1
|
Giordano N, Bolognini I, Knaflitz M, Rosati S, Balestra G. Stratification of Heart Sounds Morphology Through Unsupervised Learning. Stud Health Technol Inform 2024; 316:889-893. [PMID: 39176936 DOI: 10.3233/shti240555] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 08/24/2024]
Abstract
The use of heart sounds for the assessment of the hemodynamic condition of the heart in telemonitoring applications is object of wide research at date. Many different approaches have been tried out for the analysis of the first (S1) and second (S2) heart sounds, but their morphological interpretation is still to be explored: in fact, the sound morphology is not unique and this impact the separability of the heart sounds components with methods based on envelopes or model optimization. In this study, we propose a method to stratify S1 and S2 according to their morphology to explore their diversity and increase their morphological interpretability. The method we propose is based on unsupervised learning, which we obtain using the cascade of four Self-Organizing Maps (SOMs) of decreasing dimensions. When tested on a publicly available heart sounds dataset, the proposed clustering approach proved to be robust and consistent, with over 80% of the heartbeats of the same patient being clustered together. The identified heart sounds templates highlight differences in the time and energy domains which may open to new directions of analysis in the future.
Collapse
|
2
|
L�m HK, Wang JK, Tsai KS, Chien YH, Chang YC, Cheng CH, Tsai CY, Peng YW, Hwang JJ, Huei-Ming Ma M. Cardiac screening in school children: Combining auscultation and electrocardiography with a crowdsourcing model. J Formos Med Assoc 2023; 122:1313-1320. [PMID: 37468409 DOI: 10.1016/j.jfma.2023.07.002] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/31/2022] [Revised: 03/05/2023] [Accepted: 07/03/2023] [Indexed: 07/21/2023] Open
Abstract
BACKGROUND/PURPOSE School-based cardiac screening is useful for identifying children and adolescents with a high risk of sudden cardiac death. However, because of challenges associated with cost, distance, and human resources, cardiac screening is not widely implemented, especially in rural areas with limited medical resources. This study aims to establish a cloud-based system suitable for mass cardiac screening of schoolchildren in rural areas with limited medical resources. METHODS Students from three schools were included. They or their guardians completed a simple questionnaire, administered in paper or electronic form. Heart sounds were recorded using an electronic stethoscope. Twelve-lead electrocardiograms (ECGs) were recorded and digitalized. The signals were transmitted through Bluetooth to a tablet computer and then uploaded to a cloud server over Wi-Fi. Crowdsourced pediatric cardiologists reviewed those data from a web-based platform and provided remote consultation. In cases in which abnormal heart sounds or ECGs were noted, the students were referred to the hospital for further evaluation. RESULTS A total of 1004 students were enrolled in this study. Of the 138 students referred, 62 were diagnosed as having an abnormal heart condition and most had previously been undiagnosed. The interrater agreeability was high. CONCLUSION An innovative strategy combining a cloud-based cardiac screening system with remote consultation by crowdsourced experts was established. This system allows pediatric cardiologists to provide consultation and make reliable diagnoses. Combined with crowdsourcing, the system constitutes a viable approach for mass cardiac screening in children and adolescents living in rural areas with insufficient medical resources.
Collapse
|
3
|
Westphal P, Luo H, Shahmohammadi M, Prinzen FW, Delhaas T, Cornelussen RN. Machine learning-powered, device-embedded heart sound measurement can optimize AV delay in patients with CRT. Heart Rhythm 2023; 20:1316-1324. [PMID: 37247684 DOI: 10.1016/j.hrthm.2023.05.025] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 01/30/2023] [Revised: 04/28/2023] [Accepted: 05/17/2023] [Indexed: 05/31/2023]
Abstract
BACKGROUND Continuous optimization of atrioventricular (AV) delay for cardiac resynchronization therapy (CRT) is mainly performed by electrical means. OBJECTIVE The purpose of this study was to develop an estimation model of cardiac function that uses a piezoelectric microphone embedded in a pulse generator to guide CRT optimization. METHODS Electrocardiogram, left ventricular pressure (LVP), and heart sounds were simultaneously collected during CRT device implantation procedures. A piezoelectric alarm transducer embedded in a modified CRT device facilitated recording of heart sounds in patients undergoing a pacing protocol with different AV delays. Machine learning (ML) was used to produce a decision-tree ensemble model capable of estimating absolute maximal LVP (LVPmax) and maximal rise of LVP (LVdP/dtmax) using 3 heart sound-based features. To gauge the applicability of ML in AV delay optimization, polynomial curves were fitted to measured and estimated values. RESULTS In the data set of ∼30,000 heartbeats, ML indicated S1 amplitude, S2 amplitude, and S1 integral (S1 energy for LVdP/dtmax) as most prominent features for AV delay optimization. ML resulted in single-beat estimation precision for absolute values of LVPmax and LVdP/dtmax of 67% and 64%, respectively. For 20-30 beat averages, cross-correlation between measured and estimated LVPmax and LVdP/dtmax was 0.999 for both. The estimated optimal AV delays were not significantly different from those measured using invasive LVP (difference -5.6 ± 17.1 ms for LVPmax and +5.1 ± 6.7 ms for LVdP/dtmax). The difference in function at estimated and measured optimal AV delays was not statiscally significant (1 ± 3 mm Hg for LVPmax and 9 ± 57 mm Hg/s for LVdP/dtmax). CONCLUSION Heart sound sensors embedded in a CRT device, powered by a ML algorithm, provide a reliable assessment of optimal AV delays and absolute LVPmax and LVdP/dtmax.
