A deep learning model to predict recurrence of atrial fibrillation after pulmonary vein isolation

An Explainable AI Application (AF'fective) to Support Monitoring of Patients With Atrial Fibrillation After Catheter Ablation: Qualitative Focus Group, Design Session, and Interview Study

BACKGROUND

The opaque nature of artificial intelligence (AI) algorithms has led to distrust in medical contexts, particularly in the treatment and monitoring of atrial fibrillation. Although previous studies in explainable AI have demonstrated potential to address this issue, they often focus solely on electrocardiography graphs and lack real-world field insights.

OBJECTIVE

We addressed this gap by incorporating standardized clinical interpretation of electrocardiography graphs into the system and collaborating with cardiologists to co-design and evaluate this approach using real-world patient cases and data.

METHODS

We conducted a 3-stage iterative design process with 23 cardiologists to co-design, evaluate, and pilot an explainable AI application. In the first stage, we identified 4 physician personas and 7 explainability strategies, which were reviewed in the second stage. A total of 4 strategies were deemed highly effective and feasible for pilot deployment. On the basis of these strategies, we developed a progressive web application and tested it with cardiologists in the third stage.

RESULTS

The final progressive web application prototype received above-average user experience evaluations and effectively motivated physicians to adopt it owing to its ease of use, reliable information, and explainable functionality. In addition, we gathered in-depth field insights from cardiologists who used the system in clinical contexts.

CONCLUSIONS

Our study identified effective explainability strategies, emphasized the importance of curating actionable features and setting accurate expectations, and suggested that many of these insights could apply to other disease care contexts, paving the way for future real-world clinical evaluations.

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

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 (Mp), secondary (Ms), and combined (Mc) PV branches were built.

RESULTS

Features relating to primary PVs were found to be associated with AF recurrence. The Mp 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 Ms 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.

Beyond Clinical Factors: Harnessing Artificial Intelligence and Multimodal Cardiac Imaging to Predict Atrial Fibrillation Recurrence Post-Catheter Ablation

Atrial fibrillation (AF) is the most common type of cardiac arrhythmia, with catheter ablation being a key alternative to medical treatment for restoring normal sinus rhythm. Despite advances in understanding AF pathogenesis, approximately 35% of patients experience AF recurrence at 12 months after catheter ablation. Therefore, accurate prediction of AF recurrence occurring after catheter ablation is important for patient selection and management. Conventional methods for predicting post-catheter ablation AF recurrence, which involve the use of univariate predictors and scoring systems, have played a supportive role in clinical decision-making. In an ever-changing landscape where technology is becoming ubiquitous within medicine, cardiac imaging and artificial intelligence (AI) could prove pivotal in enhancing AF recurrence predictions by providing data with independent predictive power and identifying key relationships in the data. This review comprehensively explores the existing methods for predicting the recurrence of AF following catheter ablation from different perspectives, including conventional predictors and scoring systems, cardiac imaging-based methods, and AI-based methods developed using a combination of demographic and imaging variables. By summarising state-of-the-art technologies, this review serves as a roadmap for developing future prediction models with enhanced accuracy, generalisability, and explainability, potentially contributing to improved care for patients with AF.

Use of artificial intelligence and I-Score for prediction of recurrence before catheter ablation of atrial fibrillation

BACKGROUND

Based solely on pre-ablation characteristics, previous risk scores have demonstrated variable predictive performance. This study aimed to predict the recurrence of AF after catheter ablation by using artificial intelligence (AI)-enabled pre-ablation computed tomography (PVCT) images and pre-ablation clinical data.

METHODS

A total of 638 drug-refractory paroxysmal atrial fibrillation (AF) patients undergone ablation were recruited. For model training, we used left atria (LA) acquired from pre-ablation PVCT slices (126,288 images). A total of 29 clinical variables were collected before ablation, including baseline characteristics, medical histories, laboratory results, transthoracic echocardiographic parameters, and 3D reconstructed LA volumes. The I-Score was applied to select variables for model training. For the prediction of one-year AF recurrence, PVCT deep-learning and clinical variable machine-learning models were developed. We then applied machine learning to ensemble the PVCT and clinical variable models.

