Centre for Human Drug Research, The Netherlands; Leiden Academic Centre for Drug Research, The Netherlands.
Centre for Human Drug Research, The Netherlands.
J Electrocardiol. 2022 May-Jun;72:49-55. doi: 10.1016/j.jelectrocard.2022.03.001. Epub 2022 Mar 11.
The aim of the present study was to develop a neural network to characterize the effect of aging on the ECG in healthy volunteers. Moreover, the impact of the various ECG features on aging was evaluated.
METHODS & RESULTS: A total of 6228 healthy subjects without structural heart disease were included in this study. A neural network regression model was created to predict age of the subjects based on their ECG; 577 parameters derived from a 12‑lead ECG of each subject were used to develop and validate the neural network; A tenfold cross-validation was performed, using 118 subjects for validation each fold. Using SHapley Additive exPlanations values the impact of the individual features on the prediction of age was determined. Of 6228 subjects tested, 1808 (29%) were females and mean age was 34 years, range 18-75 years. Physiologic age was estimated as a continuous variable with an average error of 6.9 ± 5.6 years (R = 0.72 ± 0.04). The correlation was slightly stronger for men (R = 0.74) than for women (R = 0.66). The most important features on the prediction of physiologic age were T wave morphology indices in leads V4 and V5, and P wave amplitude in leads AVR and II.
The application of machine learning to the ECG using a neural network regression model, allows accurate estimation of physiologic cardiac age. This technique could be used to pick up subtle age-related cardiac changes, but also estimate the reversing of these age-associated effects by administered treatments.
本研究旨在开发一种神经网络，以描述健康志愿者心电图随年龄变化的特征。此外，还评估了各种心电图特征对衰老的影响。
本研究共纳入 6228 名无结构性心脏病的健康受试者。建立了一种神经网络回归模型，根据受试者的心电图预测其年龄；使用每位受试者的 12 导联心电图得出 577 个参数，用于开发和验证神经网络；采用 10 折交叉验证，每折使用 118 名受试者进行验证。使用 Shapley Additive exPlanations 值确定了各个特征对年龄预测的影响。在 6228 名受试者中，1808 名（29%）为女性，平均年龄为 34 岁，年龄范围为 18-75 岁。生理年龄被估计为一个连续变量，平均误差为 6.9±5.6 岁（R=0.72±0.04）。男性（R=0.74）的相关性略强于女性（R=0.66）。对生理年龄预测最重要的特征是 V4 和 V5 导联的 T 波形态指数，以及 AVR 和 II 导联的 P 波振幅。
使用神经网络回归模型对心电图进行机器学习应用，可以准确估计生理心脏年龄。这种技术可以用于发现微妙的与年龄相关的心脏变化，还可以估计通过给予治疗逆转这些与年龄相关的效应。