Collapse
|
4
|
Phonocardiogram transfer learning-based CatBoost model for diastolic dysfunction identification using multiple domain-specific deep feature fusion. Comput Biol Med 2023; 156:106707. [PMID: 36871337 DOI: 10.1016/j.compbiomed.2023.106707] [Citation(s) in RCA: 4] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/17/2022] [Revised: 02/11/2023] [Accepted: 02/19/2023] [Indexed: 02/22/2023]
Abstract
Left ventricular diastolic dyfunction detection is particularly important in cardiac function screening. This paper proposed a phonocardiogram (PCG) transfer learning-based CatBoost model to detect diastolic dysfunction noninvasively. The Short-Time Fourier Transform (STFT), Mel Frequency Cepstral Coefficients (MFCCs), S-transform and gammatonegram were utilized to perform four different representations of spectrograms for learning the representative patterns of PCG signals in two-dimensional image modality. Then, four pre-trained convolutional neural networks (CNNs) such as VGG16, Xception, ResNet50 and InceptionResNetv2 were employed to extract multiple domain-specific deep features from PCG spectrograms using transfer learning, respectively. Further, principal component analysis and linear discriminant analysis (LDA) were applied to different feature subsets, respectively, and then these different selected features are fused and fed into CatBoost for classification and performance comparison. Finally, three typical machine learning classifiers such as multilayer perceptron, support vector machine and random forest were employed to compared with CatBoost. The hyperparameter optimization of the investigated models was determined through grid search. The visualized result of the global feature importance showed that deep features extracted from gammatonegram by ResNet50 contributed most to classification. Overall, the proposed multiple domain-specific feature fusion based CatBoost model with LDA achieved the best performance with an area under the curve of 0.911, accuracy of 0.882, sensitivity of 0.821, specificity of 0.927, F1-score of 0.892 on the testing set. The PCG transfer learning-based model developed in this study could aid in diastolic dysfunction detection and could contribute to non-invasive evaluation of diastolic function.
Collapse
|
5
|
Chen X, Guo X, Zheng Y, Lv C. Heart function grading evaluation based on heart sounds and convolutional neural networks. Phys Eng Sci Med 2023; 46:279-288. [PMID: 36625996 DOI: 10.1007/s13246-023-01216-9] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/16/2022] [Revised: 11/22/2022] [Accepted: 01/02/2023] [Indexed: 01/11/2023]
Abstract
Accurate and rapid cardiac function assessment is critical for disease diagnosis and treatment strategy. However, the current cardiac function assessment methods have their adaptability and limitations. Heart sounds (HS) can reflect changes in heart function. Therefore, HS signals were proposed to assess cardiac function, and a specially designed pruning convolutional neural network (CNN) was applied to recognize subjects' cardiac function at different levels in this paper. Firstly, the adaptive wavelet denoising algorithm and logistic regression based hidden semi-Markov model were utilized for signal denoising and segmentation. Then, the continuous wavelet transform (CWT) was employed to convert the preprocessed HS signals into spectra as input to the convolutional neural network, which can extract features automatically. Finally, the proposed method was compared with AlexNet, Resnet50, Xception, GhostNet and EfficientNet to verify the superiority of the proposed method. Through comprehensive comparison, the proposed approach achieves the best classification performance with an accuracy of 94.34%. The study indicates HS analysis is a non-invasive and effective method for cardiac function classification, which has broad research prospects.
Collapse
|
6
|
Schmidt SE, Madsen LH, Hansen J, Zimmermann H, Kelbæk H, Winter S, Hammershøi D, Toft E, Struijk JJ, Clemmensen P. Coronary Artery Disease Detected by Low Frequency Heart Sounds. Cardiovasc Eng Technol 2022; 13:864-871. [PMID: 35545751 DOI: 10.1007/s13239-022-00622-6] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 08/25/2021] [Accepted: 03/28/2022] [Indexed: 01/27/2023]
Abstract
OBJECTIVES Previous studies have observed an increase in low frequency diastolic heart sounds in patients with coronary artery disease (CAD). The aim was to develop and validate a diagnostic, computerized acoustic CAD-score based on heart sounds for the non-invasive detection of CAD. METHODS Prospective study enrolling 463 patients referred for elective coronary angiography. Pre-procedure non-invasive recordings of heart sounds were obtained using a novel acoustic sensor. A CAD-score was defined as the power ratio between the 10-90 Hz frequency spectrum and the 90-300 Hz frequency spectrum of the mid-diastolic heart sound. Quantitative coronary angiography analysis was performed by a blinded core laboratory and patients grouped according to the results: obstructive CAD defined by the presence of at least one ≥ 50% stenosis, non-obstructive CAD as patients with a maximal stenosis in the 25-50% interval and non-CAD as no coronary lesions exceeding 25%. We excluded patients with potential confounders or incomplete data (n = 245). To avoid over-fitting the final cohort of 218 patients was randomly divided into to a training group for development (n = 127) and a validation group (n = 91). RESULTS In both the training and the validation group the CAD-score was significantly increased in CAD patients compared to non-CAD patients (p < 0.0001). In the validation group the area under the receiver-operating curve was 77% (95% CI 63-91%). Sensitivity was 71% (95% CI 59-82%) and specificity 64% (95% CI 45-83%). CONCLUSION The acoustic CAD-score is a new, inexpensive, non-invasive method to detect CAD, which may supplement clinical risk stratification and reduce the need for subsequent non-invasive and invasive testing.
