Cardiology Research Articles

Contrast echocardiography: What's new?

Over the last two decades, regardless of advances in cardiac imaging technologies, echocardiography has sustained its leading role as a diagnostic tool in cardiovascular medicine. The foremost reason is because of its unique advantages comprising versatility, portability, rapid availability even at the bed side, excellent temporal resolution, elucidation of images in real time, comparatively low cost and above all no risk of exposure to radiation. In multiple clinical situations 2Dimensional echocardiography (2DE) is operated for imaging in OPD clinics, indoor wards, intensive care units, operation theatres and faraway locations.

S. Padmavati: The Trailblazing Pioneer of Cardiovascular Medicine in India

S. Padmavati: The Trailblazing Pioneer of Cardiovascular Medicine in India

Can Radiological Renal Artery Parameters Predict Acute Kidney Injury in Infective Endocarditis Surgery?—From Imaging to Outcomes

Background: Infective endocarditis (IE) poses significant challenges in cardiovascular medicine, often necessitating valvular surgery to manage severe complications. Postoperative acute kidney injury (AKI) is a notable complication affecting patient outcomes. While clinical and procedural factors have been well studied, the role of radiological renal artery parameters in AKI risk remains underexplored. Methods: This retrospective study analyzed 80 patients with IE who underwent valvular surgery from 2013 to 2021, focusing on postoperative AKI as defined by the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. Radiological parameters, including renal artery calcification, renal ostial calcification, the presence of renal infarctions, and additional arteries, were assessed using preoperative computed tomography (CT). Statistical analyses included binary logistic and linear regression models, Kaplan–Meier survival curves, and Cox proportional hazard regression to explore associations between these parameters and AKI incidence, creatinine levels, and mortality. Results: Out of 80 patients, 31 (38.8%) developed AKI. No significant differences were found in baseline characteristics or radiological parameters between the AKI+ and AKI− groups. Binary logistic and linear regression analyses revealed no substantial relationship between anatomical factors and AKI risk or creatinine levels. However, Cox regression identified “additional renal artery” as a significant predictor of 1-month mortality (HR: 1.747, 95% CI: 1.024–2.979, p = 0.041) but not for 6- or 12-month mortality. Conclusions: Radiological anatomical factors, including calcifications and additional arteries, did not significantly impact AKI risk in IE patients undergoing valvular surgery. However, the presence of additional arteries was associated with increased short-term mortality. These findings suggest the need for further research to elucidate factors contributing to AKI and mortality in this context.

Abstract 4128519: Performance of a Popular Large Language Model in Answering Cardiovascular Related Queries: A Systematic Review and Pooled-Analysis

Background: The integration of large language models (LLMs) such as ChatGPT into healthcare can have significant implications for patient education and clinical decision-making. Aims: This systematic review and pooled analysis aim to evaluate the accuracy of ChatGPT 3.5 and 4 in answering simple queries across cardiovascular (CV) medicine disciplines. Methods: Literature searches were conducted in PubMed, Embase, and Cochrane Central in May 2024. Keywords included “ChatGPT”, “LLMs”, and “Chat-based artificial intelligence models”. Cross-sectional, peer-reviewed studies published in 2023 and 2024 investigating ChatGPT’s performance in CV medicine-related queries (Table/Figure) were extracted and included. All queries were evaluated by expert physicians in the corresponding fields within each study (and not by our readjudication), and a standardized grading system was employed for pooled analysis using an "accurate" and "inaccurate" grading scale for each answer. Results: Out of 127 identified and screened peer-reviewed studies, fourteen studies involving 542 CV-related queries were included. Pooled analysis revealed an overall accuracy of 84.5% (458/542) (95% CI [81.5, 87.6]). Stratification by model (ChatGPT-4 vs. ChatGPT-3.5) did not show a significant difference in accuracy (p=0.32). Furthermore, no significant differences in accuracies were seen between answers in 2023 and 2024 (p=0.07). The accuracies across the various topics were statistically comparable, except in the field of cardio-oncology, which showed significantly lower accuracy at 68% (p=0.02). Detailed performances per topic are included in the table and figure. Conclusion: ChatGPT demonstrated consistently high accuracy in answering CV-related queries with no significant differences across model versions or years. These results support the potential use of online-chat based LLMs as an informational tool in cardiology.

