Early Clinical Development, Pfizer Incorporated, Cambridge, Massachusetts.
Comparative Medicine, Pfizer Incorporated, Cambridge, Massachusetts.
Am J Physiol Heart Circ Physiol. 2022 Oct 1;323(4):H628-H639. doi: 10.1152/ajpheart.00208.2022. Epub 2022 Aug 19.
Echocardiography (echo) is a translationally relevant ultrasound imaging modality widely used to assess cardiac structure and function in preclinical models of heart failure (HF) during research and drug development. Although echo is a very valuable tool, the image analysis is a time-consuming, resource-demanding process, and is susceptible to interreader variability. Recent advancements in deep learning have enabled researchers to automate image processing and reduce analysis time and interreader variability in the field of medical imaging. In the present study, we developed a fully automated tool, mouse-echocardiography neural net (MENN), for the analysis of both long-axis brightness (B)-mode and short-axis motion (M)-mode images of left ventricle. MENN is a series of fully convolutional neural networks that were trained and validated using manually segmented B-mode and M-mode echo images of the left ventricle. The segmented images were then used to compute cardiac structural and functional metrics. The performance of MENN was further validated in two preclinical models of HF. MENN achieved excellent correlations (Pearson's = 0.85-0.99) and good-to-excellent agreement between automated and manual analyses. Further interreader variability analysis showed that MENN has better agreements with an expert analyst than both a trained analyst and a novice. Notably, the use of MENN reduced manual analysis time by >92%. In conclusion, we developed an automated echocardiography analysis tool that allows for fast and accurate analysis of B-mode and M-mode mouse echo data and mitigates the issue of interreader variability in manual analysis. Echocardiography is commonly used in preclinical research to evaluate cardiac structure and function. Despite the broad applications across therapeutic areas, the analysis of echo data is laborious and susceptible to interreader variability. In this study, we developed a fully automated mouse-echocardiography neural net (MENN). Cardiac measurements from MENN showed excellent correlations with manual analysis. Furthermore, the use of MENN leads to >92% reduction in analysis time and potentially eliminates the interobserver variability issue.
超声心动图（echo）是一种转化相关的超声成像方式，广泛用于在心力衰竭（HF）的临床前模型中评估心脏结构和功能，在研究和药物开发过程中。尽管超声心动图是一种非常有价值的工具，但图像分析是一个耗时、资源密集型的过程，并且容易受到读者间变异性的影响。深度学习的最新进展使研究人员能够在医学成像领域实现图像处理的自动化，并减少分析时间和读者间的变异性。在本研究中，我们开发了一种全自动工具，即鼠标超声心动图神经网络（MENN），用于分析左心室的长轴亮度（B）模式和短轴运动（M）模式图像。MENN 是一系列完全卷积神经网络，使用手动分割的左心室 B 模式和 M 模式超声心动图图像进行训练和验证。然后使用分割图像计算心脏结构和功能指标。MENN 的性能在两种心力衰竭的临床前模型中得到了进一步验证。MENN 实现了出色的相关性（Pearson's = 0.85-0.99）和自动分析与手动分析之间的良好到极好的一致性。进一步的读者间变异性分析表明，MENN 与专家分析员的一致性优于训练有素的分析员和新手。值得注意的是，MENN 的使用将手动分析时间减少了超过 92%。总之，我们开发了一种自动化超声心动图分析工具，允许快速准确地分析 B 模式和 M 模式鼠标超声心动图数据，并减轻手动分析中的读者间变异性问题。超声心动图常用于临床前研究，以评估心脏结构和功能。尽管在治疗领域有广泛的应用，但超声心动图数据的分析既费力又容易受到读者间变异性的影响。在这项研究中，我们开发了一种全自动的鼠标超声心动图神经网络（MENN）。MENN 的心脏测量结果与手动分析具有极好的相关性。此外，MENN 的使用将分析时间减少了超过 92%，并可能消除了观察者间变异性问题。
