AI-Enabled Echocardiography Identifies Prominent Epicardial Adipose Tissue Associated With an Adverse Cardiometabolic Profile
- 1Yale School of Medicine
Abstract
BackgroundExcess epicardial adipose tissue (EAT) is associated with cardiovascular-kidney-metabolic (CKM) dysfunction, but its assessment has traditionally required advanced imaging.
ObjectivesWe investigated whether AI-enabled echocardiography could characterize epicardial adiposity and identify residual cardiometabolic risk beyond body mass index (BMI).
MethodsWe developed PanAdipo, a video-based deep learning model in 1,114,441 videos from 28,797 studies across the Yale-New Haven Health System (YNHHS; 2016-2022), using report-derived expert reader labels of prominent EAT (2.3% of studies). PanAdipo was evaluated in five cohorts: a temporally distinct YNHHS transthoracic echocardiography (TTE) set (n=4,588 patients contributing 5,130 studies), an emergency department point-of-care ultrasound (POCUS) cohort (n=10,957), the geographically distinct MIMIC-IV dataset (n=4,549), the community-based Multi-Ethnic Study of Atherosclerosis (MESA Exam 6, n=2,740), and in held-out YNHHS patients with paired coronary CT angiography within 12 months (n=5,167). Analyses examined discrimination of the report label and of CT-defined high indexed EAT volume, correlation with paired cardiac CT phenotypes, and BMI-adjusted associations with cardiometabolic biomarkers and newly documented metabolic disease.
ResultsIn the held-out health system test set, PanAdipo identified prominent EAT documented in clinical reports with an area under the receiver operating characteristic curve (AUROC) of 0.86 (95% CI, 0.82-0.91) across all views and of up to 0.93 when combining apical 4-chamber and parasternal views. Against paired CT-derived measurements, PanAdipo discriminated age- and sex-specific high EAT (EAT/heart volume [≥]95th percentile) with AUROCs of 0.91 to 0.95 across age-sex groups and correlated more strongly with indexed EAT volume ({rho}=0.76) than with BMI ({rho}=0.41, all P<0.001); PanAdipo alone explained 53% of indexed EAT volume variance, versus 23% for age, sex, and BMI combined. In MESA, higher scores were associated with an adverse metabolic profile (higher HOMA-IR [{beta}=0.17 per SD; 95% CI, 0.11-0.22]) after adjustment for age, sex, and BMI. In health system TTE and POCUS cohorts, higher scores were associated with newly documented MASLD/MASH (BMI-adjusted HRpooled 1.31 [1.03-1.67] per SD of log[PanAdipo]) and hyperlipidemia (1.11 [1.04-1.18]).
ConclusionsAI-enabled echocardiography provides a scalable digital biomarker that characterizes prominent epicardial adiposity and is associated with an adverse cardiometabolic profile, highlighting a role for echocardiography in metabolic risk stratification.
CONDENSED ABSTRACTEpicardial adipose tissue (EAT) is linked to cardiometabolic risk, but its assessment has traditionally required advanced imaging. We developed PanAdipo, a deep learning model that derives a digital biomarker of EAT from routine echocardiography. PanAdipo tracks CT-defined epicardial adiposity ({rho}=0.76; AUROC 0.91-0.95 for age-/sex-specific high indexed-EAT), yet correlates modestly with body mass index (BMI; {rho}=0.19-0.41). After BMI adjustment, higher scores were associated with insulin resistance in a community-based cohort and newly documented metabolic disease across health systems. AI-enabled echocardiography may provide a scalable window into epicardial adiposity, with potential applications for enriching trial populations and monitoring response to cardiometabolic therapies.
Related
- Cardiometabolic Substrate Modification in Atrial Fibrillation: Obesity and Epicardial Adipose Tissue in the Era of Incretin-Based Therapies and Catheter Ablation
- Prominent R Wave in aVR Associated with Drug Poisoning
- Cryptococcal meningitis-associated cerebrovascular injury presenting with sequential multifocal cerebral infarctions and prominent autonomic dysfunction in a non-HIV patient: a case report and focused literature review.