RESULTS

The PVCT model achieved an AUC of 0.63 in the test set. Various combinations of clinical variables selected by I-Score can yield an AUC of 0.72, which is significantly better than all variables or features selected by nonparametric statistics (AUCs of 0.66 to 0.69). The ensemble model (PVCT images and clinical variables) significantly improved predictive performance up to an AUC of 0.76 (sensitivity of 86.7% and specificity of 51.0%).

CONCLUSIONS

Before ablation, AI-enabled PVCT combined with I-Score features was applicable in predicting recurrence in paroxysmal AF patients. Based on all possible predictors, the I-Score is capable of identifying the most influential combination.

Exploring interpretability in deep learning prediction of successful ablation therapy for atrial fibrillation

Background: Radiofrequency catheter ablation (RFCA) therapy is the first-line treatment for atrial fibrillation (AF), the most common type of cardiac arrhythmia globally. However, the procedure currently has low success rates in dealing with persistent AF, with a reoccurrence rate of ∼50% post-ablation. Therefore, deep learning (DL) has increasingly been applied to improve RFCA treatment for AF. However, for a clinician to trust the prediction of a DL model, its decision process needs to be interpretable and have biomedical relevance. Aim: This study explores interpretability in DL prediction of successful RFCA therapy for AF and evaluates if pro-arrhythmogenic regions in the left atrium (LA) were used in its decision process. Methods: AF and its termination by RFCA have been simulated in MRI-derived 2D LA tissue models with segmented fibrotic regions (n = 187). Three ablation strategies were applied for each LA model: pulmonary vein isolation (PVI), fibrosis-based ablation (FIBRO) and a rotor-based ablation (ROTOR). The DL model was trained to predict the success of each RFCA strategy for each LA model. Three feature attribution (FA) map methods were then used to investigate interpretability of the DL model: GradCAM, Occlusions and LIME. Results: The developed DL model had an AUC (area under the receiver operating characteristic curve) of 0.78 ± 0.04 for predicting the success of the PVI strategy, 0.92 ± 0.02 for FIBRO and 0.77 ± 0.02 for ROTOR. GradCAM had the highest percentage of informative regions in the FA maps (62% for FIBRO and 71% for ROTOR) that coincided with the successful RFCA lesions known from the 2D LA simulations, but unseen by the DL model. Moreover, GradCAM had the smallest coincidence of informative regions of the FA maps with non-arrhythmogenic regions (25% for FIBRO and 27% for ROTOR). Conclusion: The most informative regions of the FA maps coincided with pro-arrhythmogenic regions, suggesting that the DL model leveraged structural features of MRI images to identify such regions and make its prediction. In the future, this technique could provide a clinician with a trustworthy decision support tool.

The Effect of Discharge Mode on the Distribution of Myocardial Pulsed Electric Field—A Simulation Study for Pulsed Field Ablation of Atrial Fibrillation

Background: At present, the effects of discharge modes of multielectrode catheters on the distribution of pulsed electric fields have not been completely clarified. Therefore, the control of the distribution of the pulsed electric field by selecting the discharge mode remains one of the key technical problems to be solved. Methods: We constructed a model including myocardium, blood, and a flower catheter. Subsequently, by setting different positive and ground electrodes, we simulated the electric field distribution in the myocardium of four discharge modes (A, B, C, and D) before and after the catheter rotation and analyzed their mechanisms. Results: Modes B, C, and D formed a continuous circumferential ablation lesion without the rotation of the catheter, with depths of 1.6 mm, 2.7 mm, and 0.7 mm, respectively. After the catheter rotation, the four modes could form a continuous circumferential ablation lesion with widths of 10.8 mm, 10.6 mm, 11.8 mm, and 11.5 mm, respectively, and depths of 5.2 mm, 2.7 mm, 4.7 mm, and 4.0 mm, respectively. Conclusions: The discharge mode directly affects the electric field distribution in the myocardium. Our results can help improve PFA procedures and provide enlightenment for the design of the discharge mode with multielectrode catheters.