Abstract 4135265: When the Beat Drops: A Pilot Project for a Case-Based Resident’s Guide to Cardiovascular Care

Residents care for patients with a vast array of cardiovascular (CV) diagnoses in settings ranging from primary care clinic to the cardiac intensive care unit. However, the resources available are mainly intended for advanced cardiology trainees. We created a resident-driven, case-based reference for residents to serve as a companion to traditional CV medicine educational resources. Methods: We surveyed residents, fellows, and cardiology attendings to identify topics within CV care for which residents are responsible. We recruited residents from multiple ACGME-accredited Internal Medicine residency programs to write a chapter on each identified topic that would be published on a central website. The chapters included a topic overview with key learning points, a case presentation, review questions, attending pearls, and high-yield articles for further reading. Chapters were reviewed for accuracy and completeness by cardiology fellows, pharmacologists, and attendings. The chapters were sent out to residents weekly via email and through social media platforms. Website analytics were used to track usage and monitor engagement. Each chapter concluded with a brief survey regarding learning gains, comprehension, and if the reader would recommend the resources to others. Results: A total of 52 cases were submitted by resident contributors from four residency programs. Topics included management of arrhythmias, structural heart disease, ischemia, cardiogenic shock, and post-procedure complications. Within the first two months of publication, the website was being accessed throughout 30+ states and 6 countries with more than 2,500 views and 1,500 unique visitors. Topics with high traffic included chapters regarding arrhythmias and post-procedural complications. The website was viewed throughout the entire day with no specific time pattern; however, there were significantly more views on days when each chapter was released via email and social media. Preliminary survey results show that residents and other trainees find this resource useful, and many want to get involved and help create more content. Conclusion: Our resident-created and cardiologist reviewed website is a unique resource created for residents caring for patients with CV disease. Using email and social media marketing has allowed this resource to be widely accessible to a large global audience. Survey feedback suggests the importance of this resource and need for continued CV resources for trainees.

The Future of Artificial Intelligence and Genetic Insights in Precision Cardiovascular Medicine: A Comprehensive Review

Abstract A medical condition called cardiovascular disease (CVD) affects the heart or blood vessels, and about 40% of its causes can be attributed to genetic factors. The pathophysiology of CVD is still unknown despite numerous studies identifying important environmental and genetic factors. Genetic data research has significantly increased due to the application of genome-wide association studies. The utilization of artificial intelligence (AI) technology demonstrates clear advantages in managing intricate projects, outperforming traditional statistical methods in processing such data. The use of AI in the status of genetic research on CVD and medicine is briefly reviewed in the opening section of this article. Then, it gives a complete picture of how AI is used in genetic CVD research, including genetic data-driven diagnosis and prognosis, genetic variation analysis, gene expression profiles, gene interactions, and analysis of genes using knowledge bases. Even though much research has yielded significant findings, it is still early. The main disadvantages are database limitations, the underuse of AI in systematic biology analysis, and the lack of a theoretical framework for interpreting analysis results. The paper concludes with future directions and the significance of creating comprehensive, high-quality, large-sample-size data-sharing resources. Much research is going into how to use AI analysis techniques to help with development. Being creative with computers can help make new CVD intervention protocols and develop and test theoretical models.

Productivity and Pay in Cardiovascular Medicine and Surgery: Is the Financial Accounting Link Broken?

Productivity and Pay in Cardiovascular Medicine and Surgery: Is the Financial Accounting Link Broken?

Glucagon-like peptide-1 agonists in cardiovascular diseases: a bibliometric analysis from inception to 2023

