Pasar al contenido principal

Biomedical signal processing, genetics and data science tools for cardiovascular risk prediction

RYC2021-031413-I

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal para Desarrollar, Atraer y Retener Talento
Subprograma Subprograma Estatal de Incorporación
Convocatoria Ramón y Cajal
Año convocatoria 2021
Plan Estatal 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSIDAD DE ZARAGOZA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 24
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

State of the Art of Artificial Intelligence in Clinical Electrophysiology in 2025: A Scientific Statement of the European Heart Rhythm Association (EHRA) of the ESC, the Heart Rhythm Society (HRS), and the ESC Working Group on E-Cardiology

  • Svennberg, Emma
  • Han, Janet K.
  • Caiani, Enrico G.
  • Engelhardt, Sandy
  • Ernst, Sabine
  • Friedman, Paul
  • Garcia, Rodrigue
  • Ghanbari, Hamid
  • Hindricks, Gerhard
  • Man, Sharon H.
  • Narayan, Sanjiv M.
  • Ng, G. Andre
  • Noseworthy, Peter A.
  • Tjong, Fleur V. Y.
  • Millet Roig, José|||0000-0002-8879-003X
[EN] Aims Artificial intelligence (AI) has the potential to transform cardiac electrophysiology (EP), particularly in arrhythmia detection, procedural optimization, and patient outcome prediction. However, a standardized approach to reporting and understanding AI-related research in EP is lacking. This scientific statement aims to develop and apply a checklist for AI-related research reporting in EP to enhance transparency, reproducibility, and understandability in the field. Methods and results An AI checklist specific to EP was developed with expert input from the writing group and voted on using a modified Delphi process, leading to the development of a 29-item checklist. The checklist was subsequently applied to assess reporting practices to identify areas where improvements could be made and provide an overview of the state of the art in AI-related EP research in three domains from May 2021 until May 2024: atrial fibrillation (AF) management, sudden cardiac death (SCD), and EP lab applications. The EHRA AI checklist was applied to 31 studies in AF management, 18 studies in SCD, and 6 studies in EP lab applications. Results differed between the different domains, but in no domain reporting of a specific item exceeded 55% of included papers. Key areas such as trial registration, participant details, data handling, and training performance were underreported (<20%). The checklist application highlighted areas where reporting practices could be improved to promote clearer, more comprehensive AI research in EP. Conclusion The EHRA AI checklist provides a structured framework for reporting AI research in EP. Its use can improve understanding but also enhance the reproducibility and transparency of AI studies, fostering more robust and reliable integration of AI into clinical EP practice.; [GRAPHICS]; ., E.S. is supported by the Stockholm County Council (Clinical researcher appointment), the Swedish Research Council (DNR 2022-01466), the Swedish Heart and Lung foundation, CIMED, S.E. is supported by research grants from German Center for Cardiovascular Research (DZHK), the Carl-Zeiss Foundation in the project Multi-dimension AI. J Millet (e-cardiology) reports grant support PID2022-142514OB-I00 from Spanish Research Agency. Professor GA Ng reports grant support from the British Heart Foundation [Research Excellence Award (RE/24/13003) and Healthcare Innovation Fund (HI/C2/24/360038)], the National Institute for Health and Care Research Invention for Innovation Award (NIHR204553). F.V.Y.T. is supported by personal research grants from the Dutch Research Council (NWO), Amsterdam Cardiovascular Sciences, and Health Holland. J.R. acknowledges funding from the European Union-NextGenerationEU, fellowship RYC2021-031413-I from MCIN/AEI/10.13039/501100011033, and from the European Union "NextGenerationEU/PRTR" and from grants PID2021-128972OA-I00, CNS2023-143599 and PID2023-148975OB-I00, funded by MCIN/AEI/10.13039/501100011033. N.T. (HRS) is supported by NIH grants R01HL166759 and R01HL174440 and a grant from the Leducq Foundation.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Predicting left ventricular hypertrophy from the 12-lead electrocardiogram in the UK Biobank imaging study using machine learning

  • Naderi, Hafiz
  • Ramírez, Julia
  • van Duijvenboden, Stefan
  • Pujadas, Esmeralda Ruiz
  • Aung, Nay
  • Wang, Lin
  • Anwar Ahmed Chahal, Choudhary
  • Lekadir, Karim
  • Petersen, Steffen E
  • Munroe, Patricia B
Aims: Left ventricular hypertrophy (LVH) is an established, independent predictor of cardiovascular disease. Indices derived from the electrocardiogram (ECG) have been used to infer the presence of LVH with limited sensitivity. This study aimed to classify LVH defined by cardiovascular magnetic resonance (CMR) imaging using the 12-lead ECG for cost-effective patient stratification. Methods and results: We extracted ECG biomarkers with a known physiological association with LVH from the 12-lead ECG of 37 534 participants in the UK Biobank imaging study. Classification models integrating ECG biomarkers and clinical variables were built using logistic regression, support vector machine (SVM) and random forest (RF). The dataset was split into 80% training and 20% test sets for performance evaluation. Ten-fold cross validation was applied with further validation testing performed by separating data based on UK Biobank imaging centres. QRS amplitude and blood pressure (P < 0.001) were the features most strongly associated with LVH. Classification with logistic regression had an accuracy of 81% [sensitivity 70%, specificity 81%, Area under the receiver operator curve (AUC) 0.86], SVM 81% accuracy (sensitivity 72%, specificity 81%, AUC 0.85) and RF 72% accuracy (sensitivity 74%, specificity 72%, AUC 0.83). ECG biomarkers enhanced model performance of all classifiers, compared to using clinical variables alone. Validation testing by UK Biobank imaging centres demonstrated robustness of our models. Conclusion: A combination of ECG biomarkers and clinical variables were able to predict LVH defined by CMR. Our findings provide support for the ECG as an inexpensive screening tool to risk stratify patients with LVH as a prelude to advanced imaging.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Time-warping analysis of the T-wave peak-to-end Interval to quantify ventricular repolarization dispersion during ischemia

