Key result
XGBoost outperforms logistic regression in predicting LVRR in first-diagnosed DCM, achieving ~0.82 AU-ROC.
Why the study?
There is a lack of an early prediction model for left ventricular reverse remodeling (LVRR) in patients with first-diagnosed dilated cardiomyopathy.
Do tree-based machine learning models like XGBoost improve the early prediction of left ventricular reverse remodeling compared to logistic regression in patients with first-diagnosed idiopathic dilated cardiomyopathy?
Population
104 patients with idiopathic DCM
Comparison
Logistic regression vs random forests vs extreme gradient boosting (XGBoost)
Design
Single-center study
Authors
Loading...
May aid LVRR prediction in new DCM; hypothesis-generating and requires prospective validation before practice change.
Cohort (n=104)
No
Do tree-based machine learning models like XGBoost improve the early prediction of left ventricular reverse remodeling compared to logistic regression in patients with first-diagnosed idiopathic dilated cardiomyopathy?
Absolute Event Rate: 0.8205% vs 0.5909%
p-value: p=0.0119
Tree-based machine learning models like XGBoost can effectively predict early left ventricular reverse remodeling in newly diagnosed idiopathic dilated cardiomyopathy, outperforming traditional logistic regression.
Xie et al. (2021) conducted a cohort in Idiopathic Dilated Cardiomyopathy (n=104). XGBoost predictive model vs. Logistic regression was evaluated on Prediction of left ventricular reverse remodeling (AU-ROC) (95% CI 0.6775-0.9497, p=0.0119). The XGBoost machine learning model predicted left ventricular reverse remodeling in patients with first-diagnosed dilated cardiomyopathy with a significantly higher AU-ROC of 0.8205 compared to 0.5909 for logistic regression.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: