Utility of a Deep-Learning Algorithm to Guide Novices to Acquire Echocardiograms for Limited Diagnostic Use

Use of artificial intelligence-guided echocardiography to detect cardiac dysfunction and heart valve disease in rural and remote areas: Rationale and design of the AGILE-echo trial

BACKGROUND

Transthoracic echocardiography (TTE) is essential in the diagnosis of cardiovascular diseases (CVD), including but not limited to heart failure (HF) and heart valve disease (HVD). However, its dependence on expert acquisition means that its accessibility in rural areas may be limited, leading to delayed management decisions and potential missed diagnoses. Artificial intelligence-guided (AI)-TTE offers a solution by permitting non-expert image acquisition. The impact of AI-TTE on the timing of diagnosis and early initiation of cardioprotection is undefined.

METHODS

AGILE-Echo (use of Artificial intelligence-Guided echocardiography to assIst cardiovascuLar patient managEment) is a randomized-controlled trial conducted in 5 rural and remote areas around Australia. Adults with CV risk factors and exercise intolerance, or concerns regarding HVD are randomized into AI-TTE or usual care (UC). AI-TTE participants may have a cardiovascular problem excluded, identified (leading to AI-guided interventions) or unresolved (leading to conventional TTE). UC participants undergo usual management, including referral for standard TTE. The primary endpoint is a composite of HVD or HF diagnosis at 12-months. Subgroup analysis, stratified based on age range and sex, will be conducted. All statistical analyses will be conducted using R.

RESULTS

Of the first 157 participants, 78 have been randomized into AI-TTE (median age 68 [IQR 17]) and 79 to UC (median age 65 [IQR 17], P = .034). HVD was the primary concern in 37 participants (23.6%) while 84.7% (n = 133) experienced exercise intolerance. The overall 10-year HF incidence risk was 13.4% and 20.0% (P = .089) for UC and AI-TTE arm respectively. Atrial remodeling, left ventricular remodeling and valvular regurgitation were the most common findings. Thirty-three patients (42.3%) showed no abnormalities.

CONCLUSIONS

This randomized-controlled trial of AI-TTE will provide proof-of-concept for the role of AI-TTE in identifying pre-symptomatic HF or HVD when access to TTE is limited. Additionally, this could promote the usage of AI-TTE in rural or remote areas, ultimately improving health and quality of life of community dwelling adults with risks, signs or symptoms of cardiac dysfunction.

Leveraging the Capabilities of AI: Novice Neurology-Trained Operators Performing Cardiac POCUS in Patients with Acute Brain Injury

BACKGROUND

Cardiac point-of-care ultrasound (cPOCUS) can aid in the diagnosis and treatment of cardiac disorders. Such disorders can arise as complications of acute brain injury, but most neurologic intensive care unit (NICU) providers do not receive formal training in cPOCUS. Caption artificial intelligence (AI) uses a novel deep learning (DL) algorithm to guide novice cPOCUS users in obtaining diagnostic-quality cardiac images. The primary objective of this study was to determine how often NICU providers with minimal cPOCUS experience capture quality images using DL-guided cPOCUS as well as the association between DL-guided cPOCUS and change in management and time to formal echocardiograms in the NICU.

METHODS

From September 2020 to November 2021, neurology-trained physician assistants, residents, and fellows used DL software to perform clinically indicated cPOCUS scans in an academic tertiary NICU. Certified echocardiographers evaluated each scan independently to assess the quality of images and global interpretability of left ventricular function, right ventricular function, inferior vena cava size, and presence of pericardial effusion. Descriptive statistics with exact confidence intervals were used to calculate proportions of obtained images that were of adequate quality and that changed management. Time to first adequate cardiac images (either cPOCUS or formal echocardiography) was compared using a similar population from 2018.

RESULTS

In 153 patients, 184 scans were performed for a total of 943 image views. Three certified echocardiographers deemed 63.4% of scans as interpretable for a qualitative assessment of left ventricular size and function, 52.6% of scans as interpretable for right ventricular size and function, 34.8% of scans as interpretable for inferior vena cava size and variability, and 47.2% of scans as interpretable for the presence of pericardial effusion. Thirty-seven percent of screening scans changed management, most commonly adjusting fluid goals (81.2%). Time to first adequate cardiac images decreased significantly from 3.1 to 1.7 days (p < 0.001).

CONCLUSIONS

With DL guidance, neurology providers with minimal to no cPOCUS training were often able to obtain diagnostic-quality cardiac images, which informed management changes and significantly decreased time to cardiac imaging.

Advancements in Incident Heart Failure Risk Prediction and Screening Tools

Heart failure (HF) is a major cause of mortality and morbidity in the United States that carries substantial healthcare costs. Multiple risk prediction models and strategies have been developed over the past 30 years with the aim of identifying those at high risk of developing HF and of implementing preventive therapies effectively. This review highlights recent developments in HF risk prediction tools including emerging risk factors, innovative risk prediction models, and novel screening strategies from artificial intelligence to biomarkers. These developments allow more accurate prediction, but their impact on clinical outcomes remains to be investigated. Implementation of these risk models in clinical practice is a considerable challenge, but HF risk prediction tools offer a promising opportunity to improve outcomes while maintaining value.

Deep Learning Based Shear Wave Detection and Segmentation Tool for Use in Point-of-Care for Chronic Liver Disease Assessments

OBJECTIVE

As metabolic dysfunction-associated steatotic liver disease (MASLD) becomes more prevalent worldwide, it is imperative to create more accurate technologies that make it easy to assess the liver in a point-of-care setting. The aim of this study is to test the performance of a new software tool implemented in Velacur (Sonic Incytes), a liver stiffness and ultrasound attenuation measurement device, on patients with MASLD. This tool employs a deep learning-based method to detect and segment shear waves in the liver tissue for subsequent analysis to improve tissue characterization for patient diagnosis.

METHODS

This new tool consists of a deep learning based algorithm, which was trained on 15,045 expert-segmented images from 103 patients, using a U-Net architecture. The algorithm was then tested on 4429 images from 36 volunteers and patients with MASLD. Test subjects were scanned at different clinics with different Velacur operators. Evaluation was performed on both individual images (image based) and averaged across all images collected from a patient (patient based). Ground truth was defined by expert segmentation of the shear waves within each image. For evaluation, sensitivity and specificity for correct wave detection in the image were calculated. For those images containing waves, the Dice coefficient was calculated. A prototype of the software tool was also implemented on Velacur and assessed by operators in real world settings.

RESULTS

The wave detection algorithm had a sensitivity of 81% and a specificity of 84%, with a Dice coefficient of 0.74 and 0.75 for image based and patient-based averages respectively. The implementation of this software tool as an overlay on the B-Mode ultrasound resulted in improved exam quality collected by operators.

CONCLUSION

The shear wave algorithm performed well on a test set of volunteers and patients with metabolic dysfunction-associated steatotic liver disease. The addition of this software tool, implemented on the Velacur system, improved the quality of the liver assessments performed in a real world, point of care setting.

A multi-task deep learning approach for real-time view classification and quality assessment of echocardiographic images

High-quality standard views in two-dimensional echocardiography are essential for accurate cardiovascular disease diagnosis and treatment decisions. However, the quality of echocardiographic images is highly dependent on the practitioner's experience. Ensuring timely quality control of echocardiographic images in the clinical setting remains a significant challenge. In this study, we aimed to propose new quality assessment criteria and develop a multi-task deep learning model for real-time multi-view classification and image quality assessment (six standard views and "others"). A total of 170,311 echocardiographic images collected between 2015 and 2022 were utilized to develop and evaluate the model. On the test set, the model achieved an overall classification accuracy of 97.8% (95%CI 97.7-98.0) and a mean absolute error of 6.54 (95%CI 6.43-6.66). A single-frame inference time of 2.8 ms was achieved, meeting real-time requirements. We also analyzed pre-stored images from three distinct groups of echocardiographers (junior, senior, and expert) to evaluate the clinical feasibility of the model. Our multi-task model can provide objective, reproducible, and clinically significant view quality assessment results for echocardiographic images, potentially optimizing the clinical image acquisition process and improving AI-assisted diagnosis accuracy.

Challenges for augmenting intelligence in cardiac imaging

Artificial Intelligence (AI), through deep learning, has brought automation and predictive capabilities to cardiac imaging. However, despite considerable investment, tangible health-care cost reductions remain unproven. Although AI holds promise, there has been insufficient time for both methodological development and prospective clinical trials to establish its advantage over human interpretations in terms of its effect on patient outcomes. Challenges such as data scarcity, privacy issues, and ethical concerns impede optimal AI training. Furthermore, the absence of a unified model for the complex structure and function of the heart and evolving domain knowledge can introduce heuristic biases and influence underlying assumptions in model development. Integrating AI into diverse institutional picture archiving and communication systems and devices also presents a clinical hurdle. This hurdle is further compounded by an absence of high-quality labelled data, difficulty sharing data between institutions, and non-uniform and inadequate gold standards for external validations and comparisons of model performance in real-world settings. Nevertheless, there is a strong push in industry and academia for AI solutions in medical imaging. This Series paper reviews key studies and identifies challenges that require a pragmatic change in the approach for using AI for cardiac imaging, whereby AI is viewed as augmented intelligence to complement, not replace, human judgement. The focus should shift from isolated measurements to integrating non-linear and complex data towards identifying disease phenotypes-emphasising pattern recognition where AI excels. Algorithms should enhance imaging reports, enriching patients' understanding, communication between patients and clinicians, and shared decision making. The emergence of professional standards and guidelines is essential to address these developments and ensure the safe and effective integration of AI in cardiac imaging.

Application of artificial intelligence in the diagnosis and treatment of Kawasaki disease

This editorial provides commentary on an article titled "Potential and limitations of ChatGPT and generative artificial intelligence (AI) in medical safety education" recently published in the World Journal of Clinical Cases. AI has enormous potential for various applications in the field of Kawasaki disease (KD). One is machine learning (ML) to assist in the diagnosis of KD, and clinical prediction models have been constructed worldwide using ML; the second is using a gene signal calculation toolbox to identify KD, which can be used to monitor key clinical features and laboratory parameters of disease severity; and the third is using deep learning (DL) to assist in cardiac ultrasound detection. The performance of the DL algorithm is similar to that of experienced cardiac experts in detecting coronary artery lesions to promoting the diagnosis of KD. To effectively utilize AI in the diagnosis and treatment process of KD, it is crucial to improve the accuracy of AI decision-making using more medical data, while addressing issues related to patient personal information protection and AI decision-making responsibility. AI progress is expected to provide patients with accurate and effective medical services that will positively impact the diagnosis and treatment of KD in the future.

Artificial intelligence and myocarditis-a systematic review of current applications

Myocarditis, marked by heart muscle inflammation, poses significant clinical challenges. This study, guided by PRISMA guidelines, explores the expanding role of artificial intelligence (AI) in myocarditis, aiming to consolidate current knowledge and guide future research. Following PRISMA guidelines, a systematic review was conducted across PubMed, Cochrane Reviews, Scopus, Embase, and Web of Science databases. MeSH terms including artificial intelligence, deep learning, machine learning, myocarditis, and inflammatory cardiomyopathy were used. Inclusion criteria involved original articles utilizing AI for myocarditis, while exclusion criteria eliminated reviews, editorials, and non-AI-focused studies. The search yielded 616 articles, with 42 meeting inclusion criteria after screening. The identified articles, spanning diagnostic, survival prediction, and molecular analysis aspects, were analyzed in each subsection. Diagnostic studies showcased the versatility of AI algorithms, achieving high accuracies in myocarditis detection. Survival prediction models exhibited robust discriminatory power, particularly in emergency settings and pediatric populations. Molecular analyses demonstrated AI's potential in deciphering complex immune interactions. This systematic review provides a comprehensive overview of AI applications in myocarditis, highlighting transformative potential in diagnostics, survival prediction, and molecular understanding. Collaborative efforts are crucial for overcoming limitations and realizing AI's full potential in improving myocarditis care.

Assessment of an Artificial Intelligence Tool for Estimating Left Ventricular Ejection Fraction in Echocardiograms from Apical and Parasternal Long-Axis Views

This work aims to evaluate the performance of a new artificial intelligence tool (ExoAI) to compute the left ventricular ejection fraction (LVEF) in echocardiograms of the apical and parasternal long axis (PLAX) views. We retrospectively gathered echocardiograms from 441 individual patients (70% male, age: 67.3 ± 15.3, weight: 87.7 ± 25.4, BMI: 29.5 ± 7.4) and computed the ejection fraction in each echocardiogram using the ExoAI algorithm. We compared its performance against the ejection fraction from the clinical report. ExoAI achieved a root mean squared error of 7.58% in A2C, 7.45% in A4C, and 7.29% in PLAX, and correlations of 0.79, 0.75, and 0.89, respectively. As for the detection of low EF values (EF < 50%), ExoAI achieved an accuracy of 83% in A2C, 80% in A4C, and 91% in PLAX. Our results suggest that ExoAI effectively estimates the LVEF and it is an effective tool for estimating abnormal ejection fraction values (EF < 50%). Importantly, the PLAX view allows for the estimation of the ejection fraction when it is not feasible to acquire apical views (e.g., in ICU settings where it is not possible to move the patient to obtain an apical scan).

An Automated Machine Learning-Based Quantitative Multiparametric Approach for Mitral Regurgitation Severity Grading

BACKGROUND

Considering the high prevalence of mitral regurgitation (MR) and the highly subjective, variable MR severity reporting, an automated tool that could screen patients for clinically significant MR (≥ moderate) would streamline the diagnostic/therapeutic pathways and ultimately improve patient outcomes.

OBJECTIVES

The authors aimed to develop and validate a fully automated machine learning (ML)-based echocardiography workflow for grading MR severity.

METHODS

ML algorithms were trained on echocardiograms from 2 observational cohorts and validated in patients from 2 additional independent studies. Multiparametric echocardiography core laboratory MR assessment served as ground truth. The machine was trained to measure 16 MR-related parameters. Multiple ML models were developed to find the optimal parameters and preferred ML model for MR severity grading.

RESULTS

The preferred ML model used 9 parameters. Image analysis was feasible in 99.3% of cases and took 80 ± 5 seconds per case. The accuracy for grading MR severity (none to severe) was 0.80, and for significant (moderate or severe) vs nonsignificant MR was 0.97 with a sensitivity of 0.96 and specificity of 0.98. The model performed similarly in cases of eccentric and central MR. Patients graded as having severe MR had higher 1-year mortality (adjusted HR: 5.20 [95% CI: 1.24-21.9]; P = 0.025 compared with mild).

CONCLUSIONS

An automated multiparametric ML model for grading MR severity is feasible, fast, highly accurate, and predicts 1-year mortality. Its implementation in clinical practice could improve patient care by facilitating referral to specialized clinics and access to evidence-based therapies while improving quality and efficiency in the echocardiography laboratory.

Artificial Intelligence Assessment of Biological Age From Transthoracic Echocardiography: Discrepancies with Chronologic Age Predict Significant Excess Mortality

BACKGROUND

Age and sex can be estimated using artificial intelligence on the basis of various sources. The aims of this study were to test whether convolutional neural networks could be trained to estimate age and predict sex using standard transthoracic echocardiography and to evaluate the prognostic implications.

METHODS

The algorithm was trained on 76,342 patients, validated in 22,825 patients, and tested in 20,960 patients. It was then externally validated using data from a different hospital (n = 556). Finally, a prospective cohort of handheld point-of-care ultrasound devices (n = 319; ClinicalTrials.gov identifier NCT05455541) was used to confirm the findings. A multivariate Cox regression model was used to investigate the association between age estimation and chronologic age with overall survival.

RESULTS

The mean absolute error in age estimation was 4.9 years, with a Pearson correlation coefficient of 0.922. The probabilistic value of sex had an overall accuracy of 96.1% and an area under the curve of 0.993. External validation and prospective study cohorts yielded consistent results. Finally, survival analysis demonstrated that age prediction ≥5 years vs chronologic age was associated with an independent 34% increased risk for death during follow-up (P < .001).

CONCLUSIONS

Applying artificial intelligence to standard transthoracic echocardiography allows the prediction of sex and the estimation of age. Machine-based estimation is an independent predictor of overall survival and, with further evaluation, can be used for risk stratification and estimation of biological age.

