An algorithmic approach to accurate ultrasonic fetal weight estimation

Area of Science:

• Perinatology and Obstetrics
• Medical Imaging and Ultrasound
• Biometry and Anthropometry

Background:

• Accurate fetal weight estimation is crucial for optimal pregnancy management and neonatal outcomes.
• Existing fetal weight estimation models often have limitations in accuracy across diverse fetal biometry.
• Ultrasound biometry, including head and abdominal measurements, is a standard tool for fetal assessment.

Purpose of the Study:

• To develop and validate a flexible, adaptable model for fetal weight estimation.
• To improve the accuracy of predicted fetal weight using standard ultrasound-derived fetal dimensions.
• To utilize gestational age-independent decision rules based on fetal body proportions for formula selection.

Main Methods:

• A flexible mathematical model was developed for fetal weight prediction.
• Input parameters included biparietal diameter, occipitofrontal diameter, average abdominal diameter, and femoral shaft length.
• Gestational age-independent decision rules were implemented to select appropriate individual formulas based on fetal proportions.

Main Results:

• The proposed model demonstrated a high correlation coefficient (r = .994) between observed and predicted fetal weights.
• The study included 96 fetuses, with measurements taken within 36 hours of delivery.
• Prediction errors were low: 20% of cases had errors < 10 g/kg, and 90% had errors < 80 g/kg.

Conclusions:

• The flexible fetal weight estimation model offers high accuracy and reliability.
• The model's adaptability to varying fetal body proportions enhances its clinical utility.
• This approach has the potential to refine prenatal assessment and improve perinatal care.