Automated Identification of Innocent Still's Murmur in Children

Identifying pediatric heart murmurs and distinguishing innocent from pathologic using deep learning

OBJECTIVE

To develop a deep learning algorithm to perform multi-class classification of normal pediatric heart sounds, innocent murmurs, and pathologic murmurs.

METHODS

We prospectively enrolled children under age 18 being evaluated by the Division of Pediatric Cardiology. Parents provided consent for a deidentified recording of their child's heart sounds with a digital stethoscope. Innocent murmurs were validated by a pediatric cardiologist and pathologic murmurs were validated by echocardiogram. To augment our collection of normal heart sounds, we utilized a public database of pediatric heart sound recordings (Oliveira, 2022). We propose two novel approaches for this audio classification task. We train a vision transformer on either Markov transition field or Gramian angular field image representations of the frequency spectrum. We benchmark our results against a ResNet-50 CNN trained on spectrogram images.

RESULTS

Our final dataset consisted of 366 normal heart sounds, 175 innocent murmurs, and 216 pathologic murmurs. Innocent murmurs collected include Still's murmur, venous hum, and flow murmurs. Pathologic murmurs included ventricular septal defect, tetralogy of Fallot, aortic regurgitation, aortic stenosis, pulmonary stenosis, mitral regurgitation and stenosis, and tricuspid regurgitation. We find that the Vision Transformer consistently outperforms the ResNet-50 on all three image representations, and that the Gramian angular field is the superior image representation for pediatric heart sounds. We calculated a one-vs-rest multi-class ROC curve for each of the three classes. Our best model achieves an area under the curve (AUC) value of 0.92 ± 0.05, 0.83 ± 0.04, and 0.88 ± 0.04 for identifying normal heart sounds, innocent murmurs, and pathologic murmurs, respectively.

CONCLUSION

We present two novel methods for pediatric heart sound classification, which outperforms the current standard of using a convolutional neural network trained on spectrogram images. To our knowledge, we are the first to demonstrate multi-class classification of pediatric murmurs. Multiclass output affords a more explainable and interpretable model, which can facilitate further model improvement in the downstream model development cycle and enhance clinician trust and therefore adoption.

A Noise-Robust Heart Sound Segmentation Algorithm Based on Shannon Energy

Heart sound segmentation has been shown to improve the performance of artificial intelligence (AI)-based auscultation decision support systems increasingly viewed as a solution to compensate for eroding auscultatory skills and the associated subjectivity. Various segmentation approaches with demonstrated performance can be utilized for this task, but their robustness can suffer in the presence of noise. A noise-robust heart sound segmentation algorithm was developed and its accuracy was tested using two datasets: the CirCor DigiScope Phonocardiogram dataset and an in-house dataset - a heart murmur library collected at the Children's National Hospital (CNH). On the CirCor dataset, our segmentation algorithm marked the boundaries of the primary heart sounds S1 and S2 with an accuracy of 0.28 ms and 0.29 ms, respectively, and correctly identified the actual positive segments with a sensitivity of 97.44%. The algorithm also executed four times faster than a logistic regression hidden semi-Markov model. On the CNH dataset, the algorithm succeeded in 87.4% cases, achieving a 6% increase in segmentation success rate demonstrated by our original Shannon energy-based algorithm. Accurate heart sound segmentation is critical to supporting and accelerating AI research in cardiovascular diseases. The proposed algorithm increases the robustness of heart sound segmentation to noise and viability for clinical use.

