Torso Model Research Articles

A two-step inverse solution for a single dipole cardiac source.

Introduction: The inverse problem of electrocardiography noninvasively localizes the origin of undesired cardiac activity, such as a premature ventricular contraction (PVC), from potential recordings from multiple torso electrodes. However, the optimal number and placement of electrodes for an accurate solution of the inverse problem remain undetermined. This study presents a two-step inverse solution for a single dipole cardiac source, which investigates the significance of the torso electrodes on a patient-specific level. Furthermore, the impact of the significant electrodes on the accuracy of the inverse solution is studied. Methods: Body surface potential recordings from 128 electrodes of 13 patients with PVCs and their corresponding homogeneous and inhomogeneous torso models were used. The inverse problem using a single dipole was solved in two steps: First, using information from all electrodes, and second, using a subset of electrodes sorted in descending order according to their significance estimated by a greedy algorithm. The significance of electrodes was computed for three criteria derived from the singular values of the transfer matrix that correspond to the inversely estimated origin of the PVC computed in the first step. The localization error (LE) was computed as the Euclidean distance between the ground truth and the inversely estimated origin of the PVC. The LE obtained using the 32 and 64 most significant electrodes was compared to the LE obtained when all 128 electrodes were used for the inverse solution. Results: The average LE calculated for both torso models and using all 128 electrodes was 28.8 ± 11.9mm. For the three tested criteria, the average LEs were 32.6 ± 19.9mm, 29.6 ± 14.7mm, and 28.8 ± 14.5mm when 32 electrodes were used. When 64 electrodes were used, the average LEs were 30.1 ± 16.8mm, 29.4 ± 12.0mm, and 29.5 ± 12.6mm. Conclusion: The study found inter-patient variability in the significance of torso electrodes and demonstrated that an accurate localization by the inverse solution with a single dipole could be achieved using a carefully selected reduced number of electrodes.

Anatomically Detailed Simulation of Human Torso

Many existing digital human models approximate the human skeletal system using rigid bodies connected by rotational joints. While the simplification is considered acceptable for legs and arms, it significantly lacks fidelity to model rich torso movements in common activities such as dancing, Yoga, and various sports. Research from biomechanics provides more detailed modeling for parts of the torso, but their models often operate in isolation and are not fast and robust enough to support computationally heavy applications and large-scale data generation for full-body digital humans. This paper proposes a new torso model that aims to achieve high fidelity both in perception and in functionality, while being computationally feasible for simulation and optimal control tasks. We build a detailed human torso model consisting of various anatomical components, including facets, ligaments, and intervertebral discs, by coupling efficient finite-element and rigid-body simulations. Given an existing motion capture sequence without dense markers placed on the torso, our new model is able to recover the underlying torso bone movements. Our method is remarkably robust that it can be used to automatically "retrofit" the entire Mixamo motion database of highly diverse human motions without user intervention. We also show that our model is computationally efficient for solving trajectory optimization of highly dynamic full-body movements, without relying any reference motion. Physiological validity of the model is validated against established literature.

Predictive analysis of the scoliotic curve using a subject's 3D model.

A predictive analysis of the conservative scoliosis treatment is necessary, in which a 3D model of an optimal treatment algorithm is a basic part in the design of a prosthetic corset. Since CAD technology has proven to be very useful in the field of prosthetics and orthotics, we used an open-source software to plan the correction of the scoliotic curve on a virtual model of the subject's torso. The shape of the scoliosis was simplified by means of a directional polygon, which was drawn in a reverse manner depending on the directional arcs of the scoliotic curve. The resulting scoliosis correction, simulated in a predictive analysis, was defined by changing the Cobb angle, eccentricity, and torso height. With the proposed low-cost method of predictive analysis, it is possible to help CPOs to a more accurate and effective design of orthoses and corrective aids and to comprehensively determine the entire treatment procedure.

