The BSICOS group has participated in the Computing in Cardiology Congress

Clara Sales: Left Bundle Branch Area Pacing Generates More Physiological Ventricular Activation Sequences than Right Ventricular Pacing

Left bundle branch area pacing (LBBAP) has been proposed as a new physiological pacing modality to overcome ventricular dyssynchrony reported in bradycardic patients undergoing conventional right ventricular pacing (RVP). The standard non-invasive measure of depolarization synchrony is the QRS duration. However, a deeper understanding of not only the activation time but also the activation sequence is needed to evaluate the effects of RVP and LBBAP in bradycardic patients. This study aimed to estimate the precordial ventricular activation from standard 12-lead ECGs of bradycardic patients with physiological conduction and use it to compare LBBAP vs RVP. 22 RVP and 42 LBBAP ECGs recordings were collected before and after pacemaker implantation. High (HF) and low (LF) frequency-based methods for QRS complex analysis were used and the precordial activation sequence and activation delay (pAD) were estimated. Results showed more physiological activation sequences after LBBAP than after RVP, with lower pAD (p\(<\)0.01) after LBBAP [HF:12(-2,15) vs 29(6,56); LF:9(-25,13) vs 31(17,38)]. The proposed ECG methodology could be used in clinical practice to map more physiological pacing targets in pacemaker implantation.

Alba Martín: Assessment of QT Interval Dynamics Induced by Heart Rate Changes through Bivariate Phase-Rectified Signal Averaging

The aim of this study is to investigate the relationship between the RR and QT series through the bivariate phase-rectified signal averaging (BPRSA), assuming that spontaneous changes of RR (trigger signal) cause a response in QT (target signal). Moreover, the prognostic value of new deceleration-related indices in a chronic heart failure (CHF) population will be assessed.

The 24-hour QT and RR series were extracted from 650 Holter recordings acquired in a CHF population. The PRSA and BPRSA techniques were applied to assess the QT/RR dependency. The original PRSA technique is designed for detecting and quantifying recurrent components in biological series, characterized by non-stationarities and noise. First, deceleration anchor points were identified as an RR increase, and segments of 2L+1=15 samples around them were extracted from both the trigger (RR) and the target signal (QT). Finally, all segments were aligned with respect to the anchor and averaged, and four new indices were defined as differences between consecutive samples of the BPRSA series (xBPRSA(i)): ∆1,0=xBPRSA(1) − xBPRSA(0) and ∆0,−1= xBPRSA(0) − xBPRSA(−1), where i = 0 is the position of the aligned anchor points. For comparison, deceleration capacity (DC) and its bivariate analogous (BDC) were also assessed. Prognostic value of deceleration-based computed indices in predicting cardiac death was determined with univariate Cox proportional hazards analysis.

Patients suffering from pump failure death (PFD) had significant lower DC and QT increase after heart rate deceleration. As shown in Table 1, the proposed BPRSA indices were associated to pump failure mortality in the studied population.

Saúl Palacios: Spatial Dispersion of Activation and Repolarization Times Associated with Different Cardiac Pacing Modes

Various modes of ventricular pacing are currently applied to patients with an indication for permanent pacemaker implantation. The so-called physiological pacing modes, like His bundle pacing (HBP) and left bundle branch pacing (LBBP), stimulate the cardiac conduction system to induce efficient physiological activation. Other techniques, such as left ventricular septal pacing (LVSP) and right ventricular septal and apical pacing (RVSP and RVAP), stimulate the ventricular septum or right ventricular apex. 695 ultra-high-frequency electrocardiograms (UHF-ECG) from 176 patients with narrow QRS and pacemaker indication were analyzed to characterize their activation (AT) and repolarization (RT) time. AT and RT were grouped into three regions (R1: leads V1-V2; R2: V3-V4; R3: V5-V6). Overall, selective HBP (sHBP), non-selective LBBP (nsLBBP) and LVSP recordings had the closest AT and RT values to spontaneous rhythm recordings. For AT, the mean R1-R2 and R3-R2 differences with respect to spontaneous rhythm were, in absolute value, below 3, 16 and 10 ms for sHBP, nsLBBP and LVSP, respectively. For RT, the corresponding mean differences were below 11, 34 and 24 ms for sHBP, nsLBBP and LVSP. In conclusion, HBP, LBBP and LVSP render the closest ventricular AT and RT to the spontaneous rhythm in patients with physiological conduction (narrow QRS).

