Kwon JM, Lee SY, Jeon KH, et al. - By performing this retrospective cohort analysis, researchers created as well as validated a deep learning–based algorithm, integrating a multilayer perceptron and convolutional neural network, for identifying significant aortic stenosis (AS) using ECGs. Participants were adult patients who had undergone both ECG and echocardiography. Using 39,371 ECGs, a deep learning–based algorithm was constructed. Experts used demographic data, features, and 500‐Hz, 12‐lead ECG raw data as predictive variables. For the detection of significant AS, the areas under the receiver operating characteristic curve, during internal and external validation, of the deep learning–based algorithm using 12‐lead ECG were identified to be 0.884 and 0.861, respectively; those using a single‐lead ECG signal were identified to be 0.845 and 0.821, respectively. Overall, findings revealed high accuracy of deep learning–based algorithm for the detection of significant AS using both 12‐lead and single‐lead ECGs.