A fusion framework based on multi-domain features and deep learning features of phonocardiogram for coronary artery disease detection

A novel feature-level fusion scheme with multimodal attention CNN for heart sound classification

BACKGROUND AND OBJECTIVE

Most of the existing machine learning-based heart sound classification methods achieve limited accuracy. Since they primarily depend on single domain feature information and tend to focus equally on each part of the signal rather than employing a selective attention mechanism. In addition, they fail to exploit convolutional neural network (CNN) - based features with an effective fusion strategy.

METHODS

In order to overcome these limitations, a novel multimodal attention convolutional neural network (MACNN) with a feature-level fusion strategy, in which Mel-cepstral domain as well as general frequency domain features are incorporated to increase the diversity of the features, is proposed in this paper. In the proposed method, DilationAttenNet is first utilized to construct attention-based CNN feature extractors and then these feature extractors are jointly optimized in MACNN at the feature-level. The attention mechanism aims to suppress irrelevant information and focus on crucial diverse features extracted from the CNN.

RESULTS

Extensive experiments are carried out to study the efficacy of the feature level fusion in comparison to that with early fusion. The results show that the proposed MACNN method significantly outperforms the state-of-the-art approaches in terms of accuracy and score for the two publicly available Github and Physionet datasets.

CONCLUSION

The findings of our experiments demonstrated the high performance for heart sound classification based on the proposed MACNN, and hence have potential clinical usefulness in the identification of heart diseases. This technique can assist cardiologists and researchers in the design and development of heart sound classification methods.

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.

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.

Automatic Detection and Classification of Cardiovascular Disorders Using Phonocardiogram and Convolutional Vision Transformers

The major cause of death worldwide is due to cardiovascular disorders (CVDs). For a proper diagnosis of CVD disease, an inexpensive solution based on phonocardiogram (PCG) signals is proposed. (1) Background: Currently, a few deep learning (DL)-based CVD systems have been developed to recognize different stages of CVD. However, the accuracy of these systems is not up-to-the-mark, and the methods require high computational power and huge training datasets. (2) Methods: To address these issues, we developed a novel attention-based technique (CVT-Trans) on a convolutional vision transformer to recognize and categorize PCG signals into five classes. The continuous wavelet transform-based spectrogram (CWTS) strategy was used to extract representative features from PCG data. Following that, a new CVT-Trans architecture was created to categorize the CWTS signals into five groups. (3) Results: The dataset derived from our investigation indicated that the CVT-Trans system had an overall average accuracy ACC of 100%, SE of 99.00%, SP of 99.5%, and F1-score of 98%, based on 10-fold cross validation. (4) Conclusions: The CVD-Trans technique outperformed many state-of-the-art methods. The robustness of the constructed model was confirmed by 10-fold cross-validation. Cardiologists can use this CVT-Trans system to help patients with the diagnosis of heart valve problems.

Improved pre-test likelihood estimation of coronary artery disease using phonocardiography

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.

Effective features extraction by analyzing heart sound for identifying cardiovascular diseases related to COVID-19: A diagnostic model

Incidence and exacerbation of some of the cardiovascular diseases in the presence of the coronavirus will lead to an increase in the mortality rate among patients. Therefore, early diagnosis of such diseases is critical, especially during the COVID-19 pandemic (mild COVID-19 infection). Thus, for diagnosing the heart diseases related to the COVID-19, an automatic, non-invasive, and inexpensive method based on the heart sound processing approach is proposed. In the present study, a set of features related to the nature of heart signals is defined and extracted. The investigated features included morphological and statistical features in the heart sound frequencies. By extracting and selecting a set of effective features related to the mentioned diseases, and avoiding to use different segmentation and filtering techniques, dependence on a limited dataset and specific sampling procedures has been eliminated. Different classifiers with various kernels are applied for diagnosis in data unbalanced and balanced conditions. The results showed 93.15% accuracy and 93.72% F1-score using 60 effective features in data balanced conditions. The identification system using the extracted features from Azad dataset is able to achieve the desired results in a generalized dataset. In this way, in the shortest possible sampling time, the present system provided an effective and generalizable method and a practical model for diagnosing important cardiovascular diseases in the presence of coronavirus in the COVID-19 pandemic.

A multimodal parallel method for left ventricular dysfunction identification based on phonocardiogram and electrocardiogram signals synchronous analysis

Heart failure (HF) is widely acknowledged as the terminal stage of cardiac disease and represents a global clinical and public health problem. Left ventricular ejection fraction (LVEF) measured by echocardiography is an important indicator of HF diagnosis and treatment. Early identification of LVEF reduction and early treatment is of great significance to improve LVEF and the prognosis of HF. This research aims to introduce a new method for left ventricular dysfunction (LVD) identification based on phonocardiogram (ECG) and electrocardiogram (PCG) signals synchronous analysis. In the present study, we established a database called Synchronized ECG and PCG Database for Patients with Left Ventricular Dysfunction (SEP-LVDb) consisting of 1046 synchronous ECG and PCG recordings from patients with reduced (n = 107) and normal (n = 699) LVEF. 173 and 873 recordings were available from the reduced and normal LVEF group, respectively. Then, we proposed a parallel multimodal method for LVD identification based on synchronous analysis of PCG and ECG signals. Two-layer bidirectional gate recurrent unit (Bi-GRU) was used to extract features in the time domain, and the data were classified using residual network 18 (ResNet-18). This research confirmed that fused ECG and PCG signals yielded better performance than ECG or PCG signals alone, with an accuracy of 93.27%, precision of 93.34%, recall of 93.27%, and F1-score of 93.27%. Verification of the model's performance with an independent dataset achieved an accuracy of 80.00%, precision of 79.38%, recall of 80.00% and F1-score of 78.67%. The Bi-GRU model outperformed Bi-directional long short-term memory (Bi-LSTM) and recurrent neural network (RNN) models with a best selection frame length of 3.2 s. The Saliency Maps showed that SEP-LVDPN could effectively learn features from the data.

