The 12‑lead ECG provides an excellent substrate for artificial intelligence (AI) enabled prediction of various cardiovascular diseases. However, a measure of prediction certainty is lacking.
To ...assess a novel approach for estimating certainty of AI-ECG predictions.
Two convolutional neural networks (CNN) were developed to predict patient age and sex. Model 1 applied a 5 s sliding time-window, allowing multiple CNN predictions. The consistency of the output values, expressed as interquartile range (IQR), was used to estimate prediction certainty. Model 2 was trained on the full 10s ECG signal, resulting in a single CNN point prediction value. Performance was evaluated on an internal test set and externally validated on the PTB-XL dataset.
Both CNNs were trained on 269,979 standard 12‑lead ECGs (82,477 patients). Model 1 showed higher accuracy for both age and sex prediction (mean absolute error, MAE 6.9 ± 6.3 years vs. 7.7 ± 6.3 years and AUC 0.946 vs. 0.916, respectively, P < 0.001 for both). The IQR of multiple CNN output values allowed to differentiate between high and low accuracy of ECG based predictions (P < 0.001 for both). Among 10% of patients with narrowest IQR, sex prediction accuracy increased from 65.4% to 99.2%, and MAE of age prediction decreased from 9.7 to 4.1 years compared to the 10% with widest IQR. Accuracy and estimation of prediction certainty of model 1 remained true in the external validation dataset.
Sliding window-based approach improves ECG based prediction of age and sex and may aid in addressing the challenge of prediction certainty estimation.
Sliding window-based artificial intelligence-enabled ECG provides an accurate and reliable estimate of prediction certainty. Display omitted
Background: The atrial fibrillation cycle length (AFCL) is a critical parameter for the perpetuation and termination of AF. In the present study, we evaluated a new method to measure the AFCL based ...on transthoracic tissue Doppler imaging (TDI) of the right atrium (RA) and left atrium (LA).
Methods: Twenty patients with AF (6 acute AF, 14 persistent or permanent AF) were studied. A quadripolar catheter was positioned at RA or LA to measure AFCL (AFCLEGM, gold standard). Transthoracic echocardiography (apical 4‐chamber view) was used to perform pulsed wave TDI at the free wall of RA or LA. AFCLTDI was defined as the time interval between two consecutive positive to negative crossings of the baseline of the atrial time velocity curves. AFCLEGM and AFCLTDI were measured at baseline and during a 10‐minute infusion of flecainide (1.5 mg/kg).
Results: Measurement of AFCLTDI was feasible in all but one patient. At baseline, AFCLEGM was 170 ± 22 ms, AFCLTDI 172 ± 22 ms (difference 2 ± 5 ms). AFCLTDI correlated significantly with AFCLEGM (R = 0.91, P < 0.0001). Bland‐Altman analysis showed a bias of −2 ms with a 95% limit of agreement between −26 ms and +22 ms. During flecainide, the AFCLTDI method yielded an AFCL prolongation from 176 ± 23 ms at baseline to 279 ± 68 ms (P < 0.01) after 10 minutes of infusion (57 ± 26%).
Conclusions: (1) Tissue Doppler imaging of the atria during transthoracic echocardiography can be used to reliably determine the AFCL during both acute and persistent or permanent AF. (2) Continuous measurement of AFCL with TDI can be used to monitor the effect of antiarrhythmic drugs on atrial rate during AF. (3) This novel method is attractive because of the ease of acquiring the data and its noninvasive character.
Very high-power short-duration (vHPSD) via temperature-controlled ablation (TCA) is a new modality to perform radiofrequency pulmonary vein isolation (PVI), conceivably at the cost of a narrower ...safety margin towards the oesophagus. In this two-centre trial, we aimed to determine the safety of vHPSD-based PVI with specific emphasis on silent oesophageal injury.
Ninety consecutive patients with atrial fibrillation (AF) underwent vHPSD-PVI (90 W, 3-4 s, TCA) using the QDOT MICRO catheter, in conjunction with the nGEN (Bad Neustadt, n = 45) or nMARQ generator (Bruges, n = 45). All patients underwent post-ablation oesophageal endoscopy. Procedural parameters and complications were recorded. A subgroup of 21 patients from Bad Neustadt underwent cerebral magnetic resonance imaging (cMRI) to detect silent cerebral events (SCEs). Mean age was 67 ± 9 years, 59% patients were male, and 66% patients had paroxysmal AF. Pulmonary vein isolation was obtained in all cases after 96 ± 29 min. No steam pop, cardiac tamponade, stroke, or fistula was reported. None of the 90 patients demonstrated oesophageal ulceration (0%). Charring was not observed in the nMARQ cohort (0% vs. 11% in the nGEN group). In 5 out of 21 patients (24%), cMRI demonstrated SCE (exclusively nGEN cohort).