Background: In recent years, glucagon-like peptide-1 (GLP-1) agonists have garnered increasing attention for their potential cardiovascular benefits beyond glycemic control in patients with diabetes. Understanding the research landscape surrounding GLP-1 agonists and cardiovascular diseases (CVDs) is crucial for informing clinical practice and guiding future research endeavors. This bibliometric analysis aimed to comprehensively assess the scholarly output and trends in this field, shedding light on the evolving landscape of GLP-1 agonists’ role in cardiovascular health. Methods: The publications concerning GLP-1 agonists in CVDs were gathered from the Web of Science Core Collection, and visualizations were created utilizing Excel 2019, Cite Space, and VOS viewer software. Results and Conclusion: Using bibliometric and visual methods, the research hotspots and trends regarding GLP-1 agonists in cardiovascular diseases were pinpointed. Additionally, a thriving interest in GLP-1 agonists research within cardiovascular medicine was observed, with a notable surge in publications from 2016 onwards. The analysis revealed that the United States and China are the leading contributors, accounting for over 50% of the total publications. The University of Copenhagen and the University of Toronto emerged as the most prolific institutions in this field. Co-citation analysis highlighted the influential role of landmark clinical trials, such as the LEADER, ELIXA, and EXSCEL. Keyword trend analysis identified the emergence of newer GLP-1 agonists, such as tirzepatide and semaglutide, as well as a growing focus on topics like ‘healthy obesity’ and chronic kidney disease. These findings suggest that the research landscape is evolving, with a focus on expanding the therapeutic applications of GLP-1 agonists beyond glycemic control. Overall, this bibliometric analysis provided insights into the current state and future directions of research on GLP-1 agonists and their impact on cardiovascular health, guiding future research endeavors, and informing clinical practice.

Understanding European patterns of deprescribing antithrombotic medication during end-of-life care in patients with cancer

BackgroundEven though antithrombotic therapy (ATT) probably has little or even negative effect on the well-being of patients with cancer near the end of life, it is often continued until death, possibly leading to excess bleeding complications, increased disease burden, reduced quality of life and higher healthcare costs. AimTo explore and describe European practice patterns and perspectives of healthcare professionals from different disciplines and specialties on ATT in the end-of-life care (EOLC) of patients with cancer. MethodsWe performed a two-week international cross-sectional survey study using flash-mob research methodology. Eligible were healthcare professionals from different institutions across Europe, who prescribed ATT and/or dealt with EOLC of patients with cancer. The survey comprised three parts, including a series of choice sets (hypothetical scenarios involving a set of characteristics changing in level [e.g., high vs. low thrombotic risk]) on ATT management in EOLC. The discrete choice experiment analysis was conducted using multinomial logistic regression. ResultsOut of 467 pre-registrants, 208 participated in the survey from 4 to 18 July 2023. The majority (53 %) considered a patient with cancer as in EOLC when life expectancy is below 3 months. Respondents reported seeing or treating 20 patients with cancer on ATT in EOLC per year (IQR 10–50). The median estimated frequency of considering ATT deprescription per healthcare professional was 10 times per year (IQR 4–10), while the frequency of actual deprescription was 5 times per year (IQR 2–10). Twenty percent of respondents had never deprescribed ATT in the context of EOLC. Across the eight choice sets, five respondents (2.7 %) found deprescribing inappropriate in any scenario. Deprescribing was more often considered in patients with poor ECOG-performance status, high bleeding risk and low-molecular-weight heparin use as opposed to oral ATT. Haemato-oncology and cardiovascular medicine specialists were more inclined to deprescribe antiplatelet therapy than other specialties. ConclusionOur study describes medical decision-making regarding ATT in EOLC of patients with cancer. Healthcare professionals' perspectives and practice patterns vary, and some preferences appear associated with the therapists' professional focus and region of practice.

Circular RNAs in Cardiovascular Diseases: Molecular Mechanisms, Therapeutic Advances, and Innovations

Circular RNAs (circRNAs) have emerged as promising therapeutic targets due to their unique covalently closed-loop structures and their regulatory roles in gene expression. Despite their potential, challenges in circRNA-based therapies include ensuring stability, tissue specificity, and efficient intracellular delivery. This review explores the implications of circRNAs in cardiovascular diseases (CVDs), providing an overview of their biogenesis, molecular mechanisms, and roles in disease pathology. In addition to discussing molecular features, this review highlights therapeutic advances, including small-molecule drugs targeting circRNAs, synthetic circRNA sponges, and innovations in drug delivery systems that enhance the effectiveness of these therapies. Finally, current challenges and future directions are addressed, emphasizing the need for continued research to fully unlock the therapeutic potential of circRNA-based strategies in cardiovascular medicine.

