Artificial Intelligence for Detecting Fetal Orofacial Clefts and Advancing Medical Education

2026-03-06

AI Summary

This study addresses the low prenatal diagnostic accuracy of fetal orofacial cleft anomalies, a challenge exacerbated by their rarity and the scarcity of specialized clinicians. The authors developed a deep learning–based artificial intelligence system trained on 45,139 fetal ultrasound images from 22 hospitals, marking the first implementation of AI for both clinical diagnosis and medical education in this domain. The system achieved a diagnostic sensitivity of 93% and specificity of 95%, comparable to that of experienced radiologists. When deployed as an AI-assisted tool, it improved diagnostic sensitivity among junior physicians by over 6%. Furthermore, a pilot educational trial involving 24 trainees demonstrated the system’s efficacy in accelerating clinical competency development, highlighting its dual utility in diagnostic support and medical training.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Humans and AI: Learning Human Values and Preferences
• Computer Vision: Learning & Optimization for CV

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Orofacial clefts are among the most common congenital craniofacial abnormalities, yet accurate prenatal detection remains challenging due to the scarcity of experienced specialists and the relative rarity of the condition. Early and reliable diagnosis is essential to enable timely clinical intervention and reduce associated morbidity. Here we show that an artificial intelligence system, trained on over 45,139 ultrasound images from 9,215 fetuses across 22 hospitals, can diagnose fetal orofacial clefts with sensitivity and specificity exceeding 93% and 95% respectively, matching the performance of senior radiologists and substantially outperforming junior radiologists. When used as a medical copilot, the system raises junior radiologists'sensitivity by more than 6%. Beyond direct diagnostic assistance, the system also accelerates the development of clinical expertise. A pilot study involving 24 radiologists and trainees demonstrated that the model can improve the expertise development for rare conditions. This dual-purpose approach offers a scalable solution for improving both diagnostic accuracy and specialist training in settings where experienced radiologists are scarce.

Problem

• fetal orofacial clefts
• prenatal detection
• congenital craniofacial abnormalities
• medical education
• diagnostic accuracy

Innovation

• artificial intelligence
• fetal orofacial clefts
• ultrasound diagnosis
• medical education
• diagnostic assistance