Early diagnosis of pulmonary hypertension (PH) is critical for effective treatment and management. We aimed to develop and externally validate an artificial intelligence algorithm that could serve as a PH screening tool, based on analysis of a standard 12-lead ECG.

The PH Early Detection Algorithm (PH-EDA) is a convolutional neural network developed using retrospective ECG voltage–time data, with patients classified as "PH-likely" or "PH-unlikely" (controls) based on right heart catheterisation or echocardiography. In total, 39 823 PH-likely patients and 219 404 control patients from Mayo Clinic were randomly split into training (48%), validation (12%) and test (40%) sets. ECGs taken within 1 month of PH diagnosis (diagnostic dataset) were used to train the PH-EDA at Mayo Clinic. Performance was tested on diagnostic ECGs within the test sets from Mayo Clinic (n=16 175/87 998 PH-likely/controls) and Vanderbilt University Medical Center (VUMC; n=6045/24 256 PH-likely/controls). In addition, performance was tested on ECGs taken 6–18 months (pre-emptive dataset), and up to 5 years prior to a PH diagnosis at both sites.

Performance testing yielded an area under the receiver operating characteristic curve (AUC) of 0.92 and 0.88 in the diagnostic test sets at Mayo Clinic and VUMC, respectively, and 0.86 and 0.81, respectively, in the pre-emptive test sets. The AUC remained a minimum of 0.79 at Mayo Clinic and 0.73 at VUMC up to 5 years before diagnosis.

The PH-EDA can detect PH at diagnosis and 6–18 months prior, demonstrating the potential to accelerate diagnosis and management of this debilitating disease.

肺动脉高压（PH）的早期诊断对于有效的治疗和管理至关重要。我们的目标是基于对标准 12 导联心电图的分析，开发并在外部验证一种人工智能算法，该算法可以用作 PH 筛查工具。

PH 早期检测算法 (PH-EDA) 是一种使用回顾性心电图电压-时间数据开发的卷积神经网络，根据右心导管插入术或超声心动图将患者分类为“疑似 PH”或“不太可能 PH”（对照）。总共，来自 Mayo Clinic 的 39 823 名疑似 PH 患者和 219 404 名对照患者被随机分为训练组 (48%)、验证组 (12%) 和测试组 (40%)。 Mayo Clinic 使用 PH 诊断后 1 个月内采集的心电图（诊断数据集）来训练 PH-EDA。在 Mayo Clinic（n=16 175/87 998 PH 可能/对照）和范德比尔特大学医学中心（VUMC；n=6045/24 256 PH 可能/对照）的测试集中对诊断心电图进行了性能测试。此外，还对两个地点的 PH 诊断前 6-18 个月（先发制人数据集）以及长达 5 年的心电图进行了性能测试。

性能测试得出，Mayo Clinic 和 VUMC 的诊断测试集中的受试者工作特征曲线下面积 (AUC) 分别为 0.92 和 0.88，在先发制人的测试集中分别为 0.86 和 0.81。诊断前 5 年内，Mayo Clinic 的 AUC 最低为 0.79，VUMC 的 AUC 最低为 0.73。

PH-EDA 可以在诊断时和提前 6-18 个月检测 PH，这表明有可能加快这种使人衰弱的疾病的诊断和治疗。