Collapse
|
7
|
Larsen BS, Winther S, Nissen L, Diederichsen A, Bøttcher M, Renker M, Struijk JJ, Christensen MG, Schmidt SE. Improved pre-test likelihood estimation of coronary artery disease using phonocardiography. EUROPEAN HEART JOURNAL. DIGITAL HEALTH 2022; 3:600-609. [PMID: 36710896 PMCID: PMC9779903 DOI: 10.1093/ehjdh/ztac057] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 04/13/2022] [Revised: 08/22/2022] [Accepted: 09/19/2022] [Indexed: 12/24/2022]
Abstract
Aims Current early risk stratification of coronary artery disease (CAD) consists of pre-test probability scoring such as the 2019 ESC guidelines on chronic coronary syndromes (ESC2019), which has low specificity and thus rule-out capacity. A newer clinical risk factor model (risk factor-weighted clinical likelihood, RF-CL) showed significantly improved rule-out capacity over the ESC2019 model. The aim of the current study was to investigate if the addition of acoustic features to the RF-CL model could improve the rule-out potential of the best performing clinical risk factor models. Methods and results Four studies with heart sound recordings from 2222 patients were pooled and distributed into two data sets: training and test. From a feature bank of 40 acoustic features, a forward-selection technique was used to select three features that were added to the RF-CL model. Using a cutoff of 5% predicted risk of CAD, the developed acoustic-weighted clinical likelihood (A-CL) model showed significantly (P < 0.05) higher specificity of 48.6% than the RF-CL model (specificity of 41.5%) and ESC 2019 model (specificity of 6.9%) while having the same sensitivity of 84.9% as the RF-CL model. Area under the curve of the receiver operating characteristic for the three models was 72.5% for ESC2019, 76.7% for RF-CL, and 79.5% for A-CL. Conclusion The proposed A-CL model offers significantly improved rule-out capacity over the ESC2019 model and showed better overall performance than the RF-CL model. The addition of acoustic features to the RF-CL model was shown to significantly improve early risk stratification of symptomatic patients suspected of having stable CAD.
Collapse
|
8
|
Zheng Y, Guo X, Wang Y, Qin J, Lv F. A multi-scale and multi-domain heart sound feature-based machine learning model for ACC/AHA heart failure stage classification. Physiol Meas 2022; 43. [PMID: 35512699 DOI: 10.1088/1361-6579/ac6d40] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/07/2022] [Accepted: 05/05/2022] [Indexed: 11/11/2022]
Abstract
OBJECTIVE Heart sounds can reflect detrimental changes in cardiac mechanical activity that are common pathological characteristics of chronic heart failure (CHF). The ACC/AHA heart failure (HF) stage classification is essential for clinical decision-making and the management of CHF. Herein, a machine learning model that makes use of multi-scale and multi-domain heart sound features was proposed to provide an objective aid for ACC/AHA HF stage classification. APPROACH A dataset containing phonocardiogram (PCG) signals from 275 subjects was obtained from two medical institutions and used in this study. Complementary ensemble empirical mode decomposition and tunable-Q wavelet transform were used to construct self-adaptive sub-sequences and multi-level sub-band signals for PCG signals. Time-domain, frequency-domain and nonlinear feature extraction were then applied to the original PCG signal, heart sound sub-sequences and sub-band signals to construct multi-scale and multi-domain heart sound features. The features selected via the least absolute shrinkage and selection operator were fed into a machine learning classifier for ACC/AHA HF stage classification. Finally, mainstream machine learning classifiers, including least-squares support vector machine (LS-SVM), deep belief network (DBN) and random forest (RF), were compared to determine the optimal model. MAIN RESULTS The results showed that the LS-SVM, which utilized a combination of multi-scale and multi-domain features, achieved better classification performance than the DBN and RF using multi-scale or multi-domain features alone or together, with average sensitivity, specificity, and accuracy of 0.821, 0.955 and 0.820 on the testing set, respectively. SIGNIFICANCE PCG signal analysis provides efficient measurement information regarding CHF severity and is a promising noninvasive method for ACC/AHA HF stage classification.
Collapse
|
9
|
Wang H, Guo X, Zheng Y, Yang Y. An automatic approach for heart failure typing based on heart sounds and convolutional recurrent neural networks. Phys Eng Sci Med 2022; 45:475-485. [PMID: 35347667 DOI: 10.1007/s13246-022-01112-8] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/08/2021] [Accepted: 02/18/2022] [Indexed: 11/26/2022]
Abstract
Heart failure (HF) is a complex clinical syndrome that poses a major hazard to human health. Patients with different types of HF have great differences in pathogenesis and treatment options. Therefore, HF typing is of great significance for timely treatment of patients. In this paper, we proposed an automatic approach for HF typing based on heart sounds (HS) and convolutional recurrent neural networks, which provides a new non-invasive and convenient way for HF typing. Firstly, the collected HS signals were preprocessed with adaptive wavelet denoising. Then, the logistic regression based hidden semi-Markov model was utilized to segment HS frames. For the distinction between normal subjects and the HF patients with preserved ejection fraction or reduced ejection fraction, a model based on convolutional neural network and recurrent neural network was built. The model can automatically learn the spatial and temporal characteristics of HS signals. The results show that the proposed model achieved a superior performance with an accuracy of 97.64%. This study suggests the proposed method could be a useful tool for HF recognition and as a supplement for HF typing.
