Clinical outcome of children and adults with localized Ewing sarcoma

Impact of Chemotherapy Dose and Timing of Local Therapy


  • We thank Rupinder Chera, Vivian Tsung, Anthony Griffin, Uri Edell, Peter Chung, David Hodgson, and Derek Stephens.



As Ewing sarcoma (EWS) can affect children and adults, these patients can be treated at either a pediatric or an adult institution. This study investigated whether differences in therapeutic strategy undertaken in pediatric and adult specialty sarcoma centers correlated with clinical outcome.


Data from patients with localized EWS treated between 1990 and 2005 at tertiary care pediatric and adult institutions were reviewed.


Fifty-three patients (24 adult and 29 pediatric) were treated. Pediatric patients received a median of 16 cycles of chemotherapy comprised of doxorubicin, vincristine, cyclophosphamide, ifosfamide, and etoposide. Adult patients received a median of 10 cycles of treatment, and a significantly lower total cumulative dose of ifosfamide and cyclophosphamide (P < .0001). There was no difference noted with regard to the total dose of doxorubicin, or in the type of local therapy offered (surgery or radiotherapy, vs both). However, local therapy occurred earlier in pediatric patients compared with adults (3.7 months vs 7.4 months; P = .0003). The 3-year event-free survival (EFS) rate in pediatric and adult patients was 70% ± 9% and 43% ± 13% (P = 0.1), respectively. The 3-year overall survival rate was 81% ± 7.7% and 59% ± 12% (P = .02) for pediatric and adult patients, respectively. Factors found to be significantly associated with EFS on univariate analysis included pelvic site, cyclophosphamide dose, and time to local therapy. On multivariate analysis, only pelvic disease (hazard ratio [HR] 4.26; P = .018) and time to local therapy (HR, 1.19; P = .002) were found to be significant.


Adults with localized EWS have an inferior outcome compared with pediatric patients. This difference may be related to lower doses of alkylating agents and the timing of local therapy. Cancer 2010. © 2010 American Cancer Society.

The Ewing family of tumors are comprised of Ewing sarcoma (EWS), extraskeletal EWS, and primitive neuroectodermal tumor (PNET) of bone, soft tissue, and the chest wall (Askin tumor). EWS is the second most common primary bone malignancy in childhood and adolescence, with a yearly frequency in the population younger than 20 years of approximately 2.9 per million.1 The peak age for EWS is 15 to 17 years,2 and consequently, patients with this tumor are often seen at both adult and pediatric institutions.

The impact of age on the prognosis of EWS remains controversial. In certain series, older age has been associated with an inferior clinical outcome,3-6 yet others have been unable to demonstrate a significant difference based on age alone.7-9 The impact of age can be confounded by a greater proportion of large pelvic primary tumors and more advanced disease in adult patients,4 although other specific biological differences based on age may also play a role. To our knowledge, there have been no studies to date to systematically document treatment differences between adult and pediatric centers to account for differences in outcome.

The approach to therapy for pediatric patients is driven by clinical trial-based protocols, whereas adult treatment is often institution-specific. Some adult centers follow pediatric protocols, whereas concerns regarding tolerance to therapy have influenced other centers that treat adults to modify their regimens to better suit this older population. Thus, the ideal treatment strategy for adults with EWS remains undefined, and the lack of clinical trials available through adult cooperative groups for newly diagnosed patients contributes to heterogeneity in treatment.

The University of Toronto has a large tertiary care pediatric oncology program and a separate tertiary referral center for adult sarcoma. Differences in therapeutic strategy used at these institutions provide an opportunity for comparative analysis. We sought to compare the treatment-related factors that impacted outcome in patients with localized EWS, adjusting for age.


After Research Ethics Board approval was obtained from The Hospital for Sick Children (Sickkids) and Mount Sinai Hospital (MSH), hospital records of patients with newly diagnosed localized EWS from 1990 to 2005 were retrospectively reviewed. Patients with recurrent or metastatic disease at presentation or those who did not receive all systemic therapy at these institutions were excluded. Information regarding tumor site, stage, chemotherapy drugs and doses, timing of chemotherapy, type and timing of local therapy (radiotherapy [RT] or surgery), and outcome was recorded. Data on tumor size were not uniformly available.

