Key result
An artificial intelligence-enabled ECG algorithm accurately predicted coronary artery calcification scores ≥100, ≥400, and ≥1,000 with AUROCs of 0.753, 0.802, and 0.835, respectively, outperforming traditional logistic regression models.
Why the study?
Routine measurement of CAC scores is limited by cost, radiation, and availability, creating a need to predict CAC using standard, cost-effective ECGs.
Does an AI-enabled ECG algorithm accurately predict coronary artery calcification in adult patients?
Population
Adult patients with standard 12-lead ECGs measured within 60 days of CAC scores
Comparison
AI models predicting CAC score ≥100, ≥400, and ≥1,000 from raw ECG waveforms
Design
AI model development and external validation study
Authors
Loading...
May aid low-cost CAC screening via routine ECGs; hypothesis-generating and requires prospective validation before clinical use.
Observational (n=6,642)
Yes
Does an AI-enabled ECG algorithm accurately predict coronary artery calcification in adult patients?
Effect estimate: AUROC 0.753, 0.802, and 0.835
An AI-enabled algorithm using standard 12-lead ECGs can predict coronary artery calcification, potentially offering a cost-effective and accessible tool for cardiovascular risk stratification.
Han et al. (2022) conducted an observational in Coronary Artery Calcification (n=6,642). Artificial Intelligence-Enabled ECG Algorithm vs. Logistic regression models using traditional ECG features was evaluated on Prediction of Coronary Artery Calcification (CAC) score ≥ 100, ≥ 400, and ≥ 1,000 (AUROC 0.753, 0.802, and 0.835). An artificial intelligence-enabled ECG algorithm accurately predicted coronary artery calcification scores ≥100, ≥400, and ≥1,000 with AUROCs of 0.753, 0.802, and 0.835, respectively, outperforming traditional logistic regression models.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: