Purpose To evaluate whether the artificial intelligence (AI)-quantified mean thoracic skeletal muscle (TSM) attenuation from coronary artery calcium (CAC) scans predicts incident cardiovascular disease (CVD), with a focus on atrial fibrillation (AF) and heart failure (HF). Materials and Methods Data from the Multi-Ethnic Study of Atherosclerosis, including participants without baseline CVD who underwent CAC scanning, were retrospectively analyzed. Myosteatosis was defined by sex-specific, AI-quantified mean TSM attenuation cutoffs. The Cox proportional hazards model was used to compare total CVD, AF, and HF risks between the bottom and top quartiles of TSM attenuation after adjusting for CVD risk factors, inflammatory markers, insulin resistance, Agatston score, TSM volume, and social determinants of health. Results Among 5739 participants (mean age, 62.1 years ± 10.3 [SD]; 3002 [52.3%] female), 1826 CVD events occurred over 19 years, including 1139 AF and 359 HF events. Myosteatosis was independently associated with increased risks of total CVD (hazard ratio [HR], 1.48 [95% CI: 1.25, 1.75]; P = .001), AF (HR, 1.68 [95% CI: 1.37, 2.07]; P < .001), and HF (HR, 1.61 [95% CI: 1.18, 2.19]; P < .002). Participants with both myosteatosis and high Agatston scores had markedly higher cumulative incidences compared with those with high Agatston scores alone (total CVD: 84.6% vs 68.6%; AF: 52.9% vs 42.4%; HF: 22.6% vs 16.1%). Adding myosteatosis to the Agatston score significantly improved prediction (time-dependent area under the receiver operating characteristic curve, total CVD: 0.74 vs 0.80, P < .001; AF: 0.68 vs 0.76, P < .001; HF: 0.73 vs 0.78, P = .007). Conclusion AI-quantified mean TSM attenuation on CAC scans independently predicted AF and HF and enhanced the Agatston score's predictive value. Keywords: Myosteatosis, Coronary Artery Calcium Scan, Atrial Fibrillation, Heart Failure, Artificial Intelligence, Applications-CT, Cardiac, Thorax, Muscular, Heart ClinicalTrials.gov NCT00005487 Supplemental material is available for this article. © RSNA, 2026.

BACKGROUND:Heart failure (HF) remains a leading cause of hospitalization and mortality in the United States. OBJECTIVES:We aim to quantify the burden of HF, characterize epidemiological trends, assess geographic disparities, and examine variation in underlying causes and risk factors using updated national estimates. METHODS:Data on the prevalence, years lived with disability (YLDs), and attributable causes of HF across the United States from 1990 to 2023 were obtained from the Global Burden of Disease Study 2023. Risk factors were evaluated using a complementary cause-pathway framework and a data-driven machine-learning approach. Bayesian age-period-cohort models were used to project HF burden through 2040. RESULTS:In 2023, the United States had 4.65 million prevalent HF cases and 432 thousand YLDs, an approximately 80% increase in absolute burden since 1990. Age-standardized rates remained largely stable, with an age-standardized prevalence rate of 864.29 (95% uncertainty intervals [UI]: 789.90-937.31) and age-standardized YLD rate of 80.41 (95% UI: 56.41-110.96) per 100,000 in 2023. The etiologic profile of HF was dominated by ischemic, hypertensive, valvular, and cardiomyopathic pathways. Marked geographic disparities persisted, with disproportionate burden concentrated in several Midwestern and Southern states. High systolic blood pressure had the highest population-attributable fraction of 40.7%, followed by dietary risks (27.3%), high BMI (22.5%), high low-density lipoprotein cholesterol (11.5%), and kidney dysfunction (10.9%). Projections suggested broadly stable trends in HF burden through 2040. CONCLUSIONS:HF remains a substantial and persistent public health burden in the United States. These findings support more targeted prevention, improved risk-factor control, and more equitable delivery of HF care.

