Related Experiment Video
Updated: Aug 5, 2026
09:31
Hemodynamic Precision in the Neonatal Intensive Care Unit using Targeted Neonatal Echocardiography
Published on: January 27, 2023
Fully Automated Biometric Parameter Measurement in Prenatal Ultrasound Screening for Total Anomalous Pulmonary Venous
Rina Aoyama1, Naoaki Harada2,3,4, Masaaki Komatsu3,5
1Department of Obstetrics and Gynecology, Showa Medical University School of Medicine, 1-5-8 Hatanodai, Shinagawa-ku, Tokyo 142-8666, Japan.
Bioengineering (Basel, Switzerland)
|July 28, 2026
Summary
Automated analysis of fetal cardiac ultrasound videos improves prenatal detection of total anomalous pulmonary venous connection (TAPVC). Novel methods using artificial intelligence show performance comparable to experts, aiding less experienced examiners.
Area of Science:
- Medical Imaging
- Cardiology
- Artificial Intelligence in Medicine
Background:
- Total anomalous pulmonary venous connection (TAPVC) is a critical congenital heart defect with suboptimal prenatal detection rates.
- The post-left atrium space (PLAS) index and left-atrial posterior-space-to-diagonal (LAPSD) ratio in the four-chamber view (4CV) are proposed biometric parameters for TAPVC screening.
Purpose of the Study:
- To develop and evaluate an automated approach for extracting the 4CV and measuring PLAS index and LAPSD ratio from fetal cardiac ultrasound videos.
- To assess the screening performance of automated methods compared to manual measurements and expert performance.
Main Methods:
- Deep learning models (DeepLabv3+, UNet3+, SegFormer) were employed for automated segmentation of cardiac structures.
- Automated extraction of the 4CV and measurement of PLAS index and LAPSD ratio were performed.
- Performance was evaluated using Area Under the Receiver Operating Characteristic Curve (AUC) in a clinical comparison study.
Main Results:
- Automated methods achieved screening performance comparable to manual 4CV extraction.
- Fully automated approaches (AE-DeepLabv3+, AE-UNet3+, AE-SegFormer) consistently outperformed novice examiners (residents and fellows).
- AUC values for automated methods ranged from 0.903 to 0.940, approaching expert performance (0.996).
Conclusions:
- The developed automated approach shows significant potential to enhance prenatal TAPVC screening accuracy.
- This technology can support less experienced healthcare providers, streamline workflows, and improve early detection rates for TAPVC.
