Automatic Detection of Slow Conducting Channels during Substrate Ablation of Scar-Related Ventricular Arrhythmias.

BACKGROUND: Voltage mapping allows identifying the arrhythmogenic substrate during scar-related ventricular arrhythmia (VA) ablation procedures. Slow conducting channels (SCCs), defined by the presence of electrogram (EGM) signals with delayed components (EGM-DC), are responsible for sustaining VAs and constitute potential ablation targets. However, voltage mapping, as it is currently performed, is time-consuming, requiring a manual analysis of all EGMs to detect SCCs, and its accuracy is limited by electric far-field. We sought to evaluate an algorithm that automatically identifies EGM-DC, classifies mapping points, and creates new voltage maps, named "Slow Conducting Channel Maps" (SCC-Maps). METHODS: Retrospective analysis of electroanatomic maps (EAM) from 20 patients (10 ischemic, 10 with arrhythmogenic right ventricular dysplasia/cardiomyopathy) was performed. EAM voltage maps were acquired during sinus rhythm and used for ablation. Preprocedural contrast-enhanced cardiac magnetic resonance (Ce-CMR) imaging was available for the ischemic population. Three mapping modalities were analysed: (i) EAM voltage maps using standard (EAM standard) or manual (EAM screening) thresholds for defining core and border zones; (ii) SCC-Maps derived from the use of the novel SCC-Mapping algorithm that automatically identify EGM-DCs measuring the voltage of the local component; and (iii) Ce-CMR maps (when available). The ability of each mapping modality in identifying SCCs and their agreement was evaluated. RESULTS: SCC-Maps and EAM screening identified a greater number of SCC entrances than EAM standard (3.45 ± 1.61 and 2.95 ± 2.31, resp., vs. 1.05 ± 1.10; p < 0.01). SCC-Maps and EAM screening highly correlate with Ce-CMR maps in the ischemic population when compared to EAM standard (Lin's correlation = 0.628 and 0.679, resp., vs. 0.212, p < 0.01). CONCLUSION: The SCC-Mapping algorithm allows an operator-independent analysis of EGM signals showing better identification of the arrhythmogenic substrate characteristics when compared to standard voltage EAM.

In silico pace-mapping: prediction of left vs. right outflow tract origin in idiopathic ventricular arrhythmias with patient-specific electrophysiological simulations.

AIMS: A pre-operative non-invasive identification of the site of origin (SOO) of outflow tract ventricular arrhythmias (OTVAs) is important to properly plan radiofrequency ablation procedures. Although some algorithms based on electrocardiograms (ECGs) have been developed to predict left vs. right ventricular origins, their accuracy is still limited, especially in complex anatomies. The aim of this work is to use patient-specific electrophysiological simulations of the heart to predict the SOO in OTVA patients. METHODS AND RESULTS: An in silico pace-mapping procedure was designed and used on 11 heart geometries, generating for each case simulated ECGs from 12 clinically plausible SOO. Subsequently, the simulated ECGs were compared with patient ECG data obtained during the clinical tachycardia using the 12-lead correlation coefficient (12-lead ρ). Left ventricle (LV) vs. right ventricle (RV) SOO was estimated by computing the LV/RV ratio for each patient, obtained by dividing the average 12-lead ρ value of the LV- and RV-SOO simulated ECGs, respectively. Simulated ECGs that had virtual sites close to the ablation points that stopped the arrhythmia presented higher correlation coefficients. The LV/RV ratio correctly predicted LV vs. RV SOO in 10/11 cases; 1.07 vs. 0.93 P < 0.05 for 12-lead ρ. CONCLUSION: The obtained results demonstrate the potential of the developed in silico pace-mapping technique to complement standard ECG for the pre-operative planning of complex ventricular arrhythmias.

Clinical validation of automatic local activation time annotation during focal premature ventricular complex ablation procedures.

