Citations

Abstract Computational analysis of speech enables precise, objective measurements of a behavioral signal closely linked to manic episodes. In a prospective cohort study conducted in Rochester, MN ( ClinicalTrials.gov Identifier: NCT05956340 ) , eighteen English-speaking adults voluntarily hospitalized for a manic episode of Bipolar I Disorder between August 2023 and September 2024 were enrolled. Participants completed semi-structured research interviews during both acute episodes and remission, with manic symptom severity assessed at each session using the Young Mania Rating Scale (YMRS). Across 57 recorded interviews, temporal prosodic features were extracted using automated computational algorithms. After adjusting for individual differences and moderating effects, participant articulation rate (β = 0.24, p = 0.0038), speech rate variability (β = 0.23, p = 0.0041), floor transfer offset (β = -0.31, p = 0.007), and speaking duration (β = 0.54, p < 0.001) were significantly associated with manic symptom severity. These temporal elements classified manic (CV AUC = 0.85) and remission (CV AUC = 0.83) status, highlighting their potential as monitoring and prognostic biomarkers for manic episodes.

Abstract Background Elevated left ventricular filling pressure (LVFP) is associated with an increased risk of heart failure and mortality. Pulmonary artery wedge pressure (PAWP) is considered the gold standard for assessing LVFP. Recently, we validated an artificial intelligence-enabled electrocardiogram (AI-ECG) to assess LV diastolic function (LVDF) and LVFP. Herein, we compare AI-ECG with PAWP in predicting long-term mortality. Methods We included adults who had comprehensive echocardiography with available ECG within 14-days and right-heart-catheterization (RHC) within 30-days of echocardiography. Elevated-LVFP was defined as PAWP&gt;15 mmHg by RHC, grades 2-3 LVDF by echocardiography, or as adjudicated by the AI-ECG algorithm. Prognosis by AI-ECG was compared to that by echocardiography and PAWP. Results Among 4,168 patients (median age 64 years, 47.7% female), 1,841 had normal-LVFP, and 2,327 had elevated-LVFP by RHC. Echocardiography classified 1,502 patients as normal-LVFP, 1,643 as elevated-LVFP, and 1,023 as indeterminate-LVFP. AI-ECG demonstrated an area-under-the-curve of 0.879 (95%-confidence-interval 0.867-0.891) for predicting elevated-LVFP as determined by echocardiography and 0.742 (95%-confidence-interval 0.724-0.759) for predicting elevated-LVFP as determined by RHC. Over a median follow-up of 7.7 years, 2,735 (34.4%) patients died. In multivariable models, elevated-LVFP by any modality was associated with an increased risk of death (AI-ECG: adjusted-hazard-ratio 1.80, 95% confidence-interval 1.55-2.09; echocardiography: adjusted-hazard-ratio 1.74, 95% confidence-interval 1.51-2.02; PAWP: adjusted-hazard-ratio 1.26, 95% confidence-interval 1.11-1.43). AI-ECG-derived LVFP provided incremental prognostic discrimination compared to echocardiography- or RHC-derived LVFP. Conclusions AI-ECG-derived LVFP offers an accessible, affordable, and noninvasive alternative for detecting elevated LVFP, with potential independent prognostic value beyond traditional echocardiographic and invasive assessments.Survival across LVFP concordance

