- Email: [email protected]
Evaluation of brachytherapy and external beam radiation therapy for early stage, node-negative uterine carcinosarcoma
Brachytherapy 14 (2015) 606e612
Evaluation of brachytherapy and external beam radiation therapy for early stage, node-negative uterine carcinosarcoma Nirav Patel1, Sarah E. Hegarty2, Leigh A. Cantrell3, Mark V. Mishra4, Timothy N. Showalter1,* 1 Department of Radiation Oncology, University of Virginia School of Medicine, Charlottesville, VA Division of Biostatistics, Department of Pharmacology & Experimental Therapeutics, Sidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, PA 3 Division of Gynecology Oncology, Department of Obstetrics & Gynecology, University of Virginia School of Medicine, Charlottesville, VA 4 Department of Radiation Oncology, University of Maryland School of Medicine, Baltimore, MD
2
ABSTRACT
PURPOSE: There is limited evidence to guide treatment decision making for patients with early stage uterine carcinosarcoma (UCS) regarding the use of pelvic external beam radiation therapy (RT) vs. vaginal brachytherapy (BT) after hysterectomy. We analyzed a population-based database to compare survival outcomes after adjuvant BT vs. pelvic external beam RT for patients with Stages IeII UCS. METHODS AND MATERIALS: We searched the Surveillance, Epidemiology, and End Results registry to identify a cohort of patients with International Federation of Gynecology and Obstetrics I/II UCS diagnosed during 1998e2010, who received a total hysterectomy and for whom radiotherapy type was known. c2 tests were used to test associations between patient characteristics and radiotherapy type. Overall and cancer-specific survival, measured from date of diagnosis, were summarized within each covariate. Cox proportional hazards models were used to model the impact of RT type on survival while adjusting for other factors. RESULTS: A total of 1581 subjects were identified, including 803 (50.8%) no radiotherapy; 636 (40.2%) external beam radiotherapy BT; and 142 (9.0%) BT alone. The use of BT alone increased from 4.5% in 1988e1999 to 12.5% in 2005e2010. Multivariate models of overall and causespecific survival showed that radiotherapy type was not associated with survival after adjustment for other covariates. CONCLUSIONS: For patients with Stages IeII UCS, adjuvant radiotherapy type did not influence survival after hysterectomy. This study addresses an existing evidence gap and identifies a trend toward increasing utilization of BT alone. Prospective trials are warranted to provide highquality evidence to guide adjuvant therapy decisions for these patients. Ó 2015 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.
Keywords:
Brachytherapy; Radiation therapy; Uterine cancer; Carcinosarcoma
Introduction Uterine carcinosarcoma (UCS), which was previously classified as malignant mixed mullerian tumor and malignant mixed mesodermal tumor, is a rare and aggressive malignancy that arises in the uterus. In the United States, the
Received 11 May 2015; received in revised form 3 June 2015; accepted 10 June 2015. Financial disclosure: The authors have no financial disclosures or conflicts of interest to report. * Corresponding author. Department of Radiation Oncology, University of Virginia School of Medicine, PO Box 800383, Charlottesville, VA 22908. Tel.: þ1-434-982-6282; fax: þ1-434-982-3262. E-mail address: [email protected] (T.N. Showalter).
annual incidence is less than two per 100,000 women per year (1). Prognosis for UCS remains relatively poor; despite accounting for less than 5% of uterine malignancies, they are associated with greater than 15 percent of uterine cancererelated deaths (1, 2). Five-year survival rates remain less than 50%, and median survival is only 21 months (1, 3). Standard staging and treatment for UCS includes total hysterectomy and bilateral salpingo-oophorectomy with pelvic and para-aortic lymph node dissection, cytology of peritoneal washings, omentectomy, and biopsies of peritoneal surfaces (4). After surgery, adjuvant therapy options for patients with no extrauterine spread include observation, chemotherapy, pelvic external beam radiation therapy
1538-4721/$ - see front matter Ó 2015 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brachy.2015.06.003
N. Patel et al. / Brachytherapy 14 (2015) 606e612
(EBRT), and/or vaginal brachytherapy (BT) (4). Although multimodality therapy is recommended, the optimal adjuvant therapy approach has not been clearly established (2). Because UCS is a rare tumor, there are few prospective studies to guide clinical decisions for adjuvant therapy, and most of the existing evidence comes from small, retrospective studies (2). The role of adjuvant chemotherapy has increased in recent years, based on limited data that suggest a clinical benefit from adjuvant chemotherapy (2). Although radiation therapy (RT) has been shown to provide some benefit in local control within the pelvis (5, 6), it is not clear how RT should be combined with chemotherapy and which modality (BT or EBRT) should be delivered (2). The only randomized control trial of adjuvant RT for UCS evaluated pelvic EBRT vs. observation (7). However, recent trends toward increasing role of lymphadenectomy for staging and the use of vaginal BT in node-negative patients with endometrioid adenocarcinoma, as well as trends toward increased use of chemotherapy for UCS, combine to encourage potential interest in BT for UCS. Studies comparing BT to pelvic radiation as adjuvant therapy after surgery in endometrial adenocarcinoma have shown that BT has similar rates of recurrence, metastases, and disease-free and overall survival (OS) (8e11). Although vaginal BT is a potential therapeutic option, there are limited data available regarding the potential benefit of BT added to chemotherapy for UCS (12, 13). To provide evidence to inform treatment decisions regarding vaginal BT vs. pelvic EBRT for patients with early stage UCS, we used an existing large, populationbased database to evaluate outcomes after pelvic EBRT and BT. We hypothesize that for patients with pathologically determined negative lymph nodes, disease-specific outcomes after BT will be similar to those observed after pelvic EBRT.
607
to be foreign-born and more often live in urban areas (15). This study was considered exempt from Institutional Review Board approval because the data provided by the SEER database is absent of any personal identifiers. Patient cohort SEER*Stat was queried for primary diagnosis of UCS (8950/3 mullerian mixed tumor, 8951/3 mesodermal mixed tumor, 8980/3 carcinosarcoma, not otherwise specified). This initial query yielded 1731 cases. Women were then excluded if they had International Federation of Gynecology and Obstetrics (FIGO) stage other than I or II and if they did not have a total hysterectomy (53 excluded). SEER tumor extent data were mapped to current FIGO staging for consistency across versions of SEER tumor extent coding. Women with unknown RT use or missing demographic information were also excluded (97 excluded) leaving 1581 subjects for analysis. The cohort selection strategy is summarized in Fig. 1.
Methods and materials Data source The Surveillance, Epidemiology, and End Results (SEER) Program of the U.S. National Cancer Institute was used to identify a cohort of female patients who were diagnosed with UCS between 1988 and 2010 (14). The SEER database routinely collects and provides comprehensive cancer case data from various population-based registries located around the United States. The program provides data on patient demographics, tumor anatomical location, tumor morphology, disease stage, treatments, and survival (15). This retrospective analysis used the November 2012 SEER 18-Registries 1973e2010 data set, which represents 27.8% of the total US population. The study population is comparable to the general US population with respect to education level and poverty, although the patients recorded in the SEER database are more likely
Fig. 1. Cohort selection strategy to identify patients in the SEER database who received hysterectomy for Stages IeII uterine carcinosarcoma. The EBRT brachytherapy cohort includes 176 patients who received both EBRT and brachytherapy and 460 patients treated with EBRT alone. SEER Surveillance, Epidemiology, and End Results; EBRT, external beam radiation therapy; RT, radiation therapy.
