Training machine learning models with synthetic data improves the prediction of ventricular origin in outflow tract ventricular arrhythmias

Computational modelling of biological systems now and then: revisiting tools and visions from the beginning of the century

Since the turn of the millennium, computational modelling of biological systems has evolved remarkably and sees matured use spanning basic and clinical research. While the topic of the peri-millennial debate about the virtues and limitations of 'reductionism and integrationism' seems less controversial today, a new apparent dichotomy dominates discussions: mechanistic versus data-driven modelling. In light of this distinction, we provide an overview of recent achievements and new challenges with a focus on the cardiovascular system. Attention has shifted from generating a universal model of the human to either models of individual humans (digital twins) or entire cohorts of models representative of clinical populations to enable in silico clinical trials. Disease-specific parametrization, inter-individual and intra-individual variability, uncertainty quantification as well as interoperable, standardized and quality-controlled data are important issues today, which call for open tools, data and metadata standards, as well as strong community interactions. The quantitative, biophysical and highly controlled approach provided by in silico methods has become an integral part of physiological and medical research. In silico methods have the potential to accelerate future progress also in the fields of integrated multi-physics modelling, multi-scale models, virtual cohort studies and machine learning beyond what is feasible today. In fact, mechanistic and data-driven modelling can complement each other synergistically and fuel tomorrow's artificial intelligence applications to further our understanding of physiology and disease mechanisms, to generate new hypotheses and assess their plausibility, and thus to contribute to the evolution of preventive, diagnostic and therapeutic approaches.This article is part of the theme issue 'Science into the next millennium: 25 years on'.

Digital Twins for Clinical and Operational Decision-Making: Scoping Review

BACKGROUND

The health care industry must align with new digital technologies to respond to existing and new challenges. Digital twins (DTs) are an emerging technology for digital transformation and applied intelligence that is rapidly attracting attention. DTs are virtual representations of products, systems, or processes that interact bidirectionally in real time with their actual counterparts. Although DTs have diverse applications from personalized care to treatment optimization, misconceptions persist regarding their definition and the extent of their implementation within health systems.

OBJECTIVE

This study aimed to review DT applications in health care, particularly for clinical decision-making (CDM) and operational decision-making (ODM). It provides a definition and framework for DTs by exploring their unique elements and characteristics. Then, it assesses the current advances and extent of DT applications to support CDM and ODM using the defined DT characteristics.

METHODS

We conducted a scoping review following the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) protocol. We searched multiple databases, including PubMed, MEDLINE, and Scopus, for original research articles describing DT technologies applied to CDM and ODM in health systems. Papers proposing only ideas or frameworks or describing DT capabilities without experimental data were excluded. We collated several available types of information, for example, DT characteristics, the environment that DTs were tested within, and the main underlying method, and used descriptive statistics to analyze the synthesized data.

RESULTS

Out of 5537 relevant papers, 1.55% (86/5537) met the predefined inclusion criteria, all published after 2017. The majority focused on CDM (75/86, 87%). Mathematical modeling (24/86, 28%) and simulation techniques (17/86, 20%) were the most frequently used methods. Using International Classification of Diseases, 10th Revision coding, we identified 3 key areas of DT applications as follows: factors influencing diseases of the circulatory system (14/86, 16%); health status and contact with health services (12/86, 14%); and endocrine, nutritional, and metabolic diseases (10/86, 12%). Only 16 (19%) of 86 studies tested the developed system in a real environment, while the remainder were evaluated in simulated settings. Assessing the studies against defined DT characteristics reveals that the developed systems have yet to materialize the full capabilities of DTs.

CONCLUSIONS

This study provides a comprehensive review of DT applications in health care, focusing on CDM and ODM. A key contribution is the development of a framework that defines important elements and characteristics of DTs in the context of related literature. The DT applications studied in this paper reveal encouraging results that allow us to envision that, in the near future, they will play an important role not only in the diagnosis and prevention of diseases but also in other areas, such as efficient clinical trial design, as well as personalized and optimized treatments.

