Publicly Available Fetal Ultrasound Databases: Current Status and Future Directions

Area of Science:

• Medical Imaging
• Biomedical Informatics
• Obstetrics

Background:

• Fetal ultrasound imaging is essential for prenatal diagnosis and monitoring.
• Standardization and accessibility of public fetal ultrasound databases are limited, hindering research and clinical translation.
• Advancements in antenatal care rely on robust, diverse, and well-characterized imaging datasets.

Purpose of the Study:

• To review and analyze publicly available fetal ultrasound databases published between 2017 and 2024.
• To identify current trends, challenges, and opportunities in fetal ultrasound data repository development.
• To propose recommendations for enhancing future database creation and utilization in clinical practice.

Main Methods:

• A systematic literature search was conducted across Web of Science, Scopus, and Google Scholar.
• Fourteen public fetal ultrasound databases were identified and analyzed based on data type, size, and content.
• Technical validation metrics and emerging development trends were assessed.

Main Results:

• The analyzed databases vary widely in size (25-15,728 images/sequences) and data modality (primarily 2D, with emerging 3D and Doppler data).
• High technical accuracy (82-98%) was reported, but significant limitations exist regarding standardization, pathological case inclusion, and demographic representation.
• Key trends include phantom data, transfer learning for low-resource settings, and synthetic data generation.

Conclusions:

• Public fetal ultrasound databases show progress but require improved standardization and data diversity for broader clinical impact.
• Future databases should prioritize clinical workflow integration, standardized quality metrics, and inclusive data collection.
• Addressing these gaps will facilitate the translation of technical advancements into improved antenatal care.