608
N. Patel et al. / Brachytherapy 14 (2015) 606e612
Statistical analysis
Results
c tests were used to test associations between patient characteristics and RT use (no RT, BT alone, and EBRT BT). OS and cancer-specific survival (CSS) were summarized with frequency and percentages within each potential covariate. KaplaneMeier plots and log-rank tests were used to test for univariate associations and assess proportionality. Cox proportional hazards models were then used to model the impact of RT on OS and CSS while adjusting for other factors. The following variables were included in all multivariate models regardless of significance: RT type, FIGO stage, race, age at diagnosis, SEER registry, and year of diagnosis. The remaining covariatesdmarital status, % foreign-born, % with !high school education, and % families below povertydwere retained only if the p-Value was less than 0.1 through a backward selection process. Significance was set at a 5 0.05; all tests are two-sided. Analyses were performed in SAS 9.4 (SAS Institute Inc., Cary, NC).
Of the 1581 women included in the analysis, 803 received no RT (50.8%), 636 (40.2%) received EBRT either alone or in combination with BT, and 142 (9.0%) received BT alone. The cohort of patients who received EBRT BT can be further divided into 176 patients who received both EBRT and BT and 460 patients treated with EBRT alone. Table 1 summarizes the clinical and demographic characteristics of patients in the treatment cohorts. Characteristics associated with receipt of RT include: year of diagnosis, SEER registry, FIGO stage, and census level values for high school education and poverty. Later year of diagnosis was associated with increased likelihood of RT receipt: 51.8% of women diagnosed between 2005 and 2010 received RT vs. 45% diagnosed during 1988e1999 ( p ! 0.001). More specifically, women diagnosed more recently were more likely to receive BT alone with 4.5% and 12.5% of patients between 1988e1999 and 2005e2010 receiving BT alone,
2
Table 1 Cohort characteristics (n 5 1581) Variable
All subjects
None
EBRT brachy
Brachy
p-Value
Uterine carcinosarcoma cases Year of dx, n (%) 1988e1999 2000e2004 2005e2010 SEER region, n (%) West South Northeast Midwest Age at dx, n (%) !55 55e64 65e75 75þ Race, n (%) Black Nonblack Marital status, n (%) Not married Married FIGO, n (%) IA IB/I,NOS II !High school educationa, n (%) !18% 18%þ Families below povertya, n (%) !8% 8%þ Foreign borna, n (%) !14% 14%þ
1581
803 (50.8)
636 (40.2)
142 (9.0)
d !0.001
291 (18.4) 451 (28.5) 839 (53.1)
160 (55.0) 239 (53.0) 404 (48.2)
118 (40.5) 188 (41.7) 330 (39.3)
13 (4.5) 24 (5.3) 105 (12.5)
788 259 296 238
(49.8) (16.4) (18.7) (15.1)
411 123 138 131
(52.2) (47.5) (46.6) (55.0)
322 122 101 91
(40.9) (47.1) (34.1) (38.2)
55 14 57 16
(7.0) (5.4) (19.3) (6.7)
200 506 487 388
(12.7) (32.0) (30.8) (24.5)
104 250 229 220
(52.0) (49.4) (47.0) (56.7)
77 211 210 138
(38.5) (41.7) (43.1) (35.6)
19 45 48 30
(9.5) (8.9) (9.9) (7.7)
!0.001
0.176
0.902 286 (18.1) 1295 (81.9)
147 (51.4) 656 (50.7)
112 (39.2) 524 (40.5)
27 (9.4) 115 (8.9)
767 (48.5) 814 (51.5)
403 (52.5) 400 (49.1)
303 (39.5) 333 (40.9)
61 (8.0) 81 (10.0)
895 (56.6) 476 (30.1) 210 (13.3)
493 (55.1) 240 (50.4) 70 (33.3)
314 (35.1) 204 (42.9) 118 (56.2)
88 (9.8) 32 (6.7) 22 (10.5)
848 (53.6) 733 (46.4)
434 (51.2) 369 (50.3)
309 (36.4) 327 (44.6)
105 (12.4) 37 (5.0)
810 (51.2) 771 (48.8)
417 (51.5) 386 (50.1)
299 (36.9) 337 (43.7)
94 (11.6) 48 (6.2)
791 (50.0) 790 (50.0)
393 (49.7) 410 (51.9)
323 (40.8) 313 (39.6)
75 (9.5) 67 (8.5)
0.241
!0.001
!0.001 !0.001
0.617
EBRT 5 external beam radiation therapy; SEER 5 Surveillance, Epidemiology, and End Results; FIGO 5 International Federation of Gynecology and Obstetrics. a From 2000 census.
N. Patel et al. / Brachytherapy 14 (2015) 606e612
respectively. Higher FIGO stage was also associated with a higher likelihood of RT use. RT was used in 44.9% of patients with FIGO Stage IA, 49.6% of patients with FIGO Stage IB or I not other specified, and 66.7% of patients with FIGO Stage II. Women who lived in an area with a larger high school educated population (O18%) were more likely to receive RT compared with women who lived in an area with a smaller high school educated population (!18%). Women who lived in poorer areas (O8% of families below poverty) were also more likely to receive RT compared with women who lived in areas with !8% of families below poverty. Receipt of RT also varied by region with the south and northeast delivering RT in 52.5% and 53.4% of cases, respectively, whereas the west and midwest delivering RT in 47.8% and 45.0% of cases, respectively.
609
respectively. Table 3 contains the full results of the final OS multivariate model. Note that receipt of BT was associated with longer OS than no RT. Figure 2 provides the KaplaneMeier plot of OS by RT type. Cause-specific survival Younger age at diagnosis, nonblack race, married, and FIGO Stage I (vs. II) were associated with longer CSS on univariate analysis (Supplemental Table 1). These results remained significant on multivariate analysis with longer CSS for subjects with a younger age at diagnosis, nonblack race, and FIGO Stage I (vs. II) (Table 3). RT type, including BT alone vs. EBRT BT, was not significantly associated with CSS in either the univariate and multivariate analyses. See Table 3 for the full results of the final CSS multivariate model. Figure 3 provides the KaplaneMeier plot of CSS by RT type.
