Global S&T Development Trend Analysis Platform of Resources and Environment
A video-based deep learning algorithm-EchoNet-Dynamic-accurately identifies subtle changes in ejection fraction and classifies heart failure with reduced ejection fraction using information from multiple cardiac cycles.
Accurate assessment of cardiac function is crucial for the diagnosis of cardiovascular disease(1), screening for cardiotoxicity(2) and decisions regarding the clinical management of patients with a critical illness(3). However, human assessment of cardiac function focuses on a limited sampling of cardiac cycles and has considerable inter-observer variability despite years of training(4,5). Here, to overcome this challenge, we present a video-based deep learning algorithm-EchoNet-Dynamic-that surpasses the performance of human experts in the critical tasks of segmenting the left ventricle, estimating ejection fraction and assessing cardiomyopathy. Trained on echocardiogram videos, our model accurately segments the left ventricle with a Dice similarity coefficient of 0.92, predicts ejection fraction with a mean absolute error of 4.1% and reliably classifies heart failure with reduced ejection fraction (area under the curve of 0.97). In an external dataset from another healthcare system, EchoNet-Dynamic predicts the ejection fraction with a mean absolute error of 6.0% and classifies heart failure with reduced ejection fraction with an area under the curve of 0.96. Prospective evaluation with repeated human measurements confirms that the model has variance that is comparable to or less than that of human experts. By leveraging information across multiple cardiac cycles, our model can rapidly identify subtle changes in ejection fraction, is more reproducible than human evaluation and lays the foundation for precise diagnosis of cardiovascular disease in real time. As a resource to promote further innovation, we also make publicly available a large dataset of 10,030 annotated echocardiogram videos.
Single-cell RNA sequencing and spatial transcriptomics reveal that the somitogenesis clock is active in mouse gastruloids, which can be induced to generate somites with the correct rostral-caudal patterning.
Gastruloids are three-dimensional aggregates of embryonic stem cells that display key features of mammalian development after implantation, including germ-layer specification and axial organization(1-3). To date, the expression pattern of only a small number of genes in gastruloids has been explored with microscopy, and the extent to which genome-wide expression patterns in gastruloids mimic those in embryos is unclear. Here we compare mouse gastruloids with mouse embryos using single-cell RNA sequencing and spatial transcriptomics. We identify various embryonic cell types that were not previously known to be present in gastruloids, and show that key regulators of somitogenesis are expressed similarly between embryos and gastruloids. Using live imaging, we show that the somitogenesis clock is active in gastruloids and has dynamics that resemble those in vivo. Because gastruloids can be grown in large quantities, we performed a small screen that revealed how reduced FGF signalling induces a short-tail phenotype in embryos. Finally, we demonstrate that embedding in Matrigel induces gastruloids to generate somites with the correct rostral-caudal patterning, which appear sequentially in an anterior-to-posterior direction over time. This study thus shows the power of gastruloids as a model system for exploring development and somitogenesis in vitro in a high-throughput manner.