Treatment patterns and survival outcomes of Ewing sarcoma in Ethiopia: a five-year retrospective analysis from a tertiary hospital

Abstract

Background

Ewing sarcoma is a malignant small round-cell sarcoma of bone or soft tissue that is highly responsive to modern multimodal therapy and considered a highly curable disease in developed countries. However, survival remains poor in developing countries. This study aims to assess treatment patterns and survival outcomes of Ewing sarcoma at Tikur Anbessa Hospital.

Methods and materials

We conducted a retrospective study on all patients diagnosed with Ewing sarcoma at Tikur Anbessa Hospital between November 15, 2018, and November 15, 2023. Descriptive statistics and Kaplan–Meier survival analysis were used to analyze the data.

Result

The median age at diagnosis was 19 years, with most of the patients (67.2%) in the age range of 8–25 years. The main modality of treatment for localized disease was chemotherapy combined with surgery, whereas chemotherapy alone was used for metastatic disease. The Kaplan–Meier estimated 1-year and 3-year overall survival (OS) rates for the entire cohort were 96% and 68%, respectively, and the corresponding 1-year and 3-year event-free survival (EFS) rates were 70% and 40%, respectively. A sensitivity analysis treating loss to follow-up as events reduced the 3-year overall survival from 68% to 55% and the 3-year event-free survival from 40 to 33%. The 3-year OS and EFS rates were 89% and 76% for localized disease versus 10% and 0% for metastatic disease. This difference was statistically significant (log-rank p < 0.001).

Small tumor size (≤ 8 cm), surgical intervention, and localized stage were associated with statistically significant improvements in OS and EFS, whereas pelvic location was associated with a poorer prognosis.

Conclusion

Ewing sarcoma at our institution mainly affects adolescents and young adults. Survival is poor, especially for metastatic disease, while localized tumors, smaller size, and surgery are associated with better survival. Therefore, early diagnosis and improved multimodal treatment are needed to improve survival.

Introduction

Ewing sarcoma (ES) is a malignant small round-cell tumor of bone or soft tissue characterized by rearrangements of the Ewing Sarcoma RNA binding protein 1 (EWSR1) gene and members of the E26 transformation-specific (ETS) transcription factor family [1–3]. It is the third most common primary bone malignancy overall and the second most common primary bone malignancy in pediatric and adolescent populations, comprising approximately 2–3% of all malignancies in this age group. Although its incidence in older populations is not well defined, approximately 30% of cases occur in individuals older than 20 years [4–8].

Unlike most other bone and soft tissue sarcomas, ES is highly sensitive to both systemic multiagent chemotherapy and radiotherapy. Over the past few decades, data from developed countries have shown that the 5-year survival rate has improved from 10% to over 70% for localized disease and from 0 to 30% for metastatic disease [8–12]. However, data on treatment patterns and survival outcomes from low- and middle-income countries are limited. Available reports suggest that 5-year survival for localized disease ranges from 35–69%, and about 4% for metastatic disease. To the best of our knowledge, no published data exist on survival and treatment modality for Ewing sarcoma from Sub-Saharan Africa including Ethiopia [13, 14].

Therefore, the aim of this study was to evaluate treatment patterns and survival outcomes of patients with Ewing sarcoma at Tikur Anbessa Hospital, Ethiopia. The findings will provide locally relevant evidence to inform clinical decision-making and optimize resource allocation, while also establishing a foundation for future prospective research in resource-limited settings.

Methods

Study design and setting

This retrospective cohort study was conducted at Tikur Anbessa Specialized Hospital (TASH), Addis Ababa University, Ethiopia, which is the largest tertiary referral center and the country’s leading facility for comprehensive cancer care, including chemotherapy, radiotherapy, surgery, and palliative services. The study was performed in the Adult Oncology Department and the Pediatric Hematology–Oncology Unit.

Study population and eligibility criteria

Patients were included in the study if the diagnosis of Ewing sarcoma was established by the treating physician based on compatible histopathological and imaging features; if they were diagnosed and treated at Tikur Anbessa Specialized Hospital (TASH) between November 15, 2018, and November 15, 2023; if they received treatment (chemotherapy, surgery, or a combination) at TASH; and if they had an available date of diagnosis and at least one follow-up record to estimate survival. A total of 75 consecutive patients diagnosed with Ewing sarcoma at TASH between November 2018 and November 2023 were screened. After exclusions (Fig. 1), 61 patients were included in the final analysis.