Definitions and Characteristics of Patient Digital Twins Being Developed for Clinical Use: Scoping Review

BACKGROUND

The concept of digital twins, widely adopted in industry, is entering health care. However, there is a lack of consensus on what constitutes the digital twin of a patient.

OBJECTIVE

The objective of this scoping review was to analyze definitions and characteristics of patient digital twins being developed for clinical use, as reported in the scientific literature.

METHODS

We searched PubMed, Scopus, Embase, IEEE, and Google Scholar for studies claiming digital twin development or evaluation until August 2023. Data on definitions, characteristics, and development phase were extracted. Unsupervised classification of claimed digital twins was performed.

RESULTS

We identified 86 papers representing 80 unique claimed digital twins, with 98% (78/80) in preclinical phases. Among the 55 papers defining "digital twin," 76% (42/55) described a digital replica, 42% (23/55) mentioned real-time updates, 24% (13/55) emphasized patient specificity, and 15% (8/55) included 2-way communication. Among claimed digital twins, 60% (48/80) represented specific organs (primarily heart: 15/48, 31%; bones or joints: 10/48, 21%; lung: 6/48, 12%; and arteries: 5/48, 10%); 14% (11/80) embodied biological systems such as the immune system; and 26% (21/80) corresponded to other products (prediction models, etc). The patient data used to develop and run the claimed digital twins encompassed medical imaging examinations (35/80, 44% of publications), clinical notes (15/80, 19% of publications), laboratory test results (13/80, 16% of publications), wearable device data (12/80, 15% of publications), and other modalities (32/80, 40% of publications). Regarding data flow between patients and their virtual counterparts, 16% (13/80) claimed that digital twins involved no flow from patient to digital twin, 73% (58/80) used 1-way flow from patient to digital twin, and 11% (9/80) enabled 2-way data flow between patient and digital twin. Based on these characteristics, unsupervised classification revealed 3 clusters: simulation patient digital twins in 54% (43/80) of publications, monitoring patient digital twins in 28% (22/80) of publications, and research-oriented models unlinked to specific patients in 19% (15/80) of publications. Simulation patient digital twins used computational modeling for personalized predictions and therapy evaluations, mostly for one-time assessments, and monitoring digital twins harnessed aggregated patient data for continuous risk or outcome forecasting and care optimization.

CONCLUSIONS

We propose defining a patient digital twin as "a viewable digital replica of a patient, organ, or biological system that contains multidimensional, patient-specific information and informs decisions" and to distinguish simulation and monitoring digital twins. These proposed definitions and subtypes offer a framework to guide research into realizing the potential of these personalized, integrative technologies to advance clinical care.

Enhancing origin prediction: deep learning model for diagnosing premature ventricular contractions with dual-rhythm analysis focused on cardiac rotation

AIMS

Several algorithms can differentiate inferior axis premature ventricular contractions (PVCs) originating from the right side and left side on 12-lead electrocardiograms (ECGs). However, it is unclear whether distinguishing the origin should rely solely on PVC or incorporate sinus rhythm (SR). We compared the dual-rhythm model (incorporating both SR and PVC) to the PVC model (using PVC alone) and quantified the contribution of each ECG lead in predicting the PVC origin for each cardiac rotation.

METHODS AND RESULTS

This multicentre study enrolled 593 patients from 11 centres-493 from Japan and Germany, and 100 from Belgium, which were used as the external validation data set. Using a hybrid approach combining a Resnet50-based convolutional neural network and a transformer model, we developed two variants-the PVC and dual-rhythm models-to predict PVC origin. In the external validation data set, the dual-rhythm model outperformed the PVC model in accuracy (0.84 vs. 0.74, respectively; P < 0.01), precision (0.73 vs. 0.55, respectively; P < 0.01), specificity (0.87 vs. 0.68, respectively; P < 0.01), area under the receiver operating characteristic curve (0.91 vs. 0.86, respectively; P = 0.03), and F1-score (0.77 vs. 0.68, respectively; P = 0.03). The contributions to PVC origin prediction were 77.3% for PVC and 22.7% for the SR. However, in patients with counterclockwise rotation, SR had a greater contribution in predicting the origin of right-sided PVC.

