Sakamoto Akira, Kaneko Tomohiro, Sato Eiichiro, Fujita Wataru, Nakamura Yutaka, Yokotsuka Noriko, Kagiyama Nobuyuki
Department of Cardiovascular Biology and Medicine, Juntendo University Graduate School of Medicine, 2-1-1 Hongo, Tokyo, 113-0021, Japan.
Clinical Laboratory, Juntendo University Hospital, Tokyo, Japan.
J Echocardiogr. 2025 Aug 21. doi: 10.1007/s12574-025-00703-0.
Artificial intelligence (AI) is rapidly transforming the field of echocardiography. By leveraging machine learning, particularly deep learning, AI enhances image acquisition, interpretation, and diagnostic accuracy. It addresses long-standing limitations of echocardiography, such as operator dependency and inter-observer variability. AI-enabled systems, ranging from probe guidance to automated quantification tools, have improved image quality and reduced variability in key measurements such as left ventricular ejection fraction (LVEF). Recent studies show that AI can assist in disease classification, detect regional wall motion abnormalities, and predict disease progression with accuracy comparable to expert assessment. Despite these advances, several challenges remain. Concerns regarding data bias, limited generalizability across populations and devices, and the "black-box" nature of many AI models hinder clinical adoption. Ethical issues, including data privacy and unequal access to digital technologies, also require careful attention. Importantly, AI should be viewed not as a replacement for human expertise but as a tool to augment clinical decision-making and improve workflow efficiency. Looking ahead, integrating echocardiographic data with other clinical information through AI could enable earlier diagnosis and better patient management. As technology evolves, AI is expected to reinforce echocardiography's role as a non-invasive, widely available, and highly informative diagnostic modality. Continued research and rigorous validation are essential to ensure the safe, equitable, and effective use of AI in clinical echocardiography.
人工智能（AI）正在迅速改变超声心动图领域。通过利用机器学习，尤其是深度学习，人工智能提高了图像采集、解读和诊断准确性。它解决了超声心动图长期存在的局限性，如对操作者的依赖性和观察者间的变异性。从探头引导到自动定量工具的人工智能系统，提高了图像质量，减少了左心室射血分数（LVEF）等关键测量中的变异性。最近的研究表明，人工智能可以协助疾病分类，检测局部室壁运动异常，并以与专家评估相当的准确性预测疾病进展。尽管取得了这些进展，但仍存在一些挑战。对数据偏差、在不同人群和设备间的有限通用性以及许多人工智能模型的“黑匣子”性质的担忧阻碍了临床应用。包括数据隐私和数字技术获取不平等在内的伦理问题也需要仔细关注。重要的是，人工智能不应被视为取代人类专业知识，而应被视为增强临床决策和提高工作流程效率的工具。展望未来，通过人工智能将超声心动图数据与其他临床信息整合起来，可以实现更早的诊断和更好的患者管理。随着技术的发展，人工智能有望加强超声心动图作为一种无创、广泛可用且信息丰富的诊断方式的作用。持续的研究和严格的验证对于确保人工智能在临床超声心动图中的安全、公平和有效使用至关重要。
