Predicting Fetal Growth with Curve Fitting and Machine Learning

Area of Science:

• Obstetrics and Gynecology
• Medical Imaging
• Biostatistics

Background:

• Accurate fetal growth monitoring is crucial for identifying developmental issues.
• Existing fetal growth references may not accurately reflect diverse populations.
• Population-specific references are needed for precise prenatal care.

Purpose of the Study:

• To develop a Taiwan-specific fetal growth reference chart.
• To utilize a web-based platform for data collection and analysis.
• To implement real-time anomaly detection for fetal biometric parameters.

Main Methods:

• Collected ultrasound data from 980 pregnant women (8350 scans).
• Modeled six key fetal biometric parameters using polynomial regression (quadratic).
• Developed a web-based platform for data management and analysis.

Main Results:

• Achieved R-squared values over 0.95 for most modeled fetal parameters.
• Established a Taiwan-specific fetal growth reference.
• Integrated confidence intervals and real-time anomaly detection into the platform.

Conclusions:

• The developed Taiwan-specific fetal growth reference enables efficient monitoring.
• Population-specific charts improve the accuracy of fetal growth assessment.
• This approach has significant potential for clinical application in prenatal care.