Deep Learning Based Automatic Left Ventricle Segmentation from the Transgastric Short-Axis View on Transesophageal Echocardiography: A Feasibility Study

Segmenting the left ventricle from the transgastric short-axis views (TSVs) on transesophageal echocardiography (TEE) is the cornerstone for cardiovascular assessment during perioperative management. Even for seasoned professionals, the procedure remains time-consuming and experience-dependent. The current study aims to evaluate the feasibility of deep learning for automatic segmentation by assessing the validity of different U-Net algorithms. A large dataset containing 1388 TSV acquisitions was retrospectively collected from 451 patients (32% women, average age 53.42 years) who underwent perioperative TEE between July 2015 and October 2023. With image preprocessing and data augmentation, 3336 images were included in the training set, 138 images in the validation set, and 138 images in the test set. Four deep neural networks (U-Net, Attention U-Net, UNet++, and UNeXt) were employed for left ventricle segmentation and compared in terms of the Jaccard similarity coefficient (JSC) and Dice similarity coefficient (DSC) on the test set, as well as the number of network parameters, training time, and inference time. The Attention U-Net and U-Net++ models performed better in terms of JSC (the highest average JSC: 86.02%) and DSC (the highest average DSC: 92.00%), the UNeXt model had the smallest network parameters (1.47 million), and the U-Net model had the least training time (6428.65 s) and inference time for a single image (101.75 ms). The Attention U-Net model outperformed the other three models in challenging cases, including the impaired boundary of left ventricle and the artifact of the papillary muscle. This pioneering exploration demonstrated the feasibility of deep learning for the segmentation of the left ventricle from TSV on TEE, which will facilitate an accelerated and objective alternative of cardiovascular assessment for perioperative management.

Use of Artificial Intelligence Including Multimodal Systems to Improve the Management of Cardiovascular Disease

The rising prevalence of cardiovascular disease presents an escalating challenge for current health services, which are grappling with increasing demands. Innovative changes are imperative to sustain the delivery of high-quality patient care. Recent technologic advances have resulted in the emergence of artificial intelligence as a viable solution. Advanced algorithms are now capable of performing complex analysis of large volumes of data rapidly and with exceptional accuracy. Multimodality artificial intelligence systems handle a diverse range of data including images, text, video, and audio. Compared with single-modality systems, multimodal artificial intelligence systems appear to hold promise for enhancing overall performance and enabling smoother integration into existing workflows. Such systems can empower physicians with clinical decision support and enhanced efficiency. Owing to the complexity of the field, however, truly multimodal artificial intelligence is still scarce in the management of cardiovascular disease. This article aims to cover current research, emerging trends, and the future utilisation of artificial intelligence in the management of cardiovascular disease, with a focus on multimodality systems.

Designing medical artificial intelligence systems for global use: focus on interoperability, scalability, and accessibility

Advances in artificial intelligence (AI) and machine learning systems promise faster, more efficient, and more personalized care. While many of these models are built on the premise of improving access to the timely screening, diagnosis, and treatment of cardiovascular disease, their validity and accessibility across diverse and international cohorts remain unknown. In this mini-review article, we summarize key obstacles in the effort to design AI systems that will be scalable, accessible, and accurate across distinct geographical and temporal settings. We discuss representativeness, interoperability, quality assurance, and the importance of vendor-agnostic data types that will be available to end-users across the globe. These topics illustrate how the timely integration of these principles into AI development is crucial to maximizing the global benefits of AI in cardiology.

Recommendations for Cardiac Point-of-Care Ultrasound Nomenclature

Point-of-care ultrasound (POCUS) involves the acquisition, interpretation, and immediate clinical integration of ultrasonographic imaging performed by a treating clinician. The current state of cardiac POCUS terminology is heterogeneous and ambiguous, in part because it evolved through siloed specialty practices. In particular, the medical literature and colloquial medical conversation contain a wide variety of terms that equate to cardiac POCUS. While diverse terminology aided in the development and dissemination of cardiac POCUS throughout multiple specialties, it also contributes to confusion and raises patient safety concerns. This statement is the product of a diverse and inclusive Writing Group from multiple specialties, including medical linguistics, that employed an iterative process to contextualize and standardize a nomenclature for cardiac POCUS. We sought to establish a deliberate vocabulary that is sufficiently unrelated to any specialty, ultrasound equipment, or clinical setting to enhance consistency throughout the academic literature and patient care settings. This statement (1) reviews the evolution of cardiac POCUS-related terms; (2) outlines specific recommendations, distinguishing between intrinsic and practical differences in terminology; (3) addresses the implications of these recommendations for current practice; and (4) discusses the implications for novel technologies and future research.

Artificial intelligence-enhanced patient evaluation: bridging art and science

The advent of digital health and artificial intelligence (AI) has promised to revolutionize clinical care, but real-world patient evaluation has yet to witness transformative changes. As history taking and physical examination continue to rely on long-established practices, a growing pipeline of AI-enhanced digital tools may soon augment the traditional clinical encounter into a data-driven process. This article presents an evidence-backed vision of how promising AI applications may enhance traditional practices, streamlining tedious tasks while elevating diverse data sources, including AI-enabled stethoscopes, cameras, and wearable sensors, to platforms for personalized medicine and efficient care delivery. Through the lens of traditional patient evaluation, we illustrate how digital technologies may soon be interwoven into routine clinical workflows, introducing a novel paradigm of longitudinal monitoring. Finally, we provide a skeptic's view on the practical, ethical, and regulatory challenges that limit the uptake of such technologies.

Multicenter validation study for automated left ventricular ejection fraction assessment using a handheld ultrasound with artificial intelligence

We sought to validate the ability of a novel handheld ultrasound device with an artificial intelligence program (AI-POCUS) that automatically assesses left ventricular ejection fraction (LVEF). AI-POCUS was used to prospectively scan 200 patients in two Japanese hospitals. Automatic LVEF by AI-POCUS was compared to the standard biplane disk method using high-end ultrasound machines. After excluding 18 patients due to infeasible images for AI-POCUS, 182 patients (63 ± 15 years old, 21% female) were analyzed. The intraclass correlation coefficient (ICC) between the LVEF by AI-POCUS and the standard methods was good (0.81, p < 0.001) without clinically meaningful systematic bias (mean bias -1.5%, p = 0.008, limits of agreement ± 15.0%). Reduced LVEF < 50% was detected with a sensitivity of 85% (95% confidence interval 76%-91%) and specificity of 81% (71%-89%). Although the correlations between LV volumes by standard-echo and those by AI-POCUS were good (ICC > 0.80), AI-POCUS tended to underestimate LV volumes for larger LV (overall bias 42.1 mL for end-diastolic volume). These trends were mitigated with a newer version of the software tuned using increased data involving larger LVs, showing similar correlations (ICC > 0.85). In this real-world multicenter study, AI-POCUS showed accurate LVEF assessment, but careful attention might be necessary for volume assessment. The newer version, trained with larger and more heterogeneous data, demonstrated improved performance, underscoring the importance of big data accumulation in the field.

Transforming Cardiovascular Care With Artificial Intelligence: From Discovery to Practice: JACC State-of-the-Art Review

Artificial intelligence (AI) has the potential to transform every facet of cardiovascular practice and research. The exponential rise in technology powered by AI is defining new frontiers in cardiovascular care, with innovations that span novel diagnostic modalities, new digital native biomarkers of disease, and high-performing tools evaluating care quality and prognosticating clinical outcomes. These digital innovations promise expanded access to cardiovascular screening and monitoring, especially among those without access to high-quality, specialized care historically. Moreover, AI is propelling biological and clinical discoveries that will make future cardiovascular care more personalized, precise, and effective. The review brings together these diverse AI innovations, highlighting developments in multimodal cardiovascular AI across clinical practice and biomedical discovery, and envisioning this new future backed by contemporary science and emerging discoveries. Finally, we define the critical path and the safeguards essential to realizing this AI-enabled future that helps achieve optimal cardiovascular health and outcomes for all.

Artificial intelligence-guided detection of under-recognized cardiomyopathies on point-of-care cardiac ultrasound: a multi-center study

Background

Point-of-care ultrasonography (POCUS) enables cardiac imaging at the bedside and in communities but is limited by abbreviated protocols and variation in quality. We developed and tested artificial intelligence (AI) models to automate the detection of underdiagnosed cardiomyopathies from cardiac POCUS.

Methods

In a development set of 290,245 transthoracic echocardiographic videos across the Yale-New Haven Health System (YNHHS), we used augmentation approaches and a customized loss function weighted for view quality to derive a POCUS-adapted, multi-label, video-based convolutional neural network (CNN) that discriminates HCM (hypertrophic cardiomyopathy) and ATTR-CM (transthyretin amyloid cardiomyopathy) from controls without known disease. We evaluated the final model across independent, internal and external, retrospective cohorts of individuals who underwent cardiac POCUS across YNHHS and Mount Sinai Health System (MSHS) emergency departments (EDs) (2011-2024) to prioritize key views and validate the diagnostic and prognostic performance of single-view screening protocols.