  • Gómez, Neurys
  • Ramírez, Julia
  • Martínez, Juan Pablo
  • Laguna, Pablo
Variations in the dispersion of ventricular repolarization can be quantified by T-wave time-warping based index, dw . However, the early phase of the T-wave can be affected by ST-segment changes during ischemia. We hypothesized that restricting dw to the T-wave peak-to-end ( Tpe ) would circumvent this limitation while still quantifying variations in repolarization dispersion. A total of 101 ECG recordings from patients undergoing coronary occlusion, together with their control recordings, were analyzed. A series of dw values was calculated by quantifying the Tpe morphological variations between the T-waves at different occlusion stages and a baseline T-wave. We introduced a normalized version of dw , Rd , reflecting variations of dw during occlusion relative to control recordings ( Rd=1 corresponds to the same level of variation). The dw series followed a gradually increasing trend with occlusion time, reaching median [range] Rd values of 9.44 [1.01, 80.74] at the occlusion end. Rd at occlusion end was significantly higher than threshold values of 1, 2, 5, and 10 in 94.1%, 85.11%, 64.4% and 48.5% of patients, respectively. The spatial lead-wise analysis of dw showed distinct distributions depending on the occluded artery, suggesting a relation with the ischemia location. The relative variation R with ischemia of index dw (9.4) is greater than that of the T-wave amplitude (7.7), Tpe interval (2.7) and T-wave width (3.0). In conclusion, dw tracks ischemic-induced variations in repolarization dispersion in a more robust manner than classical indexes, avoiding the impact of ST segment elevation/depression or early T-wave distortions, thus warranting further clinical studies.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Contribución de congreso (ConferenceOutput)

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

  • Gómez Fonseca, Neurys
  • Ramírez, Julia
  • Martin-Yebra, Alba
  • M. Demidova, Marina
  • Platonov, Pyotr
  • Martínez, Juan Pablo
  • Laguna, Pablo
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 under- going 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) d w 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.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Long-term association of ultra-short heart rate variability with cardiovascular events

  • Orini, Michele
  • van Duijvenboden, Stefan
  • Young, William J.
  • Ramírez, Julia
  • Jones, Aled R.
  • Hughes, Alun D.
  • Tinker, Andrew
  • Munroe, Patricia B.
  • Lambiase, Pier D.
Heart rate variability (HRV) is a cardiac autonomic marker with predictive value in cardiac patients. Ultra-short HRV (usHRV) can be measured at scale using standard and wearable ECGs, but its association with cardiovascular events in the general population is undetermined. We aimed to validate usHRV measured using ≤ 15-s ECGs (using RMSSD, SDSD and PHF indices) and investigate its association with atrial fibrillation, major adverse cardiac events, stroke and mortality in individuals without cardiovascular disease. In the National Survey for Health and Development (n = 1337 participants), agreement between 15-s and 6-min HRV, assessed with correlation analysis and Bland–Altman plots, was very good for RMSSD and SDSD and good for PHF. In the UK Biobank (n = 51,628 participants, 64% male, median age 58), after a median follow-up of 11.5 (11.4–11.7) years, incidence of outcomes ranged between 1.7% and 4.3%. Non-linear Cox regression analysis showed that reduced usHRV from 15-, 10- and 5-s ECGs was associated with all outcomes. Individuals with low usHRV (< 20th percentile) had hazard ratios for outcomes between 1.16 and 1.29, p < 0.05, with respect to the reference group. In conclusion, usHRV from ≤ 15-s ECGs correlates with standard short-term HRV and predicts increased risk of cardiovascular events in a large population-representative cohort.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Integration of genetic fine-mapping and multi-omics data reveals candidate effector genes for hypertension

  • van Duijvenboden, Stefan
  • Ramírez, Julia
  • Young, William J.
  • Olczak, Kaya J.
  • Ahmed, Farah
  • Alhammadi, Mohammed J.A.Y.
  • Bell, Christopher G.
  • Morris, Andrew P.
  • Munroe, Patricia B.
Genome-wide association studies of blood pressure (BP) have identified >1,000 loci, but the effector genes and biological pathways at these loci are mostly unknown. Using published association summary statistics, we conducted annotation-informed fine-mapping incorporating tissue-specific chromatin segmentation and colocalization to identify causal variants and candidate effector genes for systolic BP, diastolic BP, and pulse pressure. We observed 532 distinct signals associated with ≥2 BP traits and 84 with all three. For >20% of signals, a single variant accounted for >75% posterior probability, 65 were missense variants in known (SLC39A8, ADRB2, and DBH) and previously unreported BP candidate genes (NRIP1 and MMP14). In disease-relevant tissues, we colocalized >80 and >400 distinct signals for each BP trait with cis-eQTLs and regulatory regions from promoter capture Hi-C, respectively. Integrating mouse, human disorder, gene expression and tissue abundance data, and literature review, we provide consolidated evidence for 436 BP candidate genes for future functional validation and discover several potential drug targets.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

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

  • Madrid, Josseline
  • Munroe, Patricia
  • Van Duijvenboden, Stefan
  • Ramírez, Julia
  • Mincholé, Ana
Aims: Coronary Artery Disease (CAD) is one of the main causes of life-threatening ventricular arrhythmias (LTVAs) leading to sudden cardiac death. CAD slows ventricular conduction across individuals, manifesting as heterogeneous 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. Methods: An average heartbeat was derived from 10-second electrocardiograms (ECGs, lead I) from 1,458 individuals diagnosed with CAD in the UK Biobank study. An unsupervised clustering algorithm based on 3-nearest neighbours was used to classify each individual, then we evaluated the association of each cluster with LTVA risk. Results: There were a total of 65 LTVA events in the population. The algorithm distinguished 3 distinct clusters of QRS-related morphological features, which significantly differed in terms of LTVA events rate. Cluster 2, characterized by the lowest QRS amplitudes and widest QRS complexes, was strongly associated with LTVA risk. Conclusions: Our analysis has identified CAD individuals at risk of LTVA using the QRS morphology. The identified cluster could be used to tailor care and provide refined risk assessment in CAD individuals to apply specific prevention measures.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