Data were collected for 29 patients from Sickkids (pediatric) and 24 patients from MSH (adult). At both centers, initial staging investigations included computed tomography (CT) scan of the chest, technetium-99 total body bone scan, magnetic resonance imaging (MRI) of the primary tumor, and bilateral bone marrow aspirations and biopsies. The chemotherapy agents and doses used for all patients at both hospitals included vincristine (maximum dose 2 mg), doxorubicin (75 mg/m2), and cyclophosphamide (1.2 g/m2) [VDC] alternating with etoposide (100 mg/m2 × 5 days) and ifosfamide (1.8 g/m2/day × 5 days) [IE] every 3 weeks. At Sickkids, before 1996, patients were treated with the intent of receiving 10 cycles of chemotherapy (5 cycles of VDC and 5 cycles of IE), with local therapy planned after cycle 4 (n = 4). From 1996 onward, patients at Sickkids were treated according to the Intergroup Study 0154, through the Children's Oncology Group (COG).10 The standard arm of this study included 5 cycles of VDC and 4 cycles of VC alternating with IE (8 cycles), for a total planned 17 cycles and the investigational arm contained dose-intense alkylating agent for a total of 11 cycles. Local treatment was planned after 12 weeks of therapy on both arms.10 Three patients from Sickkids were registered on this study: 1 on the standard arm and 2 on the investigational arm. After closure of the study, all patients were treated per the standard arm with 17 cycles of therapy. There were no clinical trials open at MSH for patients with newly diagnosed EWS during the study period, and no patients at MSH were registered on COG studies. All MSH patients were treated with VDC (total of 5 cycles) alternating with IE (total of 5 cycles), for a total of 10 planned cycles of therapy, regardless of age. The timing of local therapy was at the discretion of the multidisciplinary treating team. Information regarding treatment at recurrence was not collected.

At both institutions, local treatment was by surgical resection by orthopedic oncologists whenever possible, with radiotherapy reserved for lesions deemed to be unresectable, or for patients with microscopic residual disease after surgery. At Sickkids, RT was delivered as previously described.6 At MSH, patients who received RT alone received 5000 centigrays (cGy) in 25 fractions to the entire medullary cavity to cover MRI bone marrow changes plus a 2-cm margin, followed by a boost to 6000 to 6600 cGy in 30 to 33 fractions. Similarly, patients with microscopic residual disease postoperatively received 5000 cGy in 25 fractions to the tumor bed plus a 5-cm margin, and a boost to tumor bed plus 2-cm margin to a maximum of 6600 cGy in 33 fractions.

The date of local treatment was defined as the date of definitive surgery or RT. If a patient received both modalities of treatment, the date of local control was defined by whichever occurred first. Time to local therapy was defined as the time from the initiation of chemotherapy to local treatment. Progressive disease was defined as evidence of new disease or documented radiologic progression during or up to 3 months after completion of all therapy. Recurrence was defined as any evidence of new disease >3 months after the completion of therapy. Dose intensity of doxorubicin chemotherapy was calculated as total dose of doxorubicin/(date of last cycle of doxorubicin + 21 days − date chemotherapy started)/7, as previously described.11 Similar calculations were performed for cyclophosphamide and ifosfamide.

Statistical Analysis

Medians and means were used to estimate central values of the variables considered. Ranges and standard deviations were used to estimate the spread of the data. For the comparison of medians, the median test was used. Chi-square tests were used to assess cross-classification of categorical variables. Survival was calculated from the date of diagnosis to death or last patient follow-up. Kaplan-Meier survival curves were used to estimate proportion surviving and the log-rank test was used to compare subgroups. For univariate and multiple regression, Cox proportional hazards were used. If variables were significant at the .025 level on univariate analysis, then they were included in the multiple regression. All analyses were performed using SAS statistical software (version 9.1; SAS Institute, Inc, Cary, NC).


Patient Characteristics

The demographics and clinical characteristics of the 53 evaluable patients are depicted in Table 1. The median age for the 29 pediatric patients was 13.4 years (range, 0.29-16.2 years), and was 26.1 years for the 24 adult patients (range, 16.7-66.5 years). There were 2 patients aged <18 years who were treated at the adult hospital, but they were considered adults for the purpose of this report. Two pediatric (6.9%) and 6 adult (25%) patients had primary pelvic disease (P = .07).

