Congenital heart defects (CHD) remain a leading cause of neonatal mortality worldwide, with current detection rates as low as 35–50%. Early and accurate diagnosis is clinically challenging due to the limited availability of pediatric cardiology specialists and the complexity of echocardiographic imaging in newborns.
Ultra Echo Scan (UES) is an AI-powered diagnostic support system developed to assist medical professionals in identifying CHD during standard neonatal echocardiographic (USG) examinations. UES analyzes ultrasound recordings from four standard cardiac projections, automatically selecting optimal frames and applying advanced image stabilization, denoising, segmentation, and morphological feature enhancement techniques. The refined data is classified using one of three deep convolutional neural network models, each tailored to specific imaging characteristics and measurement variances.
With predictive performance exceeding 85%, UES reaches the diagnostic proficiency of expert pediatric cardiologists at reference centers. In early testing, the system demonstrated 96% accuracy in detecting 90% of CHD cases. Its scalable, browser-based delivery model requires no specialized hardware or operator expertise—making it a powerful candidate for future CHD screening.
By equipping neonatologists, pediatricians, and cardiologists with an accessible and reliable diagnostic tool, UES addresses a critical gap in early CHD detection—enabling timely intervention, reducing infant mortality, and preventing long-term complications.