A Multi-layer CNN Using the ECG, Age and Sex Predicts Ventricular Arrhythmias in the General Population

  • Ramírez, Julia
  • Miguel, Antonio
  • van Duijvenboden, Stefan
  • Orini, Michele
  • Young, William J.
  • Tinker, Andrew
  • Lambiase, Pier D.
  • Munroe, Patricia B.
  • Martínez, Juan Pablo
Life-threatening ventricular arrhythmias (LTVA) prediction in individuals without cardiovascular disease remains a major challenge. We tested the performance of a multilayer convolutional neural network (CNN) using ECG signals, age and sex. We split 86,603 individuals from the UK Biobank study into a training (90%) and a test (10%) set. In the training set, we trained a multilayer CNN using 15-second ECGs at rest from lead I, age and sex as inputs. The output was the probability of LTVA within a 12-year follow-up. The CNN model consisted of a four-layer CNN (128, 128, 256 and 256 channels, kernel sizes of 3) and a single attention layer. Age and sex were included as external inputs to the final layer. In the test set (0.9% LTVA events), the CNN's prediction led to a median AUC of 0.601, and a specificity of 0.287 for a sensitivity of 0.750. We set a threshold at the CNN's prediction value maximising the sum of specificity and sensitivity in the training set. Survival analyses showed a hazard ratio (HR) of 1.396(P=0.021) for individuals with a CNN's prediction value > threshold, versus those with a CNN's prediction value < threshold. A multilayer CNN model using 10-second ECG data from lead I, together with information on age and sex, can stratify individuals at risk of LTVA. Our findings support the potential utility of wearables for accessible screening in the general population.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

T-wave Peak-to-End Changes Quantified by Time-Warping Predicts Ventricular Fibrillation in a Porcine Myocardial Infarction Model

  • Gómez, Neurys
  • Ramírez, Julia
  • Martín-Yebra, Alba
  • Demidova, Marina M.
  • Platonov, Pyotr
  • Martínez, Juan Pablo
  • Laguna, Pablo
Background: The Tpe-interval time-warping-based morphology variation index, derived from a spatially transformed PCA lead, dPCAw,Tpe, has previously demonstrated utility in tracking repolarization changes induced by a 5-minute ischemia model in humans. The value of dPCAw,Tpe as a predictor of ventricular fibrillation (VF) episodes is assessed in a 40-minutes porcine infarction model. Preocclusion and occlusion ECG recordings from 10 pigs who suffered a late VF episode and 16 who did not were analyzed. The dPCAw,Tpe series was measured by comparing Tpe morphologies at different stages of the occlusion relative to a baseline. During baseline, dPCAw,Tpe remained stationary with an intra-recording median [IQR] value of 1.60 [1.33] ms. During occlusion, dPCAw,Tpe followed a well-marked gradual increasing trend, reaching a median of 14.58 [17.72] ms. dPCAw,Tpe averages were significantly higher in the VF group than in the Non-VF group at five time intervals taken every 5 minutes after occlusion onset with median values of 12.5, 18.8, 26.8, 24.0, and 31.0, vs 6.3, 7.6, 8.0, 7.8, and 7.8, and also at three consecutive 5-minute intervals prior to VF 18.6, 25.0 and 28.8 vs 8.5, 7.2 and 6.0 ms, respectively. The TPCApe interval was also significantly higher in the VF group at all analyzed time periods, but with a lower significance level. Pigs with maximum dPCAw,Tpe ≥ 20.0 ms and TPCApe ≥ 85.4 ms in the early 5-10 minutes interval after occlusion, had significantly higher risk for VF occurring with 90.0%/75.0% and 80.0%/69.0% sensitivity/ specificity, respectively. In conclusion, dPCAw,Tpe is a stronger VF predictor than TPCApe during ischemia in a porcine model.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Fine mapping of candidate effector genes for heart rate

  • Ramírez, Julia
  • van Duijvenboden, Stefan
  • Young, William J.
  • Chen, Yutang
  • Usman, Tania
  • Orini, Michele
  • Lambiase, Pier D.
  • Tinker, Andrew
  • Bell, Christopher G.
  • Morris, Andrew P.
  • Munroe, Patricia B.
An elevated resting heart rate (RHR) is associated with increased cardiovascular mortality. Genome-wide association studies (GWAS) have identified > 350 loci. Uniquely, in this study we applied genetic fine-mapping leveraging tissue specific chromatin segmentation and colocalization analyses to identify causal variants and candidate effector genes for RHR. We used RHR GWAS summary statistics from 388,237 individuals of European ancestry from UK Biobank and performed fine mapping using publicly available genomic annotation datasets. High-confidence causal variants (accounting for > 75% posterior probability) were identified, and we collated candidate effector genes using a multi-omics approach that combined evidence from colocalisation with molecular quantitative trait loci (QTLs), and long-range chromatin interaction analyses. Finally, we performed druggability analyses to investigate drug repurposing opportunities. The fine mapping pipeline indicated 442 distinct RHR signals. For 90 signals, a single variant was identified as a high-confidence causal variant, of which 22 were annotated as missense. In trait-relevant tissues, 39 signals colocalised with cis-expression QTLs (eQTLs), 3 with cis-protein QTLs (pQTLs), and 75 had promoter interactions via Hi-C. In total, 262 candidate genes were highlighted (79% had promoter interactions, 15% had a colocalised eQTL, 8% had a missense variant and 1% had a colocalised pQTL), and, for the first time, enrichment in nervous system pathways. Druggability analyses highlighted ACHE, CALCRL, MYT1 and TDP1 as potential targets. Our genetic fine-mapping pipeline prioritised 262 candidate genes for RHR that warrant further investigation in functional studies, and we provide potential therapeutic targets to reduce RHR and cardiovascular mortality.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Genetic basis of right and left ventricular heart shape

