Heart Anatomical Model Alignment for Early Failure Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting heart failure lack sufficient heart-based anatomical information, making it difficult to accurately forecast the condition, especially in its early stages.
Innovation Solution
A computer-implemented method that involves receiving imaging data of a subject's heart over time, generating anatomical models for each image, aligning these models to create a set of aligned data, and using a machine learning model to compare the aligned data and identify changes in the heart, thereby predicting heart failure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If algorithms are trained to predict heart failure based on biomarkers and clinical measures, then prediction capability is provided, but the input data lacks sufficient heart-based anatomical information for accurate prediction
Solution Approach 1:
The imaging data is segmented to extract specific anatomical structures of the heart (ventricles, atria, valves) and represent them as separate mesh models. This segmentation enables detailed analysis of individual heart components and their temporal changes, providing the missing anatomical information for accurate prediction.
Solution Approach 2:
The patent transforms 2D imaging data into 3D mesh models, adding a spatial dimension to the analysis. This dimensional transformation enables comprehensive anatomical assessment by representing heart structures in three-dimensional space, capturing geometric information that cannot be obtained from conventional 2D measurements alone.
2Measurement precision
If imaging data is collected at multiple points in time to detect early signs, then early detection capability is improved, but comparing images taken over time becomes difficult due to variations between investigations
Solution Approach 1:
The patent transforms imaging data into a standardized parameter space by extracting geometric features (volumes, areas, distances, angles) from the mesh models. This parameter transformation enables direct comparison across different time points by expressing all measurements in consistent, normalized units that are independent of imaging variations.
Solution Approach 2:
The patent creates simplified mesh model representations (copies) of the complex imaging data. These mesh models serve as standardized intermediaries that capture essential anatomical information while eliminating variations introduced by different imaging devices, protocols, and operators, enabling reliable longitudinal comparison.
3Ease of operation
If single measurements are used to assess heart function, then the assessment process is simple, but early signs of heart failure are subtle and not detectable
Solution Approach 1:
The patent implements continuous monitoring by analyzing sequences of mesh models derived from multiple imaging time points. This continuous analysis approach enables detection of subtle, progressive anatomical changes that would be invisible in single measurements, while maintaining operational simplicity through automated processing of the temporal data series.
Solution Approach 2:
The patent performs preliminary transformation of imaging data into standardized mesh models and geometric parameters before comparison. This preliminary processing step prepares the data in advance for sensitive detection of subtle changes, enabling early sign detection while keeping the final assessment simple by presenting processed results in an easily interpretable format.
Data Source
AI summary
A computer implemented method of identifying changes in a subject's heart or an adjacent region over time. The method comprising: receiving a set of imaging data relating to a subject's heart that has been obtained at a plurality of points in time; generating an anatomical model of the subject's heart for each of the images in the set of imaging data so as to provide a set of anatomical models of the subject's heart corresponding to the plurality of points in time; and aligning each of the anatomical models in the set of anatomical models relative to one another so as to provide a set of aligned data of the subject's heart. The aligned data are for identifying changes in at least one region of the subject's heart by comparing the anatomical models in the set of aligned anatomical models using a machine learning model.