Temperature-controlled vHPSD catheter ablation allows straightforward PVI without evidence of oesophageal ulcerations or symptomatic complications. Catheter tip charring and silent cerebral lesions when using the nGEN generator have led to further modification.
Accurate mapping of the compact atrioventricular (AV) node is critical during ablation of a range of arrhythmias.
The purpose of this multicenter prospective study was to test the hypothesis that ...microelectrode (ME)-embedded catheters more accurately define the near-field compact AV node compared to conventional catheters.
For the mapping phase, detailed AV junction maps were created in 47 patients using an ME-embedded catheter. His electrograms (EGMs) detected by conventional electrodes (His
) and by ME (His
) were annotated. For the ablation phase, AV nodal ablation (Qmode 50 W) was performed in 10 patients after pacemaker implantation, with initial His
-only ablation in group 1 (n = 6) and initial His
ablation in group 2 (n = 4). For the clinical phase, a prospective registry of parahisian tachycardia using QDOT was obtained.
In the mapping phase, 7.0 ± 5.4 His
and 8.0 ± 5.6 His
points were acquired per map (n = 47). His
cloud was smaller and more proximally located than His
cloud: (99.4 ± 74.7 mm
vs 197.6 ± 110.6 mm
; P = .0008). His
EGMs had larger amplitudes than His
EGMs (0.40 ± 0.38 mV vs 0.16 ± 0.1 mV; P = .0002). In the ablation phase, for group 1: His
-only ablation never resulted in AV block, whereas His
ablation resulted in AV block after limited ablation in all patients (after 13.3 ± 9.2 s); and for group 2: His
ablation always resulted in AV block after 1 application (after 14.3 ± 10.3 s). In the clinical phase, a His
-avoidance strategy could avoid AV block in a prospective registry of 11 patients.
ME more accurately defines the region of the compact node, and ablation in this region is associated with a high risk for AV block. ME-based mapping has the potential to significantly enhance ablation safety and efficacy.
In this work, we present the release of a novel easy to use software package called DGM or Directed-Graph-Mapping. DGM can automatically analyze any type of arrhythmia to find reentry or focal ...sources if the measurements are synchronized in time. Currently, DGM requires the local activation times (LAT) and the spatial coordinates of the measured electrodes. However, there is no requirement for any spatial organization of the electrodes, allowing to analyze clinical, experimental or computational data. DGM creates directed networks of the activation, which are analyzed with fast algorithms to search for reentry (cycles in the network) and focal sources (nodes with outgoing arrows). DGM has been mainly optimized to analyze atrial tachycardia, but we also discuss other applications of DGM demonstrating its wide applicability. The goal is to release a free software package which can allow researchers to save time in the analysis of cardiac data. An academic license is attached to the software, allowing only non-commercial use of the software. All updates of the software, user and installation guide will be published on a dedicated website
www.dgmapping.com
.
Graphical Abstract
Direct-Graph-Mapping is a method to automatically analyze a given arrhythmia by converting measured data of the electrodes in a directed network. DGM requires the local activation times (LAT) and the spatial coordinates of the measured electrodes. There is no requirement for any spatial organization of the electrodes, allowing to analyze clinical, experimental or computational data (see left). An example could be the LATs and coordinates from a CARTO file. DGM creates a directed network of the activation by (1) determining the neighbors of each node, 2 (2) allowing a directed arrow between two neighbors if propagation of the electrical wave is possible, (3) repeating this process for all nodes, (4) if necessary, redistributing the nodes more uniformly and repeating step (1)-(3). Two possible steps are (5) to visualize the wavefront by creating an average graph or (6) find the cycles in the network which represent the reentry loops. Focal sources are nodes with only outgoing arrows.
Early recurrence of atrial tachyarrhythmia (ERAT) is associated with ablation-induced proarrhythmogenic inflammation; however, existing studies used intermittent monitoring or nonoptimized ...radiofrequency (RF) applications (noncontiguous or without ablation index target value).
The purpose of this study was to investigate the relationship between ERAT and late recurrence based on insertable cardiac monitor (ICM) data.
We compiled data from Close-To-Cure and Close Maze studies, which enrolled patients who underwent RF ablation for paroxysmal or persistent atrial fibrillation (AF). All patients were implanted with an ICM 2-3 months before ablation.
