Application of intelligent phonocardiography in the detection of congenital heart disease in pediatric patients: A narrative review

Background: Congenital heart disease (CHD) is the most common birth defect. Phonocardiography (PCG) is the graphical representation of heart sounds. Utilizing machine learning and artificial neural networks, recent models have utilized phonocardiographic signals to predict structural heart disease. Aim of review: It is the aim of this review to summarize and organize the evidence reporting the use of intelligent phonocardiography in the diagnosis of congenital heart disease. Key scientific concepts of review: Since the early twenty-first century, models have demonstrated accuracy and sensitivity in predicting whether a murmur is pathologic or innocent using PCG. More recent evidence demonstrates similar predictive efficacy in specific lesions, including aortic stenosis and regurgitation. Further research with larger populations is required to further characterize the utility of PCG and artificial intelligence to diagnose CHD in children. However, the success of this technology has important roles in education and significant implications in resource-poor areas. Further, these studies support the expanded use of artificial intelligence in pediatric cardiology and may prove useful not only in PCG but also in electrocardiography and echocardiography.