Multi-Center Study on Deep Learning-Assisted Detection and Classification of Fetal Central Nervous System Anomalies Using Ultrasound Imaging
Yang Qi, Jiaxin Cai, Jing Lu, Runqing Xiong, Rongshang Chen, Liping Zheng, Duo Ma
TL;DR
The paper addresses the challenge of prenatal detection and classification of fetal CNS malformations from ultrasound by proposing a multicenter deep learning pipeline built on ResNet34. It employs leave-one-out cross-validation, weighted cross-entropy to handle class imbalance, and Grad-CAM heatmaps for localization to improve interpretability. Key findings include subject-level accuracy of $94.5\%$ with AUROC $0.993$, robust performance across gestational stages, and a five-class extension including a normal class that achieves $0.9825$ accuracy and micro/macro AUCs of $0.97$ and $0.95$, respectively; a retrospective reader study also demonstrates that integrating DL predictions with radiologist judgment enhances diagnostic accuracy and efficiency. The work highlights the potential for AI-assisted prenatal screening to reduce misdiagnosis and radiologist workload while maintaining clinical interpretability through heatmaps, supporting broader deployment in varied clinical settings.
Abstract
Prenatal ultrasound evaluates fetal growth and detects congenital abnormalities during pregnancy, but the examination of ultrasound images by radiologists requires expertise and sophisticated equipment, which would otherwise fail to improve the rate of identifying specific types of fetal central nervous system (CNS) abnormalities and result in unnecessary patient examinations. We construct a deep learning model to improve the overall accuracy of the diagnosis of fetal cranial anomalies to aid prenatal diagnosis. In our collected multi-center dataset of fetal craniocerebral anomalies covering four typical anomalies of the fetal central nervous system (CNS): anencephaly, encephalocele (including meningocele), holoprosencephaly, and rachischisis, patient-level prediction accuracy reaches 94.5%, with an AUROC value of 99.3%. In the subgroup analyzes, our model is applicable to the entire gestational period, with good identification of fetal anomaly types for any gestational period. Heatmaps superimposed on the ultrasound images not only provide a visual interpretation for the algorithm but also provide an intuitive visual aid to the physician by highlighting key areas that need to be reviewed, helping the physician to quickly identify and validate key areas. Finally, the retrospective reader study demonstrates that by combining the automatic prediction of the DL system with the professional judgment of the radiologist, the diagnostic accuracy and efficiency can be effectively improved and the misdiagnosis rate can be reduced, which has an important clinical application prospect.