Maxi Rosales: In Silico Assessment of Arrhythmic Risk in Infarcted Ventricles Engrafted with Engineered Heart Tissues

Engineered heart tissues (EHTs) from human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) represent novel alternatives to repair damaged cardiac tissue after myocardial infarction (MI). However, their associated slow conduction and prolonged action potential duration (APD) can favor pro-arrhythmicity. We coupled biventricular (BiV) and EHT in silico electrophysiological models. The BiV model, defined from magnetic resonance imaging, included an MI region. The EHT model was deformed to follow the epicardial surface. Different values of EHT electrical conductivity (EHTc) and degree of myocardial attachment (EHTa) were tested. Results showed that even a minimum degree of EHTa inhibited EHT automaticity. Activation time and APD were driven by myocardial tissue when engrafted. Importantly, higher EHTc and EHTa led to lower repolarization gradients (RTGs), considered as a surrogate of arrhythmic risk. Maximum RTG decreased by 52.3 ms/mm when EHTa varied from 25% to 100% under low EHTc and decreased by 98.8 ms/mm when EHTc raised to 90% of the myocardium conductivity under low EHTa. Pro-arrhythmicity in cardiac tissue engineering highly depends on EHT conductivity and degree of engraftment on the myocardium.

Josseline Madrid: ECG-Based Unsupervised Clustering in Coronary Artery Disease Associates with Ventricular Arrhythmia

Coronary Artery Disease (CAD) is one of the leading causes of life-threatening ventricular arrhythmias (LTVAs) leading to sudden cardiac death. The presence of CAD slows ventricular conduction heterogeneously across individuals, manifesting as different QRS morphologies. This study aimed to identify distinct clusters of CAD individuals based on QRS morphology using unsupervised learning, and investigate their association with LTVA risk.

An average heartbeat was derived from 10-second ECGs (lead I) from 1,458 individuals diagnosed with CAD in the UK Biobank study. QRS morphology was mathematically characterized by a combination of Hermite functions, as well as standard biomarkers, such as QRS amplitude, slopes, and duration. An unsupervised clustering algorithm based on 3-nearest neighbours was used to classify each individual into 3 distinct clusters. LTVA risk was defined as LTVA mortality or admission to hospital with a LTVA diagnosis 6-months before or after the CAD diagnosis. Statistical nonparametric tests (chi-square test) were performed to evaluate the association of each of the clusters with LTVA risk.

There were a total of 65 LTVA events in the population. The unsupervised algorithm was able to distinguish 3 distinct clusters of QRS-related morphological features in CAD, which significantly differed in terms of LTVA events rate. Cluster 1, which included the highest rate of LTVA events (6.38%), was characterized by lower QRS amplitude and down slopes, and a wider QRS than clusters 2 and 3.

Our analysis has identified in an unsupervised manner CAD individuals at risk of LTVA using information from the QRS morphology. Further studies will investigate the contribution of additional LTVA risk factors in CAD.

Laura García: Physiological variations in CX43 and fibrosis deposition affect human ventricular electrophysiology promoting arrhythmia

Connexin 43 (Cx43), the major component of gap junctions in the ventricle, is responsible for electrical impulse transmission between ventricular cardiomyocytes. Little is known about the
interindividual heterogeneity of CX43 tissue expression in the human left ventricle (LV) and its contribution to arrhythmogenecity either alone or in combination with other proarrhythmic factors like fibrosis.

