Toward Patient-Specific Prediction of Ablation Strategies for Atrial Fibrillation Using Deep Learning

Standardized 2D atrial mapping and its clinical applications

The visualization and comparison of electrophysiological information in the atrium among different patients could be facilitated by a standardized 2D atrial mapping. However, due to the complexity of the atrial anatomy, unfolding the 3D geometry into a 2D atrial mapping is challenging. In this study, we aim to develop a standardized approach to achieve a 2D atrial mapping that connects the left and right atria, while maintaining fixed positions and sizes of atrial segments across individuals. Atrial segmentation is a prerequisite for the process. Segmentation includes 19 different segments with 12 segments from the left atrium, 5 segments from the right atrium, and two segments for the atrial septum. To ensure consistent and physiologically meaningful segment connections, an automated procedure is applied to open up the atrial surfaces and project the 3D information into 2D. The corresponding 2D atrial mapping can then be utilized to visualize different electrophysiological information of a patient, such as activation time patterns or phase maps. This can in turn provide useful information for guiding catheter ablation. The proposed standardized 2D maps can also be used to compare more easily structural information like fibrosis distribution with rotor presence and location. We show several examples of visualization of different electrophysiological properties for both healthy subjects and patients affected by atrial fibrillation. These examples show that the proposed maps provide an easy way to visualize and interpret intra-subject information and perform inter-subject comparison, which may provide a reference framework for the analysis of the atrial fibrillation substrate before treatment, and during a catheter ablation procedure.

Left atrial appendage morphofunctional indices could be predictive of arrhythmia recurrence post-atrial fibrillation ablation: a meta-analysis

BACKGROUND

Left atrium changes are implicated in atrial fibrillation (AF) substrate and are predictive of AF outcomes. Left atrial appendage (LAA) is an integral component of left atrial structure and could be affected by atrial cardiomyopathy. We aimed to elucidate the association between LAA indices and late arrhythmia recurrence after atrial fibrillation catheter ablation (AFCA).

METHODS

The MEDLINE database, ClinicalTrials.gov, medRxiv and Cochrane Library were searched for studies evaluating LAA and late arrhythmia recurrence in patients undergoing AFCA. Data were pooled by meta-analysis using a random-effects model. The primary endpoint was pre-ablation difference in LAA anatomic or functional indices.

RESULTS

A total of 34 studies were found eligible and five LAA indices were analyzed. LAA ejection fraction and LAA emptying velocity were significantly lower in patients with AF recurrence post-ablation [SMD = - 0.66; 95% CI (- 1.01, - 0.32) and SMD = - 0.56; 95% CI (- 0.73, - 0.40) respectively] as compared to arrhythmia free controls. LAA volume and LAA orifice area were significantly higher in patients with AF recurrence post-ablation (SMD = 0.51; 95% CI 0.35-0.67, and SMD = 0.35; 95% CI 0.20-0.49, respectively) as compared to arrhythmia free controls. LAA morphology was not predictive of AF recurrence post-ablation (chicken wing morphology; OR 1.27; 95% CI 0.79-2.02). Moderate statistical heterogeneity and small case-control studies are the main limitations of our meta-analysis.

CONCLUSIONS

Our findings suggest that LAA ejection fraction, LAA emptying velocity, LAA orifice area and LAA volume differ between patients suffering from arrhythmia recurrence post-ablation and arrhythmia free counterparts, while LAA morphology is not predictive of AF recurrence.

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.

LncRNA HOTAIR promotes myocardial fibrosis in atrial fibrillation through binding with PTBP1 to increase the stability of Wnt5a

BACKGROUND

Atrial fibrillation (AF) is one of the most common arrhythmia in clinical practice, and atrial fibrosis is the important mediator in AF. LncRNA HOTAIR was reported to be up-regulated in AF, while the underlying mechanism of HOTAIR in AF remains unclear.

