人工智能辅助心音识别先天性心脏病的研究进展

doi:10.12372/jcp.2026.25e0866

摘要/Abstract

摘要：

先天性心脏病是威胁儿童健康的重要出生缺陷，早期筛查对降低病死率至关重要。心音听诊是识别先天性心脏病的重要方式，但其筛查效能易受医师临床经验影响。近年来，以深度学习为代表的人工智能技术在心脏杂音识别和疾病诊断中表现出较高的准确性，可辅助提升临床医师的诊断水平，助力先天性心脏病实现早发现、早诊断。当前人工智能模型仍存在数据质量、算法性能与临床验证等诸多挑战，但其应用潜力突出，仍是未来先天性心脏病筛查体系中的重要组成部分。本文综述了现阶段人工智能辅助心音识别先天性心脏病的发展现状，梳理算法进展以及其临床应用情况，并归纳总结现存挑战，以期推动其从实验向实践转化。

关键词: 人工智能, 先天性心脏病, 心音听诊, 深度学习, 疾病筛查

Abstract:

Congenital heart disease (CHD) is a major birth defect threatening children's health, and early screening is crucial for reducing mortality. Cardiac auscultation is an important method for identifying CHD, yet its screening efficacy is susceptible to clinicians' practical experience. In recent years, artificial intelligence (AI) technologies represented by deep learning have achieved high accuracy in heart murmurs detection and disease diagnosis. They can help improve clinicians' diagnostic capabilities and facilitate the early detection and diagnosis of CHD. Current AI models still face challenges in terms of data quality, algorithm performance and clinical validation. Nevertheless, they possess great application potential and will remain an essential part of the screening system for CHD. This paper reviews the current development of AI-assisted heart sound recognition for CHD, summarizes the progress of relevant algorithms and clinical applications, and concludes the existing challenges. It aims to promote the transformation of this technology from experimental research to clinical practice.

Key words: artificial intelligence, congenital heart disease, cardiac auscultation, deep learning, disease screening

中图分类号:

赵鎏丹, 周欣, 孙锟. 人工智能辅助心音识别先天性心脏病的研究进展[J]. 临床儿科杂志, 2026, 44(7): 665-672.

ZHAO Liudan, ZHOU Xin, SUN Kun. Research progress on artificial intelligence-assisted heart sound recognition for congenital heart disease[J]. Journal of Clinical Pediatrics, 2026, 44(7): 665-672.

