Prediction of acute kidney injury after total aortic arch replacement with postoperative serum lactate: a retrospective cohort study

Figures

• 

  Figure 1.. Flowchart for patient selection.

• 

  Figure 2.. ROC analysis, calibration curves and DCA for Model 1 and Model 2 in predicting postoperative AKI. Plot (A) presents the ROC analysis results of Model 1 and Model 2. The AUC-ROC for Model1 and Model2 were 0.757 (95% CI: 0.708–0.807, p < 0.001) and 0.824 (95% CI: 0.781–0.867, p < 0.001). Plots (B) and (C) display the calibration curves of Model 1 and Model 2, respectively. Calibration curves depicted the alignment between model-predicted outcomes and observed results, where a 45° diagonal indicated perfect fit. The thin dashed line and solid line denote the actual prediction curve and the bootstrap-adjusted curve, respectively. A closer fit of the solid line to the 45° line indicates better predictive accuracy of the model. Plots (D) and (E) show DCA of Model 1 and Model 2 for predicting postoperative AKI. The two curves are compared to the curves of treating none and all patients. AKI, acute kidney injury; AUC, area under the receiver operating characteristic curve; CI, confidence interval; DCA, decision curve analysis; ROC, receiver operating characteristic.