  • Burns, Richard
  • Young, William J.
  • Aung, Nay
  • Lopes, Luis R.
  • Elliott, Perry M.
  • Syrris, Petros
  • Barriales-Villa, Roberto
  • Sohrabi, Catrin
  • Petersen, Steffen E.
  • Ramírez, Julia
  • Young, Alistair
  • Munroe, Patricia B.
Heart shape captures variation in cardiac structure beyond traditional phenotypes of mass and volume. Although observational studies have demonstrated associations with cardiometabolic risk factors and diseases, its genetic basis is less understood. We utilised cardiovascular magnetic resonance images from 45,683 UK Biobank participants to construct a heart shape atlas from bi-ventricular end-diastolic surface mesh models through principal component (PC) analysis. Genome-wide association studies were performed on the first 11 PCs that captured 83.6% of shape variance. We identified 43 significant loci, 14 were previously unreported for cardiac traits. Genetically predicted PCs were associated with cardiometabolic diseases. In particular two PCs (2 and 3) linked with more spherical ventricles being associated with increased risk of atrial fibrillation. Our study explores the genetic basis of multidimensional bi-ventricular heart shape using PCA, reporting new loci and biology, as well as polygenic risk scores for exploring genetic relationships of heart shape with cardiometabolic diseases.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Leucocyte telomere length and conduction system ageing

  • van Duijvenboden, Stefan
  • Nelson, Christopher P
  • Raisi-Estabragh, Zahra
  • Ramirez, Julia
  • Orini, Michele
  • Wang, Qingning
  • Aung, Nay
  • Codd, Veryan
  • Stoma, Svetlana
  • Allara, Elias
  • Wood, Angela M
  • Di Angelantonio, Emanuele
  • Danesh, John
  • Harvey, Nicholas C
  • Petersen, Steffen E
  • Munroe, Patricia B
  • Samani, Nilesh J
Deterioration of the cardiac conduction system is an important manifestation of cardiac ageing. Cellular ageing is accompanied by telomere shortening and telomere length (TL) is often regarded as a marker of biological ageing, potentially adding information regarding conduction disease over and above chronological age. We therefore sought to evaluate the association between leucocyte telomere length (LTL) on two related, but distinct aspects of the cardiac conduction system: ECG measures of conduction (PR interval and QRS duration) and incident pacemaker implantation in a large population-based cohort.MethodsIn the UK Biobank, we measured PR interval and QRS duration from signal-averaged ECG waveforms in 59 868 and 62 266 participants, respectively. Incident pacemaker implantation was ascertained using hospital episode data from 420 071 participants. Associations with LTL were evaluated in (Cox) multivariable regression analyses adjusted for potential confounders. Putative causal effects of LTL were investigated by mendelian randomisation (MR).ResultsMean PR interval and QRS duration were 144.2 ms (± 20.4) and 92.3 ms (± 7.8), respectively, and there were 7169 (1.7%) incident pacemaker implantations, during a median follow-up period of 13.6 (IQR 1.5) years. LTL was significantly associated with PR interval (0.19 ms (95% CI: 0.03 to 0.35), per 1 SD shorter LTL, p=0.021), but not QRS duration. After adjusting for age, sex and cardiovascular risk factors, shorter LTL remained associated with an increased risk for incident pacemaker implantation (HR per SD decrease in LTL: 1.03 (95% CI: 1.01 to 1.06), p=0.012). MR analysis showed a trend towards an association of shorter LTL with longer PR interval and higher risk of pacemaker implantation but was likely to be underpowered.ConclusionsShorter LTL was significantly, and possibly causally, associated with prolongation of atrioventricular conduction and pacemaker implantation, independent of traditional cardiovascular risk factors. Our findings support further research to explore the role of ageing on cardiac conduction beyond chronological age.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Prediction of atrial and ventricular arrhythmias using multiple cardiovascular risk-factor polygenic risk scores

  • Ramírez, Julia
  • van Duijvenboden, Stefan
  • Orini, Michele
  • Lambiase, Pier D.
  • Tinker, Andrew
  • Young, William J.
  • Munroe, Patricia B.
Background. Atrial fibrillation (AF) prediction improves by combining clinical scores with a polygenic risk score (PRS) for AF (AF-PRS), but there are limited studies of PRS for ventricular arrhythmia (VA) prediction. We assessed the value of including multiple PRS for cardiovascular risk factors (CV-PRS) for incident AF and VA prediction. Methods. We used 158,733 individuals of European ancestry from UK Biobank to build 3 models for AF: CHARGE-AF (AF1), AF1 + AF-PRS (AF2), AF2 + CV-PRS (AF3). Models for VA included sex and age (VA1), VA1 + coronary artery disease (CAD) PRS (CAD-PRS, VA2), and VA2 + CV-PRS (VA3), conducting separate analyses in subjects with and without ischemic heart disease (IHD). Performance was evaluated in individuals of European (N = 158,733), African (N = 7200), South Asian (N = 9241) and East Asian (N = 2076) ancestry from UK Biobank. Results. AF2 had a higher C-index than AF1 (0.762 vs 0.746, P < .001), marginally improving to 0.765 for AF3 (P < .001, including PRS for heart failure, electrocardiogram and cardiac magnetic resonance measures). In South Asians, AF2 C-index was higher than AF1 (P < .001). For VA, the C-index for VA2 was greater than VA1 (0.692 vs 0.681, P < .001) in Europeans, which was also observed in South Asians (P < .001). VA3 improved prediction of VA in individuals with IHD. Conclusion. CV-PRS improved AF prediction compared to CHARGE-AF and AF-PRS. A CAD-PRS improved VA prediction, while CV-PRS contributed in IHD. AF- and CAD-PRS were transferable to individuals of South Asian ancestry. Our results inform of the use of CV-PRS for personalized screening.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Unsupervised clustering of songle-lead ECGs associates with prevalent and incident heart failure in coronary artery disease

