Machine Learning for ECG Diagnosis of LV Dysfunction (JACC Editorial) – June, 2021

Since the first report of the human electrocardiogram (ECG) by Augustus Waller in 1887 (1), its diagnostic use has continually expanded. Providing a window to the electric activity of the heart, the ECG allowed the documentation and classification of arrhythmias, soon followed by estimation of atrial and ventricular size (2). Some years later, dynamic changes associated with myocardial infarction were discovered and shown to correlate with clinical and histopathologic changes (2). Researchers have continued to extract a diversifying range of clinically useful data from this simple bedside test, demonstrating the wealth of information contained within it and making the ECG ubiquitous in clinical medicine.