Background: Atrial arrhythmia recurrence after surgical ablation remains challenging to predict and integrating novel biomarkers may improve risk stratification. Objective: To evaluate whether combining preoperative transthoracic echocardiography (TTE) with artificial intelligence-enabled ECG (AI-ECG) scores enhances prediction of atrial fibrillation/flutter (AF/AFl) recurrence post surgical ablation. Methods: We retrospectively analyzed 1,696 patients undergoing Surgical AF/AFl Ablation between 2006 and 2025 with preoperative TTE and post-blanking ECGs. Clinical variables, TTE indices, and AI-ECG scores (AF probability, ECG-estimated age, HFpEF, LV dysfunction, and aortic stenosis scores) were assessed. The primary outcome was time to AF/AFl recurrence. Univariate/multivariable Cox models and a Random Survival Forest (RSF) model (80:20, training: testing) were developed to identify predictors. Results: Of 1,696 patients (mean age 67.3±10.2 years; 61.7% male) undergoing surgical AF/AFl ablation, 949 (56%) experienced AF/AFl recurrence during a median follow-up of 3.14 years. Patients with recurrence had larger left atria (mean LA area 30.4 vs 24.5 cm 2 , p < 0.001), more diastolic dysfunction (mitral inflow E-wave velocity 1.015 m/s vs 0.896 m/s , p < 0.001), and adverse AI-ECG biomarkers. In multivariable analysis, independent predictors of recurrence included a higher ECG-AF score (p < 1 x 10 -300 ), an older AI-ECG age (p = 0.0002), LA area (p = 0.046), body mass index (p = 0.036), and elevated diastolic blood pressure (HR 1.008 per mmHg, 95% CI 1.002 – 1.014; p = 0.010). The final Cox model achieved a C-index of ~0.67 and stratified patients into risk quartiles with 3-year freedom-from-arrhythmia rates of ~85% (lowest-risk) vs ~43% (highest-risk). An RSF model yielded a slightly higher test C-index (≈0.69), suggesting modest improvement with non-linear modeling. Conclusions: Preoperative AI-ECG biomarkers (AF probability, age discordance) and TTE markers of atrial remodeling independently predicted AF/AFl recurrence after surgical AF/AFl ablation and integration of these metrics improved risk stratification.

Background: Disease-causative variants in LMNA -encoded lamin A/C cause the laminopathies, a heterogeneous group of diseases variably resulting in arrhythmogenic/dilated cardiomyopathy (ACM/DCM), lipodystrophy, muscular dystrophy, and progeria. The artificial intelligence (AI)-derived electrocardiographic age gap (AI-EAG), determined by the discrepancy between a patient’s artificial AI-enabled electrocardiogram (ECG) predicted biological age versus their chronological age, is accelerated in laminopathy. Thus, we sought to determine if the AI-EAG serves as a prognostic marker in patients with cardiac laminopathy. Methods: Retrospective analysis of 1,049 genotype-positive patients with genetic ACM/DCM was used to identify those with pathogenic/likely pathogenic (P/LP) variants in LMNA . After the exclusion of those who lacked a 12-lead ECG while in sinus rhythm, a previously trained AI-ECG age algorithm was used to determine the AI-EAG by subtracting the patient’s chronological age from the AI-ECG derived biological age. The AI-EAG was then correlated with a combined outcome of major ventricular arrhythmia [sudden cardiac arrest, sustained ventricular tachycardia, and appropriate implantable cardioverter-defibrillator shocks], heart transplantation, and cardiovascular death. Results: Overall, 147/1,049 (14%) patients with genetically-mediated ACM/DCM had a P/LP variant in LMNA . Of these, 80/147 (54%) LMNA variant-positive patients (52% female, mean chronological age 36 ± 15 years) had ECGs suitable for AI-EAG analysis. Most (52/80; 65%) had an AI-EAG >10 years with a mean AI-EAG of 17 ± 13 years. Of note, the AI-EAG was greater in those with a clinical cardiac phenotype (20 ± 14 vs. 13 ± 10 years; p = 0.011). As a continuous variable, an increased AI-EAG was associated with increased risk of the composite outcome at a median follow-up of 20 months (HR 1.036; 95% CI 1.004–1.068; p = 0.026) and AI-EAG > 10 and > 20 years were both associated with elevated risk (HR 4.272; 95% CI 1.388–13.147; p = 0.011 and HR 3.732; 95% CI 1.505–9.255; p = 0.004, respectively). Conclusion: In cardiac laminopathy, an increased AI-EAG was common and correlated with adverse cardiovascular outcomes. Non-invasive ascertainment of electrocardiographic aging by AI may provide a novel prognostic marker in cardiac laminopathy. Future studies are needed to validate this finding and further define the role of the AI-EAG in the risk-stratification of patients with cardiac laminopathy.