Background The AI-CVD initiative seeks to extract actionable information from coronary artery calcium (CAC) scans beyond the CAC score. We aimed to develop a heart failure (HF) prediction model, AI-CVD-HF, based on AI-derived features from non-contrast CAC scans, and compare it with PREVENT-HF. Method AI features from CAC scans of 6743 asymptomatic participants in the Multi-Ethnic Study of Atherosclerosis (MESA) and Framingham Heart Study–Offspring (FHS-O) (mean age 62.3 ± 10.1; 47.2% male; median follow-up 17.1 years; 429 HF events) were analyzed. Features were selected using random forest and modeled with random survival forest. Performance was assessed against the base PREVENT-HF using 5-fold cross-validation for area under the receiver-operating-characteristic curve (AUC), area under the precision-recall curve (AUPRC), net-benefit, and calibration. External validation experiment was conducted to assess generalizability. Results Selected features for AI-CVD-HF model included AI-CAC score, left-to-right ventricular volume ratio, left atrial volume, left ventricular mass, visceral fat volume, skeletal muscle mean density, thoracic aortic calcification, age and sex. The 10-year AUC of AI-CVD-HF was 0.83 (95% CI:0.81–0.85), compared with PREVENT-HF (0.79, 95% CI:0.75–0.83; p = 0.01) and also demonstrated 32% higher AUPRC (0.25[95% CI:0.11–0.39] vs 0.19[95% CI:0.07–0.31]; p = 0.049), and consistent performance across age, sex, and race/ethnicity. Calibration metrics favored AI-CVD-HF over PREVENT-HF (Brier score [0.0356 vs 0.0383], slope [1.02 vs 1.20]). External validation showed performance and calibration consistent with internal validation. Conclusions AI-CVD-HF, using AI-derived features from CAC scans, demonstrated more favorable discrimination and calibration than PREVENT-HF for HF prediction, extending the utility of CAC scans beyond coronary artery disease risk assessment.

BACKGROUND AND AIMS:The favorable meal timing for glycemic control in individuals with diabetes remains unclear. The primary aim of this study was to evaluate the prospective association of meal timing with hemoglobin A1C (HbA1c) in individuals with diabetes. The secondary aim was to evaluate associations of meal timing with other cardiovascular disease (CVD) risk factors, including triglycerides, low-density and high-density lipoprotein cholesterol (LDL-C, HDL-C), and systolic and diastolic blood pressure (SBP, DBP). METHODS:This cohort study included participants with diabetes at baseline (2008-2011) and attending visit 2 examination (2014-2017). Baseline dietary intakes were assessed using two 24-h dietary recalls. Linear regression was performed to estimate prospective associations of energy intake (EI) and glycemic load (GL) at each meal timing with HbA1c and other CVD risk factors at visit 2. RESULTS:This study included 1740 participants (mean age: 52.8 years; 59.2% female). Meal timings were associated with HbA1c and CVD risk factors, but in different directions and varied by sex and antidiabetic medication use. For HbA1c, there were generally inverse associations with morning meals, particular early-morning meal (6:00-8:59 AM) in males [EI: percent change, -2.86 (95% CI, -5.48 to -0.17)] and late-morning meal (9:00-11:59 AM) in individuals without antidiabetic medication use [EI: -2.83 (-5.49 to -0.09); GL: -3.31 (-6.46 to -0.05)]. For triglyceride, inverse associations were found for early-morning meals, whereas positive associations were found for afternoon meals (12:00-5:59 PM) in females and individuals without antidiabetic medication use. For LDL-C, positive associations were found for evening meals (6:00-11:59 PM) in individuals without antidiabetic medication use. There were overall positive associations of afternoon meal with SBP and DBP. CONCLUSION:A greater proportion of daily energy intake or glycemic load consumed during morning meals was associated with lower HbA1c levels and more favorable cardiovascular risk factor profiles in individuals with diabetes. Medication use and sex should be considered when discussing potential meal-timing recommendations.

