Abstract
Objective
To compare the outcomes of women with stage III uterine cancer treated with chemotherapy alone, external beam radiation alone, and combination chemotherapy and radiation.
Methods
The National Cancer Database was used to identify women with stage III endometrioid, serous, and clear cell uterine cancer treated with either chemotherapy (with or without vaginal brachytherapy) alone, external beam radiation (with or without brachytherapy) or combination chemotherapy and external beam radiation (with or without vaginal brachytherapy) from 2004–2015. Survival was estimated using Cox proportional hazards models and adjusted survival curves after propensity score analysis using inverse probability of treatment weighting to balance the clinical and demographic characteristics between the cohorts.
Results
Of the 20,873 patients identified, 9,456 (45.3%) received chemotherapy alone, 2,417 (11.6%) were treated with radiation alone, and 9,000 (43.1%) received chemotherapy in combination with external beam radiation. Use of combination therapy was 33.0% in 2004, and then rose to 50.5% in 2015. The mortality of the cohort was 33.1% and median survival was 115 months. Within the cohort, combination therapy was associated with a 23% reduction in mortality (HR=0.77; 95% CI, 0.73–0.80) compared to chemotherapy alone. Similar findings of decreased mortality were noted in subgroup analyses for both stage IIIA (HR=0.81; 95% CI, 0.68–0.98) and stage IIIC (HR=0.79; 95% CI, 0.75–0.84) tumors. Similarly, when compared to radiation alone, combination therapy was accompanied by a 19% decrease in mortality (HR=0.81; 95% CI, 0.73–0.89). Combination chemoradiation was associated with decreased mortality across all of the histologic subtypes.
Conclusion
Among women with stage III uterine cancer, combination chemotherapy and external beam radiation is associated with decreased mortality compared to chemotherapy or radiation alone.
PRECIS:
Among women with stage III uterine cancer, the combination of chemotherapy and external beam radiation is associated with improved survival compared to chemotherapy or radiation alone.
INTRODUCTION
While the prognosis for endometrial cancer is favorable, it is still estimated that over 11,000 women die annually from the disease.1 More than 50% of these deaths are attributed to patients who present with disease outside of the uterus at the time of diagnosis.2 While stage III and IV disease account for only 7% and 9% of all women diagnosed with endometrial cancer, 5-year survival is estimated to range from 47–58% for stage III tumors, and 15–17% for those with stage IV neoplasm.3,4 There is a clear need to improve therapeutic options for these women.
Several factors have been shown to impact survival among patients with stage III uterine cancer, including extent of tumor invasion, lymph node involvement, age, and non-endometrioid histology. However, the optimal management of these patients remains uncertain.5 Women with stage III uterine cancer may receive treatment with either radiation or chemotherapy or multimodal therapy with both chemotherapy and radiation.6 The National Comprehensive Cancer Network (NCCN) guidelines recommend systemic therapy with or without external beam radiation treatment with or without vaginal brachytherapy for patients with stage III and IV uterine cancer.7 While patterns of care studies have suggested that these patients are commonly treated with both chemotherapy and radiation, recent prospective clinical trials have failed to demonstrate a survival advantage for combination chemoradiation.23,24
Given the uncertainty regarding the optimal adjuvant therapy for women with stage III uterine cancer, we examined the comparative effectiveness of currently available treatment strategies. Specifically, we compared survival for women with stage III uterine cancer who received chemotherapy in combination with external beam radiation to those who received chemotherapy alone (with or without vaginal brachytherapy) or external beam radiation alone.
METHODS
The National Cancer Data Base (NCDB) was used for the analysis. The NCDB is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society.8,9 It contains data on cancer patients from over 1,500 Commission on Cancer-accredited hospitals and captures more than 70% of newly diagnosed cancer cases across the United States. The study utilized de-identified data and was deemed exempt by the Institutional Review Board of Columbia University.
We identified women with pathologically confirmed stage III uterine cancer diagnosed from 2004 to 2015. We limited the cohort to women with endometrioid, serous, clear cell or endometrial not otherwise specified histologies. We excluded women who did not undergo hysterectomy, those who had neoadjuvant chemotherapy or radiation therapy prior to hysterectomy, and those who had missing data on the time of treatment initiation. We also excluded patients who had neither chemotherapy nor radiation after hysterectomy, those who had non-pelvic radiation or brachytherapy alone, or women in whom the radiation modality was unspecified.
The primary aim of the study was to determine if combination therapy with external beam radiation therapy and chemotherapy was associated with improved survival compared to single modality therapy with either chemotherapy or external beam radiotherapy. Patients who received chemotherapy and external beam radiation (alone or combined with brachytherapy) were classified as combination therapy. Women who received chemotherapy without external beam radiation were classified as the chemotherapy only cohort. These women could have received vaginal brachytherapy as part of their treatment. Finally, women who received external beam radiation with or without brachytherapy were classified as having received radiation therapy.
Patients’ demographic characteristics included age (<50, 50–59, 60–79, 70–79, ≥80 years), race and ethnicity (white, black, Hispanic, other), year of diagnosis, insurance status (private, Medicare, Medicaid, uninsured, other governmental or unknown), median household income in the patients’ zip code (<$38,000, $38,000-$47,999, $48,000-$62,999, $63,000+), and location (metropolitan, urban, rural). Tumor stage was derived from the American Joint Committee on Cancer (AJCC) pathologic stage (IIIA, IIIB, IIIC, III NOS). Patients with missing data on AJCC pathologic stage were staged using FIGO stage captured by Collaborative Stage site-specific factor. Patients with missing data on both fields were excluded. The other tumor characteristics included histology (endometrioid, serous, clear cell, endometrial not otherwise specified) and grade (well, moderate, poorly). Regional lymph nodes examination was also noted. Patients with missing data on race or ethnicity, income, urban or rural location, grade, and lymph node status were classified into an unknown group. Patients with missing data on insurance status were combined with those with other governmental insurance due to small cell size.
Hospital characteristics included facility region (northeast, Midwest, south, west) and type of hospital as defined by the American Cancer Society’s Commission on Cancer Accreditation program (academic, community cancer, comprehensive community cancer, integrated network cancer).8
Patient characteristics between those who received combination therapy and those who received chemotherapy alone or radiotherapy alone were compared using χ2 tests. The trend in use was tested using χ2 tests. To examine predictors of combination therapy, we fit a polytomous logistic regression model with combination therapy as the reference. The model included age, race, year of diagnosis, insurance status, income, location, comorbidity, facility type and region, stage, histology, grade, and nodes examined.
To account for the effect of measured covariates on treatment selection, we performed propensity score analyses comparing women who received combination therapy to chemotherapy alone and comparing combination therapy to radiation alone separately. The propensity score is the predicted probability of receipt of a given treatment, in this case, use of combination therapy. We performed propensity score analysis with inverse probability of treatment weighting (IPTW). For each pair, we fit a log-Poisson regression model including age, race, year of diagnosis, insurance status, income, location, comorbidity, facility type and region, stage, histology, grade, lymph nodes examined, and including the hospital identifier as random intercepts to account for hospital-level clustering when possible. We calculated the weights as the reciprocal of the propensity score, stabilized the weights and truncated them at 0.1 and 10 to reduce variability due to instability in estimation that could be induced by patients with very large weights.10,11,12
We compared all-cause mortality using adjusted survival curves and the log-rank test from the IPTW cohort excluding patients diagnosed in 2015 because of unavailable survival data. We estimated the hazard ratio of combination therapy using IPT-weighted Cox proportional hazard model with bootstrap variance estimation of the 95% confidence intervals (200 resamples with replacement).13 The proportional hazards assumption was tested by examining the correlation between Schoenfield residuals and function of time (the linear, the log, and the square).