CONCLUSION

Our deep learning-based model, incorporating both PVC and SR morphologies, resulted in a higher prediction accuracy for PVC origin, considering SR is particularly important for predicting right-sided origin in patients with counterclockwise rotation.

Evaluation of five methods for the interpolation of bad leads in the solution of the inverse electrocardiography problem

Objective.This study aims to assess the sensitivity of epicardial potential-based electrocardiographic imaging (ECGI) to the removal or interpolation of bad leads.Approach.We utilized experimental data from two distinct centers. Langendorff-perfused pig (n= 2) and dog (n= 2) hearts were suspended in a human torso-shaped tank and paced from the ventricles. Six different bad lead configurations were designed based on clinical experience. Five interpolation methods were applied to estimate the missing data. Zero-order Tikhonov regularization was used to solve the inverse problem for complete data, data with removed bad leads, and interpolated data. We assessed the quality of interpolated ECG signals and ECGI reconstructions using several metrics, comparing the performance of interpolation methods and the impact of bad lead removal versus interpolation on ECGI.Main results.The performance of ECG interpolation strongly correlated with ECGI reconstruction. The hybrid method exhibited the best performance among interpolation techniques, followed closely by the inverse-forward and Kriging methods. Bad leads located over high amplitude/high gradient areas on the torso significantly impacted ECGI reconstructions, even with minor interpolation errors. The choice between removing or interpolating bad leads depends on the location of missing leads and confidence in interpolation performance. If uncertainty exists, removing bad leads is the safer option, particularly when they are positioned in high amplitude/high gradient regions. In instances where interpolation is necessary, the inverse-forward and Kriging methods, which do not require training, are recommended.Significance.This study represents the first comprehensive evaluation of the advantages and drawbacks of interpolating versus removing bad leads in the context of ECGI, providing valuable insights into ECGI performance.

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.

ECG-based estimation of respiration-induced autonomic modulation of AV nodal conduction during atrial fibrillation

Introduction: Information about autonomic nervous system (ANS) activity may offer insights about atrial fibrillation (AF) progression and support personalized AF treatment but is not easily accessible from the ECG. In this study, we propose a new approach for ECG-based assessment of respiratory modulation in atrioventricular (AV) nodal refractory period and conduction delay. Methods: A 1-dimensional convolutional neural network (1D-CNN) was trained to estimate respiratory modulation of AV nodal conduction properties from 1-minute segments of RR series, respiration signals, and atrial fibrillatory rates (AFR) using synthetic data that replicates clinical ECG-derived data. The synthetic data were generated using a network model of the AV node and 4 million unique model parameter sets. The 1D-CNN was then used to analyze respiratory modulation in clinical deep breathing test data of 28 patients in AF, where an ECG-derived respiration signal was extracted using a novel approach based on periodic component analysis. Results: We demonstrated using synthetic data that the 1D-CNN can estimate the respiratory modulation from RR series alone with a Pearson sample correlation of r = 0.805 and that the addition of either respiration signal (r = 0.830), AFR (r = 0.837), or both (r = 0.855) improves the estimation. Discussion: Initial results from analysis of ECG data suggest that our proposed estimate of respiration-induced autonomic modulation, a resp, is reproducible and sufficiently sensitive to monitor changes and detect individual differences. However, further studies are needed to verify the reproducibility, sensitivity, and clinical significance of a resp.

