G Model
CANEP-797; No. of Pages 7 Cancer Epidemiology xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention journal homepage: www.cancerepidemiology.net
Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database Kyle R. Duchman, Yubo Gao, Benjamin J. Miller * Department of Orthopaedics and Rehabilitation, University of Iowa, 200 Hawkins Drive, 01015 JPP, Iowa City, IA 52242, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 September 2014 Received in revised form 22 December 2014 Accepted 26 December 2014 Available online xxx
Background: The current study aims to determine cause-specific survival in patients with Ewing’s sarcoma while reporting clinical risk factors for survival. Methods: The Surveillance, Epidemiology, and End Results (SEER) Program database was used to identify patients with osseous Ewing’s sarcoma from 1991 to 2010. Patient, tumor, and socioeconomic variables were analyzed to determine prognostic factors for survival. Results: There were 1163 patients with Ewing’s sarcoma identified in the SEER Program database. The 10-year cause-specific survival for patients with non-metastatic disease at diagnosis was 66.8% and 28.1% for patients with metastatic disease. Black patients demonstrated reduced survival at 10 years with an increased frequency of metastatic disease at diagnosis as compared to patients of other race, while Hispanic patients more frequently presented with tumor size > 10 cm. Univariate analysis revealed that metastatic disease at presentation, tumor size > 10 cm, axial tumor location, patient age 20 years, black race, and male sex were associated with decreased cause-specific survival at 10 years. Metastatic disease at presentation, axial tumor location, tumor size > 10 cm, and age 20 years remained significant in the multivariate analysis. Conclusions: Patients with Ewing’s sarcoma have decreased cause-specific survival at 10 years when metastatic at presentation, axial tumor location, tumor size > 10 cm, and patient age 20 years. ß 2015 Elsevier Ltd. All rights reserved.
Keywords: Bone neoplasms/mortality Sarcoma Ewing/mortality Risk factors Survival analysis SEER program: survival rate
1. Introduction Ewing’s sarcoma is the second most common primary osseous sarcoma in children and adolescents [1]. Treatment of Ewing’s sarcoma is multimodal, utilizing a combination of surgery, radiation, and multidrug chemotherapy. Prior to implementation of multidrug chemotherapeutic regimens in the 1970s [2], 5-year survival for patients diagnosed with Ewing’s sarcoma was less than 25% [3,4]. Today, with intensified chemotherapeutic regimens, 5year overall survival exceeds 60% with reported recurrence rates ranging from 15 to 30% in patients with localized disease [5– 16]. Metastatic disease at presentation, particularly in the setting of bony metastases [17], continues to portend a grim prognosis [8,12,13,15,18,19].
* Corresponding author. Tel.: +1 319 385 5535; fax: +1 319 353 6754. E-mail addresses: [email protected] (K.R. Duchman), [email protected] (Y. Gao), [email protected], [email protected] (B.J. Miller).
Several studies have previously described prognostic variables in patients with Ewing’s sarcoma. Previous reports have identified increased patient age and male sex as well as metastatic disease at presentation, large tumor volume, axial tumor location, poor response to chemotherapy, and distant tumor recurrence as independent risk factors for decreased disease-free survival [5,7,9– 12,17]. Additionally, there is evidence that patient race and low socioeconomic status may negatively influence survival, suggesting that disparities may exist in the treatment of patients with Ewing’s sarcoma [12]. The Surveillance, Epidemiology, and End Results (SEER) Program database provides data from seventeen established cancer registries from across the United States, serving as a particularly useful tool for analyzing rare cancers. Utilizing the SEER Program database, the purpose of the present study was to describe cause-specific survival in patients with primary osseous Ewing’s sarcoma out to 10 years after diagnosis. Additionally, we aimed to identify patient, tumor, and socioeconomic variables associated with cause-specific survival in order to determine prognostic variables for survival in patients with Ewing’s sarcoma.
http://dx.doi.org/10.1016/j.canep.2014.12.012 1877-7821/ß 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 2
K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
Patients with Ewing’s sarcoma listed within the SEER Program database from 1991 to 2010 were identified. This time frame was selected as both therapeutic and imaging modalities, namely modern chemotherapeutic regimens and magnetic resonance imaging, were consistently employed and available. The SEER Program database is publicly available, and cases and associated variables of interest were identified using the SEER*Stat application (version 8.1.5).
Associations between cause-specific survival and patient, tumor, and socioeconomic variables were determined using the log-rank test. Several multivariate models were created using Cox proportional hazards ratios to determine independent predictors of survival. Chi-square analysis was performed in order to determine univariate measures of association between categorical variables with respect to number and frequency at presentation. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).
2.1. Human subjects determination
2.4. Multivariate model selection
The current project was reviewed by the Human Subjects Office/Institutional Review Board (IRB). It was determined that the project described in the application does not meet the regulatory definition of human subjects research and does not require review by the IRB.
Model 1 included all variables of interest regardless of univariate measures of association and included 1071/1163 (92.1%) of patients with Ewing’s sarcoma. A second model, Model 2, was developed which utilized stepwise selection to include all variables with univariate p-values < 0.1 while retaining composite SES score as a clinically significant variable based on previous literature investigating both Ewing’s sarcoma and osteosarcoma [12,21]. After a preliminary analysis revealed a discrepancy in survival between black race and all other patients, a chi-square analysis was performed which determined that black race presented with metastatic disease more frequently compared to all other patients (48.4% vs. 31.6%, p = 0.049). Because of the association between black race and metastatic disease at presentation, we chose to eliminate metastatic disease from one of our multivariate models in order to further isolate the effect of race on survival. A third multivariate model, Model 3, was subsequently developed which included all categorical variables with univariate p-values < 0.1 as well as composite SES score while excluding metastatic disease at presentation.
2. Materials and methods
2.2. Data elements Patient variables of interest included age, sex, and race. Age is recorded within the SEER Program database as a categorical variable in five-year intervals ranging from 0 to 85 years. We elected to categorize patients into two age groups, 5–10 cm), large (>10 cm), or unknown. Several socioeconomic measures based on the patient’s county of residence are provided by the SEER Program database. Countylevel socioeconomic variables of interest included median family income, the percent of persons living below the poverty line, the percent of persons at least 25 years of age with less than a high school education, and rural or urban county setting. A composite socioeconomic status (SES) variable was created using this data as has been described previously [20–22]. The composite SES score was further categorized as low (composite SES 4), middle (composite SES 5–10), or high (composite SES 11). Cause-specific survival was the primary outcome of interest and was calculated using abstracted data from death certificates to determine cause of death, which are dichotomously coded as ‘‘cancer’’ or ‘‘other causes’’. Any death attributable to disease recurrence or metastatic disease was considered death attributable to ‘‘cancer’’. Overall, 1163 patients with Ewing’s sarcoma from 1991 to 2010 had sufficient data for survival analysis.
There were 1163 patients with primary osseous Ewing’s sarcoma identified in the SEER Program database from 1991 to program database from 1991 to 2010. The vast majority of patients were other race (78.2%) and male (63.5%). Overall, 371/1163 (32.1%) patients presented with metastatic disease (Table 1). Metastatic disease was noted to present with increased frequency in patients of black race (48.4% vs. 31.6%, p = 0.049), while tumors > 10 cm were more frequent in patients of Hispanic race (57.0 vs. 35.6%, p < 0.001). Cause-specific survival for patients with non-metastatic disease at presentation was 86.5, 72.3, and 66.8% at 2, 5, and 10 years, respectively. For patients who presented with metastatic disease, cause-specific survival was 54.7, 35.7, and 28.1% at 2, 5, and 10 years, respectively.
