Citations

Background Bicuspid aortic valve (BAV) is the most common congenital heart defect, associated with a lifetime morbidity burden exceeding 80% and an estimated 7% probability of transmission to first-degree relatives. Hence, echocardiographic screening of relatives is recommended. The aim of our study is to develop an artificial intelligence-enabled electrocardiogram (AI-ECG) model for the detection of BAV. Methods Between January 1, 1990, and June 30, 2023, the Mayo Clinic healthcare system database was searched for patients aged ≥18 years with echocardiograms performed for family history of BAV, abnormal auscultation, a BAV-related diagnosis, or to rule out the condition, and with an electrocardiogram (ECG) obtained within 6 months. Cases were patients with a confirmed BAV diagnosis, and controls were those with normal aortic valves. ECGs were randomly assigned into training (80%), internal validation (10%), and test (10%) groups. The AI-ECG model was built using Keras under TensorFlow. Results A total of 13,065 valid ECG–echocardiogram pairs were included (mean [SD] age, 58 [17] years; 44% women), with a median time of 1 day (interquartile range, 0–3 days) between studies. With a 45% BAV prevalence in the test group ( n = 1301), the area under the curve was 0.704, positive predictive value (PPV) 60.1%, and negative predictive value (NPV) 69.4%. At a prevalence of 7%, the NPV increased to 96.1%, but the PPV decreased to 12.2%. Conclusions In this retrospective cohort and using the current AI-ECG model, BAV detection showed modest diagnostic performance and limited standalone clinical utility, suggesting that ECG-only screening for BAV may require additional modalities or more refined phenotyping.

Importance Timely identification of aortic stenosis (AS) is essential for appropriate clinical management, yet screening remains limited by dependence on comprehensive echocardiography and trained imaging personnel. Objective To develop and validate a deep learning algorithm for detection of moderate or greater AS and prospectively evaluate its performance using artificial intelligence (AI)–guided focused cardiac ultrasound (FoCUS) acquired by novice operators. Design, Setting, and Participants This diagnostic study included retrospective algorithm development and validation and prospective evaluation of AI-guided FoCUS across Mayo Clinic sites in the Midwest, Arizona, and Florida. The model was developed using 6753 patients and evaluated in internal validation (n = 852), internal test (n = 844), and validation (n = 1912) cohorts. Performance was assessed on FoCUS acquired by experienced sonographers (n = 602) and prospectively by novice operators (n = 1302). The retrospective model development and validation cohorts comprised studies performed from January 2005 through September 2022. Prospective study was conducted in 2 enrollment periods from June to August 2024 and from June to September 2025. Participants from both periods were combined to comprise the final prospective cohort. Exposure AI-guided FoCUS acquisition and automated deep learning–based assessment for detection of moderate or greater AS. Main Outcomes and Measures Detection of moderate or greater AS. Performance was assessed using area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and positive predictive value. Results The model demonstrated excellent discrimination in the internal test cohort (AUROC, 0.99; 95% CI, 0.98-1.00) and geographically distinct validation cohorts in Arizona (AUROC, 0.99; 95% CI, 0.97-1.00) and Florida (AUROC, 0.99; 95% CI, 0.96-1.00). Among FoCUS examinations acquired by experienced sonographers, sensitivity was 95% (95% CI, 82-99) and specificity was 97% (95% CI, 95-98). In the prospective novice-operator cohort, 1258 of 1302 examinations (96.6%) were suitable for automated analysis. Sensitivity was 93% (95% CI, 82-99) and specificity was 96% (95% CI, 95-97). Expert review of AI-positive and uninterpretable examinations increased the positive predictive value from 49.4% to 91.1%, with sensitivity of 85.4%. Conclusions and Relevance A deep learning algorithm accurately detected moderate or greater AS across validation cohorts in this study. In prospective evaluation, novice operators were able to acquire AI-guided FoCUS examinations that enabled accurate detection of moderate or greater AS. These findings support a scalable strategy that may expand access to AS detection in settings with limited echocardiography resources.