Deep Learning Algorithms to Detect Murmurs Associated With Structural Heart Disease

Background The success of cardiac auscultation varies widely among medical professionals, which can lead to missed treatments for structural heart disease. Applying machine learning to cardiac auscultation could address this problem, but despite recent interest, few algorithms have been brought to clinical practice. We evaluated a novel suite of Food and Drug Administration-cleared algorithms trained via deep learning on >15 000 heart sound recordings. Methods and Results We validated the algorithms on a data set of 2375 recordings from 615 unique subjects. This data set was collected in real clinical environments using commercially available digital stethoscopes, annotated by board-certified cardiologists, and paired with echocardiograms as the gold standard. To model the algorithm in clinical practice, we compared its performance against 10 clinicians on a subset of the validation database. Our algorithm reliably detected structural murmurs with a sensitivity of 85.6% and specificity of 84.4%. When limiting the analysis to clearly audible murmurs in adults, performance improved to a sensitivity of 97.9% and specificity of 90.6%. The algorithm also reported timing within the cardiac cycle, differentiating between systolic and diastolic murmurs. Despite optimizing acoustics for the clinicians, the algorithm substantially outperformed the clinicians (average clinician accuracy, 77.9%; algorithm accuracy, 84.7%.) Conclusions The algorithms accurately identified murmurs associated with structural heart disease. Our results illustrate a marked contrast between the consistency of the algorithm and the substantial interobserver variability of clinicians. Our results suggest that adopting machine learning algorithms into clinical practice could improve the detection of structural heart disease to facilitate patient care.

StethAid: A Digital Auscultation Platform for Pediatrics

(1) Background: Mastery of auscultation can be challenging for many healthcare providers. Artificial intelligence (AI)-powered digital support is emerging as an aid to assist with the interpretation of auscultated sounds. A few AI-augmented digital stethoscopes exist but none are dedicated to pediatrics. Our goal was to develop a digital auscultation platform for pediatric medicine. (2) Methods: We developed StethAid-a digital platform for artificial intelligence-assisted auscultation and telehealth in pediatrics-that consists of a wireless digital stethoscope, mobile applications, customized patient-provider portals, and deep learning algorithms. To validate the StethAid platform, we characterized our stethoscope and used the platform in two clinical applications: (1) Still's murmur identification and (2) wheeze detection. The platform has been deployed in four children's medical centers to build the first and largest pediatric cardiopulmonary datasets, to our knowledge. We have trained and tested deep-learning models using these datasets. (3) Results: The frequency response of the StethAid stethoscope was comparable to those of the commercially available Eko Core, Thinklabs One, and Littman 3200 stethoscopes. The labels provided by our expert physician offline were in concordance with the labels of providers at the bedside using their acoustic stethoscopes for 79.3% of lungs cases and 98.3% of heart cases. Our deep learning algorithms achieved high sensitivity and specificity for both Still's murmur identification (sensitivity of 91.9% and specificity of 92.6%) and wheeze detection (sensitivity of 83.7% and specificity of 84.4%). (4) Conclusions: Our team has created a technically and clinically validated pediatric digital AI-enabled auscultation platform. Use of our platform could improve efficacy and efficiency of clinical care for pediatric patients, reduce parental anxiety, and result in cost savings.

Artificial Intelligence in Pediatric Cardiology: A Scoping Review

The evolution of AI and data science has aided in mechanizing several aspects of medical care requiring critical thinking: diagnosis, risk stratification, and management, thus mitigating the burden of physicians and reducing the likelihood of human error. AI modalities have expanded feet to the specialty of pediatric cardiology as well. We conducted a scoping review searching the Scopus, Embase, and PubMed databases covering the recent literature between 2002-2022. We found that the use of neural networks and machine learning has significantly improved the diagnostic value of cardiac magnetic resonance imaging, echocardiograms, computer tomography scans, and electrocardiographs, thus augmenting the clinicians' diagnostic accuracy of pediatric heart diseases. The use of AI-based prediction algorithms in pediatric cardiac surgeries improves postoperative outcomes and prognosis to a great extent. Risk stratification and the prediction of treatment outcomes are feasible using the key clinical findings of each CHD with appropriate computational algorithms. Notably, AI can revolutionize prenatal prediction as well as the diagnosis of CHD using the EMR (electronic medical records) data on maternal risk factors. The use of AI in the diagnostics, risk stratification, and management of CHD in the near future is a promising possibility with current advancements in machine learning and neural networks. However, the challenges posed by the dearth of appropriate algorithms and their nascent nature, limited physician training, fear of over-mechanization, and apprehension of missing the 'human touch' limit the acceptability. Still, AI proposes to aid the clinician tomorrow with precision cardiology, paving a way for extremely efficient human-error-free health care.

