Surgical Resection of Primary Ewing's Sarcoma of Bone Improves Overall Survival in Patients Presenting with Metastasis.

BACKGROUND Metastatic Ewing's sarcoma (ES) of bone has a poor prognosis. Because there have been few previous studies on the prognostic factors and clinical outcome in patients with ES who have metastases at presentation, the aim of this study was to use the Surveillance, Epidemiology, and End Results (SEER) database to compare the clinical outcome following single and combined radiation treatment and surgery. MATERIAL AND METHODS The SEER database was used to identify patients with ES who presented with bone involvement and metastasis between 1973 to 2015. Prognostic analysis was performed using the Kaplan-Meier method and the Cox proportional hazards regression model. RESULTS There were 643 patients identified from the SEER database. The 5-year overall survival (OS) and cancer-specific survival (CSS) rates were 33.1% and 34.3%, respectively and the median OS and CSS were 29.0±1.9 and 29.0±2.1 months, respectively. Multivariate analysis identified age <20 years and surgical resection of the primary tumor to be significantly associated with improved OS. Radiation therapy was not an independent predictor of OS or CSS. Radiation therapy alone resulted in a significantly reduced the OS and CSS compared with surgical resection alone. Combined surgery and radiation therapy of the primary tumor did not significantly improve the OS or CSS of patients with ES and metastatic disease when compared with surgery alone. CONCLUSIONS Age <20 years and surgical resection of the primary tumor were significantly associated with improved OS in patients with primary ES of bone who presented with metastasis.

Background

Ewing’s sarcoma (ES) is the second most common primary malignancy of bone in children and adolescents [1,2]. Approximately one-third of patients with ES present with metastatic disease [3,4]. The demographic, prognostic, and outcome data for ES are well documented. Patient age, primary tumor site, size, grade, and stage, treatment with chemotherapy, surgery, and radiation therapy are all prognostic factors in patients with ES [3,5-10]. It has previously been reported that systemic chemotherapy combined with surgery and/or radiation therapy for local control of ES was associated with a 5-year survival rate of approximately 70% [11]. However, patients with primary ES of bone who presented with metastasis usually have a poor prognosis, with a 5-year event-free survival rate from presentation of between 20–30% [1,11]. A previously reported study that used patient data from the Surveillance, Epidemiology, and End Results (SEER) database showed that the 10-year survival was 28.1% when patients with osseous ES presented with metastasis, tumor size >10 cm, patient age ≥20 years, and axial primary tumor location [3].

Current standard localized or regional treatment for patients with ES includes systemic chemotherapy and surgery and/or radiation therapy for local control. However, the treatment for metastatic ES remains a challenge and chemotherapy has been the accepted treatment [12]. Whether surgical excision of the primary bone tumor improves the survival of patients who present with metastatic ES is unknown. ES is considered to be a radiosensitive tumor and radiation therapy may provide a survival benefit [6]. However, Arshi et al. [8] reported that radiation therapy did not improve the survival of patients with ES arising in the spine. Therefore, the role of radiation therapy on the prognosis of patients who present with primary osseous ES with metastases requires further study.

Because there have been few previous studies on prognostic factors and clinical outcome in patients with ES who have metastases at presentation, the aim of this large-scale study was to use the Surveillance, Epidemiology, and End Results (SEER) program database of the National Cancer Institute (NCI) to compare the clinical outcome following single and combined radiation treatment and surgery to identify predictors of survival.

Material and Methods

Patient population

Data on patients with primary Ewing’s sarcoma (ES) of bone with metastasis at diagnosis were obtained using the case-listing session procedure from the Surveillance, Epidemiology, and End Results (SEER) program. The SEER database is publicly available, confidential, and does not contain patient identification data. This study was approved by the Ethics Committee of Yuyao Peoples’ Hospital of Zhejiang Province, China.