Findings

We identified 33,127 patients (median age 61 [IQR: 45-75] years, n=17,276 [52·2%] female) at YNHHS and 5,624 (57 [IQR: 39-71] years, n=1,953 [34·7%] female) at MSHS with 78,054 and 13,796 eligible cardiac POCUS videos, respectively. An AI-enabled single-view screening approach successfully discriminated HCM (AUROC of 0·90 [YNHHS] & 0·89 [MSHS]) and ATTR-CM (YNHHS: AUROC of 0·92 [YNHHS] & 0·99 [MSHS]). In YNHHS, 40 (58·0%) HCM and 23 (47·9%) ATTR-CM cases had a positive screen at median of 2·1 [IQR: 0·9-4·5] and 1·9 [IQR: 1·0-3·4] years before clinical diagnosis. Moreover, among 24,448 participants without known cardiomyopathy followed over 2·2 [IQR: 1·1-5·8] years, AI-POCUS probabilities in the highest (vs lowest) quintile for HCM and ATTR-CM conferred a 15% (adj.HR 1·15 [95%CI: 1·02-1·29]) and 39% (adj.HR 1·39 [95%CI: 1·22-1·59]) higher age- and sex-adjusted mortality risk, respectively.

Interpretation

We developed and validated an AI framework that enables scalable, opportunistic screening of treatable cardiomyopathies wherever POCUS is used.

Funding

National Heart, Lung and Blood Institute, Doris Duke Charitable Foundation, BridgeBio.

Deep Learning Phenotyping of Tricuspid Regurgitation for Automated High Throughput Assessment of Transthoracic Echocardiography

Background and Aims

Diagnosis of tricuspid regurgitation (TR) requires careful expert evaluation. This study developed an automated deep learning pipeline for assessing TR from transthoracic echocardiography.

Methods

An automated deep learning workflow was developed using 47,312 studies (2,079,898 videos) from Cedars-Sinai Medical Center (CSMC) between 2011 and 2021. The pipeline was tested on a temporally distinct test set of 2,462 studies (108,138 videos) obtained in 2022 at CSMC and a geographically distinct cohort of 5,549 studies (278,377 videos) from Stanford Healthcare (SHC).

Results

In the CSMC test dataset, the view classifier demonstrated an AUC of 1.000 (0.999 - 1.000) and identified at least one A4C video with colour Doppler across the tricuspid valve in 2,410 of 2,462 studies with a sensitivity of 0.975 (0.968-0.982) and a specificity of 1.000 (1.00-1.000). In the CSMC test cohort, moderate-or-severe TR was detected with an AUC of 0.928 (0.913 - 0.943) and severe TR was detected with an AUC of 0.956 (0.940 - 0.969). In the SHC cohort, the view classifier correctly identified at least one TR colour Doppler video in 5,268 of the 5,549 studies, resulting in an AUC of 0.999 (0.998 - 0.999), a sensitivity of 0.949 (0.944 - 0.955) and specificity of 0.999 (0.999 - 0.999). The AI model detected moderate-or-severe TR with an AUC of 0.951 (0.938 - 0.962) and severe TR with an AUC of 0.980 (0.966 - 0.988).

Conclusions

We developed an automated pipeline to identify clinically significant TR with excellent performance. This approach carries potential for automated TR detection and stratification for surveillance and screening.

Structured Graphical Abstract

Key Question

Can an automated deep learning model assess tricuspid regurgitation severity from echocardiography?

Key Finding

We developed and validated an automated tricuspid regurgitation detection algorithm pipeline across two healthcare systems with high volume echocardiography labs. The algorithm correctly identifies apical-4-chamber view videos with colour Doppler across the tricuspid valve and grades clinically significant TR with strong agreement to expert clinical readers.

Take Home message

A deep learning pipeline could automate TR screening, facilitating reproducible accurate assessment of TR severity, allowing rapid triage or re-review and expand access in low-resource or primary care settings.

Artificial Intelligence for Cardiovascular Care-Part 1: Advances: JACC Review Topic of the Week

Recent artificial intelligence (AI) advancements in cardiovascular care offer potential enhancements in diagnosis, treatment, and outcomes. Innovations to date focus on automating measurements, enhancing image quality, and detecting diseases using novel methods. Applications span wearables, electrocardiograms, echocardiography, angiography, genetics, and more. AI models detect diseases from electrocardiograms at accuracy not previously achieved by technology or human experts, including reduced ejection fraction, valvular heart disease, and other cardiomyopathies. However, AI's unique characteristics necessitate rigorous validation by addressing training methods, real-world efficacy, equity concerns, and long-term reliability. Despite an exponentially growing number of studies in cardiovascular AI, trials showing improvement in outcomes remain lacking. A number are currently underway. Embracing this rapidly evolving technology while setting a high evaluation benchmark will be crucial for cardiology to leverage AI to enhance patient care and the provider experience.

A Multimodal Video-Based AI Biomarker for Aortic Stenosis Development and Progression

Importance

Aortic stenosis (AS) is a major public health challenge with a growing therapeutic landscape, but current biomarkers do not inform personalized screening and follow-up. A video-based artificial intelligence (AI) biomarker (Digital AS Severity index [DASSi]) can detect severe AS using single-view long-axis echocardiography without Doppler characterization.

Objective

To deploy DASSi to patients with no AS or with mild or moderate AS at baseline to identify AS development and progression.

Design, Setting, and Participants

This is a cohort study that examined 2 cohorts of patients without severe AS undergoing echocardiography in the Yale New Haven Health System (YNHHS; 2015-2021) and Cedars-Sinai Medical Center (CSMC; 2018-2019). A novel computational pipeline for the cross-modal translation of DASSi into cardiac magnetic resonance (CMR) imaging was further developed in the UK Biobank. Analyses were performed between August 2023 and February 2024.

Exposure

DASSi (range, 0-1) derived from AI applied to echocardiography and CMR videos.

Main Outcomes and Measures

Annualized change in peak aortic valve velocity (AV-Vmax) and late (>6 months) aortic valve replacement (AVR).

Results

A total of 12 599 participants were included in the echocardiographic study (YNHHS: n = 8798; median [IQR] age, 71 [60-80] years; 4250 [48.3%] women; median [IQR] follow-up, 4.1 [2.4-5.4] years; and CSMC: n = 3801; median [IQR] age, 67 [54-78] years; 1685 [44.3%] women; median [IQR] follow-up, 3.4 [2.8-3.9] years). Higher baseline DASSi was associated with faster progression in AV-Vmax (per 0.1 DASSi increment: YNHHS, 0.033 m/s per year [95% CI, 0.028-0.038] among 5483 participants; CSMC, 0.082 m/s per year [95% CI, 0.053-0.111] among 1292 participants), with values of 0.2 or greater associated with a 4- to 5-fold higher AVR risk than values less than 0.2 (YNHHS: 715 events; adjusted hazard ratio [HR], 4.97 [95% CI, 2.71-5.82]; CSMC: 56 events; adjusted HR, 4.04 [95% CI, 0.92-17.70]), independent of age, sex, race, ethnicity, ejection fraction, and AV-Vmax. This was reproduced across 45 474 participants (median [IQR] age, 65 [59-71] years; 23 559 [51.8%] women; median [IQR] follow-up, 2.5 [1.6-3.9] years) undergoing CMR imaging in the UK Biobank (for participants with DASSi ≥0.2 vs those with DASSi <.02, adjusted HR, 11.38 [95% CI, 2.56-50.57]). Saliency maps and phenome-wide association studies supported associations with cardiac structure and function and traditional cardiovascular risk factors.

Conclusions and Relevance

In this cohort study of patients without severe AS undergoing echocardiography or CMR imaging, a new AI-based video biomarker was independently associated with AS development and progression, enabling opportunistic risk stratification across cardiovascular imaging modalities as well as potential application on handheld devices.

Artificial intelligence: The future for multimodality imaging of right ventricle

The crucial pathophysiological and prognostic roles of the right ventricle in various diseases have been well-established. Nonetheless, conventional cardiovascular imaging modalities are frequently associated with intrinsic limitations when evaluating right ventricular (RV) morphology and function. The integration of artificial intelligence (AI) in multimodality imaging presents a promising avenue to circumvent these obstacles, paving the way for future fully automated imaging paradigms. This review aimed to address the current challenges faced by clinicians and researchers in integrating RV imaging and AI technology, to provide a comprehensive overview of the current applications of AI in RV imaging, and to offer insights into future directions, opportunities, and potential challenges in this rapidly advancing field.