Table 1. Demographic and Treatment Characteristics of Pediatric and Adult Patients
CharacteristicPediatric Patients No. (%)Adult Patients No. (%)P
Median [Range]Median [Range]
  1. NS indicates not significant.

Median age, y13.4 [0.29–16.2]26.1 [16.7-66.5] 
Site of primary tumor   
 Lower extremity13 (45)9 (37.5) 
 Upper extremity1 (3.4)1 (4.2) 
  Spine5 (17.2)3 (12.5) 
  Rib2 (6.9)2 (8.3) 
  Pelvis2 (6.9)6 (25) 
  Other4 (13.8)1 (4.2) 
 Soft tissue2 (6.9)2 (8.3) 
Local therapy   
 Radiation10 (35)7 (29) 
 Surgery16 (55)11 (46) 
 Surgery+radiation3 (10)6 (25) 
 Median no. of cycles16 [10-17]10 [6-10] 
  Total dose of doxorubicin, mg/m2370 [111-490]375 [295-375].14
  Dose intensity of doxorubicin, mg/m2/wk9.77 [5.3-12.1]12.0 [10.75-15.2].0001
  Total dose of ifosfamide, g/m269.3 [14.3-72.2]44.8 [8.8-45.1].0001
  Dose intensity of Ifosfamide, mg/m2/wk27.2 [6.58-39.0]26.7 [5.41-46.8]NS
  Total dose of cyclophosphamide, g/m210.5 [2.7-17.3]6 [4.7-6].0001
  Dose intensity of cyclophosphamide, mg/m2/wk3.79 [1.97-11.24]4.0 [3.5 – 5.75]NS
 Mo to local treatment3.38 [0.85-14.9]7.63 [3.68-20.9].0003


All patients received chemotherapy. For patients who completed planned therapy (ie, excluding those who experienced disease progression while receiving therapy), the median number of cycles of therapy was 16 and 10, respectively, for pediatric and adult patients (P <.0001). Correspondingly, the median total dose of ifosfamide and cyclophosphamide was higher in pediatric than adult patients (Table 1). The median dose of doxorubicin delivered was similar among pediatric and adult patients; however, the dose intensity of doxorubicin was higher in adult patients compared with pediatric patients (P = .0001). There was no difference in dose intensity of ifosfamide or cyclophosphamide between pediatric and adult patients. Pediatric patients who received RT at Week 12 would have had an interruption in the delivery of doxorubicin. Dose intensity of doxorubicin did not differ between pediatric and adult patients who received surgery alone for local control (data not shown).

Local Therapy

The form of local therapy used did not differ between pediatric and adult patients (P = .36); 16 of 29 (55%) pediatric patients and 11 of 24 (46%) adult patients underwent surgery only; 3 of 29 (10%) and 6 of 24 (25%), respectively, were treated with surgery and RT; and 10 of 29 (35%) and 7 of 24 (29%), respectively, received RT only. The median dose of RT was 4500 centi-Gray (cGy) (range, 4500-6940 cGy) and 60 Gy (range, 5000-6600 cGy) in pediatric and adult patients, respectively. None of the patients required amputation. The time to local therapy was shorter in pediatric patients compared with all adult patients (3.7 months vs 7.4 months; P = .0003) and compared with adults with nonpelvic primary tumors (3.7 months vs 6.0 months; P = .003).


The median follow-up time for surviving patients (n = 37) was 3.9 years (range, 0.67-10.1 years). The estimated 3-year event-free survival (EFS) and overall survival (OS) rates for all patients are listed in Table 2. After treatment for localized disease in pediatric and adult patients, the 3-year EFS rate was 70% ± 9% and 43% ± 13% (P = .1) and the OS rate was 81% ± 7.7% and 59% ± 12% (P = .02), respectively. Nine of 29 (31%) pediatric patients experienced an event (recurrent or progressive disease [PD]), 3 of whom (33%) were alive in second remission at a median follow-up of 3.4 years (range, 2.1-6.36 years). In comparison, 10 of 24 (42%) adult patients developed disease recurrence or had PD, and all had died at the time of last follow-up. The median time to disease recurrence/PD was 1.59 years and 1.43 years, respectively, in pediatric and adult patients. The estimated 3-year OS and EFS rates for all patients are listed in Table 2. In patients with disease recurrence/PD, the median time to local therapy was 6.2 months (range, 2-21 months) compared with 3.75 months (range, 3.75-9.07 months) in 31 patients without disease recurrence/PD.

