Automated Critical Congenital Heart Disease Screening Combining Non-Invasive Measurements of Oxygenation and Perfusion

Abstract

Researchers at the University of California, Davis have developed a computer-implemented method for accurately classifying congenital heart defects in newborns using pulse oximetry and machine learning.

Full Description

This technology employs dual pulse oximetry to acquire physiological measurements from both upper and lower extremities of neonates. A predictive model, powered by machine learning, analyzes these measurements to identify potential critical congenital heart defects (CCHD). The model is trained using a vast dataset of neonatal physiological readings, incorporating advanced feature selection techniques to refine its diagnostic accuracy.

Applications

• Hospital neonatal care units for routine CCHD screening. 
• Pediatric cardiac diagnostic and treatment centers. 
• Development of portable diagnostic devices for low-resource settings. 
• Integration with healthcare IT systems for predictive analytics in neonatal care.

Features/Benefits

• Enhanced early detection of CCHD, reducing mortality and morbidity. 
• Non-invasive and real-time monitoring of neonatal cardiac health. 
• Uses dual pulse oximetry for comprehensive physiological assessment. 
• Machine learning model continuously improves with more data. 
• Automated feature selection optimizes model accuracy and efficiency. 
• Reduces the number of missed CCHD diagnoses in newborns. 
• Minimizes the reliance on post-discharge diagnosis that can lead to increased risk. 
• Addresses limitations of current oxygen-saturation based CCHD screening protocols. 
• Improves detection of defects like coarctation of the aorta, often missed by conventional methods.

Patent Status