Background:Artificial intelligence (AI)–enabled analysis of 12-lead ECGs may facilitate efficient estimation of incident atrial fibrillation (AF) risk. However, it remains unclear whether AI provides meaningful and generalizable improvement in predictive accuracy beyond clinical risk factors for AF.Methods:We trained a convolutional neural network (ECG-AI) to infer 5-year incident AF risk using 12-lead ECGs in patients receiving longitudinal primary care at Massachusetts General Hospital (MGH). We then fit 3 Cox proportional hazards models, composed of ECG-AI 5-year AF probability, CHARGE-AF clinical risk score (Cohorts for Heart and Aging in Genomic Epidemiology–Atrial Fibrillation), and terms for both ECG-AI and CHARGE-AF (CH-AI), respectively. We assessed model performance by calculating discrimination (area under the receiver operating characteristic curve) and calibration in an internal test set and 2 external test sets (Brigham and Women’s Hospital [BWH] and UK Biobank). Models were recalibrated to estimate 2-year AF risk in the UK Biobank given limited available follow-up. We used saliency mapping to identify ECG features most influential on ECG-AI risk predictions and assessed correlation between ECG-AI and CHARGE-AF linear predictors.Results:The training set comprised 45 770 individuals (age 55±17 years, 53% women, 2171 AF events) and the test sets comprised 83 162 individuals (age 59±13 years, 56% women, 2424 AF events). Area under the receiver operating characteristic curve was comparable using CHARGE-AF (MGH, 0.802 [95% CI, 0.767–0.836]; BWH, 0.752 [95% CI, 0.741–0.763]; UK Biobank, 0.732 [95% CI, 0.704–0.759]) and ECG-AI (MGH, 0.823 [95% CI, 0.790–0.856]; BWH, 0.747 [95% CI, 0.736–0.759]; UK Biobank, 0.705 [95% CI, 0.673–0.737]). Area under the receiver operating characteristic curve was highest using CH-AI (MGH, 0.838 [95% CI, 0.807 to 0.869]; BWH, 0.777 [95% CI, 0.766 to 0.788]; UK Biobank, 0.746 [95% CI, 0.716 to 0.776]). Calibration error was low using ECG-AI (MGH, 0.0212; BWH, 0.0129; UK Biobank, 0.0035) and CH-AI (MGH, 0.012; BWH, 0.0108; UK Biobank, 0.0001). In saliency analyses, the ECG P-wave had the greatest influence on AI model predictions. ECG-AI and CHARGE-AF linear predictors were correlated (Pearson r: MGH, 0.61; BWH, 0.66; UK Biobank, 0.41).Conclusions:AI-based analysis of 12-lead ECGs has similar predictive usefulness to a clinical risk factor model for incident AF and the approaches are complementary. ECG-AI may enable efficient quantification of future AF risk.

中文翻译：

---

基于心电图的深度学习和临床风险因素预测心房颤动

背景：人工智能 (AI) 支持的 12 导联心电图分析可能有助于有效估计房颤 (AF) 事件的风险。然而，除了 AF 的临床风险因素之外，AI 是否在预测准确性方面提供了有意义且可推广的改进，目前尚不清楚。方法：我们训练了一个卷积神经网络 (ECG-AI)，使用患者的 12 导联心电图来推断 5 年的 AF 风险在麻省总医院 (MGH) 接受纵向初级保健。然后，我们拟合 3 个 Cox 比例风险模型，由 ECG-AI 5 年 AF 概率、CHARGE-AF 临床风险评分（基因组流行病学中的心脏和衰老队列 - 心房颤动）以及 ECG-AI 和 CHARGE- 的术语组成AF（CH-AI），分别。我们通过在内部测试集和 2 个外部测试集（布莱根妇女医院 [BWH] 和 UK Biobank）中计算辨别力（接受者操作特征曲线下的面积）和校准来评估模型性能。鉴于有限的可用后续行动，模型被重新校准以估计英国生物银行 2 年的 AF 风险。我们使用显着性映射来识别对 ECG-AI 风险预测最有影响的 ECG 特征，并评估 ECG-AI 和 CHARGE-AF 线性预测因子之间的相关性。结果：训练集包括 45770 人（年龄 55±17 岁，53% 的女性， 2171 个 AF 事件）和测试集包括 83162 个个体（年龄 59±13 岁，56% 为女性，2424 个 AF 事件）。接受者操作特征曲线下的面积使用 CHARGE-AF 进行比较（MGH，0.802 [95% CI，0.767–0.836]；BWH，0.752 [95% CI，0.741–0.763]；英国生物库，0.732 [95% CI，0.704–0.759]）和 ECG-AI（MGH，0.823 [95% CI，0.790–0.856]；BWH，0.747 [95% CI，0.736–0.759]；英国生物库，0.705 [ 95% CI，0.673–0.737]）。使用 CH-AI 时受试者工作特征曲线下的面积最高（MGH，0.838 [95% CI，0.807 至 0.869]；BWH，0.777 [95% CI，0.766 至 0.788]；UK Biobank，0.746 [95% CI，0.716]到 0.776]）。使用 ECG-AI（MGH，0.0212；BWH，0.0129；UK Biobank，0.0035）和 CH-AI（MGH，0.012；BWH，0.0108；UK Biobank，0.0001）的校准误差较低。在显着性分析中，ECG P 波对 AI 模型预测的影响最大。ECG-AI 和 CHARGE-AF 线性预测因子相关（Pearson 使用 CH-AI 时受试者工作特征曲线下的面积最高（MGH，0.838 [95% CI，0.807 至 0.869]；BWH，0.777 [95% CI，0.766 至 0.788]；UK Biobank，0.746 [95% CI，0.716]到 0.776]）。使用 ECG-AI（MGH，0.0212；BWH，0.0129；UK Biobank，0.0035）和 CH-AI（MGH，0.012；BWH，0.0108；UK Biobank，0.0001）的校准误差较低。在显着性分析中，ECG P 波对 AI 模型预测的影响最大。ECG-AI 和 CHARGE-AF 线性预测因子相关（Pearson 使用 CH-AI 时受试者工作特征曲线下的面积最高（MGH，0.838 [95% CI，0.807 至 0.869]；BWH，0.777 [95% CI，0.766 至 0.788]；UK Biobank，0.746 [95% CI，0.716]到 0.776]）。使用 ECG-AI（MGH，0.0212；BWH，0.0129；UK Biobank，0.0035）和 CH-AI（MGH，0.012；BWH，0.0108；UK Biobank，0.0001）的校准误差较低。在显着性分析中，ECG P 波对 AI 模型预测的影响最大。ECG-AI 和 CHARGE-AF 线性预测因子相关（Pearson ECG P 波对 AI 模型预测的影响最大。ECG-AI 和 CHARGE-AF 线性预测因子相关（Pearson ECG P 波对 AI 模型预测的影响最大。ECG-AI 和 CHARGE-AF 线性预测因子相关（Pearsonr : MGH, 0.61; 体重，0.66；UK Biobank，0.41）。结论：基于 AI 的 12 导联 ECG 分析与事件 AF 的临床风险因素模型具有相似的预测效用，并且这些方法是互补的。ECG-AI 可以有效量化未来的 AF 风险。