UCS. CSS was defined as the time interval between hysterectomy and death due to UCS. Cases without these survival events at last follow-up were censored.
Statistical Considerations
The primary objective of this study is to outline the clinicopathological characteristics across type A–C UCS. The secondary objective of this study is to examine the independent association of the tumor categorization with survival.
The Kaplan–Meier method was used to plot survival curves, and the log-rank test was used to assess statistical difference among the curves. In this study, an association of tumor categorization and survival was adjusted for clinicopathological factors in the four models based on a manner of practical treatment intervention for UCS. A Cox proportional hazard regression test was used for this modeling, and the relative magnitude of statistical significance is expressed as hazard ratio (HR) and 95% confidence interval (CI). Stepwise assessments were performed to examine the durability of independent association in each set of the adjustment model.
In the first model, an association of tumor categorization and survival was adjusted for patient demographics. In the second model, the association was further adjusted for surgical performance. In the third model, tumor factors were added to the adjustment model. In the fourth model, postoperative treatment types were added to the third model. The variables and their cut-point in the four models were based on a priori survival factors, as described previously.4
Sensitivity analysis was performed to examine the utility of tumor categorization on postoperative treatment response. Among women with stage I–III disease, survival was examined based on postoperative treatment type (none, chemotherapy alone, radiotherapy alone, and both chemotherapy and radiotherapy). These groupings have previously been shown to be possible effective postoperative therapy choices.12–14
P < 0.05 was considered statistically significant (two-sided hypothesis). Statistical Package for Social Science software (IBM SPSS, version 24.0, Armonk, NY, USA) was used for all analyses. We consulted the STROBE guidelines to describe the results of retrospective observational cohort studies.15
RESULTS
The two most common tumor categories were type B (n = 412, 46.3%) and type C (n = 381, 42.9%), while type A tumor was the least common category (n = 96, 10.8%). Patient demographics are presented in Table 2.
TABLE 2.
Clinicopathological demographics based on risk stratification
| Characteristic | Type A | Type B | Type C | P value |
|---|---|---|---|---|
| Number | n = 96 | n = 412 | n = 381 | |
| Age (years) | 59 (IQR 15) | 62 (IQR 14) | 65 (IQR 14) | < 0.001 |
| <60 | 49 (51.0%) | 148 (35.9%) | 95 (24.9%) | |
| ≥ 60 | 47 (49.0%) | 264 (54.1%) | 286 (75.1%) | |
| Race | < 0.001 | |||
| Caucasian | 16 (17.0%) | 120 (29.6%) | 139 (37.2%) | |
| African | 1 (1.1%) | 28 (6.9%) | 50 (13.4%) | |
| Hispanic | 2 (2.1%) | 7 (1.7%) | 13 (3.5%) | |
| Asian | 73 (77.7%) | 243 (59.9%) | 163 (43.6%) | |
| Others | 2 (2.1%) | 8 (2.0%) | 9 (2.4%) | |
| Country | < 0.001 | |||
| USA | 24 (25.0%) | 171 (41.5%) | 223 (58.5%) | |
| Japan | 72 (75.0%) | 241 (58.5%) | 158 (41.5%) | |
| BMI (kg/m2) | 22.6 (IQR 5.6) | 23.3 (IQR 5.9) | 23.9 (IQR 7.2) | 0.008 |