Comparison of dipole-based and potential-based ECGI methods for premature ventricular contraction beat localization with clinical data.

Introduction: Localization of premature ventricular contraction (PVC) origin to guide the radiofrequency ablation (RFA) procedure is one of the prominent clinical goals of non-invasive electrocardiographic imaging. However, the results reported in the literature vary significantly depending on the source model and the level of complexity in the forward model. This study aims to compare the paced and spontaneous PVC localization performances of dipole-based and potential-based source models and corresponding inverse methods using the same clinical data and to evaluate the effects of torso inhomogeneities on these performances. Methods: The publicly available EP solution data from the EDGAR data repository (BSPs from a maximum of 240 electrodes) with known pacing locations and the Bratislava data (BSPs in 128 leads) with spontaneous PVCs from patients who underwent successful RFA procedures were used. Homogeneous and inhomogeneous torso models and corresponding forward problem solutions were used to relate sources on the closed epicardial and epicardial-endocardial surfaces. The localization error (LE) between the true and estimated pacing site/PVC origin was evaluated. Results: For paced data, the median LE values were 25.2 and 13.9mm for the dipole-based and potential-based models, respectively. These median LE values were higher for the spontaneous PVC data: 30.2-33.0mm for the dipole-based model and 28.9-39.2mm for the potential-based model. The assumption of inhomogeneities in the torso model did not change the dipole-based solutions much, but using an inhomogeneous model improved the potential-based solutions on the epicardial-endocardial ventricular surface. Conclusion: For the specific task of localization of pacing site/PVC origin, the dipole-based source model is more stable and robust than the potential-based source model. The torso inhomogeneities affect the performances of PVC origin localization in each source model differently. Hence, care must be taken in generating patient-specific geometric and forward models depending on the source model representation used in electrocardiographic imaging (ECGI).

Additional coils mitigate elevated defibrillation threshold in right-sided implantable cardioverter defibrillator generator placement: a simulation study.

The standard implantable cardioverter defibrillator (ICD) generator (can) is placed in the left pectoral area; however, in certain circumstances, right-sided cans may be required which may increase defibrillation threshold (DFT) due to suboptimal shock vectors. We aim to quantitatively assess whether the potential increase in DFT of right-sided can configurations may be mitigated by alternate positioning of the right ventricular (RV) shocking coil or adding coils in the superior vena cava (SVC) and coronary sinus (CS). A cohort of CT-derived torso models was used to assess DFT of ICD configurations with right-sided cans and alternate positioning of RV shock coils. Efficacy changes with additional coils in the SVC and CS were evaluated. A right-sided can with an apical RV shock coil significantly increased DFT compared to a left-sided can [19.5 (16.4, 27.1) J vs. 13.3 (11.7, 19.9) J, P < 0.001]. Septal positioning of the RV coil led to a further DFT increase when using a right-sided can [26.7 (18.1, 36.1) J vs. 19.5 (16.4, 27.1) J, P < 0.001], but not a left-sided can [12.1 (8.1, 17.6) J vs. 13.3 (11.7, 19.9) J, P = 0.099). Defibrillation threshold of a right-sided can with apical or septal coil was reduced the most by adding both SVC and CS coils [19.5 (16.4, 27.1) J vs. 6.6 (3.9, 9.9) J, P < 0.001, and 26.7 (18.1, 36.1) J vs. 12.1 (5.7, 13.5) J, P < 0.001]. Right-sided, compared to left-sided, can positioning results in a 50% increase in DFT. For right-sided cans, apical shock coil positioning produces a lower DFT than septal positions. Elevated right-sided can DFTs may be mitigated by utilizing additional coils in SVC and CS.