Overall survival On univariate analysis (Supplemental Table 1), higher rates of OS were associated with younger age at diagnosis, nonblack race, FIGO Stage I (vs. II), married, and residing in an area with a greater foreign-born population ($14%). After adjustment for other covariates, the results were similar with higher rates of OS for subjects with a younger age at diagnosis, nonblack race, and FIGO Stage I (vs. II) (Table 2). RT type, including BT alone vs. EBRT BT, was not significantly associated with OS in both the univariate and multivariate analysis with p 5 0.153 and p 5 0.059, Table 2 Cox proportional hazards model of overall survival Covariate Radiotherapy use Brachy vs. no RT EBRT vs. no RT Brachy vs. EBRT FIGO stage IB/I,NOS vs. IA II vs. IA Race Black vs. nonblack SEER region Midwest vs. west Northeast vs. west South vs. west Year of diagnosis 2000e2004 vs. !1988e1999 2005e2010 vs. !1988e1999 Age at diagnosis 55e64 vs. !55 65e75 vs. !55 75 þ vs. !55
95% CI
0.67 0.90 0.74
(0.47, 0.95) (0.76, 1.06) (0.52, 1.05)
0.024 0.220
p
Overall p
1.49 1.84
(1.25, 1.78) (1.47, 2.29)
!0.001 !0.001
1.63
(1.33, 2.00)
!0.001
0.97 1.13 1.18
(0.78, 1.22) (0.91, 1.42) (0.93, 1.49)
0.818 0.264 0.176
0.96
(0.78, 1.19)
0.707
0.91
(0.73, 1.13)
0.399
0.059
!0.001 !0.001
0.396
0.693
(1.37, 2.84) (2.34, 4.76) (3.71, 7.54)
Total abdominal hysterectomy and bilateral salpingooophorectomy remains the mainstay of primary surgical treatment for UCS. The role of adjuvant RT and type of radiation treatment delivered remains controversial (16). This study analyzed a large cohort of 1581 women from the SEER database with FIGO Stages IeII UCS managed with Table 3 Cox proportional hazards model of cause-specific survival Covariate
HR
1.97 3.34 5.29
Discussion
!0.001 !0.001 !0.001
!0.001
HR 5 hazard ratio; CI 5 confidence interval; RT 5 radiation therapy; EBRT 5 external beam radiation therapy; FIGO 5 International Federation of Gynecology and Obstetrics; SEER 5 Surveillance, Epidemiology, and End Results.
Radiotherapy use Brachy vs. no RT EBRT vs. no RT Brachy vs. EBRT FIGO stage IB/I,NOS vs. IA II vs. IA Race Black vs. nonblack SEER region Midwest vs. west Northeast vs. west South vs. west Year of diagnosis 2000e2004 vs. 1988e1999 2005e2010 vs. 1988e1999 Age at diagnosis 55e64 vs. !55 65e75 vs. !55 75 þ vs. !55 !High school education 18% þ vs. !18%
HR
95% CI
p
Overall p
0.67 0.97 0.69
(0.44, 1.03) (0.79, 1.18) (0.45, 1.06)
0.066 0.754
1.61 2.20
(1.30, 1.99) (1.70, 2.86)
!0.001 !0.001
1.80
(1.42, 2.30)
!0.001
1.08 0.89 1.08
(0.82, 1.40) (0.67, 1.17) (0.82, 1.42)
0.591 0.398 0.607
1.06
(0.82, 1.37)
0.676
1.06
(0.82, 1.37)
0.654
1.89 2.56 3.72
(1.26, 2.82) (1.72, 3.80) (2.50, 5.53)
0.002 !0.0001 !0.0001
0.85
(0.69, 1.03)
0.099
0.183
!0.001 !0.001 0.640
0.890
!0.001
0.099
HR 5 hazard ratio; CI 5 confidence interval; RT 5 radiation therapy; EBRT 5 external beam radiation therapy; FIGO 5 International Federation of Gynecology and Obstetrics; SEER 5 Surveillance, Epidemiology, and End Results.
610
N. Patel et al. / Brachytherapy 14 (2015) 606e612
Fig. 2. Overall survival by adjuvant radiation therapy group. EBRT, external beam radiation therapy.
total hysterectomy. This included 636 (40.2%) women who received EBRT either alone or in combination with BT and 142 (9.0%) women who received BT alone. This is the first study to analyze and report practice patterns for RT in this patient population. After adjustment for clinical and demographic factors using a multivariate model, it was determined that RT type was not associated with improvements in OS and CSS. Younger age at diagnosis, nonblack race, married, and FIGO Stage I (vs. II) were found to be significant predictors of both OS and CSS. In the analysis of the cohort, BT was not observed to be associated with worse OS or CSS when compared with EBRT. However, the receipt of BT was associated with improved OS compared with no RT. The significance of this finding is not clear and should be interpreted with caution, as unmeasured confounders such as comorbidities could influence the analysis. A similar survival advantage was not seen when EBRT was compared with no RT. Previous studies have demonstrated contradictory results
Fig. 3. Cancer-specific survival by adjuvant radiation therapy group. EBRT, external beam radiation therapy.
regarding the association between adjuvant RT and survival with some studies only reporting local control without a survival benefit and other studies reporting an improvement in survival with adjuvant RT (3, 5,17e19). The reliability of the conclusions from this study is limited given its nonrandomized design. Allocation of patients to type of adjuvant RT may have been subject to channeling bias based on the presence of unmeasured confounders such medical comorbidities or more aggressive tumor features, which were not included in the analysis. Despite this potential limitation of the study, it is important to note that there is little evidence in the literature comparing BT to EBRT as adjuvant therapy for early stage UCS. Despite this, as seen in Table 1, there is an increasing trend toward using BT alone with 4.5% and 12.5% of patients receiving BT alone between 1988e1999 and 2005e2010, respectively. In the setting of increasing utilization of BT as adjuvant therapy after hysterectomy for UCS, the present study provides useful evidence to suggest that OS and CSS are not influenced by the use of BT over EBRT. Given the use of pelvic lymphadenectomy and the availability of adjuvant chemotherapy, vaginal BT may be a reasonable approach for adjuvant RT and its use in combined modality treatment of UCS warrants further investigation. Multivariate analysis of our cohort revealed several prognostic factors that are consistent with previous studies including: age at diagnosis, marital status, race, and FIGO score. Previous studies have reported that FIGO score is the most important prognostic factor for UCS (20). Our analysis also found that black women had relatively worse outcomes as compared with nonblack women. Previous studies, including gynecologic oncology group (GOG) 150, have reported race as a significant prognostic factor, and black patients were found to have a 59% and 26% higher risk of recurrence and death, respectively, when adjusted for stage, age, and treatment (21). A 2008 SEER analysis assessing the effects of lymphadenectomy and RT on uterine carcinoma found a similar association between race and survival after accounting for socioeconomic demographics, stage, and treatment. African Americans have been found to have a higher rate of advanced stage tumors and lower survival rates for tumors of the uterus (22). These differences in outcomes could be due to inherently more aggressive tumors in this select population, which could not be accounted for in our analysis. Our analysis also revealed that marital status was significantly associated with OS and CCS with married females having better outcomes. Previous studies have also described the significance of marriage as a prognostic factor for UCS. A 2013 SEER analysis of a large cohort of female and male patients with a variety of different cancers found that marriage has a positive influence on survival (23). The exact mechanism for this beneficial effect has not been determined; however, it has been found that unmarried cancer patients are more likely to refuse surgery and RT (24). Our study supports the association between marital status and survival.
