Murmur identification and outcome prediction in phonocardiograms using deep features based on Stockwell transform.

Traditionally, heart murmurs are diagnosed through cardiac auscultation, which requires specialized training and experience. The purpose of this study is to predict patients' clinical outcomes (normal or abnormal) and identify the presence or absence of heart murmurs using phonocardiograms (PCGs) obtained at different auscultation points. A semi-supervised model tailored to PCG classification is introduced in this study, with the goal of improving performance using time-frequency deep features. The study begins by investigating the behavior of PCGs in the time-frequency domain, utilizing the Stockwell transform to convert the PCG signal into two-dimensional time-frequency maps (TFMs). A deep network named AlexNet is then used to derive deep feature sets from these TFMs. In feature reduction, redundancy is eliminated and the number of deep features is reduced to streamline the feature set. The effectiveness of the extracted features is evaluated using three different classifiers using the CinC/Physionet challenge 2022 dataset. For Task I, which focuses on heart murmur detection, the proposed approach achieved an average accuracy of 93%, sensitivity of 91%, and F1-score of 91%. According to Task II of the CinC/Physionet challenge 2022, the approach showed a clinical outcome cost of 5290, exceeding the benchmark set by leading methods in the challenge.

Major congenital heart defects are rarely diagnosed after newborns' hospital discharge with modern screening.

AIM: Our aim was to assess undiagnosed congenital heart defects (CHD) after newborns' hospital discharge in patients with a murmur or CHD suspicion, to find out the signs that predict CHDs and to estimate the costs of the examinations. METHODS: We reviewed retrospective medical records of patients (n = 490) referred for the evaluation of CHD suspicion during 2017-2018. RESULTS: The median age of the patients was 2.5 (IQR 0.5-7.4) years. Sixty-three (13%) patients had an abnormal echocardiography. Neither ductal-dependent nor cyanotic CHDs were found. Cardiac interventions were performed for 14 out of 63 (22%) patients. Clinical signs indicating CHDs were murmur grade ≥3 (10/11 [91%] vs. 53/479 [11%], p < 0.001) and harsh murmur (15/44 [34%] vs. 48/446 [11%], p < 0.001). Abnormal electrocardiography did not indicate CHD (8/40 [20%] vs. 55/447 [12%], p = 0.165). The total cost of the examinations was 259 700€. The share of the cost of studies assessed as benign was 59%. CONCLUSION: Only a few CHDs were found after newborn hospital discharge among patients who received foetal and newborn screening and were examined due to CHD suspicion. The high number of benign murmurs in children leads to many referrals, resulting in unnecessary healthcare costs.

Early-warning of Cardiac Condition through Detection of Murmur in Heart Sound - A Case Study.

Cardiovascular disease, particularly Rheumatic Heart Disease (RHD), is one of the leading causes of death in many developing countries. RHD is manageable and treatable with early detection. However, multiple countries across the globe suffer from a scarcity of experienced physicians who can perform screening at large scales. Advancements in machine learning and signal processing have paved way for Phonocardiogram (PCG)-based automatic heart sound classification. The direct implication of such methods is that it is possible to enable a person without specialized training to detect potential cardiac conditions with just a digital stethoscope. Hospitalization or life-threatening situations can be dramatically reduced via such early screenings. Towards this, we conducted a case study amongst a population from a particular geography using machine learning and deep learning methods for the detection of murmur in heart sounds. The methodology consists of first pre-processing and identifying normal vs. abnormal heart sound signals using 3 state-of-the-art methods. The second step further identifies the murmur to be systolic or diastolic by capturing the auscultation location. Abnormal findings are then sent for early attention of clinicians for proper diagnosis. The case study investigates the efficacy of the automated method employed for early screening of potential RHD and initial encouraging results of the study are presented.

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.

Aortic valve insufficiency caused by a hypoplastic aortic valve in a calf.

A 1-month-old crossbred calf was referred for examination due to marked systolic heart murmurs and poor growth. The heart murmur was most audible on the right side of the cranial thorax. Cardiomegaly was evident on chest radiography, and echocardiography demonstrated aortic regurgitation and decreased fractional shortening. Cardiomegaly, aortic root dilation and cardiac displacement were confirmed by computed tomography. At necropsy, the heart was enlarged, and all three aortic valve leaflets were irregularly shaped. In calves with chronic aortic insufficiency, remodeling displacement of the heart and aorta causes changes in the location and timing of heart murmurs. Therefore, aortic insufficiency cannot be ruled out when a systolic heart murmur can be observed in the right chest wall.