Numerical simulation of Non-Lethal Kinetic Energy Projectile Impact on Human Torso Model

In order to analyse the dynamic response and damage characteristics of human tissues and organs under the non-lethal kinetic energy projectiles, a numerical simulation model of human torso is established based on the Chinese visual human (CVH) data sets. By the dummy established with CVH data sets, the calculation accuracy of movement of human torso impact by non-Lethal kinetic energy projectile of finite element model were validated, and the skin perforation performance were validated by animal test. The non-lethal kinetic energy projectile, which is rubber ball in diameter 8mm and 50 shores hardness, impact on human torso with different velocity were calculated, the dynamic responses of sternum, rib and lung were captured and analysed. The results shows that, if the projectile speed is near the probability perforation velocity, due to the hardness of projectile and structure of organs and ribs, the projectile will rebound no matter the skin is perforation or not, and the vulnerable organ lung may not be damaged. At the same time, the pressure of lungs without skin penetration were significantly higher than skin without penetration, the human damage risk under soft projectiles is greater than hard projectiles when skin with damage but without perforation.

Study of blunt impact effects of rifle bullet penetrating the SiC/UHMWPE armor covered human torso

To study the behind armor blunt trauma (BABT), a finite element model of a 5.8 mm rifle bullet penetrating the SiC/UHMWPE armor covered human torso was established. The human torso model comprises internal organs (heart, lung, liver, and stomach), skeleton (sternum, ribs, and spine), and muscle. It is assumed that the skeleton and muscle is linear elastic, and all internal organs are viscoelastic and hyperelastic. The impact response of a rifle bullet penetrating a human target with bulletproof armor at a speed of 900m/s is numerically simulated. The deformation data of bullet and body armor were obtained. In addition, the von Mises stresses of muscle, skeleton, and internal organs were analyzed. The numerical model provides an effective method for studying BABT and the improvement of bulletproof armor.

PO-01-207 VERNIX CASEOSA DISTRIBUTION MODULATES NON-INVASIVE FETAL ECG SIGNAL: A COMPUTATIONAL STUDY

Non-invasive fetal ECG (FECG) shows promise as a diagnostic method to monitor fetal health. However, extracting the fetal cardiac signal remains a challenge due to low signal-to-noise ratio. An important factor is a low-conducting layer which forms around the fetus from week 28 of gestation called the vernix caseosa (VC). We developed an anatomically detailed and biophysically accurate computational model of a pregnant torso to investigate the effects of the VC on the FECG. The model was constructed by combining geometries of a pregnant woman with a 24 week infant and a torso with ventricles (a). Ionic current properties of the maternal heart is modelled using the Ten Tusscher human ventricular cardiomyocyte model, while a modified version adapted to match fetal electrophysiology is used for the fetal heart. 22 distribution patterns of the VC was generated using combinations of 3 regions and 4 thickness. No VC case was used as control. Electrical propagation was simulated using the pseudo-bidomain formulation. Sinus rhythm was simulated with a baseline cycle of 500 ms and 450 ms for the maternal and fetal hearts, respectively. When VC was represented as homogeneous, changing the VC thickness from 0.5 mm to 3 mm resulted in reduction of QRS maximum amplitude from 6 x 10-5 to 6 x 10-6 mV (c). However, in the heterogeneous VC models, there is no significant difference in signal as thickness was changed (d). The location of the VC significantly affected the signal (e). VC only covering the head was almost identical to control with a relative difference measure of only 1.35%. However, introduction of VC on the back resulted in greatest difference in potential distribution with a relative difference measure ranging between 34.5 and 49%. We have developed a novel, highly-detailed computational model of a pregnant torso that shows how VC distribution affects the FECG. This model can be used to improve techniques in extracting FECG at different gestational stages.

Open-source software library for real-time inertial measurement unit data-based inverse kinematics using OpenSim.

