INFANT Research Centre, University College Cork, Cork, Ireland.
Department of Physiology, School of Medicine, College of Medicine and Health, University College Cork, Cork, Ireland.
J Physiol. 2024 Nov;602(22):6347-6360. doi: 10.1113/JP287001. Epub 2024 Oct 19.
The present study was designed to test the potential utility of regional cerebral oxygen saturation (rcSO) in detecting term infants with brain injury. The study also examined whether quantitative rcSO features are associated with grade of hypoxic ischaemic encephalopathy (HIE). We analysed 58 term infants with HIE (>36 weeks of gestational age) enrolled in a prospective observational study. All newborn infants had a period of continuous rcSO monitoring and magnetic resonance imaging (MRI) assessment during the first week of life. rcSOSignals were pre-processed and quantitative features were extracted. Machine-learning and deep-learning models were developed to detect adverse outcome (brain injury on MRI or death in the first week) using the leave-one-out cross-validation approach and to assess the association between rcSO and HIE grade (modified Sarnat - at 1 h). The machine-learning model (rcSO excluding prolonged relative desaturations) significantly detected infant MRI outcome or death in the first week of life [area under the curve (AUC) = 0.73, confidence interval (CI) = 0.59-0.86, Matthew's correlation coefficient = 0.35]. In agreement, deep learning models detected adverse outcome with an AUC = 0.64, CI = 0.50-0.79. We also report a significant association between rcSO features and HIE grade using a machine learning approach (AUC = 0.81, CI = 0.73-0.90). We conclude that automated analysis of rcSO using machine learning methods in term infants with HIE was able to determine, with modest accuracy, infants with adverse outcome. De novo approaches to signal analysis of NIRS holds promise to aid clinical decision making in the future. KEY POINTS: Hypoxic-induced neonatal brain injury contributes to both short- and long-term functional deficits. Non-invasive continuous monitoring of brain oxygenation using near-infrared- spectroscopy offers a potential new insight to the development of serious injury. In this study, characteristics of the NIRS signal were summarised using either predefined features or data-driven feature extraction, both were combined with a machine learning approach to predict short-term brain injury. Using data from a cohort of term infants with hypoxic ischaemic encephalopathy, the present study illustrates that automated analysis of regional cerebral oxygen saturation rcSO, using either machine learning or deep learning methods, was able to determine infants with adverse outcome.
本研究旨在测试局部脑氧饱和度（rcSO）检测足月脑损伤婴儿的潜在效用。本研究还研究了定量 rcSO 特征是否与缺氧缺血性脑病（HIE）的严重程度有关。我们分析了 58 例胎龄大于 36 周的 HIE 足月婴儿，这些婴儿参与了一项前瞻性观察性研究。所有新生儿在生命的第一周都进行了连续的 rcSO 监测和磁共振成像（MRI）评估。预处理 rcSO 信号并提取定量特征。使用留一法交叉验证方法，开发了机器学习和深度学习模型，以检测不良结局（MRI 上的脑损伤或第一周内死亡），并评估 rcSO 与 HIE 严重程度（改良 Sarnat-1 小时）之间的关联。机器学习模型（排除延长相对缺氧的 rcSO）显著检测到婴儿 MRI 结局或第一周内死亡[曲线下面积（AUC）= 0.73，置信区间（CI）= 0.59-0.86，马修相关系数= 0.35]。一致地，深度学习模型使用 AUC=0.64，CI=0.50-0.79，检测到不良结局。我们还报告了使用机器学习方法，rcSO 特征与 HIE 严重程度之间存在显著关联（AUC=0.81，CI=0.73-0.90）。我们的结论是，使用机器学习方法对 HIE 足月婴儿的 rcSO 进行自动分析能够以中等准确性确定不良结局的婴儿。NIRS 信号的新方法有望在未来为临床决策提供帮助。关键点：缺氧引起的新生儿脑损伤导致短期和长期功能缺陷。使用近红外光谱对脑氧合进行非侵入性连续监测，为严重损伤的发展提供了新的潜在见解。在这项研究中，使用预定义特征或基于数据的特征提取来总结 NIRS 信号的特征，两者都与机器学习方法相结合，以预测短期脑损伤。使用缺氧缺血性脑病足月婴儿队列的数据，本研究表明，使用机器学习或深度学习方法对局部脑氧饱和度 rcSO 的自动分析能够确定不良结局的婴儿。