To test the robustness of our findings, we examined the effect after adjusting for brachytherapy in the IPT-weighted Cox proportional hazard model. We also performed propensity score analyses estimating the propensity of brachytherapy and the IPT-weighted hazard ratios of brachytherapy among chemotherapy alone and combination patients separately. We performed a series of sensitivity analyses among the subgroups based on stage (IIIA and IIIC) and histology (endometrioid and serous), and limiting to patients who received multiagent chemotherapy. For each subgroup, similar propensity score models were fit excluding the stratifying covariate, and the hazard ratio of combination was estimated from IPT-weighted Cox-proportional hazards models adjusting for brachytherapy. All hypothesis testing was two-sided and a P-value of <0.05 was considered statistically significant. All analyses were conducted using SAS version 9.4 (SAS Institute Inc, Cary, North Carolina).
RESULTS
We identified a total of 20,873 patients including 9,456 (45.3%) who received chemotherapy alone, 2,417 (11.6%) treated with radiation alone, and 9,000 (43.1%) who received chemotherapy in combination with external beam radiation (Figure 1). The rate of combination therapy was 33.0% (95% CI, 29.2–37.0%) in 2004, and then rose year after year to 50.5% (95% CI, 48.2–52.7%) in 2015 (P<0.001, Figure 2). The mortality of the cohort was 33.1% and median survival was 115 months (95% CI, 110–123).
Figure 1.
Flowchart of cohort selection.
Figure 2.
Trend of combination therapy. Error bars indicate 95% confidence intervals. Trend was tested using Chi-square test (P-<.001). RT, radiation therapy.
Combination therapy was used in 36.1% of patients with stage IIIA tumors, 40.5% of those with stage IIIB tumors, and 45.4% of women with stage IIIC neoplasms (P<0.001) (Table 1). Combination therapy was more commonly used in women with endometrioid (44.8%) tumors than those with serous (36.5%) and clear cell (41.1%) histologies (P<0.001). In a polytomous logistic regression model, in addition to stage and histology, older women (vs. younger women), black patients (vs. white patients) and Medicaid recipients (vs. private insurance) and those with more medical comorbidities were less likely to receive combination therapy (P<0.05 for all) (Table 2). Patients treated at community cancer centers were more likely to receive combination therapy (OR comparing chemotherapy to combination=0.59; 95% CI 0.51–0.68).
Table 1.
Demographic and clinical characteristics of patients who had stage III uterine cancer, comparing combination therapy, chemotherapy and pelvic radiation.
| Combination | Chemotherapy | Pelvic radiation | |||||
|---|---|---|---|---|---|---|---|
| N | (%) | N | (%) | N | (%) | P-value | |
| All | 9,000 | (43.1) | 9,456 | (45.3) | 2,417 | (11.6) | |
| Age (in years) | <0.001 | ||||||
| <50 | 1,083 | (49.6) | 925 | (42.4) | 176 | (8.1) | |
| 50–59 | 2,832 | (47.2) | 2,603 | (43.4) | 559 | (9.3) | |
| 60–69 | 3,412 | (44.4) | 3,564 | (46.4) | 702 | (9.1) | |
| 70–79 | 1,453 | (37.9) | 1,812 | (47.2) | 573 | (14.9) | |
| ≥80 | 220 | (18.7) | 552 | (46.8) | 407 | (34.5) | |
| Race/ethnicity | <0.001 | ||||||
| White | 7,097 | (43.9) | 7,119 | (44.0) | 1,965 | (12.1) | |
| Black | 954 | (38.6) | 1,273 | (51.5) | 244 | (9.9) | |
| Hispanic | 480 | (41.4) | 564 | (48.6) | 116 | (10.0) | |
| Other | 401 | (45.6) | 406 | (46.1) | 73 | (8.3) | |
| Unknown | 68 | (37.6) | 94 | (51.9) | 19 | (10.5) | |
| Year of diagnosis | <0.001 | ||||||
| 2004 | 271 | (33.0) | 229 | (27.9) | 322 | (39.2) | |
| 2005 | 347 | (37.4) | 308 | (33.2) | 274 | (29.5) | |
| 2006 | 400 | (38.8) | 392 | (38.1) | 238 | (23.1) | |
| 2007 | 427 | (38.3) | 476 | (42.7) | 211 | (18.9) | |
| 2008 | 455 | (38.2) | 559 | (47.0) | 176 | (14.8) | |
| 2009 | 519 | (43.2) | 556 | (46.3) | 127 | (10.6) | |
| 2010 | 807 | (40.7) | 982 | (49.5) | 194 | (9.8) | |
| 2011 | 893 | (41.2) | 1,085 | (50.0) | 191 | (8.8) | |
| 2012 | 992 | (42.9) | 1,153 | (49.8) | 168 | (7.3) | |
| 2013 | 1,141 | (44.3) | 1,276 | (49.5) | 160 | (6.2) | |
| 2014 | 1,325 | (48.6) | 1,247 | (45.8) | 153 | (5.6) | |
| 2015 | 1,423 | (50.5) | 1,193 | (42.3) | 203 | (7.2) | |
| Insurance status | <0.001 | ||||||
| Private | 4,822 | (47.9) | 4,352 | (43.2) | 896 | (8.9) | |
| Medicare | 3,038 | (37.9) | 3,735 | (46.6) | 1,243 | (15.5) | |
| Medicaid | 596 | (43.1) | 662 | (47.8) | 126 | (9.1) | |
| Uninsured | 344 | (37.8) | 474 | (52.0) | 93 | (10.2) | |
| Other government/unknown | 200 | (40.7) | 233 | (47.4) | 59 | (12.0) | |
| Income | <0.001 | ||||||
| <$38,000 | 1,371 | (39.1) | 1,700 | (48.5) | 435 | (12.4) | |
| $38,000-$47,999 | 2,126 | (42.9) | 2,206 | (44.5) | 625 | (12.6) | |
| $48,000-$62,999 | 2,533 | (44.2) | 2,566 | (44.7) | 637 | (11.1) | |
| $63,000+ | 2,901 | (44.7) | 2,899 | (44.7) | 690 | (10.6) | |
| Unknown | 69 | (37.5) | 85 | (46.2) | 30 | (16.3) | |
| Location | 0.04 | ||||||
| Metropolitan | 7,332 | (43.5) | 7,583 | (45.0) | 1,931 | (11.5) | |
| Urban | 1,262 | (42.1) | 1,361 | (45.4) | 373 | (12.4) | |
| Rural | 145 | (39.6) | 176 | (48.1) | 45 | (12.3) | |
| Unknown | 261 | (39.2) | 336 | (50.5) | 68 | (10.2) | |
| Comorbidity | <0.001 | ||||||
| 0 | 6,993 | (44.0) | 7,070 | (44.5) | 1,827 | (11.5) | |
| 1 | 1,681 | (41.1) | 1,951 | (47.7) | 462 | (11.3) | |
| ≥2 | 326 | (36.7) | 435 | (48.9) | 128 | (14.4) | |
| Facility type | <0.001 | ||||||