Fig. 1.

Flow diagram of patient selection. Of 75 patients diagnosed with Ewing sarcoma at Tikur Anbessa Specialized Hospital between November 15, 2018, and November 15, 2023, 14 were excluded, leaving 61 patients for final analysis

Diagnostic approach

As this study was conducted in a resource-limited setting with limited access to immunohistochemistry and molecular testing (FISH), the diagnosis of Ewing sarcoma was established by the treating physician based on compatible histopathological and radiological findings. This assessment formed the basis for patient inclusion, and all cases in the final analysis were diagnosed and managed as Ewing sarcoma by the treating physician.

Variables and data collection

Clinical and treatment variables

Data were extracted from medical records, operative reports, and oncology charts. Treatment variables included chemotherapy regimen, surgical intervention, radiotherapy, and treatment sequence. Treatment pattern was defined as the type and sequence of treatments received, including surgery, chemotherapy, and radiotherapy.

Chemotherapy regimens were abstracted from patient records. No formal institutional protocol was documented. The most common curative-intent regimen was vincristine, doxorubicin, and cyclophosphamide alternating with ifosfamide and etoposide (VDC/IE). For palliative intent, the most common regimen was vincristine, actinomycin-D, and cyclophosphamide (VAC). Dosing schedules varied by department: every 2 weeks (pediatric hemato-oncology) [15] versus every 3 weeks (clinical oncology).

Surgical procedures were categorized from operative reports as curative-intent surgery (excision of the primary tumor, not biopsy). Margin status (positive or negative) was recorded from pathology reports when available. Radiotherapy data were obtained from oncology records. Most patients received 3D conformal radiotherapy (95.7%), while 4.3% received 2D RT. Treatment intent was mainly definitive (70%) with the remainder adjuvant for positive margins (30%). The mean dose was 46 Gy (20–55.8 Gy) for curative intent and 25 Gy (20–30 Gy) for palliative intent.

Adherence to standard treatment guidelines

Treatment adherence was operationally defined based on receipt of multi-agent chemotherapy and appropriate local control therapy (surgery and/or radiotherapy) as documented in patient records. This was assessed in comparison with standard treatment principles outlined in the National Comprehensive Cancer Network guidelines for Ewing sarcoma applicable during the study period. For localized disease, adherence included receipt of appropriate local therapy (surgery and/or radiotherapy) and completion of the recommended perioperative chemotherapy cycles. For metastatic disease, adherence was defined based on receipt of an appropriate systemic chemotherapy regimen and completion of the recommended number of chemotherapy cycles according to disease stage.

Staging and imaging

Baseline imaging included contrast-enhanced CT of the primary site and chest CT for metastatic workup. MRI was used for local tumor extent assessment when available. Staging followed the AJCC TNM system (8th edition) [16]

Response assessment and follow-up

Treatment response was assessed retrospectively using RECIST criteria (version 1.1) based on radiology reports. Radiologic assessment was performed using computed tomography (CT) scans or magnetic resonance imaging (MRI), depending on availability. Response categories were defined as: complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD). The timing of response assessment was recorded as the interval between treatment completion and the follow-up imaging study (range: 2 weeks to 4 months).

Follow-up evaluation, including imaging, was performed at variable intervals based on clinical indication and treating physician discretion in routine practice. Recurrence or disease progression was identified from clinical evaluation and radiological findings documented in patient records.

Outcome definition

Overall survival (OS) was defined as the time from diagnosis to death from any cause or last follow-up. Event-free survival (EFS) was defined as the time from diagnosis to the first occurrence of disease progression, recurrence, death from any cause, or last follow-up.