Inertial measurements (IMUs) facilitate the measurement of human motion outside the motion laboratory. A commonly used open-source software for musculoskeletal simulation and analysis of human motion, OpenSim, includes a tool to enable kinematics analysis of IMU data. However, it only enables offline analysis, i.e., analysis after the data has been collected. Extending OpenSim's functionality to allow real-time kinematics analysis would allow real-time feedback for the subject during the measurement session and has uses in e.g., rehabilitation, robotics, and ergonomics. We developed an open-source software library for real-time inverse kinematics (IK) analysis of IMU data using OpenSim. The software library reads data from IMUs and uses multithreading for concurrent calculation of IK. Its operation delays and throughputs were measured with a varying number of IMUs and parallel computing IK threads using two different musculoskeletal models, one a lower-body and torso model and the other a full-body model. We published the code under an open-source license on GitHub. A standard desktop computer calculated full-body inverse kinematics from treadmill walking at 1.5 m/s with data from 12 IMUs in real-time with a mean delay below 55 ms and reached a throughput of more than 90 samples per second. A laptop computer had similar delays and reached a throughput above 60 samples per second with treadmill walking. Minimal walking kinematics, motion of lower extremities and torso, were calculated from treadmill walking data in real-time with a throughput of 130 samples per second on the laptop and 180 samples per second on the desktop computer, with approximately half the delay of full-body kinematics. The software library enabled real-time inverse kinematical analysis with different numbers of IMUs and customizable musculoskeletal models. The performance results show that subject-specific full-body motion analysis is feasible in real-time, while a laptop computer and IMUs allowed the use of the method outside the motion laboratory.

Smart Wireless-Powering-Enabling IoT in Inhomogeneous Environments: A Case Study on Biomedical Applications

Wireless powering holds the key to unlock the use of batteryless micro/nanodevices inside the human body for research, monitoring, and treating health conditions, and enables precise and personalized healthcare technological solutions. As the era of IoT emerges, wireless devices become more prevalent and capable than ever. In the medical field, this transformation is also highly anticipated. From in-body networks of tiny static devices, to freely moving implants inside the body, e.g., using the blood stream, the possibilities are endless. Their limits, however, are miniaturization capabilities and power availability. Despite the great potential, most wireless-powering solutions are developed for a single static device in a virtually immutable environment. In contrast, we present a system with the ability to power different static places or a moving device, without prior knowledge of its location nor the medium. The system focuses power on the target location(s) through beamforming. The methodology introduced by this work increased the power provided to a wireless test device by up to 18 dB in torso and leg models, and 25 dB in a skull model, relatively to a traditional wireless power transfer (WPT) system and without surpassing specific absorption rate (SAR) limits.

Non-invasive localization of post-infarct ventricular tachycardia exit sites to guide ablation planning: a computational deep learning platform utilizing the 12-lead electrocardiogram and intracardiac electrograms from implanted devices.

Existing strategies that identify post-infarct ventricular tachycardia (VT) ablation target either employ invasive electrophysiological (EP) mapping or non-invasive modalities utilizing the electrocardiogram (ECG). Their success relies on localizing sites critical to the maintenance of the clinical arrhythmia, not always recorded on the 12-lead ECG. Targeting the clinical VT by utilizing electrograms (EGM) recordings stored in implanted devices may aid ablation planning, enhancing safety and speed and potentially reducing the need of VT induction. In this context, we aim to develop a non-invasive computational-deep learning (DL) platform to localize VT exit sites from surface ECGs and implanted device intracardiac EGMs. A library of ECGs and EGMs from simulated paced beats and representative post-infarct VTs was generated across five torso models. Traces were used to train DL algorithms to localize VT sites of earliest systolic activation; first tested on simulated data and then on a clinically induced VT to show applicability of our platform in clinical settings. Localization performance was estimated via localization errors (LEs) against known VT exit sites from simulations or clinical ablation targets. Surface ECGs successfully localized post-infarct VTs from simulated data with mean LE = 9.61 ± 2.61 mm across torsos. VT localization was successfully achieved from implanted device intracardiac EGMs with mean LE = 13.10 ± 2.36 mm. Finally, the clinically induced VT localization was in agreement with the clinical ablation volume. The proposed framework may be utilized for direct localization of post-infarct VTs from surface ECGs and/or implanted device EGMs, or in conjunction with efficient, patient-specific modelling, enhancing safety and speed of ablation planning.