Harnessing the Power of ChatGPT in Cardiovascular Medicine: Innovations, Challenges, and Future Directions

Harnessing the Power of ChatGPT in Cardiovascular Medicine: Innovations, Challenges, and Future Directions

Induced Pluripotent Stem Cells: A New Dawn for the Treatment of Ischemic Cardiomyopathy

Ischemic cardiomyopathy, a leading cause of heart failure globally, arises from the irreversible damage inflicted upon cardiac tissue following myocardial infarction. The limited regenerative capacity of the adult human heart necessitates the exploration of novel therapeutic strategies to restore cardiac function and improve outcomes. Induced pluripotent stem cells (iPSCs), with their remarkable ability to differentiate into virtually any cell type in the body, have emerged as a beacon of hope for regenerative medicine, including the treatment of ischemic cardiomyopathy. This review delves into the pathophysiology of ischemic cardiomyopathy, highlighting the cellular and molecular mechanisms underlying cardiac dysfunction. We provide a comprehensive overview of the current state of iPSC technology, focusing on its potential applications in disease modeling, drug discovery, and cell-based therapies for cardiac regeneration. Furthermore, we discuss the challenges and opportunities associated with translating iPSC-based therapies to the clinic, emphasizing the need for rigorous preclinical studies and well-designed clinical trials to ensure safety and efficacy. Finally, we offer perspectives on the future directions of iPSC research in the context of ischemic cardiomyopathy, highlighting the transformative potential of this technology to revolutionize cardiovascular medicine and usher in an era of personalized regenerative therapies for patients with heart failure.

Cardiovascular Imaging in the Era of Precision Medicine: Insights from Advanced Technologies - A Narrative Review.

Cardiovascular diseases are responsible for a high mortality rate globally. Precision medicine has emerged as an essential tool for improving cardiovascular disease outcomes. In this context, using advanced imaging exams is fundamental in cardiovascular precision medicine, enabling more accurate diagnoses and customized treatments. This review aims to provide a concise review on how advanced cardiovascular imaging supports precision medicine, highlighting its benefits, challenges, and future directions. A literature review was carried out using the Pubmed and Google Scholar databases, using search strategies that combined terms such as precision medicine, cardiovascular diseases, and imaging tests. More advanced analysis aimed at diagnosing and describing cardiovascular diseases in greater detail is made possible by tests such as cardiac computed tomography, cardiac magnetic resonance imaging, and cardiac positron emission tomography. In addition, the aggregation of imaging data with other omics data allows for more personalized treatment and a better description of patient profiles. The use of advanced imaging tests is essential in cardiovascular precision medicine. Although there are still technical and ethical obstacles, it is essential that there is collaboration between health professionals, as well as investments in technology and education to better disseminate cardiovascular precision medicine and consequently promote improved patient outcomes.

Heart disease detection using an acceleration-deceleration curve-based neural network with consumer-grade smartwatch data

Cardiovascular disease (CVD) is the most important cause of morbidity and mortality worldwide. Early detection, prevention or even prediction is of pivotal importance to reduce the burden of cardiovascular disease and its associated costs. Low cost, consumer-grade smartwatches have the potential to revolutionize cardiovascular medicine by enabling continuous monitoring of heart rate and activity. When combined with machine learning(ML), the resulting large amounts of time series data hold the potential of detection, or exclusion of CVD. However, analyzing such large datasets is challenging due to the sparse presence of informative segments. Efficient selection of these segments is essential for developing predictive models for clinical deployment. The objective of this paper was to investigate the potential of an acceleration-deceleration curvebased ML model as a novel clinical indicator for the detection of cardiovascular diseases. We used data from the ME-TIME study; 42 participants from which 21 have a cardiovascular disease and 21 are health controls. Data from each subject was normalized to decrease inter-subject variability. A neural network model aggregated predictions per week. We showed that per-subject normalization by the peak value of curves during inactivity, aggregation of model predictions over a week, and using a contrastive loss, resulted in a predictive model with 99 % ± 3 % specificity and 40 % ± 49 % sensitivity on the development set, and 100 % specificity with 67 % ± 47 % sensitivity on the test set. Acceleration-deceleration curves are effective patterns for ruling out the presence of cardiovascular disease, but caution must be taken to properly pre-process the curves and carefully choosing a model that reduces the variability in the extracted curves.

Artificial Intelligence Algorithms in Cardiovascular Medicine: An Attainable Promise to Improve Patient Outcomes or an Inaccessible Investment?