A total of 2,436 patients were identified who were diagnosed between 1973 and 2015 with primary ES of bone with metastasis at initial presentation. The International Classification of Diseases for Oncology, 3rd edition (ICD-O-3) was used to identify patients with ES (ICD-O-3 histologic type: 9260; ICD-O-3 site code: C40.0–40.3, C40.8–41.4, C41.8–41.9), using the case-listing procedure. Only patients with metastatic disease were enrolled, based on tumor stage at diagnosis. Patients with ES arising from the extra-skeletal locations were excluded. The diagnosis in all cases was confirmed histologically, based either on biopsy results or from the surgical resection specimen. Thirty-seven patients were excluded who were diagnosed only by clinical presentation, radiography, or identified as unknown. Forty-one patients who did not receive chemotherapy, 17 patients with unknown therapy, and four patients with a survival time of less than one month were also excluded from the study (Figure 1).

Figure 1

Study flowchart of the selection of the study population. ES – Ewing’s sarcoma; SEER – Surveillance, Epidemiology, and End Results; ICD-O-3 – International Classification of Diseases for Oncology, 3rd Edition.

Data extracted from the SEER database included patient age, gender, race, year of diagnosis, tumor location, grade, stage, type, size, surgical treatment, radiation treatment, chemotherapy, the cause of death, and survival time. In this study, surgery or radiation treatment referred to treatment for the local primary bone tumor. Tumor location was identified by four categories: axial (pelvis and spine; appendicular (long and short bones of the upper and lower extremities); rib, sternum, and clavicle, and; other locations (mandible, skull, and other atypical locations).

Statistical analysis

Statistical analysis was performed using Microsoft Excel 2016 (Microsoft Corp., Redmond, WA, USA) and SPSS software version 21.0 (SPSS Inc., Chicago, IL, USA). Overall survival (OS) was defined as the time from diagnosis to death from any cause and cancer-specific survival (CSS) was defined as the time from diagnosis to death due specifically to ES. The Kaplan-Meier method was used to draw the OS and CSS curves and calculate the median survival. Observations were statistically censored if the patient was alive at the time of the last follow-up. Univariate analysis was performed using the Kaplan-Meier method with the log-rank test. Variables with a P-value <0.05 from the univariate analysis were included from multivariate analysis. Multivariate analysis was used to determine the independent predictors of OS and CSS with a Cox proportional hazards regression model. The hazard ratios (HR) and corresponding 95% confidence interval (CI) were used to show the effect of patient factors on OS and CSS. Differences were deemed statistically significant if the P-value was <0.05.

Results

Clinical characteristics of 643 patients with metastatic Ewing’s sarcoma (ES)

In total, 643 patients with metastatic Ewing’s sarcoma (ES) identified from the Surveillance, Epidemiology, and End Results (SEER) database were eligible for the study. In terms of primary tumor location, 34.5% of primary tumors were located in the extremities, 41.5% in the axial skeleton, and 14.2% in the rib, sternum, or clavicle. Information on the tumor size was available in 54% of the cases and was categorized based on the mean tumor size (10 cm). All patients underwent chemotherapy. About one-third of the patients (37.8%) received local surgery and more than half (61.4%) received radiation treatment. There were 412 (64.1%) patients who died, and 387 patients died from metastatic ES. The 3-year and 5-year OS rates for the entire cohort were 43.4% and 33.1%, respectively. The 3-year and 5-year CSS rates were 44.7% and 34.3%, respectively (Table 1). The median OS and CSS were 29.0±1.9 and 29.0±2.1 months, respectively, indicating a poor prognosis for this cohort (Table 2).

Table 1

Demographic and clinical characteristics of 643 patients with primary Ewing’s sarcoma of bone and metastatic disease at presentation identified in the Surveillance, Epidemiology, and End Results (SEER) database from 1973 to 2015.

Category	Value
Mean age (years)	20
Median age (years)	17
Age (years)
<20	411 (63.9%)
≥20	232 (36.1%)
Gender
Female	236 (36.7%)
Male	407 (63.3%)
Location
Appendicular	222 (34.5%)
Axial	267 (41.5%)
Rib, sternum and clavicle	91 (14.2%)
Other locations	63 (9.8%)
Tumor size
Mean (cm)	10
Median (cm)	9
<10 cm	181 (28.1%)
≥10 cm	166 (25.8%)
Unknown	296 (46%)
Surgery
Yes	243 (37.8%)
No	400 (62.2%)
Radiation treatment
Yes	395 (61.4%)
No	248 (38.6%)
Local treatment
Surgery + radiation	147 (22.9%)
Surgery only	96 14.9%)
Radiation only	248 (38.6%)
No therapy	152 (23.6%)
Alive
No	412 (64.1%)
Yes	231 (35.9%)
3-year OS rate	43.4%
3-year CSS rate	44.7%
5-year OS rate	33.1%
5-year CSS rate	34.3%

OS – overall survival; CSS – cancer-specific survival; SEER – Surveillance, Epidemiology, and End Results.