Effect of torso reflections from simplified torso models on head-related transfer function simulation and ipsilateral perception of elevation

The effect of torso reflection on head-related transfer functions (HRTFs) is considered a suitable elevation localization cue for lateral sound sources below 3.0 kHz, and various torso models have been hitherto adopted in HRTF simulation. However, their torso reflection effects have not been compared in detail. This work analyzes the torso reflection effects of the representative torso models on HRTFs, and further evaluates their elevation localization performance via psychoacoustic experiments. The results show that the semi-ellipsoidal torso and truncated Knowles Electronics Mannequin for Acoustic Research (KEMAR) torso, both of which have considerable advantages over the spherical torso, can achieve better elevation localization performance, which validates the torso with human-like shape provide enough torso reflection and have a better ability in low-frequency lateral elevation perception.

The Importance of ECG Offset Correction for Premature Ventricular Contraction Origin Localization from Clinical Data

Abstract In this study, the inverse solution with a single dipole was computed to localize the premature ventricular contraction (PVC) origin from long term multiple leads ECG measurements on fourteen patients. The stability of the obtained results was studied with respect to the preprocessing of signals used as an input to the inverse solution and the complexity of the torso model. Two methods were used for the baseline drift removal. After an averaging of the heartbeats, the influence of the retention or elimination of the remaining offset at the beginning of the PVC signal was examined. The inverse computations were performed using both homogeneous and inhomogeneous patient-specific torso models. It was shown that the remaining offset in the averaged signals at the beginning of the PVC signal had the most significant impact on the stability of the resulting position within the ventricles. Its elimination stabilizes the location of the results, decreases the sensitivity to the torso model complexity and decreases the sensitivity to the primary baseline drift removal method. The additional offset correction decreased the mean distance between the results for all patients from 17-18 mm to 1-2 mm, regardless of the baseline drift removal method or the torso model complexity.

An ECG generative model of myocardial infarction

Background and Objective Computer-aided diagnosis (CAD) of Myocardial Infarction (MI) using machine learning depends on a large amount of clinical Electrocardiogram (ECG) data. Existing infarct ECG databases face the problem of class imbalance. Data augmentation using generative simulation models is a new approach to effectively address this problem.Methods A multiscale ECG generative model was established for ECG data augmentation. In the cellular layer, an ischemic Action Potential (AP) model was established to generate APs in cardiomyocytes with different transmural regions of infraction or different ischemic durations. In the tissue layer, a probability-driven cellular automata excitation propagation model was established to simulate the propagation speed and direction of excitation. An infarct tissue model and a coronary artery model were established to describe the spatiotemporal diversity of MI. A ventricle model, a human torso model, and a computational model of surface ECG based on field source theory were established in the heart-torso layer.Results The model generated pathological 12-lead ECGs of MI with different topography and different extent. When simulating different ventricular wall infarction, the lesions appear in the same leads as the clinical 12-lead ECG. The ST-segment decreases and the T-wave amplitude decreases, similar to the clinical ECG features when simulating subendocardial ischemia. The average fidelity of the 12-lead ECG the model generated is 95.6%, according to the designed DTW-GRA distance algorithm.Conclusions The generative model considers the electrophysiological properties of the natural heart, the pathology of myocardial infarction, and the diversity of clinical ECGs. The model can provide many reliable samples for machine learning of MI.