图/表 1

参考文献 47

[1]	GBD 2019 Diseases and Injuries Collaborators. Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019[J/OL]. Lancet, 2020, 396(10258): 1204-1222. https://doi.org/10.1016/S0140-6736(20)30925-9.
[2]	Bai Z, Han J, An J, et al. The global, regional, and national patterns of change in the burden of congenital birth defects, 1990-2021:an analysis of the global burden of disease study 2021 and forecast to 2040[J/OL]. EClinicalMedicine, 2024, 77: 102873. https://doi.org/10.1016/j.eclinm.2024.102873.
[3]	Pan F, Li J, Lou H, et al. Geographical and socioeconomic factors influence the birth prevalence of congenital heart disease:a population-based cross-sectional study in eastern China[J/OL]. Curr Probl Cardiol, 2022, 47(11): 101341. https://doi.org/10.1016/j.cpcardiol.2022.101341.
[4]	Yan H, Zhai B, Feng R, et al. Prevalence of congenital heart disease in Chinese children with different birth weights and its relationship to the neonatal birth weight[J/OL]. Front Pediatr, 2022, 10: 828300. https://doi.org/10.3389/fped.2022.828300.
[5]	Zhao QM, Liu F, Wu L, et al. Prevalence of congenital heart disease at live birth in China[J/OL]. J Pediatr, 2019, 204: 53-58. https://doi.org/10.1016/j.jpeds.2018.08.040.
[6]	Zhang Y, Wang J, Zhao J, et al. Current status and challenges in prenatal and neonatal screening, diagnosis, and management of congenital heart disease in China[J/OL]. The Lancet Child & Adolescent Health, 2023, 7(7): 479-489. https://doi.org/10.1016/S2352-4642(23)00176-1.
[7]	Ho TC, Ouyang H, Lu Y, et al. Postprocedural outcomes of rural children undergoing correction of congenital heart lesions in Yunnan Province, China[J/OL]. Pediatr Cardiol, 2011, 32(6): 811-814. https://doi.org/10.1007/s00246-011-9991-y.
[8]	Hu XJ, Ma XJ, Zhao QM, et al. Pulse oximetry and auscultation for congenital heart disease detection[J/OL]. Pediatrics, 2017, 140(4): e20171154. https://doi.org/10.1542/peds.2017-1154.
[9]	Zhao QM, Ma XJ, Ge XL, et al. Pulse oximetry with clinical assessment to screen for congenital heart disease in neonates in China: a prospective study[J/OL]. Lancet, 2014, 384(9945): 747-754. https://doi.org/10.1016/S0140-6736(14)60198-7.
[10]	Geva T, Hegesh J, Frand M. Reappraisal of the approach to the child with heart murmurs: is echocardiography mandatory?[J/OL]. Int J Cardiol, 1988, 19(1): 107-113. https://doi.org/10.1016/0167-5273(88)90196-9.
[11]	Mackie AS, Jutras LC, Dancea AB, et al. Can cardiologists distinguish innocent from pathologic murmurs in neonates?[J/OL]. J Pediatr, 2009, 154(1): 50-54. https://doi.org/10.1016/j.jpeds.2008.06.017.
[12]	Griebsch I, Knowles RL, Brown J, et al. Comparing the clinical and economic effects of clinical examination, pulse oximetry, and echocardiography in newborn screening for congenital heart defects: a probabilistic cost-effectiveness model and value of information analysis[J/OL]. Int J Technol Assess Health Care, 2007, 23(2): 192-204. https://doi.org/10.1017/S0266462307070304.
[13]	Singh Y, Chen SE. Impact of pulse oximetry screening to detect congenital heart defects:5 years' experience in a UK regional neonatal unit[J/OL]. Eur J Pediatr, 2022, 181(2): 813-821. https://doi.org/10.1007/s00431-021-04275-w.
[14]	Lee C, Rankin KN, Zuo KJ, et al. Computer-aided auscultation of murmurs in children: evaluation of commercially available software[J/OL]. Cardiol Young, 2016, 26(7): 1359-1364. https://doi.org/0.1017/S1047951115002656.
[15]	Brunetti ND, Rosania S, D'antuono C, et al. Diagnostic accuracy of heart murmur in newborns with suspected congenital heart disease[J/OL]. J Cardiovasc Med (Hagerstown), 2015, 16(8): 556-561. https://doi.org/10.2459/JCM.0b013e3283649953.
[16]	Gokmen Z, Tunaoglu FS, Kula S, et al. Comparison of initial evaluation of neonatal heart murmurs by pediatrician and pediatric cardiologist[J/OL]. J Matern Fetal Neonatal Med, 2009, 22(11): 1086-1091. https://doi.org/10.3109/14767050903009230.