Table 2

Median survival data (in months) of patients with Ewing’s sarcoma of bone and metastatic disease at presentation.

Category	Overall survival	95% CI	Cancer-specific survival	95% CI
Overall	29.0±1.9	25.2–32.8	29.0±2.1	24.9–33.1
Age (years)
<20	33.0±2.7	27.7–38.3	34.0±2.7	28.6–39.4
≥20	21.0±1.9	17.2–24.8	22.0±3.2	15.7–28.3
Gender
Female	34.0±4.6	25.0–43.0	36.0±5.4	25.5–46.5
Male	28.0±1.9	24.2–31.8	28.0±2.2	23.6–32.4
Location
Appendicular	32.0±3.1	26.0–38.0	32.0±3.2	25.8–38.2
Axial	25.0±2.0	21.0–29.0	26.0±2.1	21.9–30.1
Rib, sternum and clavicle	57.0±10.9	35.7–78.3	57.0±11.7	34.2–79.8
Other locations	21.0±4.9	11.4–30.6	26.0±6.1	14.1–37.9
Tumor size
<10 cm	43.0±8.1	27.2–58.8	47.0±14.3	18.9–75.1
≥10 cm	32.0±3.4	25.4–38.6	33.0±3.9	25.4–40.6
Surgery
Yes	40.0±6.7	27.0–53.0	43.0±7.6	28.1–57.9
No	22.0±1.7	18.6–25.4	22.0±1.9	18.4–25.6
Radiation treatment
Yes	32.0±2.3	27.6–36.4	34.0±2.4	29.3–38.7
No	23.0±2.4	18.3–27.7	26.0±2.3	21.4–30.6
Local treatment
Surgery + radiation	41.0±8.1	25.1–56.9	41.0±8.0	25.4–56.6
Surgery only	39.0±14.7	10.2–67.8	50.0±16.3	18.0–82.0
Radiation only	28.0±3.0	22.1–33.9	28.0±3.3	21.5–34.5
No therapy	18.0±1.9	14.3–21.7	18.0±1.9	14.2–21.8

CI – confidence interval.

Univariate analysis of variables associated with OS and CSS for patients with metastatic ES

Table 2 summarizes the median survival data, in months, for patients with ES of bone and metastatic disease at presentation. Table 3 shows the results of Kaplan-Meier univariate survival analysis. Univariate analysis showed that gender and tumor size were not associated with OS or CSS. Younger patients had a significantly better outcome, with a longer median survival time, compared with older patients (Figures 2A, 3A). Patients with axial tumors had poorer outcomes than those with appendicular, rib, sternum, or clavicle tumors.

Table 3

Univariate analysis of variables in patients with Ewing’s sarcoma of bone and metastatic disease at presentation using the Kaplan-Meier method.

Category	Overall survival (log-rank p-value)	Cancer-specific survival (log-rank p-value)
Age at diagnosis	0.000	0.000
Gender	0.080	0.065
Location	0.003	0.008
Axial vs. appendicular	0.019	0.030
Axial vs. rib, sternum, and clavicle	0.001	0.003
Axial vs. other locations	0.639	0.792
Appendicular vs. rib, sternum and clavicle	0.222	0.186
Appendicular vs. other locations	0.055	0.113
Rib, sternum and clavicle vs. other locations	0.008	0.018
Tumor size (<10 cm vs. ≥10 cm)	0.095	0.124
Surgery	0.000	0.000
Radiation treatment	0.048	0.067
Local treatment	0.000	0.000
Surgery + radiation vs. surgery only	0.706	0.966
Surgery + radiation vs. radiation only	0.004	0.008
Surgery + radiation vs. no therapy	0.000	0.000
Surgery only vs. radiation only	0.044	0.024
Surgery only vs. no therapy	0.000	0.000
Radiation only vs. no therapy	0.013	0.010

Figure 2

Kaplan-Meier plots of the overall survival (OS) in patients with primary Ewing’s sarcoma of the bone and metastatic disease at presentation stratified by age, surgery, and radiation therapy. (A) OS in patients with primary Ewing’s sarcoma of the bone and metastatic disease at presentation stratified by age (years) at diagnosis. (B) OS in patients with primary Ewing’s sarcoma of the bone and metastatic disease at presentation stratified by treatment with surgery alone. (C) OS in patients with primary Ewing’s sarcoma of the bone and metastatic disease at presentation stratified by treatment with surgery and radiation therapy.

