利用智能手机采集心音分析的肥厚型心肌病合并射血分数保留心衰的辅助筛查

董先鹏; 孟祥彬; 张阔; 房冠辰; 盖威蒿; 王文尧; 汪京嘉; 高峻; 潘俊君; 唐振超; 宋震

doi:10.11999/JEIT250830

利用智能手机采集心音分析的肥厚型心肌病合并射血分数保留心衰的辅助筛查

doi: 10.11999/JEIT250830 cstr: 32379.14.JEIT250830

董先鹏¹,
孟祥彬²,
张阔³,
房冠辰⁴,
盖威蒿⁵,
王文尧⁶,
汪京嘉⁶,
高峻⁶,
潘俊君⁷,
唐振超^{1, 8, 9},
宋震^{1, 2, ,}

1.
北京航空航天大学生物与医学工程学院北京 100191
2.
鹏城实验室网络智能研究部深圳 518055
3.
中国医学科学院阜外医院心血管内科北京 100037
4.
北京大学软件与微电子学院北京 102600
5.
中国科学院自动化研究所分子影像重点实验室北京 100190
6.
北京大学第三医院心血管内科北京 100191
7.
北京航空航天大学计算机学院北京 100191
8.
北京航空航天大学医学科学与工程学院北京 100191
9.
北京航空航天大学大数据精准医疗工业和信息化部重点实验室北京 100191

基金项目: 国家自然科学基金(32571731, 82172067, 82272068)，北京市自然科学基金(L256013)，中国科协青年人才托举工程(2023QNRC001)

详细信息

作者简介:
董先鹏：男，硕士生，研究方向为医学人工智能等

孟祥彬：男，研究员，研究方向为医疗多模态大型语言模型等

张阔：男，主治医师，研究方向为心血管代谢疾病及治疗等

房冠辰：男，硕士生，研究方向为生物医学影像处理等

盖威蒿：男，博士生，研究方向为智能生物医学影像等

王文尧：男，副研究员，研究方向为心血管代谢疾病及治疗等

汪京嘉：男，主治医师，研究方向为心血管代谢失稳及防治等

高峻：男，住院医师，研究方向为心力衰竭、心肌病诊疗等

潘俊君：男，教授，研究方向为虚拟手术、医学可视化等

唐振超：男，副教授，研究方向为心血管影像智能分析等

宋震：男，副研究员，研究方向为高性能计算与模型优化等

通讯作者:
宋震　songzh01@pcl.ac.cn

中图分类号: TN911.7; R318; TP18
计量
- 文章访问数: 88
- HTML全文浏览量: 27
- PDF下载量: 16
- 被引次数: 0
出版历程
- 收稿日期: 2025-09-01
- 修回日期: 2025-12-16
- 录用日期: 2025-12-17
- 网络出版日期: 2025-12-25

Auxiliary Screening for Hypertrophic Cardiomyopathy With Heart Failure with Preserved Ejection Fraction Utilizing Smartphone-Acquired Heart Sound Analysis

1.
School of Biological Science and Medical Engineering, Beihang University, Beijing 100191, China
2.
Department of Network Intelligence, Peng Cheng Laboratory, Shenzhen 518055, China
3.
Department of Cardiovascular Medicine, Fuwai Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing 100037, China
4.
School of Software and Microelectronics, Peking University, Beijing 102600, China
5.
Key Laboratory of Molecular Imaging, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
6.
Department of Cardiology and Institute of Vascular Medicine, Peking University Third Hospital, Beijing 100191, China
7.
School of Computer Science and Engineering, Beihang University, Beijing 100191, China
8.
School of Engineering Medicine, Beihang University, Beijing 100191, China
9.
Key Laboratory of Big Data-Based Precision Medicine, Ministry of Industry and Information Technology, Beihang University, Beijing 100191, China

Funds: The National Natural Science Foundation of China(32571731, 82172067, 82272068), Beijing Municipal Natural Science Foundation(L256013), The Young Elite Scientists Sponsorship Program by CAST (2023QNRC001)