2.3. Statistical analysis
3.2. Univariate survival analysis
Using the Kaplan–Meier method, estimated cause-specific survival was calculated at 2, 5, and 10 years after diagnosis.
Analysis of Kaplan–Meier cause-specific survival curves with the log-rank test revealed that metastatic disease at presentation
2.5. Missing data Tumor size was the least consistently reported variable in the SEER Program database, with values provided for 646/1163 (55.5%) patients in the cohort. There was no difference in 10-year survival in patients with or without tumor size data available (p = 0.562). Patients without tumor size data available were classified as unknown for statistical analysis. Other missing data from the analysis included 92/1163 (7.9%) patients with a missing socioeconomic variable. 3. Results 3.1. Cause-specific survival
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
3
Table 1 Univariate 10-year survival estimates of patient, tumor, and socioeconomic characteristics for patients with Ewing’s sarcoma identified in the SEER Program database from 1991 to 2010. Variable Age (years) 5–10 cm >10 cm Unknown Composite SESd 4 5–10 11 Rural or urban Rural Urban a b c d e
Survival (95% CI)a
Dead/censoredb
782 (67.2) 381 (32.8)
60.0 (55.9–64.1) 43.5 (37.4–49.7)
245/537 168/213
738 (63.5) 425 (36.5)
53.2 (48.9–57.4) 57.6 (51.7–63.4)
278/460 135/290
909 (78.2) 31 (2.7) 223 (19.2)
55.6 (51.7–59.4) 38.4 (18.6–60.4) 54.0 (45.5–62.4)
324/585 14/17 75/148
550 (47.3) 613 (52.7)
61.6 (56.7–66.5) 48.8 (44.0–53.6)
165/385 248/365
373 (32.1) 790 (67.9)
28.1 (22.4–34.0) 66.8 (62.8–70.7)
212/161 201/585
142 280 283 458
66.8 60.3 41.7 54.9
(56.7–76.1) (52.9–67.4) (34.1–49.5) (49.7–59.9)
34/108 81/199 121/162 177/281
286 (26.7) 513 (47.9) 272 (25.4)
55.4 (48.5–62.1) 53.1 (47.9–58.3) 56.4 (49.2–63.5)
103/183 188/325 93/179
131 (11.3) 1032 (88.7)
61.4 (51.2–71.2) 54.0 (50.3–57.7)
40/91 373/659
Number (%)
p-Valuec 10 cm in size at presentation. Several of these findings warrant further discussion. Metastatic disease at presentation results in a particularly grave prognosis in patients with Ewing’s sarcoma. Previous studies report 5-year survival estimates of approximately 30% in the era of multidrug chemotherapeutic regimens [11,13,17–19]. Patients with metastatic disease isolated to the lung tend to have improved survival as compared to patients with involvement of other skeletal sites [17]. Esiashvili et al. [8] reported on the changes in incidence and survival in patients with Ewing’s sarcoma over the last three decades using the SEER Program database. According to their report, survival for patients with metastatic disease has improved over the last three decades, with 39% 5-year survival estimates reported for the most recent decade. Our findings are consistent with both the incidence and 5-year survival estimates of previous reports, while also identifying metastatic disease at presentation as the independent risk factor that portends the poorest prognosis in patients with Ewing’s sarcoma. Increasing tumor size has also been consistently reported as a poor prognostic indicator throughout the literature. Using the California Cancer Registry database, Lee et al. identified tumor size 8 cm as an independent risk factor for decreased survival [12]. Utilizing the SEER Program database, Jawad et al. noted
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 4
K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
Fig. 1. Kaplan–Meier cause-specific survival curves comparing (a) metastatic versus non-metastatic disease, (b) patient age, (c) axial versus extremity tumor location, (d) tumor size, (e) patient race, and (f) male versus female sex. Significance was determined using the log-rank test, with p-values < 0.05 considered statistically significant.
similar findings in tumors measuring >8 cm [11]. While large tumor size may be frequently associated with metastatic disease at presentation and axial tumor location, both of which have been shown to be risk factors for decreased survival, our multivariate analysis suggests that large (>10 cm) and unknown tumor size confers a poor prognosis, independent of metastatic disease and tumor location. The overwhelming majority of patients diagnosed with Ewing’s sarcoma are children or adolescents [12]. Gupta et al. [23] compared survival in pediatric and adult patients diagnosed with local disease and noted significantly decreased overall survival at 3 years in adult
patients. The authors partially attributed the poorer prognosis in adults to lower doses of alkylating chemotherapeutic agents. Verrill et al. [24] described the use of full dose pediatric chemotherapeutic protocols in adult patients with moderate success. Utilizing the SEER Program database, Karski et al. [25] identified patient age 40 years as an independent risk factor for decreased survival. It should be noted that the SEER Program database does not provide data on the use, type, or intensity of chemotherapy used in the treatment of Ewing’s sarcoma. Our results are consistent with previous reports that identify increased patient age as an independent risk factor for decreased survival in patients with Ewing’s sarcoma.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
5
Table 2 Multivariate analysis of patient, tumor, and socioeconomic variables at 10 years in patients with Ewing’s sarcoma identified in the SEER Program database from 1991 to 2010a. Variable
Model 1b
Model 2c
Model 3d
n included in model Age (years) 5–10 cm >10 cm Unknown
Ref 1.119 (0.737–1.698) 2.013 (1.352–2.999) 1.679 (1.149–2.454)
Ref 1.119 (0.737–1.698) 2.010 (1.349–2.994) 1.682 (1.151–2.459)
Ref 1.259 (0.830–1.910) 2.269 (1.522–3.382) 1.885 (1.290–2.754)
Composite SESe 11 4 5–10
Ref 1.045 (0.773–1.413) 1.165 (0.904–1.500)
Ref 0.989 (0.738–1.326) 1.131 (0.880–1.453)
Ref 1.057 (0.790–1.414) 1.130 (0.880–1.452)
Rural or urban Urban Rural
Ref 0.779 (0.546–1.110)
a b c d e
Calculated using Cox proportional hazards model. Values listed as hazard ratios and 95% confidence intervals. Includes all categorical variables. Includes categorical variables with univariate p-values < 0.1 and composite SES. Includes categorical variables with univariate p-values < 0.1 and composite SES while eliminating metastatic disease as a variable. Composite socioeconomic status.