Statistical analysis

All analyses were performed using IBM SPSS Statistics (Version 25). Descriptive statistics were used to summarize patient demographics, clinical characteristics, and treatment patterns. Overall survival (OS) and event-free survival (EFS) were estimated using Kaplan–Meier method and compared between groups using the log-rank test. Patients who were alive without an event at last contact or lost to follow-up were censored at the date of last known follow-up. Cox proportional hazards regression was used to analyze the association between patient demographics, disease characteristics, and different treatment modalities. Due to the small sample size, only clinically relevant covariates with a p ≤ 0.25 on univariate analysis were included in the multivariable model to reduce the risk of overfitting. Statistical significance was defined as p < 0.05. Given the limited number of outcome events relative to the number of covariates, the multivariable Cox regression analysis was considered exploratory and intended for hypothesis generation rather than definitive inference. Missing data were minimal (< 6%), therefore, descriptive and primary analyses were based on complete cases.

Results

Of 75 patients diagnosed with Ewing sarcoma at TASH between November 15, 2018, and November 15, 2023, 61 met the inclusion criteria and were included in this study (Fig. 1).

Socio-demographic and disease characteristics

The median age at diagnosis was 19 years, with minimum and maximum age range of 5–50 years. Of this, 39.3% (n = 24) of the studied population were in the pediatric age group and 60.7% (n = 37) of them were adults (≥ 18 years) (Table 1). Among the 61 patients included in the study, 4 underwent CD99 immunohistochemistry at diagnosis, all of which were positive. No other immunohistochemical or molecular tests were performed or documented.

Table 1.

Socio-demographic and disease characteristics

Patient characteristics (N = 61)		Frequency (n)	%
Age group (years)	0–7	5	8.2
	8–17	19	31.1
	18–25	22	36.1
	> 25	15	24.6
Sex	Female	40	65.6
	Male	21	34.4
Stage	Localized	48	78.7
	Metastatic	13	21.3
Metastatic sitea	Lung	9	69.2
	Lung + bone/bone marrow	1	7.6
	Lung and other	2	15.3
	Bone or bone marrow	1	7.6
Primary tumor size	≤ 8 cm	26	42.6
	> 8 cm	35	57.4
Disease origin	Bone	55	90.1
	Soft tissue	6	9.9

^aPercentages are based on the number of patients with metastatic disease (n = 13)

Treatment patterns

Among the entire cohort, 73.8% received treatment with curative intent and 26.2% as palliative intent. For localized cancer, the goal was curative in 94% of patients; however, 6% received palliative care. Among patients treated with curative intent, the most common treatment modalities were chemotherapy combined with surgery (42.2%), followed by chemotherapy combined with radiotherapy (40%), and trimodality therapy (chemotherapy, surgery, and radiotherapy) (17.8%). All patients with metastatic disease received palliative treatment.

Among patients treated for localized disease with curative intent, the majority (70%) received radiotherapy as definitive treatment, whereas 30% received adjuvant radiotherapy due to positive surgical margins.

Adherence to standard treatment was 16.7% for localized and 20% for metastatic disease. Among patients treated with curative intent, 71.1% received neoadjuvant chemotherapy. Chemotherapy dosing schedules and treatment intensity varied by treating department: the pediatric hemato-oncology department administered an intensive regimen every two weeks, while the clinical oncology department used a conventional regimen every three weeks. The average number of chemotherapy cycles also varied by department. In pediatric hemato-oncology, it was 12 cycles, compared to 6 cycles in the clinical oncology department. The most common type of combination chemotherapy used in curative-intent treatment (97%) was vincristine, doxorubicin and cyclophosphamide alternating with ifosfamide and etoposide (VDC/IE).

At treatment completion, based on radiological assessment by RECIST 1.1 criteria, 49.2% (n = 30) of the patients had complete response (CR), 39.3% (n = 24) had partial response (PR), and 6.6% had stable disease (SD).

Survival outcome

Median follow-up time of the study by using the reverse Kaplan–Meier method is 34.0 months (95% CI: 28.7–39.2) with a minimum and maximum follow-up time of 8.0 to 59.0 months, respectively. At the end of follow-up, 57.4% (n = 35) of the patients were alive, 32.8% (n = 20) died and the survival status unknown in 9.8% (n = 6) of the cases.

