Age, tumor size, and therapy timing are associated with survival in primary spinal Ewing sarcoma: a nationwide analysis.

BACKGROUND: Primary spinal Ewing sarcoma (EwS) is a rare, aggressive small round cell tumor. While age, tumor size, and metastatic status are established prognostic factors in extremity disease, their impact in spinal primaries remains poorly defined, and the optimal timing of multimodal therapy is uncertain. OBJECTIVE: To evaluate the prognostic significance of age, tumor size, and treatment timing in primary spinal EwS using a large national database. METHODS: The National Cancer Database was queried for patients diagnosed with primary spinal EwS between 2004 and 2017. Demographic, clinical, and treatment variables including surgery, chemotherapy, radiation dose/fractionation, and time to local therapy were collected. Overall survival (OS) was assessed by Kaplan-Meier and log-rank testing. Of the five machine-learning survival models that were trained, Extra Survival Trees demonstrated the highest discrimination (concordance index 0.80, integrated Brier score 0.13) and was used for feature importance analysis via SHapley Additive exPlanations (SHAP). Receiver operating characteristic analyses identified prognostic thresholds for age, tumor size, and initiation of chemotherapy. RESULTS: Among 413 patients (204 pediatric, 219 adult; mean age 24 years), 37% died at latest follow-up. Adults had higher mortality (48% vs. 25%, p < 0.001) and longer delays to chemotherapy initiation (41 vs. 19 days, p < 0.001). SHAP analysis highlighted tumor size (HR 1.01 per mm, p = 0.001), age (HR 1.04 per year, p < 0.001), and metastasis (HR 4.41, p < 0.001) as dominant mortality predictors. Optimal size thresholds were < 64 mm for children and < 46 mm for adults. In adults, surgery with radiation was associated with longer OS in large tumors (HR 0.77, p = 0.044), whereas surgery with chemotherapy was associated with the greatest benefit in small tumors (HR 0.46, p = 0.002). Earlier chemotherapy initiation (< 18 days) was associated with longer survival in non-metastatic and large-tumor subgroups. CONCLUSIONS: Age, tumor size, and timely initiation of chemotherapy are strongly associated with survival in primary spinal EwS. Local therapy associations varied by tumor size and age, but these non-randomized comparisons are subject to confounding by indication. Because this cohort spans 2004 to 2017 and predates uniform adoption of interval-compressed chemotherapy and contemporary radiation practice, these findings may not fully reflect current outcomes. Early multidisciplinary coordination to expedite chemotherapy remains a reasonable target. Of note, these modality comparisons are non-randomized and are subject to confounding by indication.