The clinical characteristics, novel predictive tool, and risk classification system for primary Ewing sarcoma patients that underwent chemotherapy: A large population-based retrospective cohort study

Abstract

Background: This study aims to determine the independent prognostic predictors of cancer-specific survival (CSS) in patients with primary Ewing sarcoma (ES) that underwent chemotherapy and create a novel prognostic nomogram and risk stratification system.

Methods: Demographic and clinicopathologic characteristics related to patients with primary ES that underwent chemotherapy between 2000 and 2018 were extracted from the Surveillance, Epidemiology, and End Results (SEER) database. CSS was the primary endpoint of this study. First, independent prognostic predictors of CSS identified from univariate and multivariate Cox regression analyses were used to construct a prognostic nomogram for predicting 1-, 3-, and 5-year CSS of patients with primary ES that underwent chemotherapy. Then, calibration curves and receiver operating characteristic (ROC) curves were used to evaluate the nomogram's prediction accuracy, while decision curve analysis (DCA) was used to evaluate the nomogram's clinical utility. Finally, a mortality risk stratification system was constructed for this subpopulation.

Results: A total of 393 patients were included in this study. Age, tumor size, bone metastasis, and surgery were independent prognostic predictors of CSS. The calibration curves, ROC, and DCA showed that the nomogram had excellent discrimination and clinical value, with the 1-, 3-, and 5-year AUCs higher than 0.700. Moreover, the mortality risk stratification system could effectively divide all patients into three risk subgroups and achieve targeted patient management.

Conclusions: Based on the SEER database, a novel prognostic nomogram for predicting 1-, 3-, and 5- year CSS in patients with primary ES that underwent chemotherapy has been constructed and validated. The nomogram showed relatively good performance, which could be used in clinical practice to assist clinicians in individualized treatment strategies.

Keywords: Ewing sarcoma; SEER; cancer-specific survival; chemotherapy; nomogram; risk stratification system.

FIGURE 1

The flowchart for patient selection in this study.

FIGURE 2

Kaplan–Meier survival curves were performed on cancer‐specific survival (CSS)‐related risk variables in primary Ewing sarcoma (ES) patients treated with chemotherapy. (A) age, (B) sex, (C) race, (D) marital status, (E) tumor size, (F) primary site, (G) tumor stage, (H) T stage, (I) N stage, (J) surgery, (K) bone metastasis, (L) brain metastasis, (M) liver metastasis, (N) lung metastasis, and (O) radiation.

FIGURE 3

The prognostic nomogram predicts the 1‐, 3‐ and 5‐year CSS in primary ES patients treated with chemotherapy. Specifically, when such a patient consults about his CSS, we can sum the scores of the obtained independent prognostic predictors to get a total score and draw a vertical line from the total score to the bottom timeline. For example, the primary tumor, a 25 years old male patient with a 150 mm diameter ES, had received chemotherapy with no surgical treatment and no bone metastasis. The corresponding total score of he is 41 (25 years old) + 97 (150 mm of tumor size) + 33 (no bone metastasis) + 33 (no surgery) = 204, and the corresponding death risk possible at 1‐, 3‐, and 5‐ year are 0.0947, 0.351, and 0.537, respectively, while the corresponding CSS of the patient at 1‐, 3‐, and 5‐ year are 0.9053, 0.649, and 0.463, respectively.

FIGURE 4

The calibration curves of the nomogram were used to predict CSS in primary ES patients treated with chemotherapy at 1‐(A), 3‐(B), and 5‐(C) year in the training cohort and 1‐(D), 3‐(E), and 5‐(F) year in the validation cohort, respectively.

FIGURE 5

The 1‐, 3‐, and 5‐year receiver operating characteristic curves of primary ES patients treated with chemotherapy in the training (A) and validation (B) cohorts.

FIGURE 6

Comparison of prediction accuracy between the constructed novel nomogram and each CSS‐related independent prognostic predictor in primary ES patients treated with chemotherapy at 1‐(A), 3‐(B), and 5‐(C) year in the training cohort and 1‐(D), 3‐(E), and 5‐(F) year in the validation cohort, respectively.

FIGURE 7

The decision curve analysis of the constructed novel nomogram was used to predict the 1‐(A), 3‐(B), and 5‐(C) year CSS in the training cohort, and the 1‐(D), 3‐(E), and 5‐(F) year CSS in the validation cohort for primary ES patients treated with chemotherapy.

FIGURE 8

Kaplan–Meier survival analysis and log‐rank tests were performed to compare CSS of primary ES patients treated with chemotherapy in the three risk subgroups in the training cohort (A) and validation cohort (B). The high‐risk subgroup had a worse prognosis than those in the low‐risk subgroup.