| Academic | 3,733 | (40.8) | 4,500 | (49.2) | 908 | (9.9) | |
| Community cancer | 552 | (49.1) | 386 | (34.3) | 186 | (16.5) | |
| Comprehensive community cancer | 3,590 | (44.5) | 3,414 | (42.3) | 1,060 | (13.1) | |
| Integrated network cancer | 1,125 | (44.2) | 1,156 | (45.4) | 263 | (10.3) | |
| Facility location | <0.001 | ||||||
| Northeast | 1,938 | (42.9) | 2,103 | (46.5) | 480 | (10.6) | |
| Midwest | 2,740 | (44.9) | 2,634 | (43.1) | 734 | (12.0) | |
| South | 2,867 | (40.7) | 3,330 | (47.3) | 847 | (12.0) | |
| West | 1,455 | (45.5) | 1,389 | (43.4) | 356 | (11.1) | |
| Stage | <0.001 | ||||||
| IIIA | 1,327 | (36.1) | 2,025 | (55.1) | 325 | (8.8) | |
| IIIB | 515 | (40.5) | 495 | (38.9) | 262 | (20.6) | |
| IIIC | 6,971 | (45.4) | 6,649 | (43.3) | 1,726 | (11.2) | |
| III NOS | 187 | (32.4) | 287 | (49.7) | 104 | (18.0) | |
| Histology | <0.001 | ||||||
| Endometrioid | 5,625 | (44.8) | 5,227 | (41.6) | 1,715 | (13.6) | |
| Serous | 1,224 | (36.5) | 1,979 | (59.1) | 146 | (4.4) | |
| Clear cell | 263 | (41.1) | 311 | (48.6) | 66 | (10.3) | |
| EM NOS | 1,888 | (43.7) | 1,939 | (44.9) | 490 | (11.4) | |
| Grade | <0.001 | ||||||
| Well | 1,278 | (44.8) | 1,119 | (39.2) | 458 | (16.0) | |
| Moderate | 2,515 | (44.2) | 2,326 | (40.9) | 852 | (15.0) | |
| Poorly | 3,726 | (41.4) | 4,455 | (49.5) | 814 | (9.0) | |
| Unknown | 1,481 | (44.5) | 1,556 | (46.7) | 293 | (8.8) | |
| Regional lymph nodes examined* | 8,502 | (94.5) | 8,638 | (91.3) | 2,196 | (90.9) | <0.001 |
| Brachytherapy | 3,312 | (36.8) | 2,283 | (24.1) | 833 | (34.5) | <0.001 |
Regional lymph nodes examined was analyzed as yes, no, unknown. Only patients that had nodes examined were reported because of small cells (<10) in the unknown group.
Table 2.
Polytomous logistic regression model for predictors of treatment.
| Chemotherapy vs. Combination | Pelvic radiation vs. Combination | |
|---|---|---|
| aOR | aOR | |
| Age | ||
| <50 | Referent | Referent |
| 50–59 | 1.12 (1.01–1.24)* | 1.49 (1.23–1.81)* |
| 60–69 | 1.20 (1.08–1.34)* | 1.74 (1.43–2.12)* |
| 70–79 | 1.37 (1.20–1.57)* | 3.05 (2.43–3.81)* |
| ≥80 | 2.64 (2.17–3.21)* | 15.37 (11.75–20.11)* |
| Race/ethnicity | ||
| White | Referent | Referent |
| Black | 1.11 (1.001–1.22)* | 1.23 (1.04–1.47)* |
| Hispanic | 1.11 (0.97–1.27) | 1.05 (0.84–1.33) |
| Other | 1.06 (0.91–1.23) | 0.88 (0.66–1.16) |
| Unknown | 1.34 (0.97–1.85) | 0.90 (0.51–1.58) |
| Year of diagnosis | ||
| 2004 | Referent | Referent |
| 2005 | 1.07 (0.84–1.35) | 0.67 (0.53–0.85)* |
| 2006 | 1.17 (0.93–1.47) | 0.50 (0.39–0.63)* |
| 2007 | 1.30 (1.04–1.62)* | 0.39 (0.31–0.50)* |
| 2008 | 1.48 (1.19–1.83)* | 0.29 (0.23–0.37)* |
| 2009 | 1.27 (1.02–1.57)* | 0.18 (0.14–0.23)* |
| 2010 | 1.26 (1.03–1.55)* | 0.14 (0.11–0.18)* |
| 2011 | 1.27 (1.04–1.56)* | 0.12 (0.10–0.16)* |
| 2012 | 1.19 (0.98–1.46) | 0.10 (0.08–0.13)* |
| 2013 | 1.15 (0.94–1.40) | 0.08 (0.06–0.10)* |
| 2014 | 0.97 (0.80–1.18) | 0.07 (0.05–0.09)* |
| 2015 | 0.87 (0.72–1.07) | 0.08 (0.06–0.10)* |
| Insurance status | ||
| Private | Referent | Referent |
| Medicare | 1.09 (1.01–1.18)* | 1.35 (1.18–1.54)* |
| Medicaid | 1.18 (1.04–1.33)* | 1.27 (1.02–1.58)* |
| Uninsured | 1.42 (1.23–1.65)* | 1.55 (1.20–2.00)* |
| Other government/unknown | 1.15 (0.94–1.40) | 1.30 (0.94–1.79) |
| Income | ||
| <$38,000 | Referent | Referent |
| $38,000-$47,999 | 0.92 (0.83–1.01) | 0.87 (0.75–1.02) |
| $48,000-$62,999 | 0.92 (0.83–1.01) | 0.79 (0.67–0.92)* |
| $63,000+ | 0.92 (0.83–1.01) | 0.79 (0.67–0.92)* |
| Unknown | 0.91 (0.63–1.31) | 0.86 (0.49–1.52) |
| Location | ||
| Metropolitan | Referent | Referent |
| Urban | 1.09 (0.997–1.20) | 1.04 (0.90–1.20) |
| Rural | 1.24 (0.98–1.56) | 0.98 (0.68–1.42) |
| Unknown | 1.25 (1.04–1.51)* | 0.76 (0.53–1.08) |
| Comorbidity | ||
| 0 | Referent | Referent |
| 1 | 1.11 (1.03–1.20)* | 1.06 (0.94–1.21) |
| ≥2 | 1.23 (1.06–1.43)* | 1.45 (1.16–1.83)* |
| Facility type | ||
| Academic | Referent | Referent |
| Community cancer | 0.59 (0.51–0.68)* | 1.16 (0.94–1.41) |
| Comprehensive community cancer | 0.79 (0.74–0.84)* | 1.11 (0.99–1.24) |
| Integrated network cancer | 0.84 (0.76–0.92)* | 0.89 (0.75–1.05) |
| Facility region | ||
| Northeast | Referent | Referent |
| Midwest | 0.91 (0.84–0.996)* | 1.06 (0.91–1.22) |
| South | 1.07 (0.99–1.17) | 1.27 (1.10–1.47)* |
| West | 0.97 (0.87–1.07) | 1.20 (1.01–1.42)* |
| Stage | ||
| IIIA | Referent | Referent |
| IIIB | 0.57 (0.50–0.66)* | 1.44 (1.17–1.77)* |
| IIIC | 0.60 (0.55–0.65)* | 0.54 (0.46–0.63)* |
| III NOS | 0.90 (0.73–1.10) | 1.08 (0.80–1.46) |
| Histology | ||
| Endometrioid | Referent | Referent |
| Serous | 1.51 (1.38–1.66)* | 0.32 (0.26–0.39)* |
| Clear Cell | 1.14 (0.96–1.36) | 0.71 (0.52–0.96)* |
| EM NOS | 1.05 (0.97–1.13) | 0.79 (0.70–0.90)* |
| Grade | ||
| Well | Referent | Referent |
| Moderate | 1.05 (0.95–1.16) | 0.78 (0.67–0.90)* |
| Poorly | 1.13 (1.03–1.25)* | 0.49 (0.42–0.57)* |
| Unknown | 1.06 (0.95–1.19) | 0.67 (0.56–0.80)* |
| Regional lymph nodes examined | ||
| No | Referent | Referent |
| Yes | 0.83 (0.73–0.95)* | 0.80 (0.66–0.98)* |
| Unknown | 2.47 (1.15–5.30)* | 0.75 (0.20–2.72) |
Polytomous logistic regression model was fitted for predictors of treatment using combination as the reference. The model included age, race, year, insurance status, income, location, comorbidity, facility type and region, stage, histology, grade, and regional lymph nodes examined.
