AI-derived cardiovascular age from point-of-care ultrasound uncovers accelerated cardiometabolic aging
R Amster, D Simanovsky, R Loutati, R Raphael, A Goshen, E Bischof, T Strauss, E MaorAbstract
Background
Artificial intelligence (AI)-based biological age models offer a non-invasive means to quantify cardiovascular (CV) aging, potentially capturing risk beyond chronological age. We evaluated the clinical and cardiometabolic correlates of accelerated CV aging estimated from AI-augmented focused cardiac ultrasound (FoCUS) in a preventive screening cohort.
Methods
Asymptomatic self-referred adults undergoing annual screening in a hospital-based program underwent FoCUS imaging. An AI model estimated CV biological age from FoCUS images, with accelerated aging defined as the top quintile of the delta (CV biological age minus chronological age) after Beheshti bias correction. Participants were stratified into accelerated (quintile 5, Q5) versus expected aging (quintiles 1–4). Demographic, anthropometric, body composition, and laboratory parameters were compared between groups.
Results
Among 1,506 participants (mean age 66 years; 38% women), chronological age was similar between groups (p=0.659). As expected, the Accelerated Aging group had higher CV biological age (68±5 vs. 60±6 years, p<0.001).
Compared with expected aging, accelerated agers showed higher BMI (28±4 vs. 27±4 kg/m², p<0.001), waist circumference (98±14 vs. 95±12 cm, p=0.005), and body fat percentage (36±8% vs. 33±8%, p<0.001). They also had higher resting heart rate (71±11 vs. 68±11 bpm, p<0.001), diastolic blood pressure (81±9 vs. 79±8 mmHg, p=0.001) and triglycerides (120±62 vs. 105±50 mg/dL, p<0.001). Metabolic syndrome was more prevalent in Q5 (27% vs. 17%, p<0.001). In participants aged 40–69 years (N=963), accelerated agers trended toward higher SCORE2 CV risk (5.6±2 vs. 5.3±2, p=0.07).
Conclusions
AI-derived CV biological age from FoCUS identifies subclinical cardiometabolic vulnerability independent of chronological age. This rapid, point-of-care tool may enhance personalized risk profiling and guide early preventive interventions in asymptomatic adults.Table