Heart Strain Determination via 3D Deformation Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining heart strain, particularly 2D strain measures, are prone to errors due to foreshortening effects and out-of-plane deformations, lacking uncertainty measures and corrections, which negatively impacts the predictive value of clinical studies.
Innovation Solution
A system and method that utilize a series of 2D-slice images to estimate 3D heart strain, incorporating a pose estimation module to correct for heart and patient movements, and a 3D deformation estimation module to calculate accurate 3D deformation fields, thereby providing a statistical characterization of heart strain.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D ultrasound or MRI is used for heart strain measurement, then the installation base is large and temporal/spatial resolution is high, but measurement precision deteriorates due to foreshortening effects and out-of-plane deformations
Solution Approach 1:
The patent transforms 2D strain measurements into 3D strain estimates by introducing a third dimension through statistical modeling. The system takes 2D deformation fields from ultrasound or MRI and applies 3D statistical models to compensate for out-of-plane deformations and foreshortening effects, effectively converting limited 2D data into accurate 3D strain characterizations without requiring actual 3D imaging.
Solution Approach 2:
The patent changes the parameter representation from 2D strain values to 3D strain distributions with uncertainty measures. By modeling strain as a 3D statistical field rather than simple 2D measurements, the system accounts for measurement uncertainties and anatomical variations, transforming the nature of the measurement parameters to include mean strain values and uncertainty distributions.
2Measurement precision
If 3D ultrasound or MRI is used for heart strain measurement, then measurement precision is improved, but device complexity increases and upgrade costs are significant
Solution Approach 1:
The patent introduces statistical 3D models as an intermediary between 2D imaging data and 3D strain measurement. Instead of directly acquiring 3D images, the system uses 2D images combined with pre-computed 3D statistical models of heart deformation to infer 3D strain values. This intermediary approach allows 3D measurement capabilities through 2D imaging systems.
Solution Approach 2:
The patent creates virtual 3D strain measurements by copying and adapting 3D statistical models to individual patient 2D data. The system uses population-based 3D deformation models as templates and adapts them to individual patients through their 2D imaging data, effectively copying 3D measurement capabilities without requiring actual 3D imaging hardware.
3Reliability
If 2D strain measurements are used, then ease of operation is maintained, but reliability deteriorates due to hidden errors in clinical studies
Solution Approach 1:
The patent implements feedback by providing uncertainty estimates alongside strain measurements. The system calculates not only mean 3D strain values but also uncertainty distributions that feedback into the interpretation of measurement reliability. This allows clinicians to assess the confidence level of each measurement and adjust clinical decisions accordingly.
Solution Approach 2:
The patent performs preliminary statistical modeling and uncertainty quantification before clinical interpretation. By pre-computing 3D statistical models and uncertainty distributions from training data, the system prepares the analytical framework in advance, allowing rapid and reliable 3D strain estimation with uncertainty measures during actual clinical measurements.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
The invention describes a system (12) for heart strain determination comprising: - a computing unit (13), designed to receive a series of 2D-slice images (SI) as input, to host several modules (20, 21, 22), to pass intermediate result-data from an output of one module (20, 21, 22) to an input of another module (20, 21, 22) and to output a result, - a pose estimation module (20), designed to estimate a slicing-pose of the inputted series of 2D-slice images (SI) in the heart, - a 3D deformation estimation module (21), designed for estimating a 3D deformation field (DF) from the series of 2D-slice images (SI) and the estimated slicing-pose, - a strain measurement module (22), designed for computing a heart strain measure from the 3D deformation field (DF) and predefined definitions for strain computation. The invention further describes a related method, a data generation device for training the models of this system, the method for manufacturing such system and a related medical imaging system.