{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,19]],"date-time":"2026-06-19T17:13:02Z","timestamp":1781889182256,"version":"3.54.5"},"reference-count":31,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,4,16]],"date-time":"2025-04-16T00:00:00Z","timestamp":1744761600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Digit. Health"],"abstract":"<jats:sec><jats:title>Background<\/jats:title><jats:p>12-lead electrocardiograms (ECGs) are a cornerstone for diagnosing and monitoring cardiovascular diseases (CVDs). They play a key role in detecting abnormalities such as arrhythmias and myocardial infarction, enabling early intervention and risk stratification. However, traditional analysis relies heavily on manual interpretation, which is time-consuming and expertise-dependent. Moreover, existing machine learning models often lack personalization, as they fail to integrate subject-specific anatomical and demographic information. Advances in deep generative models offer an opportunity to overcome these challenges by synthesizing personalized ECGs and extracting clinically relevant features for improved risk assessment.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>We propose a conditional Variational Autoencoder (cVAE) framework to generate realistic, subject-specific 12-lead ECGs by incorporating demographic metadata, anatomical heart features, and ECG electrodes\u2019 positions as conditioning factors. This allows for physiologically consistent and personalized ECG synthesis. Furthermore, we introduce a revised Cox proportional-hazards regression model that utilizes the latent embeddings learned by the cVAE to predict future CVD risk. This approach not only enhances the interpretability of ECG-derived risk factors but also demonstrates the potential of deep generative models in personalized cardiac assessment.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Our model is trained and validated on the UK Biobank dataset and <jats:italic>in silico<\/jats:italic> simulation data. By incorporating heart position and electrodes\u2019 positions, the generated ECGs demonstrate strong consistency with <jats:italic>in silico<\/jats:italic> simulations, providing insights into the relationship between cardiac anatomy and ECG morphology. Furthermore, our CVD risk prediction model achieves a C-index of 0.65, indicating that ECG signals, together with demographic and anatomical information, contain valuable prognostic information for stratifying subjects based on future cardiovascular risk.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusion<\/jats:title><jats:p>This work marks a significant advancement in ECG analysis by providing a conditional VAE framework that not only improves ECG generation but also enriches our understanding of the relationship between ECG patterns and subject-specific information. Importantly, our approach enables clinically significant information to be extracted from 12-lead ECGs, providing valuable insights for predicting future CVD risks.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fdgth.2025.1558589","type":"journal-article","created":{"date-parts":[[2025,4,16]],"date-time":"2025-04-16T05:28:54Z","timestamp":1744781334000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["Deep conditional generative model for personalization of 12-lead electrocardiograms and cardiovascular risk prediction"],"prefix":"10.3389","volume":"7","author":[{"given":"Yuling","family":"Sang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Abhirup","family":"Banerjee","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Marcel","family":"Beetz","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Vicente","family":"Grau","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,4,16]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"229","DOI":"10.1113\/jphysiol.1887.sp000257","article-title":"A demonstration on man of electromotive changes accompanying the heart\u2019s beat","volume":"8","author":"Waller","year":"1887","journal-title":"J Physiol (Lond)"},{"key":"B2","doi-asserted-by":"publisher","first-page":"1891","DOI":"10.1177\/0091270011430505","article-title":"Computerized extraction of electrocardiograms from continuous 12-lead holter recordings reduces measurement variability in a thorough QT study","volume":"52","author":"George","year":"2012","journal-title":"J Clin Pharmacol"},{"key":"B3","doi-asserted-by":"publisher","first-page":"886723","DOI":"10.3389\/fphys.2022.886723","article-title":"Multi-domain variational autoencoders for combined modeling of MRI-based biventricular anatomy and ECG-based cardiac electrophysiology","volume":"13","author":"Beetz","year":"2022","journal-title":"Front Physiol"},{"key":"B4","doi-asserted-by":"crossref","DOI":"10.1109\/ISBI52829.2022.9761590","article-title":"Combined generation of electrocardiogram and cardiac anatomy models using multi-modal variational autoencoders","author":"Beetz","year":""},{"key":"B5","doi-asserted-by":"publisher","first-page":"883","DOI":"10.1007\/s10916-010-9551-7","article-title":"Automatic classification of heartbeats using wavelet neural network","volume":"36","author":"Benali","year":"2012","journal-title":"J Med Syst"},{"key":"B6","doi-asserted-by":"publisher","first-page":"512","DOI":"10.1109\/TITB.2008.2003323","article-title":"Heartbeat time series classification with support vector machines","volume":"13","author":"Kampouraki","year":"2008","journal-title":"IEEE Trans Inf Technol Biomed"},{"key":"B7","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-031-23443-9_34","article-title":"Deep computational model for the inference of ventricular activation properties","author":"Li","year":""},{"key":"B8","doi-asserted-by":"publisher","first-page":"2466","DOI":"10.1109\/TMI.2024.3367409","article-title":"Toward enabling cardiac digital twins of myocardial infarction using deep computational models for inverse inference","volume":"43","author":"Li","year":"2024","journal-title":"IEEE Trans Med Imaging"},{"key":"B9","doi-asserted-by":"crossref","DOI":"10.1109\/WCSP.2010.5633782","article-title":"An approach for ECG classification based on wavelet feature extraction and decision