Background The utility of coronary artery calcium (CAC) scoring in individuals with elevated lipoprotein(a) [Lp(a)] for atherosclerotic cardiovascular disease (ASCVD) risk assessment is currently unclear given the propensity of Lp(a) toward noncalcified plaque. Objectives The authors aimed to evaluate the interaction between elevated Lp(a) (>50 mg/dL) and CAC score, and the association of Lp(a) with ASCVD risk across strata of CAC. Methods A pooled cohort of participants without known ASCVD from 4 U.S.-based prospective cohort studies with baseline Lp(a) and CAC measurements was used. The association between elevated Lp(a) across CAC strata and incident ASCVD (myocardial infarction, stroke, coronary revascularization) was evaluated in multivariable Cox regression models. Results The study included 11,319 participants (mean age 56 years, 54% women) with 1,569 incident ASCVD events over 14.8 year mean follow-up. Lp(a) >50 mg/dL (HR: 1.24; 95% CI: 1.09-1.41) and CAC >0 (HR: 2.44; 95% CI: 2.14-2.77) were independently associated with ASCVD risk (P interaction = 0.80). Among individuals with CAC = 0, ASCVD incidence rates were low overall, but higher with Lp(a) >50 mg/dL vs ≤50 mg/dL (4.9 vs 3.8/1,000 person-years, HR: 1.28; 95% CI: 1.01-1.60). Among those with CAC >0, increased risk was again noted with elevated Lp(a) (21.2 vs 18.2/1,000 person-years, HR: 3.03; 95% CI: 2.52-3.64). Similar results were observed when examining further CAC strata with the greatest risk noted with both CAC ≥300 and Lp(a) >50 mg/dL (HR: 6.12; 95% CI: 4.80-7.81). Consistent results were noted by age and sex with greater absolute risk in general among individuals >50 years of age and men. Conclusions Elevated Lp(a) is associated with higher relative risk across CAC strata, including CAC of 0. Among individuals with CAC of 0, absolute event rates remain low even when Lp(a) is elevated. CAC scoring remains a powerful tool for risk assessment among individuals with elevated Lp(a).

Background Chronic inflammation is increasingly recognized as a driver of residual cardiovascular risk, yet its integration into routine practice remains unclear. Methods A 20-item International Lipid Expert Panel (ILEP) online questionnaire, circulated March–October 2024, assessed healthcare professionals' (HCPs) knowledge, perceptions, and practices regarding inflammation in atherosclerotic cardiovascular disease (ASCVD), with subgroup analyses by geography (Europe vs. US) and specialty (cardiology vs. other). Results 453 HCPs from 58 countries responded. Nearly all (96.9%) regarded inflammation as a key mechanism of atherosclerosis, but this was not matched in practice: only about one-third routinely measured high-sensitivity C-reactive protein (hsCRP) and roughly 40% never did. Interpretation was inconsistent: 46.8% considered CRP itself causal, whereas only about one-fifth identified interleukin (IL)-6 as the causal mediator. Agents that lower hsCRP (statins, colchicine, bempedoic acid) were often equated with anti-inflammatory therapies (65.8%), and anti-inflammatory effects were incorrectly attributed to ezetimibe (24.1%) and PCSK9 inhibitors (38.1%). For elevated hsCRP, most respondents (72.4%) favoured intensifying lifestyle and background CVD therapy. In exploratory, unadjusted analyses, US respondents measured hsCRP more often (56.9% vs. 29.5%), were more likely to know recommended cutoffs (74.5% vs. 44.4%), and more frequently endorsed statins, colchicine, and bempedoic acid for elevated hsCRP (94.1% vs. 58.6%). HCPs specializing in cardiology outperformed other respondents on most items, including correct reference intervals and recognition of anti‑inflammatory drugs; they also more often intensified therapy for elevated hsCRP (78.3% vs. 67.5%), whereas other HCPs favored nutraceuticals or monitoring. Conclusions In this self-selected survey, there was a discordance between reported awareness of inflammation and its reported clinical use: routine hsCRP measurement was infrequent and the distinction between biomarkers and causal mediators was inconsistently understood, with exploratory differences by region and specialty. These hypothesis-generating findings support extensive, well-designed education to improve assessment and management of residual inflammatory risk.

