Fetal Weight Estimation Using Multivariate Gaussian Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating fetal weight using ultrasonic biometric parameters and pregnancy-related information are unreliable, particularly for microsomic and macrosomic fetuses, resulting in significant errors and increased healthcare risks due to human and instrumental measurement inaccuracies.
Innovation Solution
A method utilizing a mathematical multivariate Gaussian probabilistic model, combined with Artificial Neural Networks (ANNs), to predict fetal weight by estimating the mean vector and covariance matrix, providing unbiased estimates and evaluating the reliability of measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If statistical regression models are used for fetal weight estimation, then the method is simple to implement, but the reliability and accuracy are insufficient with average absolute error never below 7-8%
Solution Approach 1:
The patent replaces statistical regression models with a physics-based biomechanical model that simulates fetal growth according to physical principles. This substitution transforms the estimation approach from empirical curve-fitting to a mechanistic model that calculates fetal weight based on biometric measurements and growth dynamics, thereby improving reliability while maintaining implementation feasibility through computer-based calculations
Solution Approach 2:
The patent introduces additional parameters beyond standard biometric measurements, including maternal physiological parameters (uterine artery resistance, cardiac output) and fetal physiological parameters (heart rate variability, metabolic rate). By expanding the parameter set and using them in a physics-based model, the system achieves higher accuracy (reducing absolute error below 7-8%) while managing complexity through systematic parameter integration
2Loss of information
If correction factors and new information such as amniotic fluid amount and maternal pathologies are introduced, then more data is available for estimation, but measurement difficulties increase especially for operators with little experience
Solution Approach 1:
The patent develops a unified biomechanical model that simultaneously processes multiple types of data (biometric measurements, maternal physiological parameters, fetal physiological parameters) through a single physics-based framework. This multi-functional approach integrates diverse information sources without requiring separate measurement protocols, thereby reducing measurement difficulty while maintaining data completeness
Solution Approach 2:
The patent introduces computational algorithms as intermediaries that automatically process and integrate multiple parameters. The computer-based system performs complex calculations involving biometric measurements, maternal physiological data, and fetal physiological data, eliminating the need for operators to manually apply correction factors or interpret multiple parameters, thus reducing measurement difficulty while preserving information completeness
3Ease of operation
If standard ultrasound biometric parameters are used, then the measurement procedure is simple, but the accuracy is insufficient particularly for microsomic and macrosomic fetuses
Solution Approach 1:
The patent segments the fetal weight estimation problem into multiple independent measurement components (head circumference, abdominal circumference, femur length, uterine artery resistance, cardiac output) that are processed separately through the biomechanical model. This segmentation allows each parameter to be measured using standard simple procedures while the model integrates them to achieve high precision for both microsomic and macrosomic fetuses
Solution Approach 2:
The patent transitions from two-dimensional ultrasound biometric measurements to a multi-dimensional physics-based model that incorporates spatial, temporal, and physiological dimensions. By adding dimensions such as maternal cardiac output, uterine artery resistance, and fetal metabolic rate to the traditional biometric measurements, the system achieves improved precision for extreme fetal weights while maintaining simplicity of individual measurements
Data Source
AI summary
A method for fetal weight estimation wherein fetal biometric parameters are ultrasonically measured and further physiological and phenomenological pregnancy parameters are determined for a plurality of sample cases. The fetal weight for each of the sample cases is then determined by precision weighing at birth. A database is then created based on the known sample cases and mathematical prediction models are generated from the database. For the case under examination for which fetal weight is to be predicted, the fetal biometric parameters are ultrasonically measured and further physiological and phenomenological pregnancy parameters are determined. The fetal weight of the case under examination is predicted by using a mathematical prediction model and a multinormal probabilistic model is used as a mathematical model.