Figure 3

Kaplan-Meier plots of the cancer-specific survival (CSS) in patients with primary Ewing’s sarcoma of the bone and metastatic disease at presentation stratified by age, surgery, and radiation therapy. (A) CSS in patients with primary Ewing’s sarcoma of the bone and metastatic disease at presentation stratified by age (years) at diagnosis. (B) CSS in patients with primary Ewing’s sarcoma of the bone and metastatic disease at presentation stratified by treatment with surgery alone. (C) CSS in patients with primary Ewing’s sarcoma of the bone and metastatic disease at presentation stratified by treatment with surgery and radiation therapy.

In terms of treatment, patients who underwent surgical treatment had a better OS and CSS than those who did not (Figures 2B, 3B). Radiation treatment significantly prolonged the OS of patients with ES who presented with metastasis. However, radiation treatment had no significant effect on CSS and the addition of radiation treatment did not significantly improve the OS and CSS of patients who underwent surgery (Figures 2C, 3C). The group treated with surgery or radiation therapy alone had considerably better outcomes when compared with the non-treated group (Figures 2C, 3C).

Multivariate analysis of independent predictors of OS and CSS for patients with metastatic ES

Multivariate analysis of all patients identified age at diagnosis and surgery for primary tumors to be independent predictors of OS and CSS. Multivariate analysis identified no significant difference in either OS or CSS based on tumor location and showed that radiation treatment was not an independent prognostic factor for OS (Table 4).

Table 4

Multivariate analysis of overall survival (OS) and cancer-specific survival (CSS) for patients with primary Ewing’s sarcoma of bone and metastatic disease at presentation.

Variable	Overall survival		Cancer-specific survival
	Hazard ratio (95% CI)	P-value	Hazard ratio (95% CI)	P-value
Age (years)		0.001		0.001
<20	1		1
≥20	1.396 (1.138–1.712)	0.001	1.416 (1.149–1.744)	0.001
Location		0.117		0.197
Appendicular	1		1
Axial	1.214 (0.967–1.524)	0.095	1.180 (0.936–1.488)	0.162
Rib, sternum and clavicle	0.890 (0.642–1.233)	0.482	0.871 (0.618–1.227)	0.428
Other locations	1.284 (0.904–1.822)	0.162	1.240 (0.861–1.785)	0.248
Surgery		0.000		0.000
Yes	1		1
No	1.477 (1.192–1.829)	0.000	1.501 (1.203–1.873)	0.000
Radiation treatment		0.095		–
Yes	1		–
No	1.187 (0.970–1.453)	0.095	–	–

Discussion

In childhood and adolescence, Ewing’s sarcoma (ES) is the second most common primary sarcoma of bone and this tumor metastasizes to the lungs, bones, and other organs at an early stage [1,2]. Despite intensive treatment, patients with ES of bone and metastasis usually have a poor prognosis, with a 5-year OS rate <30% [13-15]. The findings of the present study showed that the 5-year OS and CSS rates for the entire cohort were 30.9% and 37.6%, respectively. Therefore, there is a significant need to improve these clinical outcomes. Few studies have documented the prognosis of patients with metastatic primary ES of bone. To our knowledge, this study is the first to assess the demographic information with surgical and radiation treatment information of patients with metastatic primary ES of bone and to explore possible predictors of survival using the Surveillance, Epidemiology, and End Results (SEER) database.

