心肌应变是反映应力作用下整体或局部心肌形变程度的无量纲参数，可以量化检测心肌损伤，指导心脏疾病的早期诊断、干预与预后评估。心脏超声、心脏CT、心脏磁共振均可被用来进行应变成像与分析，其中二维斑点追踪超声心动图是当下应用最为广泛的心肌应变检测手段。但由于人工分析心肌应变存在观察者间差异且所使用的成像系统和分析软件各有不同，测得的应变值在不同供应商中一致性和可重复性欠佳，限制了心肌应变参数的临床应用。而人工智能可以通过自动应变计算和图像质量评估等方式在一定程度上克服应变测量的缺陷，具备广阔的发展前景。该文重点介绍人工智能在超声、磁共振等影像学手段中辅助测量心肌应变的研究进展，及其在疾病诊断与患者预后评估中的应用，希望助力于提高应变测量的效率和一致性，推动心肌应变常规应用于临床，在心肌损伤及心功能评估中发挥增量作用。然而，目前大部分研究涉及的样本量较小，并且缺乏外部验证，其结果的可靠性还需进一步证实。

Myocardial strain is a dimensionless parameter reflecting the degree of deformation of the whole or local myocardium under stress, which can quantitatively detect myocardial injury and guide the early diagnosis, intervention and prognostic assessment of cardiac diseases. Cardiac ultrasound, cardiac CT and cardiac magnetic resonance can all be used for strain imaging and analysis, with two-dimensional speckle-tracking echocardiography being the most widely used means of myocardial strain detection today. However, due to inter-observer variations in manual analysis of myocardial strain and differences in the imaging systems and analysis software, the consistency and reproducibility of measured strain values among vendors are poor, limiting the clinical application of myocardial strain. Artificial intelligence (AI) can overcome the defects of strain measurement to a certain extent through automatic strain calculation and image quality assessment, which has a broad developmental prospect. This review focuses on the progress of AI-assisted measurement of myocardial strain in ultrasound, magnetic resonance, and other imaging modalities, as well as its application to disease diagnosis and patient prognosis assessment. This will improve the efficiency and consistency of strain measurement and promote the routine application of myocardial strain to clinical practice, which will play an incremental role in assessing myocardial injury and cardiac function. However, most of the current studies involve small sample sizes and lack external validation, and the reliability of their results needs to be further verified.