We studied 165 patients (104 with paroxysmal AF, 61 with persistent AF). Over the 1-year follow-up period, 41 of the patients experienced late recurrence. The risk of late recurrence was higher in patients experiencing ERAT (hazard ratio HR 6.2; 95% confidence interval CI 3.0-13.0), with negative and positive predictive values of 90.5% and 45.7%, respectively. Median burden of AF during the blanking period was significantly higher in patients with late recurrence (7.9% 0.0%-99.6%) compared to those without recurrence (0.0% 0.0%-6.0; P <.001). For each 1% increase in AF burden during the blanking period, late recurrence increased by 4.6% (HR 1.046; 95% CI 1.035-1.059). The best tradeoff for predicting AF from ERAT occurrence was AF burden of 0.6% and last ERAT at 64 days.
In patients ablated for paroxysmal and persistent AF with a durable RF lesion set and implanted with a continuous monitoring device, postablation early AF recurrence and burden significantly predict late recurrence. The post-AF ablation blanking period should be reduced to 2 months.
Purpose. We aimed to study long-term outcome after pulmonary vein isolation (PVI) guided by remote magnetic navigation (RMN) and provided comparative data to outcome after manual navigation (MAN). ...Methods. Three hundred thirty-six patients with symptomatic paroxysmal AF underwent PVI by irrigated point-by-point radiofrequency (RF) ablation (RMN, n=114 versus MAN, n=222). Patients were followed up with symptom guided rhythm monitoring for a period up to 43 months. The end point of the study was freedom from repeat ablation after a single procedure and without antiarrhythmic drug treatment (ADT). Results. At the end of follow-up (median 26.3 months), freedom from repeat ablation was comparable between RMN and MAN (70.9% versus 69.5%, p=0.61). At repeat, mean number of reconnected veins was 2.4 ± 1.2 in RMN versus 2.6 ± 1.0 in MAN (p=0.08). The majority of repeat procedures occurred during the first year (82.1% in RMN versus 78.5% in MAN; p=0.74). Conclusion. On the long term (up to 3 years) and in a large cohort of patients with paroxysmal AF, RMN-guided PVI is as effective as MAN guided PVI. In both strategies the majority of repeat procedures occurred during the first year after index procedure.
We report on the entrapment of a circular mapping catheter in the mitral valve apparatus during segmental pulmonary vein isolation. In an attempt to remove the catheter by using mild traction in ...combination with clockwise and counter‐clockwise rotation, the circular part and the shaft of the catheter separated. The circular part could be removed without sequelae using a transseptally positioned vascular extraction nipper in combination with a basket catheter. (PACE 2003; 26:1774–1775)
Mitral annular disjunction (MAD) has received particular interest in patients with mitral valve prolapse, ventricular tachycardia, and sudden cardiac death. The clinical significance of MAD for ...patients with Marfan syndrome (MFS) remains largely unexplored.
To define the prevalence of MAD and examine its association with cardiovascular outcomes and arrhythmia among patients with MFS.
This retrospective, single-center cohort study included 142 patients with a diagnosis of MFS based on the revised Ghent criteria and a confirmed (likely) pathogenic variant in the FBN1 gene who underwent regular follow-up between January 1, 2004, and December 31, 2019.
The presence of MAD was assessed by echocardiography, and the extent of MAD was categorized in tertiles. Patients also underwent resting electrocardiography and 24-hour Holter monitoring. Outcomes included aortic events (aortic dissection or prophylactic aortic surgery), arrhythmic events (defined as sustained ventricular tachycardia or sudden cardiac death), and mitral valve surgery.
A total of 142 patients (72 female patients 51%; median age at first examination, 25 years range, 2-64 years) were evaluated. Forty-eight patients (34%) had MAD. Patients with MAD had larger aortic root z scores than patients without MAD (4.1 interquartile range, 2.8-5.7 vs 3.0 interquartile range, 1.8-4.0; P < .001) and more often had mitral valve prolapse (34 of 48 71% vs 14 of 94 15%; P < .001), ventricular ectopy (14 of 33 42% vs 15 of 70 21%; P = .03), and nonsustained ventricular tachycardia (13 of 33 39% vs 12 of 70 17%; P = .01). During follow-up, aortic events occurred at similar rates among patients with vs without MAD (15 of 43 35% vs 21 of 84 25%; P = .24), but patients in the upper MAD tertile (>10 mm) showed a higher occurrence of aortic events compared with patients with MAD of 10 mm or smaller (9 of 15 60% vs 6 of 28 21%; P = .01). Patients with arrhythmic events (n = 5) and patients requiring mitral valve surgery (n = 7) were observed exclusively in the group displaying MAD.
This study suggests that MAD among patients with MFS is associated with the occurrence of arrhythmic events, a higher need for mitral valve intervention, and, among patients with extensive MAD, more aortic events. Cardiac imaging for patients with MFS should consider the assessment of MAD as a potential marker for adverse outcomes.
PDF