Table 2. Event-Free and Overall Survival of Pediatric and Adult Patients
  3-Year EFSPa3-Year OSPa
  • EFS indicates event-free survival; OS, overall survival.

  • a

    Comparing pediatric versus adult patients.

LocalizedAdult43±13 59±12 
Localized nonpelvicAdult48±17 67±13 

Univariate and Multiple Regression Analyses

Univariate analysis demonstrated that time to local therapy, total dose of cyclophosphamide, and primary tumor site in the pelvis were statistically significantly associated with EFS (Table 3). On multivariate analysis, primary pelvic tumor site (HR, 4.26; P = .018) and time to local therapy (HR, 1.19; P = .002) were found to be significant predictors of EFS (Table 4).

Table 3. Univariate Analysis of Prognostic Features for EFS in Pediatric and Adult Patients
FeatureHR95% CIP
  1. EFS indicates event-free survival; HR, hazard ratio; 95% CI, 95% confidence interval.

Mo to local treatment1.131.04–1.23.003
Total dose of cyclophosphamide0.770.61-0.97.03
Total dose of ifosfamide0.980.95-1.0.077
Dose intensity of doxorubicin1.300.978-1.73.071
Table 4. Multivariate Analysis of Prognostic Features for EFS in Pediatric and Adult Patients
ParameterHR95% CIP
  1. EFS indicates event-free survival; HR, hazard ratio; 95% CI, 95% confidence interval.

Pelvic primary tumor4.261.28-14.1.018
Mo to local treatment1.191.1-1.31.002


We compared the treatment-related factors that affected outcome of localized EWS in adult and pediatric patients treated at 2 tertiary care institutions. The major advantage of this study was the homogeneity in chemotherapy delivered to the adult patients, because only 1 medical oncologist was involved over the study period. Furthermore, both the pediatric and adult institutions are large tertiary care centers with dedicated orthopedic, medical, and radiation oncologists with expertise in the management of sarcoma. The current study is subject to the inherent bias of a retrospective chart review, with small numbers of patients in 2 groups that were very different in their characteristics. Furthermore, specific limitations in data collection were encountered. Information regarding toxicities, including febrile neutropenia, was not collected because many patients were managed at local community hospitals for these events. Unfortunately, tumor volume was also not recorded due to the difficulty in locating the reports of initial MRI scans from patients treated in earlier years.

Adult patients received 5 cycles of IE alternating with 5 cycles of VDC, resulting in a 3-year OS rate of 59% for patients with localized disease. If patients with pelvic primary tumors were excluded, OS improved to 67%. The survival results in the current study compare favorably with many previous reports of localized EWS in adults treated with conventional chemotherapy, and a median age of >19 years with 3-year to 5-year OS rates of 40%,12 49%,13 52%,7 54%,8 59%,14 and 60%.15

The outcome of patients treated at the pediatric institution in the current study was better than that of patients treated at the adult institution. We compared the treatment-related factors that may have contributed to this difference, independent of age. Adults received fewer cycles of chemotherapy compared with children, with a consequential difference in the total dose of alkylating agent. On univariate analysis, the total dose of cyclophosphamide significantly impacted EFS, whereas the total dose of ifosfamide did not. The INT-0091 study6 demonstrated that the addition of IE improved survival in pediatric patients with localized EWS, but was not sufficiently powered to determine if this was beneficial to patients aged >18 years. Because there is no a priori reason to expect different responses in young adults, inclusion of these agents is reasonable. In the European cooperative group study EICESS-92, in which 23% of patients were ages 20-35 years, ifosfamide was not found to be superior to cyclophosphamide in patients with localized disease of <100 mL, and the improvement in EFS with the addition of etoposide to ifosfamide-based therapy was not statistically significant in patients with large-volume disease.16 The treatment strategy in the current study of using 10 cycles of chemotherapy in adults offered an acceptable outcome; however, it is conceivable that increasing the dose of alkylating agents by increasing the total number of cycles of chemotherapy may further improve survival.