RF-CNN-F: random forest with convolutional neural network features for coronary artery disease diagnosis based on cardiac magnetic resonance

Coronary artery disease (CAD) is a prevalent disease with high morbidity and mortality rates. Invasive coronary angiography is the reference standard for diagnosing CAD but is costly and associated with risks. Noninvasive imaging like cardiac magnetic resonance (CMR) facilitates CAD assessment and can serve as a gatekeeper to downstream invasive testing. Machine learning methods are increasingly applied for automated interpretation of imaging and other clinical results for medical diagnosis. In this study, we proposed a novel CAD detection method based on CMR images by utilizing the feature extraction ability of deep neural networks and combining the features with the aid of a random forest for the very first time. It is necessary to convert image data to numeric features so that they can be used in the nodes of the decision trees. To this end, the predictions of multiple stand-alone convolutional neural networks (CNNs) were considered as input features for the decision trees. The capability of CNNs in representing image data renders our method a generic classification approach applicable to any image dataset. We named our method RF-CNN-F, which stands for Random Forest with CNN Features. We conducted experiments on a large CMR dataset that we have collected and made publicly accessible. Our method achieved excellent accuracy (99.18%) using Adam optimizer compared to a stand-alone CNN trained using fivefold cross validation (93.92%) tested on the same dataset.

A multi-scale and multi-domain heart sound feature-based machine learning model for ACC/AHA heart failure stage classification

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.

Ensembled Transfer Learning and Multiple Kernel Learning for Phonocardiogram based Atherosclerotic Coronary Artery Disease Detection

Conventional machine learning has paved the way for a simple, affordable, non-invasive approach for Coronary artery disease (CAD) detection using phonocardiogram (PCG). It leaves a scope to explore improvement of performance metrics by fusion of learned representations from deep learning. In this study, we propose a novel, multiple kernel learning (MKL) for their fusion using deep embeddings transferred from pre-trained convolutional neural network (CNN). The proposed MKL, finds optimal kernel combination by maximizing the similarity with ideal kernel and minimizing the redundancy with other basis kernels. Experiments are performed on 960 PCG epochs collected from 40 CAD and 40 normal subjects. The transferred embeddings attain maximum subject-level accuracy of 89.25% with kappa of 0.7850. Later, their fusion with handcrafted features using the proposed MKL gives an accuracy of 91.19% and kappa 0.8238. The study shows the potential of development of high accuracy CAD detection system by using easy to acquire, non-invasive PCG signal.

Phonocardiogram signal analysis for classification of Coronary Artery Diseases using MFCC and 1D adaptive local ternary patterns

Coronary Artery Diseases (CADs) are a dominant cause of worldwide fatalities. The development of accurate and timely diagnosis routines is imperative to reduce these risks and mortalities. Coronary angiography, an invasive and expensive technique, is currently used as a diagnostic tool for the detection of CAD but it has some procedural hazards, i.e., it requires arterial puncture, and the subject gets exposed to iodinated radiation. Phonocardiography (PCG), a non-invasive and inexpensive technique, is a modality employing heart sounds to diagnose heart diseases but it requires only trained medical personnel to apprehend cardiac murmurs in clinical environments. Furthermore, there is a strong compulsion to characterize CAD into its types, such as Single vessel coronary artery disease (SVCAD), Double vessel coronary artery disease (DVCAD), and Triple vessel coronary artery disease (TVCAD) to assist the cardiologist in decision making about the treatment procedure followed. This paper presents a computer-aided diagnosis system for the categorization of CAD and its types based on Phonocardiogram (PCG) signal analysis. The raw PCG signals were denoised via empirical mode decomposition (EMD) to remove redundant information and noise. Next, we extract MFCC and proposed 1D-Adaptive Local Ternary Patterns (1D-ALTP) and fused them serially to get a strong feature representation of multiple PCG signal classes. Features were further reduced through Multidimensional Scaling (MDS) and subjected to several classification methods such as support vector machines (SVM), Decision Tree (DT), and K-nearest neighbors (KNN) in a comparative fashion. The best classification performances of 98.3% and 97.2% mean accuracies were obtained through SVM with the cubic kernel for binary and multiclass experiments, respectively. The performance of the proposed system is comprehensively tested through 10-fold cross-validation and hold-out train-test techniques to avoid model overfitting. Comparative analysis with existing approaches advocates the superiority of the proposed approach.