  • Madrid, Josseline
  • Young, William
  • van Duijvenboden, Stefan
  • Orini, Michele
  • Munroe, Patricia
  • Ramírez García, Julia*
  • Mincholé, Ana
Clinical consequences of coronary artery disease (CAD) are varied [e.g. atrial fibrillation (AF) and heart failure (HF)], and current risk stratification tools are ineffective. We aimed to identify clusters of individuals with CAD exhibiting unique pat-terns on the electrocardiogram (ECG) in an unsupervised manner and assess their association with cardiovascular risk. Methods and results Twenty-one ECG markers were derived from single-lead median-beat ECGs of 1928 individuals with CAD without a previous diagnosis of AF, HF, or ventricular arrhythmia (VA) from the imaging study in UK Biobank (CAD-IMG-UKB). An un-supervised clustering algorithm was used to group these markers into distinct clusters. We characterized each cluster according to their demographic and ECG characteristics, as well as their prevalent and incident risk of AF, HF, and VA(4-year median follow-up). Validation and association with prevalent diagnoses were performed in an independent cohort of 1644 individuals. The model identified two clusters within the CAD-IMG-UKB cohort. Cluster 1 (n = 359) exhibited pro-longed QRS duration and QT intervals, along with greater morphological variations in QRS and T-waves, compared with Cluster 2 (n = 1569). Cluster 1, relative to Cluster 2, had a significantly higher risk of incident HF [hazard ratio (HR):2.40, 95% confidence interval (CI): 1.51–3.83], confirmed by independent validation (HR: 1.77, CI: 1.31–2.41). It also showed a higher association with prevalent HF (odds ratio: 4.10, CI: 2.02–8.29), independent of clinical risk factors. Conclusion Our approach identified a cluster of individuals with CAD sharing ECG characteristics indicating HF risk, holding significant implications for targeted treatment and prevention enabling accessible large-scale screening.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

RoePINNs: An integration of advanced CFD solvers with Physics-Informed Neural Networks and application in arterial flow modeling

  • Orera, J.
  • Ramírez, J.
  • García-Navarro, P.
  • Murillo, J.
The characterization of forward and inverse problems describing blood flow dynamics plays a decisive role in numerous biomedical applications. These systems can be modeled using one-dimensional (1D) approaches leading to a hyperbolic system of equations with source terms. Their numerical discretization, associated to the spatial variation of mechanical and geometrical properties, requires advanced numerical solvers that ensure both stability and an accurate description of the dynamics of the system. In this work, we present RoePINNs, a hybrid framework for the embedding of advanced Computational Fluid Dynamics (CFD) solvers into Physics-Informed Neural Networks (PINNs), and give examples of application to Burgers’ equation as well as the propagation of nonlinear waves in elastic arteries, both under the presence of geometric-type source terms, for forward and inverse problems. We demonstrate that Augmented Riemann solvers can be incorporated into the PINN framework with straightforward adjustments to the hyperparameters, providing a promising alternative to automatic differentiation (AD), especially in cases where the solution exhibits strong nonlinearities and physical constraints are required. Benefits of the proposed RoePINN compared with the vanilla PINN based in AD are twofold: on the one hand, this hybrid approach employs numerical differentiation by means of support points in the surroundings of the collocation points, hence the robustness, generalization capacity and tunability of the PINNs are, in most cases, largely enhanced. On the other hand, the RoePINN incorporates the numerical solver, hence it is also capable of capturing sharp discontinuities with an order-of-magnitude improvement in accuracy compared with the vanilla version.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Using Genomics to Develop Personalized Cardiovascular Treatments

  • Sanghvi, Mihir M.
  • Young, William J.
  • Naderi, Hafiz
  • Burns, Richard
  • Ramírez, Julia
  • Bell, Christopher G.
  • Munroe, Patricia B.
Advances in genomic technologies have significantly enhanced our understanding of both monogenic and polygenic etiologies of cardiovascular disease. In this review, we explore how the utilization of genomic information is bringing personalized medicine approaches to the forefront of cardiovascular disease management. We describe how genomic data can resolve diagnostic uncertainty, support cascade screening, and inform treatment strategies. We discuss how genome-wide association studies have identified thousands of genetic variants associated with polygenic cardiovascular diseases, and how integrating these insights into polygenic risk scores can enhance personalized risk prediction beyond traditional clinical algorithms. We detail how pharmacogenomics approaches leverage genotype information to guide drug selection and mitigate adverse events. Finally, we present the paradigm-shifting approach of gene therapy, which holds the promise of being a curative intervention for cardiovascular conditions.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

State of the Art of Artificial Intelligence in Clinical Electrophysiology in 2025: A Scientific Statement of the European Heart Rhythm Association (EHRA) of the ESC, the Heart Rhythm Society (HRS), and the ESC Working Group on E-Cardiology

