在这项最新的研究中，人工智能模型接受了23万名患者超过120万对的心电图-超声心动图的对照训练，以检测多种形式的结构性心脏病。此后，团队对近 85000名此前未做过超声心动图检查且正在接受心电图检查的患者使用了EchoNext筛查工具，EchoNext识别出超过7500名存在未确诊的结构性心脏病的高风 险。在这些被EchoNext认定为高风险的人群中，有超过一半人接受了超声心动图检查。结果显示，AI工具对结构性心脏病诊断的准确率高达77.3%。 研究团队将EchoNext模型与人类心脏病专家进行了比较，共有13位心脏病专家对3200份心电图的评估（其中1600份使用AI辅助，1600份未使用AI辅助，每 位心脏病专家平均评估了246份心电图）来诊断结构性心脏病。结果显示，在未使用AI辅助的情况下，心脏病专家的准确率为64%；在使用AI辅助的情况 下，心脏病专家的准确率提升至69.2%，但仍低于EchoNext模型检测的77.3%的准确率。 由于缺乏常规且经济有效的筛查测试，许多有结构性心脏问题的人直到心脏功能严重受损时才被发现。 早期发现结构性心脏病对于改善预后至关重要，但广泛筛查仍然受限于超声心动图等影像工 ...

Group 1 - The prevalence of heart valve diseases in China is increasing, with approximately 25 million people affected, making it the third most common cardiovascular disease after coronary heart disease and hypertension [1] - The most common type of heart valve disease is mitral valve disease, which is particularly significant due to the high pressure it endures [1] - Aging and modern lifestyle factors contribute to the rising incidence of heart valve diseases, leading to a younger demographic being affected [1] Group 2 - Treatment for valve diseases typically requires surgical intervention, either through repair or replacement, as medications alone are insufficient [2] - Advances in technology have led to minimally invasive procedures for mitral valve treatment, reducing surgical trauma [2] - The introduction of new products, such as the MITRIS RESILIA dry mitral valve, enhances treatment options and durability for patients [2] Group 3 - Improved diagnostic capabilities, particularly the widespread use of echocardiography, allow for early detection of valve diseases even before symptoms appear [3] - Regular health check-ups and screenings are essential for early diagnosis and management of heart valve conditions [3]