A Machine Learning-Based Predictive Model for Predicting Lymph Node Metastasis in Patients With Ewing's Sarcoma.

Area of Science:

• Oncology
• Medical Informatics
• Machine Learning

Background:

• Ewing sarcoma (ES) is a rare bone cancer.
• Lymph node metastasis (LNM) is a critical prognostic factor in ES.
• Accurate prediction of LNM is essential for guiding clinical treatment decisions.

Purpose of the Study:

• To develop and validate a machine learning-based risk prediction model for lymph node metastasis (LNM) in Ewing sarcoma (ES).
• To provide a convenient tool for clinicians to assess LNM risk in ES patients.
• To improve the clinical management of ES by identifying patients at high risk for LNM.

Main Methods:

• Retrospective collection of clinicopathological data from 923 ES patients (SEER database) and 51 external validation patients.
• Application of various machine learning algorithms including random forest (RF), naive Bayes classifier (NBC), decision tree (DT), xgboost (XGB), gradient boosting machine (GBM), and logistic regression (LR).
• Model performance evaluation using 10-fold cross-validation and receiver operating characteristic (ROC) curve analysis with area under the curve (AUC) calculation.

Main Results:

• 13.86% of ES patients had confirmed or unevaluable LNM.
• Race, T stage, M stage, and lung metastases were identified as independent predictors of LNM.
• The random forest (RF) model demonstrated the best predictive performance with AUC ranging from 0.612 to 0.727 in external validation.
• A web-based calculator was developed for clinical application.

Conclusions:

• Clinicopathological data can effectively aid clinicians in identifying LNM in ES patients.
• The developed machine learning models, particularly the RF model, offer robust performance for LNM risk prediction in ES.
• The web-based calculator facilitates the practical application of these predictive models in clinical settings.