No abstract

Echocardiography is essential to cardiology. However, the need for human interpretation has limited echocardiography’s full potential for precision medicine. Deep learning is an emerging tool for analyzing images but has not yet been widely applied to echocardiograms, partly due to their complex multi-view format. The essential first step toward comprehensive computer-assisted echocardiographic interpretation is determining whether computers can learn to recognize these views. We trained a convolutional neural network to simultaneously classify 15 standard views (12 video, 3 still), based on labeled still images and videos from 267 transthoracic echocardiograms that captured a range of real-world clinical variation. Our model classified among 12 video views with 97.8% overall test accuracy without overfitting. Even on single low-resolution images, accuracy among 15 views was 91.7% vs. 70.2–84.0% for board-certified echocardiographers. Data visualization experiments showed that the model recognizes similarities among related views and classifies using clinically relevant image features. Our results provide a foundation for artificial intelligence-assisted echocardiographic interpretation.

Vertebrate hearts depend on highly specialized cardiomyocytes that form the cardiac conduction system (CCS) to coordinate chamber contraction and drive blood efficiently and unidirectionally throughout the organism. Defects in this specialized wiring system can lead to syncope and sudden cardiac death. Thus, a greater understanding of cardiac conduction development may help to prevent these devastating clinical outcomes. Utilizing a cardiac-specific fluorescent calcium indicator zebrafish transgenic line, Tg(cmlc2:gCaMP)s878, that allows for in vivo optical mapping analysis in intact animals, we identified and analyzed four distinct stages of cardiac conduction development that correspond to cellular and anatomical changes of the developing heart. Additionally, we observed that epigenetic factors, such as hemodynamic flow and contraction, regulate the fast conduction network of this specialized electrical system. To identify novel regulators of the CCS, we designed and performed a new, physiology-based, forward genetic screen and identified for the first time, to our knowledge, 17 conduction-specific mutations. Positional cloning of hobgoblins634 revealed that tcf2, a homeobox transcription factor gene involved in mature onset diabetes of the young and familial glomerulocystic kidney disease, also regulates conduction between the atrium and the ventricle. The combination of the Tg(cmlc2:gCaMP)s878 line/in vivo optical mapping technique and characterization of cardiac conduction mutants provides a novel multidisciplinary approach to further understand the molecular determinants of the vertebrate CCS.