METHODS

In vitro and in vivo AF model was established. qRT-PCR and Western blotting were used to assess the mRNA expression (HOTAIR, Wnt5a and PTBP1) and protein levels (Wnt5a, collagen I/III, α-SMA, CTGF, p-ERK, ERK, p-JNK, and JNK), respectively. MTT, CCK8, transwell assay was used to test cell viability, proliferation and migration, respectively. RIP assay assessed the correlation among HOTAIR, PTBP1 and Wnt5a. The level of α-SMA was detected by immunofluorescence. HE and Masson staining detected the histological changes and fibrosis in mouse heart tissues.

RESULTS

Ang II significantly increased the viability of atrial fibroblasts. The levels of HOTAIR and Wnt5a in fibroblasts were up-regulated by Ang II. HOTAIR silencing or Wnt5a significantly inhibited Ang II-induced proliferation, migration and fibrosis in fibroblasts. HOTAIR silencing repressed Wnt5a-mediated ERK and JNK signaling pathway, and Wnt5a partially abolished the effect of HOTAIR silencing on cell proliferation, migration and fibrosis. Meanwhile, HOTAIR could increase the mRNA stability of Wnt5a via recruiting PTBP1. Furthermore, HOTAIR knockdown notably inhibited the fibrosis in heart tissues of AF mice via regulation of Wnt signaling.

CONCLUSION

HOTAIR could promote atrial fibrosis in AF through binding with PTBP1 to increase Wnt5a stability. Our study might shed new insights on exploring new strategies against AF.

Interpretable machine learning of action potential duration restitution kinetics in single-cell models of atrial cardiomyocytes

Action potential duration (APD) restitution curve and its maximal slope (Smax) reflect single cell-level dynamic instability for inducing chaotic heart rhythms. However, conventional parameter sensitivity analysis often fails to describe nonlinear relationships between ion channel parameters and electrophysiological phenotypes, such as Smax. We explored the parameter-phenotype mapping in a population of 5000 single-cell atrial cell models through interpretable machine learning (ML) approaches. Parameter sensitivity analyses could explain the linear relationships between parameters and electrophysiological phenotypes, including APD₉₀, resting membrane potential, Vmax, refractory period, and APD/calcium alternans threshold, but not for Smax. However, neural network models had better prediction performance for Smax. To interpret the ML model, we evaluated the parameter importance at the global and local levels by computing the permutation feature importance and the local interpretable model-agnostic explanations (LIME) values, respectively. Increases in I_CaL, I_NCX, and I_Kr, and decreases in I_K1, I_b,Cl, I_Kur, I_SERCA, and I_to are correlated with higher Smax values. The LIME algorithm determined that I_NaK plays a significant role in determining Smax as well as I_to and I_Kur. The atrial cardiomyocyte population was hierarchically clustered into three distinct groups based on the LIME values and the single-cell simulation confirmed that perturbations in I_NaK resulted in different behaviors of APD restitution curves in three clusters. Our combined top-down interpretable ML and bottom-up mechanistic simulation approaches uncovered the role of I_NaK in heterogeneous behaviors of Smax in the atrial cardiomyocyte population.

Artificial Intelligence-Driven Algorithm for Drug Effect Prediction on Atrial Fibrillation: An in silico Population of Models Approach