Design and Analysis of a 5G Wideband Antenna for Wireless Body-Centric Network

A compact 5G wideband antenna for body-centric network (BCN) operating on Ka band has been presented in this paper. The design of the antenna consists of a very simple key-shaped radiator patch with a vertical slot for better impedance matching. The antenna was designed and simulated with the help of the Computer Simulation Technology (CST) Microwave Studio Suite, a well-liked and dependable electromagnetic simulation program running on Microsoft Windows. Free-space simulation produces a resonant frequency at 28 GHz, which falls under the Ka band and 5G’s n257, more precisely n261. The proposed antenna has a size of 1.24 λ × 0.6 λ × 0.153 λ and has a wider impedance bandwidth of more than 20 GHz. The antenna’s gain and radiation efficiency are 3.87 dBi and 70%, respectively, at the resonant point. Further parametric studies reveal that the antenna can be activated in the V-band by increasing the feedline width. The antenna is proposed for the application of BCN. Therefore, a three-dimensional human torso phantom was developed virtually to test on-body performance. The on-body findings of this antenna were resimulated by positioning the antenna in close proximity to the three-layer human body model, where 22.5 dB of on-body reflection coefficient was recorded at 28 GHz. Simulated on-body gain and efficiency were 4.56 dBi and 61.33 percent, respectively. A distance-based investigation was conducted to investigate the impacts of the human body’s presence by positioning the antenna at five different distances from the human torso model. The findings were compared to assess how distance affects its behaviors. The antenna’s gap was kept at 6 mm for the optimum results, which included 4.83 dBi of gain with a 66 percent efficiency and a recorded RL value of about 23 dB. The on-body simulations produced very consistent results with a slight deviation after 26.5 GHz, even though the distance was varied.

A Microwave Imaging Procedure for Lung Lesion Detection: Preliminary Results on Multilayer Phantoms

In this work, a feasibility study for lung lesion detection through microwave imaging based on Huygens’ principle (HP) has been performed using multilayer oval shaped phantoms mimicking human torso having a cylindrically shaped inclusion simulating lung lesion. First, validation of the proposed imaging method has been performed through phantom experiments using a dedicated realistic human torso model inside an anechoic chamber, employing a frequency range of 1–5 GHz. Subsequently, the miniaturized torso phantom validation (using both single and double inclusion scenarios) has been accomplished using a microwave imaging (MWI) device, which operates in free space using two antennas in multi-bistatic configuration. The identification of the target’s presence in the lung layer has been achieved on the obtained images after applying both of the following artifact removal procedures: (i) the “rotation subtraction” method using two adjacent transmitting antenna positions, and (ii) the “ideal” artifact removal procedure utilizing the difference between received signals from unhealthy and healthy scenarios. In addition, a quantitative analysis of the obtained images was executed based on the definition of signal to clutter ratio (SCR). The obtained results verify that HP can be utilized successfully to discover the presence and location of the inclusion in the lung-mimicking phantom, achieving an SCR of 9.88 dB.

Wall-Meshed Cavity Resonator-Based Wireless Power Transfer Without Blocking Wireless Communications With Outside World

In order to address the issue of wireless communication signal blocking in the traditional cavity resonance wireless power transfer (CR WPT), a wall-meshed cavity resonator constructed of the meshed metallic walls is proposed together with an analytical design method. This cavity resonator can not only generate and confine the electromagnetic waves of natural resonant modes inside the cavity resonator, but also support wireless communications with outside. The wall-meshed cavity resonator is analytically designed by the known theoretical equations without time-consuming optimizations and verified by electromagnetic simulations and experiments. The results show that the proposed CR WPT system can achieve simultaneous wireless power transfer in the cavity and wireless communications with outside. The maximum power transfer efficiency is higher than 85%. The power transfer efficiency exceeds 60% within about 70% areas inside the wall-meshed cavity. For wireless communication signals passing through the wall-meshed cavity, the attenuation at 2.1 GHz (4G), 2.45 GHz (Wi-Fi), and 3.5 GHz (5G) is about 6.0 dB, 5.1 dB, and 3.8 dB, respectively. Additionally, the safety performance of the proposed CR WPT system is numerically investigated by the specific absorption rate analysis of a human torso model.