  • Svennberg, Emma
  • Han, Janet K.
  • Caiani, Enrico G.
  • Engelhardt, Sandy
  • Ernst, Sabine
  • Friedman, Paul
  • Garcia, Rodrigue
  • Ghanbari, Hamid
  • Hindricks, Gerhard
  • Man, Sharon H.
  • Millet, José
  • Narayan, Sanjiv M.
  • Ng, G. André
  • Noseworthy, Peter A.
  • Tjong, Fleur V.Y.
  • Ramírez, Julia
  • Singh, Jagmeet P.
  • Trayanova, Natalia
  • Duncker, David
  • Tfelt Hansen, Jacob
  • Barker, Joseph
  • Casado-Arroyo, Ruben
  • Chatterjee, Neal A.
  • Conte, Giulio
  • Diederichsen, Søren Zöga
  • Linz, Dominik
  • Mahtani, Arun Umesh
  • Zorzi, Alessandro
Aims Artificial intelligence (AI) has the potential to transform cardiac electrophysiology (EP), particularly in arrhythmia detection, procedural optimization, and patient outcome prediction. However, a standardized approach to reporting and understanding AI-related research in EP is lacking. This scientific statement aims to develop and apply a checklist for AI-related research reporting in EP to enhance transparency, reproducibility, and understandability in the field. Methods and results An AI checklist specific to EP was developed with expert input from the writing group and voted on using a modified Delphi process, leading to the development of a 29-item checklist. The checklist was subsequently applied to assess reporting practices to identify areas where improvements could be made and provide an overview of the state of the art in AI-related EP research in three domains from May 2021 until May 2024: atrial fibrillation (AF) management, sudden cardiac death (SCD), and EP lab applications. The EHRA AI checklist was applied to 31 studies in AF management, 18 studies in SCD, and 6 studies in EP lab applications. Results differed between the different domains, but in no domain reporting of a specific item exceeded 55% of included papers. Key areas such as trial registration, participant details, data handling, and training performance were underreported (<20%). The checklist application highlighted areas where reporting practices could be improved to promote clearer, more comprehensive AI research in EP. Conclusion The EHRA AI checklist provides a structured framework for reporting AI research in EP. Its use can improve understanding but also enhance the reproducibility and transparency of AI studies, fostering more robust and reliable integration of AI into clinical EP practice.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Phenotypic expression of rare progressive cardiac conduction disease variants in the general population

  • Shah, Ravi A
  • Ramírez, Julia
  • Kirkby, Claire
  • Ives, Charlotte
  • Lowe, Martin
  • Munroe, Patricia B
  • Lambiase, Pier D
  • Young, William J
Aims Familial progressive cardiac conduction disease (PCCD) is a heritable condition leading to conduction defects that may require pacemaker implantation. The penetrance of rare PCCD variants in general populations and relationship with electrocardiogram (ECG) trait polygenic risk scores (PRS) is unknown. We investigated the prevalence and phenotypic expression of rare variants linked with PCCD in a population cohort and to establish whether ECG-trait PRSs improve risk prediction. Methods and results Carriers of known rare pathogenic/likely pathogenic (P/LP) PCCD variants, and variants of uncertain significance (VUS) were identified in 469 511 UK Biobank participants. Primary (any conduction disease) and secondary (high-grade AV block and pacemaker implantation) outcomes were evaluated in lifetime-risk Cox proportional hazard models including rare variant status, sex, and age. Additional models including PR and QRS PRSs were tested. There were 25 P/LP carriers (5 genes) and 3174 VUS carriers (4 genes). Conduction disease was more prevalent in P/LP individuals compared with non-carriers (28% vs. 5.3%, P < 0.001) with a hazard ratio (HR) of 6.60 (95% CI = 3.14–13.8) over 6.5 million person-years of follow-up and C-index 0.602 (0.599–0.605). This was driven by AV block (HR 23.2 [8.7–61.8]) and pacemaker implantation (HR 13.4 [6.01–29.8]). All individuals were aged >50 at diagnosis. Combined with P/LP status, PR-PRS and QRS-PRS improved model performance (C-index 0.618 [0.615–0.622]). Conclusion In a population-based cohort, PCCD P/LP variant carriers were at greater risk of conduction disease. Including PRSs for the PR and QRS improved risk prediction, supporting the combination of rare and common variants in risk assessment.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Relationships between heart shape, function, and disease in 38,858 UK biobank participants

  • Burns, Richard
  • Dal Toso, Laura
  • Mauger, Charlène A.
  • Sojoudi, Alireza
  • Suinesiaputra, Avan
  • Petersen, Steffen E.
  • Ramírez, Julia
  • Munroe, Patricia B.
  • Young, Alistair A.
Background. Cardiac functional metrics such as ejection fraction, strain, and valve excursion are important diagnostic and prognostic measures of cardiac disease. However, they ignore a large amount of systolic shape change information available from modern cardiovascular magnetic resonance (CMR) examinations. We aimed to automatically quantify multidimensional shape and motion scores from CMR, investigate covariates, and test their discrimination of disease in the UK Biobank compared against standard functional metrics. Methods. An automated analysis pipeline was used to obtain quality-controlled three-dimensional left and right ventricular shape models in 38,858 UK Biobank participants, 5149 of whom had one or more diagnoses of cardiovascular or cardiometabolic disease. Principal component analysis was used to obtain a statistical shape atlas and quantify each participant’s left and right ventricular shape at both end-diastole and end-systole simultaneously. Systolic strain was obtained from arc length changes computed from the shape model, and mitral/tricuspid annular plane systolic excursion (MAPSE/TAPSE) was computed from the displacement of the valves. Discrimination for prevalent disease was quantified using linear discriminant analysis area under the receiver operating characteristic curve. Results. The first 25 principal component scores captured >90% of the total shape variance. Significantly stronger discrimination for atrial fibrillation, heart failure, diabetes, ischemic disease, and conduction disorders (p<0.001 for each) was obtained using shape scores compared with volumes, ejection fractions, strains, MAPSE, and TAPSE. Conclusion. Automatically derived shape and motion z-scores capture more discriminative information on disease effects than standard metrics, including volumes, ejection fraction, strain and valve excursions.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Diagnostic and prognostic value of ECG-predicted hypertension-mediated left ventricular hypertrophy using machine learning