This opinion paper highlights the advancements in artificial intelligence (AI) technology for cardiovascular disease (CVD), presents best practices and transformative impacts, and addresses current concerns that must be resolved for broader adoption. With the evolution of digitization in data collection, large amounts of data have become available, surpassing the human capacity for processing and analysis, thus enabling the application of AI. These models can learn complex spatial and temporal patterns from large amounts of data, providing patient-specific outputs. These advantages have resulted, at the moment, in more than 900 AI-based devices being approved, today, by regulatory entities, for clinical use, with similar to improved performance and efficiency compared to traditional technologies. However, issues such as model generalization, bias, transparency, interpretability, accountability, and data privacy remain significant barriers for broad adoption of these technologies. AI shows great promise in enhancing CVD care through more accurate and efficient approaches. Yet, widespread adoption is hindered by unresolved concerns of interested stakeholders. Addressing these challenges is crucial for fully integrating AI into clinical practice and shaping the future of CVD prevention, diagnosis and treatment.

Factors associated with non-publication of clinical trials in cardiovascular medicine: a cross-sectional analysis

Abstract Background/introduction Cardiovascular medicine is the fourth most-funded area of clinical research as it received an estimate for 7% of the total research funding. Nearly 8% of all clinical trials are cardiovascular-related (1,2). Non-publication of the registered clinical trials wastes money and time, encourages dissemination bias, and creates a body of possibly inaccurate literature, all of which obstruct scientific advancement. (3) Purpose To study the factors associated with the non-publication of cardiovascular clinical trials between January 2000 and June 2022. Methods We performed a cross-sectional analysis using ClinicalTrials.gov to determine all completed, discontinued, published, and unpublished cardiovascular medicine clinical trials between January 2000 and June 2022. For each trial, data on the trial phase, funding source, intervention type, enrolment size, trial completion, and publication status were extracted and subjected to logistic regression to determine clinical trial publication predictor factors. Results Among 3946 trials in cardiovascular medicine registered in ClinicalTrials.gov, there were 3215 completed trials included in the analysis. Of these, 34.2% (n=1100) were unpublished. On multivariate logistic regression, triple-blinded and quadruple-blinded trials were less likely to be unpublished compared to open-label trials (odds ratio [OR] 0.622; 95% confidence interval [CI] 0.45-0.85; p=0.003) and (OR 0.54; CI 0.42-0.69; p<0.001), respectively. Enrolment of 100 participants or more was associated with a decreased likelihood of non-publication (OR 0.46; CI 0.39-0.55; p<0.001). In addition, randomised trials were less likely to be unpublished than non-randomised trials (OR 0.67; CI 0.50-0.88; p=0.004). There was no significant effect of the intervention type, phase of the trial, trial participants' age and gender, or the trial funding resource on the publication status of cardiovascular medicine clinical trials. Conclusions It is observed that a considerable number of clinical trials in cardiovascular medicine remain unpublished. This issue raises some concerns regarding the risks and discomfort participants may encounter during the trial, while no results are added to the literature. Blinding of the trials, enrolment size, and allocation methods are the main predictors of trial publication.Forest plot of logistic regression

Artificial intelligence derived electrocardiograms: a comparative study of network architectures predicting sex and left ventricular dysfunction