[17]	Gaskin PR, Owens SE, Tainer NS, et al. Clinical auscultation skills in pediatric residents[J/OL]. Pediatrics, 2000, 105(6): 1184-1187. https://doi.org/10.1542/peds.105.6.1184.
[18]	Prince J, Maidens J, Kieu S, et al. Deep learning algorithms to detect murmurs associated with structural heart disease[J/OL]. J Am Heart Assoc, 2023, 12(20): e030377. https://doi.org/10.1161/JAHA.123.030377.
[19]	Papunen I, Ylanen K, Lundqvist O, et al. Automated analysis of heart sound signals in screening for structural heart disease in children[J/OL]. Eur J Pediatr, 2024, 183(11): 4951-4958. https://doi.org/10.1007/s00431-024-05773-3.
[20]	Zhao Q, Geng S, Wang B, et al. Deep learning in heart sound analysis: from techniques to clinical applications[J/OL]. Health Data Sci, 2024, 4: 0182. https://doi.org/10.34133/hds.0182.
[21]	Ren Z, Chang Y, Nguyen TT, et al. A comprehensive survey on heart sound analysis in the deep learning era[J/OL]. IEEE Computational Intelligence Magazine, 2024, 19(3): 42-57. https://doi.org/10.1109/MCI.2024.3401309.
[22]	Noman F, Salleh SH, Ting CM, et al. A markov-switching model approach to heart sound segmentation and classification[J/OL]. IEEE J Biomed Health Inform, 2020, 24(3): 705-716. https://doi.org/10.1109/JBHI.2019.2925036.
[23]	Ghosh SK, Tripathy RK, Ponnalagu RN. Evaluation of performance metrics and denoising of PCG signal using wavelet based decomposition: proceedings of the 2020 IEEE 17th India council international conference (INDICON)[C/OL]. New Delhi, 2020. https://doi.org/10.1109/INDICON49873.2020.9342464.
[24]	Sharan TS, Bhattacharjee R, Sharma S, et al. Evaluation of deep learning methods (DnCNN and U-Net) for denoising of heart auscultation signals: proceedings of the 2020 3rd international conference on communication system, computing and IT applications (CSCITA)[C/OL]. Mumbai, 2020. https://doi.org/10.1109/CSCITA47329.2020.9137813.
[25]	Ali SN, Shuvo SB, Al-manzo MIS, et al. An end-to-end deep learning framework for real-time denoising of heart sounds for cardiac disease detection in unseen noise[J/OL]. IEEE Access, 2023, 11: 87887-87901. https://doi.org/10.1109/ACCESS.2023.3292551. doi: 10.1109/ACCESS.2023.3292551
[26]	Thompson WR, Reinisch AJ, Unterberger MJ, et al. Artificial intelligence-assisted auscultation of heart murmurs: validation by virtual clinical trial[J/OL]. Pediatr Cardiol, 2019, 40(3): 623-629. https://doi.org/10.1007/s00246-018-2036-z.
[27]	Messner E, ZÖhrer M, Pernkopf F. Heart sound segmentation—an event detection approach using deep recurrent neural networks[J/OL]. IEEE Transactions on Biomedical Engineering, 2018, 65(9): 1964-1974. https://doi.org/10.1109/TBME.2018.2843258. doi: 10.1109/TBME.10
[28]	Bourouhou A, Jilbab A, Nacir C, et al. Heart sound signals segmentation and multiclass classification[J/OL]. International Journal of Online and Biomedical Engineering, 2020, 16(15): 64-79. https://doi.org/10.3991/ijoe.v16i15.16817.
[29]	Tang H, Wang M, Hu Y, et al. Automated signal quality assessment for heart sound signal by novel features and evaluation in open public datasets[J/OL]. Biomed Res Int, 2021, 2021: 7565398. https://doi.org/10.1155/2021/7565398.
[30]	Wang JK, Chang YF, Tsai KH, et al. Automatic recognition of murmurs of ventricular septal defect using convolutional recurrent neural networks with temporal attentive pooling[J/OL]. Sci Rep, 2020, 10(1): 21797. https://doi.org/10.1038/s41598-020-77994-z.
[31]	Yang C, Hu N, Xu D, et al. Monaural cardiopulmonary sound separation via complex-valued deep autoencoder and cyclostationarity[J/OL]. Biomed Phys Eng Express, 2023, 9(3): 035002. https://doi.org/10.1088/2057-1976/acbc7f.