Atrial Septal Defect Detection in Children Based on Ultrasound Video Using Multiple Instances Learning

Thoracic echocardiography (TTE) can provide sufficient cardiac structure information, evaluate hemodynamics and cardiac function, and is an effective method for atrial septal defect (ASD) examination. This paper aims to study a deep learning method based on cardiac ultrasound video to assist in ASD diagnosis. We chose four standard views in pediatric cardiac ultrasound to identify atrial septal defects; the four standard views were as follows: subcostal sagittal view of the atrium septum (subSAS), apical four-chamber view (A4C), the low parasternal four-chamber view (LPS4C), and parasternal short-axis view of large artery (PSAX). We enlist data from 300 children patients as part of a double-blind experiment for five-fold cross-validation to verify the performance of our model. In addition, data from 30 children patients (15 positives and 15 negatives) are collected for clinician testing and compared to our model test results (these 30 samples do not participate in model training). In our model, we present a block random selection, maximal agreement decision, and frame sampling strategy for training and testing respectively, resNet18 and r3D networks are used to extract the frame features and aggregate them to build a rich video-level representation. We validate our model using our private dataset by five cross-validation. For ASD detection, we achieve 89.33 ± 3.13 AUC, 84.95 ± 3.88 accuracy, 85.70 ± 4.91 sensitivity, 81.51 ± 8.15 specificity, and 81.99 ± 5.30 F1 score. The proposed model is a multiple instances learning-based deep learning model for video atrial septal defect detection which effectively improves ASD detection accuracy when compared to the performances of previous networks and clinical doctors.

Artificial intelligence-assisted evaluation of cardiac function by oncology staff in chemotherapy patients

Aims

Left ventricular ejection fraction (LVEF) calculation by echocardiography is pivotal in evaluating cancer patients' cardiac function. Artificial intelligence (AI) can facilitate the acquisition of optimal images and automated LVEF (autoEF) calculation. We sought to evaluate the feasibility and accuracy of LVEF calculation by oncology staff using an AI-enabled handheld ultrasound device (HUD).

Methods and results

We studied 115 patients referred for echocardiographic LVEF estimation. All patients were scanned by a cardiologist using standard echocardiography (SE), and biplane Simpson's LVEF was the reference standard. Hands-on training using the Kosmos HUD was provided to the oncology staff before the study. Each patient was scanned by a cardiologist, a senior oncologist, an oncology resident, and a nurse using the TRIO AI and KOSMOS EF deep learning algorithms to obtain autoEF. The correlation between autoEF and SE-ejection fraction (EF) was excellent for the cardiologist (r = 0.90), the junior oncologist (r = 0.82), and the nurse (r = 0.84), and good for the senior oncologist (r = 0.79). The Bland-Altman analysis showed a small underestimation by autoEF compared with SE-EF. Detection of impaired LVEF < 50% was feasible with a sensitivity of 95% and specificity of 94% for the cardiologist; sensitivity of 86% and specificity of 93% for the senior oncologist; sensitivity of 95% and specificity of 91% for the junior oncologist; and sensitivity of 94% and specificity of 87% for the nurse.

Conclusion

Automated LVEF calculation by oncology staff was feasible using AI-enabled HUD in a selected patient population. Detection of LVEF < 50% was possible with good accuracy. These findings show the potential to expedite the clinical workflow of cancer patients and speed up a referral when necessary.

Vision-language foundation model for echocardiogram interpretation

The development of robust artificial intelligence models for echocardiography has been limited by the availability of annotated clinical data. Here, to address this challenge and improve the performance of cardiac imaging models, we developed EchoCLIP, a vision-language foundation model for echocardiography, that learns the relationship between cardiac ultrasound images and the interpretations of expert cardiologists across a wide range of patients and indications for imaging. After training on 1,032,975 cardiac ultrasound videos and corresponding expert text, EchoCLIP performs well on a diverse range of benchmarks for cardiac image interpretation, despite not having been explicitly trained for individual interpretation tasks. EchoCLIP can assess cardiac function (mean absolute error of 7.1% when predicting left ventricular ejection fraction in an external validation dataset) and identify implanted intracardiac devices (area under the curve (AUC) of 0.84, 0.92 and 0.97 for pacemakers, percutaneous mitral valve repair and artificial aortic valves, respectively). We also developed a long-context variant (EchoCLIP-R) using a custom tokenizer based on common echocardiography concepts. EchoCLIP-R accurately identified unique patients across multiple videos (AUC of 0.86), identified clinical transitions such as heart transplants (AUC of 0.79) and cardiac surgery (AUC 0.77) and enabled robust image-to-text search (mean cross-modal retrieval rank in the top 1% of candidate text reports). These capabilities represent a substantial step toward understanding and applying foundation models in cardiovascular imaging for preliminary interpretation of echocardiographic findings.

Evaluation of Focused Assessment With Sonography for Trauma Completeness of Children in the Clinical Setting

OBJECTIVES

We evaluated the completeness of real-world Focused Assessment with Sonography for Trauma (FAST) in children after blunt abdominal trauma by benchmarking against established expert guidelines.

METHODS

We conducted a retrospective cohort study, analyzing a random sample of FASTs from two urban pediatric emergency departments. Two experts reviewed and labeled all FASTs for completeness using a predefined guideline of 5 anatomic views and 30 landmarks. We compared frequencies of views and landmarks as medians with interquartile ranges.

RESULTS

We analyzed 200 FASTs, consisting of 1636 video clips, performed by 31 clinicians representing 198 children with a median age of 10 years (IQR 5,14). Over half of FASTs (52%) had all 5 views. The right upper quadrant view was most commonly visualized (96.5%), and suprapubic sagittal was least (65%). None of the FASTs included all 30 landmarks, ranging from 0 to 28 and median of 19 (IQR 15,23). The least visualized landmark of the right and left upper quadrants was caudal liver edge (60%) and splenic tip (64%), respectively. In the pericardial view, it was left atrium (45%). In both transverse and sagittal pelvic views, retro-uterine space was least visualized in girls, 21 and 29% respectively.

CONCLUSIONS

In our study, most FAST views and landmarks were visualized. However, the pelvic sagittal view was the least frequently visualized view, and caudal liver edge was the least visualized landmark. Future research should evaluate if variability in visualizing FAST views and landmarks correlates with inconsistencies in diagnostic test performance.

Deep learning supported echocardiogram analysis: A comprehensive review

An echocardiogram is a sophisticated ultrasound imaging technique employed to diagnose heart conditions. The transthoracic echocardiogram, one of the most prevalent types, is instrumental in evaluating significant cardiac diseases. However, interpreting its results heavily relies on the clinician's expertise. In this context, artificial intelligence has emerged as a vital tool for helping clinicians. This study critically analyzes key state-of-the-art research that uses deep learning techniques to automate transthoracic echocardiogram analysis and support clinical judgments. We have systematically organized and categorized articles that proffer solutions for view classification, enhancement of image quality and dataset, segmentation and identification of cardiac structures, detection of cardiac function abnormalities, and quantification of cardiac functions. We compared the performance of various deep learning approaches within each category, identifying the most promising methods. Additionally, we highlight limitations in current research and explore promising avenues for future exploration. These include addressing generalizability issues, incorporating novel AI approaches, and tackling the analysis of rare cardiac diseases.

The future of valvular heart disease assessment and therapy

Valvular heart disease (VHD) is becoming more prevalent in an ageing population, leading to challenges in diagnosis and management. This two-part Series offers a comprehensive review of changing concepts in VHD, covering diagnosis, intervention timing, novel management strategies, and the current state of research. The first paper highlights the remarkable progress made in imaging and transcatheter techniques, effectively addressing the treatment paradox wherein populations at the highest risk of VHD often receive the least treatment. These advances have attracted the attention of clinicians, researchers, engineers, device manufacturers, and investors, leading to the exploration and proposal of treatment approaches grounded in pathophysiology and multidisciplinary strategies for VHD management. This Series paper focuses on innovations involving computational, pharmacological, and bioengineering approaches that are transforming the diagnosis and management of patients with VHD. Artificial intelligence and digital methods are enhancing screening, diagnosis, and planning procedures, and the integration of imaging and clinical data is improving the classification of VHD severity. The emergence of artificial intelligence techniques, including so-called digital twins-eg, computer-generated replicas of the heart-is aiding the development of new strategies for enhanced risk stratification, prognostication, and individualised therapeutic targeting. Various new molecular targets and novel pharmacological strategies are being developed, including multiomics-ie, analytical methods used to integrate complex biological big data to find novel pathways to halt the progression of VHD. In addition, efforts have been undertaken to engineer heart valve tissue and provide a living valve conduit capable of growth and biological integration. Overall, these advances emphasise the importance of early detection, personalised management, and cutting-edge interventions to optimise outcomes amid the evolving landscape of VHD. Although several challenges must be overcome, these breakthroughs represent opportunities to advance patient-centred investigations.