P-value<0.05
After inverse probability of treatment weighting, the clinical and pathologic characteristics of the chemotherapy and combination therapy cohorts were well balanced (Table 3). Within this cohort, the median follow-up time was 39 months (IQR, 23–63). Women who received chemotherapy alone were significantly more likely to die than those women who received combination therapy (P<0.001) (Figure 3A). In the IPT-weighted cohort, women who received combination therapy vs. chemotherapy alone had a 23% reduction in mortality (HR=0.77; 95% CI, 0.73–0.80). Five year survival was 68.5% (95% CI, 67.2–69.7%) in women who received combination therapy compared to 62.0% (95% CI, 60.7–63.3%) in those who received chemotherapy alone (P<0.001) (Figure 3A). Brachytherapy further reduced the risk of death in the combination therapy and chemotherapy cohorts. When the analysis was limited to just the cohort of women who received combination therapy, brachytherapy was associated with a decreased risk of death (HR=0.89; 95% CI, 0.82–0.97). Similarly, among women who only received chemotherapy without external beam therapy, brachytherapy decreased the risk of death (HR=0.67; 9% CI, 0.61–0.75).
Table 3.
Propensity score weighted demographic and tumor characteristics, comparing combination therapy to chemotherapy.
| Crude | Propensity Score Weighted | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Chemotherapy | Combination | Chemotherapy | Combination | |||||||
| N | (%) | N | (%) | P-value | N | (%) | N | (%) | P-value | |
| All | 9,456 | (51.2) | 9,000 | (48.8) | 9,886 | (51.2) | 9,418 | (48.8) | ||
| Age (in years) | <0.001 | 1.00 | ||||||||
| <50 | 925 | (9.8) | 1,083 | (12.0) | 1,088 | (11.0) | 1,035 | (11.0) | ||
| 50–59 | 2,603 | (27.5) | 2,832 | (31.5) | 2,913 | (29.5) | 2,782 | (29.5) | ||
| 60–69 | 3,564 | (37.7) | 3,412 | (37.9) | 3,732 | (37.7) | 3,565 | (37.8) | ||
| 70–79 | 1,812 | (19.2) | 1,453 | (16.1) | 1,737 | (17.6) | 1,649 | (17.5) | ||
| ≥80 | 552 | (5.8) | 220 | (2.4) | 416 | (4.2) | 387 | (4.1) | ||
| Race/ethnicity | <0.001 | 0.83 | ||||||||
| White | 7,119 | (75.3) | 7,097 | (78.9) | 7,595 | (76.8) | 7,242 | (76.9) | ||
| Black | 1,273 | (13.5) | 954 | (10.6) | 1,177 | (11.9) | 1,153 | (12.2) | ||
| Hispanic | 564 | (6.0) | 480 | (5.3) | 568 | (5.7) | 533 | (5.7) | ||
| Other | 406 | (4.3) | 401 | (4.5) | 460 | (4.7) | 411 | (4.4) | ||
| Unknown | 94 | (1.0) | 68 | (0.8) | 86 | (0.9) | 80 | (0.8) | ||
| Year of diagnosis | <0.001 | 1.00 | ||||||||
| 2004 | 229 | (2.4) | 271 | (3.0) | 270 | (2.7) | 249 | (2.6) | ||
| 2005 | 308 | (3.3) | 347 | (3.9) | 350 | (3.5) | 331 | (3.5) | ||
| 2006 | 392 | (4.1) | 400 | (4.4) | 419 | (4.2) | 396 | (4.2) | ||
| 2007 | 476 | (5.0) | 427 | (4.7) | 480 | (4.9) | 454 | (4.8) | ||
| 2008 | 559 | (5.9) | 455 | (5.1) | 548 | (5.5) | 511 | (5.4) | ||
| 2009 | 556 | (5.9) | 519 | (5.8) | 583 | (5.9) | 544 | (5.8) | ||
| 2010 | 982 | (10.4) | 807 | (9.0) | 968 | (9.8) | 909 | (9.6) | ||
| 2011 | 1,085 | (11.5) | 893 | (9.9) | 1,065 | (10.8) | 1,020 | (10.8) | ||
| 2012 | 1,153 | (12.2) | 992 | (11.0) | 1,140 | (11.5) | 1,101 | (11.7) | ||
| 2013 | 1,276 | (13.5) | 1,141 | (12.7) | 1,309 | (13.2) | 1,235 | (13.1) | ||
| 2014 | 1,247 | (13.2) | 1,325 | (14.7) | 1,374 | (13.9) | 1,319 | (14.0) | ||
| 2015 | 1,193 | (12.6) | 1,423 | (15.8) | 1,382 | (14.0) | 1,350 | (14.3) | ||
| Insurance status | <0.001 | 0.99 | ||||||||
| Private | 4,352 | (46.0) | 4,822 | (53.6) | 4,917 | (49.7) | 4,711 | (50.0) | ||
| Medicare | 3,735 | (39.5) | 3,038 | (33.8) | 3,619 | (36.6) | 3,439 | (36.5) | ||
| Medicaid | 662 | (7.0) | 596 | (6.6) | 682 | (6.9) | 646 | (6.9) | ||
| Uninsured | 474 | (5.0) | 344 | (3.8) | 438 | (4.4) | 409 | (4.3) | ||
| Other government/unknown | 233 | (2.5) | 200 | (2.2) | 229 | (2.3) | 213 | (2.3) | ||
| Income | <0.001 | 0.94 | ||||||||
| <$38,000 | 1,700 | (18.0) | 1,371 | (15.2) | 1,637 | (16.6) | 1,574 | (16.7) | ||
| $38,000-$47,999 | 2,206 | (23.3) | 2,126 | (23.6) | 2,290 | (23.2) | 2,223 | (23.6) | ||
| $48,000-$62,999 | 2,566 | (27.1) | 2,533 | (28.1) | 2,736 | (27.7) | 2,581 | (27.4) | ||
| $63,000+ | 2,899 | (30.7) | 2,901 | (32.2) | 3,143 | (31.8) | 2,960 | (31.4) | ||
| Unknown | 85 | (0.9) | 69 | (0.8) | 81 | (0.8) | 79 | (0.8) | ||
| Location | 0.03 | 0.47 | ||||||||
| Metropolitan | 7,583 | (80.2) | 7,332 | (81.5) | 7,971 | (80.6) | 7,656 | (81.3) | ||