Aims: Current navigation systems incorporate algorithms for automatic identification of local activation time (LAT). However, data about their utility and accuracy in premature ventricular complex (PVC) ablation procedures are scarce. This study analyses the accuracy of an algorithmic method based on automatic annotation of the maximal negative slope of the unipolar electrogram within the window demarcated by the bipolar electrogram compared with conventional manual annotation during PVC ablation procedures. Methods and results: Forty patients with successful ablation of focal PVC in three centres were included. Electroanatomical activation maps obtained with the automatic system (WF-map) were compared with manual annotation maps (M-map). Correlation and concordance of LAT obtained with both methods were assessed at 3536 points. The distance between the earliest activation site (EAS) and the effective radiofrequency application point (e-RFp) were determined in M-map and WF-map. The distance between WF-EAS and M-EAS was assessed. Successful ablation sites included left ventricular outflow tract (LVOT; 55%), right ventricular outflow tract (40%), and tricuspid annulus (5%). Good correlation was observed between the two annotation approaches (r = 0.655; P < 0.0001). Bland-Altman analysis revealed a systematic delayed detection of LAT by WF-map (bias 33.8 ± 30.9 ms), being higher in LVOT than in the right ventricle (42.6 ± 29.2 vs. 27.2 ± 30.5 ms, respectively; P < 0.0001). No difference in EAS-eRFp distance was observed between M-map and WF-map (1.8 ± 2.8 vs. 1.8 ± 3.4 mm, respectively; P = 0.986). The median (interquartile range) distance between WF-EAS and M-EAS was 2.2(0-6) mm. Conclusion: Good correlation was found between M-map and WF-map. Local activation time detection was systematically delayed in WF-map, especially in LVOT. Accurate identification of e-RFp was achieved with both annotation approaches.

Automatic activation mapping and origin identification of idiopathic outflow tract ventricular arrhythmias.

PURPOSE: Activation mapping is used to guide ablation of idiopathic outflow tract ventricular arrhythmias (OTVAs). Isochronal activation maps help to predict the site of origin (SOO): left vs right outflow tract (OT). We evaluate an algorithm for automatic activation mapping based on the onset of the bipolar electrogram (EGM) signal for predicting the SOO and the effective ablation site in OTVAs. METHODS: Eighteen patients undergoing ablation due to idiopathic OTVAs were studied (12 with left ventricle OT origin). Right ventricle activation maps were obtained offline with an automatic algorithm and compared with manual annotation maps obtained during the intervention. Local activation time (LAT) accuracy was assessed, as well as the performance of the 10ms earliest activation site (EAS) isochronal area in predicting the SOO. RESULTS: High correlation was observed between manual and automatic LATs (Spearman's: 0.86 and Lin's: 0.85, both p<0.01). The EAS isochronal area were closely located in both map modalities (5.55 ± 3.56mm) and at a similar distance from the effective ablation site (0.15±2.08mm difference, p=0.859). The 10ms isochronal area longitudinal/perpendicular diameter ratio measured from automatic maps showed slightly superior SOO identification (67% sensitivity, 100% specificity) compared with manual maps (67% sensitivity, 83% specificity). CONCLUSIONS: Automatic activation mapping based on the bipolar EGM onset allows fast, accurate and observer-independent identification of the SOO and characterization of the spreading of the activation wavefront in OTVAs.

Generalising electrocardiogram detection and delineation: training convolutional neural networks with synthetic data augmentation.

Introduction: Extracting beat-by-beat information from electrocardiograms (ECGs) is crucial for various downstream diagnostic tasks that rely on ECG-based measurements. However, these measurements can be expensive and time-consuming to produce, especially for long-term recordings. Traditional ECG detection and delineation methods, relying on classical signal processing algorithms such as those based on wavelet transforms, produce high-quality delineations but struggle to generalise to diverse ECG patterns. Machine learning (ML) techniques based on deep learning algorithms have emerged as promising alternatives, capable of achieving similar performance without handcrafted features or thresholds. However, supervised ML techniques require large annotated datasets for training, and existing datasets for ECG detection/delineation are limited in size and the range of pathological conditions they represent. Methods: This article addresses this challenge by introducing two key innovations. First, we develop a synthetic data generation scheme that probabilistically constructs unseen ECG traces from "pools" of fundamental segments extracted from existing databases. A set of rules guides the arrangement of these segments into coherent synthetic traces, while expert domain knowledge ensures the realism of the generated traces, increasing the input variability for training the model. Second, we propose two novel segmentation-based loss functions that encourage the accurate prediction of the number of independent ECG structures and promote tighter segmentation boundaries by focusing on a reduced number of samples. Results: The proposed approach achieves remarkable performance, with a F 1 -score of 99.38% and delineation errors of 2.19 ± 17.73 ms and 4.45 ± 18.32 ms for ECG segment onsets and offsets across the P, QRS, and T waves. These results, aggregated from three diverse freely available databases (QT, LU, and Zhejiang), surpass current state-of-the-art detection and delineation approaches. Discussion: Notably, the model demonstrated exceptional performance despite variations in lead configurations, sampling frequencies, and represented pathophysiology mechanisms, underscoring its robust generalisation capabilities. Real-world examples, featuring clinical data with various pathologies, illustrate the potential of our approach to streamline ECG analysis across different medical settings, fostered by releasing the codes as open source.