N. Patel et al. / Brachytherapy 14 (2015) 606e612
There are several limitations to our study that must be mentioned by virtue of this being a retrospective study using the SEER database (25, 26). The analysis is limited by the variables provided by the SEER database, which does not include information such as medical comorbidities, tumor size, chemotherapy, and hormonal therapy. A multivariate analysis was performed to account for the clinical and demographic information provided by the SEER database; however, it is possible that the treatment groups were not homogenous, and without further information, it is not possible to account for these other variables in our analysis. This study could be complicated by confounding in which patients are allocated to treatment groups on a nonrandom basis with adjuvant RT more likely to be delivered to patients with larger, more aggressive tumors. It is possible that patients with complex comorbidities would be allocated to the EBRT group or no RT group over the BT group, potentially introducing another confounder to the analysis. It is also possible that the patients in the group that did not receive RT had lower grade tumors in which hormonal therapy, chemotherapy, or observation was chosen. Conversely, the patients in the group that did not receive RT may have had poor prognoses in which treatment would not have provided benefit. The SEER database also does not provide information regarding the timing of delivery of RT in the course of disease thus potentially introduces the immortal time bias for estimation of survival (27). Without information regarding treatment failure and time to relapse, it is also impossible to analyze event-free survival time. Furthermore, information regarding patterns of failure would be very important in considering the relative merits of BT and EBRT. Despite these limitations, this study is powered by a large cohort of patients, treated at both academic and community centers across the United States, and contributes to the paucity of evidence regarding the use of BT vs. EBRT in early stage UCS. Prospective trials are certainly warranted to further contribute to the evidence that guides RT therapy in these patients. In conclusion, the results of this large retrospective study of a population-based registry provide evidence regarding prognostic factors and comparative effectiveness of EBRT vs. BT for patients with Stages IeII UCS. BT was not shown to have worse OS or CSS than EBRT, suggesting that vaginal BT may be an effective alternative to EBRT for adjuvant therapy after hysterectomy for UCS. Age at diagnosis, race, marital status, and tumor stage were found to be significant predictors of survival. This study identifies an increasing trend toward using BT alone despite limited evidence to support its efficacy. The conclusions of this study must be interpreted with caution due to the potential biases and confounders present in this population study with nonrandom allocation of treatment. Given the lack of evidence, prospective trials are needed to help guide optimal therapeutic choices for patients with early stage UCS.
611
Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.brachy.2015.06.003. References [1] El-Nashar S, Mariani A. Uterine carcinosarcoma. Clin Obstet Gynecol 2011;54:292e304. [2] Cantrell LA, Blank SV, Duska LR. Uterine carcinosarcoma: a review of the literature. Gynecol Oncol 2015;137:581e588. [3] Gerszten K, Faul C, Kounelis S, et al. The impact of adjuvant radiotherapy on carcinosarcoma of the uterus. Gynecol Oncol 1998;68:8e13. [4] National Comprehensive Cancer Network. Uterine Neoplasms. NCCN Clinical Practice Guidelines in Oncology. Fort Washington, PA: National Comprehensive Cancer Network; 2015. [5] Chi DS, Mychalczak B, Saigo PE, et al. The role of whole-pelvic irradiation in the treatment of early-stage uterine carcinosarcoma. Gynecol Oncol 1997;65:493e498. [6] Sorbe B, Paulsson G, Andersson S, et al. A population-based series of uterine carcinosarcoma with long-term follow-up. Acta Oncol 2013;52: 759e766. [7] Reed NS, Mangioni C, Malmstrom H, et al. Phase III randomised study to evaluate the role of adjuvant pelvic radiotherapy in the treatment of uterine sarcomas stages I and II: an European Organisation for Research and Treatment of Cancer Gynaecological Cancer Group Study (protocol 55874). Eur J Cancer 2008;44:808e818. [8] Jolly S, Vargas C, Kumar T, et al. Vaginal brachytherapy alone: an alternative to adjuvant whole pelvis radiation for early stage endometrial cancer. Gynecol Oncol 2005;97:887e892. [9] Nout RA, Smit VT, Putter H, et al. Vaginal brachytherapy versus pelvic external beam radiotherapy for patients with endometrial cancer of high-intermediate risk (PORTEC-2): an open-label, non-inferiority, randomised trial. Lancet 2010;375:816e823. [10] Rittenberg PV, Lotocki RJ, Heywood MS, et al. Stage II endometrial carcinoma: limiting post-operative radiotherapy to the vaginal vault in node-negative tumors. Gynecol Oncol 2005;98:434e438. [11] Anderson JM, Stea B, Hallum AV, et al. High-dose-rate postoperative vaginal cuff irradiation alone for stage IB and IC endometrial cancer. Int J Radiat Oncol Biol Phys 2000;46:417e425. [12] Cantrell LA, Havrilesky L, Moore DT, et al. A multi-institutional cohort study of adjuvant therapy in stage I-II uterine carcinosarcoma. Gynecol Oncol 2012;127:22e26. [13] Makker V, Abu-Rustum NR, Alektiar KM, et al. A retrospective assessment of outcomes of chemotherapy-based versus radiationonly adjuvant treatment for completely resected stage I-IV uterine carcinosarcoma. Gynecol Oncol 2008;111:249e254. [14] Surveillance Research Program. National Cancer Institute SEER*Stat software version 8.1.5.; 2014. http://seer.cancer.gov/seerstat. [15] Yu JB, Gross CP, Wilson LD, et al. NCI SEER public-use data: applications and limitations in oncology research. Oncology (Williston Park) 2009;23:288e295. [16] Gadducci A, Cosio S, Romanini A, et al. The management of patients with uterine sarcoma: a debated clinical challenge. Crit Rev Oncol Hematol 2008;65:129e142. [17] Smith DC, Macdonald OK, Gaffney DK. The impact of adjuvant radiation therapy on survival in women with uterine carcinosarcoma. Radiother Oncol 2008;88:227e232. [18] Hornback NB, Omura G, Major FJ. Observations on the use of adjuvant radiation therapy in patients with stage I and II uterine sarcoma. Int J Radiat Oncol Biol Phys 1986;12:2127e2130. [19] Knocke TH, Kucera H, Dorfler D, et al. Results of postoperative radiotherapy in the treatment of sarcoma of the corpus uteri. Cancer 1998;83:1972e1979. [20] Callister M, Ramondetta LM, Jhingran A, et al. Malignant mixed M€ullerian tumors of the uterus: analysis of patterns of failure,
612
N. Patel et al. / Brachytherapy 14 (2015) 606e612
prognostic factors, and treatment outcome. Int J Radiat Oncol Biol Phys 2004;58:786e796. [21] Wolfson AH, Brady MF, Rocereto T, et al. A gynecologic oncology group randomized phase III trial of whole abdominal irradiation (WAI) vs. cisplatin-ifosfamide and mesna (CIM) as post-surgical therapy in stage I-IV carcinosarcoma (CS) of the uterus. Gynecol Oncol 2007;107:177e185. [22] Sherman ME, Devesa SS. Analysis of racial differences in incidence, survival, and mortality for malignant tumors of the uterine corpus. Cancer 2003;98:176e186. [23] Aizer AA, Chen MH, McCarthy EP, et al. Marital status and survival in patients with cancer. J Clin Oncol 2013;31:3869e3876.
[24] Aizer AA, Chen MH, Parekh A, et al. Refusal of curative radiation therapy and surgery among patients with cancer. Int J Radiat Oncol Biol Phys 2014;89:756e764. [25] Park HS, Lloyd S, Decker RH, et al. Limitations and biases of the Surveillance, Epidemiology, and End Results database. Curr Probl Cancer 2012;36:216e224. [26] Armstrong K. Methods in comparative effectiveness research. J Clin Oncol 2012;30:4208e4214. [27] Park HS, Gross CP, Makarov DV, et al. Immortal time bias: a frequently unrecognized threat to validity in the evaluation of postoperative radiotherapy. Int J Radiat Oncol Biol Phys 2012;83: 1365e1373.