Novel non-invasive 12-lead ECG-based imaging method with potential to guide and optimise right ventricular lead implantation

Abstract Funding Acknowledgements Type of funding sources: Private grant(s) and/or Sponsorship. Main funding source(s): Svend Andersens Foundation Karl G Andersens Foundation Helsefonden Background Right ventricular (RV) pacing may induce electrical and mechanical dyssynchrony which may lead to heart failure. Often, the physician aims for lead implantation at the RV septum as this is expected to result in a more physiologic activation compared to alternative RV lead locations. Fluoroscopy, often in combination with QRS morphology derived from the 12-lead ECG, is used to guide lead implantation. However, this method is inaccurate and can result in a non-optimal RV lead location. We present a novel non-invasive method to create patient-specific 3D electrical activation maps from the 12-lead ECG that has potential to support RV lead implantation. Methods Data from 34 patients with an implanted dual chamber pacemaker were used. A contrast-enhanced cardiac CT scan showing the RV lead implantation site was obtained as well as recording of a 12-lead ECG during RV pacing together with a 3D photo documenting the ECG electrode positions. Discrete patient-specific torso and heart models were created from the CT scans. Each torso model was merged with the 3D photo for precise placement of the ECG electrodes in relation to the heart (figure 1). Combining the 12-lead ECGs and the heart/torso models, patient-specific 3D electrical activation maps originating from the RV were created using a novel inverse-ECG technique applying electrophysiological rules. The accuracy of the inverse-ECG method was determined by comparing the earliest site of activation from the 3D activation map with the known RV insertion site marked on the CT scan. Results Documented by the implanting physician in the medical records, 33 RV leads were estimated to be septal and one apical. Estimated from the CT scan 9 leads were placed septal, 18 apical and 7 on the free wall. The mean geodesic distance between the initial site of activation in the 3D activation map and the marked RV insertion site from CT was 13.6 ±5.7 mm (range 4.3-28.6). The distance for each patient is shown in figure 2. The initial site of activation was constrained to the discrete nodes of the ventricular model whereas the marker for RV lead position was localized freely on the CT scan. The average distance from the RV CT marker to the nearest discrete node was 4.3 ±2.2 mm. Correcting for this error, the geodesic distance between the initial site of activation and RV CT marker was 9.3 ±5.4 mm (range 0.0-24.6). The average time used for 3D activation map computation was 1.1 ±0.4 s per ECG. Conclusions We demonstrated a novel non-invasive 12-lead ECG-based method to accurately and effectively localize the RV lead in relation to the ventricular anatomy during RV pacing. Furthermore, we confirmed that the RV lead was often implanted in an unintended position. With further advancements, this method has the potential to support physicians during pacemaker implantation to ensure optimal RV lead positioning. Further studies are needed to confirm the accuracy.

Automated detection of scar-related ventricular tachycardia origins from implanted device electrograms: a combined computational-AI platform