Atrial fibrosis identification with unipolar electrogram eigenvalue distribution analysis in multi-electrode arrays.

Atrial fibrosis plays a key role in the initiation and progression of atrial fibrillation (AF). Atrial fibrosis is typically identified by a peak-to-peak amplitude of bipolar electrograms (b-EGMs) lower than 0.5 mV, which may be considered as ablation targets. Nevertheless, this approach disregards signal spatiotemporal information and b-EGM sensitivity to catheter orientation. To overcome these limitations, we propose the dominant-to-remaining eigenvalue dominance ratio (EIGDR) of unipolar electrograms (u-EGMs) within neighbor electrode cliques as a waveform dispersion measure, hypothesizing that it is correlated with the presence of fibrosis. A simulated 2D tissue with a fibrosis patch was used for validation. We computed EIGDR maps from both original and time-aligned u-EGMs, denoted as [Formula: see text] and [Formula: see text], respectively, also mapping the gain in eigenvalue concentration obtained by the alignment, [Formula: see text]. The performance of each map in detecting fibrosis was evaluated in scenarios including noise and variable electrode-tissue distance. Best results were achieved by [Formula: see text], reaching 94% detection accuracy, versus the 86% of b-EGMs voltage maps. The proposed strategy was also tested in real u-EGMs from fibrotic and non-fibrotic areas over 3D electroanatomical maps, supporting the ability of the EIGDRs as fibrosis markers, encouraging further studies to confirm their translation to clinical settings. Upper panels: map of [Formula: see text] from 3×3 cliques for Ψ= 0∘ and bipolar voltage map Vb-m, performed assuming a variable electrode-to-tissue distance and noisy u-EGMs (noise level σv = 46.4 µV ). Lower panels: detected fibrotic areas (brown), using the thresholds that maximize detection accuracy of each map.

Delineation of the electrocardiogram with a mixed-quality-annotations dataset using convolutional neural networks.

Detection and delineation are key steps for retrieving and structuring information of the electrocardiogram (ECG), being thus crucial for numerous tasks in clinical practice. Digital signal processing (DSP) algorithms are often considered state-of-the-art for this purpose but require laborious rule readaptation for adapting to unseen morphologies. This work explores the adaptation of the the U-Net, a deep learning (DL) network employed for image segmentation, to electrocardiographic data. The model was trained using PhysioNet's QT database, a small dataset of 105 2-lead ambulatory recordings, while being independently tested for many architectural variations, comprising changes in the model's capacity (depth, width) and inference strategy (single- and multi-lead) in a fivefold cross-validation manner. This work features several regularization techniques to alleviate data scarcity, such as semi-supervised pre-training with low-quality data labels, performing ECG-based data augmentation and applying in-built model regularizers. The best performing configuration reached precisions of 90.12%, 99.14% and 98.25% and recalls of 98.73%, 99.94% and 99.88% for the P, QRS and T waves, respectively, on par with DSP-based approaches. Despite being a data-hungry technique trained on a small dataset, a U-Net based approach demonstrates to be a viable alternative for this task.

Characterization of Atrial Propagation Patterns and Fibrotic Substrate With a Modified Omnipolar Electrogram Strategy in Multi-Electrode Arrays.

Introduction: The omnipolar electrogram method was recently proposed to try to generate orientation-independent electrograms. It estimates the electric field from the bipolar electrograms of a clique, under the assumption of locally plane and homogeneous propagation. The local electric field evolution over time describes a loop trajectory from which omnipolar signals in the propagation direction, substrate and propagation features, are derived. In this work, we propose substrate and conduction velocity mapping modalities based on a modified version of the omnipolar electrogram method, which aims to reduce orientation-dependent residual components in the standard approach. Methods: A simulated electrical propagation in 2D, with a tissue including a circular patch of diffuse fibrosis, was used for validation. Unipolar electrograms were calculated in a multi-electrode array, also deriving bipolar electrograms along the two main directions of the grid. Simulated bipolar electrograms were also contaminated with real noise, to assess the robustness of the mapping strategies against noise. The performance of the maps in identifying fibrosis and in reproducing unipolar reference voltage maps was evaluated. Bipolar voltage maps were also considered for performance comparison. Results: Results show that the modified omnipolar mapping strategies are more accurate and robust against noise than bipolar and standard omnipolar maps in fibrosis detection (accuracies higher than 85 vs. 80% and 70%, respectively). They present better correlation with unipolar reference voltage maps than bipolar and original omnipolar maps (Pearson's correlations higher than 0.75 vs. 0.60 and 0.70, respectively). Conclusion: The modified omnipolar method improves fibrosis detection, characterization of substrate and propagation, also reducing the residual sensitivity to directionality over the standard approach and improving robustness against noise. Nevertheless, studies with real electrograms will elucidate its impact in catheter ablation interventions.