Left Ventricular Segmentation, Warping, and Myocardial Registration for Automated Strain Measurement

The left ventricular global longitudinal strain (LVGLS) is a crucial prognostic indicator. However, inconsistencies in measurements due to the speckle tracking algorithm and manual adjustments have hindered its standardization and democratization. To solve this issue, we proposed a fully automated strain measurement by artificial intelligence-assisted LV segmentation contours. The LV segmentation model was trained from echocardiograms of 368 adults (11,125 frames). We compared the registration-like effects of dynamic time warping (DTW) with speckle tracking on a synthetic echocardiographic dataset in experiment-1. In experiment-2, we enrolled 80 patients to compare the DTW method with commercially available software. In experiment-3, we combined the segmentation model and DTW method to create the artificial intelligence (AI)-DTW method, which was then tested on 40 patients with general LV morphology, 20 with dilated cardiomyopathy (DCMP), and 20 with transthyretin-associated cardiac amyloidosis (ATTR-CA), 20 with severe aortic stenosis (AS), and 20 with severe mitral regurgitation (MR). Experiments-1 and -2 revealed that the DTW method is consistent with dedicated software. In experiment-3, the AI-DTW strain method showed comparable results for general LV morphology (bias - 0.137 ± 0.398%), DCMP (- 0.397 ± 0.607%), ATTR-CA (0.095 ± 0.581%), AS (0.334 ± 0.358%), and MR (0.237 ± 0.490%). Moreover, the strain curves showed a high correlation in their characteristics, with R-squared values of 0.8879-0.9452 for those LV morphology in experiment-3. Measuring LVGLS through dynamic warping of segmentation contour is a feasible method compared to traditional tracking techniques. This approach has the potential to decrease the need for manual demarcation and make LVGLS measurements more efficient and user-friendly for daily practice.

The effect of real-time EF automatic tool on cardiac ultrasound performance among medical students

PURPOSE

Point-of-care ultrasound (POCUS) is a sensitive, safe, and efficient tool used in many clinical settings and is an essential part of medical education in the United States. Numerous studies present improved diagnostic performances and positive clinical outcomes among POCUS users. However, others stress the degree to which the modality is user-dependent, rendering high-quality POCUS training necessary in medical education. In this study, the authors aimed to investigate the potential of an artificial intelligence (AI) based quality indicator tool as a teaching device for cardiac POCUS performance.

METHODS

The authors integrated the quality indicator tool into the pre-clinical cardiac ultrasound course for 4th-year medical students and analyzed their performances. The analysis included 60 students who were assigned to one of two groups as follows: the intervention group using the AI-based quality indicator tool and the control group. Quality indicator users utilized the tool during both the course and the final test. At the end of the course, the authors tested the standard echocardiographic views, and an experienced clinician blindly graded the recorded clips. Results were analyzed and compared between the groups.

RESULTS

The results showed an advantage in quality indictor users' median overall scores (P = 0.002) with a relative risk of 2.3 (95% CI: 1.10, 4.93, P = 0.03) for obtaining correct cardiac views. In addition, quality indicator users also had a statistically significant advantage in the overall image quality in various cardiac views.

CONCLUSIONS

The AI-based quality indicator improved cardiac ultrasound performances among medical students who were trained with it compared to the control group, even in cardiac views in which the indicator was inactive. Performance scores, as well as image quality, were better in the AI-based group. Such tools can potentially enhance ultrasound training, warranting the expansion of the application to more views and prompting further studies on long-term learning effects.

Accuracy and Efficacy of Artificial Intelligence-Derived Automatic Measurements of Transthoracic Echocardiography in Routine Clinical Practice

Background: Transthoracic echocardiography (TTE) is the gold standard modality for evaluating cardiac morphology, function, and hemodynamics in clinical practice. While artificial intelligence (AI) is expected to contribute to improved accuracy and is being applied clinically, its impact on daily clinical practice has not been fully evaluated. Methods: We retrospectively examined 30 consecutive patients who underwent AI-equipped TTE at a single institution. All patients underwent manual and automatic measurements of TTE parameters using the AI-equipped TTE. Measurements were performed by three sonographers with varying experience levels: beginner, intermediate, and expert. Results: A comparison between the manual and automatic measurements assessed by the experts showed extremely high agreement in the left ventricular (LV) filling velocities (E wave: r = 0.998, A wave: r = 0.996; both p < 0.001). The automated measurements of LV end-diastolic and end-systolic diameters were slightly smaller (-2.41 mm and -1.19 mm) than the manual measurements, although without significant differences, and both methods showing high agreement (r = 0.942 and 0.977, both p < 0.001). However, LV wall thickness showed low agreement between the automated and manual measurements (septum: r = 0.670, posterior: r = 0.561; both p < 0.01), with automated measurements tending to be larger. Regarding interobserver variabilities, statistically significant agreement was observed among the measurements of expert, intermediate, and beginner sonographers for all the measurements. In terms of measurement time, automatic measurement significantly reduced measurement time compared to manual measurement (p < 0.001). Conclusions: This preliminary study confirms the accuracy and efficacy of AI-equipped TTE in routine clinical practice. A multicenter study with a larger sample size is warranted.

Harnessing the Power of AI: A Comprehensive Review of Left Ventricular Ejection Fraction Assessment With Echocardiography

The quantification of left ventricular ejection fraction (LVEF) has important clinical utility in the assessment of cardiac function and is vital for the diagnosis of cardiovascular diseases. A transthoracic echocardiogram serves as the most commonly used tool for LVEF assessment for several reasons, including, its noninvasive nature, great safety profile, real-time image processing ability, portability, and cost-effectiveness. However, transthoracic echocardiogram is highly dependent on the clinical skill of the sonographer and interpreting physician. Moreover, even amongst well-trained clinicians, significant interobserver variability exists in the quantification of LVEF. In search of possible solutions, the usage of artificial intelligence (AI) has been increasingly tested in the clinical setting. While AI-derived ejection fraction is in the preliminary stages of development, it has shown promise in its ability to rapidly quantify LVEF, decrease variability, increase accuracy, and utilize higher-order processing capabilities. This review will delineate the latest advancements of AI in evaluating LVEF through echocardiography and explore the challenges and future trajectory of this emerging domain.

Preliminary Evidence of the Use of Generative AI in Health Care Clinical Services: Systematic Narrative Review

BACKGROUND

Generative artificial intelligence tools and applications (GenAI) are being increasingly used in health care. Physicians, specialists, and other providers have started primarily using GenAI as an aid or tool to gather knowledge, provide information, train, or generate suggestive dialogue between physicians and patients or between physicians and patients' families or friends. However, unless the use of GenAI is oriented to be helpful in clinical service encounters that can improve the accuracy of diagnosis, treatment, and patient outcomes, the expected potential will not be achieved. As adoption continues, it is essential to validate the effectiveness of the infusion of GenAI as an intelligent technology in service encounters to understand the gap in actual clinical service use of GenAI.

OBJECTIVE

This study synthesizes preliminary evidence on how GenAI assists, guides, and automates clinical service rendering and encounters in health care The review scope was limited to articles published in peer-reviewed medical journals.

METHODS

We screened and selected 0.38% (161/42,459) of articles published between January 1, 2020, and May 31, 2023, identified from PubMed. We followed the protocols outlined in the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines to select highly relevant studies with at least 1 element on clinical use, evaluation, and validation to provide evidence of GenAI use in clinical services. The articles were classified based on their relevance to clinical service functions or activities using the descriptive and analytical information presented in the articles.

RESULTS

Of 161 articles, 141 (87.6%) reported using GenAI to assist services through knowledge access, collation, and filtering. GenAI was used for disease detection (19/161, 11.8%), diagnosis (14/161, 8.7%), and screening processes (12/161, 7.5%) in the areas of radiology (17/161, 10.6%), cardiology (12/161, 7.5%), gastrointestinal medicine (4/161, 2.5%), and diabetes (6/161, 3.7%). The literature synthesis in this study suggests that GenAI is mainly used for diagnostic processes, improvement of diagnosis accuracy, and screening and diagnostic purposes using knowledge access. Although this solves the problem of knowledge access and may improve diagnostic accuracy, it is oriented toward higher value creation in health care.