BACKGROUND:Echocardiographic findings suggesting hemodynamic compromise can precede clinical deterioration in cardiac tamponade. We aim to validate echocardiographic predictors for tamponade and construct a simple imaging-based risk stratification model to guide prompt management plans. METHODS:We retrospectively studied consecutive patients who presented with at least moderate pericardial effusion. Patients with atrial fibrillation, greater than moderate pulmonary hypertension, or mechanical ventilation were excluded. Patients were identified as having cardiac tamponade if the intrapericardial pressure was ≥7 mmHg. The following echocardiographic parameters, each assigned equal weight, were evaluated to create the E-TAP (Echocardiography in Tamponade Assessment to recommend Pericardiocentesis) score: (1) right atrial late diastolic collapse, (2) right ventricular early diastolic collapse, (3) discordant left and right ventricular measurement variation indicating ventricular interdependence, (4) mitral valve inflow E velocity variation >30%, (5) tricuspid valve inflow E velocity variation >60%, (6) left ventricular outflow velocity variation >20%, (7) inferior vena cava plethora, (8) reversal of diastolic flows of the hepatic vein in expiration, and (9) large pericardial effusion. RESULTS:Of the 254 patients, 134 had tamponade, with a mean E-TAP score of 6.0 ± 1.8, compared to 3.0 ± 1.7 in those without tamponade (P < .001). Adjusted logistic regression analyses showed that inferior vena cave plethora, tricuspid valve inflow variation >60%, and left ventricular outflow tract flow variation >20% were independently associated with tamponade. An E-TAP score ≥5 best identified patients with tamponade with 80% specificity and 81% sensitivity. CONCLUSION:The novel E-TAP score was strongly associated with cardiac tamponade and may facilitate timely decision-making for pericardiocentesis.

Background:Opportunistic cardiac chamber volumetry derived from coronary artery calcium (CAC) scans using the AI-CVD platform predicts heart failure (HF) independent of conventional risk factors. Type 2 diabetes mellitus (T2DM), which classifies individuals as Stage A HF, is associated with chamber enlargement; however, the HF risk associated with chamber enlargement in the absence of T2DM has not been characterized. Methods:We analyzed left atrial (LA) and left ventricular (LV) volumes and mass, indexed to body surface area, using AI-CVD chamber volumetry of 7585 asymptomatic participants in the pooled cohort of Multi-Ethnic Study of Atherosclerosis (MESA) and Framingham Heart Study (FHS) second, and third generation (MESA & FHS; mean age 62.7 ± 14.6 years, 48.6 % male, 10.6 % with T2DM). Chambers were classified as enlarged (≥95th) or normal (<50th percentile). Cox regression and Kaplan-Meier analyses with log-rank tests were performed. Results:Over a median follow-up of 17.1 years, 438 HF events occurred. Individuals without T2DM, with enlarged chambers had HF incidence rates comparable to or higher than individuals with T2DM and normal chambers: LA volume 16.4 vs 8.7 (p = 0.001), LV volume 8.1 vs 8.9 (p = 0.66), and LV mass 10.1 vs 8.9 (p = 0.55) per 1000 person-years. After multivariable adjustment, compared with normal chambers, enlarged LA (HR 2.5[1.9-3.3]), LV (HR 3.5[2.3-5.4]), and LV mass (HR 3.2[2.2-4.9]) remained independently associated with HF in individuals without T2DM. Conclusion:AI-derived cardiac chamber enlargement measured in CAC scans is associated with HF incidence in individuals without T2DM, supporting its potential utility in HF risk stratification.

