A novel nomogram and risk classification system predicting the Ewing sarcoma: a population-based study

Abstract

Ewing sarcoma (ES) is a rare disease that lacks a prognostic prediction model. This study aims to develop a nomogram and risk classification system for estimating the probability of overall survival (OS) of patients with ES. The clinicopathological data of ES were collected from the Surveillance, Epidemiology and Final Results (SEER) database from 2010 to 2018. The primary cohort was randomly assigned to the training set and the validation set. Univariate and multiple Cox proportional hazard analyses based on the training set were performed to identify independent prognostic factors. A nomogram was established to generate individualized predictions of 3- and 5-year OS and evaluated by the concordance index (C-index), the receiver operating characteristic curve (ROC), the calibration curve, the integrated discrimination improvement (IDI) and the net reclassification improvement (NRI). Based on the scores calculated with the nomogram, ES patients were divided into three risk groups to predict their survival. A total of 935 patients were identified, and a nomogram consisting of 6 variables was established. The model provided better C-indices of OS (0.788). The validity of the Cox model assumptions was evaluated through the Schönfeld test and deviance residual. The ROC, calibration curve, IDI and NRI indicated that the nomogram exhibited good performance. A risk classification system was built to classify the risk group of ES patients. The nomogram compares favourably and accurately to the traditional SEER tumour staging systems, and risk stratification provides a more convenient and effective tool for clinicians to optimize treatment options.