Evaluation of brachytherapy and external beam radiation therapy for early stage, node-negative uterine carcinosarcoma Nirav Patel1, Sarah E. Hegarty2, Leigh A. Cantrell3, Mark V. Mishra4, Timothy N. Showalter1,* 1 Department of Radiation Oncology, University of Virginia School of Medicine, Charlottesville, VA Division of Biostatistics, Department of Pharmacology & Experimental Therapeutics, Sidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, PA 3 Division of Gynecology Oncology, Department of Obstetrics & Gynecology, University of Virginia School of Medicine, Charlottesville, VA 4 Department of Radiation Oncology, University of Maryland School of Medicine, Baltimore, MD
2
ABSTRACT
PURPOSE: There is limited evidence to guide treatment decision making for patients with early stage uterine carcinosarcoma (UCS) regarding the use of pelvic external beam radiation therapy (RT) vs. vaginal brachytherapy (BT) after hysterectomy. We analyzed a population-based database to compare survival outcomes after adjuvant BT vs. pelvic external beam RT for patients with Stages IeII UCS. METHODS AND MATERIALS: We searched the Surveillance, Epidemiology, and End Results registry to identify a cohort of patients with International Federation of Gynecology and Obstetrics I/II UCS diagnosed during 1998e2010, who received a total hysterectomy and for whom radiotherapy type was known. c2 tests were used to test associations between patient characteristics and radiotherapy type. Overall and cancer-specific survival, measured from date of diagnosis, were summarized within each covariate. Cox proportional hazards models were used to model the impact of RT type on survival while adjusting for other factors. RESULTS: A total of 1581 subjects were identified, including 803 (50.8%) no radiotherapy; 636 (40.2%) external beam radiotherapy BT; and 142 (9.0%) BT alone. The use of BT alone increased from 4.5% in 1988e1999 to 12.5% in 2005e2010. Multivariate models of overall and causespecific survival showed that radiotherapy type was not associated with survival after adjustment for other covariates. CONCLUSIONS: For patients with Stages IeII UCS, adjuvant radiotherapy type did not influence survival after hysterectomy. This study addresses an existing evidence gap and identifies a trend toward increasing utilization of BT alone. Prospective trials are warranted to provide highquality evidence to guide adjuvant therapy decisions for these patients. Ó 2015 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.
Keywords:
Brachytherapy; Radiation therapy; Uterine cancer; Carcinosarcoma
Introduction Uterine carcinosarcoma (UCS), which was previously classified as malignant mixed mullerian tumor and malignant mixed mesodermal tumor, is a rare and aggressive malignancy that arises in the uterus. In the United States, the
Received 11 May 2015; received in revised form 3 June 2015; accepted 10 June 2015. Financial disclosure: The authors have no financial disclosures or conflicts of interest to report. * Corresponding author. Department of Radiation Oncology, University of Virginia School of Medicine, PO Box 800383, Charlottesville, VA 22908. Tel.: þ1-434-982-6282; fax: þ1-434-982-3262. E-mail address: [email protected] (T.N. Showalter).
annual incidence is less than two per 100,000 women per year (1). Prognosis for UCS remains relatively poor; despite accounting for less than 5% of uterine malignancies, they are associated with greater than 15 percent of uterine cancererelated deaths (1, 2). Five-year survival rates remain less than 50%, and median survival is only 21 months (1, 3). Standard staging and treatment for UCS includes total hysterectomy and bilateral salpingo-oophorectomy with pelvic and para-aortic lymph node dissection, cytology of peritoneal washings, omentectomy, and biopsies of peritoneal surfaces (4). After surgery, adjuvant therapy options for patients with no extrauterine spread include observation, chemotherapy, pelvic external beam radiation therapy
1538-4721/$ - see front matter Ó 2015 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brachy.2015.06.003
N. Patel et al. / Brachytherapy 14 (2015) 606e612
(EBRT), and/or vaginal brachytherapy (BT) (4). Although multimodality therapy is recommended, the optimal adjuvant therapy approach has not been clearly established (2). Because UCS is a rare tumor, there are few prospective studies to guide clinical decisions for adjuvant therapy, and most of the existing evidence comes from small, retrospective studies (2). The role of adjuvant chemotherapy has increased in recent years, based on limited data that suggest a clinical benefit from adjuvant chemotherapy (2). Although radiation therapy (RT) has been shown to provide some benefit in local control within the pelvis (5, 6), it is not clear how RT should be combined with chemotherapy and which modality (BT or EBRT) should be delivered (2). The only randomized control trial of adjuvant RT for UCS evaluated pelvic EBRT vs. observation (7). However, recent trends toward increasing role of lymphadenectomy for staging and the use of vaginal BT in node-negative patients with endometrioid adenocarcinoma, as well as trends toward increased use of chemotherapy for UCS, combine to encourage potential interest in BT for UCS. Studies comparing BT to pelvic radiation as adjuvant therapy after surgery in endometrial adenocarcinoma have shown that BT has similar rates of recurrence, metastases, and disease-free and overall survival (OS) (8e11). Although vaginal BT is a potential therapeutic option, there are limited data available regarding the potential benefit of BT added to chemotherapy for UCS (12, 13). To provide evidence to inform treatment decisions regarding vaginal BT vs. pelvic EBRT for patients with early stage UCS, we used an existing large, populationbased database to evaluate outcomes after pelvic EBRT and BT. We hypothesize that for patients with pathologically determined negative lymph nodes, disease-specific outcomes after BT will be similar to those observed after pelvic EBRT.
607
to be foreign-born and more often live in urban areas (15). This study was considered exempt from Institutional Review Board approval because the data provided by the SEER database is absent of any personal identifiers. Patient cohort SEER*Stat was queried for primary diagnosis of UCS (8950/3 mullerian mixed tumor, 8951/3 mesodermal mixed tumor, 8980/3 carcinosarcoma, not otherwise specified). This initial query yielded 1731 cases. Women were then excluded if they had International Federation of Gynecology and Obstetrics (FIGO) stage other than I or II and if they did not have a total hysterectomy (53 excluded). SEER tumor extent data were mapped to current FIGO staging for consistency across versions of SEER tumor extent coding. Women with unknown RT use or missing demographic information were also excluded (97 excluded) leaving 1581 subjects for analysis. The cohort selection strategy is summarized in Fig. 1.
Methods and materials Data source The Surveillance, Epidemiology, and End Results (SEER) Program of the U.S. National Cancer Institute was used to identify a cohort of female patients who were diagnosed with UCS between 1988 and 2010 (14). The SEER database routinely collects and provides comprehensive cancer case data from various population-based registries located around the United States. The program provides data on patient demographics, tumor anatomical location, tumor morphology, disease stage, treatments, and survival (15). This retrospective analysis used the November 2012 SEER 18-Registries 1973e2010 data set, which represents 27.8% of the total US population. The study population is comparable to the general US population with respect to education level and poverty, although the patients recorded in the SEER database are more likely
Fig. 1. Cohort selection strategy to identify patients in the SEER database who received hysterectomy for Stages IeII uterine carcinosarcoma. The EBRT brachytherapy cohort includes 176 patients who received both EBRT and brachytherapy and 460 patients treated with EBRT alone. SEER Surveillance, Epidemiology, and End Results; EBRT, external beam radiation therapy; RT, radiation therapy.