| Urban | 1,361 | (14.4) | 1,262 | (14.0) | 1,406 | (14.2) | 1,320 | (14.0) | ||
| Rural | 176 | (1.9) | 145 | (1.6) | 176 | (1.8) | 159 | (1.7) | ||
| Unknown | 336 | (3.6) | 261 | (2.9) | 333 | (3.4) | 284 | (3.0) | ||
| Comorbidity | <0.001 | 0.81 | ||||||||
| 0 | 7,070 | (74.8) | 6,993 | (77.7) | 7,538 | (76.2) | 7,165 | (76.1) | ||
| 1 | 1,951 | (20.6) | 1,681 | (18.7) | 1,936 | (19.6) | 1,873 | (19.9) | ||
| ≥2 | 435 | (4.6) | 326 | (3.6) | 412 | (4.2) | 381 | (4.0) | ||
| Facility type | <0.001 | 0.07 | ||||||||
| Academic | 4,500 | (47.6) | 3,733 | (41.5) | 4,321 | (43.7) | 4,240 | (45.0) | ||
| Community cancer | 386 | (4.1) | 552 | (6.1) | 545 | (5.5) | 472 | (5.0) | ||
| Comprehensive community cancer | 3,414 | (36.1) | 3,590 | (39.9) | 3,822 | (38.7) | 3,522 | (37.4) | ||
| Integrated network cancer | 1,156 | (12.2) | 1,125 | (12.5) | 1,197 | (12.1) | 1,184 | (12.6) | ||
| Facility location | <0.001 | 0.65 | ||||||||
| Northeast | 2,103 | (22.2) | 1,938 | (21.5) | 2,146 | (21.7) | 2,052 | (21.8) | ||
| Midwest | 2,634 | (27.9) | 2,740 | (30.4) | 2,906 | (29.4) | 2,715 | (28.8) | ||
| South | 3,330 | (35.2) | 2,867 | (31.9) | 3,276 | (33.1) | 3,193 | (33.9) | ||
| West | 1,389 | (14.7) | 1,455 | (16.2) | 1,558 | (15.8) | 1,458 | (15.5) | ||
| Stage | <0.001 | 0.97 | ||||||||
| IIIA | 2,025 | (21.4) | 1,327 | (14.7) | 1,798 | (18.2) | 1,695 | (18.0) | ||
| IIIB | 495 | (5.2) | 515 | (5.7) | 542 | (5.5) | 521 | (5.5) | ||
| IIIC | 6,649 | (70.3) | 6,971 | (77.5) | 7,292 | (73.8) | 6,967 | (74.0) | ||
| III NOS | 287 | (3.0) | 187 | (2.1) | 255 | (2.6) | 235 | (2.5) | ||
| Histology | <0.001 | 0.94 | ||||||||
| Endometrioid | 5,227 | (55.3) | 5,625 | (62.5) | 5,848 | (59.2) | 5,553 | (59.0) | ||
| Serous | 1,979 | (20.9) | 1,224 | (13.6) | 1,710 | (17.3) | 1,613 | (17.1) | ||
| Clear cell | 311 | (3.3) | 263 | (2.9) | 301 | (3.0) | 295 | (3.1) | ||
| EM NOS | 1,939 | (20.5) | 1,888 | (21.0) | 2,027 | (20.5) | 1,957 | (20.8) | ||
| Grade | <0.001 | 0.79 | ||||||||
| Well | 1,119 | (11.8) | 1,278 | (14.2) | 1,286 | (13.0) | 1,233 | (13.1) | ||
| Moderate | 2,326 | (24.6) | 2,515 | (27.9) | 2,631 | (26.6) | 2,456 | (26.1) | ||
| Poorly | 4,455 | (47.1) | 3,726 | (41.4) | 4,377 | (44.3) | 4,173 | (44.3) | ||
| Unknown | 1,556 | (16.5) | 1,481 | (16.5) | 1,592 | (16.1) | 1,556 | (16.5) | ||
| Regional lymph nodes examined | <0.001 | 0.90 | ||||||||
| No | 787 | (8.3) | 489 | (5.4) | 683 | (6.9) | 646 | (6.9) | ||
| Yes | 8,638 | (91.3) | 8,502 | (94.5) | 9,182 | (92.9) | 8,749 | (92.9) | ||
| Unknown | 31 | (0.3) | * | * | 22 | (0.2) | 24 | (0.3) | ||
Cells <10.
Mixed-effect log-Poisson model with robust error variance was fit for propensity of combination therapy, including age, race, year, insurance status, income, location, comorbidity, facility type and region, stage, histology, grade, and regional lymph nodes examined. Hospital identifiers were included as random intercepts to accounted for data clustering. The inverse probability of treatment weights (IPTW) were calculated the reciprocal of the propensity score. The IPTW was stabilized to reduce variability due to instability in estimation that could be induced by patients with extreme weights.
Because of weighting of the propensity score analysis, the number of subjects is greater than in the original cohort.
Figure 3.
Propensity score weighted survival curves. Shades indicate 95% confidence intervals. Chemotherapy compared with combination (P<.001, adjusted log-rank test) (A); Pelvic radiation compared with combination (P<.001, adjusted log-rank test) (B).
These findings were robust in a number of sensitivity analyses. When limited to women with stage IIIA tumors, mortality was reduced by 19% in those who received combination therapy compared to chemotherapy after further adjusting for brachytherapy (HR=0.81; 95% CI, 0.68–0.98). In a similar analysis of only women with stage IIIC tumors, mortality was 21% lower with combination therapy (HR=0.79; 95% CI, 0.75–0.84). Finally, within the cohort of women treated with combination therapy, 6.6% received single agent therapy and 90.0% received multiagent therapy. When this cohort was restricted to only those women who had received multiagent chemotherapy as part of treatment, mortality was lower in those who received combination therapy compared to chemotherapy alone (HR=0.78; 95% CI, 0.74–0.82) (Supplemental Figure 1).