Automated identification of innocent Still's murmur using a convolutional neural network

Background

Still's murmur is the most prevalent innocent heart murmur of childhood. Auscultation is the primary clinical tool to identify this murmur as innocent. Whereas pediatric cardiologists routinely perform this task, primary care providers are less successful in distinguishing Still's murmur from the murmurs of true heart disease. This results in a large number of children with a Still's murmur being referred to pediatric cardiologists.

Objectives

To develop a computer algorithm that can aid primary care providers to identify the innocent Still's murmur at the point of care, to substantially decrease over-referral.

Methods

The study included Still's murmurs, pathological murmurs, other innocent murmurs, and normal (i.e., non-murmur) heart sounds of 1,473 pediatric patients recorded using a commercial electronic stethoscope. The recordings with accompanying clinical diagnoses provided by a pediatric cardiologist were used to train and test the convolutional neural network-based algorithm.

Results

A comparative analysis showed that the algorithm using only the murmur sounds recorded at the lower left sternal border achieved the highest accuracy. The developed algorithm identified Still's murmur with 90.0% sensitivity and 98.3% specificity for the default decision threshold. The area under the receiver operating characteristic curve was 0.943.

Conclusions

Still's murmur can be identified with high accuracy with the algorithm we developed. Using this approach, the algorithm could help to reduce the rate of unnecessary pediatric cardiologist referrals and use of echocardiography for a common benign finding.

On the analysis of data augmentation methods for spectral imaged based heart sound classification using convolutional neural networks

Background

The application of machine learning to cardiac auscultation has the potential to improve the accuracy and efficiency of both routine and point-of-care screenings. The use of convolutional neural networks (CNN) on heart sound spectrograms in particular has defined state-of-the-art performance. However, the relative paucity of patient data remains a significant barrier to creating models that can adapt to a wide range of potential variability. To that end, we examined a CNN model’s performance on automated heart sound classification, before and after various forms of data augmentation, and aimed to identify the most optimal augmentation methods for cardiac spectrogram analysis.

Results

We built a standard CNN model to classify cardiac sound recordings as either normal or abnormal. The baseline control model achieved a PR AUC of 0.763 ± 0.047. Among the single data augmentation techniques explored, horizontal flipping of the spectrogram image improved the model performance the most, with a PR AUC of 0.819 ± 0.044. Principal component analysis color augmentation (PCA) and perturbations of saturation-value (SV) of the hue-saturation-value (HSV) color scale achieved a PR AUC of 0.779 ± 045 and 0.784 ± 0.037, respectively. Time and frequency masking resulted in a PR AUC of 0.772 ± 0.050. Pitch shifting, time stretching and compressing, noise injection, vertical flipping, and applying random color filters negatively impacted model performance. Concatenating the best performing data augmentation technique (horizontal flip) with PCA and SV perturbations improved model performance.

Conclusion

Data augmentation can improve classification accuracy by expanding and diversifying the dataset, which protects against overfitting to random variance. However, data augmentation is necessarily domain specific. For example, methods like noise injection have found success in other areas of automated sound classification, but in the context of cardiac sound analysis, noise injection can mimic the presence of murmurs and worsen model performance. Thus, care should be taken to ensure clinically appropriate forms of data augmentation to avoid negatively impacting model performance.

Classification of Children's Heart Sounds With Noise Reduction Based on Variational Modal Decomposition

Purpose

Children's heart sounds were denoised to improve the performance of the intelligent diagnosis.