CONCLUSIONS

GenAI informs rather than assisting or automating clinical service functions in health care. There is potential in clinical service, but it has yet to be actualized for GenAI. More clinical service-level evidence that GenAI is used to streamline some functions or provides more automated help than only information retrieval is needed. To transform health care as purported, more studies related to GenAI applications must automate and guide human-performed services and keep up with the optimism that forward-thinking health care organizations will take advantage of GenAI.

Deep learning from latent spatiotemporal information of the heart: Identifying advanced bioimaging markers from echocardiograms

A key strength of echocardiography lies in its integration of comprehensive spatiotemporal cardiac imaging data in real-time, to aid frontline or bedside patient risk stratification and management. Nonetheless, its acquisition, processing, and interpretation are known to all be subject to heterogeneity from its reliance on manual and subjective human tracings, which challenges workflow and protocol standardization and final interpretation accuracy. In the era of advanced computational power, utilization of machine learning algorithms for big data analytics in echocardiography promises reduction in cost, cognitive errors, and intra- and inter-observer variability. Novel spatiotemporal deep learning (DL) models allow the integration of temporal arm information based on unlabeled pixel echocardiographic data for convolution of an adaptive semantic spatiotemporal calibration to construct personalized 4D heart meshes, assess global and regional cardiac function, detect early valve pathology, and differentiate uncommon cardiovascular disorders. Meanwhile, data visualization on spatiotemporal DL prediction models helps extract latent temporal imaging features to develop advanced imaging biomarkers in early disease stages and advance our understanding of pathophysiology to support the development of personalized prevention or treatment strategies. Since portable echocardiograms have been increasingly used as point-of-care imaging tools to aid rural care delivery, the application of these new spatiotemporal DL techniques show the potentials in streamlining echocardiographic acquisition, processing, and data analysis to improve workflow standardization and efficiencies, and provide risk stratification and decision supporting tools in real-time, to prompt the building of new imaging diagnostic networks to enhance rural healthcare engagement.

Implementing focused echocardiography and AI-supported analysis in a population-based survey in Lesotho: implications for community-based cardiovascular disease care models

In settings where access to expert echocardiography is limited, focused echocardiography, combined with artificial intelligence (AI)-supported analysis, may improve diagnosis and monitoring of left ventricular hypertrophy (LVH). Sixteen nurses/nurse-assistants without prior experience in echocardiography underwent a 2-day hands-on intensive training to learn how to assess parasternal long axis views (PLAX) using an inexpensive hand-held ultrasound device in Lesotho, Southern Africa. Loops were stored on a cloud-drive, analyzed using deep learning algorithms at the University Hospital Basel, and afterwards confirmed by a board-certified cardiologist. The nurses/nurse-assistants obtained 756 echocardiograms. Of the 754 uploaded image files, 628 (83.3%) were evaluable by deep learning algorithms. Of those, results of 514/628 (81.9%) were confirmed by a cardiologist. Of the 126 not evaluable by the AI algorithm, 46 (36.5%) were manually evaluable. Overall, 660 (87.5%) uploaded files were evaluable and confirmed. Following short-term training of nursing cadres, a high proportion of obtained PLAX was evaluable using AI-supported analysis. This could be a basis for AI- and telemedical support in hard-to-reach areas with minimal resources.

Is 1 Day of Focused Training in Echocardiographic Assessment Using Subxiphoid-Only (EASy) Examination Enough? A Tertiary Hospital Response to the COVID-19 Crisis and the Use of the EASy Examination to Support Unit-Wide Image Acquisition

OBJECTIVES

We assessed the efficacy of 1-day training in echocardiography assessment using subxiphoid-only (EASy) followed by supervised image interpretation and decision-making during patient rounds as a novel approach to scaling up the use of point-of-care ultrasound (POCUS) in critically ill patients.

DESIGN

Retrospective analysis of medical records and EASy examination images.

SETTING

Tertiary care academic hospital.

PATIENTS

A total of 14 adults (> 18 yr old) with COVID-19-associated respiratory failure under the care of Albany Medical Center's surge response team from April 6-17, 2020 who received at least one EASy examination.

INTERVENTIONS

Residents (previously novice sonographers) were trained in EASy examination using 1 day of didactic and hands-on training, followed by independent image acquisition and supervised image interpretation, identification of hemodynamic patterns, and clinical decision-making facilitated by an echocardiography-certified physician during daily rounds.

MEASUREMENTS AND MAIN RESULTS

We recorded the quality of resident-obtained EASy images, scanning time, and frequency with which the supervising physician had to repeat the examination or obtain additional images. A total of 63 EASy examinations were performed; average scanning time was 4.3 minutes. Resident-obtained images were sufficient for clinical decision-making on 55 occasions (87%), in the remaining 8 (13%) the supervising physician obtained further images.

CONCLUSIONS

EASy examination is an efficient, valuable tool under conditions of scarce resources. The educational model of 1-day training followed by supervised image interpretation and decision-making allows rapid expansion of the pool of sonographers and implementation of bedside echocardiography into routine ICU patient management.

Value Creation Through Artificial Intelligence and Cardiovascular Imaging: A Scientific Statement From the American Heart Association

Multiple applications for machine learning and artificial intelligence (AI) in cardiovascular imaging are being proposed and developed. However, the processes involved in implementing AI in cardiovascular imaging are highly diverse, varying by imaging modality, patient subtype, features to be extracted and analyzed, and clinical application. This article establishes a framework that defines value from an organizational perspective, followed by value chain analysis to identify the activities in which AI might produce the greatest incremental value creation. The various perspectives that should be considered are highlighted, including clinicians, imagers, hospitals, patients, and payers. Integrating the perspectives of all health care stakeholders is critical for creating value and ensuring the successful deployment of AI tools in a real-world setting. Different AI tools are summarized, along with the unique aspects of AI applications to various cardiac imaging modalities, including cardiac computed tomography, magnetic resonance imaging, and positron emission tomography. AI is applicable and has the potential to add value to cardiovascular imaging at every step along the patient journey, from selecting the more appropriate test to optimizing image acquisition and analysis, interpreting the results for classification and diagnosis, and predicting the risk for major adverse cardiac events.

Machine Learning and Bias in Medical Imaging: Opportunities and Challenges

Bias in health care has been well documented and results in disparate and worsened outcomes for at-risk groups. Medical imaging plays a critical role in facilitating patient diagnoses but involves multiple sources of bias including factors related to access to imaging modalities, acquisition of images, and assessment (ie, interpretation) of imaging data. Machine learning (ML) applied to diagnostic imaging has demonstrated the potential to improve the quality of imaging-based diagnosis and the precision of measuring imaging-based traits. Algorithms can leverage subtle information not visible to the human eye to detect underdiagnosed conditions or derive new disease phenotypes by linking imaging features with clinical outcomes, all while mitigating cognitive bias in interpretation. Importantly, however, the application of ML to diagnostic imaging has the potential to either reduce or propagate bias. Understanding the potential gain as well as the potential risks requires an understanding of how and what ML models learn. Common risks of propagating bias can arise from unbalanced training, suboptimal architecture design or selection, and uneven application of models. Notwithstanding these risks, ML may yet be applied to improve gain from imaging across all 3A's (access, acquisition, and assessment) for all patients. In this review, we present a framework for understanding the balance of opportunities and challenges for minimizing bias in medical imaging, how ML may improve current approaches to imaging, and what specific design considerations should be made as part of efforts to maximize the quality of health care for all.

Development of artificial intelligence-based slow-motion echocardiography and clinical usefulness for evaluating regional wall motion abnormalities

The diagnostic accuracy of exercise stress echocardiography (ESE) for myocardial ischemia requires improvement, given that it currently depends on the physicians' experience and image quality. To address this issue, we aimed to develop artificial intelligence (AI)-based slow-motion echocardiography using inter-image interpolation. The clinical usefulness of this method was evaluated for detecting regional wall-motion abnormalities (RWMAs). In this study, an AI-based echocardiographic image-interpolation pipeline was developed using optical flow calculation and prediction for in-between images. The accuracy for detecting RWMAs and image readability among 25 patients with RWMA and 25 healthy volunteers was compared between four cardiologists using slow-motion and conventional ESE. Slow-motion echocardiography was successfully developed for arbitrary time-steps (e.g., 0.125×, and 0.5×) using 1,334 videos. The RWMA detection accuracy showed a numerical improvement, but it was not statistically significant (87.5% in slow-motion echocardiography vs. 81.0% in conventional ESE; odds ratio: 1.43 [95% CI: 0.78-2.62], p = 0.25). Interreader agreement analysis (Fleiss's Kappa) for detecting RWMAs among the four cardiologists were 0.66 (95%CI: 0.55-0.77) for slow-motion ESE and 0.53 (95%CI: 0.42-0.65) for conventional ESE. Additionally, subjective evaluations of image readability using a four-point scale showed a significant improvement for slow-motion echocardiography (2.11 ± 0.73 vs. 1.70 ± 0.78, p < 0.001).In conclusion, we successfully developed slow-motion echocardiography using in-between echocardiographic image interpolation. Although the accuracy for detecting RWMAs did not show a significant improvement with this method, we observed enhanced image readability and interreader agreement. This AI-based approach holds promise in supporting physicians' evaluations.