Introduction: Diastolic dysfunction is a hallmark of heart failure with preserved ejection fraction, yet echocardiographic grading is challenging in patients with hypertrophic or amyloid cardiomyopathies due to variable structural and functional remodeling. Consequently, there is a need for additional diagnostic tools to identify patients at increased risk of diastolic dysfunction and enhanced filling pressure. Research Question: What is the accuracy of an artificial intelligence (AI)-enabled electrocardiogram (ECG) model to predict diastolic dysfunction in hypertrophic and amyloid cardiomyopathy? Methods: We included patients with diverse phenotypes of hypertrophic cardiomyopathy (HCM) and cardiac amyloidosis and validated Mayo Clinic's AI-enabled ECG model to grade diastolic function (normal: grade 0/1; abnormal: grade 2/3). Echocardiographic analysis of systolic (LVEF) and diastolic function was performed within 14 days of ECG recording. Diastolic function was graded based on E/e’ (>14), left atrial volume index (>34 mL/m2), pulmonary venous atrial reversal flow duration (≥30 ms), and peak tricuspid regurgitation velocity (>2.8 m/s). Filling pressure was considered increased if ≥3 parameters were abnormal, inconclusive if 2 were abnomal, and normal if ≥3 were normal. Echocardiograms with severe mitral regurgitation were excluded. Results: We identified and matched 840 ECGs from non-obstructive HCM patients (n=125, age 58±19y, 70% male), 145 ECGs from obstructive HCM patients (n=29, age 56±18y, 66% male), 225 ECGs from apical HCM patients (n=33, age 66±14y, 61% male), and 382 ECGs from cardiac amyloidosis patients (n=49, age 70±13y, 57% male). The AI-enabled ECG model predicted diastolic dysfunction with an accuracy of 65% (non-obstructive HCM), 67% (obstructive HCM), 65% (apical HCM), and 75% (cardiac amyloidosis), with sensitivities of 94% and 97% for obstructive HCM and cardiac amyloidosis, respectively, and high positive and negative predictive values for cardiac amyloidosis. Systolic dysfunction (LVEF≤40%) was more common in cardiac amyloidosis (29%) than in all HCM patients (5%). Conclusion: The AI-enabled ECG model demonstrates moderate accuracy but high sensitivity for detecting diastolic dysfunction and increased filling pressure, especially in obstructive HCM and cardiac amyloidosis. The AI-enabled ECG is a useful, widely scalable, and low cost tool to identify patients at increased risk for diastolic dysfunction.

Background: Arrhythmogenic and dilated cardiomyopathies (ACM/DCM) are associated with adverse cardiovascular outcomes, including sudden death. Artificial intelligence (AI)-enabled electrocardiographic aging algorithms estimate biological age from 12-lead tracings and predict cardiovascular outcomes in the general population. However, their utility in genetic ACM/DCM has not been studied. Methods: Retrospective review of 1,045 genotype-positive ACM/DCM patients was used to identify those with a pathogenic/likely pathogenic variant in ClinGen-classified definitive/strong evidence ACM/DCM-susceptibility genes. Following exclusion of patients without a baseline ECG for analysis (e.g. paced rhythms), the AI-ECG age gap (AI-EAG) was calculated by subtracting chronological from AI-ECG age. Genes were grouped into functional categories: desmosomal ( PKP2, DSP, DSG2, DSC2, JUP ), nuclear envelope ( LMNA ), sarcomeric ( TTN, MYH7, TNNT2 ), cytoskeleton/Z-disc ( FLNC, DES, DMD, BAG3 ), and regulatory/ion channel ( PLN, RBM20, SCN5A ). Cardiac MRI data included left ventricular structure, function, and late gadolinium enhancement (LGE) presence and pattern. Mann-Whitney U was used for group comparisons; results are shown as median (IQR). Results: Overall, 824 patients (52.8% male; median age of 40.7 years, IQR 26.5–53.4; and median AI-EAG of 7.49 years, IQR 0.43–15.26) were included. DES variant-positive patients had the highest (19.98 years, IQR 9.66–23.21) and BAG3 variant-positive patients the lowest (–0.95 years, IQR -6.05–7.61) AI-EAGs. Among gene groups, nuclear envelope had the highest (15.12 years, IQR 6.79–25.17) and sarcomeric the lowest (5.75 years, IQR -0.69-14.01) AI-EAGs. Cardiac MRI was available for 538/824 (65%) of which 233/538 (43.3%) had LGE. No significant difference in AI-EAG was observed between LGE-positive and -negative patients (6.99 vs. 7.94 years; p=0.831). However, several LGE patterns were associated with significantly higher AI-EAGs: subepicardial (9.88 vs. 6.84; p=0.004), transmural (13.83 vs. 7.25; p=0.012), apical (12.89 vs. 7.17; p=0.002), and septal (8.91 vs. 7.19; p=0.016). Conclusion: AI-EAG varies by gene and functional group, possibly reflecting gene-specific myocardial remodeling. While certain LGE patterns were linked to increased biological aging, overall LGE status was not. These findings suggest AI-EAG may capture disease activity that is distinct from LGE and could serve as a complementary, non-invasive biomarker in genetic ACM/DCM.