We processed LV fluorescent immunostaining images from living-donors and characterized the population heterogeneity of CX43 expression and fibrosis deposition. The lowest CX43 expression and the highest fibrosis deposition values in the population were implemented in 2D computational models of human LV electrophysiology. We measured conduction velocity (CV) and areas of High Repolarization Gradient (HRG) from the simulated action potentials (APs), and the amplitudes, duration and areas of calculated unipolar electrograms (EGMs). Simulations showed that CV was highly influenced by reduced %CX43, whereas HRG area was more affected by increased fibrosis. The combination of both factors led to the highest decrease in CV, the largest amount of HRG area and the greatest dispersion of repolarization duration (ARI dispersion).

In conclusion, decreased %CX43 and increased fibrosis, to extents measured in human LV tissues contribute to a substrate for the generation of reentrant arrhythmias, which can be quantified from ventricular EGMs.

Jimena Rodríguez: ECG-based characterization of the extent, severity and spatial location of acute ischemia in patients with and without prior myocardial infarction

Ischemic heart disease is the leading cause of death in the world. Its diagnosis involves monitoring of the dynamic changes in the ST segment of the ECG, although changes in other intervals and waves of the ECG have been observed. One important aspect that is usually not considered is the presence of a prior myocardial infarction (prior MI) and how this affects the response to an ischemic event. Our aim was to quantify the changes in the ECG during acute myocardial ischemia and the effects a prior MI has on them. 12-lead ECG recordings acquired during and prior the acute ischemia induced by percutaneous coronary intervention (PCI) in one of the three main coronary arteries were analyzed. Averaged heartbeats were computed and ECG depolarization and repolarization features were extracted.

Our results show that, in addition to ST deviation, there is a temporal evolution in other ECG parameters during artery occlusion. Repolarization parameters show a faster and stronger change than depolarization ones. The relationship between the extent and severity of the ischemia and the ECG changes was more pronounced in patients with prior MI. A spatial lead profile was described as a function of the occluded artery and the presence of a prior MI. In conclusion, the presence of a prior MI affects the myocardial response to acute ischemia, resulting in more pronounced changes and a stronger relationship with the extension and severity of the ischemia.

Cristina Pérez: Evaluation of a QT Adaptation Time Estimator for ECG Exercise Stress Test in Controlled Simulation

Slowed adaptation of the QT interval to sudden abrupt heart rate (HR) changes has been identified as a marker of ventricular arrhythmic risk. However, abrupt HR changes are difficult to induce in patients. Quantifying the QT adaptation time in gradual HR changes, as observed in ECGs recording during an exercise stress test, has been recently proposed. The time lag between the QT series and an instantaneous memoryless HR-dependent QT series along stress test was computed as QT memory. Here, this method was evaluated in a control scenario using simulated exercise stress test ECG signals presenting different QT adaptation times. The method robustness was studied by contaminating the ECGs with muscular noise (MN) signals with different Signal-to-Noise ratio (SNR) values, either synthetic or extracted from real recordings. We found that delineation of the T-wave end point in the first transformed lead from Periodic Component Analysis offers the best performance for low SNR. Moreover, we confirmed that the estimator provides an unbiased estimate of the QT memory introduced in the simulations for the studied range of SNR values (25 to 50 dB).

Neurys Gómez: Changes in T-peak-to-T-end Morphology Measured by Time-Warping Are Associated with Ischemia-Induced Ventricular Fibrillation in a Porcine Model

In this work, we use a time-warping-based morphology variation index, dw, computed between the peak and the end of the T-wave, and assess its association with the occurrence of ventricular fibrillation (VF) episodes in ischemic conditions. ECG recordings from 26 pigs undergoing a 40-minute coronary occlusion were analyzed. The dw series was obtained by quantifying the morphological differences between the final part of the T wave at different stages of the occlusion and a reference T wave in the control recording. During control recordings, dw remained stationary with a median value along each recording of 1.76 ms, IQR of 1.80, while during artery occlusion followed a well-marked gradual increasing trend as ischemia progressed, with median of 15.47 ms, IQR of 18.53. At the 20-to-25 min period from occlusion onset (and during 5 min prior to VF episode) dw averages in the VF group was significantly higher than in the non-VF group with median values of 40.0 (and 34.4) vs 7.8 (and 7.7) ms, with p-values of 0.002 (and 0.001), respectively. In conclusion, dynamic increases of the dw index during ischaemia progression in pigs are associated with VF occurrence.