The mean and median age at diagnosis of this cohort was 20 and 17 years, respectively, with a male predominance (1.7: 1.0), which was similar to previously reported findings [3]. Univariate analysis of the cohort of patients studied showed that gender was not associated with significant differences in either OS or CSS, which was also supported by the findings of previous studies [3,8,16-18]. However, Bacci et al. [19] reported that female gender was an independent favorable prognostic factor for patients with non-metastatic ES. Several previously published studies have shown that young age at diagnosis was associated with improved patient prognosis [3,16-19]. Bacci et al. [19] reported that an age ≤14 years was an independent predictor of positive outcomes for patients with non-metastatic ES. The findings of the present study also showed that patient age ≥20 years was independently associated with decreased OS and CSS in patients with primary ES of bone with metastasis. Although age is an important prognostic factor for patients with ES, with or without metastasis at presentation, some recent studies have found age not to be a predictor of outcome for patients with ES, which has been attributed to the similar treatment of adults and children [20-22].

ES frequently presents as an axial mass, which was also the case for the cohort of patients in the present study [13,14]. An appendicular location of ES has previously been shown to be associated with a better outcome compared with an axial location [3]. In the present study, tumor location was significantly associated with both OS and CSS but was not an independent prognostic factor. Tumor size has been shown to be an important predictor of outcome in patients with ES [3,9,18,23]. Although the findings of the present study showed that tumor size was not associated with either OS or CSS, trends toward increased OS and CSS were observed for patients with tumor size <10 cm compared with patients with a tumor size ≥10 cm.

Current treatment for ES arising in bone consists of systemic chemotherapy in addition to local control with surgical excision and/or radiation therapy. However, little is known about the optimal standard treatment for patients with ES of bone who have metastasis at presentation. Surgical resection is the main local treatment for patients with ES and can prolong survival [24]. In the cohort in the present study, surgical resection of the primary tumors also prolonged the survival of patients with ES who presented with metastasis. Also, in patients with metastasis, removing the primary tumor can reduce pain, improve quality of life, and prolong survival. Therefore, local surgery is an appropriate treatment for patients with metastatic ES. Raciborska et al. [25] reported that treatment of isolated lung metastases may have a role in improving prognosis in patients with ES. Letourneau et al. [26] reported that resection of pulmonary metastases in pediatric patients with ES improved survival. Liu et al. [27] reported that radiation therapy for local control of metastatic sites was effective and tolerable in children with metastatic ES. Therefore, for these patients, both primary and metastatic lesions should be actively treated in order to maximize survival time.

Although ES is considered to be a radiosensitive tumor, radiation therapy is favored for central or unresectable tumors [28]. The effect of radiation therapy on the prognosis for metastatic ES is unclear. Previous studies have shown that patients with ES who received only radiation therapy as the local treatment, had significantly reduced OS and CSS compared with patients who underwent surgery alone [24,29,30], which is consistent with the findings of the present study. However, some studies have reported that radiation therapy alone improved local control and survival that was comparable to surgery [31-33]. In the present study, radiation therapy was not an independent predictor of OS and was not associated with CSS and patients who received local radiation therapy alone or no local treatment had the worst outcomes (Table 2) (Figures 2C, 3C). Also, the combination of surgery and radiation therapy appeared to be unnecessary as the addition of radiation treatment did not significantly improve the OS and CSS of patients with metastatic ES who underwent surgery (Table 3).

This study had several limitations. Data on other known prognostic factors for cancer survival, including the surgical method used and the use of neoadjuvant chemotherapy were not available in the database. Also, data indicating whether surgery was performed before or after chemotherapy was not available, and tumor size was only available for 54% of the patients. Despite these limitations, the SEER database provides important insights into rare cancers, including ES, and was an important resource for data on patients with primary ES of bone who presented with late-stage disease and tumor metastasis.

Conclusions

To our knowledge, this was the largest population-based study that included patient demographics, clinical presentation, treatment, and clinical outcome to analyze the prognostic factors in patients with primary Ewing’s sarcoma (ES) of bone and metastatic disease at diagnosis. The study included for 643 patients, and the 5-year overall survival (OS) and cancer-specific survival (CSS) rates were 33.1% and 34.3%, respectively. The median OS and CSS were 29.0±1.9 and 29.0±2.1 months, respectively. The independent predictors of both OS and CSS were age at diagnosis and surgical excision of the primary tumors. It is hoped that the findings for this study on the OS, CSS, and risk factors for patients with primary ES of bone who present with metastasis, will provide the basis for future research on ES and for the development of standardized treatment.