Ewing’s sarcoma frequently presents as a pelvic mass, which is often associated with increased tumor size at the time of diagnosis [11,12]. The literature remains relatively inconclusive with respect to the independent effect axial tumor location has on survival. Cotterill et al. [17] identified axial tumor location in patients with non-metastatic disease as an independent risk factor for decreased disease-free survival. Similarly, using the SEER Program database, Jawad et al. [11] identified axial tumor location as an independent risk factor for decreased overall survival as compared to primary extremity tumor location. On the other hand, Bacci et al. [6] did not identify axial tumor location as an independent risk factor for decreased survival in patients with non-metastatic Ewing’s sarcoma. Axial tumor location presents many challenges for the treating physician, as the anatomic constraints and neurovascular structures of the axial skeleton make excision with adequate surgical margins difficult. In an investigation of the role of surgical margins in the treatment of Ewing’s sarcoma, Bacci et al. [26] noted that surgical excision for tumors located within the axial skeleton was significantly less likely to achieve adequate surgical margins as compared to tumors located in the extremities. Axial tumors may also be noticed later by patients, potentially resulting in a delayed presentation and increased risk of disease spread prior to initiation of treatment. In the present study, we identified axial tumor location as an independent risk factor for decreased causespecific survival while controlling for tumor size and metastatic disease at presentation. Ewing’s sarcoma occurs with the highest frequency in patients of white race and is rare in blacks [8,11,12,27]. In a
study utilizing the SEER Program database, Jawad et al. [11] found that survival was not impacted by race while noting that Ewing’s sarcoma was nine times more likely to occur in patients with white race as compared to black race. Using the California Cancer Registry database, Lee et al. [12] described decreased survival in patients with Hispanic race, while also noting that Hispanic patients more frequently presented with tumors 8 cm in size. In the present study, black race was associated with an increased frequency of metastatic disease at diagnosis as compared to patients of all other races (48.4 vs. 31.6%, p = 0.049), while patients with Hispanic race presented more frequently with tumors > 10 cm in size compared to patients of all other races (57.0 vs. 35.6%, p < 0.001). Further analysis revealed that the 10-year cause-specific survival for patients with black race was decreased compared to patients of all other races (38.4 vs. 55.5%, p = 0.024). No survival difference was noted for patients with Hispanic race compared to patients of all other races. When excluding metastatic disease at diagnosis from the multivariate model, black race became an independent risk factor for decreased cause-specific survival at 10 years, but failed to become an independent risk factor for decreased causespecific survival following elimination of any other categorical variable. While the low incidence of Ewing’s sarcoma in Hispanic and black patients has limited comparisons, future research should focus on factors intrinsic to the tumor itself as well as variable access to health care resources for the disproportionate presentation with advanced disease, large tumor size, and discrepancy in survival.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
6
A patient’s environment and access to health care are both factors that can potentially be modified. Several studies have identified decreased socioeconomic status and/or limited access to health care as factors that can affect treatment selection, disease stage at presentation, and survival in patients with cancer [20– 22,28]. Using a composite socioeconomic status index based on census block-level data from the state of California, Lee et al. [12] identified the lowest composite socioeconomic status category as an independent risk factor for mortality in patients with Ewing’s sarcoma. The present study utilizes county-level data from the nationwide SEER Program database. Using this database, low socioeconomic status was not associated with decreased causespecific survival in patients with Ewing’s sarcoma. The effect of socioeconomic status on survival in patients with Ewing’s sarcoma is likely highly dependent on the heterogeneity of the geographic divisions for which data is provided as well as the inclusive elements of the composite socioeconomic status variable. It is possible that socioeconomic factors influence Ewing’s sarcoma, and we were unable to observe these differences using the recorded county-level data. This is an area that would benefit from further study. 4.1. Limitations The present study does have several limitations. The SEER Program database does not collect information regarding local recurrence or metastatic spread after initial diagnosis, both of which have been reported to occur while affecting prognosis in patients with Ewing’s sarcoma [17]. This limitation prevents us from reporting disease-free survival in addition to cause-specific survival. Cause-specific survival was calculated utilizing death certificates to determine cause of death, which can be inaccurate or difficult to interpret if multiple causes of death are listed. However, this method of survival analysis has been used previously throughout the literature [13,29–31]. Additionally, the SEER Program database does not collect information on the use of chemotherapy, which is frequently employed for patients with Ewing’s sarcoma, or specific chemotherapeutic agents. Age is reported as a categorical variable in five-year intervals within the SEER Program database, and an age of 20 years was arbitrarily chosen to compare children and adults with cancer. Certainly, other age cutoffs could yield varying results. While the SEER Program database provides data from 17 regional cancer registries that comprise 26% of the United States population, the regions included in the database tend to be more affluent compared with the nation as a whole [32,33]. Additionally, the geographic division of socioeconomic data provided by the SEER Program database is county-level data, which may not be as sensitive as census tracts or zip codes, particularly in heterogeneous counties [34,35]. Despite these shortcomings, the SEER Program database serves as an unparalleled resource when investigating rare cancers such as Ewing’s sarcoma. 5. Conclusions Primary osseous Ewing’s sarcoma is a rare neoplasm with the highest incidence in adolescent white males. The present study is the largest study during the modern era of multidrug chemotherapeutic regimens to report survival and prognostic factors for patients with Ewing’s sarcoma of bone. Survival analysis identified several poor prognostic factors, including metastatic disease at diagnosis, tumor size > 10 cm, unknown tumor size, patient age 20 years, and axial tumor location as independent risk factors for decreased cause-specific survival. Black patients were noted to have decreased survival while having an increased rate of metastatic disease at presentation, while Hispanic patients were
noted to present more frequently with large tumors as compared to patients of other races. Future studies should aim to identify both modifiable and non-modifiable risk factors that result in these noted discrepancies. Authorship contribution Kyle R. Duchman, MD: (1) Conception and design, acquisition of data, and analysis and interpretation of data; (2) drafting the article and revising it critically for important intellectual content; and (3) final approval of the version to be published. Yubo Gao, PhD: (1) Acquisition of data, and analysis and interpretation of data; (2) revising it critically for important intellectual content; (3) final approval of the version to be published. Benjamin J. Miller, MD, MS: (1) Conception and design, acquisition of data, and analysis and interpretation of data; (2) drafting the article and revising it critically for important intellectual content; and (3) final approval of the version to be published. Conflict of interest statement None to declare. No competing financial interests exist. Acknowledgement None. References [1] Cotterill S, Parker L, Malcolm A, Reid M, More L, Craft A. Incidence and survival for cancer in children and young adults in the North of England, 1968–1995: a report from the Northern Region Young Persons’ Malignant Disease Registry. Br J Cancer 2000;83:397. [2] Jaffe N, Paed D, Traggis D, Salian S, Cassady JR. Improved outlook for Ewing’s sarcoma with combination chemotherapy (vincristine, actinomycin D and cyclophosphamide) and radiation therapy. Cancer 1976;38:1925–30. [3] Ewing J. Diffuse