Methods

A combined noise reduction method based on variational modal decomposition (VMD) and wavelet soft threshold algorithm (WST) was proposed, and used to denoise 103 phonocardiogram samples. Features were extracted after denoising and employed for an intelligent diagnosis model to verify the effect of the denoising method.

Results

The noise in children's phonocardiograms, especially crying noise, was suppressed. The signal-to-noise ratio obtained by the method for normal heart sounds was 14.69 dB at 5 dB Gaussian noise, which was higher than that obtained by WST only and the other VMD denoising method. Intelligent classification showed that the accuracy, sensitivity and specificity of the classification system for congenital heart diseases were 92.23, 92.42, and 91.89%, respectively and better than those with WST only.

Conclusion

The proposed noise reduction method effectively eliminates noise in children's phonocardiograms and improves the performance of intelligent screening for the children with congenital heart diseases.

Embedded platform based heart murmur classification using deep learning approach

Ubiquitous Perturbations in cardiac auscultation properties, cardiovascular diseases (CVDs) are widely recognized. In the auscultation procedure, the appearance of pathological cardiac murmurs is linked to heart disorders. A noble automated detection system using 1-D Convolutional Neural Network (CNN) for the detection of pathological heart murmurs is proposed in this study, which removes the difficult task of extracting and selecting features. It directly acts on the phonocardiogram (PCG) signals. The fundamental purpose of this research is to develop a classification model for consistent recognition of cardiac murmurs when the data-set is imbalanced. In view of this, the proposed study for the imbalanced data-set incorporates the Adaptive Synthetic (ADASYN) approach to generate synthetic data for the minority class. The outcome analysis illustrates the positive result in the identification of heart murmurs on both balanced and imbalanced data-sets. Therefore, the developed deep learning model will learn better from the minority class and classify heart murmurs accurately.

Cardiovascular Disease Recognition Based on Heartbeat Segmentation and Selection Process

Assessment of heart sounds which are generated by the beating heart and the resultant blood flow through it provides a valuable tool for cardiovascular disease (CVD) diagnostics. The cardiac auscultation using the classical stethoscope phonological cardiogram is known as the most famous exam method to detect heart anomalies. This exam requires a qualified cardiologist, who relies on the cardiac cycle vibration sound (heart muscle contractions and valves closure) to detect abnormalities in the heart during the pumping action. Phonocardiogram (PCG) signal represents the recording of sounds and murmurs resulting from the heart auscultation, typically with a stethoscope, as a part of medical diagnosis. For the sake of helping physicians in a clinical environment, a range of artificial intelligence methods was proposed to automatically analyze PCG signal to help in the preliminary diagnosis of different heart diseases. The aim of this research paper is providing an accurate CVD recognition model based on unsupervised and supervised machine learning methods relayed on convolutional neural network (CNN). The proposed approach is evaluated on heart sound signals from the well-known, publicly available PASCAL and PhysioNet datasets. Experimental results show that the heart cycle segmentation and segment selection processes have a direct impact on the validation accuracy, sensitivity (TPR), precision (PPV), and specificity (TNR). Based on PASCAL dataset, we obtained encouraging classification results with overall accuracy 0.87, overall precision 0.81, and overall sensitivity 0.83. Concerning Micro classification results, we obtained Micro accuracy 0.91, Micro sensitivity 0.83, Micro precision 0.84, and Micro specificity 0.92. Using PhysioNet dataset, we achieved very good results: 0.97 accuracy, 0.946 sensitivity, 0.944 precision, and 0.946 specificity.

Artificial intelligence-assisted auscultation in detecting congenital heart disease

Aims

Computer-assisted auscultation has become available to assist clinicians with physical examinations to detect congenital heart disease (CHD). However, its accuracy and effectiveness remain to be evaluated. This study seeks to evaluate the accuracy of auscultations of abnormal heart sounds of an artificial intelligence-assisted auscultation (AI-AA) platform we create.

