Beyond Heart Murmur Detection: Automatic Murmur Grading from Phonocardiogram

This study aims to develop an algorithmic approach for automatic murmur grading from phonocardiograms (PCGs) of a large population of pediatric patients from a low-resource rural area. Murmurs are abnormal heart sounds that can be identified by experts through cardiac auscultation, and the murmur grade, which is a quantitative measure of the murmur intensity, is strongly correlated with the patient's clinical condition. The proposed method uses Mel spectrogram representation of each PCG recording and 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 method was cross-validated on a dataset consisting of 3456 PCG recordings from 1007 patients using a stratified ten-fold cross-validation and tested on a hidden test set comprised of 1538 PCG recordings from 442 patients. The overall cross-validation performances for patient-level murmur gradings are 86.3% in terms of unweighted average sensitivities and 81.6% in terms of F1-scores. On the test set, the algorithm achieved an unweighted average sensitivity score of 80.4% and an F1-score of 75.8%. The sensitivities (and F1 scores) for absent, soft, and loud murmurs were 90.7% (93.6%), 75.8% (66.8%), and 92.3% (84.2%) respectively; indicating that the proposed method provides characterization beyond detection that may enhance classification accuracy in resource limited settings where expert screening costs are relatively high.

Automatic Murmur Grading from Phonocardiograms of Pediatric Patients: An Algorithmic Approach

Cardiac auscultation is an important tool used by medical professionals to detect and diagnose heart conditions. One of the most common findings during cardiac auscultation is the presence of murmurs, which are abnormal heart sounds that can be indicative of underlying health issues. The murmur grade, which is a quantitative measure of the murmur intensity, can provide valuable information about the patient's clinical condition. However, accurately grading murmurs requires expertise and experience that may not be available in low-resource rural areas.

In order to address this issue, researchers have developed an algorithmic approach for automatic murmur grading from phonocardiograms (PCGs) of pediatric patients from such settings. PCGs are recordings made with specialized microphones placed on the chest wall to capture heart sounds and vibrations associated with various cardiac events. In this study, they proposed a method using Mel spectrogram representation of each PCG recording and an ensemble of 15 convolutional residual neural networks with channel-wise attention mechanisms to classify each PCG recording into one of three categories: absent (no murmur), soft (low intensity), or loud (high intensity).

The dataset used for cross-validation consisted of 3456 PCG recordings from 1007 patients using stratified ten-fold cross-validation; while the test set was comprised of 1538 PCG recordings from 442 patients. The overall cross-validation performances for patient-level murmur gradings were 86.3% in terms unweighted average sensitivities and 81.6% in terms F1 scores; while on the test set it achieved an unweighted average sensitivity score 80.4% and F1 score 75.8%. The sensitivities (and F1 scores) for absent, soft, and loud murmurs were 90.7% (93%), 75%, 8% (66%), 8%) and 92%, 3%(84%), 2%) respectively; indicating that their proposed method provides characterization beyond detection that may enhance classification accuracy even in resource limited settings where expert screening costs are relatively high..

The results suggest that automated algorithms could potentially help reduce workloads for clinicians as well as improve diagnostic accuracy in low resource settings where access to experts may be limited or unavailable due to geographical constraints or financial limitations . Furthermore , these algorithms could also enable more accurate diagnosis through improved characterization beyond mere detection , thus providing better care options for pediatric patients living in rural areas .