A wearable cardiac ultrasound imager can noninvasively capture real-time images of the human heart for an extended period of time. The patch, which is about the size of a postage stamp, has comparable performance to a commercial ultrasound device. The imager can be worn during exercise, providing valuable cardiac information when the heart is under stress.
The researchers evaluated how their wearable patch compared with a traditional ultrasound device. Using a human tissue phantom model, they characterized the properties of their patch, such as its spatial resolution and its signal-to-noise ratio and found that these characteristics were similar to a commercial ultrasound device. Next, they used their patch to image the heart of a human subject. Looking at four standard views of cardiac anatomy, they found that the images generated by their patch were similar to those generated by the commercial ultrasound device.
After characterizing the performance of their patch, the researchers evaluated its utility during exercise — specifically, how it fares during a stress test. Traditional stress echocardiography evaluates images of the heart before and after intensive exercise (as holding an ultrasound probe over the chest by hand and maintaining a stable position is impossible during exercise).
The patch was used to monitor the cardiac performance of one healthy participant during a rigorous exercise session. The researchers found that the patch could capture the activities of the left ventricle (the heart chamber that pumps oxygenated blood throughout the body) without any interruption.
Finally, the researchers wanted to determine whether their patch could be used to calculate key cardiac functions, such as stroke volume, cardiac output, and ejection fraction (which are all related to how much blood is pumped out of the heart and how efficiently the heart is working). Using deep learning, they extracted specific features from ultrasound images taken with their patch and trained a model to reliably extrapolate these cardiac metrics.