Pediatric Murmurs.

Many healthy children may be found to have a murmur on physical exam. Whether this murmur is discovered at a routine health maintenance visit or as a result of a focused exam on a child with illness, it is just one finding and must be considered in the context of the child's history and other physical exam findings. Murmurs associated with heart defect or dysfunction occur most often in infancy. Most murmurs discovered in children, especially after infancy, between ages 3 to 6 and in young-adulthood, are innocent or benign murmurs and less likely a symptom of cardiac dysfunction or defect.

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.

Feasibility of Digital Stethoscopes in Telecardiology Visits for Interstage Monitoring in Infants with Palliated Congenital Heart Disease.

Infants with staged surgical palliation for congenital heart disease are at high-risk for interstage morbidity and mortality. Interstage telecardiology visits (TCV) have been effective in identifying clinical concerns and preventing unnecessary emergency department visits in this high-risk population. We aimed to assess the feasibility of implementing auscultation with digital stethoscopes (DSs) during TCV and the potential impact on interstage care in our Infant Single Ventricle Monitoring & Management Program. In addition to standard home-monitoring practice for TCV, caregivers received training on use of a DS (Eko CORE attachment assembled with Classic II Infant Littman stethoscope). Sound quality of the DS and comparability to in-person auscultation were evaluated based on two providers' subjective assessment. We also evaluated provider and caregiver acceptability of the DS. From 7/2021 to 6/2022, the DS was used during 52 TCVs in 16 patients (median TCVs/patient: 3; range: 1-8), including 7 with hypoplastic left heart syndrome. Quality of heart sounds and murmur auscultation were subjectively equivalent to in-person findings with excellent inter-rater agreement (98%). All providers and caregivers reported ease of use and confidence in evaluation with the DS. In 12% (6/52) of TCVs, the DS provided additional significant information compared to a routine TCV; this expedited life-saving care in two patients. There were no missed events or deaths. Use of a DS during TCV was feasible in this fragile cohort and effective in identifying clinical concerns with no missed events. Longer term use of this technology will further establish its role in telecardiology.

Evaluation of Heart Murmurs in Children.

This Patient Page describes how heart murmurs may be found in children and what treatment and follow-up may look like if an abnormal murmur is found.

Evaluation of a new smartphone-based digital stethoscope featuring phonocardiography and electrocardiography in dogs and cats.

This study assessed a new smartphone-based digital stethoscope (DS) featuring simultaneous phonocardiographic and one-lead electrocardiogram (ECG) recording in dogs and cats. The audio files and ECG traces obtained by the device were compared with conventional auscultation and standard ECG. A total of 99 dogs and nine cats were prospectively included. All cases underwent conventional auscultation using an acoustic stethoscope, standard six-lead ECG, standard echocardiography and recordings with the DS. All the audio recordings, phonocardiographic files and ECG traces were then blind reviewed by an expert operator. The agreement between methods was assessed using Cohen's kappa and the Bland-Altman test. Audio recordings were considered interpretable in 90% animals. Substantial agreement was found in the diagnosis of heart murmur (κ = 0.691) and gallop sound (k = 0.740). In nine animals with an echocardiographic diagnosis of heart disease, only the DS detected a heart murmur or gallop sound. ECG traces recorded with the new device were deemed interpretable in 88 % animals. Diagnosis of heart rhythm showed moderate agreement in the identification of atrial fibrillation (k = 0.596). The detection of ventricular premature complexes and bundle branch blocks revealed an almost perfect agreement (k = 1). Overall, the DS showed a good diagnostic accuracy in detecting heart murmurs, gallop sounds, ventricular premature complexes and bundle branch blocks. A clinically relevant overdiagnosis of atrial fibrillation was found but without evidence of false negatives. The DS could represent a useful screening tool for heart sound abnormalities and cardiac arrhythmias..

Beyond Heart Murmur Detection: Automatic Murmur Grading From Phonocardiogram.

