Multimodal Nomogram for the Prenatal Risk Assessment of Hypoplastic Left Heart Syndrome Using Self-Supervised Learning

The authors aim to develop and validate a multimodal nomogram for the prenatal risk assessment of hypoplastic left heart syndrome (HLHS) by identifying significant risk factors through a retrospective study of pregnancies. They utilized a self-supervised learning approach to extract features from cardiac ultrasound images, resulting in a nomogram that demonstrated high accuracy and outperformed traditional models and less experienced sonographers. The study highlights key risk factors, such as the area ratios of the heart chambers, which warrant further clinical investigation.