摘要

摘要: 射血分数保留心衰(HFpEF)是一种高度异质性的临床综合征，在肥厚型心肌病患者(HCM)中较为常见。由于其诊断流程复杂，开展初步筛查与早期识别具有重要意义。对此，该文基于患者智能手机所采集的心音信号，提取了梅尔频率倒谱系数和短时傅里叶变换时频谱特征，并基于此分别构建了支持向量机与卷积神经网络两个基分类器。随后将两者预测概率作为新特征，构建并训练了以逻辑回归为元分类器的集成学习模型，用于HCM合并HFpEF的识别。结果显示，集成模型在测试集的曲线下面积(AUC)达到了0.900，准确率、灵敏度和特异度分别达到了0.813, 0.768和0.854，有效提升了预测性能。结果表明，该文设计的分类模型可以基于智能手机采集的心音实现HCM合并HFpEF的高效识别，有望用于HCM患者对自身病情的动态监测和HFpEF初步筛查，从而缩短诊断延迟。
- 心音信号分析 /
- 人工智能 /
- 肥厚型心肌病 /
- 射血分数保留心衰 /
- 辅助筛查
Abstract: Objective Heart Failure with preserved Ejection Fraction (HFpEF) is highly prevalent among patients with Hypertrophic CardioMyopathy (HCM), and early identification is critical for improving disease management. However, early screening for HFpEF remains challenging because symptoms are non-specific, diagnostic procedures are complex, and follow-up costs are high. Smartphones, owing to their wide accessibility, low cost, and portability, provide a feasible means to support heart sound-based screening. In this study, smartphone-acquired heart sounds from patients with HCM are used to develop and train an ensemble learning classification model for early detection and dynamic self-monitoring of HFpEF in the HCM population. Methods The proposed HFpEF screening framework consists of three components: preprocessing, feature extraction, and model training and fusion based on ensemble learning (Fig. 1). During preprocessing, smartphone-acquired heart sounds are subjected to bandpass filtering and wavelet denoising to improve signal quality, followed by segmentation into individual cardiac cycles. For feature extraction, Mel-Frequency Cepstral Coefficients (MFCCs) and Short-Time Fourier Transform (STFT) time-frequency spectra are calculated (Fig. 3). For classification, a stacking ensemble strategy is applied. Base learners, including a Support Vector Machine (SVM) and a Convolutional Neural Network (CNN), are trained, and their predicted probabilities are combined to construct a new feature space. A Logistic Regression (LR) meta-learner is then trained on this feature space to identify HFpEF in patients with HCM. Results and Discussions The classification performance of the three models is evaluated using the same patient-level independent test set. The SVM base learner achieves an Area Under the Curve (AUC) of 0.800, with an accuracy of 0.766, sensitivity of 0.659, and specificity of 0.865 (Table 5). The CNN base learner attains an AUC of 0.850, with an accuracy of 0.789, sensitivity of 0.622, and specificity of 0.944 (Table 5). By comparison, the ensemble-based LR classifier demonstrates superior performance, reaching an AUC of 0.900, with an accuracy of 0.813, sensitivity of 0.768, and specificity of 0.854 (Table 5). Relative to the base learners, the ensemble model exhibits a significant overall performance improvement after probability-based feature fusion (Fig. 5). Compared with existing clinical HFpEF risk scores, the proposed method shows higher predictive performance and stronger dynamic monitoring capability, supporting its suitability for risk stratification and follow-up warning in home settings. Compared with professional heart sound acquisition devices, the smartphone-acquired approach provides greater accessibility and cost efficiency, supporting its application in auxiliary HFpEF screening for high-risk HCM populations. Conclusions The challenges of clinical HFpEF screening in patients with HCM are addressed by proposing a smartphone-acquired heart sound analysis approach combined with an ensemble learning prediction model, resulting in an accessible and easily implemented auxiliary screening pipeline. The effectiveness of smartphone-based heart sound analysis for initial HFpEF screening in patients with HCM is validated, demonstrating its feasibility as an economical auxiliary tool for early HFpEF detection. This approach provides a non-invasive, convenient, and efficient screening strategy for patients with HCM complicated by HFpEF.
- Heart sound signal analysis /
- Artificial Intelligence (AI) /
- Hypertrophic CardioMyopathy (HCM) /
- Heart Failure with preserved Ejection Fraction (HFpEF) /
- Auxiliary screening