endothelioma of bone. CA Cancer J Clin 1972;22:95–8. [4] Phillips RF, Higinbotham NL. The curability of Ewing’s endothelioma of bone in children. J Pediatr 1967;70:391–7. [5] Arpaci E, Yetisyigit T, Seker M, Uncu D, Uyeturk U, Oksuzoglu B, et al. Prognostic factors and clinical outcome of patients with Ewing’s sarcoma family of tumors in adults: multicentric study of the Anatolian Society of Medical Oncology. Med Oncol 2013;30:1–7. [6] Bacci G, Ferrari S, Bertoni F, Rimondini S, Longhi A, Bacchini P, et al. Prognostic factors in nonmetastatic Ewing’s sarcoma of bone treated with adjuvant chemotherapy: analysis of 359 patients at the Istituto Ortopedico Rizzoli. J Clin Oncol 2000;18:4–11. [7] Bacci G, Longhi A, Ferrari S, Mercuri M, Versari M, Bertoni F. Prognostic factors in non-metastatic Ewing’s sarcoma tumor of bone: an analysis of 579 patients treated at a single institution with adjuvant or neoadjuvant chemotherapy between 1972 and 1998. Acta Oncol 2006;45:469–75. [8] Esiashvili N, Goodman M, Marcus RB. Changes in incidence and survival of Ewing sarcoma patients over the past 3 decades: Surveillance Epidemiology and End Results data. J Pediatr Hematol Oncol 2008;30:425–30. [9] Gaspar N, Rey A, Be´rard PM, Michon J, Gentet JC, Tabone MD, et al. Risk adapted chemotherapy for localised Ewing’s sarcoma of bone: the French EW93 study. Eur J Cancer 2012;48:1376–85. [10] Lopez Guerra JL, Marquez-Vega C, Ramirez-Villar GL, Cabrera P, Ordonez R, Praena-Fernandez JM, et al. Prognostic factors for overall survival in paediatric patients with Ewing sarcoma of bone treated according to multidisciplinary protocol. Clin Transl Oncol 2012;14:294–301. [11] Jawad MU, Cheung MC, Min ES, Schneiderbauer MM, Koniaris LG, Scully SP. Ewing sarcoma demonstrates racial disparities in incidence-related and sexrelated differences in outcome. Cancer 2009;115:3526–36. [12] Lee J, Hoang BH, Ziogas A, Zell JA. Analysis of prognostic factors in Ewing sarcoma using a population-based cancer registry. Cancer 2010;116:1964–73. [13] Miller BJ, Lynch CF, Buckwalter JA. Conditional survival is greater than overall survival at diagnosis in patients with osteosarcoma and Ewing’s sarcoma. Clin Orthop Relat Res 2013;471:3398–404. [14] O’Connor MI, Pritchard DJ. Ewing’s sarcoma: prognostic factors, disease control, and the reemerging role of surgical treatment. Clin Orthop Relat Res 1991;262:78–87. [15] Wilkins RM, Pritchard DJ, Omer EB, Unni KK. Ewing’s sarcoma of bone. Experience with 140 patients. Cancer 1986;58:2551–5.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx [16] Paulussen M, Ahrens S, Dunst J, Winkelmann W, Exner G, Kotz R, et al. Localized Ewing tumor of bone: final results of the cooperative Ewing’s Sarcoma Study CESS 86. J Clin Oncol 2001;19:1818–29. [17] Cotterill S, Ahrens S, Paulussen M, Ju¨rgens H, Voute P, Gadner H, et al. Prognostic factors in Ewing’s tumor of bone: analysis of 975 patients from the European Intergroup Cooperative Ewing’s Sarcoma Study Group. J Clin Oncol 2000;18:3108–14. [18] Cangir A, Vietti TJ, Gehan EA, Burgert EO, Thomas P, Tefft M, et al. Ewing’s sarcoma metastatic at diagnosis results and comparisons of two intergroup Ewing’s sarcoma studies. Cancer 1990;66:887–93. [19] Paulussen M, Ahrens S, Burdach S, Craft A, Dockhorn-Dworniczak B, Dunst J, et al. Primary metastatic (stage IV) Ewing tumor: survival analysis of 171 patients from the EICESS studies. Ann Oncol 1998;9:275–81. [20] Du XL, Fang S, Coker AL, Sanderson M, Aragaki C, Cormier JN, et al. Racial disparity and socioeconomic status in association with survival in older men with local/regional stage prostate carcinoma. Cancer 2006;106:1276–85. [21] Miller BJ, Cram P, Lynch CF, Buckwalter JA. Risk factors for metastatic disease at presentation with osteosarcoma: an analysis of the SEER database. J Bone Joint Surg Am 2013;95:e891–8. [22] Sun M, Abdollah F, Liberman D, Abdo A, Thuret R, Tian Z, et al. Racial disparities and socioeconomic status in men diagnosed with testicular germ cell tumors. Cancer 2011;117:4277–85. [23] Gupta AA, Pappo A, Saunders N, Hopyan S, Ferguson P, Wunder J, et al. Clinical outcome of children and adults with localized Ewing sarcoma. Cancer 2010;116:3189–94. [24] Verrill M, Judson I, Wiltshaw E, Thomas J, Harmer C, Fisher C. The use of paediatric chemotherapy protocols at full dose is both a rational and feasible treatment strategy in adults with Ewing’s family tumours. Ann Oncol 1997;8:1099–105. [25] Karski EE, Matthay KK, Neuhaus JM, Goldsby RE, DuBois SG. Characteristics and outcomes of patients with Ewing sarcoma over 40 years of age at diagnosis. Cancer Epidemiol 2013;37:29–33.
7
[26] Bacci G, Longhi A, Briccoli A, Bertoni F, Versari M, Picci P. The role of surgical margins in treatment of Ewing’s sarcoma family tumors: experience of a single institution with 512 patients treated with adjuvant and neoadjuvant chemotherapy. Int J Radiat Oncol Biol Phys 2006;65:766–72. [27] Polednak AP. Primary bone cancer incidence in black and white residents of New York State. Cancer 1985;55:2883–8. [28] Crawford SM, Sauerzapf V, Haynes R, Zhao H, Forman D, Jones AP. Social and geographical factors affecting access to treatment of lung cancer. Br J Cancer 2009;101:897–901. [29] Duchman KR, Lynch CF, Buckwalter JA, Miller BJ. Estimated cause-specific survival continues to improve over time in patients with chondrosarcoma. Clin Orthop Relat Res 2014;472(8):2516–25. [30] Giuffrida AY, Burgueno JE, Koniaris LG, Gutierrez JC, Duncan R, Scully SP. Chondrosarcoma in the United States (1973 to 2003): an analysis of 2890 cases from the SEER database. J Bone Joint Surg Am 2009;91:1063–72. [31] Johnson DR, Ma DJ, Buckner JC, Hammack JE. Conditional probability of longterm survival in glioblastoma. Cancer 2012;118:5608–13. [32] Bach PB, Guadagnoli E, Schrag D, Schussler N, Warren JL. Patient demographic and socioeconomic characteristics in the SEER-Medicare database: applications and limitations. Med Care 2002;40:IV-19–25. [33] Nattinger AB, McAuliffe TL, Schapira MM. Generalizability of the surveillance, epidemiology, and end results registry population: factors relevant to epidemiologic and health care research. J Clin Epidemiol 1997;50:939–45. [34] Krieger N, Chen JT, Waterman PD, Soobader M-J, Subramanian S, Carson R. Geocoding and monitoring of US socioeconomic inequalities in mortality and cancer incidence: does the choice of area-based measure and geographic level matter? the Public Health Disparities Geocoding Project. Am J Epidemiol 2002;156:471–82. [35] Subramanian S, Chen J, Rehkopf D, Waterman P, Krieger N. Comparing individual-and area-based socioeconomic measures for the surveillance of health disparities: a multilevel analysis of Massachusetts births, 1989–1991. Am J Epidemiol 2006;164:823–34.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
CANEP-797; No. of Pages 7 Cancer Epidemiology xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention journal homepage: www.cancerepidemiology.net
Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database Kyle R. Duchman, Yubo Gao, Benjamin J. Miller * Department of Orthopaedics and Rehabilitation, University of Iowa, 200 Hawkins Drive, 01015 JPP, Iowa City, IA 52242, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 September 2014 Received in revised form 22 December 2014 Accepted 26 December 2014 Available online xxx
Background: The current study aims to determine cause-specific survival in patients with Ewing’s sarcoma while reporting clinical risk factors for survival. Methods: The Surveillance, Epidemiology, and End Results (SEER) Program database was used to identify patients with osseous Ewing’s sarcoma from 1991 to 2010. Patient, tumor, and socioeconomic variables were analyzed to determine prognostic factors for survival. Results: There were 1163 patients with Ewing’s sarcoma identified in the SEER Program database. The 10-year cause-specific survival for patients with non-metastatic disease at diagnosis was 66.8% and 28.1% for patients with metastatic disease. Black patients demonstrated reduced survival at 10 years with an increased frequency of metastatic disease at diagnosis as compared to patients of other race, while Hispanic patients more frequently presented with tumor size > 10 cm. Univariate analysis revealed that metastatic disease at presentation, tumor size > 10 cm, axial tumor location, patient age 20 years, black race, and male sex were associated with decreased cause-specific survival at 10 years. Metastatic disease at presentation, axial tumor location, tumor size > 10 cm, and age 20 years remained significant in the multivariate analysis. Conclusions: Patients with Ewing’s sarcoma have decreased cause-specific survival at 10 years when metastatic at presentation, axial tumor location, tumor size > 10 cm, and patient age 20 years. ß 2015 Elsevier Ltd. All rights reserved.