Collapse
|
10
|
Liu J, Wang H, Yang Z, Quan J, Liu L, Tian J. Deep learning-based computer-aided heart sound analysis in children with left-to-right shunt congenital heart disease. Int J Cardiol 2021; 348:58-64. [PMID: 34902505 DOI: 10.1016/j.ijcard.2021.12.012] [Citation(s) in RCA: 11] [Impact Index Per Article: 3.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 07/03/2021] [Revised: 11/20/2021] [Accepted: 12/08/2021] [Indexed: 12/21/2022]
Abstract
OBJECTIVE The purpose of this study was to explore a new algorithm model capable of leverage deep learning to screen and diagnose specific types of left-to-right shunt congenital heart disease (CHD) in children. METHODS Using deep learning, screening models were constructed to identify 884 heart sound recordings from children with left-to-right shunt CHD. The most suitable model for each type was summarized and compared with expert auscultation. An exploratory analysis was conducted to assess whether there were correlations between heart sounds and left ventricular ejection fraction (LVEF), pulmonary artery pressure, and malformation size. RESULTS The residual convolution recurrent neural network (RCRnet) classification model had higher accuracy than other models with respect to atrial septal defect (ASD), ventricular septum defect (VSD), patent ductus arteriosus (PDA) and combined CHD, and the best auscultation sites were determined to be the 4th, 5th, 2nd and 3rd auscultation areas, respectively. The diagnostic results of this model were better than those derived from expert auscultation, with sensitivity values of 0.932-1.000, specificity values of 0.944-0.997, precision values of 0.888-0.997 and accuracy values of 0.940-0.994. Absolute Pearson correlation coefficient values between heart sounds of the four types of CHD and LVEF, right ventricular systolic pressure (RVSP) and malformation size were all less than 0.3. CONCLUSIONS The RCRnet model can preliminarily determine types of left-to-right shunt CHD and improve diagnostic efficiency, which may provide a new choice algorithmic CHD screening in children.
Collapse
|
11
|
Kagaya Y, Tabata M, Arata Y, Kameoka J, Ishii S. Employment of color Doppler echocardiographic video clips in a cardiac auscultation class with a cardiology patient simulator: discrepancy between students' satisfaction and learning. BMC MEDICAL EDUCATION 2021; 21:600. [PMID: 34872540 PMCID: PMC8647442 DOI: 10.1186/s12909-021-03033-8] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 08/11/2021] [Accepted: 11/19/2021] [Indexed: 06/13/2023]
Abstract
BACKGROUND We have provided fourth-year medical students with a three-hour cardiac auscultation class using a cardiology patient simulator since 2010. The test results of 2010-2012 revealed that as compared with aortic stenosis murmur, students correctly identified murmurs of other valvular diseases less often. We investigated whether employment of color Doppler echocardiographic video clips would improve proficiency in identifying murmurs of aortic regurgitation and mitral regurgitation, and whether students' favorable responses to a questionnaire were associated with improved proficiency. METHODS A total of 250 fourth-year medical students were divided into groups of 7-9 students in 2014 and 2015. Each group attended a three-hour cardiac auscultation class comprising a mini-lecture, facilitated training, two different auscultation tests (the second test being closer to clinical setting than the first) and a questionnaire. We provided each student with color Doppler echocardiographic videos of aortic regurgitation and mitral regurgitation using a tablet computer, which they freely referred to before and after listening to corresponding murmurs. The test results were compared with those in 2010-2012. The students had already completed the course of cardiovascular medicine, comprising lectures including those of physical examination, echocardiography, and valvular heart diseases, before participating in this auscultation training class. RESULTS Most students indicated that the videos were useful or somewhat useful regarding aortic regurgitation (86.3%) and mitral regurgitation (85.7%). The accuracy rates were 78.4% (81.2% in 2010-2012) in aortic regurgitation and 76.0% (77.8%) in mitral regurgitation in the first test, and 83.3% (71.4%) in aortic regurgitation and 77.1% (77.6%) in mitral regurgitation in the second test, showing no significant differences as compared to 2010-2012. Overall accuracy rate of all heart sounds and murmurs in the first test and that of second/third/fourth sounds in the first and second tests were significantly lower in 2014-2015 than in 2010-2012. CONCLUSIONS Referring to color Doppler echocardiographic video clips in the way employed in the present study, which most students regarded as useful, did not improve their proficiency in identifying the two important regurgitant murmurs, revealing a discrepancy between students' satisfaction and learning. Video clips synchronized with their corresponding murmurs may contribute toward improving students' proficiency.