Similarly, after inverse probability of treatment weighting, the radiation therapy and combination therapy cohorts were more balanced, however, there remained imbalances in the age, histology, and grade distributions (Table 4). Within this cohort, the median follow-up time was 43 months (IQR, 25–70). Women who received radiation alone were significantly more likely to die than those women who received combination therapy (P<0.001) (Figure 3B). The IPT-weighted hazard ratio of death for women who received combination therapy vs. radiation alone was 0.81 (95% CI, 0.73–0.89). Five-year survival was 69.4% (95% CI, 68.1–70.7%) in women who received combination therapy compared to 62.9% (95% CI, 59.8–65.8%) in those who received radiation alone (P<0.001). These findings were robust in a series of sensitivity analyses stratified by stage and histology (Appendix 1, available online at http://links.lww.com/xxx).
Table 4.
Propensity score weighted demographic and tumor characteristics, comparing combination therapy to pelvic radiation.
| Crude | Propensity Score Weighted | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Pelvic radiation | Combination | Pelvic radiation | Combination | |||||||
| N | (%) | N | (%) | P-value | N | (%) | N | (%) | P-value | |
| All | 2,417 | (21.2) | 9,000 | (78.8) | 3,869 | (29.1) | 9,418 | (70.9) | ||
| Age (in years) | <0.001 | 0.005 | ||||||||
| <50 | 176 | (7.3) | 1,083 | (12.0) | 445 | (11.5) | 1,060 | (11.3) | ||
| 50–59 | 559 | (23.1) | 2,832 | (31.5) | 1,241 | (32.1) | 2,820 | (29.9) | ||
| 60–69 | 702 | (29.0) | 3,412 | (37.9) | 1,280 | (33.1) | 3,414 | (36.3) | ||
| 70–79 | 573 | (23.7) | 1,453 | (16.1) | 676 | (17.5) | 1,644 | (17.5) | ||
| ≥80 | 407 | (16.8) | 220 | (2.4) | 226 | (5.9) | 480 | (5.1) | ||
| Race | 0.01 | 0.35 | ||||||||
| White | 1,965 | (81.3) | 7,097 | (78.9) | 3,089 | (79.8) | 7,466 | (79.3) | ||
| Black | 244 | (10.1) | 954 | (10.6) | 410 | (10.6) | 994 | (10.6) | ||
| Hispanic | 116 | (4.8) | 480 | (5.3) | 207 | (5.4) | 489 | (5.2) | ||
| Other | 73 | (3.0) | 401 | (4.5) | 134 | (3.5) | 400 | (4.2) | ||
| Unknown | 19 | (0.8) | 68 | (0.8) | 29 | (0.7) | 70 | (0.7) | ||
| Year of diagnosis | <0.001 | 0.07 | ||||||||
| 2004 | 322 | (13.3) | 271 | (3.0) | 213 | (5.5) | 466 | (4.9) | ||
| 2005 | 274 | (11.3) | 347 | (3.9) | 218 | (5.6) | 490 | (5.2) | ||
| 2006 | 238 | (9.8) | 400 | (4.4) | 232 | (6.0) | 521 | (5.5) | ||
| 2007 | 211 | (8.7) | 427 | (4.7) | 225 | (5.8) | 516 | (5.5) | ||
| 2008 | 176 | (7.3) | 455 | (5.1) | 230 | (5.9) | 518 | (5.5) | ||
| 2009 | 127 | (5.3) | 519 | (5.8) | 239 | (6.2) | 528 | (5.6) | ||
| 2010 | 194 | (8.0) | 807 | (9.0) | 377 | (9.7) | 826 | (8.8) | ||
| 2011 | 191 | (7.9) | 893 | (9.9) | 370 | (9.6) | 897 | (9.5) | ||
| 2012 | 168 | (7.0) | 992 | (11.0) | 368 | (9.5) | 964 | (10.2) | ||
| 2013 | 160 | (6.6) | 1,141 | (12.7) | 409 | (10.6) | 1,091 | (11.6) | ||
| 2014 | 153 | (6.3) | 1,325 | (14.7) | 461 | (11.9) | 1,244 | (13.2) | ||
| 2015 | 203 | (8.4) | 1,423 | (15.8) | 528 | (13.6) | 1,356 | (14.4) | ||
| Insurance status | <0.001 | 0.75 | ||||||||
| Private | 896 | (37.1) | 4,822 | (53.6) | 1,920 | (49.6) | 4,774 | (50.7) | ||
| Medicare | 1,243 | (51.4) | 3,038 | (33.8) | 1,443 | (37.3) | 3,475 | (36.9) | ||
| Medicaid | 126 | (5.2) | 596 | (6.6) | 265 | (6.9) | 599 | (6.4) | ||
| Uninsured | 93 | (3.8) | 344 | (3.8) | 153 | (4.0) | 359 | (3.8) | ||
| Other government/unknown | 59 | (2.4) | 200 | (2.2) | 87 | (2.3) | 210 | (2.2) | ||
| Income | <0.001 | 0.39 | ||||||||
| <$38,000 | 435 | (18.0) | 1,371 | (15.2) | 602 | (15.6) | 1,477 | (15.7) | ||
| $38,000-$47,999 | 625 | (25.9) | 2,126 | (23.6) | 970 | (25.1) | 2,260 | (24.0) | ||
| $48,000-$62,999 | 637 | (26.4) | 2,533 | (28.1) | 1,104 | (28.5) | 2,621 | (27.8) | ||
| $63,000+ | 690 | (28.5) | 2,901 | (32.2) | 1,161 | (30.0) | 2,979 | (31.6) | ||
| Unknown | 30 | (1.2) | 69 | (0.8) | 33 | (0.8) | 80 | (0.8) | ||
| Location | 0.26 | 0.90 | ||||||||
| Metropolitan | 1,931 | (79.9) | 7,332 | (81.5) | 3,159 | (81.7) | 7,662 | (81.4) | ||
| Urban | 373 | (15.4) | 1,262 | (14.0) | 528 | (13.7) | 1,328 | (14.1) | ||
| Rural | 45 | (1.9) | 145 | (1.6) | 68 | (1.8) | 156 | (1.7) | ||
| Unknown | 68 | (2.8) | 261 | (2.9) | 113 | (2.9) | 272 | (2.9) | ||
| Comorbidity | <0.001 | 0.92 | ||||||||
| 0 | 1,827 | (75.6) | 6,993 | (77.7) | 2,995 | (77.4) | 7,282 | (77.3) | ||
| 1 | 462 | (19.1) | 1,681 | (18.7) | 719 | (18.6) | 1,771 | (18.8) | ||
| ≥ 2 | 128 | (5.3) | 326 | (3.6) | 154 | (4.0) | 364 | (3.9) | ||
| Facility type | <0.001 | 0.80 | ||||||||
| Academic | 908 | (37.6) | 3,733 | (41.5) | 1,593 | (41.2) | 3,850 | (40.9) | ||
| Community cancer | 186 | (7.7) | 552 | (6.1) | 262 | (6.8) | 601 | (6.4) | ||
| Comprehensive community cancer | 1,060 | (43.9) | 3,590 | (39.9) | 1,553 | (40.1) | 3,820 | (40.6) | ||
| Integrated network cancer | 263 | (10.9) | 1,125 | (12.5) | 461 | (11.9) | 1,146 | (12.2) | ||
| Facility location | 0.01 | 0.90 | ||||||||
| Northeast | 480 | (19.9) | 1,938 | (21.5) | 830 | (21.4) | 2,003 | (21.3) | ||
| Midwest | 734 | (30.4) | 2,740 | (30.4) | 1,195 | (30.9) | 2,859 | (30.4) | ||
| South | 847 | (35.0) | 2,867 | (31.9) | 1,239 | (32.0) | 3,060 | (32.5) | ||
| West | 356 | (14.7) | 1,455 | (16.2) | 605 | (15.6) | 1,495 | (15.9) | ||
| Stage | <0.001 | 0.40 | ||||||||
| IIIA | 325 | (13.4) | 1,327 | (14.7) | 572 | (14.8) | 1,350 | (14.3) | ||
| IIIB | 262 | (10.8) | 515 | (5.7) | 286 | (7.4) | 639 | (6.8) | ||
| IIIC | 1,726 | (71.4) | 6,971 | (77.5) | 2,905 | (75.1) | 7,192 | (76.4) | ||
| III NOS | 104 | (4.3) | 187 | (2.1) | 106 | (2.7) | 238 | (2.5) | ||
| Histology | <0.001 | <0.001 | ||||||||
| Endometrioid | 1,715 | (71.0) | 5,625 | (62.5) | 2,615 | (67.6) | 5,989 | (63.6) | ||
| Serous | 146 | (6.0) | 1,224 | (13.6) | 362 | (9.4) | 1,192 | (12.7) | ||
| Clear cell | 66 | (2.7) | 263 | (2.9) | 91 | (2.3) | 277 | (2.9) | ||
| EM NOS | 490 | (20.3) | 1,888 | (21.0) | 801 | (20.7) | 1,959 | (20.8) | ||
| Grade | <0.001 | <0.001 | ||||||||
| Well | 458 | (18.9) | 1,278 | (14.2) | 606 | (15.7) | 1,402 | (14.9) | ||
| Moderate | 852 | (35.3) | 2,515 | (27.9) | 1,214 | (31.4) | 2,746 | (29.2) | ||
| Poorly | 814 | (33.7) | 3,726 | (41.4) | 1,346 | (34.8) | 3,790 | (40.2) | ||
| Unknown | 293 | (12.1) | 1,481 | (16.5) | 702 | (18.2) | 1,480 | (15.7) | ||
| Regional lymph nodes examined | <0.001 | 0.87 | ||||||||