OBJECTIVE: Murmurs are abnormal heart sounds, identified by experts through cardiac auscultation. The murmur grade, a quantitative measure of the murmur intensity, is strongly correlated with the patient's clinical condition. This work aims to estimate each patient's murmur grade (i.e., absent, soft, loud) from multiple auscultation location phonocardiograms (PCGs) of a large population of pediatric patients from a low-resource rural area. METHODS: The Mel spectrogram representation of each PCG recording is given to an ensemble of 15 convolutional residual neural networks with channel-wise attention mechanisms to classify each PCG recording. The final murmur grade for each patient is derived based on the proposed decision rule and considering all estimated labels for available recordings. The proposed method is cross-validated on a dataset consisting of 3456 PCG recordings from 1007 patients using a stratified ten-fold cross-validation. Additionally, the method was tested on a hidden test set comprised of 1538 PCG recordings from 442 patients. RESULTS: The overall cross-validation performances for patient-level murmur gradings are 86.3% and 81.6% in terms of the unweighted average of sensitivities and F1-scores, respectively. The sensitivities (and F1-scores) for absent, soft, and loud murmurs are 90.7% (93.6%), 75.8% (66.8%), and 92.3% (84.2%), respectively. On the test set, the algorithm achieves an unweighted average of sensitivities of 80.4% and an F1-score of 75.8%. CONCLUSIONS: This study provides a potential approach for algorithmic pre-screening in low-resource settings with relatively high expert screening costs. SIGNIFICANCE: The proposed method represents a significant step beyond detection of murmurs, providing characterization of intensity, which may provide an enhanced classification of clinical outcomes.

Diagnosis of Coexisting Valvular Heart Diseases Using Image-to-Sequence Translation of Contact Microphone Recordings.

OBJECTIVE: Development of a contact microphone-driven screening framework for the diagnosis of coexisting valvular heart diseases (VHDs). METHODS: A sensitive accelerometer contact microphone (ACM) is employed to capture heart-induced acoustic components on the chest wall. Inspired by the human auditory system, ACM recordings are initially transformed into Mel-frequency cepstral coefficients (MFCCs) and their first and second derivatives, resulting in 3-channel images. An image-to-sequence translation network based on the convolution-meets-transformer (CMT) architecture is then applied to each image to find local and global dependencies in images, and predict a 5-digit binary sequence, where each digit corresponds to the presence of a specific type of VHD. The performance of the proposed framework is evaluated on 58 VHD patients and 52 healthy individuals using a 10-fold leave-subject-out cross-validation (10-LSOCV) approach. RESULTS: Statistical analyses suggest an average sensitivity, specificity, accuracy, positive predictive value, and F1 score of 93.28%, 98.07%, 96.87%, 92.97%, and 92.4% respectively, for the detection of coexisting VHDs. Furthermore, areas under the curve (AUC) of 0.99 and 0.98 are respectively reported for the validation and test sets. CONCLUSION: The high performances achieved prove that local and global features of ACM recordings effectively characterize heart murmurs associated with valvular abnormalities. SIGNIFICANCE: Limited access of primary care physicians to echocardiography machines has resulted in a low sensitivity of 44% when using a stethoscope for the identification of heart murmurs. The proposed framework provides accurate decision-making on the presence of VHDs, thus reducing the number of undetected VHD patients in primary care settings.

Remote Teaching of Canine Cardiac Auscultation Using Digital Phonocardiograms.

Veterinary students often struggle to correctly interpret heart sounds. This study sought to evaluate if additional online training using digital phonocardiograms (DPCGs) improves students' ability to identify normal and pathologic heart sounds in dogs. Thirty-six randomly assigned veterinary students listened to and interpreted 10 audio recordings of normal heart sounds (2), heart murmurs (4), and arrhythmias (4) at the start and the end of a 4-week period. Twenty-two students participated in training with DPCGs, including those created from these recordings during this period, via a self-study website (n = 12) or online webinar (n = 10). Their results were compared with those of a control group (n = 14) that did not undergo additional training. Although pre- and post-training test scores did not differ between groups, both training groups showed within-group improvement between the two tests (p = .024, p = .037); the control group did not (p = .49). Although neither training group showed differences in ability to differentiate normal heart sounds from arrhythmias, both showed increased ability to detect and specify heart murmurs and provide refined diagnoses of detected arrhythmias. These results suggest additional training, even without actual patients, improves students' ability to identify heart murmurs and provide specific diagnoses for arrhythmias. Further study with a larger sample size and an additional group without DPCG-based training would help evaluate the effectiveness of DPCGs regarding arrhythmias. Studying a larger sample size would also allow for a training group participating in both training methods, measuring cumulative effectiveness of both methods.

Quantifying the Effect of Body Habitus on Cardiac Auscultation Via Computational Hemoacoustics.

