Artificial Intelligence-Based Feature Analysis of Pulmonary Vein Morphology on Computed Tomography Scans and Risk of Atrial Fibrillation Recurrence After Catheter Ablation: A Multi-Site Study

Abstract

Background: Atrial fibrillation (AF) recurrence is common after catheter ablation. Pulmonary vein (PV) isolation is the cornerstone of AF ablation, but PV remodeling has been associated with the risk of AF recurrence. We aimed to evaluate whether artificial intelligence-based morphological features of primary and secondary PV branches on computed tomography images are associated with AF recurrence post-ablation.

Methods: Two artificial intelligence models were trained for the segmentation of computed tomography images, enabling the isolation of PV branches. Patients from Cleveland Clinic (N=135) and Vanderbilt University (N=594) were combined and divided into 2 sets for training and cross-validation (D₁, n=218) and internal testing (D₂, n=511). An independent validation set (D₃, N=80) was obtained from University Hospitals of Cleveland. We extracted 48 fractal-based and 12 shape-based radiomic features from primary and secondary PV branches of patients with AF recurrence (AF+) and without recurrence after catheter ablation of AF (AF-). To predict AFrecurrence, 3 Gradient Boosting classification models based on significant features from primary (M_p), secondary (M_s), and combined (M_c) PV branches were built.

Results: Features relating to primary PVs were found to be associated with AF recurrence. The M_p classifier achieved area under the curve values of 0.73, 0.71, and 0.70 across the 3 datasets. AF+ cases exhibited greater surface complexity in their primary PV area, as evidenced by higher fractal dimension values compared with AF- cases. The M_s classifier results revealed a weaker association with AF+, suggesting higher relevance to AF recurrence post-ablation from primary PV branch morphology.

Conclusions: This largest multi-institutional study to date revealed associations between artificial intelligence-extracted morphological features of the primary PV branches with AF recurrence in 809 patients from 3 sites. Future work will focus on enhancing the predictive ability of the classifier by integrating clinical, structural, and morphological features, including left atrial appendage and left atrium-related characteristics.

Keywords: artificial intelligence; atrial fibrillation; catheter ablation; fractals; pulmonary veins.

Figure 1.

Flowchart of the method. The workflow includes 3 stages of segmentation, feature extraction and model evaluation. AF indicates atrial fibrillation; PV, Pulmonary vein; LA, Left atrium; AUC, Area under the curve.

Figure 2.

Flow diagram of the dataset with exclusion criteria. Data from each cohort is shown with a specific color.

Figure 3.

Workflow of PV and LA segmentation. An ensemble UNnet was trained on 90 cases for PV segmentation. A low resolution UNet was trained on 120 cases for LA segmentation. CT indicates computed tomography; PV, Pulmonary vein; LA, Left atrium

Figure 4.

PV and LA segmentations of a CT image. LA segmentation, primary PV branches, and secondary PV branches are shown in yellow, orange and blue colors, respectively. (a) shows the segmentations overlaid on the CT images, while (b) presents the segments in a 3D space.