| No | 217 | (9.0) | 489 | (5.4) | 238 | (6.1) | 563 | (6.0) | ||
| Yes | 2,196 | (90.9) | 8,502 | (94.5) | 3,627 | (93.8) | 8,844 | (93.9) | ||
| Unknown | * | * | * | * | * | * | 11 | (0.1) | ||
Cells <10.
Log-Poisson model was fit for propensity of combination therapy, including age, race, year, insurance status, income, location, comorbidity, facility type and region, stage, histology, grade, and lymph nodes examined. Hospital-level clustering was not accounted for due to convergence and overfitting concerns. The inverse probability of treatment weights (IPTW) were calculated the reciprocal of the propensity score. The IPTW was stabilized and truncated at 0.1 and 10 to reduce variability due to instability in estimation that could be induced by patients with extreme weights.
Because of weighting of the propensity score analysis, the number of subjects is greater than in the original cohort.
DISCUSSION
These findings suggest that among women with stage III uterine cancer who underwent hysterectomy, survival is improved with the combination of chemotherapy and external beam radiation compared to either chemotherapy or radiation alone. Combination therapy resulted in a clinically significant 6.5% improvement in 5-year survival compared to either radiation or chemotherapy alone. The survival advantage conferred by multimodal therapy was noted across all histologic subtypes and for all of the individual substages. Over the years of study, approximately 42% of the women with stage III tumors received combination therapy.
The optimal treatment for patients with metastatic uterine cancer remains an area of uncertainty.14,15 In 2006, Gynecologic Oncology Group (GOG) protocol 122 established that chemotherapy was superior to whole abdominal radiation for women with stage III-IV endometrial cancer.16 This protocol led to the increased uptake of chemotherapy for women with advanced stage endometrial cancer. While chemotherapy use became more common, many women were treated with chemotherapy in combination with radiation, with several studies suggesting a possible trend towards improved survival among these women.17,18,19, 20 In our cohort, use of chemotherapy, with or without brachytherapy, was slightly more common than use of combination therapy.
Recently, cooperative group trials have tested the benefit of combination therapy compared to single modality treatment in women with high-risk and metastatic endometrial cancer.21 The PORTEC 3 study randomized women with high-risk stage I-III endometrial cancer to external beam radiation alone or to chemotherapy with external beam radiation.22 Combination therapy was associated with improved progression free survival, but there was no difference in overall survival. On the other hand, GOG 258 compared chemotherapy alone to chemotherapy with external beam radiation in those with stage III-IVA endometrial cancer. Though the data is not yet mature, combination therapy was more toxic and there was no difference in recurrence free or overall survival.23,24 Differing study designs have led to continued uncertainty regarding the optimal therapy for women with advanced stage uterine cancer.
We found that combination therapy was associated with a significant improvement in overall survival compared to either chemotherapy alone or radiation therapy alone. In our cohort, mortality was reduced in women who received multimodal therapy. A prior multi-institutional observational study of women with stage IIIC endometrial cancer noted that women treated with chemotherapy alone had a four-fold increase in the risk of death compared to women treated with chemotherapy in combination with radiation. In this report there was no difference in survival between combination therapy and radiation alone.20
There are a number of possible explanations for why our findings differ from those of GOG 258 and PORTEC 3. First, unlike GOG 258, our study was limited to women with stage III tumors. The benefits of radiation are likely greatest in women with disease localized to the pelvis or regional nodes as opposed to women with more diffuse metastatic disease. Second, unlike the highly selected patients in clinical trials, our cohort includes a population-based sample of women with stage III uterine cancer. These patients may be at greater risk, and thus derive more benefit from combination therapy than those patients treated in a clinical trial.
While our study benefits from the inclusion of a large number of patients, we recognize several important limitations. First, while we rigorously adjusted for measurable confounders, there are likely a number of unmeasured factors that influenced both treatment and outcome. We limited our analysis to patients with stage III tumors to exclude women with metastatic disease and to limit the potential confounding of the effect of extent of disease. Second, NCDB lacks data on the specific chemotherapeutic agents used and the number of therapeutic cycles delivered. We performed a sensitivity analysis in which the cohort was limited to those women who received multiagent therapy to exclude patients who may have received only low dose, radiosensitizing chemotherapy. Third, some subgroup analyses may have been underpowered to detect differences in outcomes for the groups. Fourth, over during the years of study the staging criteria have been slightly modified for uterine cancer. We performed a series of sensitivity analyses based on stage and our findings were robust across a variety of stage groups. Lastly, we are unable to measure toxicity and quality of life. Combination therapy is clearly accompanied by a substantial rate of adverse events and cooperative group trials have consistently reported increased toxicity for multimodal therapies.
These data highlight the need for individualized treatment planning for women with stage III uterine cancer. While recent prospective trials have failed to demonstrate an overall survival benefit for combination chemotherapy, our findings suggest that in some women with stage III tumors there may be a survival benefit for a combination of chemotherapy and external beam radiation. The potential survival benefits must be balanced against the increased risk of toxicity. Going forward, further comparative effectiveness research is needed to help develop individualized, evidence-based strategies for the allocation of adjuvant therapy in women with stage III uterine cancer.