Keywords: Bone neoplasms/mortality Sarcoma Ewing/mortality Risk factors Survival analysis SEER program: survival rate
1. Introduction Ewing’s sarcoma is the second most common primary osseous sarcoma in children and adolescents [1]. Treatment of Ewing’s sarcoma is multimodal, utilizing a combination of surgery, radiation, and multidrug chemotherapy. Prior to implementation of multidrug chemotherapeutic regimens in the 1970s [2], 5-year survival for patients diagnosed with Ewing’s sarcoma was less than 25% [3,4]. Today, with intensified chemotherapeutic regimens, 5year overall survival exceeds 60% with reported recurrence rates ranging from 15 to 30% in patients with localized disease [5– 16]. Metastatic disease at presentation, particularly in the setting of bony metastases [17], continues to portend a grim prognosis [8,12,13,15,18,19].
* Corresponding author. Tel.: +1 319 385 5535; fax: +1 319 353 6754. E-mail addresses: [email protected] (K.R. Duchman), [email protected] (Y. Gao), [email protected], [email protected] (B.J. Miller).
Several studies have previously described prognostic variables in patients with Ewing’s sarcoma. Previous reports have identified increased patient age and male sex as well as metastatic disease at presentation, large tumor volume, axial tumor location, poor response to chemotherapy, and distant tumor recurrence as independent risk factors for decreased disease-free survival [5,7,9– 12,17]. Additionally, there is evidence that patient race and low socioeconomic status may negatively influence survival, suggesting that disparities may exist in the treatment of patients with Ewing’s sarcoma [12]. The Surveillance, Epidemiology, and End Results (SEER) Program database provides data from seventeen established cancer registries from across the United States, serving as a particularly useful tool for analyzing rare cancers. Utilizing the SEER Program database, the purpose of the present study was to describe cause-specific survival in patients with primary osseous Ewing’s sarcoma out to 10 years after diagnosis. Additionally, we aimed to identify patient, tumor, and socioeconomic variables associated with cause-specific survival in order to determine prognostic variables for survival in patients with Ewing’s sarcoma.
http://dx.doi.org/10.1016/j.canep.2014.12.012 1877-7821/ß 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 2
K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
Patients with Ewing’s sarcoma listed within the SEER Program database from 1991 to 2010 were identified. This time frame was selected as both therapeutic and imaging modalities, namely modern chemotherapeutic regimens and magnetic resonance imaging, were consistently employed and available. The SEER Program database is publicly available, and cases and associated variables of interest were identified using the SEER*Stat application (version 8.1.5).
Associations between cause-specific survival and patient, tumor, and socioeconomic variables were determined using the log-rank test. Several multivariate models were created using Cox proportional hazards ratios to determine independent predictors of survival. Chi-square analysis was performed in order to determine univariate measures of association between categorical variables with respect to number and frequency at presentation. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).
2.1. Human subjects determination
2.4. Multivariate model selection
The current project was reviewed by the Human Subjects Office/Institutional Review Board (IRB). It was determined that the project described in the application does not meet the regulatory definition of human subjects research and does not require review by the IRB.
Model 1 included all variables of interest regardless of univariate measures of association and included 1071/1163 (92.1%) of patients with Ewing’s sarcoma. A second model, Model 2, was developed which utilized stepwise selection to include all variables with univariate p-values < 0.1 while retaining composite SES score as a clinically significant variable based on previous literature investigating both Ewing’s sarcoma and osteosarcoma [12,21]. After a preliminary analysis revealed a discrepancy in survival between black race and all other patients, a chi-square analysis was performed which determined that black race presented with metastatic disease more frequently compared to all other patients (48.4% vs. 31.6%, p = 0.049). Because of the association between black race and metastatic disease at presentation, we chose to eliminate metastatic disease from one of our multivariate models in order to further isolate the effect of race on survival. A third multivariate model, Model 3, was subsequently developed which included all categorical variables with univariate p-values < 0.1 as well as composite SES score while excluding metastatic disease at presentation.
2. Materials and methods
2.2. Data elements Patient variables of interest included age, sex, and race. Age is recorded within the SEER Program database as a categorical variable in five-year intervals ranging from 0 to 85 years. We elected to categorize patients into two age groups, 5–10 cm), large (>10 cm), or unknown. Several socioeconomic measures based on the patient’s county of residence are provided by the SEER Program database. Countylevel socioeconomic variables of interest included median family income, the percent of persons living below the poverty line, the percent of persons at least 25 years of age with less than a high school education, and rural or urban county setting. A composite socioeconomic status (SES) variable was created using this data as has been described previously [20–22]. The composite SES score was further categorized as low (composite SES 4), middle (composite SES 5–10), or high (composite SES 11). Cause-specific survival was the primary outcome of interest and was calculated using abstracted data from death certificates to determine cause of death, which are dichotomously coded as ‘‘cancer’’ or ‘‘other causes’’. Any death attributable to disease recurrence or metastatic disease was considered death attributable to ‘‘cancer’’. Overall, 1163 patients with Ewing’s sarcoma from 1991 to 2010 had sufficient data for survival analysis.
There were 1163 patients with primary osseous Ewing’s sarcoma identified in the SEER Program database from 1991 to program database from 1991 to 2010. The vast majority of patients were other race (78.2%) and male (63.5%). Overall, 371/1163 (32.1%) patients presented with metastatic disease (Table 1). Metastatic disease was noted to present with increased frequency in patients of black race (48.4% vs. 31.6%, p = 0.049), while tumors > 10 cm were more frequent in patients of Hispanic race (57.0 vs. 35.6%, p < 0.001). Cause-specific survival for patients with non-metastatic disease at presentation was 86.5, 72.3, and 66.8% at 2, 5, and 10 years, respectively. For patients who presented with metastatic disease, cause-specific survival was 54.7, 35.7, and 28.1% at 2, 5, and 10 years, respectively.
