Deep learning evaluation of echocardiograms to identify occult atrial fibrillation.
Deep learning evaluation of echocardiograms to identify occult atrial fibrillation.
复制标题
超声心动图的深度学习评估以识别隐匿性心房颤动。
DOI：
10.1038/s41746-024-01090-z
复制
发表时间：
2024
影响因子：
15.2
通讯作者：
Ouyang,David
中科院分区：
文献类型：
--
作者：
Yuan,Neal;Stein,NathanR;Duffy,Grant;Sandhu,RoopinderK;Chugh,SumeetS;Chen,Peng-Sheng;Rosenberg,Carine;Albert,ChristineM;Cheng,Susan;Siegel,RobertJ;Ouyang,David
由于心房颤动 (AF) 频繁出现阵发性且无症状，因此常常难以被发现。经胸超声心动图 (TTE) 的深度学习具有结构信息，可以帮助识别隐匿性 AF。我们使用基于视频的卷积神经网络模型创建了一个两阶段深度学习算法，该算法 (1) 区分 TTE 是窦性心律还是 AF，然后 (2) 预测哪些 TTE 处于窦性心律是在 90 天内经历过 AF 的患者。我们的模型在 111,319 个 TTE 视频上进行训练，在测试队列中以高精度区分 AF 中的 TTE 和窦性心律中的 TTE（AUC 0.96 (0.95–0.96)、AUPRC 0.91 (0.90–0.92)）。在窦性心律的 TTE 中，模型预测并发阵发性 AF 的存在（AUC 0.74 (0.71–0.77)，AUPRC 0.19 (0.16–0.23)）。在 10,203 个 TTE 的外部队列中，模型歧视仍然相似（AUC 为 0.69 (0.67–0.70)，AUPRC 0.34 (0.31–0.36)）。女性患者 (AUC 0.76 (0.72–0.81))、65 岁以上患者 (0.73 (0.69–0.76)) 或 CHA2DS2VASc ≥2 (0.73 (0.79–0.77)) 患者的表现均保持不变。该模型的表现优于使用临床危险因素 (AUC 0.64 (0.62–0.67))、TTE 测量值 (0.64 (0.62–0.67))、左心房大小 (0.63 (0.62–0.64)) 或 CHA2DS2VASc (0.61 (0.60–0.62))。将 TTE 模型与心电图 (ECG) 深度学习模型相结合的队列子集中的集成模型比单独使用 ECG 模型表现更好（AUC 0.81 vs. 0.79，p = 0.01）。使用 TTE 的深度学习可以预测患有活动性或隐匿性 AF 的患者，并可用于机会性 AF 筛查，从而实现早期治疗。
Atrial fibrillation (AF) often escapes detection, given its frequent paroxysmal and asymptomatic presentation. Deep learning of transthoracic echocardiograms (TTEs), which have structural information, could help identify occult AF. We created a two-stage deep learning algorithm using a video-based convolutional neural network model that (1) distinguished whether TTEs were in sinus rhythm or AF and then (2) predicted which of the TTEs in sinus rhythm were in patients who had experienced AF within 90 days. Our model, trained on 111,319 TTE videos, distinguished TTEs in AF from those in sinus rhythm with high accuracy in a held-out test cohort (AUC 0.96 (0.95–0.96), AUPRC 0.91 (0.90–0.92)). Among TTEs in sinus rhythm, the model predicted the presence of concurrent paroxysmal AF (AUC 0.74 (0.71–0.77), AUPRC 0.19 (0.16–0.23)). Model discrimination remained similar in an external cohort of 10,203 TTEs (AUC of 0.69 (0.67–0.70), AUPRC 0.34 (0.31–0.36)). Performance held across patients who were women (AUC 0.76 (0.72–0.81)), older than 65 years (0.73 (0.69–0.76)), or had a CHA2DS2VASc ≥2 (0.73 (0.79–0.77)). The model performed better than using clinical risk factors (AUC 0.64 (0.62–0.67)), TTE measurements (0.64 (0.62–0.67)), left atrial size (0.63 (0.62–0.64)), or CHA2DS2VASc (0.61 (0.60–0.62)). An ensemble model in a cohort subset combining the TTE model with an electrocardiogram (ECGs) deep learning model performed better than using the ECG model alone (AUC 0.81 vs. 0.79, p = 0.01). Deep learning using TTEs can predict patients with active or occult AF and could be used for opportunistic AF screening that could lead to earlier treatment.