Birth weight prediction from remote ultrasonographic examination

Area of Science:

• Medical imaging
• Obstetrics
• Biostatistics

Background:

• Accurate fetal birth weight estimation is crucial for clinical management.
• Existing static formulas have limitations in predicting birth weight over time.
• Remote ultrasonographic parameters offer a potential alternative for birth weight assessment.

Purpose of the Study:

• To develop and validate new equations for estimating fetal birth weight using remote ultrasonographic parameters.
• To incorporate the examination-to-delivery interval (lapse time) into birth weight estimation models.
• To assess the accuracy and reliability of these new equations compared to existing methods.

Main Methods:

• Multiple linear regression (least squares method) was employed to remodel existing static equations.
• Five existing equations were adapted by incorporating lapse time.
• The developed equations were validated on 167 independent cases.

Main Results:

• R-squared values approached 1.0 for all developed models.
• 65-67% of birth weight estimates were within 100 mg/kg using the best equations.
• Mean absolute error ranged from 86 to 91 gm/kg.
• Accuracy was maintained across various lapse times and birth weights.

Conclusions:

• New dynamic equations accurately estimate fetal birth weight from remote ultrasound parameters up to 35 days post-examination.
• The developed method provides reliable birth weight estimations, comparable or superior to existing static formulas.
• This approach is valuable in intrapartum settings where direct ultrasonography may be challenging or impossible.