[32]	Makimoto H, Shiraga T, Kohlmann B, et al. Efficient screening for severe aortic valve stenosis using understandable artificial intelligence: a prospective diagnostic accuracy study[J/OL]. Eur Heart J Digit Health, 2022, 3(2): 141-152. https://doi.org/10.1093/ehjdh/ztac029.
[33]	Gradolewski D, Magenes G, Johansson S, et al. A wavelet transform-based neural network denoising algorithm for mobile phonocardiography[J/OL]. Sensors (Basel), 2019, 19(4): 957. https://doi.org/10.3390/s19040957.
[34]	Shuvo SB, Ali SN, Swapnil SI, et al. CardioXNet: a novel lightweight deep learning framework for cardiovascular disease classification using heart sound recordings[J/OL]. IEEE Access, 2021, 9: 36955-36967. https://doi.org/10.1109/ACCESS.2021.3063129. doi: 10.1109/Access.6287639
[35]	Alqudah AM. Towards classifying non-segmented heart sound records using instantaneous frequency based features[J/OL]. Journal of Medical Engineering & Technology, 2019, 43(7): 418-430. https://doi.org/10.1080/03091902.2019.1688408.
[36]	Tuncer T, Dogan S, Tan R-S, et al. Application of petersen graph pattern technique for automated detection of heart valve diseases with PCG signals[J/OL]. Information Sciences, 2021, 565: 91-104. https://doi.org/10.1016/j.ins.2021.01.088.
[37]	Zhang W, Hang J, Deng S. Heart sound classification based on scaled spectrogram and tensor decomposition[J/OL]. Expert Systems with Applications, 2017, 84: 220-231. https://doi.org/10.1016/j.eswa.2017.05.014.
[38]	Renna F, Oliveira J, Coimbra MT. Deep convolutional neural networks for heart sound segmentation[J/OL]. IEEE J Biomed Health Inform, 2019, 23(6): 2435-2445. https://doi.org/10.1109/JBHI.2019.2894222.
[39]	Fan T, Zhu J, Cheng Y, et al. A new direct heart sound segmentation approach using bi-directional GRU: proceedings of the 2018 24th international conference on automation and computing (ICAC)[C/OL]. Newcastle Upon Tyne, 2018. https://doi.org/10.23919/IConAC.2018.8749010.
[40]	Chen Y, Sun Y, Lv J, et al. End-to-end heart sound segmentation using deep convolutional recurrent network[J/OL]. Complex & Intelligent Systems, 2021, 7: 2103-2117. https://doi.org/10.1007/S40747-021-00325-W.
[41]	Wang Y, Yang X, Qian X, et al. Assistive diagnostic technology for congenital heart disease based on fusion features and deep learning[J/OL]. Front Physiol, 2023, 14: 1310434. https://doi.org/10.3389/fphys.2023.1310434.
[42]	Koike T, Qian K, Kong Q, et al. Audio for audio is better? An investigation on transfer learning models for heart sound classification: proceedings of the 2020 42nd annual international conference of the IEEE engineering in medicine & biology society (EMBC)[C/OL]. Montreal, QC, 2020. https://doi.org/10.1109/EMBC44109.2020.9175450.
[43]	Guan T, Chen Z, Xu D, et al. LaCHeST: An AI-assisted auscultation tool for pediatric congenital heart diseases screening and validated via large-scale screening tasks[J/OL]. Biomedical Signal Processing and Control, 2025, 103: 107474. https://doi.org/10.1016/j.bs/c.2024.107474.
[44]	Lv J, Dong B, Lei H, et al. Artificial intelligence-assisted auscultation in detecting congenital heart disease[J/OL]. Eur Heart J Digit Health, 2021, 2(1): 119-124. https://doi.org/10.1016/j.bspc.2024.107474.
[45]	Zhou G, Chien C, Chen J, et al. Identifying pediatric heart murmurs and distinguishing innocent from pathologic using deep learning[J/OL]. Artif Intell Med, 2024, 153: 102867. https://doi.org/10.1016/j.artmed.2024.102867.
[46]	Liu J, Wang H, Yang Z, et al. Deep learning-based computer-aided heart sound analysis in children with left-to-right shunt congenital heart disease[J/OL]. Int J Cardiol, 2022, 348: 58-64. https://doi.org/10.1016/j.ijcard.2021.12.012.
[47]	Baghel N, Dutta MK, Burget R. Automatic diagnosis of multiple cardiac diseases from PCG signals using convolutional neural network[J/OL]. Comput Methods Programs Biomed, 2020, 197: 105750. https://doi.org/10.1016/j.cmpb.2020.105750.