Abstract Aims To assess the performance of an artificial intelligence-enabled electrocardiogram (AI-ECG) algorithm in identifying patients with moderate to severe aortic stenosis (AS) in an Asian cohort from a tertiary care centre. Methods and results We identified a randomly selected patients ≥60 years old who underwent echocardiography and ECG within in 31 days between 2012 and 2021 at the Samsung Medical Center in Korea. Patients with previous cardiac surgery, prosthetic valves, or pacemakers were excluded. The AI-ECG model, originally developed and validated by Mayo Clinic in the USA, was applied without fine-tuning. Performance metrics, including the area under the curve (AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy, were calculated to compare AI-ECG predictions with TTE-confirmed AS status. Among 5425 patients, 1095 had moderate to severe AS, and 4330 age- and sex-matched patients without AS were included as controls. The AI-ECG model achieved an AUC of 0.85 (95% CI: 0.84–0.87) in detecting moderate to severe AS. Sensitivity, specificity, PPV, NPV, and accuracy were 0.83, 0.65, 0.37, 0.94, and 68.29%, respectively. The model's performance was consistent across various age and sex subgroups, with sensitivity increasing in older patients. Conclusion The AI-ECG model developed in the USA demonstrated comparable performance in detecting moderate to severe AS in an Asian cohort compared with its original validation population. These findings highlight the potential utility of AI-ECG as a non-invasive screening tool for AS across diverse patient populations.

Objective To assess the association between educational level and cardiovascular age acceleration metric derived from ECG, and to determine whether this association is mediated by established cardiovascular disease (CVD) risk factors. Design Prospective population-based cohort study (the Tromsø Study). Setting General population of the Tromsø municipality, Norway. Participants The study sample consisted of 4367 participants of the Tromsø Study, who took part in both Tromsø6 (2007–2008) and Tromsø7 (2015–2016), had a 12-lead ECG obtained at Tromsø7 and did not report a history of heart attack, stroke or atrial fibrillation. Primary outcome measures δ-age, a biomarker of cardiovascular ageing, is defined as the difference (in years) between an individual’s ECG-predicted heart age and their chronological age. ECG-predicted heart age was estimated using a previously validated deep neural network. Results Our findings indicate an inverse association between education and δ-age, with a regression coefficient per increment increase in education of −0.24 (95% CI −0.41 to –0.07) in the overall sample, −0.38 (95% CI −0.59 to –0.16) for women and −0.04 (95% CI −0.31 to 0.23) for men. Participants with the highest level of education (university/college for 4 or more years) had the lowest estimated δ-age with a regression coefficient of −0.69 years (95% CI −1.23 to –0.16) compared with the group with primary education for the overall sample, −1.05 years (95% CI −1.73 to –0.37) for women and −0.15 years (95% CI −1.03 to 0.73) for men. CVD risk factors mediated up to 75% of the association between overall education and δ-age, and 80% of the association among those with the highest education level (university/college for 4 or more years). Among women, 50% of the effect of overall education on δ-age was mediated by CVD risk factors, rising to 53% in the category with the highest level of education. However, in the subsample of men, there was no significant association between education and δ-age, and the mediation analysis produced natural direct and indirect effects pointing in opposite directions. Conclusions Cardiovascular ageing is inversely associated with educational level, an effect that appears to be largely mediated through established risk factors.