Use of Clinical and Echocardiographic Evaluation to Assess the Risk of Heart Failure

BACKGROUND

Clinical and echocardiographic features predict incident heart failure (HF), but the optimal strategy for combining them is unclear.

OBJECTIVES

This study sought to define an effective means of using echocardiography in HF risk evaluation.

METHODS

The same clinical and echocardiographic evaluation was obtained in 2 groups with HF risk factors: a training group (n = 926, followed to 7 years) and a validation group (n = 355, followed to 10 years). Clinical risk was categorized as low, intermediate, and high using 4-year ARIC (Atherosclerosis Risk In Communities) HF risk score cutpoints of 9% and 33%. A risk stratification algorithm based on clinical risk and echocardiographic markers of stage B HF (SBHF) (abnormal global longitudinal strain [GLS], diastolic dysfunction, or left ventricular hypertrophy) was developed using a classification and regression tree analysis and was validated.

RESULTS

HF developed in 12% of the training group, including 9%, 18%, and 73% of low-, intermediate-, and high-risk patients. HF occurred in 8.6% of stage A HF and 19.4% of SBHF (P < 0.001), but stage A HF with clinical risk of ≥9% had similar outcome to SBHF. Abnormal GLS (HR: 2.92 [95% CI: 1.95-4.37]; P < 0.001) was the strongest independent predictor of HF. Normal GLS and diastolic function reclassified 61% of the intermediate-risk group into the low-risk group (HF incidence: 12%). In the validation group, 11% developed HF over 4.5 years; 4%, 17%, and 39% of low-, intermediate-, and high-risk groups. Similar results were obtained after exclusion of patients with known coronary artery disease. The echocardiographic parameters also provided significant incremental value to the ARIC score in predicting new HF admission (C-statistic: 0.78 [95% CI: 0.71-0.84] vs 0.83 [95% CI: 0.77-0.88]; P = 0.027).

CONCLUSIONS

Clinical risk assessment is adequate to classify low and high HF risk. Echocardiographic evaluation reclassifies 61% of intermediate-risk patients.

Training, Competency, and Interdisciplinary Collaboration in Point-of-Care Ultrasound

Point-of-care ultrasound can provide critical diagnostic information and add to procedural safety. As with any clinical skill, however, it must be applied by an adequately trained provider, with systems to ensure safety. Training can include a mix of hands-on training with traditional didactics, online coursework, and simulation, but each carries its own costs and benefits worth of review. Following training it is essential to think about assessment of competency in point-of-care ultrasound to reflect the combination of cognitive and procedural skills that makes up the practice. Within the frameworks described, expansion can be expected across specialty and professional boundaries.

Multimodality Cardiac Imaging and the Imaging Workforce in the United States: Diversity, Disparities, and Future Directions

Innovations in cardiac imaging have fundamentally advanced the understanding and treatment of cardiovascular disease. These advances in noninvasive cardiac imaging have also expanded the role of the cardiac imager and dramatically increased the demand for imagers who are cross-trained in multiple modalities. However, we hypothesize that there is significant variation in the availability of cardiac imaging expertise and a disparity in the adoption of advanced imaging technologies across the United States. To evaluate this, we have brought together the leaders of cardiovascular imaging societies, imaging trainees, as well as collaborated with national imaging accreditation commissions and imaging certification boards to assess the state of cardiac imaging and the diversity of the imaging workforce in the United States. Aggregate data confirm the presence of critical gaps, such as limited access to imaging and imaging expertise in rural communities, as well as disparities in the imaging workforce, notably among women and underrepresented minorities. Based on these results, we have proposed solutions to promote and maintain a robust and diverse community of cardiac imagers and improve equity and accessibility for cardiac imaging technologies.

Closing the Last Mile Gap in Access to Multimodality Imaging in Rural Settings: Design of the Imaging Core of the Risk Underlying Rural Areas Longitudinal Study

Achieving optimal cardiovascular health in rural populations can be challenging for several reasons including decreased access to care with limited availability of imaging modalities, specialist physicians, and other important health care team members. Therefore, innovative solutions are needed to optimize health care and address cardiovascular health disparities in rural areas. Mobile examination units can bring imaging technology to underserved or remote communities with limited access to health care services. Mobile examination units can be equipped with a wide array of assessment tools and multiple imaging modalities such as computed tomography scanning and echocardiography. The detailed structural assessment of cardiovascular and lung pathology, as well as the detection of extracardiac pathology afforded by computed tomography imaging combined with the functional and hemodynamic assessments acquired by echocardiography, yield deep phenotyping of heart and lung disease for populations historically underrepresented in epidemiological studies. Moreover, by bringing the mobile examination unit to local communities, innovative approaches are now possible including engagement with local professionals to perform these imaging assessments, thereby augmenting local expertise and experience. However, several challenges exist before mobile examination unit-based examinations can be effectively integrated into the rural health care setting including standardizing acquisition protocols, maintaining consistent image quality, and addressing ethical and privacy considerations. Herein, we discuss the potential importance of cardiac multimodality imaging to improve cardiovascular health in rural regions, outline the emerging experience in this field, highlight important current challenges, and offer solutions based on our experience in the RURAL (Risk Underlying Rural Areas Longitudinal) cohort study.

Remote Expert DVT Triaging of Novice-User Compression Sonography with AI-Guidance

BACKGROUND

Compression ultrasonography of the leg is established for triaging proximal lower extremity deep vein thrombosis (DVT). AutoDVT, a machine-learning software, provides a tool for nonspecialists in acquiring compression sequences to be reviewed by an expert for patient triage. The purpose of this study was to test image acquisition and remote triaging in a clinical setting.

METHODS

Patients with a suspected DVT were recruited at 2 centers in Germany and Greece. Enrolled patients underwent an artificial intelligence-guided two-point compression examination by a nonspecialist using a handheld ultrasound device prior to a standard scan. Images collected by the software were uploaded for blind review by 5 qualified physicians. All reviewers rated the quality of all sequences on the American College of Emergency Physicians (ACEP) image quality scale (score 1-5, ≥ 3 defined as adequate imaging quality) and for an ACEP score ≥3, chose "Compressible", "Incompressible", or "Other". Sensitivity and specificity were calculated for adequate quality scans with an assessment as "Compressible" or "Incompressible". We define this group as diagnostic quality. To simulate a triaging clinical algorithm, a post hoc analysis was performed merging the "incomplete", the "low quality", and the "Incompressible" into a high-risk group for proximal DVT.

RESULTS

Seventy-three patients (average age 64.2 years, 44% females) were eligible for inclusion and scanned by 3 nonultrasound-qualified healthcare professionals. Three patients were excluded from further analysis due to incomplete scans. Sixty two of 70 (88.57%) of the completed scans were judged to be of adequate image quality with an average ACEP score of 3.35. Forty seven of 62 adequate AutoDVT scans were assessed as diagnostic quality, of which 8 were interpreted as positive for proximal DVT by the reviewers resulting in a sensitivity of 100% and specificity of 95.12%. When simulating a triaging algorithm, 34/73 (46.58%) of patients would be triaged as high risk and 8 would be confirmed as positive for proximal DVT (6 in the diagnostic and 2 in the low-quality cohort). Of 39/73 patients triaged as low risk, all were negative for proximal DVT in standard duplex; thus, this triaging algorithm could potentially save 53.42% of standard duplex scans.

CONCLUSIONS

Machine learning software was able to aid nonexperts in acquiring valid ultrasound images of venous compressions and allowed remote triaging. This strategy allows faster diagnosis and treatment of high-risk patients and can spare the need for multiple unnecessary duplex scans, the vast majority being negative.