The median overall survival and event-free survival in the entire cohort were 45.0 months and 24.0 months, respectively. The Kaplan–Meier estimated 1-year and 3-year OS rates for the entire cohort were 96% and 68%, respectively (Fig. 2A), whereas the corresponding 1-year and 3-year EFS rates were 70% and 40%, respectively (Fig. 2B).

Fig. 2.

Kaplan–Meier survival curves for the entire cohort of patients with Ewing sarcoma. A Overall survival (OS) and B Event-free survival (EFS)

Survival differed significantly by disease stage at presentation. Patients with localized disease had superior outcomes, with a median OS not reached and median EFS of 32.0 months, whereas those with metastatic disease had a median OS of 23.0 months and a median EFS of 12.0 months. The 3-year OS and EFS rates were 89% and 76% for localized disease versus 10% and 0% for metastatic disease. These differences for both OS and EFS reached statistical significance (log-rank p < 0.001) (Fig. 3A and B).

Fig. 3.

Kaplan–Meier survival curves stratified by disease stage at diagnosis in patients with Ewing sarcoma. A Overall survival (OS) and B Event-free survival (EFS). (log-rank p < 0.001, for both)

A sensitivity analysis treating loss to follow-up as events reduced the 3-year overall survival from 68% to 55% and the 3-year event-free survival from 40 to 33%.

Subgroup survival analysis comparing pediatric (< 18 years, n = 24) and adult (≥ 18 years, n = 37) patients showed higher 3-year overall survival among pediatric patients (73% vs. 62%); however, the difference was not statistically significant (log-rank p = 0.257). Median survival was 37 months in adults, while it was not reached in the pediatric group due to censoring (> 50% alive at last follow-up).

Prognostic features

We evaluated tumor size, disease stage at diagnosis, whether resection (surgery) was performed, pelvic location of the primary tumor, and number of chemotherapy cycles using univariable and multivariable analyses to identify prognostic factors for OS and EFS.

On multivariable analysis, tumor size ≤ 8 cm, localized stage, and receipt of surgery were independently associated with improved OS and EFS. Conversely, pelvic location in patients with metastatic disease was associated with worse outcomes. The number of chemotherapy cycles in localized disease showed a non-significant trend toward better survival (HR = 0.42; 95% CI: 0.12–1.20; p = 0.17) (Table 2).

Table 2.

^aMultivariable Cox regression analysis for overall and Event-free survival

Variable	Category	HR (95% CI), OS	P-value	HR (95% CI), EFS	P- value
Tumor size	≤ 8 cm	0.13 (0.028–0.59)	0.045	0.28 (0.083–0.925)	0.037
	> 8 cm	1 (Reference)	_	1 (Reference)	_
Surgery as component of treatment	Yes	0.22 (0.06- 0.83)	0.025	0.2 (0.07–0.7)	0.01
	No	1 (Reference)	_	1 (Reference)	_
Stage of disease	Localized	0.26 (0.103–0.655)	0.02	0.33 (0.1–0.8)	0.015
	Metastatic	1 (Reference)	_	1 (Reference)	_
Pelvic location (for metastatic disease)	Yes	3.04 (1.26 −7.4)	0.014	2.08 (1.2—4.4)	0.05
	No	1 (Reference)	_	1 (Reference)	_
Number of cycles of CT (for localized)	≥ 14	0.34 (0.07—1.5)	0.16	0.42 (0.12—1.2)	0.17
	< 14	1 (Reference)	_	1 (Reference)	_

HR Hazard ratio, CI Confidence interval, OS Overall survival, EFS Event-free survival, CT Chemotherapy

^aMultivariable analysis is exploratory due to the limited number of outcome events relative to covariates; results are hypothesis-generating

Discussion

To our knowledge, this is the first study from Sub-Saharan Africa and from Ethiopia to report on the treatment patterns and survival outcomes of patients with Ewing sarcoma. This study therefore represents the first in the region to provide critical insight into the clinical characteristics, management, and outcomes of patients with Ewing sarcoma.