A rule-based method to model myocardial fiber orientation in cardiac biventricular geometries with outflow tracts.

Rule-based methods are often used for assigning fiber orientation to cardiac anatomical models. However, existing methods have been developed using data mostly from the left ventricle. As a consequence, fiber information obtained from rule-based methods often does not match histological data in other areas of the heart such as the right ventricle, having a negative impact in cardiac simulations beyond the left ventricle. In this work, we present a rule-based method where fiber orientation is separately modeled in each ventricle following observations from histology. This allows to create detailed fiber orientation in specific regions such as the endocardium of the right ventricle, the interventricular septum, and the outflow tracts. We also carried out electrophysiological simulations involving these structures and with different fiber configurations. In particular, we built a modeling pipeline for creating patient-specific volumetric meshes of biventricular geometries, including the outflow tracts, and subsequently simulate the electrical wavefront propagation in outflow tract ventricular arrhythmias with different origins for the ectopic focus. The resulting simulations with the proposed rule-based method showed a very good agreement with clinical parameters such as the 10 ms isochrone ratio in a cohort of nine patients suffering from this type of arrhythmia. The developed modeling pipeline confirms its potential for an in silico identification of the site of origin in outflow tract ventricular arrhythmias before clinical intervention.

Image-based criteria to identify the presence of epicardial arrhythmogenic substrate in patients with transmural myocardial infarction.

BACKGROUND: Patients with transmural myocardial infarction (MI) who undergo endocardial-only substrate ablation are at increased risk for ventricular tachycardia recurrence. Late gadolinium enhancement cardiac magnetic resonance (LGE-CMR) can be used to assess infarct transmurality (IT). However, the degree of IT associated with an epicardial arrhythmogenic substrate (AS) has not been determined. OBJECTIVE: The purpose of this study was to determine the degree of IT observed by LGE-CMR and multidetector computed tomography (MDCT) that predicts the presence of epicardial AS. METHODS: The study included 38 post-MI patients. Ten patients with a subendocardial infarction underwent endocardial-only mapping, and 28 with a classic transmural MI (C-TMI), defined as hyperenhancement ≥75% of myocardial wall thickness (WT), underwent endo-epicardial mapping. LGE-CMR/MDCT data were registered to high-density endocardial or epicardial maps to be analyzed for the presence of AS. RESULTS: Of the 28 post-MI patients with C-TMI, 18 had epicardial AS (64%) and 10 (36%) did not. An epicardial scar area ≥14 cm2 on LGE-CMR identified patients with epicardial AS (sensitivity 1, specificity 1). Mean WT in the epicardial scar area in these patients was lower than in patients without epicardial AS (3.14 ± 1.16 mm vs 5.54 ± 1.78 mm; P = .008). A mean WT cutoff value ≤3.59 mm identified patients with epicardial AS (sensitivity 0.91, specificity 0.93). CONCLUSION: An epicardial scar area ≥14 cm2 on LGE-CMR and mean CT-WT ≤3.59 mm predict epicardial AS in post-MI patients.

A Multi-Variate Predictability Framework to Assess Invasive Cardiac Activity and Interactions During Atrial Fibrillation.

OBJECTIVE: This study introduces a predictability framework based on the concept of Granger causality (GC), in order to analyze the activity and interactions between different intracardiac sites during atrial fibrillation (AF). METHODS: GC-based interactions were studied using a three-electrode analysis scheme with multi-variate autoregressive models of the involved preprocessed intracardiac signals. The method was evaluated in different scenarios covering simulations of complex atrial activity as well as endocardial signals acquired from patients. RESULTS: The results illustrate the ability of the method to determine atrial rhythm complexity and to track and map propagation during AF. CONCLUSION: The proposed framework provides information on the underlying activation and regularity, does not require activation detection or postprocessing algorithms and is applicable for the analysis of any multielectrode catheter. SIGNIFICANCE: The proposed framework can potentially help to guide catheter ablation interventions of AF.

Electrocardiogram derived respiratory rate from QRS slopes and R-wave angle.