- Echocardiographic Left Ventricular Hypertrophy Among Patients with Type 2 Diabetes Undergoing Clinically Indicated Echocardiography at a Saudi Cardiac Center: Prevalence and Associated Factors
- Waist-to-Height Ratio Identifies Cardiometabolic Risk Not Captured by Body Mass Index: The All of Us Research Program
BibTeX
@misc{aminorroaya2026aienabled,
title = {{AI-Enabled Echocardiography Identifies Prominent Epicardial Adipose Tissue Associated With an Adverse Cardiometabolic Profile}},
author = {A. Aminorroaya and A. Coppi and R. L. McNamara and J. A. C. Lima and H. M. Krumholz and C. Antoniades and R. Khera and E. K. Oikonomou},
howpublished = {medRxiv},
year = {2026},
month = jul,
doi = {10.64898/2026.07.22.26358690},
url = {https://www.medrxiv.org/content/10.64898/2026.07.22.26358690v2},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - AI-Enabled Echocardiography Identifies Prominent Epicardial Adipose Tissue Associated With an Adverse Cardiometabolic Profile
AU - A. Aminorroaya
AU - A. Coppi
AU - R. L. McNamara
AU - J. A. C. Lima
AU - H. M. Krumholz
AU - C. Antoniades
AU - R. Khera
AU - E. K. Oikonomou
PY - 2026
DA - 2026/07/22
PB - medRxiv
DO - 10.64898/2026.07.22.26358690
UR - https://www.medrxiv.org/content/10.64898/2026.07.22.26358690v2
AB - BackgroundExcess epicardial adipose tissue (EAT) is associated with cardiovascular-kidney-metabolic (CKM) dysfunction, but its assessment has traditionally required advanced imaging. ObjectivesWe investigated whether AI-enabled echocardiography could characterize epicardial adiposity and identify residual cardiometabolic risk beyond body mass index (BMI). MethodsWe developed PanAdipo, a video-based deep learning model in 1,114,441 videos from 28,797 studies across the Yale-New Haven Health System (YNHHS; 2016-2022), using report-derived expert reader labels of prominent EAT (2.3% of studies). PanAdipo was evaluated in five cohorts: a temporally distinct YNHHS transthoracic echocardiography (TTE) set (n=4,588 patients contributing 5,130 studies), an emergency department point-of-care ultrasound (POCUS) cohort (n=10,957), the geographically distinct MIMIC-IV dataset (n=4,549), the community-based Multi-Ethnic Study of Atherosclerosis (MESA Exam 6, n=2,740), and in held-out YNHHS patients with paired coronary CT angiography within 12 months (n=5,167). Analyses examined discrimination of the report label and of CT-defined high indexed EAT volume, correlation with paired cardiac CT phenotypes, and BMI-adjusted associations with cardiometabolic biomarkers and newly documented metabolic disease. ResultsIn the held-out health system test set, PanAdipo identified prominent EAT documented in clinical reports with an area under the receiver operating characteristic curve (AUROC) of 0.86 (95% CI, 0.82-0.91) across all views and of up to 0.93 when combining apical 4-chamber and parasternal views. Against paired CT-derived measurements, PanAdipo discriminated age- and sex-specific high EAT (EAT/heart volume [≥]95th percentile) with AUROCs of 0.91 to 0.95 across age-sex groups and correlated more strongly with indexed EAT volume ({rho}=0.76) than with BMI ({rho}=0.41, all P<0.001); PanAdipo alone explained 53% of indexed EAT volume variance, versus 23% for age, sex, and BMI combined. In MESA, higher scores were associated with an adverse metabolic profile (higher HOMA-IR [{beta}=0.17 per SD; 95% CI, 0.11-0.22]) after adjustment for age, sex, and BMI. In health system TTE and POCUS cohorts, higher scores were associated with newly documented MASLD/MASH (BMI-adjusted HRpooled 1.31 [1.03-1.67] per SD of log[PanAdipo]) and hyperlipidemia (1.11 [1.04-1.18]). ConclusionsAI-enabled echocardiography provides a scalable digital biomarker that characterizes prominent epicardial adiposity and is associated with an adverse cardiometabolic profile, highlighting a role for echocardiography in metabolic risk stratification. CONDENSED ABSTRACTEpicardial adipose tissue (EAT) is linked to cardiometabolic risk, but its assessment has traditionally required advanced imaging. We developed PanAdipo, a deep learning model that derives a digital biomarker of EAT from routine echocardiography. PanAdipo tracks CT-defined epicardial adiposity ({rho}=0.76; AUROC 0.91-0.95 for age-/sex-specific high indexed-EAT), yet correlates modestly with body mass index (BMI; {rho}=0.19-0.41). After BMI adjustment, higher scores were associated with insulin resistance in a community-based cohort and newly documented metabolic disease across health systems. AI-enabled echocardiography may provide a scalable window into epicardial adiposity, with potential applications for enriching trial populations and monitoring response to cardiometabolic therapies.
ER -