In our study, among the entire cohort, 73.8% received treatment with curative intent and 26.2% received palliative treatment. The type of treatment modality used varied by stage of disease at diagnosis. For localized disease, the goal was curative in 94% of patients, whereas all patients with metastatic disease received palliative treatment. Palliative treatment in 6% of localized cases was noted during analysis. Although no structured data were collected, clinical observation suggests this was likely due to challenging tumor sites (spine/pelvis) and limited radiotherapy access (delays, machine downtime). The use of palliative treatment in all metastatic cases was likely related to limited access to surgical expertise and intensive care resources required for complex metastatic management.

Among patients treated with curative intent, the most common treatment modalities were chemotherapy combined with surgery (42.2%), followed by chemotherapy combined with radiotherapy (40%), and trimodality therapy (chemotherapy, surgery, and radiotherapy) (17.8%). Adherence to standard treatment guidelines was 16.7% for localized and 20% for metastatic disease. When compared with reports from other settings, including low- and middle-income countries, the level of treatment protocol adherence in our cohort was relatively low. For example, a study from Egypt by Morsy et al. (2021) reported that, 75.3% of patients received treatment in accordance with standard treatment recommendations [13]. Although no structured data were collected, factors contributing to low adherence likely included long travel distances, financial constraints, and lack of intensive supportive care for treatment-related toxicity.

In our study, the median overall survival (OS) and event-free survival (EFS) in the entire cohort were 45.0 and 24.0 months, respectively. The Kaplan–Meier estimated 3-year OS and EFS rates for the entire cohort were 68% and 40%, respectively. Survival differed significantly by stage at diagnosis. Patients with localized disease had better outcomes, with a median OS not reached and a median EFS of 32.0 months, whereas those with metastatic disease had a median OS of 23.0 months and a median EFS of 12.0 months. The 3-year OS and EFS rates were 89% and 76% for localized disease versus 10% and 0% for metastatic disease. The differences for both OS and EFS reached statistical significance (log-rank p < 0.001). The relatively high 3-year OS in localized disease, despite low treatment adherence, may be partly explained by the high rate of loss to follow-up (9.8%), which can lead to overestimation of survival since patients were censored at last contact and may have experienced unrecorded events. Additionally, the small number of events may further contribute to this overestimation.

A sensitivity analysis treating loss to follow-up as events reduced the 3-year overall survival from 68 to 55% and the 3-year event-free survival from 40 to 33%, suggesting that the primary analysis may underestimate adverse outcomes in this resource-limited setting.

Exploratory subgroup analysis suggested better survival among pediatric patients compared with adults (3-year overall survival: 73% vs. 62%); however, the difference was not statistically significant (p = 0.257). Given the small sample size (n = 24) and the limited number of events in the pediatric subgroup, these findings should be interpreted cautiously, and this analysis should be considered hypothesis‑generating rather than definitive.

Smaller tumor size (≤ 8 cm) was the strongest favorable prognostic factor of OS and EFS followed by surgical intervention as part of curative-intent treatment for patients with localized disease. However, this finding should be interpreted with caution, as the observed benefit of surgery may reflect indication bias, since surgical intervention was likely more frequently performed in patients with more favorable tumor characteristics rather than representing an independent causal effect. For patients with metastatic disease, pelvic location of the primary tumor was independently associated with a statistically significant poorer prognosis. Completion of the standard number of chemotherapy cycles was associated with a non-significant trend toward improved OS and EFS in patients with localized disease treated with curative intent. However, the results of the multivariable analysis should be interpreted with caution due to the limited number of events and small sample size, which increase the risk of overfitting. This may affect the stability and generalizability of the estimated associations. We also acknowledge that treatment heterogeneity in our cohort may introduce confounding and should be considered when interpreting these findings.

The Finding from our study is consistent with the result from low- and middle-income countries (LMIC). Egyptian study done by Morsy et.al. (2021), reported a 3-year OS and EFS of 42.5% and 50%, respectively across entire cohort of study. Direct cross-study comparisons are limited by differences in patient populations. Notably, metastatic disease at diagnosis was 30% in Morsy et al. versus 21% in ours study. Small tumor size and surgical intervention were identified as favorable prognostic factors for OS and EFS in this Egyptian study. However, unlike the Egyptian study, response to neoadjuvant chemotherapy was not statistically significant factors in this study [13]. Factors responsible for these differences may be non- adherence to standard treatment, which is 75% in the Egyptian study and less than 17% in our study. Finding of our study also comparable to the findings from studies conducted in North east India and Iranian study [17, 18].

