Predictive modeling for acute kidney injury after percutaneous coronary intervention in patients with acute coronary syndrome: a machine learning approach

Table 2 Models’ evaluation for predicting AKI after PCI

	Pre-procedural features only			All features
	Sensitivity	Specificity	AUC (95% CI)	Sensitivity	Specificity	AUC (95% CI)
Random Forest	75.97%	61.44%	0.740 (0.694–0.783)	81.4%	60.56%	0.775 (0.730–0.818)
Logistic Regression	75.19%	59.29%	0.740 (0.689–0.785)	76.74%	60.81%	0.770 (0.725–0.811)
Naïve Bayes	73.46%	57.9%	0.727 (0.682–0.768)	48.06%	82.93%	0.763 (0.715–0.804)
CatBoost	73.64%	64.1%	0.755 (0.713–0.797)	82.95%	62.58%	0.774 (0.728–0.816)
Multi-layer Perceptron	69.77%	64.35%	0.732 (0.687–0.778)	75.19%	65.36%	0.757 (0.708–0.804)

AKI acute kidney injury, PCI percutaneous coronary intervention, AUC area under the curve, CI confidence interval

ISSN: 2047-783X