  • Naderi, Hafiz
  • Ramírez, Julia
  • Van Duijvenboden, Stefan
  • Ruiz Pujadas, Esmeralda
  • Aung, Nay
  • Wang, Lin
  • Chamling, Bishwas
  • Dörr, Marcus
  • Markus, Marcello R.P.
  • Chahal, Choudhary Anwar A.
  • Lekadir, Karim
  • Petersen, Steffen E.
  • Munroe, Patricia B.
Objective: Four hypertension-mediated left ventricular hypertrophy (LVH) phenotypes have been reported using cardiac magnetic resonance (CMR): normal LV, LV remodelling, eccentric and concentric LVH, with varying prognostic implications. The electrocardiogram (ECG) is routinely used to detect LVH; however, its capacity to differentiate between LVH phenotypes is unknown. This study aimed to classify hypertension-mediated LVH from the ECG using machine learning and test for associations of ECG-predicted phenotypes with incident cardiovascular outcomes. Methods: ECG biomarkers were extracted from the 12-lead ECG of 20 439 hypertensive patients in UK Biobank (UKB). Classification models integrating ECG and clinical variables were built using logistic regression, support vector machine (SVM), and random forest. The models were trained in 80% of the participants, and the remaining 20% formed the test set. External validation was sought in 877 hypertensive participants from the Study of Health in Pomerania (SHIP). In the UKB test set, we tested for associations between ECG-predicted LVH phenotypes and incident major adverse cardiovascular events (MACE) and heart failure. Results: Among UKB participants 19 408 had normal LV, 758 LV remodelling, 181 eccentric and 92 concentric LVH. Classification performance of the three models was comparable in UKB. SVM (accuracy 0.79, sensitivity 0.59, specificity 0.87, AUC 0.69) was taken forward for external validation with similar results in SHIP. There was superior prediction of eccentric LVH in both cohorts. In the UKB test set, ECG-predicted eccentric LVH was associated with heart failure (hazard ratio 3.42, 95% CI 1.06–9.86). Conclusion: ECG-based ML classifiers represent a potentially accessible screening strategy for the early detection of hypertension-mediated LVH phenotypes.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Sudden cardiac death risk prediction from QT–RR adaptation time lag computed from exercise stress test ECG

  • Pérez, Cristina
  • Martínez, Juan Pablo
  • Ramírez, Julia
  • Kenttä, Tuomas
  • Junttila, Juhani
  • Kiviniemi, Antti
  • Perkiömäki, Juha
  • Huikuri, Heikki
  • Pueyo, Esther
  • Laguna, Pablo
Objetive: This study aims to evaluate the QT adaptation time following gradual heart rate changes estimated from exercise stress test (EST) ECGs as a marker of sudden cardiac death (SCD) risk. The predicted risk value for any cardiovascular death (CVD) is also evaluated. Methods: Three ECG-derived markers related to QT-RR adaptation time were estimated during the exercise phase of EST, τe, during the recovery phase, τr, and as the difference between them, Δτ. The values were computed from patients with coronary artery disease (CAD) from ARTEMIS study (N=1472; median follow-up of 8.9 years). These markers are calculated as the delay between the observed RR interval series and a patient-specific memoryless RR series computed from the QT interval (RR-based strategy). Alternatively, the estimates were calculated as the delay between the observed QT intervals and a patient-specific instantaneous QT series computed from the RR intervals (QT-based strategy). The relation between the markers and CAD or SCD was evaluated. Results: The marker τr is more robustly estimateed with the RR-based strategy and is able to stratify survivors and victims of either SCD or CVD (p-value equal to 0.022 and 0.065, respectively). Multivariable regression models for predicting SCD and CVD include the QT-RR adaptation time estimated in the recovery phase of EST using the RR-based strategy. Conclusion: A prolonged QT-RR adaptation time during the recovery phase of the EST ECG, calculated with the RR-based strategy, can predict SCD and CVD, providing complementary information to other clinical markers
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Deep learning to predict left ventricular hypertrophy from the electrocardiogram

  • Naderi, Hafiz
  • Kaplan, Thomas
  • van Duijvenboden, Stefan
  • Pujadas, Esmeralda Ruiz
  • Aung, Nay
  • Anwar A Chahal, C
  • Lekadir, Karim
  • Chamling, Bishwas
  • Dörr, Marcus
  • Markus, Marcello R P
  • Petersen, Steffen E
  • Ramírez, Julia
  • Munroe, Patricia B
Aims: Left ventricular hypertrophy (LVH) is a strong predictor of cardiovascular disease. We previously compared supervised machine learning techniques to classify cardiac magnetic resonance (CMR)-derived LVH using electrocardiogram (ECG) and clinical variables in 37 534 UK Biobank participants, obtaining an area under the receiving operating curve (AUROC) of 0.85, but with limited specificity and requiring external validation. In this study, we develop a deep learning (DL) model to improve classification with external evaluation in the Study of Health in Pomerania (SHIP). Methods and results: We analysed 12-lead ECGs of 48 835 participants from the UK Biobank imaging study. The dataset was split into a training set (70%), validation set (15%), and test set (15%) for performance evaluation. The model architecture was a fully convolutional network, for which the input was the participants’ median ECG and clinical variables and the predicted indexed left ventricular mass (iLVM) as the output. A subsequent logistic regression model was used to recalibrate iLVM predictions. In UK Biobank, 717 (1.5%) participants had CMR-derived LVH and the AUROC for the DL model was 0.97. The ECG components most predictive of LVH were the QRS complex and ventricular rate. The DL model outperformed our supervised algorithms, previous DL modelling efforts and clinical ECG benchmarks. There was modest generalizability of the DL model to 1423 participants in SHIP (AUROC 0.78), with differences in clinical profile, ECG acquisition, and CMR labelling as important factors. Conclusion: Our findings support the feasibility of scalable DL-based screening tools for the prediction of LVH from the ECG, whilst highlighting the need for model development using larger datasets with greater diversity to ensure generalizability.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

