Prediction of left ventricular ejection fraction improvement in patients with ischemic cardiomyopathy after coronary artery bypass grafting based on cardiac magnetic resonance

Figures

• 

  Figure 1. Patient screening flow chart. ICM, ischemic cardiomyopathy; CMR-LGE, cardiac magnetic resonance-late gadolinium enhancement; CABG, coronary artery bypass grafting; LVEF, left ventricular ejection fraction.

• 

  Figure 2. CMR-LGE imaging data for patients with LVEF non-improvement in 6 months after CABG. (A) Obtaining a typical LGE image on a continuous short axis section; (B) the software outlined the contours of the left ventricular endocardium (red circle) and epicardium (green circle). The insertion points of the interventricular septum were located through purple and blue circles. The selected blue contours in the black myocardium are normal myocardium as a contrast, and the yellow area is the scar myocardium identified by the gray threshold method of 5 standard deviations. (C) A 16-segment diagram of the left ventricle, values on each segment represent the degree of penetration of the myocardial scar at each segment. From the diagram, the number of transmural scar segments (LGE =76–100%) =5. CMR-LGE, cardiac magnetic resonance-late gadolinium enhancement; LVEF, left ventricular ejection fraction; CABG, coronary artery bypass grafting.

• 

  Figure 3. CMR-LGE imaging data for patients with LVEF improvement 6 months after CABG. (A) A typical LGE image was obtained on a continuous short axis section; (B) the software outlined the contours of the left ventricular endocardium (red circle) and epicardium (green circle). The insertion points of the interventricular septum were located through purple and blue circles. The selected blue contours in the black myocardium are normal myocardium as a contrast, and the yellow area is the scar myocardium identified by the gray threshold method of 5 standard deviations. (C) A 16-segment diagram of the left ventricle. Values on each segment represent the degree of wall penetration of each segment of the myocardial scar. From the diagram, all segments are non-wall permeable (all LGE ≤75%). CMR-LGE, cardiac magnetic resonance-late gadolinium enhancement; LVEF, left ventricular ejection fraction; CABG, coronary artery bypass grafting.

• 

  Figure 4. Clinical variables selection using the LASSO binary logistic regression model. (A) LASSO regression cross validation curve. The left dotted line is λ-min and the right dotted line is λ-SE. Due to the small sample size in this study, λ-SE on the right was selected as the screening criteria for the final model. (B) LASSO coefficient distribution of preoperative 32 variables. A dotted vertical line was drawn at the value selected using tenfold cross-validation, where optimal λ-SE criteria resulted in two predictive variables the same as variables selected by the stepwise regression method. LASSO, least absolute shrinkage and selection operator; SE, standard error.

• 

  Figure 5. The nomogram was used to predict the probability of LVEF non-improvement after CABG in ICM patients. LVESVI, left ventricular end-systolic volume index; LVEF, left ventricular ejection fraction; CABG, coronary artery bypass grafting; ICM, ischemic cardiomyopathy.

• 

  Figure 6. The nomogram is verified by drawing the ROC curve, calibration curve and decision curve analysis. (A) ROC curve of the nomogram generated by bootstrap resampling 200 times, AUC =0.866, 95% CI: 0.792–0.940. (B) The calibration curve of the nomogram was drawn by resampling 200 times using the bootstrap method. (C) Decision curve analysis of the model. The analysis of decision curve shows that there is a good net benefit when the threshold probability predicted by the nomogram is in the range of 10–80%. (D) Comparison of differentiation between LVESVI, the number of transmural scar segments and the nomogram. AUC, area under curve; CI, confidence interval; LVESVI, left ventricular end-systolic volume index; ROC, receiver operating characteristic.

• 

  Figure 7. MACCE-free survival curves of the LVEF improvement group (blue) and the LVEF non-improvement group (red) developed by the Kaplan-Meier method, as well as log-rank test, P=0.009. MACCE, major adverse cardiovascular and cerebrovascular events; LVEF, left ventricular ejection fraction; CI, confidence interval.