HTML全文

图 1 HCM合并HFpEF辅助筛查流程图

下载: 全尺寸图片幻灯片

图 2 智能手机心音采集位置示意图

下载: 全尺寸图片幻灯片

图 3 心音信号的时域波形及其对应时频谱

下载: 全尺寸图片幻灯片

图 4 CNN分类器的模型结构图

下载: 全尺寸图片幻灯片

图 5 不同模型在测试集的混淆矩阵及ROC曲线

下载: 全尺寸图片幻灯片

表 1 被纳入HCM患者群体的基本信息和临床特征

患者类别	数量	年龄(岁)	心率(bpm)	LVEF(%)	E/e’
不患HFpEF	27	40 ± 17	73.5 ± 16.6	66.5 ± 7.2	7.4 ± 2.3
患有HFpEF	29	52 ± 15	67.8 ± 9.7	73.8 ± 7.2	12.9 ± 5.0

下载: 导出CSV

表 2 SVM分类器数据集划分详情

类别	训练集		测试集
类别	患者数	音频数	患者数	音频数
患有HFpEF	24	337	5	82
不患HFpEF	22	333	5	89
合计	46	670	10	171

下载: 导出CSV

表 3 卷积块的具体结构细节

网络层	序号	结构组成	核大小	核数量	步长	参数量	输出大小
卷积块1	1	一维卷积	5	64	1	20544	×1
	2	批标准化	-	-	-	128	×1
	3	ReLU	-	-	-	-	×1
卷积块2	4	一维卷积	5	64	1	20544	×1
	5	批标准化	-	-	-	128	×1
	6	ReLU	-	-	-	-	×1
卷积块3	7	一维卷积	5	64	1	20544	×1
	8	批标准化	-	-	-	128	×1
	9	ReLU	-	-	-	-	×1
卷积块4	10	一维卷积	5	64	2	20544	×0.5
	11	批标准化	-	-	-	128	×0.5
	12	ReLU	-	-	-	-	×0.5

下载: 导出CSV

表 4 CNN分类器数据集划分详情

类别	训练集		验证集		测试集
类别	患者数	音频数	患者数	音频数	患者数	音频数
患有HFpEF	18	254	6	83	5	82
不患HFpEF	17	253	5	80	5	89
合计	35	506	11	164	10	171

下载: 导出CSV

表 5 不同模型在测试集的预测性能对比(%)

分类模型	AUC	ACC	SEN	SPE
SVM	0.800	0.766	0.659	0.865
CNN	0.850	0.789	0.622	0.944
集成模型	0.900	0.813	0.768	0.854

下载: 导出CSV

参考文献(31)