In the St. Jude Children’s Research Hospital study review, the 3-year OS and EFS were about 70% and 58%, respectively which is higher than our finding. This study had a larger sample (220 patients) and included only pediatric patients, Unlike the St. Jude study, our cohort included adult patients, which adds value to our work but may partly explain the differences in outcomes. In addition to this, a radical treatment approach was applied for the majority of the patients with metastatic disease. However, size of the tumor, location, and stage of the disease were the independent predictors of outcome in that study, which is similar to our finding [19].

Conclusion

In conclusion, this first report from Ethiopia and potentially from Sub-Saharan Africa demonstrates that survival outcomes of patients with Ewing sarcoma are comparable to those in other resource-limited countries. However, outcomes remain substantially inferior to those in high-income settings. Disease stage at diagnosis, smaller tumor size, and surgical intervention as part of curative-intent treatment were key favorable prognostic factors, whereas pelvic primary tumors were associated with poor outcomes in metastatic disease. However, the multivariable analysis should be interpreted cautiously due to small sample size and few events, with risk of overfitting and limited generalizability. Treatment heterogeneity may also introduce confounding. In addition, this study identified low adherence to standard treatment guidelines, limited completion of recommended chemotherapy cycles, and the exclusive use of palliative care for metastatic disease. Therefore, strengthening adherence to standardized treatment guidelines, promoting early diagnosis, and improving access to comprehensive multimodal therapy are essential to improve outcomes in resource-limited settings. We acknowledge that comprehensive structured data on reasons for non-adherence were not collected; therefore, this limitation should be addressed in future prospective studies.

Limitations of the study

This study has several limitations inherent to its retrospective, single-center design and relatively small sample size, which may limit the generalizability of the findings. The inclusion of both pediatric and adult patients introduces heterogeneity in treatment protocols, dosing schedules, and clinical outcomes. In addition, multivariable analyses should be interpreted with caution due to the limited number of events and potential risk of overfitting; the model included multiple predictors relative to the number of events, making the analysis exploratory. Findings should be validated in larger prospective cohorts. Treatment heterogeneity and unmeasured confounders may also have influenced the results. Furthermore, structured data on reasons for non-adherence to standard treatment and deviations from guideline-based care were not systematically collected, limiting deeper exploration of these factors. Finally, due to the lack of immunohistochemical or molecular confirmation, diagnostic misclassification cannot be excluded; interpret findings with caution.

Authors’ contributions

Conception/design: Tegene Gizaw Provision of study material or patients: Tegene Gizaw, Birhanu Tesfaye, Adugna Fikadu, Gashaw Arega and Sonia worku Collection and/or assembly of data: Tegene Gizaw, Birhanu Tisfaye, Feleke Hailemariam, Damena Teshome Data analysis and interpretation: Tegene Gizaw, Birhanu Tesfaye, Adugna Fikadu, Gashaw Arega, Sonia worku, Feleke Hailemariam and Damena Teshome Manuscript writing: Tegene Gizaw, Birhanu Tesfaye, Sonia worku Manuscript editing;- Tegene Gizaw, Birhanu Tesfaye, Adugna Fikadu, Gashaw Arega, Sonia worku and Damena Teshome Final approval of manuscript: All authors Accountable for all aspects of the work: All authors.

Funding

The author(s) declare that no financial support was received for the research, authorship, and/or publication of this article.

Data availability

The datasets analysed during the current study are available from the corresponding author on request.

Declarations

Ethics approval and consent to participate

Ethical approval was obtained from Research and Ethics Committee of Addis Ababa University, College of Health Sciences, Ethiopia before initiating the study. The study was conducted in accordance with the principles of 1964 Helsinki declaration and its later amendments. Due to the retrospective nature of the study and the use of anonymized patient records, the requirement for informed consent to participate was waived by Research and Ethics Committee of Addis Ababa University, College of Health Sciences. The authors ensure that the rights of the patients are not violated in any manner.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.