2.3. Statistical analysis
3.2. Univariate survival analysis
Using the Kaplan–Meier method, estimated cause-specific survival was calculated at 2, 5, and 10 years after diagnosis.
Analysis of Kaplan–Meier cause-specific survival curves with the log-rank test revealed that metastatic disease at presentation
2.5. Missing data Tumor size was the least consistently reported variable in the SEER Program database, with values provided for 646/1163 (55.5%) patients in the cohort. There was no difference in 10-year survival in patients with or without tumor size data available (p = 0.562). Patients without tumor size data available were classified as unknown for statistical analysis. Other missing data from the analysis included 92/1163 (7.9%) patients with a missing socioeconomic variable. 3. Results 3.1. Cause-specific survival
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
3
Table 1 Univariate 10-year survival estimates of patient, tumor, and socioeconomic characteristics for patients with Ewing’s sarcoma identified in the SEER Program database from 1991 to 2010. Variable Age (years) 5–10 cm >10 cm Unknown Composite SESd 4 5–10 11 Rural or urban Rural Urban a b c d e
Survival (95% CI)a
Dead/censoredb
782 (67.2) 381 (32.8)
60.0 (55.9–64.1) 43.5 (37.4–49.7)
245/537 168/213
738 (63.5) 425 (36.5)
53.2 (48.9–57.4) 57.6 (51.7–63.4)
278/460 135/290
909 (78.2) 31 (2.7) 223 (19.2)
55.6 (51.7–59.4) 38.4 (18.6–60.4) 54.0 (45.5–62.4)
324/585 14/17 75/148
550 (47.3) 613 (52.7)
61.6 (56.7–66.5) 48.8 (44.0–53.6)
165/385 248/365
373 (32.1) 790 (67.9)
28.1 (22.4–34.0) 66.8 (62.8–70.7)
212/161 201/585
142 280 283 458
66.8 60.3 41.7 54.9
(56.7–76.1) (52.9–67.4) (34.1–49.5) (49.7–59.9)
34/108 81/199 121/162 177/281
286 (26.7) 513 (47.9) 272 (25.4)
55.4 (48.5–62.1) 53.1 (47.9–58.3) 56.4 (49.2–63.5)
103/183 188/325 93/179
131 (11.3) 1032 (88.7)
61.4 (51.2–71.2) 54.0 (50.3–57.7)
40/91 373/659
Number (%)
p-Valuec 10 cm in size at presentation. Several of these findings warrant further discussion. Metastatic disease at presentation results in a particularly grave prognosis in patients with Ewing’s sarcoma. Previous studies report 5-year survival estimates of approximately 30% in the era of multidrug chemotherapeutic regimens [11,13,17–19]. Patients with metastatic disease isolated to the lung tend to have improved survival as compared to patients with involvement of other skeletal sites [17]. Esiashvili et al. [8] reported on the changes in incidence and survival in patients with Ewing’s sarcoma over the last three decades using the SEER Program database. According to their report, survival for patients with metastatic disease has improved over the last three decades, with 39% 5-year survival estimates reported for the most recent decade. Our findings are consistent with both the incidence and 5-year survival estimates of previous reports, while also identifying metastatic disease at presentation as the independent risk factor that portends the poorest prognosis in patients with Ewing’s sarcoma. Increasing tumor size has also been consistently reported as a poor prognostic indicator throughout the literature. Using the California Cancer Registry database, Lee et al. identified tumor size 8 cm as an independent risk factor for decreased survival [12]. Utilizing the SEER Program database, Jawad et al. noted
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 4
K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
Fig. 1. Kaplan–Meier cause-specific survival curves comparing (a) metastatic versus non-metastatic disease, (b) patient age, (c) axial versus extremity tumor location, (d) tumor size, (e) patient race, and (f) male versus female sex. Significance was determined using the log-rank test, with p-values < 0.05 considered statistically significant.
similar findings in tumors measuring >8 cm [11]. While large tumor size may be frequently associated with metastatic disease at presentation and axial tumor location, both of which have been shown to be risk factors for decreased survival, our multivariate analysis suggests that large (>10 cm) and unknown tumor size confers a poor prognosis, independent of metastatic disease and tumor location. The overwhelming majority of patients diagnosed with Ewing’s sarcoma are children or adolescents [12]. Gupta et al. [23] compared survival in pediatric and adult patients diagnosed with local disease and noted significantly decreased overall survival at 3 years in adult
patients. The authors partially attributed the poorer prognosis in adults to lower doses of alkylating chemotherapeutic agents. Verrill et al. [24] described the use of full dose pediatric chemotherapeutic protocols in adult patients with moderate success. Utilizing the SEER Program database, Karski et al. [25] identified patient age 40 years as an independent risk factor for decreased survival. It should be noted that the SEER Program database does not provide data on the use, type, or intensity of chemotherapy used in the treatment of Ewing’s sarcoma. Our results are consistent with previous reports that identify increased patient age as an independent risk factor for decreased survival in patients with Ewing’s sarcoma.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
5
Table 2 Multivariate analysis of patient, tumor, and socioeconomic variables at 10 years in patients with Ewing’s sarcoma identified in the SEER Program database from 1991 to 2010a. Variable
Model 1b
Model 2c
Model 3d
n included in model Age (years) 5–10 cm >10 cm Unknown
Ref 1.119 (0.737–1.698) 2.013 (1.352–2.999) 1.679 (1.149–2.454)
Ref 1.119 (0.737–1.698) 2.010 (1.349–2.994) 1.682 (1.151–2.459)
Ref 1.259 (0.830–1.910) 2.269 (1.522–3.382) 1.885 (1.290–2.754)
Composite SESe 11 4 5–10
Ref 1.045 (0.773–1.413) 1.165 (0.904–1.500)
Ref 0.989 (0.738–1.326) 1.131 (0.880–1.453)
Ref 1.057 (0.790–1.414) 1.130 (0.880–1.452)
Rural or urban Urban Rural
Ref 0.779 (0.546–1.110)
a b c d e
Calculated using Cox proportional hazards model. Values listed as hazard ratios and 95% confidence intervals. Includes all categorical variables. Includes categorical variables with univariate p-values < 0.1 and composite SES. Includes categorical variables with univariate p-values < 0.1 and composite SES while eliminating metastatic disease as a variable. Composite socioeconomic status.