608
N. Patel et al. / Brachytherapy 14 (2015) 606e612
Statistical analysis
Results
c tests were used to test associations between patient characteristics and RT use (no RT, BT alone, and EBRT BT). OS and cancer-specific survival (CSS) were summarized with frequency and percentages within each potential covariate. KaplaneMeier plots and log-rank tests were used to test for univariate associations and assess proportionality. Cox proportional hazards models were then used to model the impact of RT on OS and CSS while adjusting for other factors. The following variables were included in all multivariate models regardless of significance: RT type, FIGO stage, race, age at diagnosis, SEER registry, and year of diagnosis. The remaining covariatesdmarital status, % foreign-born, % with !high school education, and % families below povertydwere retained only if the p-Value was less than 0.1 through a backward selection process. Significance was set at a 5 0.05; all tests are two-sided. Analyses were performed in SAS 9.4 (SAS Institute Inc., Cary, NC).
Of the 1581 women included in the analysis, 803 received no RT (50.8%), 636 (40.2%) received EBRT either alone or in combination with BT, and 142 (9.0%) received BT alone. The cohort of patients who received EBRT BT can be further divided into 176 patients who received both EBRT and BT and 460 patients treated with EBRT alone. Table 1 summarizes the clinical and demographic characteristics of patients in the treatment cohorts. Characteristics associated with receipt of RT include: year of diagnosis, SEER registry, FIGO stage, and census level values for high school education and poverty. Later year of diagnosis was associated with increased likelihood of RT receipt: 51.8% of women diagnosed between 2005 and 2010 received RT vs. 45% diagnosed during 1988e1999 ( p ! 0.001). More specifically, women diagnosed more recently were more likely to receive BT alone with 4.5% and 12.5% of patients between 1988e1999 and 2005e2010 receiving BT alone,
2
Table 1 Cohort characteristics (n 5 1581) Variable
All subjects
None
EBRT brachy
Brachy
p-Value
Uterine carcinosarcoma cases Year of dx, n (%) 1988e1999 2000e2004 2005e2010 SEER region, n (%) West South Northeast Midwest Age at dx, n (%) !55 55e64 65e75 75þ Race, n (%) Black Nonblack Marital status, n (%) Not married Married FIGO, n (%) IA IB/I,NOS II !High school educationa, n (%) !18% 18%þ Families below povertya, n (%) !8% 8%þ Foreign borna, n (%) !14% 14%þ
1581
803 (50.8)
636 (40.2)
142 (9.0)
d !0.001
291 (18.4) 451 (28.5) 839 (53.1)
160 (55.0) 239 (53.0) 404 (48.2)
118 (40.5) 188 (41.7) 330 (39.3)
13 (4.5) 24 (5.3) 105 (12.5)
788 259 296 238
(49.8) (16.4) (18.7) (15.1)
411 123 138 131
(52.2) (47.5) (46.6) (55.0)
322 122 101 91
(40.9) (47.1) (34.1) (38.2)
55 14 57 16
(7.0) (5.4) (19.3) (6.7)
200 506 487 388
(12.7) (32.0) (30.8) (24.5)
104 250 229 220
(52.0) (49.4) (47.0) (56.7)
77 211 210 138
(38.5) (41.7) (43.1) (35.6)
19 45 48 30
(9.5) (8.9) (9.9) (7.7)
!0.001
0.176
0.902 286 (18.1) 1295 (81.9)
147 (51.4) 656 (50.7)
112 (39.2) 524 (40.5)
27 (9.4) 115 (8.9)
767 (48.5) 814 (51.5)
403 (52.5) 400 (49.1)
303 (39.5) 333 (40.9)
61 (8.0) 81 (10.0)
895 (56.6) 476 (30.1) 210 (13.3)
493 (55.1) 240 (50.4) 70 (33.3)
314 (35.1) 204 (42.9) 118 (56.2)
88 (9.8) 32 (6.7) 22 (10.5)
848 (53.6) 733 (46.4)
434 (51.2) 369 (50.3)
309 (36.4) 327 (44.6)
105 (12.4) 37 (5.0)
810 (51.2) 771 (48.8)
417 (51.5) 386 (50.1)
299 (36.9) 337 (43.7)
94 (11.6) 48 (6.2)
791 (50.0) 790 (50.0)
393 (49.7) 410 (51.9)
323 (40.8) 313 (39.6)
75 (9.5) 67 (8.5)
0.241
!0.001
!0.001 !0.001
0.617
EBRT 5 external beam radiation therapy; SEER 5 Surveillance, Epidemiology, and End Results; FIGO 5 International Federation of Gynecology and Obstetrics. a From 2000 census.
N. Patel et al. / Brachytherapy 14 (2015) 606e612
respectively. Higher FIGO stage was also associated with a higher likelihood of RT use. RT was used in 44.9% of patients with FIGO Stage IA, 49.6% of patients with FIGO Stage IB or I not other specified, and 66.7% of patients with FIGO Stage II. Women who lived in an area with a larger high school educated population (O18%) were more likely to receive RT compared with women who lived in an area with a smaller high school educated population (!18%). Women who lived in poorer areas (O8% of families below poverty) were also more likely to receive RT compared with women who lived in areas with !8% of families below poverty. Receipt of RT also varied by region with the south and northeast delivering RT in 52.5% and 53.4% of cases, respectively, whereas the west and midwest delivering RT in 47.8% and 45.0% of cases, respectively.
609
respectively. Table 3 contains the full results of the final OS multivariate model. Note that receipt of BT was associated with longer OS than no RT. Figure 2 provides the KaplaneMeier plot of OS by RT type. Cause-specific survival Younger age at diagnosis, nonblack race, married, and FIGO Stage I (vs. II) were associated with longer CSS on univariate analysis (Supplemental Table 1). These results remained significant on multivariate analysis with longer CSS for subjects with a younger age at diagnosis, nonblack race, and FIGO Stage I (vs. II) (Table 3). RT type, including BT alone vs. EBRT BT, was not significantly associated with CSS in either the univariate and multivariate analyses. See Table 3 for the full results of the final CSS multivariate model. Figure 3 provides the KaplaneMeier plot of CSS by RT type.