Background: Antiarrhythmic drugs are the first-line treatment for atrial fibrillation (AF), but their effect is highly dependent on the characteristics of the patient. Moreover, anatomical variability, and specifically atrial size, have also a strong influence on AF recurrence. Objective: We performed a proof-of-concept study using artificial intelligence (AI) that enabled us to identify proarrhythmic profiles based on pattern identification from in silico simulations. Methods: A population of models consisting of 127 electrophysiological profiles with a variation of nine electrophysiological variables (G _Na , I _NaK , G _K1, G _CaL , G _Kur , I _KCa , [Na] _ext , and [K] _ext and diffusion) was simulated using the Koivumaki atrial model on square planes corresponding to a normal (16 cm²) and dilated (22.5 cm²) atrium. The simple pore channel equation was used for drug implementation including three drugs (isoproterenol, flecainide, and verapamil). We analyzed the effect of every ionic channel combination to evaluate arrhythmia induction. A Random Forest algorithm was trained using the population of models and AF inducibility as input and output, respectively. The algorithm was trained with 80% of the data (N = 832) and 20% of the data was used for testing with a k-fold cross-validation (k = 5). Results: We found two electrophysiological patterns derived from the AI algorithm that was associated with proarrhythmic behavior in most of the profiles, where G _K1 was identified as the most important current for classifying the proarrhythmicity of a given profile. Additionally, we found different effects of the drugs depending on the electrophysiological profile and a higher tendency of the dilated tissue to fibrillate (Small tissue: 80 profiles vs Dilated tissue: 87 profiles). Conclusion: Artificial intelligence algorithms appear as a novel tool for electrophysiological pattern identification and analysis of the effect of antiarrhythmic drugs on a heterogeneous population of patients with AF.

Reinforcement Learning to Improve Image-Guidance of Ablation Therapy for Atrial Fibrillation

Atrial fibrillation (AF) is the most common cardiac arrhythmia and currently affects more than 650,000 people in the United Kingdom alone. Catheter ablation (CA) is the only AF treatment with a long-term curative effect as it involves destroying arrhythmogenic tissue in the atria. However, its success rate is suboptimal, approximately 50% after a 2-year follow-up, and this high AF recurrence rate warrants significant improvements. Image-guidance of CA procedures have shown clinical promise, enabling the identification of key patient anatomical and pathological (such as fibrosis) features of atrial tissue, which require ablation. However, the latter approach still suffers from a lack of functional information and the need to interpret structures in the images by a clinician. Deep learning plays an increasingly important role in biomedicine, facilitating efficient diagnosis and treatment of clinical problems. This study applies deep reinforcement learning in combination with patient imaging (to provide structural information of the atria) and image-based modelling (to provide functional information) to design patient-specific CA strategies to guide clinicians and improve treatment success rates. To achieve this, patient-specific 2D left atrial (LA) models were derived from late-gadolinium enhancement (LGE) MRI scans of AF patients and were used to simulate patient-specific AF scenarios. Then a reinforcement Q-learning algorithm was created, where an ablating agent moved around the 2D LA, applying CA lesions to terminate AF and learning through feedback imposed by a reward policy. The agent achieved 84% success rate in terminating AF during training and 72% success rate in testing. Finally, AF recurrence rate was measured by attempting to re-initiate AF in the 2D atrial models after CA with 11% recurrence showing a great improvement on the existing therapies. Thus, reinforcement Q-learning algorithms can predict successful CA strategies from patient MRI data and help to improve the patient-specific guidance of CA therapy.

Machine Learning Identification of Pro-arrhythmic Structures in Cardiac Fibrosis

Cardiac fibrosis and other scarring of the heart, arising from conditions ranging from myocardial infarction to ageing, promotes dangerous arrhythmias by blocking the healthy propagation of cardiac excitation. Owing to the complexity of the dynamics of electrical signalling in the heart, however, the connection between different arrangements of blockage and various arrhythmic consequences remains poorly understood. Where a mechanism defies traditional understanding, machine learning can be invaluable for enabling accurate prediction of quantities of interest (measures of arrhythmic risk) in terms of predictor variables (such as the arrangement or pattern of obstructive scarring). In this study, we simulate the propagation of the action potential (AP) in tissue affected by fibrotic changes and hence detect sites that initiate re-entrant activation patterns. By separately considering multiple different stimulus regimes, we directly observe and quantify the sensitivity of re-entry formation to activation sequence in the fibrotic region. Then, by extracting the fibrotic structures around locations that both do and do not initiate re-entries, we use neural networks to determine to what extent re-entry initiation is predictable, and over what spatial scale conduction heterogeneities appear to act to produce this effect. We find that structural information within about 0.5 mm of a given point is sufficient to predict structures that initiate re-entry with more than 90% accuracy.