Synthetic Force-Field Database for Training Machine Learning Models to Predict Mobility-Preserving Coarse-Grained Molecular-Simulation Potentials

Balancing accuracy and efficiency is a common problem in molecular simulation. This tradeoff is evident in coarse-grained molecular dynamics simulation, which prioritizes efficiency, and all-atom molecular simulation, which prioritizes accuracy. Despite continuous efforts, creating a coarse-grained model that accurately captures both the system's structure and dynamics remains elusive. In this article, we present a data-driven approach for constructing coarse-grained models that aim to describe both the structure and dynamics of the system equally well. While the development of machine learning models is well-received in the scientific community, the significance of dataset creation for these models is often overlooked. However, data-driven approaches cannot progress without a robust dataset. To address this, we construct a database of synthetic coarse-grained potentials generated from unphysical all-atom models. A neural network is trained with the generated database to predict the coarse-grained potentials of real liquids. We evaluate their quality by calculating the combined loss of structural and dynamical accuracy upon coarse-graining. When we compare our machine learning-based coarse-grained potential with the one from iterative Boltzmann inversion, the machine learning prediction turns out better for all eight hydrocarbon liquids we studied. As all-atom surfaces turn more nonspherical, both ways of coarse-graining degrade. Still, the neural network outperforms iterative Boltzmann inversion in constructing good quality coarse-grained models for such cases. The synthetic database and the developed machine learning models are freely available to the community, and we believe that our approach will generate interest in efficiently deriving accurate coarse-grained models for liquids.

Automatic and interpretable prediction of the site of origin in outflow tract ventricular arrhythmias: machine learning integrating electrocardiograms and clinical data

The treatment of outflow tract ventricular arrhythmias (OTVA) through radiofrequency ablation requires the precise identification of the site of origin (SOO). Pinpointing the SOO enhances the likelihood of a successful procedure, reducing intervention times and recurrence rates. Current clinical methods to identify the SOO are based on qualitative analysis of pre-operative electrocardiograms (ECG), heavily relying on physician's expertise. Although computational models and machine learning (ML) approaches have been proposed to assist OTVA procedures, they either consume substantial time, lack interpretability or do not use clinical information. Here, we propose an alternative strategy for automatically predicting the ventricular origin of OTVA patients using ML. Our objective was to classify ventricular (left/right) origin in the outflow tracts (LVOT and RVOT, respectively), integrating ECG and clinical data from each patient. Extending beyond differentiating ventricle origin, we explored specific SOO characterization. Utilizing four databases, we also trained supervised learning models on the QRS complexes of the ECGs, clinical data, and their combinations. The best model achieved an accuracy of 89%, highlighting the significance of precordial leads V1-V4, especially in the R/S transition and initiation of the QRS complex in V2. Unsupervised analysis revealed that some origins tended to group closer than others, e.g., right coronary cusp (RCC) with a less sparse group than the aortic cusp origins, suggesting identifiable patterns for specific SOOs.

Non-invasive localization of the ventricular excitation origin without patient-specific geometries using deep learning

Cardiovascular diseases account for 17 million deaths per year worldwide. Of these, 25% are categorized as sudden cardiac death, which can be related to ventricular tachycardia (VT). This type of arrhythmia can be caused by focal activation sources outside the sinus node. Catheter ablation of these foci is a curative treatment in order to inactivate the abnormal triggering activity. However, the localization procedure is usually time-consuming and requires an invasive procedure in the catheter lab. To facilitate and expedite the treatment, we present two novel localization support techniques based on convolutional neural networks (CNNs) that address these clinical needs. In contrast to existing methods, our approaches were designed to be independent of the patient-specific geometry and directly applicable to surface ECG signals, while also delivering a binary transmural position. Moreover, one of the method's outputs can be interpreted as several ranked solutions. The CNNs were trained on a dataset containing only simulated data and evaluated both on simulated test data and clinical data. On a novel large and open simulated dataset, the median test error was below 3 mm. The median localization error on the unseen clinical data ranged from 32 mm to 41 mm without optimizing the pre-processing and CNN to the clinical data. Interpreting the output of one of the approaches as ranked solutions, the best median error of the top-3 solutions decreased to 20 mm on the clinical data. The transmural position was correctly detected in up to 82% of all clinical cases. These results demonstrate a proof of principle to utilize CNNs to localize the activation source without the intrinsic need for patient-specific geometrical information. Furthermore, providing multiple solutions can assist physicians in identifying the true activation source amongst more than one possible location. With further optimization to clinical data, these methods have high potential to accelerate clinical interventions, replace certain steps within these procedures and consequently reduce procedural risk and improve VT patient outcomes.