Overall survival On univariate analysis (Supplemental Table 1), higher rates of OS were associated with younger age at diagnosis, nonblack race, FIGO Stage I (vs. II), married, and residing in an area with a greater foreign-born population ($14%). After adjustment for other covariates, the results were similar with higher rates of OS for subjects with a younger age at diagnosis, nonblack race, and FIGO Stage I (vs. II) (Table 2). RT type, including BT alone vs. EBRT BT, was not significantly associated with OS in both the univariate and multivariate analysis with p 5 0.153 and p 5 0.059, Table 2 Cox proportional hazards model of overall survival Covariate Radiotherapy use Brachy vs. no RT EBRT vs. no RT Brachy vs. EBRT FIGO stage IB/I,NOS vs. IA II vs. IA Race Black vs. nonblack SEER region Midwest vs. west Northeast vs. west South vs. west Year of diagnosis 2000e2004 vs. !1988e1999 2005e2010 vs. !1988e1999 Age at diagnosis 55e64 vs. !55 65e75 vs. !55 75 þ vs. !55
95% CI
0.67 0.90 0.74
(0.47, 0.95) (0.76, 1.06) (0.52, 1.05)
0.024 0.220
p
Overall p
1.49 1.84
(1.25, 1.78) (1.47, 2.29)
!0.001 !0.001
1.63
(1.33, 2.00)
!0.001
0.97 1.13 1.18
(0.78, 1.22) (0.91, 1.42) (0.93, 1.49)
0.818 0.264 0.176
0.96
(0.78, 1.19)
0.707
0.91
(0.73, 1.13)
0.399
0.059
!0.001 !0.001
0.396
0.693
(1.37, 2.84) (2.34, 4.76) (3.71, 7.54)
Total abdominal hysterectomy and bilateral salpingooophorectomy remains the mainstay of primary surgical treatment for UCS. The role of adjuvant RT and type of radiation treatment delivered remains controversial (16). This study analyzed a large cohort of 1581 women from the SEER database with FIGO Stages IeII UCS managed with Table 3 Cox proportional hazards model of cause-specific survival Covariate
HR
1.97 3.34 5.29
Discussion
!0.001 !0.001 !0.001
!0.001
HR 5 hazard ratio; CI 5 confidence interval; RT 5 radiation therapy; EBRT 5 external beam radiation therapy; FIGO 5 International Federation of Gynecology and Obstetrics; SEER 5 Surveillance, Epidemiology, and End Results.
Radiotherapy use Brachy vs. no RT EBRT vs. no RT Brachy vs. EBRT FIGO stage IB/I,NOS vs. IA II vs. IA Race Black vs. nonblack SEER region Midwest vs. west Northeast vs. west South vs. west Year of diagnosis 2000e2004 vs. 1988e1999 2005e2010 vs. 1988e1999 Age at diagnosis 55e64 vs. !55 65e75 vs. !55 75 þ vs. !55 !High school education 18% þ vs. !18%
HR
95% CI
p
Overall p
0.67 0.97 0.69
(0.44, 1.03) (0.79, 1.18) (0.45, 1.06)
0.066 0.754
1.61 2.20
(1.30, 1.99) (1.70, 2.86)
!0.001 !0.001
1.80
(1.42, 2.30)
!0.001
1.08 0.89 1.08
(0.82, 1.40) (0.67, 1.17) (0.82, 1.42)
0.591 0.398 0.607
1.06
(0.82, 1.37)
0.676
1.06
(0.82, 1.37)
0.654
1.89 2.56 3.72
(1.26, 2.82) (1.72, 3.80) (2.50, 5.53)
0.002 !0.0001 !0.0001
0.85
(0.69, 1.03)
0.099
0.183
!0.001 !0.001 0.640
0.890
!0.001
0.099
HR 5 hazard ratio; CI 5 confidence interval; RT 5 radiation therapy; EBRT 5 external beam radiation therapy; FIGO 5 International Federation of Gynecology and Obstetrics; SEER 5 Surveillance, Epidemiology, and End Results.
610
N. Patel et al. / Brachytherapy 14 (2015) 606e612
Fig. 2. Overall survival by adjuvant radiation therapy group. EBRT, external beam radiation therapy.
total hysterectomy. This included 636 (40.2%) women who received EBRT either alone or in combination with BT and 142 (9.0%) women who received BT alone. This is the first study to analyze and report practice patterns for RT in this patient population. After adjustment for clinical and demographic factors using a multivariate model, it was determined that RT type was not associated with improvements in OS and CSS. Younger age at diagnosis, nonblack race, married, and FIGO Stage I (vs. II) were found to be significant predictors of both OS and CSS. In the analysis of the cohort, BT was not observed to be associated with worse OS or CSS when compared with EBRT. However, the receipt of BT was associated with improved OS compared with no RT. The significance of this finding is not clear and should be interpreted with caution, as unmeasured confounders such as comorbidities could influence the analysis. A similar survival advantage was not seen when EBRT was compared with no RT. Previous studies have demonstrated contradictory results
Fig. 3. Cancer-specific survival by adjuvant radiation therapy group. EBRT, external beam radiation therapy.
regarding the association between adjuvant RT and survival with some studies only reporting local control without a survival benefit and other studies reporting an improvement in survival with adjuvant RT (3, 5,17e19). The reliability of the conclusions from this study is limited given its nonrandomized design. Allocation of patients to type of adjuvant RT may have been subject to channeling bias based on the presence of unmeasured confounders such medical comorbidities or more aggressive tumor features, which were not included in the analysis. Despite this potential limitation of the study, it is important to note that there is little evidence in the literature comparing BT to EBRT as adjuvant therapy for early stage UCS. Despite this, as seen in Table 1, there is an increasing trend toward using BT alone with 4.5% and 12.5% of patients receiving BT alone between 1988e1999 and 2005e2010, respectively. In the setting of increasing utilization of BT as adjuvant therapy after hysterectomy for UCS, the present study provides useful evidence to suggest that OS and CSS are not influenced by the use of BT over EBRT. Given the use of pelvic lymphadenectomy and the availability of adjuvant chemotherapy, vaginal BT may be a reasonable approach for adjuvant RT and its use in combined modality treatment of UCS warrants further investigation. Multivariate analysis of our cohort revealed several prognostic factors that are consistent with previous studies including: age at diagnosis, marital status, race, and FIGO score. Previous studies have reported that FIGO score is the most important prognostic factor for UCS (20). Our analysis also found that black women had relatively worse outcomes as compared with nonblack women. Previous studies, including gynecologic oncology group (GOG) 150, have reported race as a significant prognostic factor, and black patients were found to have a 59% and 26% higher risk of recurrence and death, respectively, when adjusted for stage, age, and treatment (21). A 2008 SEER analysis assessing the effects of lymphadenectomy and RT on uterine carcinoma found a similar association between race and survival after accounting for socioeconomic demographics, stage, and treatment. African Americans have been found to have a higher rate of advanced stage tumors and lower survival rates for tumors of the uterus (22). These differences in outcomes could be due to inherently more aggressive tumors in this select population, which could not be accounted for in our analysis. Our analysis also revealed that marital status was significantly associated with OS and CCS with married females having better outcomes. Previous studies have also described the significance of marriage as a prognostic factor for UCS. A 2013 SEER analysis of a large cohort of female and male patients with a variety of different cancers found that marriage has a positive influence on survival (23). The exact mechanism for this beneficial effect has not been determined; however, it has been found that unmarried cancer patients are more likely to refuse surgery and RT (24). Our study supports the association between marital status and survival.
