Application of Artificial Intelligence in Detecting Fetal Abnormalities from Ultrasound Images

Ultrasound is one of the most important and widely used imaging modalities during pregnancy, playing a fundamental role in assessing fetal growth, examining anatomical structures, and identifying congenital anomalies. Despite its immense importance, the interpretation of ultrasound images relies heavily on the experience of the physician and the sonographer; factors such as fetal position and movement, equipment quality, maternal abdominal wall thickness, gestational age, and inter-observer variability can influence the examination results. In recent years, advancements in artificial intelligence, machine learning, and—specifically—deep learning have paved the way for the automated analysis of fetal ultrasound images. AI models are capable of classifying images, identifying standard ultrasound planes, localizing and segmenting anatomical structures, performing fetal measurements, and ultimately detecting findings suggestive of abnormalities. Recent research indicates that AI shows promise in areas such as the automated detection of standard fetal brain, abdominal, thoracic, and cardiac planes, biometric measurements, and the identification of specific markers of abnormality. In a multi-center evaluation published in ۲۰۲۶, an AI system designed to detect eight ultrasound findings associated with congenital anomalies achieved an average sensitivity of ۹۳.۲% and a specificity of ۹۰.۸% across more than ۶,۴۰۰ images from ۱,۱۱۵ examinations. However, artificial intelligence is not yet a substitute for a specialist physician, and its widespread clinical application requires external validation, prospective studies, bias assessment, model interpretability, and expert oversight. By reviewing the research background, technical foundations, and clinical applications of artificial intelligence in the analysis of fetal ultrasound images, this article examines the capabilities, limitations, and future of this technology.

کلیدواژه ها:

Artificial intelligence ، deep learning ، fetal ultrasound ، congenital anomalies ، prenatal diagnosis ، machine learning ، convolutional neural network ، medical imaging ، automated fetal diagnosis