RESUMEN

We propose a new method for the classification task of distinguishing atrial fibrillation (AFib) from regular atrial tachycardias including atrial flutter (AFlu) based on a surface electrocardiogram (ECG). Recently, many approaches for an automatic classification of cardiac arrhythmia were proposed and to our knowledge none of them can distinguish between these two. We discuss reasons why deep learning may not yield satisfactory results for this task. We generate new and clinically interpretable features using mathematical optimization for subsequent use within a machine learning (ML) model. These features are generated from the same input data by solving an additional regression problem with complicated combinatorial substructures. The resultant can be seen as a novel machine learning model that incorporates expert knowledge on the pathophysiology of atrial flutter. Our approach achieves an unprecedented accuracy of 82.84% and an area under the receiver operating characteristic (ROC) curve of 0.9, which classifies as "excellent" according to the classification indicator of diagnostic tests. One additional advantage of our approach is the inherent interpretability of the classification results. Our features give insight into a possibly occurring multilevel atrioventricular blocking mechanism, which may improve treatment decisions beyond the classification itself. Our research ideally complements existing textbook cardiac arrhythmia classification methods, which cannot provide a classification for the important case of AFib↔AFlu. The main contribution is the successful use of a novel mathematical model for multilevel atrioventricular block and optimization-driven inverse simulation to enhance machine learning for classification of the arguably most difficult cases in cardiac arrhythmia. A tailored Branch-and-Bound algorithm was implemented for the domain knowledge part, while standard algorithms such as Adam could be used for training.

Asunto(s)

Arritmias Cardíacas/diagnóstico , Aprendizaje Automático , Algoritmos , Arritmias Cardíacas/clasificación , Fibrilación Atrial/clasificación , Fibrilación Atrial/diagnóstico , Aleteo Atrial/clasificación , Aleteo Atrial/diagnóstico , Electrocardiografía/métodos , Humanos

RESUMEN

BACKGROUND: Catheter ablation of right ventricular outflow tract ventricular arrhythmias from above the pulmonary valve is being increasingly reported. OBJECTIVE: The purpose of this study was to systematically analyze the spatial relationship between the pulmonary trunk and the left coronaries. METHODS: Contrast-enhanced computed tomographic scans from 58 patients were analyzed. After segmentation of the pulmonary trunk and the proximal left coronaries, 3-dimensional geometries were generated. Minimal distance between the pulmonary trunk and the coronaries was automatically determined using a newly developed mathematical algorithm. RESULTS: The minimal distance between the pulmonary trunk and the coronaries was 1.4 ± 0.11 mm. Closest relationship was detected 13.8 ± 0.87 mm above the pulmonary valve annulus. Considering a safety margin of 5 mm to render coronary damage unlikely, 84% of patients were found to be at potential risk within the bottom 10 mm of the left sinus cusp. In contrast, positions within or above the right and anterior cusps were less likely to exhibit a close relationship. We identified the anterior aspect of the left cusp as the most critical region. Positions 10-20 mm above the left cusp were found to be critical in 97% of patients. Clinical parameters such as gender, age, height, weight, and body mass index were not predictive of a close spatial relationship. CONCLUSION: Our data provide evidence for a close spatial relationship between the pulmonary trunk and coronary arteries. These results should be considered when performing catheter ablation from above the pulmonary valve.