Abstract Aims To develop a deep learning model that: (i) utilizes transthoracic echocardiography (TTE) clips to detect left ventricular (LV) enlargement without being provided information regarding a patient’s sex and body size; and (ii) can be accurately applied to clips acquired using either standard comprehensive TTE or handheld cardiac ultrasound (HCU). Methods and results Using retrospective TTE data (training: 8722 patients; internal validation: 468 patients), we developed a deep learning model that estimates a patient’s end-diastolic LV volume (indexed to body surface area and normalized across the sexes), and then thresholds this estimate to perform the following classifications: (1) normally sized LV vs. ≥ mild LV enlargement; (2) normal/mildly enlarged LV vs. ≥ moderate LV enlargement. For retrospective datasets, the model showed strong performance in TTE across three geographically distinct locations (Minnesota and Wisconsin: 1082 patients, AUC = 0.925 and 0.953 for classifications 1 and 2, respectively; Arizona: 1475 patients, AUC = 0.935 and 0.969; and Florida: 1481 patients, AUC = 0.934 and 0.970). Additionally, performance was strong for both TTE and HCU clips collected from a prospective cohort of 410 patients who underwent HCU immediately following TTE (TTE: AUC = 0.925 and 0.971; HCU: AUC = 0.874 and 0.902, for classifications 1 and 2, respectively). Conclusion An automated deep learning model applied to TTE or HCU images accurately categorizes LV volumes. These results lay a foundation for future work aimed at optimizing clinical outcomes for heart failure patients by enabling early detection of LV enlargement across various point-of-care settings.

Background: The gap between artificial intelligence-derived age (ECG-Age) and chronological age has been validated as a strong predictor of increased cardiovascular morbidity and mortality, independent of age, and is a potential biomarker of biological aging. Acute stress may cause real-time, transient changes in autonomic function via electrophysiological signals that, when repeated over time, may contribute to biological aging and represent mechanisms linking stress and Major Adverse Cardiovascular Events (MACE). We hypothesized that acute mental stress induced in a laboratory is associated with increased ECG and that this increase is associated with a higher risk of MACE among individuals with stable coronary heart disease (CHD). Methods: We pooled two prospective cohorts of participants with stable CHD from a university-based hospital network in Atlanta. Stress was induced using a standardized mental stress test (public speaking task) and measured 12-lead ECGs during rest, mental stress, and recovery periods. To measure ECG-age, we utilized a published algorithm using a neural network on >500,000 independent patients. The study endpoint was MACE including nonfatal myocardial infarction, hospitalization for heart failure, or cardiovascular death. The difference in ECG-Age between resting and acute mental stress was used to define the stress Age-Gap. Cox proportional hazard regression modeled the association of stress Age-Gap operationalized as tertiles adjusting for chronological age, sex, race, cardiovascular risk factors, and baseline Cohen perceived stress scale. Results: 442 participants in the total sample pool (38% women). Over a 5-year median follow-up, 95 participants (21%) experienced MACE. At rest, the mean chronologic age was 55.7±9.9 years, and ECG-Age was 57.1 ±10.7 years, with rest Age-Gap of 0. 9±9.6 (p=0.16). During stress, the stress Age-Gap of 1.95±9.8 (p<0.0001), while recovery Age-Gap was 1.1±10.3 (p=0.13). Those that demonstrated the largest ECG-Age positive change between rest and stress by ECG (top tertile) had a higher risk for MACE when compared to those with the lowest or negative ECG age change (bottom tertile), see Figure. Conclusion: Among patients with stable CHD, acute mental stress caused changes in biological age as measured by AI-ECG consistent with accelerated biological aging. Such changes were associated with MACE. These results uncover accelerated biological aging as a mechanism linking acute stress to CVD outcomes.