Department of Computer Science, University of Oxford, Oxford, United Kingdom.
Department of Computer Science, University of Oxford, Oxford, United Kingdom.
J Electrocardiol. 2019 Nov-Dec;57S:S61-S64. doi: 10.1016/j.jelectrocard.2019.08.008. Epub 2019 Aug 8.
The electrocardiogram is the most widely used diagnostic tool that records the electrical activity of the heart and, therefore, its use for identifying markers for early diagnosis and detection is of paramount importance. In the last years, the huge increase of electronic health records containing a systematised collection of different type of digitalised medical data, together with new tools to analyse this large amount of data in an efficient way have re-emerged the field of machine learning in healthcare innovation. This review describes the most recent machine learning-based systems applied to the electrocardiogram as well as pros and cons in the use of these techniques. Machine learning, including deep learning, have shown to be powerful tools for aiding clinicians in patient screening and risk stratification tasks. However, they do not provide the physiological basis of classification outcomes. Computational modelling and simulation can help in the interpretation and understanding of key physiologically meaningful ECG biomarkers extracted from machine learning techniques.
心电图是最广泛使用的诊断工具，用于记录心脏的电活动，因此，它在识别早期诊断和检测标记物方面的应用至关重要。在过去几年中，电子健康记录的数量大幅增加，其中包含了不同类型的数字化医疗数据的系统收集，以及以高效的方式分析这些大量数据的新工具，这些都使得机器学习在医疗保健创新领域重新兴起。本综述描述了最近应用于心电图的基于机器学习的系统，以及使用这些技术的优缺点。机器学习，包括深度学习，已被证明是辅助临床医生进行患者筛查和风险分层任务的有力工具。然而，它们并未提供分类结果的生理基础。计算建模和模拟可以帮助解释和理解从机器学习技术中提取的关键生理意义上的 ECG 生物标志物。