Supplementary Material
Acknowledgments
Financial Disclosure: Dr. Wright has served as a consultant for Tesaro and Clovis Oncology. Dr. Neugut has served as a consultant to Pfizer, Teva, Otsuka, Hospira, and United Biosource Corporation. He is on the scientific advisory board of EHE, Intl. The other authors did not report any potential conflicts of interest.
Dr. Wright (NCI R01CA169121-01A1) is the recipient of a grant from the National Cancer Institute. Dr. Hershman is the recipient of a grant from the Breast Cancer Research Foundation/Conquer Cancer Foundation.
REFERENCES
- 1.Siegel R, Miller K, Jemal A. Cancer statistics, 2018. CA Cancer J Clin. 2018;68(1):7–30 [DOI] [PubMed] [Google Scholar]
- 2.Albuquerque K, Chino J, Klopp A, Kamrava M, Beriwal S. Defining the Place of Adjuvant Chemotherapy and Radiation for High-Risk Endometrial Cancer From Recent Randomized Clinical Trials: Some Answers, More Questions. International Journal of Radiation Oncology Biology Physics. 2018;102(3):473–477 [DOI] [PubMed] [Google Scholar]
- 3.Ueda S, Kapp D, Cheung M et al. Trends in demographic and clinical characteristics in women diagnosed with corpus cancer and their potential impact on the increasing number of deaths. Am J Obstet Gynecol. 2008;198(2):218.e1–218.e6 [DOI] [PubMed] [Google Scholar]
- 4.American Cancer Society. Endometrial Cancer Survival Rates, by Stage. Available at: https://www.cancer.org/cancer/endometrial-cancer/detection-diagnosis-staging/survival-rates.html. Retrieved October 30, 2018
- 5.Hoekstra A, Kim R, Small W, Rademaker A, Helenowski I, Singh D, Schink J and Lurain J (2009). FIGO stage IIIC endometrial carcinoma: Prognostic factors and outcomes. Gynecologic Oncology, 114(2), pp.273–278. [DOI] [PubMed] [Google Scholar]
- 6.Lum M, Belnap T, Frandsen J, Brown A, Sause W, Soisson A, Dodson M, Werner T and Gaffney D (2015). Survival Analysis of Cancer Patients With FIGO Stage IIIA Endometrial Cancer. American Journal of Clinical Oncology, 38(3), pp.283–288. [DOI] [PubMed] [Google Scholar]
- 7.National Comprehensive Cancer Network. Uterine Neoplasms. (Version 1.2019). https://www.nccn.org/professionals/physician_gls/pdf/uterine.pdf
- 8.The National Cancer Database. Available at: https://www.facs.org/qualityprograms/cancer/ncdb. Retrieved October 30, 2018.
- 9.Bilimoria KY, Stewart AK, Winchester DP, Ko CY. The National Cancer Data Base: a powerful initiative to improve cancer care in the United States. Annals of surgical oncology. 2008;15(3):683–690 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lunceford JK, Davidian M. Stratification and weighting via the propensity score in estimation of causal treatment effects: a comparative study. Statistics in Medicine 2004; 23(19):2937–2960. [DOI] [PubMed] [Google Scholar]
- 11.Cole SR, Hernan MA. Adjusted survival curves with inverse probability weights. Computer Methods and Programs in Biomedicine 2004; 75:45–49. [DOI] [PubMed] [Google Scholar]
- 12.Cole SR, Hernan MA. Constructing inverse probability weights for marginal structural models. American Journal of Epidemiology 2008; 168(6):656–664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Austin PC. Variance estimation when using inverse probability of treatment weighting (IPTW) with survival analysis. Stat Med. 2016; 35:5642–5655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.De Boer S, Nout R, Bosse T, Creutzberg C. Adjuvant therapy for high-risk endometrial cancer: recent evidence and future directions. Expert Rev Anticancer Ther. 2018:1–10 [DOI] [PubMed] [Google Scholar]
- 15.Yoon M, Park W, Huh S, Kim H, Kim Y, Kim Y, Kim J, Lee J, Kim H, Cha J, Kim J, Kim J, Yoon W, Choi J, Chun M, Choi Y, Chang S, Lee K and Kim M (2015). A multicenter analysis of adjuvant therapy after surgery for stage IIIC endometrial adenocarcinoma: A Korean Radiation Oncology Group study (KROG 13–17). Gynecologic Oncology, 138(3), pp.519–525. [DOI] [PubMed] [Google Scholar]
- 16.Randall M, Filiaci V, Muss H et al. Randomized Phase III Trial of Whole-Abdominal Irradiation Versus Doxorubicin and Cisplatin Chemotherapy in Advanced Endometrial Carcinoma: A Gynecologic Oncology Group Study. Journal of Clinical Oncology. 2006;24(1):36–44 [DOI] [PubMed] [Google Scholar]
- 17.Nakayama K, Nagai Y, Ishikawa M, Aoki Y and Miyazaki K (2010). Concomitant postoperative radiation and chemotherapy following surgery was associated with improved overall survival in patients with FIGO stages III and IV endometrial cancer. International Journal of Clinical Oncology, 15(5), pp.440–446. [DOI] [PubMed] [Google Scholar]
- 18.Wong A, Rineer J, Lee Y, Schwartz D, Safdieh J, Weiner J, Choi K and Schreiber D (2016). Utilization of adjuvant therapies and their impact on survival for women with stage IIIC endometrial adenocarcinoma. Gynecologic Oncology, 142(3), pp.514–519. [DOI] [PubMed] [Google Scholar]
- 19.Wang C, Christie A, Folkert M, Xie X and Albuquerque K (2018). Value of combined adjuvant chemotherapy and radiation on survival for stage III uterine cancer: is less radiation equal to more?. Journal of Gynecologic Oncology, 29(4). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Secord A, Geller M, Broadwater G, Holloway R, Shuler K, Dao N, Gehrig P, O’Malley D, Finkler N and Havrilesky L (2013). A multicenter evaluation of adjuvant therapy in women with optimally resected stage IIIC endometrial cancer. Gynecologic Oncology, 128(1), pp.65–70. [DOI] [PubMed] [Google Scholar]
- 21.Maggi R, Lissoni A, Spina F et al. Adjuvant chemotherapy vs radiotherapy in high-risk endometrial carcinoma: results of a randomised trial. Br J Cancer. 2006;95(3):266–271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.de Boer S, Powell M, 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. The Lancet Oncology. 2018;19(3):295–309. doi: 10.1016/s1470-2045(18)30079-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wolford J, Tewari K. Highlights from the Gynecologic Oncology Track at the 2017 Annual Meeting of the American Society of Clinical Oncology. J Gynecol Oncol. 2017;28(5). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Matei D, Filiaci V, Randall M et al. A randomized phase III trial of cisplatin and tumor volume directed irradiation followed by carboplatin and paclitaxel vs. carboplatin and paclitaxel for optimally debulked, advanced endometrial carcinoma. Journal of Clinical Oncology. 2017;35(15):5505–5505 [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.