Estimating Gestational Age

This study developed a neural network model to estimate gestational age using blind ultrasound sweeps from pregnant women in North Carolina and Zambia. The model demonstrated greater accuracy in estimating gestational age than traditional fetal biometry, with a mean absolute error of 3.9 days compared to 4.7 days. These findings suggest the potential for AI-based ultrasound to improve access to quality obstetric care, particularly in low-resource settings.