Background Radionuclide SPECT/computed tomography (CT) is standard for diagnosing transthyretin cardiac amyloidosis (ATTR-CM), but the prognostic value of quantitative metrics remains incompletely defined. We investigated their prognostic utility and additive value beyond existing staging systems. Methods Retrospective study of patients undergoing technetium-99m-pyrophosphate (PYP) SPECT/CT, including consecutive clinically referred patients (October 2023 to 2024) and trial participants with positive scans from January 2022. Myocardial PYP was quantified using standardized uptake values (SUV mean and SUV max ) and volumetric measures [percent injected dose (%ID), cardiac amyloid activity (CAA)]. Major adverse cardiovascular events (MACEs) comprised heart failure hospitalization or death. Cox and Kaplan–Meier analyses assessed associations; likelihood ratio testing and change in area under the curve (AUC) evaluated the additive prognostic value to existing staging systems. Results Among 43 patients with ATTR-CM, 16 experienced MACE over a median 518 days. AUCs ranged from 0.74 to 0.88 at 3 months and 0.56–0.62 at 12 months. All variables trended toward association with MACE: SUV mean ( P = 0.02), SUV max ( P = 0.03), %ID ( P = 0.06), and CAA ( P = 0.06). After age adjustment, cut-points for SUV max (5.8) and SUV mean (3.5) remained associated with MACE. Adding imaging variables to the National Amyloidosis Centre and Columbia staging systems did not significantly improve prognostic accuracy (Δ χ 2 : 0.40–2.48, P = 0.11–0.53). Interobserver reproducibility was high (intraclass correlation coefficients = 0.89–0.99). Conclusion Quantitative Tc-99m–PYP SPECT/CT metrics were associated with short-term outcomes in ATTR-CM, but did not improve established staging systems. Findings are hypothesis-generating and require validation in larger cohorts.

Abstract Aims Heart failure (HF), a major global health challenge, affects millions worldwide and poses substantial healthcare and economic burdens. It is estimated that a large proportion of those with early systolic dysfunction remain asymptomatic at a stage when guideline-directed medical therapies have been shown to prevent disease progression. To develop an artificial intelligence (AI) model capable of predicting abnormal left ventricular ejection fraction (EF) directly from static, non-gated, non-contrast chest computed tomography (CT) scans as a form of opportunistic screening, Methods and results Using a multi-institutional dataset of 34 058 paired non- contrast CT images and echocardiogram reports from two academic centres, we trained our model of classification for predicting left-ventricle ejection fraction (LVEF) categories: abnormal EF (EF < 50%) vs. normal on 25 948 studies. We validated the model on 8110 paired chest CT and echocardiogram results from a separate institution. The model achieved an area under the receiver operating characteristic (AUROC) curve of 0.786 on the hold-out test set and 0.762 on external validation to detect an abnormal EF (<50%). Beyond strong predictive performance, the AI model surpassed expert radiologists in both accuracy and efficiency and provided interpretable visualizations highlighting imaging features linked to reduced LVEF. Conclusion In this study, we developed and validated an AI model capable of predicting abnormal LVEF directly from static, non-gated, non-contrast chest CT scans, a novel application for an imaging modality typically used for unrelated indications as a form of opportunistic screening. This technology holds significant promise for early detection of systolic HF, reducing the diagnostic gap, and improving outcomes in asymptomatic HF patients.