Ewing’s sarcoma frequently presents as a pelvic mass, which is often associated with increased tumor size at the time of diagnosis [11,12]. The literature remains relatively inconclusive with respect to the independent effect axial tumor location has on survival. Cotterill et al. [17] identified axial tumor location in patients with non-metastatic disease as an independent risk factor for decreased disease-free survival. Similarly, using the SEER Program database, Jawad et al. [11] identified axial tumor location as an independent risk factor for decreased overall survival as compared to primary extremity tumor location. On the other hand, Bacci et al. [6] did not identify axial tumor location as an independent risk factor for decreased survival in patients with non-metastatic Ewing’s sarcoma. Axial tumor location presents many challenges for the treating physician, as the anatomic constraints and neurovascular structures of the axial skeleton make excision with adequate surgical margins difficult. In an investigation of the role of surgical margins in the treatment of Ewing’s sarcoma, Bacci et al. [26] noted that surgical excision for tumors located within the axial skeleton was significantly less likely to achieve adequate surgical margins as compared to tumors located in the extremities. Axial tumors may also be noticed later by patients, potentially resulting in a delayed presentation and increased risk of disease spread prior to initiation of treatment. In the present study, we identified axial tumor location as an independent risk factor for decreased causespecific survival while controlling for tumor size and metastatic disease at presentation. Ewing’s sarcoma occurs with the highest frequency in patients of white race and is rare in blacks [8,11,12,27]. In a
study utilizing the SEER Program database, Jawad et al. [11] found that survival was not impacted by race while noting that Ewing’s sarcoma was nine times more likely to occur in patients with white race as compared to black race. Using the California Cancer Registry database, Lee et al. [12] described decreased survival in patients with Hispanic race, while also noting that Hispanic patients more frequently presented with tumors 8 cm in size. In the present study, black race was associated with an increased frequency of metastatic disease at diagnosis as compared to patients of all other races (48.4 vs. 31.6%, p = 0.049), while patients with Hispanic race presented more frequently with tumors > 10 cm in size compared to patients of all other races (57.0 vs. 35.6%, p < 0.001). Further analysis revealed that the 10-year cause-specific survival for patients with black race was decreased compared to patients of all other races (38.4 vs. 55.5%, p = 0.024). No survival difference was noted for patients with Hispanic race compared to patients of all other races. When excluding metastatic disease at diagnosis from the multivariate model, black race became an independent risk factor for decreased cause-specific survival at 10 years, but failed to become an independent risk factor for decreased causespecific survival following elimination of any other categorical variable. While the low incidence of Ewing’s sarcoma in Hispanic and black patients has limited comparisons, future research should focus on factors intrinsic to the tumor itself as well as variable access to health care resources for the disproportionate presentation with advanced disease, large tumor size, and discrepancy in survival.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx
6
A patient’s environment and access to health care are both factors that can potentially be modified. Several studies have identified decreased socioeconomic status and/or limited access to health care as factors that can affect treatment selection, disease stage at presentation, and survival in patients with cancer [20– 22,28]. Using a composite socioeconomic status index based on census block-level data from the state of California, Lee et al. [12] identified the lowest composite socioeconomic status category as an independent risk factor for mortality in patients with Ewing’s sarcoma. The present study utilizes county-level data from the nationwide SEER Program database. Using this database, low socioeconomic status was not associated with decreased causespecific survival in patients with Ewing’s sarcoma. The effect of socioeconomic status on survival in patients with Ewing’s sarcoma is likely highly dependent on the heterogeneity of the geographic divisions for which data is provided as well as the inclusive elements of the composite socioeconomic status variable. It is possible that socioeconomic factors influence Ewing’s sarcoma, and we were unable to observe these differences using the recorded county-level data. This is an area that would benefit from further study. 4.1. Limitations The present study does have several limitations. The SEER Program database does not collect information regarding local recurrence or metastatic spread after initial diagnosis, both of which have been reported to occur while affecting prognosis in patients with Ewing’s sarcoma [17]. This limitation prevents us from reporting disease-free survival in addition to cause-specific survival. Cause-specific survival was calculated utilizing death certificates to determine cause of death, which can be inaccurate or difficult to interpret if multiple causes of death are listed. However, this method of survival analysis has been used previously throughout the literature [13,29–31]. Additionally, the SEER Program database does not collect information on the use of chemotherapy, which is frequently employed for patients with Ewing’s sarcoma, or specific chemotherapeutic agents. Age is reported as a categorical variable in five-year intervals within the SEER Program database, and an age of 20 years was arbitrarily chosen to compare children and adults with cancer. Certainly, other age cutoffs could yield varying results. While the SEER Program database provides data from 17 regional cancer registries that comprise 26% of the United States population, the regions included in the database tend to be more affluent compared with the nation as a whole [32,33]. Additionally, the geographic division of socioeconomic data provided by the SEER Program database is county-level data, which may not be as sensitive as census tracts or zip codes, particularly in heterogeneous counties [34,35]. Despite these shortcomings, the SEER Program database serves as an unparalleled resource when investigating rare cancers such as Ewing’s sarcoma. 5. Conclusions Primary osseous Ewing’s sarcoma is a rare neoplasm with the highest incidence in adolescent white males. The present study is the largest study during the modern era of multidrug chemotherapeutic regimens to report survival and prognostic factors for patients with Ewing’s sarcoma of bone. Survival analysis identified several poor prognostic factors, including metastatic disease at diagnosis, tumor size > 10 cm, unknown tumor size, patient age 20 years, and axial tumor location as independent risk factors for decreased cause-specific survival. Black patients were noted to have decreased survival while having an increased rate of metastatic disease at presentation, while Hispanic patients were
