Development and Validation of a Machine Learning-Based Radiomics Model on Cardiac Computed Tomography of Epicardial Adipose Tissue in Predicting Characteristics and Recurrence of Atrial Fibrillation

Abstract

Purpose: This study aimed to evaluate the feasibility of differentiating the atrial fibrillation (AF) subtype and preliminary explore the prognostic value of AF recurrence after ablation using radiomics models based on epicardial adipose tissue around the left atrium (LA-EAT) of cardiac CT images.

Method: The cardiac CT images of 314 patients were collected wherein 251 and 63 cases were randomly enrolled in the training and validation cohorts, respectively. Mutual information and the random forest algorithm were used to screen for the radiomic features and construct the radiomics signature. Radiomics models reflecting the features of LA-EAT were built to differentiate the AF subtype, and the multivariable logistic regression model was adopted to integrate the radiomics signature and volume information. The same methodology and algorithm were applied to the radiomic features to explore the ability for predicting AF recurrence.

Results: The predictive model constructed by integrating the radiomic features and volume information using a radiomics nomogram showed the best ability in differentiating AF subtype in the training [AUC, 0.915; 95% confidence interval (CI), 0.880-0.951] and validation (AUC, 0.853; 95% CI, 0.755-0.951) cohorts. The radiomic features have shown convincible predictive ability of AF recurrence in both training (AUC, 0.808; 95% CI, 0.750-0.866) and validation (AUC, 0.793; 95% CI, 0.654-0.931) cohorts.

Conclusions: The LA-EAT radiomic signatures are a promising tool in the differentiation of AF subtype and prediction of AF recurrence, which may have clinical implications in the early diagnosis of AF subtype and disease management.

Keywords: atrial fibrillation; computed tomography angiography; epicardial adipose tissue; radiomics approach; recurrence.

Figure 1

The patient enrollment flow diagram.

Figure 2

Flowchart of the current study.

Figure 3

(A) EAT. (B) LA-EAT. EAT, epicardial adipose tissue; LA-EAT, epicardial adipose tissue surrounding the left atrium.

Figure 4

The radiomics nomogram of a personalized predictive model.

Figure 5

Comparison of ROCs between the Rmodel, Cmodel, Vmodel, and Commodel for the prediction of AF subtype in the (A) training and (B) validation cohorts. Rmodel, radiomics signature model; Cmodel, clinical model; Vmodel, the model based on volume values; Commodel, the model combined with both radiomic features and volume information.

Figure 6

(A) Calibration curves of the Rmodel, Cmodel, Vmodel, and Commodel. (B) The decision curve analysis for the Rmodel, Cmodel, Vmodel, and Commodel. The decision curve analysis showed that the Commodel had the highest overall net benefit ratio compared with the Rmodel, Cmodel, and Vmodel. Rmodel, radiomics signature model; Cmodel, clinical model; Vmodel, the model based on volume values; Commodel, the model combined with both radiomic features and volume information.

Figure 7

The ROC of the radiomics signatures model for predicting AF recurrence in the training and validation cohorts.