Abstract Funding Acknowledgements Type of funding sources: Public Institution(s). Main funding source(s): EPSRC Background Existing strategies that identify ventricular tachycardia (VT) ablation targets either employ invasive and time-consuming electrophysiological (EP) mapping, or non-invasive modalities that utilise standard electrocardiogram (ECG) signals. Success of these pre-procedure ablation approaches in localising re-entrant VTs often relies on VT induction, which could be avoided by utilising recordings of clinical VT episodes stored as electrograms (EGMs) in implanted devices. Such a non-invasive approach that localises VT substrates from EGMs may aid ablation planning, enhancing safety and speed. Purpose Our goal is to automate scar-related VT localisation by utilising EGM recordings of VT episodes from implanted devices. To achieve this, deep-learning algorithms will be trained on computational data to return VT sites of origin from implanted device EGMs. Ultimately, we intend to utilise this computational-artificial intelligence (AI) framework to detect ablation targets of clinical VT episodes and guide pre-procedure ablation planning non-invasively. Methods A comprehensive library of ECGs and EGMs from simulated paced beats (~15000) and scar-related VTs (500) was generated across five detailed torso models within a fast EP computational environment, combining reaction-eikonal and lead field methods. ECG (or EGM) traces from simulated paced beats were used to initially pre-train two convolutional neural network (CNN) long short-term (LSTM) attention-based architectures. Subsequently, signals of the in-silico, re-entrant VTs were used to re-train the networks to output the sites of origin of these episodes in a standardised ventricular coordinate space. Finally, the retrained CNN architectures were tested on re-entrant VTs of unseen models, and median localisation errors (LEs) were estimated against known VT origins from simulations. Results The performance of the networks to localise scar-related VT episodes was asserted for each torso model. When a torso model was only seen during initial training on simulated paced beats, implanted device EGMs and ECGs successfully localised VT sources with LEs 10.04 – 16.36 mm and 10.05 – 12.79 mm, respectively. When a torso model was not seen during pacing or VT training, recreating potential clinical application settings where ECGs or EGMs of clinical VTs would be the only inputs to the networks, LEs ranged 12.42 - 22.79 mm and 12.41 - 19.68 mm for EGM and ECG-based testing, respectively. Conclusions Our study successfully detected VT ablation substrates with accuracy that could be beneficial in clinical ablation settings. The proposed computational-AI framework may be used to automate the localisation of scar-related VTs from clinical ECGs or EGM recordings from implanted devices, ultimately aiding ablation planning.

In-silico optimisation of ICD defibrillation efficacy by modifying lead/can configurations using a cohort of high-resolution whole-torso heart models

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Medical Research Council, UK Background ICD is an effective direct therapy against VT/VF by applying a strong electrical shock across the heart between the shocking coil and can. Conventionally, patients will have a shocking coil inside the right ventricle (RV) and a can at the upper left chest. However, due to infections or other conditions, the can may need to place towards the right chest. The placement of the RV coil may also vary in different cases, for example avoiding scar. However, it is unclear how defibrillation efficacy may be altered by these unavoidable modifications to conventional lead/can configurations and whether optimisation may be possible. Purpose To compare defibrillation efficacy of modifications of ICD configurations in a cohort of whole-torso models. Methods A cohort of 15 whole torso models was generated from high resolution CT data and contrast CT cardiac scans, including 5 dilated cardiomyopathy (DCM), 5 hypertrophic cardiomyopathy (HCM) and 5 structurally normal patients (Fig A). Transvenous ICDs were represented by a shocking coil inside the RV (near apex) and a (ground) can at the upper left chest as default settings. Configurations were then varied by moving the can to the right chest, moving the RV coil up the mid-septum or adding extra grounds (Superior Vena Cava (SVC) coil, coronary sinus (CS) coil (Fig A)). Defibrillation-strength shocks were applied to all models (Fig B). DFTs and mean electrical field were evaluated across the whole heart as well as specific LV, RV, RV insertion regions, along with overall impedance. Results Shifting the can from left to right significantly increased DFT for the whole heart (23 J vs 15 J, P=0.03) and LV (25 J vs 17 J, P=0.03) (Fig C) and reduced the mean electrical field. Moving the RV coil further up the septum did not significantly alter DFT (Fig D), but did reduce mean electrical field for all regions and reduce impedance significantly. Additional separate coils significantly reduced DFT for all regions (Fig D) by increasing mean electrical field, whilst adding both coils significantly reduced DFT the most (whole heart: 15 J vs 6 J, P=0.03) (Fig E). Impedance was increased significantly by adding SVC coil, but reduced significantly by adding CS coil. Adding both coils increased impedance slightly. Conclusions Although a right-sided can increases DFT by over 50%, additional leads (grounds) may mitigate this increase by increasing mean electrical field. Moving the RV coil closer to the mid-septum reduces DFT slightly, but also reduces mean electrical field and impedance significantly.