Abstract Background Within cardiovascular medicine, AI-driven electrocardiogram (ECG) analysis has emerged as a powerful tool, capable of predicting clinically significant abnormalities. The increasing growth of wearable technologies for ECG monitoring underscores the importance of advancing AI methodologies. While Convolutional Neural Networks (CNNs) have traditionally been the go-to choice for ECG analysis, the rise of Vision Transformers (ViTs), a novel computer vision paradigm, raises the question of their potential superiority. Purpose To compare the performance of CNNs and ViTs in ECG analysis, and to assess the efficiency of lean models using truncated ECG recordings. Methods We trained a model using 12-lead ECGs as input for the binary classification of two outputs: Sex (Male or Female) and Left ventricular dysfunction (LVD, defined as below 35%), determined by echocardiography within 2 weeks of the ECG. Models were developed using full standard ECGs, single lead and truncated recordings (1,2 and 4 10 seconds). We compared CNNs (PyTorch) and ViTs (HuggingFace) performance. We analyzed the explainability of our findings, representing the AI model's focus of interest on ECGs from normal cases versus those with reduced ejection fraction, thereby validating its diagnostic discernment (Fig 1) Results We identified 150,691 and 29,422 patients with valid ECG tests for the Sex and LVD prediction models, respectively. For the Sex outcome, the AUROCs were 0.911 (95% CI: 0.908-0.914) and 0.898 (95% CI: 0.894-0.901) for the CNN-based model and the ViT-based models, respectively, with the difference statistically significant (P-Value < 0.001). For the LVD outcome, the AUROCs were comparable at 0.878 (95% CI: 0.864-0.892) and 0.866 (95% CI: 0.853-0.879) for the CNN-based model and the ViT-based models, respectively, (P-Value = 0.056). Furthermore, we found that 98.8% for the maximal predictive power of the models was achieved within a 2 second time frame of a 12 lead ECG. Similarly, a single lead based model achieved a relatively high AUC for the prediction of LVD. Conclusions AI-ECG models developed with ViT architecture did not surpass CNN-based models in sex classification and LVD identification. Our study highlights the significance of lean models allowing for rapid prediction within 2-second tracings or using a single lead, demonstrates the potential for streamlined AI-ECG applications in clinical practice.

Comparative analysis of vitamin K antagonists and direct oral anticoagulants for left ventricular thrombus management: a retrospective study utilising TriNetX

Abstract Introduction Despite progress in cardiovascular medicine, guidelines for addressing left ventricular thrombus (LVT) remain limited. Furthermore, incorporating oral anticoagulants into dual antiplatelet therapy for acute myocardial infarction (MI) adds complexity to treatment decisions. While conventional guidance favors vitamin K antagonists (VKAs), emerging evidence suggests favorable outcomes with direct oral anticoagulants (DOACs). Purpose This study aims to assess the safety and efficacy of DOACs versus VKAs in managing LVT among patients with and without recent acute coronary syndrome (ACS). Methods This was a retrospective observational study conducted within TriNetX, a global federated health research network with access to electronic medical records (EMRs) from participating health care organizations including academic medical centres and community hospitals covering approximately 70 million individuals, mainly in the United States. The search was conducted on 26th Februrary 2024 with a study cohort comprising patients with LV thrombus treated with either DOAC or VKA between 1st December 2013 to 1st December 2023. Subgroup analyses were conducted for patients with ACS within one month and those without ACS within a month of treatment. Cohort data were subject to propensity matching (PSM) for age, gender, ethnicities, medical and drug history. Risk analysis and Kaplan-Meier survival analysis were computed for each subgroup at 90 days since the indexed event. Results Following PSM, a total of 39,770 patients were included and 14,302 were in the ACS group and 24,162 in the non-ACS group. DOAC treatment exhibited favourable outcomes for stroke, bleeding, and systemic embolism (p<0.05) compared to VKA in the overall cohort (Table 1). In the ACS group, DOAC was linked to a reduced risk of stroke and systemic embolism (p<0.05) but not bleeding (p=0.066). In the non-ACS group, DOAC was associated with a decreased risk of stroke and bleeding (p<0.05) but not systemic embolism (p=0.113). Across all groups, there were no disparities in overall all causes of mortalities. Conclusion DOACs demonstrated better safety and efficacy outcomes when compared to VKAs in LVT treatment. Interestingly, DOACs utilisation within the ACS context resulted in a diminished risk of embolic complications without the benefit of reduced bleeding risk seen in patients without ACS, suggesting potential implications of triple therapy within first month of ACS treatment, removing the bleeding risk reduction of DOACs. This shows a fundamental difference in LVT between the 2 groups of patients so future studies should aim to differentiate between the 2.

Sex related cardiovascular risk is non-dichotomous: artificial intelligence enhanced electrocardiography reveals continuum of risk in females

