This week we review a recent work in the artificial intelligence/deep learning space when we review a recent report from the team at Boston Children's Hospital on a novel model trained to predict LV characteristics in children from the ECG. How does such a program identify LV dilation, dysfunction or hypertrophy from a test (the ECG) not designed to provide most of this information? How well did the model perform when it was tested in a different center entirely? Why is it important to understand how a model of this sort is making its determinations? How might such a model be used clinically? These are amongst the many questions reviewed with the work's first author, Dr. Joshua Mayourian MD, PhD who is a first year cardiology fellow at Boston Children's Hospital.

https://doi.org/10.1161/circulationaha.123.067750