T-wave Peak-to-End Changes, Quantified by Time-Warping, Surrogates Ventricular Repolarization Dispersion for Ventricular Fibrillation Prediction in a Porcine Myocardial Infarction Model

  • Gómez, Neurys
  • Abramochkina, Alena S.
  • Ramírez, Julia
  • Ovechkin, Alexey O.
  • Bernikova, Olesya G.
  • Platonov, Pyotr G.
  • Azarov, Jan E.
  • Laguna, Pablo
  • Martínez, Juan Pablo
Background: Ventricular fibrillation (VF) during acute ischemia is linked to increased dispersion of ventricular repolarization (DOR), which is considered a robust marker of arrhythmic susceptibility. In previous works, we showed that a time-warping-based index quantifying morphological changes from the T-wave peak to the T-wave end over a PCA transform lead, dPCAw,Tpe, outperformed the T-peak-to-T-end interval (Tpe) in arrhythmic risk prediction in a closed-chest porcine infarction model. In this study, we compare dPCAw,Tpe with the DOR measured in an open-chest porcine infarction model, to test the hypothesis that dPCAw,Tpe is a noninvasive surrogate for DOR, investigating its predictive value for VF occurrence. Methods: Myocardial ischemia was induced by coronary ligation in 55 domestic pigs. Surface ECG and intramyocardial electrograms were recorded simultaneously. DOR was computed as the range of the end-of-repolarization times across intramyocardial electrograms, while dPCAw,Tpe was extracted from Tpe morphology, measured in a spatially transformed PCA lead and using a warping-based approach. Additionally, the classical Tpe interval was computed as a reference. Indices were compared between pigs developing delayed VF (Delayed-VF group) and those who did not (Non-VF group). Results: dPCAw,Tpe, DOR and Tpe remained stable at baseline but showed progressive changes during ischemia, with larger changes in animals that developed VF. At 5, 10, and 15 minutes after occlusion, ROC analysis showed significant discriminatory capacity, with AUC values consistently ≥0.79 for all indexes and an early predictive value with AUC of 0.870 for dPCAw,Tpe at 5 minutes, accompanied by balanced sensitivity and specificity values. Cox regression for VF risk prediction resulted in hazard ratios ranging from 5.86 to 11.95 across time points. Kaplan-Meier survival curves stratified by ROC-derived thresholds showed significant separation between groups at all three time points. Linear regression analyses between dPCAw,Tpe and DOR showed consistent correlation in VF-susceptible animals (median R2=0.33). This association was stronger than that observed between Tpe and DOR (median R2=0.16). Conclusions: The time-warping-based index, dPCAw,Tpe, is a reliable non-invasive surrogate of DOR, with comparable predictive value for delayed VF risk under ischemic conditions, providing earlier and more robust prediction than the Tpe index. These findings highlight their translational potential and require validation in human cohort studies.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Imaging-Derived Sarcopenic Obesity and Cardiovascular Outcomes

  • Sanghvi, Mihir M.
  • Nicholls, Hannah L.
  • Kaplan, Thomas M.
  • Chadalavada, Sucharitha
  • Ramírez, Julia
  • Stokes, Tanner
  • Lim, Changhyun
  • Mcleod, Jonathan C.
  • Phillips, Stuart M.
  • Phillips, Bethan E.
  • Atherton, Philip J.
  • Khanji, Mohammed Y.
  • Petersen, Steffen E.
  • Timmons, James A.
  • Munroe, Patricia B.
  • Aung, Nay
Background The prognostic value of obesity in cardiovascular disease is complex. Measures such as body mass index and waist-to-height ratio show differing associations with outcomes, especially in heart failure. Assessing sarcopenic obesity, the coexistence of excess fat and low muscle mass, may clarify this relationship; however, quantifying sarcopenia in clinical practice remains challenging. Objectives The goal was to establish a translatable method of assessing sarcopenic obesity from cardiovascular imaging and assess its clinical relevance using long-term follow-up, genomic, and transcriptomic data. Methods We developed a deep learning pipeline to quantify pectoralis major muscle mass from 55,768 cardiovascular magnetic resonance examinations and combined this with body weight to derive a novel sarcopenic obesity index. Associations with cardiac remodeling phenotypes and adverse cardiovascular and mortality outcomes were tested in multivariable models. Genome-wide association analysis, colocalization, and polygenic risk score evaluation were performed for the sarcopenic obesity index. Transcriptomic profiling of skeletal muscle across 7 pathophysiological states assessed differential gene expression. Results A higher sarcopenic obesity index was associated with adverse cardiac remodeling, and with increased risk of incident heart failure (HR: 1.31; 95% CI: 1.16-1.49), cardiovascular death (HR: 1.51; 95% CI: 1.25-1.81), and all-cause mortality (HR: 1.37; 95% CI: 1.26-1.49). Genome-wide association analysis identified 16 loci for sarcopenic obesity. Loci included genes associated with heart failure and nonischemic cardiomyopathy, with colocalization implicating shared causal variants in heart failure. Transcriptomic profiling demonstrated that sarcopenic obesity loci were specifically modulated during muscle atrophy. Analyses identified ACVR2B, the target of bimagrumab, which is being tested with semaglutide to enhance fat loss while maintaining lean mass. Conclusions These results establish sarcopenic obesity as a clinically meaningful cardiovascular risk phenotype and point to viable therapeutic targets.