Machine learning-based estimation of trunk fat percentage and its association with cardiometabolic risk leveraging two large national cohorts - PubMed
3 months ago
- #trunk fat percentage
- #machine learning
- #cardiometabolic risk
- Developed and validated a machine learning model to estimate trunk fat percentage using anthropometric measures.
- Utilized data from NHANES (n=30,443) for development and CHARLS (n=13,524) for external validation.
- XGBoost model showed superior performance with R² of 0.8509 on the test set.
- Simplified model with five variables (sex, waist circumference, height, weight, age) retained 99.3% accuracy.
- Trunk fat percentage outperformed whole-body fat percentage in predicting cardiometabolic risks.
- Highest AUC improvement observed for diabetes (3.22% relative improvement).
- Average relative improvement in AUC across all endpoints was 2.77%.
- Provides a non-invasive tool for central adiposity assessment in clinical and epidemiological studies.