The effect of body habitus on auscultation of heart murmurs is investigated via computational hemoacoustic modeling. The source of the heart murmur is first obtained from a hemodynamic simulation of blood flow through a stenosed aortic valve. This sound source is then placed at the aortic valve location in four distinct human thorax models, and the propagation of the murmur in each thorax model is simulated by solving the elastic wave equations in the time-domain. Placing the same sound source in different thorax models allows for the disambiguation of the effect of body habitus on cardiac auscultation. The surface acceleration resulting from the murmur on each subject's chest surface shows that subjects with higher body-mass index and thoracic cross-sectional area yield smaller acceleration values for the S1 sound. Moreover, the spectral analysis of the signal shows that slope from linear regression in the normal heart sound frequency range (10-150 Hz) is larger for children at the aortic, pulmonic, and mitral auscultation points compared to that for adults. The slope in the murmur frequency range (150-400 Hz) was larger for female subjects at the mitral point compared to that for male subjects. The trends from the results show the potential of the proposed computational method to provide quantitative insights regarding the effect of various anatomical factors on cardiac auscultation.

Correlation between Murmurs and Echocardiographic Findings; From an Imaging Cardiologist Point of View.

A heart murmur in adults is a common reason for referral for echocardiography at most general cardiology clinics in Europe. A murmur may indicate either a mild age-related valvular calcification or regurgitation, or represent a significant heart valve disease requiring valvular intervention. Generally, the correlation between murmurs by auscultation and severity of heart valve disease by echocardiography is poor. Particularly, the severity and characterization of diastolic murmurs by auscultation may poorly correlate with echocardiographic findings. This narrative review aims to summarize the differential diagnoses of physiological and pathological murmurs, describes the current referral practice of murmur patients for echocardiography, and presents a single-center experience on the correlation of auscultation and echocardiographic findings with a particular focus on aortic and mitral valve diseases. A careful auscultation of the heart prior to the echocardiogram is mandatory and may help to predict the echocardiographic findings and their interpretation in view of the clinical information. The correlation between clinical examination, point of care ultrasound and standard echocardiography is a matter of continued exploration.

A new approach to phonocardiogram severity analysis.

Phonocardiogram signal (PCG) has been the subject of several signal processing studies, where researchers applied various analysis techniques and extracted numerous features for different purposes, like cardiac pathologies identification, healthy/pathologic case discrimination, and severity assessment. When talking about cardiac severity, many think directly about the intensity or energy of the signal as the most reliable parameter. However, cardiac severity is not always reflected by the intensity or energy of the signal but includes other variables as well. In this paper, we will discuss the probability of having a Discrete Wavelet Transform (DWT) parameter that discriminates, identifies, and assesses the pathological cardiac severity levels, a parameter that takes into consideration other variables and elements for the severity study. For this purpose, we studied six PCGs signals that contain reduced murmurs (clicks) and eight murmur signals with four different cardiac severity levels. We extracted the Entropy of Approximation Coefficients (EAC) from the Discrete Wavelet Transform (DWT) sub-bands as the feature to study in this novel approach. The Energetic Ratio (ER) served as a reference parameter to evaluate the EAC evolution, due to its proven efficiency in cardiac severity tracking. While the DWT-EAC algorithm results revealed that the EAC provides better results for the paper purposes, the One versus All Support Vector Machine (OVA-SVM) classifier affirmed the efficiency of the Entropy of Approximation Coefficients (EAC) for cardiac severity assessment and proved the accuracy of this novel approach.

Challenging echocardiographic pictures of aortic arch: when heart murmur leads to a rare diagnosis.

We describe the case of a 2-week-old boy referred for systolic murmur. His echocardiography showed challenging pictures of the aortic arch, which led to the rare diagnosis of arterial tortuosity syndrome.

Synchrosqueezed Transform based Click Event Segmentation in Phonocardiogram of Mitral Valve Prolapse.

Phonocardiogram (PCG) signal of the mitral valve prolapse (MVP) patients is characterized by transient audio events which include a systolic click (SC) followed by a murmur of varying intensity. Physicians detect these auscultation clues in regular auscultation before ordering expensive echocardio-graphy test. But auscultation is often error prone and even physicians with considerable experience might end up missing these clues. Therefore developing machine learning techniques to help clinicians is the need of the hour. A segmentation technique using Fourier synchrosqueezed transform (FSST) features with a long short term memory (LSTM) network is proposed in this study. An accuracy of 99.8% on MVP dataset demonstrates the potential of the proposed method in clinical diagnosis.