Collapse
|
12
|
Zhang Y, Zheng Y, Wang M, Guo X. Prediction of exercise sudden death in rabbit exhaustive swimming using deep neural network. Biomed Eng Online 2021; 20:87. [PMID: 34461905 PMCID: PMC8404258 DOI: 10.1186/s12938-021-00925-0] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/28/2021] [Accepted: 08/19/2021] [Indexed: 11/30/2022] Open
Abstract
Background and objective Moderate exercise contributes to good health. However, excessive exercise may lead to cardiac fatigue, myocardial damage and even exercise sudden death. Monitoring the heart health has important implication to prevent exercise sudden death. Diagnosis methods such as electrocardiogram, echocardiogram, blood pressure and histological analysis have shown that arrhythmia and left ventricular fibrosis are early warning symptoms of exercise sudden death. Heart sounds (HS) can reflect the changes of cardiac valve, cardiac blood flow and myocardial function. Deep learning has drawn wide attention because of its ability to recognize disease. Therefore, a deep learning method combined with HS was proposed to predict exercise sudden death in New Zealand rabbits. The objective is to develop a method to predict exercise sudden death in New Zealand rabbits. Methods This paper proposed a method to predict exercise sudden death in New Zealand rabbits based on convolutional neural network (CNN) and gated recurrent unit (GRU). The weight-bearing exhaustive swimming experiment was conducted to obtain the HS of exercise sudden death and surviving New Zealand rabbits (n = 11/10) at four different time points. Then, the improved Viola integral method and double threshold method were employed to segment HS signals. The segmented HS frames at different time points were taken as the input of a combined CNN and GRU called CNN–GRU network to complete the prediction of exercise sudden death. Results In order to evaluate the performance of proposed network, CNN and GRU were used for comparison. When the fourth time point segmented HS frames were taken as input, the result shows that the proposed network has better performance with an accuracy of 89.57%, a sensitivity of 89.38% and a specificity of 92.20%. In addition, the segmented HS frames at different time points were input into CNN–GRU network, and the result shows that with the progress of the experiment, the prediction accuracy of exercise sudden death in New Zealand rabbits increased from 50.98 to 89.57%. Conclusion The proposed network shows good performance in classifying HS, which proves the feasibility of deep learning in exploring exercise sudden death. Further, it may have important implications in helping humans explore exercise sudden death.
Collapse
|
13
|
Levin AD, Ragazzi A, Szot SL, Ning T. Extraction and assessment of diagnosis-relevant features for heart murmur classification. Methods 2021; 202:110-116. [PMID: 34245871 DOI: 10.1016/j.ymeth.2021.07.002] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/27/2021] [Revised: 06/10/2021] [Accepted: 07/02/2021] [Indexed: 12/21/2022] Open
Abstract
This paper presents a heart murmur detection and multi-class classification approach via machine learning. We extracted heart sound and murmur features that are of diagnostic importance and developed additional 16 features that are not perceivable by human ears but are valuable to improve murmur classification accuracy. We examined and compared the classification performance of supervised machine learning with k-nearest neighbor (KNN) and support vector machine (SVM) algorithms. We put together a test repertoire having more than 450 heart sound and murmur episodes to evaluate the performance of murmur classification using cross-validation of 80-20 and 90-10 splits. As clearly demonstrated in our evaluation, the specific set of features chosen in our study resulted in accurate classification consistently exceeding 90% for both classifiers.
Collapse
|
14
|
Prospective validation of an acoustic-based system for the detection of obstructive coronary artery disease in a high-prevalence population. Heart Vessels 2021; 36:1132-1140. [PMID: 33582860 DOI: 10.1007/s00380-021-01800-7] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 11/13/2020] [Accepted: 01/29/2021] [Indexed: 11/27/2022]
Abstract
Recent guidelines recommend a risk-adjusted, non-invasive work-up in patients presenting with chest discomfort to exclude coronary artery disease (CAD). However, a risk-adjusted diagnostic approach remains challenging in clinical practice. An acoustic detection device for analyzing micro-bruits induced by stenosis-generated turbulence in the coronary circulation has shown potential for ruling out CAD in patients with low-to-intermediate likelihood. We examined the diagnostic value of this acoustic detection system in a high-prevalence cohort. In total, 226 patients scheduled for clinically indicated invasive coronary angiography (ICA) were prospectively enrolled at two centers and examined using a portable, acoustic detection system. The acoustic analysis was performed in double-blinded fashion prior to quantitative ICA and following percutaneous coronary intervention (PCI). An acoustic detection result (CAD score) was obtained in 94% of all patients. The mean baseline CAD score was 41.2 ± 11.9 in patients with obstructive CAD and 33.8 ± 13.4 in patients without obstructive CAD (p < 0.001). ROC analysis revealed an AUC of 0.661 (95% CI 0.584-0.737). Sensitivity was 97.6% (95% confidence interval (CI) 91.5-99.7%), specificity was 14.5% (CI 9.0-21.7%), negative predictive value was 90.5% (CI 69.6-98.8%), and positive predictive value was 41.7% (CI 34.6-49.0%). Following PCI, the mean CAD score decreased from 40.5 ± 11.2 to 38.3 ± 13.7 (p = 0.039). Using an acoustic detection device identified individuals with CAD in a high-prevalence cohort with high sensitivity but relatively low specificity. The negative predictive value was within the predicted range and may be of value for a fast rule-out of obstructive CAD even in a high-prevalence population.
