Predicting overall survival in synchronous metastatic nasopharyngeal carcinoma using a stacking ensemble machine learning model: a multicenter retrospective study - Summary - MDSpire
Advertisement
Predicting overall survival in synchronous metastatic nasopharyngeal carcinoma using a stacking ensemble machine learning model: a multicenter retrospective study
To develop and validate a stacking ensemble machine learning model for predicting 3-year overall survival (OS) in patients with synchronous metastatic nasopharyngeal carcinoma (smNPC).
Approach:
Study Design: A multicenter retrospective study involving 413 patients from three institutions, with a training set of 289 and an internal validation set of 124.
Predictor Variables: Candidate predictors included demographic, histopathologic, serological, and treatment-related variables.
Model Development: Six base learners were trained, and stacking ensembles were constructed and compared using various validation techniques.
Key Findings:
The stacking ensemble achieved an internal held-out validation AUC of 0.8358 for predicting 3-year OS.
The 3-year death event rate was 60.0% overall, with 59.9% in the training set and 60.5% in the internal held-out validation set.
Stacking showed significantly higher AUC than GBDT, AdaBoost, and hard voting after multiplicity adjustment, while differences versus other comparators were not significant.
Overfitting diagnostics indicated a training AUC of 0.9150, an optimism gap of 0.0792, and an optimism-corrected AUC of 0.8415.
Interpretation:
The stacking ensemble model provided an interpretable approach for predicting 3-year OS in smNPC.
Limitations:
Prospective validation of the model is needed.
Conclusion:
The stacking ensemble model supports prognostic assessment in smNPC, although prospective validation is needed.