tree","author":"Zhang","year":""},{"key":"B10","doi-asserted-by":"crossref","DOI":"10.1109\/TBDATA.2025.3536922","article-title":"Large language model-informed ECG dual attention network for heart failure risk prediction","author":"Chen","year":""},{"key":"B11","doi-asserted-by":"crossref","DOI":"10.1109\/ICHI.2016.83","article-title":"Using the caremap with health incidents statistics for generating the realistic synthetic electronic healthcare record","author":"McLachlan","year":""},{"key":"B12","article-title":"Real-valued (medical) time series generation with recurrent conditional gans","author":"Esteban","year":""},{"key":"B13","article-title":"Electrocardiogram generation and feature extraction using a variational autoencoder","author":"Kuznetsov","year":""},{"key":"B14","doi-asserted-by":"publisher","first-page":"6734","DOI":"10.1038\/s41598-019-42516-z","article-title":"Electrocardiogram generation with a bidirectional LSTM-CNN generative adversarial network","volume":"9","author":"Zhu","year":"2019","journal-title":"Sci Rep"},{"key":"B15","doi-asserted-by":"publisher","first-page":"458916","DOI":"10.3389\/fphys.2019.01103","article-title":"MRI-based computational torso\/biventricular multiscale models to investigate the impact of anatomical variability on the ECG QRS complex","volume":"10","author":"Minchol\u00e9","year":"2019","journal-title":"Front Physiol"},{"key":"B16","doi-asserted-by":"publisher","first-page":"811029","DOI":"10.3389\/fphys.2021.811029","article-title":"Deriving the bidomain model of cardiac electrophysiology from a cell-based model; properties and comparisons","volume":"12","author":"J\u00e6ger","year":"2022","journal-title":"Front Physiol"},{"key":"B17","doi-asserted-by":"crossref","DOI":"10.1109\/ISBI52829.2022.9761431","article-title":"Generation of 12-lead electrocardiogram with subject-specific, image-derived characteristics using a conditional variational autoencoder","author":"Sang","year":""},{"key":"B18","doi-asserted-by":"crossref","DOI":"10.1109\/EMBC48229.2022.9871643","article-title":"Automated torso contour extraction from clinical cardiac MR slices for 3D torso reconstruction","author":"Smith","year":""},{"key":"B19","article-title":"Anatomical basis of sex differences in human post-myocardial infarction ECG phenotypes identified by novel automated torso-cardiac 3D reconstruction","author":"Smith","year":""},{"key":"B20","doi-asserted-by":"publisher","first-page":"e0213653","DOI":"10.1371\/journal.pone.0213653","article-title":"Cardiovascular disease risk prediction using automated machine learning: a prospective study of 423,604 UK Biobank participants","volume":"14","author":"Alaa","year":"2019","journal-title":"PLoS ONE"},{"key":"B21","doi-asserted-by":"publisher","first-page":"h7013","DOI":"10.1136\/bmj.h7013","article-title":"Atrial fibrillation as risk factor for cardiovascular disease and death in women compared with men: systematic review and meta-analysis of cohort studies","volume":"352","author":"Emdin","year":"2016","journal-title":"BMJ"},{"key":"B22","doi-asserted-by":"publisher","first-page":"ehad655\u20132380","DOI":"10.1093\/eurheartj\/ehad655.2380","article-title":"Population prevalence of ECG abnormalities and risk of incident CVD outcomes: 5-year follow-up of 25,000 Chinese adults","volume":"44","author":"Turnbull","year":"2023","journal-title":"Eur Heart J"},{"key":"B23","doi-asserted-by":"publisher","first-page":"1110424","DOI":"10.3389\/fcvm.2023.1110424","article-title":"Association between cardiovascular risk factors and atrial fibrillation","volume":"10","author":"Wu","year":"2023","journal-title":"Front Cardiovasc Med"},{"key":"B24","doi-asserted-by":"publisher","first-page":"e1001779","DOI":"10.1371\/journal.pmed.1001779","article-title":"UK Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age","volume":"12","author":"Sudlow","year":"2015","journal-title":"PLoS Med"},{"key":"B25","doi-asserted-by":"crossref","DOI":"10.22489\/CinC.2020.107","article-title":"A wide and deep transformer neural network for 12-lead ECG classification","author":"Natarajan","year":""},{"key":"B26","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s12968-016-0227-4","article-title":"UK Biobank\u2019s cardiovascular magnetic resonance protocol","volume":"18","author":"Petersen","year":"2015","journal-title":"J Cardiovasc Magn Reson"},{"key":"B27","doi-asserted-by":"publisher","first-page":"162","DOI":"10.1093\/ije\/dyz212","article-title":"Impact of follow-up time and analytical approaches to account for reverse causality on the association between physical activity and health outcomes in UK Biobank","volume":"49","author":"Strain","year":"2020","journal-title":"Int J Epidemiol"},{"key":"B28","doi-asserted-by":"publisher","first-page":"20200257","DOI":"10.1098\/rsta.2020.0257","article-title":"A completely automated pipeline for 3D reconstruction of human heart from 2D cine magnetic resonance slices","volume":"379","author":"Banerjee","year":"2021","journal-title":"Philos Trans R Soc A"},{"key":"B29","article-title":"Adam: a method for stochastic optimization","author":"Kingma","year":""},{"key":"B30","doi-asserted-by":"publisher","first-page":"143","DOI":"10.1002\/sim.4780030207","article-title":"Regression modelling strategies for improved prognostic prediction","volume":"3","author":"Harrell Jr","year":"1984","journal-title":"Stat Med"},{"key":"B31","doi-asserted-by":"publisher","first-page":"2109","DOI":"10.1002\/sim.1802","article-title":"Overall c as a measure of discrimination in survival analysis: model specific population value and confidence interval estimation","volume":"23","author":"Pencina","year":"2004","journal-title":"Stat Med"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2025.1558589\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,4,16]],"date-time":"2025-04-16T05:28:55Z","timestamp":1744781335000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2025.1558589\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,4,16]]},"references-count":31,"alternative-id":["10.3389\/fdgth.2025.1558589"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2025.1558589","relation":{},"ISSN":["2673-253X"],"issn-type":[{"value":"2673-253X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,4,16]]},"article-number":"1558589"}}