Abstract Aims Heart failure is prevalent; however, there is no cost-effective screening option. Against formal echocardiography, we assessed the diagnostic performance of focused cardiac ultrasound (FoCUS), performed by novice users guided and interpreted by artificial intelligence (AI) for the assessment of left ventricular ejection fraction (LVEF). Methods and results We prospectively enrolled 496 adults referred for diagnostic echocardiography. Novice operators (without clinical or imaging experience) underwent a 4-h imaging workshop, then performed FoCUS using a point-of-care device (Philips Lumify) utilizing real-time AI-image guidance (UltraSight). Images were assessed by AI-image interpretation software (Mayo Clinic), and two expert blinded echocardiologists. Median age was 67, 39% female, and body mass index range 17–56 kg/m2. Forty-one subjects (8.3%) exhibited moderate or greater LV dilatation and 28 (5.6%) had LVEF <40%. The median scan time was 4 min (IQR 3–5). Adequate views were achieved in 95.0% and 97.4% of subjects for AI interpretation and expert analysis, respectively. A two-step diagnostic screening process for low EF in which only abnormal AI reads or uninterpretable scans were reviewed by experts (15.1%) had a sensitivity 96.2%, specificity 95.4%, PPV 54.3%, and NPV 99.8%. A subsequent prospective validation study of 344 subjects demonstrated similar results (sensitivity 100%, specificity 99.4%). Conclusion In this early feasibility investigation, AI-guidance technology embedded on a handheld device, enabled novice users without clinical or imaging experience to acquire sufficient quality images to accurately assess LVEF with minimal training. The technology evaluated in this study may hold promise for AI-guidance and interpretation to facilitate low-cost screening for cardiac dysfunction.

Abstract Echocardiography is the most widely used cardiac imaging modality, yet artificial intelligence-enabled interpretation remains limited by the inability of existing models to integrate visual assessment, quantitative measurement, and clinical reasoning within a unified framework. Here we present EchoAtlas, the first autoregressive vision-language model developed for echocardiographic interpretation. Trained on over 12.9 million question-answer pairs derived from approximately 2 million echocardiogram videos, EchoAtlas achieves 0.966 accuracy on multiple-choice questions in our internal test set and establishes a new state-of-the-art on the public MIMIC-EchoQA benchmark (0.699 vs. 0.508 previously). EchoAtlas also provides accurate quantitative measurements, segment-level regional wall motion assessment, longitudinal comparison, and diagnostic reasoning across diverse question formats — capabilities not previously demonstrated in this domain. These results highlight the potential of autoregressive vision-language models as a foundation for interactive echocardiographic interpretation, representing an early step toward scalable, auditable artificial intelligence systems in cardiology practice.

Abstract Valvular heart disease (VHD) remains significantly underdiagnosed and undertreated. This review examines an artificial intelligence (AI)-enhanced ‘spoke–hub–node’ care model designed to improve the early detection, risk stratification, and treatment of VHD. In this model, AI tools—such as automated ECG interpretation, digital stethoscopes, and point-of-care ultrasound—facilitate decentralized screening and referral for cardiac imaging at the community level. During the transition from outpatient settings to tertiary care centres, AI-integrated echocardiography, cardiac tomography, and magnetic resonance imaging facilitate advanced diagnostic evaluation and inform procedural planning. We review emerging innovations that can enhance this model of care delivery—including unsupervised machine learning to uncover novel VHD phenotypes, generative AI for automated reporting, the use of digital twins to simulate interventions, and the integration of multiple AI agents to support heart team meetings. These advances are followed by the emerging use of AI in robotic transoesophageal and intracardiac echocardiography, as well as in fusion fluoroscopy imaging, to guide valve interventions. While outlining the challenges inherent in this rapidly evolving field, the review’s central contribution is its vision to connect the continuum—from AI-enabled community screening to personalized, image-guided therapies at tertiary care centres—offering a scalable and equitable model for VHD care.