[1]	ZHAO Mengya, HE Xianzhen, MIN Xinwen, et al. Recent clinical updates of hypertrophic cardiomyopathy and future therapeutic strategies[J]. Reviews in Cardiovascular Medicine, 2025, 26(2): 25132. doi: 10.31083/RCM25132.
[2]	KRITTANAWONG C, BRITT W M, RIZWAN A, et al. Clinical update in heart failure with preserved ejection fraction[J]. Current Heart Failure Reports, 2024, 21(5): 461–484. doi: 10.1007/s11897-024-00679-5.
[3]	LIU Jie, WANG Dong, RUAN Jieyun, et al. Identification of heart failure with preserved ejection fraction helps risk stratification for hypertrophic cardiomyopathy[J]. BMC Medicine, 2022, 20(1): 21. doi: 10.1186/s12916-021-02219-7.
[4]	CHEN Qinfen, HU Jiandong, HU Jie, et al. Clinical characteristics and prognosis of patients with hypertrophic cardiomyopathy and heart failure with preserved ejection fraction[J]. Clinical Research in Cardiology, 2024, 113(5): 761–769. doi: 10.1007/s00392-023-02371-5.
[5]	YANG Yang, GUO Xingming, WANG Hui, et al. Deep learning-based heart sound analysis for left ventricular diastolic dysfunction diagnosis[J]. Diagnostics, 2021, 11(12): 2349. doi: 10.3390/diagnostics11122349.
[6]	HUIS IN ’T VELD A E, DE MAN F S, VAN ROSSUM A C, et al. How to diagnose heart failure with preserved ejection fraction: The value of invasive stress testing[J]. Netherlands Heart Journal, 2016, 24(4): 244–251. doi: 10.1007/s12471-016-0811-0.
[7]	MICHAUD M, MAURIN V, SIMON M, et al. Patients with high left ventricular filling pressure may be missed applying 2016 echo guidelines: A pilot study[J]. The International Journal of Cardiovascular Imaging, 2019, 35(12): 2157–2166. doi: 10.1007/s10554-019-01667-w.
[8]	BANTHIYA S, CHECK L, and ATKINS J. Hypertrophic cardiomyopathy as a form of heart failure with preserved ejection fraction: Diagnosis, drugs, and procedures[J]. US Cardiology Review, 2024, 18: e17. doi: 10.15420/usc.2023.21.
[9]	中华医学会心血管病学分会, 中国医师协会心血管内科医师分会, 中国医师协会心力衰竭专业委员会, 等. 中国心力衰竭诊断和治疗指南2024[J]. 中华心血管病杂志, 2024, 52(3): 235–275. doi: 10.3760/cma.j.cn112148-20231101-00405. Chinese Society of Cardiology, Chinese Medical Association, Chinese College of Cardiovascular Physician, Chinese Heart Failure Association of Chinese Medical Doctor Association, et al. Chinese guidelines for the diagnosis and treatment of heart failure 2024[J]. Chinese Journal of Cardiology, 2024, 52(3): 235–275. doi: 10.3760/cma.j.cn112148-20231101-00405.
[10]	CAMPBELL P, RUTTEN F H, LEE M M, et al. Heart failure with preserved ejection fraction: Everything the clinician needs to know[J]. The Lancet, 2024, 403(10431): 1083–1092. doi: 10.1016/S0140-6736(23)02756-3.
[11]	GHARAGOZLOO K, MEHDIZADEH M, HECKMAN G, et al. Heart failure with preserved ejection fraction in the elderly population: Basic mechanisms and clinical considerations[J]. Canadian Journal of Cardiology, 2024, 40(8): 1424–1444. doi: 10.1016/j.cjca.2024.04.006.
[12]	LUO Hongxing, WEERTS J, BEKKERS A, et al. Association between phonocardiography and echocardiography in heart failure patients with preserved ejection fraction[J]. European Heart Journal - Digital Health, 2023, 4(1): 4–11. doi: 10.1093/ehjdh/ztac073.
[13]	ALTSTIDL J M, ALTSTIDL T R, ANNEKEN L, et al. Detection of aortic stenosis using built-in microphones of commercially available smartphones[J]. European Heart Journal, 2023, 44(S2): ehad655.2983. doi: 10.1093/eurheartj/ehad655.2983.
[14]	REN Huiying, QIN Qirong, DONG Quanbin, et al. Smartphone-measured heart sounds for atrial fibrillation screening in community populations[J]. Circulation, 2025, 152(4): 283–285. doi: 10.1161/CIRCULATIONAHA.125.073828.
[15]	KANG S H, JOE B, YOON Y, et al. Cardiac auscultation using smartphones: Pilot study[J]. JMIR mHealth and uHealth, 2018, 6(2): e49. doi: 10.2196/mhealth.8946.
[16]	LUO Hongxing, LAMATA P, BAZIN S, et al. Smartphone as an electronic stethoscope: Factors influencing heart sound quality[J]. European Heart Journal - Digital Health, 2022, 3(3): 473–480. doi: 10.1093/ehjdh/ztac044.
