Development and Validation of a Frailty Risk Prediction Model for Preoperative Non-Small-Cell Lung Cancer Patients: A Cross-Sectional Study - PubMed
3 days ago
- #Machine learning
- #Non-small cell lung cancer
- #Frailty
- Study aimed to develop and validate a frailty risk prediction model for preoperative NSCLC patients using clinical parameters and machine learning.
- 489 preoperative NSCLC patients were enrolled, divided into training (n=342) and validation (n=147) sets.
- Frailty was assessed using the FRAIL scale, with a prevalence of 36.1% for frailty/pre-frailty.
- Key predictors included age, BMI, comorbidity grade, fatigue, walking difficulty, lung function, and TyG index.
- The Light Gradient Boosting Machine (LGBM) model outperformed others with an AUC of 0.965 (training) and 0.807 (validation).
- Top predictors identified were TyG index, comorbidity grade, and maximal voluntary ventilation.
- ML integration with physiological markers improved predictive accuracy over traditional methods.