noted to present more frequently with large tumors as compared to patients of other races. Future studies should aim to identify both modifiable and non-modifiable risk factors that result in these noted discrepancies. Authorship contribution Kyle R. Duchman, MD: (1) Conception and design, acquisition of data, and analysis and interpretation of data; (2) drafting the article and revising it critically for important intellectual content; and (3) final approval of the version to be published. Yubo Gao, PhD: (1) Acquisition of data, and analysis and interpretation of data; (2) revising it critically for important intellectual content; (3) final approval of the version to be published. Benjamin J. Miller, MD, MS: (1) Conception and design, acquisition of data, and analysis and interpretation of data; (2) drafting the article and revising it critically for important intellectual content; and (3) final approval of the version to be published. Conflict of interest statement None to declare. No competing financial interests exist. Acknowledgement None. References [1] Cotterill S, Parker L, Malcolm A, Reid M, More L, Craft A. Incidence and survival for cancer in children and young adults in the North of England, 1968–1995: a report from the Northern Region Young Persons’ Malignant Disease Registry. Br J Cancer 2000;83:397. [2] Jaffe N, Paed D, Traggis D, Salian S, Cassady JR. Improved outlook for Ewing’s sarcoma with combination chemotherapy (vincristine, actinomycin D and cyclophosphamide) and radiation therapy. Cancer 1976;38:1925–30. [3] Ewing J. Diffuse endothelioma of bone. CA Cancer J Clin 1972;22:95–8. [4] Phillips RF, Higinbotham NL. The curability of Ewing’s endothelioma of bone in children. J Pediatr 1967;70:391–7. [5] Arpaci E, Yetisyigit T, Seker M, Uncu D, Uyeturk U, Oksuzoglu B, et al. Prognostic factors and clinical outcome of patients with Ewing’s sarcoma family of tumors in adults: multicentric study of the Anatolian Society of Medical Oncology. Med Oncol 2013;30:1–7. [6] Bacci G, Ferrari S, Bertoni F, Rimondini S, Longhi A, Bacchini P, et al. Prognostic factors in nonmetastatic Ewing’s sarcoma of bone treated with adjuvant chemotherapy: analysis of 359 patients at the Istituto Ortopedico Rizzoli. J Clin Oncol 2000;18:4–11. [7] Bacci G, Longhi A, Ferrari S, Mercuri M, Versari M, Bertoni F. Prognostic factors in non-metastatic Ewing’s sarcoma tumor of bone: an analysis of 579 patients treated at a single institution with adjuvant or neoadjuvant chemotherapy between 1972 and 1998. Acta Oncol 2006;45:469–75. [8] Esiashvili N, Goodman M, Marcus RB. Changes in incidence and survival of Ewing sarcoma patients over the past 3 decades: Surveillance Epidemiology and End Results data. J Pediatr Hematol Oncol 2008;30:425–30. [9] Gaspar N, Rey A, Be´rard PM, Michon J, Gentet JC, Tabone MD, et al. Risk adapted chemotherapy for localised Ewing’s sarcoma of bone: the French EW93 study. Eur J Cancer 2012;48:1376–85. [10] Lopez Guerra JL, Marquez-Vega C, Ramirez-Villar GL, Cabrera P, Ordonez R, Praena-Fernandez JM, et al. Prognostic factors for overall survival in paediatric patients with Ewing sarcoma of bone treated according to multidisciplinary protocol. Clin Transl Oncol 2012;14:294–301. [11] Jawad MU, Cheung MC, Min ES, Schneiderbauer MM, Koniaris LG, Scully SP. Ewing sarcoma demonstrates racial disparities in incidence-related and sexrelated differences in outcome. Cancer 2009;115:3526–36. [12] Lee J, Hoang BH, Ziogas A, Zell JA. Analysis of prognostic factors in Ewing sarcoma using a population-based cancer registry. Cancer 2010;116:1964–73. [13] Miller BJ, Lynch CF, Buckwalter JA. Conditional survival is greater than overall survival at diagnosis in patients with osteosarcoma and Ewing’s sarcoma. Clin Orthop Relat Res 2013;471:3398–404. [14] O’Connor MI, Pritchard DJ. Ewing’s sarcoma: prognostic factors, disease control, and the reemerging role of surgical treatment. Clin Orthop Relat Res 1991;262:78–87. [15] Wilkins RM, Pritchard DJ, Omer EB, Unni KK. Ewing’s sarcoma of bone. Experience with 140 patients. Cancer 1986;58:2551–5.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
G Model
CANEP-797; No. of Pages 7 K.R. Duchman et al. / Cancer Epidemiology xxx (2015) xxx–xxx [16] Paulussen M, Ahrens S, Dunst J, Winkelmann W, Exner G, Kotz R, et al. Localized Ewing tumor of bone: final results of the cooperative Ewing’s Sarcoma Study CESS 86. J Clin Oncol 2001;19:1818–29. [17] Cotterill S, Ahrens S, Paulussen M, Ju¨rgens H, Voute P, Gadner H, et al. Prognostic factors in Ewing’s tumor of bone: analysis of 975 patients from the European Intergroup Cooperative Ewing’s Sarcoma Study Group. J Clin Oncol 2000;18:3108–14. [18] Cangir A, Vietti TJ, Gehan EA, Burgert EO, Thomas P, Tefft M, et al. Ewing’s sarcoma metastatic at diagnosis results and comparisons of two intergroup Ewing’s sarcoma studies. Cancer 1990;66:887–93. [19] Paulussen M, Ahrens S, Burdach S, Craft A, Dockhorn-Dworniczak B, Dunst J, et al. Primary metastatic (stage IV) Ewing tumor: survival analysis of 171 patients from the EICESS studies. Ann Oncol 1998;9:275–81. [20] Du XL, Fang S, Coker AL, Sanderson M, Aragaki C, Cormier JN, et al. Racial disparity and socioeconomic status in association with survival in older men with local/regional stage prostate carcinoma. Cancer 2006;106:1276–85. [21] Miller BJ, Cram P, Lynch CF, Buckwalter JA. Risk factors for metastatic disease at presentation with osteosarcoma: an analysis of the SEER database. J Bone Joint Surg Am 2013;95:e891–8. [22] Sun M, Abdollah F, Liberman D, Abdo A, Thuret R, Tian Z, et al. Racial disparities and socioeconomic status in men diagnosed with testicular germ cell tumors. Cancer 2011;117:4277–85. [23] Gupta AA, Pappo A, Saunders N, Hopyan S, Ferguson P, Wunder J, et al. Clinical outcome of children and adults with localized Ewing sarcoma. Cancer 2010;116:3189–94. [24] Verrill M, Judson I, Wiltshaw E, Thomas J, Harmer C, Fisher C. The use of paediatric chemotherapy protocols at full dose is both a rational and feasible treatment strategy in adults with Ewing’s family tumours. Ann Oncol 1997;8:1099–105. [25] Karski EE, Matthay KK, Neuhaus JM, Goldsby RE, DuBois SG. Characteristics and outcomes of patients with Ewing sarcoma over 40 years of age at diagnosis. Cancer Epidemiol 2013;37:29–33.
7
[26] Bacci G, Longhi A, Briccoli A, Bertoni F, Versari M, Picci P. The role of surgical margins in treatment of Ewing’s sarcoma family tumors: experience of a single institution with 512 patients treated with adjuvant and neoadjuvant chemotherapy. Int J Radiat Oncol Biol Phys 2006;65:766–72. [27] Polednak AP. Primary bone cancer incidence in black and white residents of New York State. Cancer 1985;55:2883–8. [28] Crawford SM, Sauerzapf V, Haynes R, Zhao H, Forman D, Jones AP. Social and geographical factors affecting access to treatment of lung cancer. Br J Cancer 2009;101:897–901. [29] Duchman KR, Lynch CF, Buckwalter JA, Miller BJ. Estimated cause-specific survival continues to improve over time in patients with chondrosarcoma. Clin Orthop Relat Res 2014;472(8):2516–25. [30] Giuffrida AY, Burgueno JE, Koniaris LG, Gutierrez JC, Duncan R, Scully SP. Chondrosarcoma in the United States (1973 to 2003): an analysis of 2890 cases from the SEER database. J Bone Joint Surg Am 2009;91:1063–72. [31] Johnson DR, Ma DJ, Buckner JC, Hammack JE. Conditional probability of longterm survival in glioblastoma. Cancer 2012;118:5608–13. [32] Bach PB, Guadagnoli E, Schrag D, Schussler N, Warren JL. Patient demographic and socioeconomic characteristics in the SEER-Medicare database: applications and limitations. Med Care 2002;40:IV-19–25. [33] Nattinger AB, McAuliffe TL, Schapira MM. Generalizability of the surveillance, epidemiology, and end results registry population: factors relevant to epidemiologic and health care research. J Clin Epidemiol 1997;50:939–45. [34] Krieger N, Chen JT, Waterman PD, Soobader M-J, Subramanian S, Carson R. Geocoding and monitoring of US socioeconomic inequalities in mortality and cancer incidence: does the choice of area-based measure and geographic level matter? the Public Health Disparities Geocoding Project. Am J Epidemiol 2002;156:471–82. [35] Subramanian S, Chen J, Rehkopf D, Waterman P, Krieger N. Comparing individual-and area-based socioeconomic measures for the surveillance of health disparities: a multilevel analysis of Massachusetts births, 1989–1991. Am J Epidemiol 2006;164:823–34.
Please cite this article in press as: Duchman KR, et al. Prognostic factors for survival in patients with Ewing’s sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology (2015), http://dx.doi.org/10.1016/ j.canep.2014.12.012