Collapse
|
15
|
Coronary artery disease risk reclassification using an acoustic-based score in view of the new European Society of Cardiology 2019 guidelines on Chronic Coronary Syndromes. Int J Cardiovasc Imaging 2019; 36:383-384. [PMID: 31853822 PMCID: PMC7080710 DOI: 10.1007/s10554-019-01746-y] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 12/05/2019] [Accepted: 12/09/2019] [Indexed: 11/21/2022]
Abstract
In August 2019, ESC published new guidelines on Chronic Coronary Syndromes including a new risk assessment paradigm for estimation of pre-test-probability. The CAD-score is an acoustic-based score for ruling-out of coronary artery disease (CAD). In the current letter to the editor we re-evaluate the re-classification potential the CAD-score in the view of the new guidelines.
Collapse
|
16
|
Alqudah AM. Towards classifying non-segmented heart sound records using instantaneous frequency based features. J Med Eng Technol 2019; 43:418-430. [PMID: 31769312 DOI: 10.1080/03091902.2019.1688408] [Citation(s) in RCA: 20] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/11/2023]
Abstract
Heart sound and its recorded signal which is known as phonocardiograph (PCG) are one of the most important biosignals that can be used to diagnose cardiac diseases alongside electrocardiogram (ECG). Over the past few years, the use of PCG signals has become more widespread and researchers pay their attention to it and aim to provide an automated heart sound analysis and classification system that supports medical professionals in their decision. In this paper, a new method for heart sound features extraction for the classification of non-segmented signals using instantaneous frequency was proposed. The method has two major phases: the first phase is to estimate the instantaneous frequency of the recorded signal; the second phase is to extract a set of eleven features from the estimated instantaneous frequency. The method was tested into two different datasets, one for binary classification (Normal and Abnormal) and the other for multi-classification (Five Classes) to ensure the robustness of the extracted features. The overall accuracy, sensitivity, specificity, and precision for binary classification and multi-classification were all above 95% using both random forest and KNN classifiers.
Collapse
|
17
|
Abstract
For many years, heart function has been measured by the electrocardiogram (ECG) signal, while sounds produced in the heart can also contain information indicating normal or abnormal heart function. What has caused to restrict the use of the phonocardiography (PCG) signal was the lack of mastery of experts in the interpretation of these sounds, as well as its high potential for noise pollution. PCG is a non-invasive signal for monitoring physiological parameters of cardiac, which can make heart disease diagnostics more efficient. In recent years, attempts have been made to use PCG to detect heart disease independently without a need to match with the ECG. We propose a hybrid algorithm including empirical mode decomposition (EMD), Hilbert transform and Gaussian function for detecting heart sounds to distinguish first (S1) and second (S2) cardiac sounds by eliminating the effect of cardiac murmurs. In this article, 250 normal and 250 abnormal sound signals were examined. The overall positive predictivity of normal and abnormal S1 and S2 is 98.98%, 98.78, 98.78 and 98.37, respectively. Our results showed that the proposed method has a high potential for heart sounds determination, while maintains its simplicity and has a reasonable computational time.
Collapse
|
18
|
Arangalage D, Abtan J, Gaschignard J, Ceccaldi PF, Remini SA, Etienne I, Ruszniewski P, Plaisance P, De Lastours V, Lefort A, Faye A. Implementation of a large-scale simulation-based cardiovascular clinical examination course for undergraduate medical students - a pilot study. BMC MEDICAL EDUCATION 2019; 19:361. [PMID: 31533700 PMCID: PMC6751897 DOI: 10.1186/s12909-019-1750-4] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 05/26/2019] [Accepted: 08/13/2019] [Indexed: 06/10/2023]
Abstract
BACKGROUND We report the implementation of a large-scale simulation-based cardiovascular diagnostics course for undergraduate medical students. METHODS A simulation-based course was integrated into the curriculum of second-year medical students (> 400 students/year). The first session aimed at teaching cardiac auscultation skills on mannequins and the second at teaching blood pressure measurement, peripheral arterial examination, and the clinical examination of heart failure in a technical skill-based manner and in a scenario. RESULTS A total of 414 (99.8%) and 402 (98.5%) students, as well as 102 and 104 educators, participated during the 2016-2017 and 2017-2018 academic years across both types of sessions. The number of positive appreciations by students was high and improved from the first to the second year (session 1: 77% vs. 98%, session 2: 89% vs. 98%; p < 0.0001). Similar results were observed for educators (session 1: 84% vs. 98%, p = 0.007; session 2: 82% vs. 98%, p = 0.01). Feedbacks by students were positive regarding the usefulness of the course, fulfillment of pedagogical objectives, quality of the teaching method, time management, and educator-student interactivity. In contrast, 95% of students criticized the quality of the mannequins during the first year leading to the replacement of the simulation material the following year. Students most appreciated the auscultation workshop (25%), the practical aspect of the course (22%), and the availability of educators (21%). CONCLUSIONS Despite the need to commit significant human and material resources, the implementation of this large-scale program involving > 400 students/year was feasible, and students and educators reacted favorably.