Pablo Armañac: Characterization of Cardiopulmonary Coupling in Pediatric Patients with Obstructive Sleep Apnea

This study aims to investigate the use of cardiopulmonary coupling (CPC), as biomarker for characterizing obstructive sleep apnea (OSA) severity in children. CPC analysis is based on the time-frequency coherence (TFC) between the respiratory effort signal and heart rate variability. We analyzed 255 children with no, mild, moderate, and severe OSA during wake, rapid eye movement (REM) sleep, and non-REM (NREM) sleep. Results showed that the TFC in the low-frequency (LF) band increased significantly with the severity of OSA in both NREM (p<0.001) and REM sleep (p<0.001). Conversely, the TFC in the HF band, the parameter estimating CPC, is significantly lower for increasing OSA severity categories during NREM (p=0.02) and REM (p=0.03). The findings suggest that TFC could be a useful biomarker for assessing OSA severity in children, and could provide additional information about underlying pathological mechanisms.

Julia Ramírez: A Multi-layer CNN using the ECG, Age and Sex Predicts Ventricular Arrhythmias in the General Population

Life-threatening ventricular arrhythmias (LTVA) prediction in individuals without cardiovascular disease remains a major challenge. We tested the performance of a multi-layer convolutional neural network (CNN) on ECG signals, including additional information (sex and gender) to predict LTVA.

We split 51,608 individuals from the UK Biobank study into training (90%) and internal test (10%) sets. In the training set, we trained a multi-layer CNN using 15-second ECGs at rest from lead I, age and gender as input. The output was the probability of LTVA within a 10-year follow-up. The CNN model consisted of a four-layer CNN (128, 128, 256 and 256 channels, kernel sizes of 3, groups of 1) and a single attention layer. Age and gender were included in the final layer. Performance was then tested in an external cohort of 32,209 individuals
from UK Biobank (3.4-year follow-up) with 10-second ECGs.

In the internal test cohort, 22 subjects had an LTVA, and the area under the curve (AUC) of the CNN was 0.760, with a specificity of 0.526 for a sensitivity of 0.750. In the external test cohort (60 LTVA events), the CNN’s prediction led to an AUC of 0.699, and a specificity of 0.551 for a sensitivity of 0.750. We set a threshold at the CNN’s prediction value maximising the sum of specificity and sensitivity above median values. Survival analyses showed a hazard ratio (HR) of 8.383 (P = 1.3 x 10-7) for individuals with a CNN’s prediction value > threshold, versus those with a CNN’s prediction value < threshold.

A multi-layer CNN model using 10-second ECG data from lead I, together with information on age and gender, can stratify individuals at risk of LTVA. Our findings support the potential utility of wearables for accessible screening in the general population.

Inés Noguero: ECG Morphology-Based Markers for Risk Stratification in Hypertrophic Cardiomyopathy

Hypertrophic cardiomyopathy (HCM) is the leading cause of sudden cardiac death in young adults. Current risk markers for this heterogeneous disease lack performance and, thus, new approaches are needed. This study aims to give more insight into risk assessment in HCM patients analyzing ECG-based markers in 24-hour Holter signals in a retrospective study dividing patients in asymptomatic, at-risk of a cardiac event and after a cardiac event. We studied conventional ECG markers such as RR interval, QRS width (QRSw) and corrected QT and T peak- T end intervals. ECG markers were computed from representative median beats for each patient in each hour. First, the median marker values in the 24 hours were compared between groups. Second, differences in the markers between day and night were studied. All patients showed marked circadian variations in RR and QT time series. Patients at risk of suffering cardiac events were found to have wider QRS complexes, with statistically significant differences between day and night. This QRS prolongation in HCM patients months before suffering a cardiac event might reflect anomalies in ventricular conduction. Regarding variations between day and night, they were slightly greater in patients before suffering an event.
This work provides new evidence on ECG-based markers and encourages further research on QRS-wave and T-wave assessment.

For further information: https://cinc2023.org/

News