| < 30 | 77 (81.1%) | 307 (76.6%) | 260 (73.7%) | |
| ≥ 30 | 18 (18.9%) | 94 (23.4%) | 93 (26.3%) | |
| Gravida | 0.002 | |||
| Nulli | 27 (28.4%) | 57 (14.1%) | 54 (14.8%) | |
| Multi | 68 (71.6%) | 346 (85.9%) | 310 (85.2%) | |
| History of tamoxifen use | 0.36 | |||
| No | 91 (95.8%) | 389 (95.6%) | 356 (93.4%) | |
| Yes | 4 (4.2%) | 18 (4.4%) | 25 (6.6%) | |
| Prior pelvic irradiation | 0.87 | |||
| No | 95 (99.0%) | 407 (98.8%) | 375 (98.4%) | |
| Yes | 1 (1.0%) | 5 (1.2%) | 6 (1.6%) | |
| Preop. CA-125 (IU/L) | 20.5 (IQR 28) | 22 (IQR 42) | 28 (IQR 52) | 0.31 |
| <30 | 47 (59.5%) | 176 (61.3%) | 127 (52.9%) | |
| ≥ 30 | 32 (40.5%) | 111 (38.7%) | 113 (47.1%) | |
| Residual disease | 0.32 | |||
| No | 86 (91.5%) | 355 (89.6%) | 318 (86.9%) | |
| Yes | 8 (8.5%) | 41 (10.4%) | 48 (13.1%) | |
| Postop. radiotherapy | 0.10 | |||
| No | 77 (80.2%) | 312 (76.7%) | 268 (71.3%) | |
| Yes | 19 (19.8%) | 95 (23.3%) | 108 (28.7%) | |
| Postop. chemotherapy | 0.40 | |||
| No | 27 (28.1%) | 129 (31.7%) | 131 (34.7%) | |
| Yes | 69 (71.9%) | 278 (68.3%) | 246 (65.3%) | |
| Stage | < 0.001 | |||
| I | 58 (60.4%) | 220 (53.4%) | 159 (41.7%) | |
| II | 8 (8.3%) | 35 (8.5%) | 21 (5.5%) | |
| III | 23 (24.0%) | 113 (27.4%) | 135 (35.4%) | |
| IV | 7 (7.3%) | 44 (10.7%) | 66 (17.3%) | |
| Tumor size (cm) | < 0.001 | |||
| <5 | 47 (50.0%) | 157 (39.7%) | 109 (29.1%) | |
| 5–9.9 | 42 (44.7%) | 200 (50.6%) | 191 (50.9%) | |
| ≥ 10 | 5 (5.3%) | 38 (9.6%) | 75 (20.0%) | |
| Myometrial invasion | 0.22 | |||
| Inner half | 56 (59.6%) | 216 (52.4%) | 187 (49.6%) | |
| Outer half | 38 (40.4%) | 196 (47.6%) | 190 (50.4%) | |
| LVSI | 0.013 | |||
| None | 46 (50.0%) | 153 (39.5%) | 148 (42.2%) | |
| Carcinoma only | 41 (44.6%) | 187 (48.3%) | 143 (40.7%) | |
| Sarcoma | 5 (5.4%) | 47 (12.1%) | 60 (17.1%) | |
| PLN metastasisa | 0.014 | |||
| No | 62 (82.8%) | 239 (76.1%) | 180 (68.4%) | |
| Yes | 12 (16.2%) | 75 (23.9%) | 83 (31.6%) | |
| PAN metastasisa | 0.007 | |||
| No | 35 (83.3%) | 168 (85.3%) | 110 (71.9%) | |
| Yes | 7 (16.7%) | 29 (14.7%) | 43 (28.1%) | |
| Lymph node ratio (%)b | ||||
| PLN | 7.6 (IQR 6.4) | 11.8 (IQR 28.3) | 21.4 (IQR 41.7) | 0.028 |
| PAN | 9.5 (IQR 61.0) | 45.0 (IQR 89.1) | 50.0 (IQR 80.0) | 0.26 |
Median (IQR) or number (percent per column) is shown. Kruskal–Wallis H test or Chi square test for P values. Significant covariates are emboldened
IQR interquartile range, BMI body mass index, CA-125 cancer antigen 125, LVSI lymphovascular space invasion, PLN pelvic lymph node, PAN paraaortic lymph node
Examined only staged cases
Examined only positive lymph node cases
Women in the type C category were more likely to be older, whereas those in the type A category were more likely to be younger (median age: type A 59, type B 62, and type C 65 years, P < 0.001). Women in the type C category were more likely to be Caucasian (37.2%), whereas those in the type A category were more likely to be Asian (77.7%) (P < 0.001). Women in the type C category were more likely to be obese, whereas those in the type A category were least likely (type A 18.9%, type B 23.4%, and type C 26.3%, P = 0.008). Among the three groups, women in the type A category were most likely to be nulligravida (type A 28.4%, type B 14.1%, and type C 14.8%, P = 0.002).