Methods and results

Initially, 1397 patients with CHD were enrolled in the study. The samples of their heart sounds were recorded and uploaded to the platform using a digital stethoscope. By the platform, both remote auscultation by a team of experienced cardiologists from Shanghai Children's Medical Center and automatic auscultation of the heart sound samples were conducted. Samples of 35 patients were deemed unsuitable for the analysis; therefore, the remaining samples from 1362 patients (mean age-2.4 ± 3.1 years and 46% female) were analysed. Sensitivity, specificity, and accuracy were calculated for remote auscultation compared to experts' face-to-face auscultation and for artificial intelligence automatic auscultation compared to experts' face-to-face auscultation. Kappa coefficients were measured. Compared to face-to-face auscultation, remote auscultation detected abnormal heart sound with 98% sensitivity, 91% specificity, 97% accuracy, and kappa coefficient 0.87. AI-AA demonstrated 97% sensitivity, 89% specificity, 96% accuracy, and kappa coefficient 0.84.

Conclusions

The remote auscultations and automatic auscultations, using the AI-AA platform, reported high auscultation accuracy in detecting abnormal heart sound and showed excellent concordance to experts' face-to-face auscultation. Hence, the platform may provide a feasible way to screen and detect CHD.

Intelligent Diagnosis of Heart Murmurs in Children with Congenital Heart Disease

Heart auscultation is a convenient tool for early diagnosis of heart diseases and is being developed to be an intelligent tool used in online medicine. Currently, there are few studies on intelligent diagnosis of pediatric murmurs due to congenital heart disease (CHD). The purpose of the study was to develop a method of intelligent diagnosis of pediatric CHD murmurs. Phonocardiogram (PCG) signals of 86 children were recorded with 24 children having normal heart sounds and 62 children having CHD murmurs. A segmentation method based on the discrete wavelet transform combined with Hadamard product was implemented to locate the first and the second heart sounds from the PCG signal. Ten features specific to CHD murmurs were extracted as the input of classifier after segmentation. Eighty-six artificial neural network classifiers were composed into a classification system to identify CHD murmurs. The accuracy, sensitivity, and specificity of diagnosis for heart murmurs were 93%, 93.5%, and 91.7%, respectively. In conclusion, a method of intelligent diagnosis of pediatric CHD murmurs is developed successfully and can be used for online screening of CHD in children.

An Approach for Automatic Identification of Fundamental and Additional Sounds from Cardiac Sounds Recordings

This paper presents an approach for automatic segmentation of cardiac events from non-invasive sounds recordings, without the need of having an auxiliary signal reference. In addition, methods are proposed to subsequently differentiate cardiac events which correspond to normal cardiac cycles, from those which are due to abnormal activity of the heart. The detection of abnormal sounds is based on a model built with parameters which are obtained following feature extraction from those segments that were previously identified as normal fundamental heart sounds. The proposed algorithm achieved a sensitivity of 91.79% and 89.23% for the identification of normal fundamental, S₁ and S₂ sounds, and a true positive (TP) rate of 81.48% for abnormal additional sounds. These results were obtained using the PASCAL Classifying Heart Sounds challenge (CHSC) database.

Fully-Automated Diagnosis of Aortic Stenosis Using Phonocardiogram-Based Features

The irreversible damage and eventual heart failure caused by untreated aortic stenosis (AS) can be prevented by early detection and timely intervention. Prior work in the field of phonocardiogram (PCG) signal analysis has provided proof of concept for using heart-sound data in AS diagnosis. However, such systems either require operation by trained technicians, fail to address a diverse subject set, or involve unwieldy configuration procedures that challenge real-world application. This paper presents an end-to-end, fully-automated system that uses noise-subtraction, heartbeat-segmentation and quality-assurance algorithms to extract physiologically-motivated features from PCG signals to diagnose AS. When tested on n=96 patients showing a diverse set of cardiac and non-cardiac conditions, the system was able to diagnose AS with 92% sensitivity and 95% specificity.