Congenital Heart Disease Classification Using Phonocardiograms: A Scalable Screening Tool for Diverse Environments

2025-03-28

AI Summary

To address the challenge of early congenital heart disease (CHD) screening in resource-limited settings, this study proposes a robust deep learning model for phonocardiogram (PCG) analysis tailored to global health needs. Methodologically, we employ a temporal modeling architecture and conduct cross-domain training using multi-source PCG data—including a locally collected cohort from Bangladesh and the PhysioNet 2016/2022 benchmarks. Notably, we demonstrate for the first time that the model maintains high robustness on single auscultation-site recordings and clinically labeled “non-diagnostic-quality” audio, achieving >85% and >80% accuracy, respectively. Key contributions include: state-of-the-art performance on the Bangladeshi dataset (94.1% accuracy, 92.7% sensitivity, 96.3% specificity); strong cross-dataset generalizability; and significantly enhanced deployability in primary-care settings and real-world clinical environments.

Technology Category

• Machine Learning: Hardware-aware ML
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Cognitive Modeling & Cognitive Systems: Affective Computing

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Congenital heart disease (CHD) is a critical condition that demands early detection, particularly in infancy and childhood. This study presents a deep learning model designed to detect CHD using phonocardiogram (PCG) signals, with a focus on its application in global health. We evaluated our model on several datasets, including the primary dataset from Bangladesh, achieving a high accuracy of 94.1%, sensitivity of 92.7%, specificity of 96.3%. The model also demonstrated robust performance on the public PhysioNet Challenge 2022 and 2016 datasets, underscoring its generalizability to diverse populations and data sources. We assessed the performance of the algorithm for single and multiple auscultation sites on the chest, demonstrating that the model maintains over 85% accuracy even when using a single location. Furthermore, our algorithm was able to achieve an accuracy of 80% on low-quality recordings, which cardiologists deemed non-diagnostic. This research suggests that an AI- driven digital stethoscope could serve as a cost-effective screening tool for CHD in resource-limited settings, enhancing clinical decision support and ultimately improving patient outcomes.

Problem

• Detects congenital heart disease using phonocardiogram signals
• Ensures high accuracy across diverse datasets and populations
• Provides cost-effective CHD screening in resource-limited settings

Innovation

• Deep learning model analyzes PCG signals
• High accuracy across diverse datasets
• Effective with single auscultation site