Collapse
|
19
|
Schmidt SE, Winther S, Larsen BS, Groenhoej MH, Nissen L, Westra J, Frost L, Holm NR, Mickley H, Steffensen FH, Lambrechtsen J, Nørskov MS, Struijk JJ, Diederichsen ACP, Boettcher M. Coronary artery disease risk reclassification by a new acoustic-based score. Int J Cardiovasc Imaging 2019; 35:2019-2028. [PMID: 31273633 PMCID: PMC6805823 DOI: 10.1007/s10554-019-01662-1] [Citation(s) in RCA: 11] [Impact Index Per Article: 2.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 03/22/2019] [Accepted: 06/27/2019] [Indexed: 01/08/2023]
Abstract
To determine the potential of a non-invasive acoustic device (CADScor®System) to reclassify patients with intermediate pre-test probability (PTP) and clinically suspected stable coronary artery disease (CAD) into a low probability group thereby ruling out significant CAD. Audio recordings and clinical data from three studies were collected in a single database. In all studies, patients with a coronary CT angiography indicating CAD were referred to coronary angiography. Audio recordings of heart sounds were processed to construct a CAD-score. PTP was calculated using the updated Diamond-Forrester score and patients were classified according to the current ESC guidelines for stable CAD: low < 15%, intermediate 15–85% and high > 85% PTP. Intermediate PTP patients were re-classified to low probability if the CAD-score was ≤ 20. Of 2245 patients, 212 (9.4%) had significant CAD confirmed by coronary angiography ( ≥ 50% diameter stenosis). The average CAD-score was higher in patients with significant CAD (38.4 ± 13.9) compared to the remaining patients (25.1 ± 13.8; p < 0.001). The reclassification increased the proportion of low PTP patients from 13.6% to 41.8%, reducing the proportion of intermediate PTP patients from 83.4% to 55.2%. Before reclassification 7 (3.1%) low PTP patients had CAD, whereas post-reclassification this number increased to 28 (4.0%) (p = 0.52). The net reclassification index was 0.209. Utilization of a low-cost acoustic device in patients with intermediate PTP could potentially reduce the number of patients referred for further testing, without a significant increase in the false negative rate, and thus improve the cost-effectiveness for patients with suspected stable CAD.
Collapse
|
20
|
Sotaquir� M, Alvear D, Mondrag�n M. Phonocardiogram classification using deep neural networks and weighted probability comparisons. J Med Eng Technol 2019; 42:510-517. [PMID: 30773957 DOI: 10.1080/03091902.2019.1576789] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/11/2023]
Abstract
Cardiac auscultation is one of the most conventional approaches for the initial assessment of heart disease, however the technique is highly user-dependent and with low repeatability. Several computational approaches based on the analysis of the phonocardiograms (PCG) have been proposed to classify heart sounds into normal or abnormal, but most often do not achieve acceptable levels of sensitivity (Se) and specificity (Sp) or require the use of special hardware. We propose a novel approach for classification of PCG. First, the system makes use of deep neural networks for computing individual cardiac cycle probabilities, followed by classification using weighted probability comparisons. The system was tested on an extended dataset consisting of a balanced sample of 18179 normal and abnormal cycles, achieving Se and Sp values of 91.3% and 93.8% respectively. In addition, the system overcomes previous limitations since it was trained with a balanced sample; also, the decision factor used during the classification stage allows to control the trade-off between Se and Sp, making the proposed system suitable for clinical applications.
Collapse
|
21
|
Kalkbrenner C, Eichenlaub M, R�diger S, Kropf-Sanchen C, Rottbauer W, Brucher R. Apnea and heart rate detection from tracheal body sounds for the diagnosis of sleep-related breathing disorders. Med Biol Eng Comput 2017; 56:671-681. [PMID: 28849304 DOI: 10.1007/s11517-017-1706-y] [Citation(s) in RCA: 19] [Impact Index Per Article: 2.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/16/2016] [Accepted: 08/03/2017] [Indexed: 11/28/2022]
Abstract
Sleep apnea is one of the most common sleep disorders. Here, patients suffer from multiple breathing pauses longer than 10�s during the night which are referred to as apneas. The standard method for the diagnosis of sleep apnea is the attended cardiorespiratory polysomnography (PSG). However, this method is expensive and the extensive recording equipment can have a significant impact on sleep quality falsifying the results. To overcome these problems, a comfortable and novel system for sleep monitoring based on the recording of tracheal sounds and movement data is developed. For apnea detection, a unique signal processing method utilizing both signals is introduced. Additionally, an algorithm for extracting the heart rate from body sounds is developed. For validation, ten subjects underwent a full-night PSG testing, using the developed sleep monitor in concurrence. Considering polysomnography as gold standard the developed instrumentation reached a sensitivity of 92.8% and a specificity of 99.7% for apnea detection. Heart rate measured with the proposed method was strongly correlated with heart rate derived from conventional ECG (r 2�=�0.8164). No significant signal losses are reported during the study. In conclusion, we demonstrate a novel approach to reliably and noninvasively detect both apneas and heart rate during sleep.