Importance Transthyretin amyloid cardiomyopathy (ATTR-CM) is underdiagnosed despite expanding treatment options. Objective To develop and evaluate an artificial intelligence (AI)–augmented clinical program for ATTR-CM detection. Design, Setting, and Participants This nonrandomized clinical trial involved constructing an AI model, ATTRACTnet, using electrocardiogram waveforms, echocardiographic measurements, demographics, and diagnosis codes for orthopedic manifestations of amyloidosis. A single-system, multisite, single-arm, open-label trial was conducted to evaluate its real-world performance. The model was trained and validated at a large academic referral site for ATTR-CM with external validation at an academic site. The trial was conducted as a single-system, multisite trial. Patients with left ventricular (LV) wall thickness 12 mm or more and an ATTRACTnet score 0.5 or higher were eligible. Exclusions included prior ATTR-CM testing, hypertrophic cardiomyopathy, expected life span less than 1 year, nursing home residence, advanced dementia, or LV wall thickness less than 14 mm explained by uncontrolled hypertension or moderate/severe aortic stenosis. Intervention Eligible patients were notified and offered nuclear scintigraphy testing, monoclonal protein testing, and follow-up care on agreement from the treating physician. Main Outcomes and Measure The primary outcome was a diagnosis of ATTR-CM by consensus criteria. ATTR-CM testing positivity was compared with historical and contemporary controls. Results ATTRACTnet was developed in an internal test set of 799 patients (mean [SD] age, 75.1 [11.1] years; 516 [64.7%] male and 283 [35.3%] female) using 5-fold cross-validation with an additional external test set of 422 patients. It had good discrimination for ATTR-CM detection with an area under the receiver operator characteristic curve of 0.85 (5-fold cross-validation, 0.77-0.85) in the internal set and 0.82 (95% CI, .81-0.83) in the external test set with similar performance in Hispanic, non-Hispanic Black, and non-Hispanic White patients. A total of 1471 patients were identified with positive AI model scores 0.5 or more during the study period, with 256 eligible patients who met study criteria. Of these patients, 50 underwent amyloidosis testing after physician and patient approval; 24 (48%) were diagnosed with ATTR-CM, and 21 (88%) initiated treatment within 3 months. The positivity rate was more than 2.8 times higher than historical controls (15.3%; 95% CI, 13.1%-17.9%; P < .001), with an 18% relative increase in new diagnoses vs the prior year. Conclusions and Relevance AI-augmented screening may improve ATTR-CM detection and identify patients who are missed by usual care. Prospective randomized trials are needed to determine if outcomes are improved. Trial Registration ClinicalTrials.gov Identifier: NCT06469372

Abstract Background and Aims Classification and risk stratification in aortic (AR), mitral (MR), and tricuspid regurgitation (TR) remains a significant clinical challenge. This study aimed to develop an artificial intelligence (AI) system to assess valvular regurgitation and stratify MR-progression risk. Methods Using transthoracic echocardiograms (TTEs) at two sites (internal development/test, external test), the DELINEATE-Regurgitation system was developed to classify AR, MR, and TR severity using colour Doppler videos. Methods of summating video-level classifications into study-level predictions were tested, comparing single-view with multiview approaches integrating predictions across multiple videos. Model agreement with cardiologists was assessed by weighted kappa. A separate AI system (DELINEATE-MR-Progression) analysing colour Doppler videos was developed to predict which patients with mild, mild–moderate, and moderate MR were most likely to progress to moderate–severe or severe MR with analysis by Kaplan–Meier and Cox proportional hazards models. Results A total of 71 660 TTEs with 1 203 980 colour Doppler videos were included. The weighted kappa in internal/external test sets for regurgitation classification was 0.81/0.76 for AR, 0.76/0.72 for MR, and 0.73/0.64 for TR using a multiview approach taking all colour Doppler videos in a study, demonstrating substantial agreement with cardiologist interpretation with superiority of multiview over single view approaches. In the progression analysis, the AI score stratified MR-progression risk even when controlled for clinical factors known to be associated with MR progression [internal test set hazard ratio 4.1 (95% confidence interval 2.5–6.6)]. Conclusions An AI system can accurately classify AR, MR, and TR and predict MR progression beyond currently known risk factors.