A method for estimating respiratory rate from electrocardiogram (ECG) signals is presented. It is based on QRS slopes and R-wave angle, which reflect respiration-induced beat morphology variations. The 12 standard leads, 3 leads from vectorcardiogram (VCG), and 2 additional non-standard leads derived from VCG loops were analyzed. The following series were studied as ECG derived respiration (EDR) signals: slope between the peak of Q and R waves, slope between the peak of R and S waves, and the R-wave angle. Information from several EDR signals was combined in order to increase the robustness of estimation. Evaluation is performed over two databases containing ECG and respiratory signals simultaneously recorded during two clinical tests with different characteristics: tilt test, representing abrupt cardiovascular changes, and stress test representing a highly non-stationary and noisy environment. A combination of QRS slopes and R-wave angle series derived from VCG leads obtained a respiratory rate estimation relative error of 0.50 ± 4.11% (measuring 99.84% of the time) for tilt test and 0.52 ± 8.99% (measuring 96.09% of the time) for stress test. These results outperform those obtained by other reported methods, both in tilt and stress testing.

A wavelet-based electrogram onset delineator for automatic ventricular activation mapping.

Electroanatomical mapping (EAM) systems are commonly used in clinical practice for guiding catheter ablation treatments of common arrhythmias. In focal tachycardias, the ablation target is defined by locating the earliest activation area determined by the joint analysis of electrogram (EGM) signals at different sites. However, this is currently a manual time-consuming and experience-dependent task performed during the intervention and thus prone to stress-related errors. In this paper, we present an automatic delineation strategy that combines electrocardiogram (ECG) information with the wavelet decomposition of the EGM signal envelope to identify the onset of each EGM signal for activation mapping. Fourteen electroanatomical maps corresponding to ten patients suffering from non-tolerated premature ventricular contraction (PVC) beats and admitted for ablation procedure were used for evaluation. We compared the results obtained automatically with two types of manual annotations: one during the intervention by an expert technician (on-procedure) and other after the intervention (off-procedure), free from time and procedural constraints, by two other technicians. The automatic annotations show a significant correlation (0.95, p 0.01) with the evaluation reference (off-procedure annotation sets combination) and has an error of 2.1 ± 10.9 ms, around the order of magnitude of the on-procedure annotations error ( - 2.6 ± 6.8 ms). The results suggest that the proposed methodology could be incorporated into EAM systems to considerably reduce processing time during ablation interventions.

Very low frequency modulation in QRS slopes and its relation with respiration and heart rate variability during hemodialysis.

In this work, we study the very low frequency (VLF) modulation (range 0.01-0.03 Hz) in QRS slopes, heart rate variability (HRV) and ECG-derived respiration in hemodialysis patients. First, the relation between QRS slopes and HRV in the VLF band is measured using ordinary coherence. Then, partial coherence is used to measure the former relationship once the effect related to respiration is removed. Ordinary coherence values above a statistical threshold revealed linear relationship between VLF modulation in QRS slopes and HRV in about 10% of analyzed segments, with mean ± SD values of 0.79 ± 0.07 for upward slope and 0.77 ± 0.06 for downward slope. For these segments, partial coherence values drop below the threshold for 64% of the cases for upward slope and 76% for downward slope, suggesting that the origin of the VLF modulation in QRS slopes is mainly driven by respiration or linearly related to it. In the rest of the cases, partial coherence values dropped with respect to ordinary coherence from 0.89 to 0.77 for upward slope and from 0.86 to 0.75 for downward slope, suggesting that other ANS effects non-linearly related to respiration also contribute to the VLF modulation in QRS slopes.

Electrocardiogram derived respiration from QRS slopes.

A method for estimation of respiratory rate from electrocardiogram (ECG) signals, based on variations in slopes of QRS complexes, is presented. 12 standard leads, 3 leads from vectorcardiogram (VCG), and 2 additional non-standard leads derived from VCG loops were analysed. A total of 34 slope series were studied, 2 for each analysed lead: slopes between the peak of Q and R waves, and between the peak of R and S waves. Information of QRS slopes series was combined in order to increase the robustness of estimation. Evaluation is performed over a database containing ECG and respiratory signals simultaneously recorded in 17 subjects spontaneously breathing during a tilt table test. Respiratory rate estimation is performed with information of 4 different combinations of QRS slope series. The best results in respiratory rate estimation error terms are 0.72 ± 4.34%(0.46 ± 7.59 mHz). These results outperform those obtained with other known methods, motivating the use of QRS slopes to obtain reliable respiratory rate estimates.