JACC. Advances · 2025
Automatic Diagnosis of Left Valvular Heart Disease Based on Artificial Intelligence Stethoscope.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Zhou Ziwei, Xie Kaipeng, Huang Yiquan, Zhang Wenjing, Li Bingzhen, Zhong Junqi, Liao Xinxue, Wang Changhong, Zhuang Xiaodong
Аннотация
BACKGROUND: Valvular heart disease (VHD) remains underdiagnosed and results in serious complications. Early screening for VHD facilitates enhanced clinical management.
OBJECTIVES: This study aim to develop an artificial intelligence-based stethoscope model for detecting left-sided VHD, including aortic stenosis, aortic regurgitation, mitral stenosis, and mitral regurgitation.
METHODS: Using an electronic stethoscope, we recorded heart sounds from derivation group to construct a machine learning algorithm. Then, the algorithm was tested on a testing group. Echocardiography was referred as the gold standard. Model performance was assessed using area under the receiver-operating characteristic (AU-ROC).
RESULTS: A total of 514 patients were included in the final analyses (304 in the algorithm training group and 210 in the result testing group). The diagnostic performance of machine learning model was as follows: aortic stenosis (AU-ROC: 0.7621), aortic regurgitation (AU-ROC: 0.7075), mitral stenosis (AU-ROC: 0.6426), mitral regurgitation (AU-ROC: 0.7906), and left-sided VHD (AU-ROC: 0.8541; sensitivity 83.07%, specificity 78.26%). When applied to the testing group, the sensitivity, specificity, and AU-ROC for identifying left-sided VHD were 70.00%, 73.68%, and 0.7554, respectively.
CONCLUSIONS: Artificial intelligence-based stethoscope is capable of diagnosing left-sided VHD accurately and may make routine screening for VHD more practical.
OBJECTIVES: This study aim to develop an artificial intelligence-based stethoscope model for detecting left-sided VHD, including aortic stenosis, aortic regurgitation, mitral stenosis, and mitral regurgitation.
METHODS: Using an electronic stethoscope, we recorded heart sounds from derivation group to construct a machine learning algorithm. Then, the algorithm was tested on a testing group. Echocardiography was referred as the gold standard. Model performance was assessed using area under the receiver-operating characteristic (AU-ROC).
RESULTS: A total of 514 patients were included in the final analyses (304 in the algorithm training group and 210 in the result testing group). The diagnostic performance of machine learning model was as follows: aortic stenosis (AU-ROC: 0.7621), aortic regurgitation (AU-ROC: 0.7075), mitral stenosis (AU-ROC: 0.6426), mitral regurgitation (AU-ROC: 0.7906), and left-sided VHD (AU-ROC: 0.8541; sensitivity 83.07%, specificity 78.26%). When applied to the testing group, the sensitivity, specificity, and AU-ROC for identifying left-sided VHD were 70.00%, 73.68%, and 0.7554, respectively.
CONCLUSIONS: Artificial intelligence-based stethoscope is capable of diagnosing left-sided VHD accurately and may make routine screening for VHD more practical.