Collapse
|
22
|
Automatic heart activity diagnosis based on Gram polynomials and probabilistic neural networks. Biomed Eng Lett 2017; 8:77-85. [PMID: 30603192 DOI: 10.1007/s13534-017-0046-z] [Citation(s) in RCA: 17] [Impact Index Per Article: 2.4] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/27/2017] [Revised: 07/13/2017] [Accepted: 08/14/2017] [Indexed: 10/19/2022] Open
Abstract
The paper proposes a new approach to heart activity diagnosis based on Gram polynomials and probabilistic neural networks (PNN). Heart disease recognition is based on the analysis of phonocardiogram (PCG) digital sequences. The PNN provides a powerful tool for proper classification of the input data set. The novelty of the proposed approach lies in a powerful feature extraction based on Gram polynomials and the Fourier transform. The proposed system presents good performance obtaining overall sensitivity of 93%, specificity of 91% and accuracy of 94%, using a public database of over 3000 heart beat sound recordings, classified as normal and abnormal heart sounds. Thus, it can be concluded that Gram polynomials and PNN prove to be a very efficient technique using the PCG signal for characterizing heart diseases.
Collapse
|
23
|
Moghaddasi H, Almasganj F, Zoroufian A. Imaging of heart acoustic based on the sub-space methods using a microphone array. COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 2017; 146:133-142. [PMID: 28688483 DOI: 10.1016/j.cmpb.2017.04.017] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 02/14/2016] [Revised: 04/03/2017] [Accepted: 04/11/2017] [Indexed: 06/07/2023]
Abstract
BACKGROUND AND OBJECTIVES Heart disease is one of the leading causes of death around the world. Phonocardiogram (PCG) is an important bio-signal which represents the acoustic activity of heart, typically without any spatiotemporal information of the involved acoustic sources. The aim of this study is to analyze the PCG by employing a microphone array by which the heart internal sound sources could be localized, too. METHOD In this paper, it is intended to propose a modality by which the locations of the active sources in the heart could also be investigated, during a cardiac cycle. In this way, a microphone array with six microphones is employed as the recording set up to be put on the human chest. In the following, the Group Delay MUSIC algorithm which is a sub-space based localization method is used to estimate the location of the heart sources in different phases of the PCG. RESULTS We achieved to 0.14cm mean error for the sources of first heart sound (S1) simulator and 0.21cm mean error for the sources of second heart sound (S2) simulator with Group Delay MUSIC algorithm. The acoustical diagrams created for human subjects show distinct patterns in various phases of the cardiac cycles such as the first and second heart sounds. Moreover, the evaluated source locations for the heart valves are matched with the ones that are obtained via the 4-dimensional (4D) echocardiography applied, to a real human case. CONCLUSIONS Imaging of heart acoustic map presents a new outlook to indicate the acoustic properties of cardiovascular system and disorders of valves and thereby, in the future, could be used as a new diagnostic tool.
Collapse
|
24
|
Mamorita N, Arisaka N, Isonaka R, Kawakami T, Takeuchi A. Development of a Smartphone App for Visualizing Heart Sounds and Murmurs. Cardiology 2017; 137:193-200. [PMID: 28441656 DOI: 10.1159/000466683] [Citation(s) in RCA: 7] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 10/13/2016] [Accepted: 02/28/2017] [Indexed: 11/19/2022]
Abstract
BACKGROUND Auscultation is one of the basic techniques for the diagnosis of heart disease. However, the interpretation of heart sounds and murmurs is a highly subjective and difficult skill. OBJECTIVES To assist the auscultation skill at the bedside, a handy phonocardiogram was developed using a smartphone (Samsung Galaxy J, Android OS 4.4.2) and an external microphone attached to a stethoscope. METHODS AND RESULTS The Android app used Java classes, "AudioRecord," "AudioTrack," and "View," that recorded sounds, replayed sounds, and plotted sound waves, respectively. Sound waves were visualized in real-time, simultaneously replayed on the smartphone, and saved to WAV files. To confirm the availability of the app, 26 kinds of heart sounds and murmurs sounded on a human patient simulator were recorded using three different methods: a bell-type stethoscope, a diaphragm-type stethoscope, and a direct external microphone without a stethoscope. The recorded waveforms were subjectively confirmed and were found to be similar to the reference waveforms. CONCLUSIONS The real-time visualization of the sound waves on the smartphone may help novices to readily recognize and learn to distinguish the various heart sounds and murmurs in real-time.
Collapse
|
25
|
Karar ME, El-Khafif SH, El-Brawany MA. Automated Diagnosis of Heart Sounds Using Rule-Based Classification Tree. J Med Syst 2017; 41:60. [PMID: 28247307 DOI: 10.1007/s10916-017-0704-9] [Citation(s) in RCA: 20] [Impact Index Per Article: 2.9] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/29/2016] [Accepted: 02/09/2017] [Indexed: 10/20/2022]
Abstract
In order to assist the diagnosis procedure of heart sound signals, this paper presents a new automated method for classifying the heart status using a rule-based classification tree into normal and three abnormal cases; namely the aortic valve stenosis, aortic insufficient, and ventricular septum defect. The developed method includes three main steps as follows. First, one cycle of the heart sound signals is automatically detected and segmented based on time properties of the heart signals. Second, the segmented cycle is preprocessed with the discrete wavelet transform and then largest Lyapunov exponents are calculated to generate the dynamical features of heart sound time series. Finally, a rule-based classification tree is fed by these Lyapunov exponents to give the final decision of the heart health status. The developed method has been tested successfully on twenty-two datasets of normal heart sounds and murmurs with success rate of 95.5%. The resulting error can be easily corrected by modifying the classification rules; consequently, the accuracy of automated heart sounds diagnosis is further improved.
Collapse
|