Tumor characteristics were examined across the three categories (Table 2). Type C tumors were more likely to be advanced stage, whereas type A tumors were least likely (proportion of stage III–IV disease: type A 31.3%, type B 38.1%, and type C 52.7%, P < 0.001). On the contrary, type A tumors were more likely to be confined to the uterus, whereas type C tumors were least likely (type A 60.4%, type B 53.4%, and type C 41.7%, P < 0.001). Type C tumors were more likely to be large, whereas type A tumors were least likely (proportion of tumor ≥ 10 cm: type A 5.3%, type B 9.6%, and type C 20.0%, P < 0.001).
Prevalence of any LVSI was similar across the three categories (P = 0.15). However, type C tumors were more likely to have LVSI with sarcoma cells, whereas type A tumors were least likely (type A 5.4%, type B 12.1%, and type C 17.1%, P = 0.013). Among staged cases, type C tumors were more likely to have pelvic/paraaortic nodal metastasis, whereas type A tumors were least likely (both P < 0.05; Table 2). Moreover, among pelvic nodal metastatic cases, type C tumors had the highest LNR, whereas type A tumors had the lowest (type A 7.6%, type B 11.8%, and type C 21.4%, P = 0.028).
The median follow-up time of censored cases was 38.6 (interquartile range, 58.5) months. During the follow-up period, there were 419 survival events recorded. On univariate analysis, tumor category was significantly associated with PFS (5-year rates: type A 70.1%, type B 48.3%, and type C 35.9%, P < 0.001; Fig. 1a) and CSS (5-year rates: type A 82.8%, type B 63.0%, and type C 47.1%, P < 0.001; Fig. 1b).
FIG. 1.
Survival curves based on risk stratification. Log-rank test for P values. Definitions of risk groups are shown in Table 1. a Progression-free survival and b cause-specific survival
On multivariate analysis (Table 3), the type C category was independently associated with decreased PFS (adjusted HR 2.38, 95% CI 1.49–3.78) and CSS (adjusted HR 2.04, 95% CI 1.17–3.53) compared with the type A category after adjusting for patient demographics, surgical factors, tumor factors, and postoperative treatment type (both adjusted P < 0.05). Similarly, the type B category was independently associated with decreased PFS and CSS compared with the type A category (both adjusted P < 0.05).
TABLE 3.
Association models for survival outcomes
| Adjustment model | Progression-free survival) |
Cause-specific survival |
||
|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | |
| Unadjusted | ||||
| Type A | 1 | 1 | ||
| Type B | 2.28 (1.48–3.51) | < 0.001 | 2.18 (1.30–3.67) | 0.003 |
| Type C | 3.16 (2.05–4.86) | < 0.001 | 3.29 (1.96–5.51) | < 0.001 |
| Patient demographics | ||||
| Type A | 1 | 1 | ||
| Type B | 2.08 (1.35–3.22) | 0.001 | 1.97 (1.17–3.33) | 0.011 |
| Type C | 2.79 (1.80–4.33) | < 0.001 | 2.89 (1.71–4.88) | < 0.001 |
| Patient demographics, surgical factor | ||||
| Type A | 1 | 1 | ||
| Type B | 1.96 (1.25–3.06) | 0.003 | 1.74 (1.03–2.95) | 0.038 |
| Type C | 2.63 (1.68–4.13) | < 0.001 | 2.50 (1.47–4.26) | 0.001 |
| Patient demographics, surgical factor, tumor factors | ||||
| Type A | 1 | 1 | ||
| Type B | 1.99 (1.26–3.15) | 0.003 | 1.86 (1.08–3.20) | 0.025 |
| Type C | 2.54 (1.60–4.03) | < 0.001 | 2.35 (1.36–4.06) | 0.002 |
| Patient demographics, surgical factor, tumor factors, postop. treatment types | ||||
| Type A | 1 | 1 | ||
| Type B | 2.05 (1.30–3.25) | 0.002 | 1.84 (1.07–3.18 | 0.029 |
| Type C | 2.38 (1.49–3.78) | < 0.001 | 2.04 (1.17–3.53) | 0.011 |
Cox proportional hazard regression models for HRs and P values. Significant covariates are emboldened. Patient demographics included age (< 60 vs ≥ 60 years) and race/ethnicity (Caucasian, African, Hispanic, Asian, and others). Surgical factor included residual disease at surgery (yes vs no). Tumor factors included cancer stage (I, II, III, or IV), tumor size (< 5 vs ≥ 5 cm), depth or myometrial invasion (inner half vs outer half), and lymphovascular space invasion (yes vs no). Postoperative treatment types included radiotherapy (yes vs no) and chemotherapy (yes vs no)
HR hazard ratio, CI confidence interval
Stage-specific survival was examined (Supplementary Table S1). Women with stage I type A tumors had a 5-year CSS rate exceeding 90%, whereas those with stage II–IV type C tumors had nearly 30%. Absolute PFS difference among the three types was 28.5% (range 54.8–83.3%) for stage I disease and 25.5% (range 21.7–47.2%) for stage II–IV disease.