N. Patel et al. / Brachytherapy 14 (2015) 606e612
There are several limitations to our study that must be mentioned by virtue of this being a retrospective study using the SEER database (25, 26). The analysis is limited by the variables provided by the SEER database, which does not include information such as medical comorbidities, tumor size, chemotherapy, and hormonal therapy. A multivariate analysis was performed to account for the clinical and demographic information provided by the SEER database; however, it is possible that the treatment groups were not homogenous, and without further information, it is not possible to account for these other variables in our analysis. This study could be complicated by confounding in which patients are allocated to treatment groups on a nonrandom basis with adjuvant RT more likely to be delivered to patients with larger, more aggressive tumors. It is possible that patients with complex comorbidities would be allocated to the EBRT group or no RT group over the BT group, potentially introducing another confounder to the analysis. It is also possible that the patients in the group that did not receive RT had lower grade tumors in which hormonal therapy, chemotherapy, or observation was chosen. Conversely, the patients in the group that did not receive RT may have had poor prognoses in which treatment would not have provided benefit. The SEER database also does not provide information regarding the timing of delivery of RT in the course of disease thus potentially introduces the immortal time bias for estimation of survival (27). Without information regarding treatment failure and time to relapse, it is also impossible to analyze event-free survival time. Furthermore, information regarding patterns of failure would be very important in considering the relative merits of BT and EBRT. Despite these limitations, this study is powered by a large cohort of patients, treated at both academic and community centers across the United States, and contributes to the paucity of evidence regarding the use of BT vs. EBRT in early stage UCS. Prospective trials are certainly warranted to further contribute to the evidence that guides RT therapy in these patients. In conclusion, the results of this large retrospective study of a population-based registry provide evidence regarding prognostic factors and comparative effectiveness of EBRT vs. BT for patients with Stages IeII UCS. BT was not shown to have worse OS or CSS than EBRT, suggesting that vaginal BT may be an effective alternative to EBRT for adjuvant therapy after hysterectomy for UCS. Age at diagnosis, race, marital status, and tumor stage were found to be significant predictors of survival. This study identifies an increasing trend toward using BT alone despite limited evidence to support its efficacy. The conclusions of this study must be interpreted with caution due to the potential biases and confounders present in this population study with nonrandom allocation of treatment. Given the lack of evidence, prospective trials are needed to help guide optimal therapeutic choices for patients with early stage UCS.
611
Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.brachy.2015.06.003. References [1] El-Nashar S, Mariani A. Uterine carcinosarcoma. Clin Obstet Gynecol 2011;54:292e304. [2] Cantrell LA, Blank SV, Duska LR. Uterine carcinosarcoma: a review of the literature. Gynecol Oncol 2015;137:581e588. [3] Gerszten K, Faul C, Kounelis S, et al. The impact of adjuvant radiotherapy on carcinosarcoma of the uterus. Gynecol Oncol 1998;68:8e13. [4] National Comprehensive Cancer Network. Uterine Neoplasms. NCCN Clinical Practice Guidelines in Oncology. Fort Washington, PA: National Comprehensive Cancer Network; 2015. [5] Chi DS, Mychalczak B, Saigo PE, et al. The role of whole-pelvic irradiation in the treatment of early-stage uterine carcinosarcoma. Gynecol Oncol 1997;65:493e498. [6] Sorbe B, Paulsson G, Andersson S, et al. A population-based series of uterine carcinosarcoma with long-term follow-up. Acta Oncol 2013;52: 759e766. [7] Reed NS, Mangioni C, Malmstrom H, et al. Phase III randomised study to evaluate the role of adjuvant pelvic radiotherapy in the treatment of uterine sarcomas stages I and II: an European Organisation for Research and Treatment of Cancer Gynaecological Cancer Group Study (protocol 55874). Eur J Cancer 2008;44:808e818. [8] Jolly S, Vargas C, Kumar T, et al. Vaginal brachytherapy alone: an alternative to adjuvant whole pelvis radiation for early stage endometrial cancer. Gynecol Oncol 2005;97:887e892. [9] Nout RA, Smit VT, Putter H, et al. Vaginal brachytherapy versus pelvic external beam radiotherapy for patients with endometrial cancer of high-intermediate risk (PORTEC-2): an open-label, non-inferiority, randomised trial. Lancet 2010;375:816e823. [10] Rittenberg PV, Lotocki RJ, Heywood MS, et al. Stage II endometrial carcinoma: limiting post-operative radiotherapy to the vaginal vault in node-negative tumors. Gynecol Oncol 2005;98:434e438. [11] Anderson JM, Stea B, Hallum AV, et al. High-dose-rate postoperative vaginal cuff irradiation alone for stage IB and IC endometrial cancer. Int J Radiat Oncol Biol Phys 2000;46:417e425. [12] Cantrell LA, Havrilesky L, Moore DT, et al. A multi-institutional cohort study of adjuvant therapy in stage I-II uterine carcinosarcoma. Gynecol Oncol 2012;127:22e26. [13] Makker V, Abu-Rustum NR, Alektiar KM, et al. A retrospective assessment of outcomes of chemotherapy-based versus radiationonly adjuvant treatment for completely resected stage I-IV uterine carcinosarcoma. Gynecol Oncol 2008;111:249e254. [14] Surveillance Research Program. National Cancer Institute SEER*Stat software version 8.1.5.; 2014. http://seer.cancer.gov/seerstat. [15] Yu JB, Gross CP, Wilson LD, et al. NCI SEER public-use data: applications and limitations in oncology research. Oncology (Williston Park) 2009;23:288e295. [16] Gadducci A, Cosio S, Romanini A, et al. The management of patients with uterine sarcoma: a debated clinical challenge. Crit Rev Oncol Hematol 2008;65:129e142. [17] Smith DC, Macdonald OK, Gaffney DK. The impact of adjuvant radiation therapy on survival in women with uterine carcinosarcoma. Radiother Oncol 2008;88:227e232. [18] Hornback NB, Omura G, Major FJ. Observations on the use of adjuvant radiation therapy in patients with stage I and II uterine sarcoma. Int J Radiat Oncol Biol Phys 1986;12:2127e2130. [19] Knocke TH, Kucera H, Dorfler D, et al. Results of postoperative radiotherapy in the treatment of sarcoma of the corpus uteri. Cancer 1998;83:1972e1979. [20] Callister M, Ramondetta LM, Jhingran A, et al. Malignant mixed M€ullerian tumors of the uterus: analysis of patterns of failure,
612
N. Patel et al. / Brachytherapy 14 (2015) 606e612
prognostic factors, and treatment outcome. Int J Radiat Oncol Biol Phys 2004;58:786e796. [21] Wolfson AH, Brady MF, Rocereto T, et al. A gynecologic oncology group randomized phase III trial of whole abdominal irradiation (WAI) vs. cisplatin-ifosfamide and mesna (CIM) as post-surgical therapy in stage I-IV carcinosarcoma (CS) of the uterus. Gynecol Oncol 2007;107:177e185. [22] Sherman ME, Devesa SS. Analysis of racial differences in incidence, survival, and mortality for malignant tumors of the uterine corpus. Cancer 2003;98:176e186. [23] Aizer AA, Chen MH, McCarthy EP, et al. Marital status and survival in patients with cancer. J Clin Oncol 2013;31:3869e3876.
[24] Aizer AA, Chen MH, Parekh A, et al. Refusal of curative radiation therapy and surgery among patients with cancer. Int J Radiat Oncol Biol Phys 2014;89:756e764. [25] Park HS, Lloyd S, Decker RH, et al. Limitations and biases of the Surveillance, Epidemiology, and End Results database. Curr Probl Cancer 2012;36:216e224. [26] Armstrong K. Methods in comparative effectiveness research. J Clin Oncol 2012;30:4208e4214. [27] Park HS, Gross CP, Makarov DV, et al. Immortal time bias: a frequently unrecognized threat to validity in the evaluation of postoperative radiotherapy. Int J Radiat Oncol Biol Phys 2012;83: 1365e1373.