Skip to main content
Fig. 4 | Cardiovascular Diabetology

Fig. 4

From: A novel 6-metabolite signature for prediction of clinical outcomes in type 2 diabetic patients undergoing percutaneous coronary intervention

Fig. 4

External validation of the 6-metabolite signature by targeted metabolite analyses in a prospective cohort of 301 type 2 diabetic patients undergoing PCI. Comparisons of the absolute concentrations of ADPR (A), 1-MNAM (B), D-Ribose 5-phosphate (C), NAM (D), L-Tryptophan (E), and PC(36:2) (F) between patients with and without the incidence of MACEs at 1-year after PCI. G The improvements in AUCs of the 6-metabolite models based on random forest or XGBoost compared with the logistic regression model. H The greater clinical benefit of 6-metabolite model based on random forest or XGBoost compared with the logistic regression model. I The ROC analysis of the 6-metabolite models for predicting target lesion failure at 1 year after PCI. ADPR adenosine diphosphate ribose; 1-MNAM 1-Methylnicotinamide; PC phosphatidylcholine; NAM nicotinamide; XGBoost gradient-boosted decision tree

Back to article page