AIMS:The AI-CVD initiative seeks to extract actionable insights from coronary artery calcium (CAC) scans beyond the traditional CAC score. We previously demonstrated that AI-derived cardiac chamber volumes from CAC scans predict incident heart failure (HF). We aimed to evaluate whether left-to-right cardiac chamber volume ratios outperform chamber volumes in predicting HF. METHODS AND RESULTS:We used AI-CVD cardiac chambers volumetry data from CAC scans of 5732 asymptomatic Multi-Ethnic Study of Atherosclerosis (MESA) participants (age 62.2 ± 10.3 years; 47.7% male). Left-to-right ventricular (LV/RV), atrial (LA/RA), and left atrial-to-right ventricular (LA/RV) volume ratios were evaluated using multivariable Cox models and feature selection techniques. External validation was performed in the Framingham Heart Study Offspring (FHS-O) cohort (N = 1,052, age:58.3 ± 8.3, 42.9% male). During a median follow-up of 17.7 years in MESA, 369 participants (6.3%) developed HF. Elevated ratios (≥75th & ≥95th percentile) of LV/RV, LA/RA, and LA/RV were strongly associated with incident HF: hazard ratio (HR) for ≥95th percentile were 4.04 (95% CI: 2.89-5.65), 2.90 (95% CI: 2.07-4.06), and 2.61 (95% CI: 1.87-3.46), respectively. Among participants with normal LV sizes (interquartile-range), LV/RV ≥95th significantly predicted HF (HR: 2.34; 95% CI: 1.29-4.25). In FHS-O (median follow-up 14.4 years), 56 HF events (5.3%) occurred. LV/RV ≥75th percentile was significantly associated with HF (HR: 2.23; 95% CI: 1.16-4.30), whereas LA/RA was not (HR: 1.22; 95% CI: 0.65-2.29). Feature selection techniques identified LV/RV as the strongest predictor. CONCLUSION:In these two prospective cohorts, AI-derived LV/RV ratio from CAC scans strongly predicted HF. New clinical trials guided by these imaging biomarkers are warranted to establish their clinical utility.

Background: Machine learning (ML) may improve prediction of atrial fibrillation (AF), but its value compared with traditional models such as Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE-AF) in patients with diabetes remains unclear. Methods: Among 9,307 patients in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) with type 2 diabetes and no prior AF, a random forest (RF) classifier using clinical and metabolic variables was compared with a CHARGE-AF Cox model. Discrimination was assessed by five-fold cross-validated area under receiver operating curve (AUC). Results: Over 6.26 years, 175 patients developed AF. The RF model (AUC = 0.731) performed comparably to CHARGE-AF (AUC = 0.756; p = 0.18). Age, waist circumference, race, total cholesterol, and estimated glomerular filtration rate were the top predictors. Conclusion: ML matched CHARGE-AF performance and revealed distinct predictors supporting personalized AF risk prevention.

Abstract Background The degree to which background genetic risk adds to self-report of family history of coronary heart disease (CHD) remains an important question. Methods We utilized data from the multiethnic Genetic Epidemiology Resource in Adult Health and Aging (GERA) cohort of 60,070 Kaiser Permanente of Northern California (KPNC) members (mean ± SD age=59 ± 9 years; 67% female; 18% non-white). We characterized the cohort at baseline in 2003-2007. Excluding those with missing self-report of family history of CHD (n=1,718) resulted in a final sample of 61,352. We stratified the cohort into not having (n=42,866; 70%) and having (n=18,486; 30%) self-reported family history of CHD and then (within stratum of family history of CHD) by three groups of CHD polygenic risk (low: quintile 1; intermediate: quintiles 2, 3 and 4 combined; and high: quintile 5) using a validated 12-SNP polygenic risk score (PRS) for CHD (CARDIO inCode-Score®). Incident CHD was based on ICD-9/10 codes for angina pectoris, myocardial infarction, revascularization procedures or CHD death (n=3,040) through 12/31/2022; mean follow-up was 14.5 years. Results Age-adjusted CHD incidence rates per 10,000 person years were estimated using Poisson regression (accounting for death and health plan disenrollment) by absence/presence of family history of CHD and polygenic risk groups. As shown in the Figure, polygenic risk provided additional risk stratification within categories of family history. The hazard ratio of high PRS (vs low PRS) adjusted for age, sex, 10 principal components of ancestry, smoking, body mass index, diabetes, hypertension, total cholesterol/HDL ratio and cholesterol lowering therapy was 1.62 (95% CI, 1.39-1.87; p<0.001) in those with no family history of CHD, and 1.66 (95% CI, 1.37-2.00; p<0.001) in those with family history of CHD (p-interaction PRS continuous*family history of CHD=0.93). Conclusions Our results show that polygenic risk predicts future CHD regardless of the presence of family history of CHD. Thus, it is not enough to rely of family history to fully characterize potential genetic burden. These data are therefore important for the future of precision medicine.