Coronary Tree Graph Neural Networks for Non-Invasive Plaque Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for identifying vulnerable plaques in coronary arteries are inaccurate and often require invasive procedures, limiting their availability and effectiveness in assessing plaque vulnerability and predicting adverse cardiovascular events.
Innovation Solution
A graph neural network-based system processes coronary tree-level and patient-level data to generate a coronary tree model, determining feature embeddings, and outputs assessments of vulnerable plaque risk using a message-passing mechanism, integrating patient-specific data for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods for identifying vulnerable plaques are used, then plaque vulnerability assessment can be performed, but the accuracy is poor and invasive procedures are required
Solution Approach 1:
The patent replaces invasive mechanical procedures (catheter-based imaging, intravascular ultrasound) with non-invasive computational methods using graph neural networks that process standard coronary angiography images and clinical data to identify vulnerable plaque characteristics through pattern recognition and machine learning
Solution Approach 2:
The patent transforms the assessment approach by changing from direct anatomical measurement to multi-parameter integration, combining angiographic image features, clinical risk factors, and hemodynamic parameters through a graph neural network to predict plaque vulnerability with higher accuracy than traditional single-modality methods
2Measurement precision
If graph neural network-based assessment is implemented, then diagnostic precision is enhanced, but system complexity increases
Solution Approach 1:
The patent segments the diagnostic task into distinct modular components: coronary tree model generation from angiography images, feature extraction and embedding at node level, graph neural network inference for vulnerability assessment, and risk stratification output. Each module processes specific data types and produces standardized outputs that feed into the next stage, managing overall system complexity through functional decomposition
Solution Approach 2:
The patent introduces a coronary tree model as an intermediary data structure that bridges raw angiographic images and the graph neural network analysis. This intermediate representation organizes vascular anatomy into a graph format with nodes and edges, enabling efficient feature extraction and network processing while simplifying the interface between image processing and diagnostic inference
3Reliability
If traditional functional or anatomical significance assessment is used, then treatment decisions can be made, but vulnerable plaques in non-significant lesions are missed
Solution Approach 1:
The patent creates a universal vulnerability assessment framework that simultaneously evaluates both functionally significant and non-significant coronary lesions. The graph neural network applies the same vulnerability prediction algorithm across all coronary segments regardless of stenosis severity, enabling consistent identification of at-risk plaques throughout the entire coronary tree in a single integrated assessment
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods for vulnerable plaque assessment and outcome prediction in coronary artery disease. Medical imaging data is used to generate a coronary tree model (300) of coronary centerlines (310) of a patient (210). The coronary tree model (300) includes a plurality of nodes that represent locations in the coronary tree model (300). Feature embedding associated with each node are determined from a plurality of features derived from the medical imaging data. The feature embeddings are input into a trained graph neural network that is configured to output an assessment at a node level, a segment level, and/or a coronary tree level for vulnerable plaque.