Dose intensity of chemotherapy over time was not found to significantly contribute to outcome in the current report. Recently, a randomized trial demonstrated that increasing the dose intensity of chemotherapy by increasing the total doses per cycle administered 3-weekly did not improve survival in patients with localized EWS.10 However, interval compression of chemotherapy given twice-weekly improved EFS in patients less than age 18 from 65% to 76% (P = .028).17 The inclusion of adults with EWS in future studies is important to ensure that trials are powered sufficiently to extrapolate conclusions to older patients.

On multivariate analysis, a longer time interval between the initiation of treatment and definitive local therapy unfavorably impacted EFS, adjusting for age. Current pediatric protocols for EWS dictate that local therapy should begin at 12 weeks from the initiation of chemotherapy.6 By contrast, in adults, the timing of local therapy is at the discretion of the multidisciplinary oncology team, and usually depends on tumor size, location, and response to neoadjuvant chemotherapy. As a result, the decision as to when to offer local therapy for large pelvic tumors, often involves a detailed multidisciplinary discussion among medical, radiation, and orthopedic oncologists. In patients with chemo-responsive disease, it is unclear whether the maximal benefit of chemotherapy should be exploited to facilitate the best possible surgical resection or radiation of the smallest tumor volume.

The delivery of concomitant RT and chemotherapy is sometimes difficult in older patients, especially for axial sites. Recovery from surgery for axial tumors is also often prolonged, rendering it difficult to resume chemotherapy in a timely manner. Furthermore, there may be merit in maintaining dose intensity of chemotherapy in a patient whose tumor is responding clinically. Such a response might persuade clinicians to complete all 10 cycles, both to avoid risking an interruption with local therapy as well as to reduce total tumor volume. However, our data suggest that delaying local therapy may not be beneficial and that it should be undertaken earlier in the treatment course. In a study reviewing tolerance of chemotherapy in older versus younger patients treated with the same chemotherapy regimen, older patients were more likely to have grade 4 hematologic toxicity, compared with younger patients; however, this did not translate into a difference in frequency of dose delays or mean dose intensities.14 In the current study, the dose intensity of doxorubicin was in fact higher in adult patients than in pediatric patients, due to the lack of interruption for RT. Adult patients treated at MSH received higher doses and larger volume of RT compared with pediatric patients treated at Sickkids, mainly due to greater concern for recurrent disease and less concern for growth-related complications and second malignancies. Careful review of the toxicity experienced by the adult population is warranted, and if feasible, earlier local therapy should be considered. It is conceivable that a delay in local therapy will allow for the development of resistant disease, although this has not been previously systematically documented. However, a previous review plotted the duration of chemotherapy before RT (in patients who were treated with RT alone) against OS and found that a significant association was identified between a delay to RT and reduced survival.18

One of the strongest prognostic factors associated with a poor outcome of patients with localized EWS is primary disease in the pelvis. The increased prevalence of pelvic EWS in adult patients is often cited as 1 of the main contributors to poor outcome. The optimal strategy of addressing local therapy in pelvic disease with surgery, RT, or both, has been the subject of many reports, yet remains highly controversial.19-22 Adults in the current study were more likely to present with primary pelvic disease (6 of 24 patients vs 2 of 29 patients; P = .07), and also fared poorly. Our results suggest that introducing local therapy earlier, even in patients with large pelvic primary disease, may improve outcome.

In summary, adults with localized EWS treated at an adult institution can achieve reasonable outcomes, although OS remains inferior to pediatric patients. There may be value to offering local therapy earlier, and perhaps revisiting the total dose of alkylating agents planned as part of adult EWS chemotherapy regimens. For a disease that is just as common in young adults as in children, a concerted effort is required to capture these patients in clinical trials to systematically address questions regarding biological factors and therapeutic strategies to improve survival. Future initiatives through the COG to bring trials in EWS to the clinical trials support unit (available at: may provide a platform for adult patients to access prospective studies.


Supported by the AFLAC Adolescents and Young Adult Children's Oncology Group Grant 15961.