[17]	LI Yun, ZHAO Zhanjiang, AINIWAER A, et al. Smartphone for heart sound measurement in hospital: Feasibility and influencing factors[J]. European Heart Journal - Digital Health, 2025, 6(3): 486–495. doi: 10.1093/ehjdh/ztaf007.
[18]	GARDEZI S K M, MYERSON S G, CHAMBERS J, et al. Cardiac auscultation poorly predicts the presence of valvular heart disease in asymptomatic primary care patients[J]. Heart, 2018, 104(22): 1832–1835. doi: 10.1136/heartjnl-2018-313082.
[19]	CHOU J C, YEH C C, and LEE H Y. One-shot voice conversion by separating speaker and content representations with instance normalization[J]. arXiv preprint arXiv: 1904.05742, 2019.
[20]	MIENYE I D and SUN Yanxia. A survey of ensemble learning: Concepts, algorithms, applications, and prospects[J]. IEEE Access, 2022, 10: 99129–99149. doi: 10.1109/ACCESS.2022.3207287.
[21]	射血分数保留的心力衰竭诊断与治疗中国专家共识制定工作组. 射血分数保留的心力衰竭诊断与治疗中国专家共识2023[J]. 中国循环杂志, 2023, 38(4): 375–393. doi: 10.3969/j.issn.1000-3614.2023.04.001. Chinese Expert Consensus Working Group on Diagnosis and Treatment of Heart Failure With Preserved Ejection Fraction. Diagnosis and treatment of heart failure with preserved ejection fraction: Chinese expert consensus 2023[J]. Chinese Circulation Journal, 2023, 38(4): 375–393. doi: 10.3969/j.issn.1000-3614.2023.04.001.
[22]	PAULUS W J. H₂FPEF score: At last, a properly validated diagnostic algorithm for heart failure with preserved ejection fraction[J]. Circulation, 2018, 138(9): 871–873. doi: 10.1161/CIRCULATIONAHA.118.035711.
[23]	PIESKE B, TSCHÖPE C, DE BOER R A, et al. How to diagnose heart failure with preserved ejection fraction: The HFA-PEFF diagnostic algorithm: A consensus recommendation from the Heart Failure Association (HFA) of the European Society of Cardiology (ESC)[J]. European Heart Journal, 2019, 40(40): 3297–3317. doi: 10.1093/eurheartj/ehz641.
[24]	贾晓艳, 刘离香, 王东伟, 等. H₂FPEF和HFA-PEFF评分在我国射血分数保留心力衰竭及射血分数保留心力衰竭合并心房颤动患者中的适用性分析[J]. 中国医学科学院学报, 2024, 46(2): 154–160. doi: 10.3881/j.issn.1000-503X.15826. JIA Xiaoyan, LIU Lixiang, WANG Dongwei, et al. Applicability of H₂FPEF and HFA-PEFF scores in Chinese patients suffering from heart failure with preserved ejection fraction and heart failure with preserved ejection fraction complicated with atrial fibrillation[J]. Acta Academiae Medicinae Sinicae, 2024, 46(2): 154–160. doi: 10.3881/j.issn.1000-503X.15826.
[25]	AKERMAN A P, AL-ROUB N, ANGELL-JAMES C, et al. External validation of artificial intelligence for detection of heart failure with preserved ejection fraction[J]. Nature Communications, 2025, 16(1): 2915. doi: 10.1038/s41467-025-58283-7.
[26]	REDDY Y N V, CARTER R E, SUNDARAM V, et al. An evidence-based screening tool for heart failure with preserved ejection fraction: The HFpEF-ABA score[J]. Nature Medicine, 2024, 30(8): 2258–2264. doi: 10.1038/s41591-024-03140-1.
[27]	LAENENS D, ZEGKOS T, KAMPERIDIS V, et al. Heart failure risk assessment in patients with hypertrophic cardiomyopathy based on the H₂FPEF score[J]. European Journal of Heart Failure, 2024, 26(10): 2173–2182. doi: 10.1002/ejhf.3413.
[28]	GAO Yipeng, LIU Hongyun, BI Xiaojun, et al. H₂FPEF and HFA-PEFF scores for heart failure risk stratification in hypertrophic cardiomyopathy patients[J]. ESC Heart Failure, 2025, 12(3): 2225–2238. doi: 10.1002/ehf2.15247.
[29]	LIU Yongmin, GUO Xingming, and ZHENG Yineng. An automatic approach using ELM classifier for HFpEF identification based on heart sound characteristics[J]. Journal of Medical Systems, 2019, 43(9): 285. doi: 10.1007/s10916-019-1415-1.
[30]	ZHENG Yineng, GUO Xingming, YANG Yang, et al. Phonocardiogram transfer learning-based CatBoost model for diastolic dysfunction identification using multiple domain-specific deep feature fusion[J]. Computers in Biology and Medicine, 2023, 156: 106707. doi: 10.1016/j.compbiomed.2023.106707.
[31]	ZHAO Qinghao, GENG Shijia, WANG Boya, et al. Deep learning in heart sound analysis: From techniques to clinical applications[J]. Health Data Science, 2024, 4: 0182. doi: 10.34133/hds.0182.