3D OCT Embryo Classification With Quantitative Morphology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating blastocyst embryos rely on subjective two-dimensional microscopic images, leading to variability in classification based on embryologist skill, hindering the success rate of assisted reproductive technology procedures.
Innovation Solution
A method using optical coherence tomography to acquire a three-dimensional image of blastocyst embryos, calculating specific parameters, and employing a trained machine learning model to classify the embryos based on these parameters, including volume, thickness, and cell count of the inner cell mass and trophectoderm.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a two-dimensional microscopic image is used for embryo evaluation, then the evaluation process is simple and quick, but the classification accuracy is low and varies depending on embryologist skill
Solution Approach 1:
The patent transitions from two-dimensional microscopic images to three-dimensional imaging using optical coherence tomography (OCT). This dimensional change enables comprehensive observation of embryo structure including inner cell mass, trophectoderm, and overall morphology, thereby improving classification accuracy while reducing dependence on embryologist expertise
Solution Approach 2:
The patent replaces the subjective mechanical evaluation process with an automated machine learning system. The machine learning model objectively analyzes three-dimensional embryo images and provides quantitative classification, eliminating variability introduced by different embryologists' skills and experience levels
2Reliability
If subjective determination by embryologist is used, then the evaluation process is simple to implement, but the reliability of classification results varies
Solution Approach 1:
The patent implements an automated evaluation system where the machine learning model independently performs embryo classification without requiring continuous human intervention. The system processes three-dimensional images, extracts relevant features, and generates classification results autonomously, ensuring consistent and reliable outcomes
Solution Approach 2:
The patent employs a trained machine learning model that has been developed using training data with ground truth labels. The model learns from this feedback during training and applies the learned patterns to new embryo images, improving classification reliability through data-driven decision-making rather than subjective judgment
3Measurement precision
If three-dimensional imaging by optical coherence tomography is used, then comprehensive embryo structure observation is enabled, but the complexity of quantitative classification increases
Solution Approach 1:
The patent segments the three-dimensional embryo image into distinct structural components including inner cell mass, trophectoderm, and overall embryo boundaries. This segmentation enables precise measurement of specific parameters such as inner cell mass volume, trophectoderm thickness, and cell distribution patterns
Solution Approach 2:
The patent transforms complex three-dimensional image data into quantifiable parameters such as volume measurements, thickness distributions, and cell counts. By converting visual information into numerical parameters, the system simplifies the classification process while maintaining high measurement precision
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables objective and quantitative classification of blastocyst embryos, improving the success rate of assisted reproductive technology by reducing reliance on embryologist skill and providing accurate classification results.
Implementation Method 1
acquiring a three-dimensional image of the blastocyst embryo by optical coherence tomography
Data Source
Figure 1
Figure 2
Figure 3
AI summary
First, a three-dimensional image (D2) of a blastocyst embryo is acquired by optical coherence tomography. Secondly, a parameter (P) representing at least one of a volume of an inner cell mass, a thickness of an inner cell mass, a thickness distribution of an inner cell mass, a volume of a trophectoderm, a thickness of a trophectoderm, a thickness distribution of a trophectoderm, or the number of cells in a trophectoderm is calculated on the basis of the three-dimensional image (D2). Then, the parameter (P) of a blastocyst embryo to be classified is input to a trained machine learning model (M1, M2, M3) that has the parameter (P) as an input variable and has a classification result of the blastocyst embryo as an output variable, and the blastocyst embryo is classified on the basis of a classification result output from the machine learning model. Thus, classification of the blastocyst embryo can be performed quantitatively on the basis of the three-dimensional image (D2).