Abstract Background Females are typically underserved in cardiovascular medicine and often considered lower risk of cardiovascular disease. The use of sex as a dichotomous variable for risk stratification fails to capture the heterogeneity of risk within each sex. Purpose We aimed to develop an artificial intelligence-enhanced ECG (AI-ECG)-derived continuous sex score that can capture the continuum of risk and sex phenotypes within each sex, with the particular goal of addressing the female disadvantage. Methods We trained a convolution neural network to identify sex using the 12-lead ECG. The derivation cohort was 1,163,401 ECGs from 189,538 subjects from a USA secondary care hospital cohort and validated inthe UK Biobank (UKB). The model outputs a continuous value from 0 to 1 for sex prediction. The AI-ECG sex discordance score is the difference between predicted sex and biological sex. Results AI-ECG accurately identified sex in both the USA cohort (AUROC 0.943 (0.942-0.943)) and UKB datasets (n = 42,386, AUROC 0.971 (0.969-0.972)). In explainability analyses we found QRS duration, T wave morphology, QT interval, heart rate and QRS voltage as the most important factors in contributing to AI-ECG sex identification. In females only but not in males, sex discordance score was associated with a covariate-adjusted increased risk of cardiovascular death in outpatients with normal ECGs from the USA cohort, (Females hazard ratio (HR) 1.64 (1.18-2.29) p = 0.003, Males HR 1.04 (0.66-1.63), p = 0.86). This pattern persisted in the UKB: Females HR 1.39 (1.08-1.78) p = 0.01, Males HR 0.97 (0.77-1.22), p = 0.80, Fig 1). In phenome-wide association studies, we found females with a higher sex score discordance were more likely to have heart failure or myocardial infarction. Covariate-adjusted cox models confirmed the association of sex score discordance with increased future heart failure (HR 1.21 (1.07-1.37) p = 0.002 and future myocardial infarction (HR 1.32 (1.11-1.56) p = 0.001). In age- and body size-adjusted analyses, females with higher sex discordance score had increased left ventricular mass, and chamber volumes as well as increased muscle mass, reduced body fat percentage (Fig 2). In a genome-wide association study of sex discordance score we identified variants adjacent to IGF1R and NDRG4 in females, which have been previously associated with LV mass and sex hormone levels (Fig 2). Conclusion We describe sex discordance score as a novel AI-ECG biomarker capable of identifying females with disproportionately elevated cardiovascular risk. Sex discordance score may help uncover potential underlying mechanisms of cardiovascular disease among females, and future clinical studies are needed to determine its potential utility in identification of females at higher risk of cardiovascular events. Rather than exacerbating biases, AI-ECG has the potential to reduce female cardiovascular health inequalities.

Assessing the need for coronary angiography in high-risk acute coronary syndrome patients using artificial intelligence and computed tomography

Abstract Background/Introduction The use of coronary computed tomography angiography (CCTA) in managing high-risk acute coronary syndrome (ACS) patients is increasingly common. Yet, its role remains unvalidated in this context, posing challenges to traditional clinical decision-making algorithms. Physicians, accustomed to a hierarchical and structured approach involving symptom assessment, electrocardiograms, and biomarkers for deciding on the necessity of invasive coronary angiography (ICA), now face dilemmas when CCTA results challenge established diagnostic pathways and contradict clinical decisions based on usual criteria. Meanwhile, the potential of artificial intelligence (AI) which processes information in a way fundamentally different from human clinical reasoning, has been shown to aid decision-making in cardiovascular medicine. Yet, it has not been tested in this acute setting. Purpose This study aimed to evaluate the efficacy of the Chat Generative Pre-trained Transformer (ChatGPT) in guiding the need for ICA in high-risk ACS patients based on both standard clinical data and CCTA findings. Methods This investigation is a sub-study of a larger research project where high-risk ACS patients (all having a theoretical guideline-based indication for ICA) underwent CCTA prior to ICA. ChatGPT analyzed clinical vignettes containing patient data, electrocardiograms, troponin levels, and CCTA results to determine the necessity of ICA. The AI's recommendations were then compared to actual ICA findings to assess its decision-making accuracy. Results Among 86 patients with a clear indication for ICA, ChatGPT recommended against it for 19 patients, 16 of whom indeed had no significant findings at ICA (Figure). For 67 patients, ChatGPT advised proceeding with ICA, and a significant lesion was confirmed in 58 of them. Consequently, ChatGPT's overall accuracy stood at 86%, with a sensitivity of 95% and a specificity of 64%. The model's negative predictive value was 84%, and its positive predictive value was 87%. Conclusion These preliminary data suggest that ChatGPT can effectively assist in guiding the need for ICA among high-risk ACS patients, potentially reducing unnecessary procedures but further clinical studies to refine AI's role in clinical decision-making are required.ChatGPT's decision as per ICA's results