An association between tumor category and postoperative treatment response was examined in women with stage I–III disease (Table 4; Supplementary Table S2). When compared with no adjuvant therapy, chemotherapy alone or chemotherapy with radiotherapy significantly reduced recurrence risk and disease mortality in type B and C tumors (both P < 0.05). When compared with radiotherapy alone, adding chemotherapy to radiotherapy reduced recurrence risk of type C tumors (HR 0.44, 95% CI 0.22–0.85, P = 0.015). When compared with chemotherapy alone, adding radiotherapy to chemotherapy significantly reduced recurrence risk (HR 0.53, 95% CI 0.34–0.85), and disease mortality (HR 0.38, 95% CI 0.21–0.70) in type C tumors (both P < 0.01).
TABLE 4.
Progression-free survival based on adjuvant therapy types for stage I–III disease (n = 772)
| Characteristic | Versus no treatment |
Versus radiotherapy alone |
Versus chemotherapy alone |
|||
|---|---|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | HR (95% CI) | P value | |
| Type A category | ||||||
| None | 1 | 0.68 (0.13–3.50) | 0.64 | 1.52 (0.50–4.67) | 0.46 | |
| Radiotherapy only | 1.48 (0.29–7.64) | 0.64 | 1 | 2.25 (0.48–10.6) | 0.31 | |
| Chemotherapy only | 0.66 (0.21–2.01) | 0.46 | 0.44 (0.09–2.10) | 0.31 | 1 | |
| Botha | 1.19 (0.28–4.99) | 0.81 | 0.80 (0.13–4.82) | 0.81 | 1.81 (0.48–6.82) | 0.38 |
| Type B category | ||||||
| None | 1 | 1.46 (0.76–2.82) | 0.26 | 1.77 (1.22–2.57) | 0.003 | |
| Radiotherapy only | 0.69 (0.35–1.32) | 0.26 | 1 | 1.21 (0.64–2.28) | 0.56 | |
| Chemotherapy only | 0.57 (0.39–0.82) | 0.003 | 0.83 (0.44–1.56) | 0.56 | 1 | |
| Botha | 0.51 (0.31–0.84) | 0.008 | 0.74 (0.36–1.53) | 0.42 | 0.90 (0.56–1.44) | 0.66 |
| Type C category | ||||||
| None | 1 | 1.23 (0.67–2.26) | 0.50 | 1.51 (1.05–2.18) | 0.027 | |
| Radiotherapy only | 0.81 (0.44–1.49) | 0.50 | 1 | 1.23 (0.68–2.23) | 0.50 | |
| Chemotherapy only | 0.66 (0.46–0.95) | 0.027 | 0.81 (0.45–1.48) | 0.50 | 1 | |
| Botha | 0.35 (0.22–0.57) | < 0.001 | 0.44 (0.22–0.85) | 0.015 | 0.53 (0.34–0.85) | 0.008 |
Cox proportional hazard regression models for unadjusted HRs and P values. Significant covariates are emboldened
HR hazard ratio, CI confidence interval
Both chemotherapy and radiotherapy
DISCUSSION
We previously examined the combination patterns of carcinoma/sarcoma and found that survival outcomes differ based on histology type.4 We since learned that sarcoma dominance has significant prognostic implications,4,5 so this current analysis examines the three principal factors in patient outcomes for UCS (carcinoma, sarcoma, and sarcoma dominance).
