Development and validation of a prognostic score at baseline diagnosis for Ewing sarcoma family of tumors: a retrospective single institution analysis of 860 patients

Introduction: Prognostic scores in Ewing sarcoma including baseline clinical and laboratory characteristics are necessary for pre-treatment risk stratification. In this study, we formulated and validated a prognostic model for baseline risk categorization in Ewing sarcoma. Materials and methods: A retrospective single-institutional study was conducted on Ewing sarcoma patients treated uniformly between January 2003 and December 2018. Baseline clinical/pathological characteristics and survival outcomes were noted from medical records. The cohort was randomised into a derivation and validation cohort. A prognostic score was formulated by including independent prognostic factors from the derivation cohort by multivariable analysis. The prognostic model was validated in the validation cohort along with estimation of its predictive ability. Results: A total of 860 patients were included with 40.3% having baseline metastases. Tumor diameter 0.001, score 2), and total leucocyte count 11000/mm(3) (HR 1.44; P=0.015; score 1) were independent predictors of overall survival in derivation cohort and included for prognostic score calculation. Patients were categorized into low (score 0), intermediate (score 1-3) and high-risk (score 4-5) groups. Harrell's c-indexes of the model were 0.625, 0.622 and 0.624 in the derivation, validation and whole cohort respectively. The timed AUC of ROC of the prognostic score-group for 5-year survival was 0.72, 0.71 and 0.73 in the derivation, validation and whole cohort respectively. Conclusions: We have formulated and validated a prognostic score for Ewing sarcoma incorporating baseline clinical and laboratory parameters, with fair predictive ability for risk stratification and facilitating riskadapted personalized therapy.

Automated summary

A prognostic score for Ewing sarcoma incorporating baseline clinical and laboratory parameters has been formulated and validated in this study. It has fair predictive ability for risk stratification and facilitates risk-adapted personalized therapy.