Although more patients with newly diagnosed Ewing sarcoma get into remission with modern multidisciplinary treatment approaches and are potential long-term survivors, relapse remains a considerable clinical obstacle against maintaining a favorable frontline therapy response. About 25% of patients with an initially localized disease will relapse after first-line treatment, and the recurrence rate is even higher in an initially metastatic disease [10].

Several prognostic factors have been identified, which can help in establishing future therapy plans. Several salvage chemotherapy regimens have been used with variable responses in patients with recurrent Ewing sarcoma, but the superiority of one regimen over another has not yet been proven, and toxicities in those previously heavily treated patients should be considered [6]. It has not yet been established what the optimal number of chemotherapy cycles is for relapsed patients, which depends in part on response, tolerance of therapy, toxicities, and quality of life. Likewise, despite the local control that is usually offered, it is mostly non-feasible, especially for widespread metastatic recurrences [6].

In the present retrospective study of 50 patients with recurrent/progressive ES, we identified some risk factors that have effects on survival outcomes after disease recurrence/progression. Only 20% of patients who had disease recurrence achieved a second remission that was maintained in almost half of them with a median OS of 7.5 months. The age at presentation, gender, and initial disease site did not affect survival outcome. In concordance with these findings, previous studies did not demonstrate an effect of the initial disease site on survival outcome [11, 12].

Primary local control in Ewing sarcoma is achieved through surgery and/or radiotherapy. In the present study, patients who had their primary local control by both surgery and RT had an inferior outcome compared to patients who had local control by either surgery or RT alone. According to Becker et al., a significantly better EFS was associated with local control by surgery alone (71.7%) followed by postoperative RT (64.1%) rather than definitive RT alone (30%) (p = 0.009). While in tune with our study, the cumulative incidence of local and distant failure in the Becker et al. study at 2 and 5 years was 25% for surgery alone, 11% for radiotherapy alone, and 16.7% for surgery plus radiotherapy, but this did not reach statistical significance (p = 0.64) [11]. This might be explained in part by the fact that patients who had local control by both modalities (surgery and RT) in our study were those patients whose tumors showed poor response to chemotherapy [TN: tumor necrosis (< 90%)] and/or surgical margins positive for tumor infiltration, and might also be due to interrupted courses of RT.

A COG study conducted by Womer et al. [13] on a large number of Ewing sarcoma patients tested whether chemotherapy intensification through interval compression could improve the outcome. They concluded that for localized Ewing sarcoma, chemotherapy administered every 2 weeks is more effective than chemotherapy administered every 3 weeks, with no increase in toxicity. Interval-compressed therapy had no apparent effect on local recurrence alone but decreased the incidence of distant and combined relapses. The 5-year OS and EFS were significantly better in every 2-week regimen [13]. Our study did not show a significant effect of interval compression of first-line CTH on survival outcome, possibly due to the small number of cases, yet it should be noted that the majority of our cases (n = 41) received their chemotherapy every 3 weeks.

Patients with combined local and distant relapses have the worst outcomes, while those with isolated local recurrences appear to do better, as shown in our study and other studies [4]. This is understandable given that these patients may have a lower total burden of disease at recurrence and may be amenable to further local therapies in addition to systemic treatment.

More debatable is the prognostic impact of isolated pulmonary recurrence, as some series have reported that these patients do better than those with other distant site metastases. Stahl et al. [14] and Bacci et al. [4] found a better prognosis for patients diagnosed with pulmonary metastases only at the time of relapse compared to extrapulmonary lesions. Other studies did not find such an association with survival improvement [15], in agreement with the current study. However, the small number of patients with isolated pulmonary recurrence in our study (n = 10) precluded a definitive conclusion.

In multivariate analysis, the time to first recurrence was the only independent predictor of post-recurrence survival outcome. Many studies showed that most recurrences present within 2 years after primary diagnosis, while rarely occurring after the fifth year [4, 14]. In our study, 15 patients (30%) who had a late recurrence (> 2 years from the initial diagnosis) had a better survival outcome compared with 35 (70%) who had recurrences within 2 years. This is consistent with other reports that also confirm the importance of time to recurrence as a prognostic factor affecting the survival outcome of recurrent ES patients [4, 12, 15,16,17,18]. Our study showed that local recurrences occur with higher frequency after 2 years, which may explain in part the favorable prognosis of later recurrences.

A variety of chemotherapy regimens have produced responses in patients with recurrent Ewing sarcoma, with no evident superiority of one over another [6]. The retrospective nature of our study and the small number of patients did not permit an extensive analysis of each salvage regimen separately. However, since the ICE combination reportedly is an effective salvage treatment for patients with recurrent ES [6, 19] and is the most commonly used regimen in our study, we analyzed the impact of this regimen on outcomes versus other regimens used. There was no significant relation between the type of salvage CTH (ICE vs. others) and survival outcome. On the other hand, aggressive local measures along with intensive chemotherapy improved the outcomes compared to chemotherapy alone.

Overall, the prognosis of recurrent Ewing sarcoma is constantly dismal, and the outcome remains poor. The standard approach to their management has not yet been established, and many patients may initially benefit from salvage chemotherapy in terms of reducing symptoms and prolonging time to further progression, but consistent cures remain elusive. Knowledge of prognostic factors that affect the survival of these patients may help guide future therapy decisions.

In conclusion, our study showed that the main factors predictive of improved post-relapse survival in relapsing/progressive ES were prolonged time to first recurrence, local site of recurrence rather than metastatic or combined, relapse rather than progression and salvage CTH combined with surgery and/or RT, compared to CTH alone. The time to the first recurrence was the only independent factor predicting post-relapse survival. There was no prognostic significance for sites of metastasis (isolated pulmonary or other sites), type of salvage CTH (ICE or other regimen), or metastatic status at initial presentation.