Prior analysis was limited to evaluation of histologic pattern on chemotherapy response, while the effects of histologic pattern on response to radiation therapy were not investigated.4 Given that sarcoma cells appear to be sensitive to radiotherapy,6,16 and that a multimodality approach with chemotherapy and radiotherapy is common in postoperative management of UCS,17–19 this investigation adds useful information on the role of radiation therapy based on histology and dominant patterns in UCS.
This study found that type A UCS represents a less aggressive tumor, whereas type C UCS exhibits more aggressive behavior. Additionally, patient baseline characteristics were largely different across the three defined types of UCS. Our results show that UCS may be better categorized by histologic type and sarcoma dominance rather than as a single disease entity. We suggest that there may be various underlying etiologies, each with unique background characteristics.
Clinically, young Asian women have more favorable UCS tumors (type A), whereas old Caucasian women may have aggressive UCS tumors (type C). Histologically, early-stage type A UCS can have survival almost comparable to that of low-grade endometrial cancer (5-year CSS rate 90.1%), while type B UCS, the most common type (high-grade carcinoma with nondominant homologous sarcoma, 28.7%), and advanced-stage type C tumors belong to a group with much worse survival (5-year CSS rate 33.9%).
Recent molecular analyses have shown that UCS originally arises from endometrial cancer by means of epithelial–mesenchymal transition (EMT) within the tumor, and that the signature of EMT is more prominent in heterologous-type UCS than in the homologous counterpart.2,3 Further study is warranted to determine what triggers the initiation of EMT in endometrial cancer causing development of UCS; and additionally what role this plays, if any, in causing homologous versus heterologous dedifferentiation. The results of our analysis may be useful in providing a link between basic clinicopathological characteristics and molecular characteristics in UCS.
This study is the first to propose a categorization of UCS utilizing relevant clinical information with a large sample size and comprehensive histopathology review, enhancing the study’s quality and reliability. However, this is a retrospective study with the inherent potential for confounding factors missing in the analysis. For example, we lacked information regarding the exact indications for postoperative therapy, introducing the possibility of selection bias. The majority of the study population was Asian, and generalizability to other population needs to be examined.
A major limitation in the interpretation of our results is the lack of central pathology review with predefined criteria (definitions and guidelines utilized for histopathology evaluation). Additionally, interobserver agreement among pathologists has not been validated. Many high-grade endometrial cancers share similar clinicopathological characteristics, making clear diagnosis difficult.20 This study did not contain a molecular analysis. Given recent analyses of molecular classifications in endometrial cancer,21 interactions between our histology pattern-based categorization and molecular characteristics are of interest for further exploration.2,3
This study proposes a new categorization of UCS that can facilitate communication between clinicians and pathologists with regards to risk stratification. Furthermore, this study elucidates the potential benefits of postoperative therapy with chemotherapy and/or radiotherapy in type B–C UCS. This is particularly applicable in type C UCS, where addition of radiotherapy to chemotherapy seems to have added benefit. As these findings were only demonstrated in retrospective analysis, further study with a prospective design is necessary to confirm this association.
In summary, we showed that UCS may represent several disease entities rather than a single one, and that survival in UCS can vary widely based on tumor characteristics. While UCS is a type of high-risk endometrial cancer, it is paramount to recognize that certain subtypes of UCS behave similarly to low-risk cancers. Our preliminary attempt at UCS classification has led to several useful observations that warrant further validation and investigation.
Supplementary Material
Acknowledgments
FUNDING Ensign Endowment for Gynecologic Cancer Research (K.M.)
Footnotes
DISCLOSURE The authors declare that there is no conflict of interest for all authors.
Electronic supplementary material The online version of this article (https://doi.org/10.1245/s10434-018-6695-z) contains supplementary material, which is available to authorized users.
REFERENCES
- 1.Matsuo K, Ross MS, Machida H, Blake EA, Roman LD. Trends of uterine carcinosarcoma in the United States. J Gynecol Oncol. 2018;29:e22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cherniack AD, Shen H, Walter V, et al. Integrated molecular characterization of uterine carcinosarcoma. Cancer Cell. 2017;31:411–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Zhao S, Bellone S, Lopez S, et al. Mutational landscape of uterine and ovarian carcinosarcomas implicates histone genes in epithelial-mesenchymal transition. Proc Natl Acad Sci U S A. 2016;113:12238–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Matsuo K, Takazawa Y, Ross MS, et al. Significance of histologic pattern of carcinoma and sarcoma components on survival outcomes of uterine carcinosarcoma. Ann Oncol. 2016;27:1257–66. [DOI] [PubMed] [Google Scholar]
- 5.Matsuo K, Takazawa Y, Ross MS, et al. Characterizing sarcoma dominance pattern in uterine carcinosarcoma: homologous versus heterologous element. Surg Oncol. 2018;27:433–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Matsuo K, Omatsu K, Ross MS, et al. Impact of adjuvant therapy on recurrence patterns in stage I uterine carcinosarcoma. Gynecol Oncol. 2017;145:78–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Matsuo K, Ross MS, Bush SH, et al. Tumor characteristics and survival outcomes of women with tamoxifen-related uterine carcinosarcoma. Gynecol Oncol. 2017;144:329–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Matsuo K, Johnson MS, Im DD, et al. Survival outcome of women with stage IV uterine carcinosarcoma who received neoadjuvant chemotherapy followed by surgery. J Surg Oncol. 2017;117:488–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Matsuo K, Ross MS, Im DD, et al. Significance of venous thromboembolism in women with uterine carcinosarcoma. Gynecol Oncol. 2018;148:267–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Matsuo K, Ross MS, Yunokawa M, et al. Salvage chemotherapy with taxane and platinum for women with recurrent uterine carcinosarcoma. Gynecol Oncol. 2018;147:565–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Matsuo K, Takazawa Y, Ross MS, et al. Significance of lymphovascular space invasion by sarcomatous component in uterine carcinosarcoma. Ann Surg Oncol. 2018;25:2756–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hogberg T, Signorelli M, de Oliveira CF, et al. Sequential adjuvant chemotherapy and radiotherapy in endometrial cancer-results from two randomised studies. Eur J Cancer. 2010;46:2422–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.de Boer SM, Powell ME, Mileshkin L, et al. Adjuvant chemoradiotherapy versus radiotherapy alone for women with high-risk endometrial cancer (PORTEC-3): final results of an international, open-label, multicentre, randomised, phase 3 trial. Lancet Oncol. 2018;19:295–09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.https://meetinglibrary.asco.org/record/147011/abstract Accessed 24 Jan 2018.
- 15.von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335:806–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sampath S, Gaffney DK. Role of radiotherapy treatment of uterine sarcoma. Best Pract Res Clin Obstet Gynaecol. 2011;25:761–72. [DOI] [PubMed] [Google Scholar]
- 17.Rauh-Hain JA, Starbuck KD, Meyer LA, et al. Patterns of care, predictors and outcomes of chemotherapy for uterine carcinosarcoma: a National Cancer Database analysis. Gynecol Oncol. 2015;139:84–9. [DOI] [PubMed] [Google Scholar]
- 18.Manzerova J, Sison CP, Gupta D, et al. Adjuvant radiation therapy in uterine carcinosarcoma: a population-based analysis of patient demographic and clinical characteristics, patterns of care and outcomes. Gynecol Oncol. 2016;141:225–30. [DOI] [PubMed] [Google Scholar]
- 19.Seagle BL, Kanis M, Kocherginsky M, Strauss JB, Shahabi S. Stage I uterine carcinosarcoma: Matched cohort analyses for lymphadenectomy, chemotherapy, and brachytherapy. Gynecol Oncol. 2017;145:71–7. [DOI] [PubMed] [Google Scholar]
- 20.Voss MA, Ganesan R, Ludeman L, McCarthy K, Gornall R, Schaller G, Wei W, Sundar S. Should grade 3 endometrioid endometrial carcinoma be considered a type 2 cancer-a clinical and pathological evaluation. Gynecol Oncol. 2012;124:15–20. [DOI] [PubMed] [Google Scholar]
- 21.Kandoth C, Schultz N, Cherniack AD, et al. Integrated genomic characterization of endometrial carcinoma. Nature 2013;497: 67–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.