Real-Time Coronary Angiography Analysis Using Dynamic Acquisition Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing coronary disease using X-ray angiography face challenges in achieving reliable, repeatable, and objective assessments due to patient variability and suboptimal acquisition settings, leading to inefficient and potentially harmful radiation and contrast agent doses.
Innovation Solution
A method employing a trained classifying device, such as a convolutional neural network, to analyze diagnostic image data during acquisition, adjusting acquisition settings in real-time to optimize image quality and reduce variability, while minimizing radiation and contrast agent use.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pre-defined acquisition settings are used to reduce variability in acquired data, then measurement precision is improved, but adaptability to individual patient anatomy deteriorates
Solution Approach 1:
The system dynamically adjusts acquisition settings during the imaging process based on real-time analysis of quantitative features extracted from acquired images. The classifying device processes images as they are acquired and modifies parameters such as projection angles, contrast agent injection rates, and radiation doses adaptively, transforming a static pre-defined acquisition protocol into a dynamic, patient-specific optimized process.
Solution Approach 2:
The system implements a feedback loop where quantitative features are continuously extracted from acquired angiographic images, compared against reference values or thresholds, and used to automatically adjust subsequent acquisition parameters. This closed-loop control ensures that the acquisition settings are optimized based on actual image quality and diagnostic information obtained so far, resolving the contradiction between standardized protocols and patient-specific adaptation.
2Reliability
If complex calculations are performed to account for all variabilities in data from different patients, then reliability of analysis is improved, but device complexity increases
Solution Approach 1:
The system replaces complex manual analysis and sophisticated calculation algorithms with a trained classifying device, likely a neural network or machine learning model. This classifying device has been pre-trained on diverse patient data to automatically recognize patterns and extract quantitative features, substituting complex computational mechanisms with a trained intelligent system that provides reliable analysis with simpler operational complexity.
Solution Approach 2:
The classifying device is pre-trained using extensive training data that encompasses various patient anatomies and pathologies. This preliminary training phase performs the complex calculations and pattern recognition work in advance, allowing the deployed system to make reliable classifications and extractions with minimal real-time computational complexity, thus resolving the contradiction between reliability and operational simplicity.
3Manufacturing precision
If acquisition settings are optimized for each patient individually, then image quality is improved, but loss of time increases due to manual adjustment
Solution Approach 1:
The system performs self-service by automatically analyzing acquired images, extracting quantitative features, and adjusting acquisition parameters without requiring manual intervention from operators. The classifying device autonomously optimizes the acquisition process in real-time, eliminating the time-consuming manual adjustment phase while maintaining high image quality through patient-specific adaptation.
Solution Approach 2:
The optimization process occurs continuously during the acquisition itself rather than requiring separate setup and adjustment phases. The system analyzes images as they are acquired and makes real-time parameter adjustments, ensuring that the useful action of image acquisition and optimization proceeds without interruption or time loss, resolving the contradiction between quality optimization and time efficiency.
4Measurement precision
If high doses of radiation and contrast agent are used to ensure diagnostic quality images, then measurement precision is improved, but object-affected harmful factors increase
Solution Approach 1:
The system applies partial action by acquiring and analyzing images incrementally, extracting quantitative features from the subset of images already acquired. Based on this partial information, the system determines whether sufficient diagnostic quality has been achieved or if additional images are needed, avoiding the excessive application of radiation and contrast agent that would result from acquiring complete image sets regardless of diagnostic sufficiency.
Solution Approach 2:
The real-time feedback loop continuously evaluates image quality metrics and quantitative features extracted from acquired images, comparing them against diagnostic requirements. This feedback mechanism allows the system to terminate acquisition early when diagnostic quality is sufficient, preventing unnecessary exposure to radiation and contrast agent while maintaining measurement precision through adaptive optimization of the acquisition process.
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 reliable, efficient, and patient-independent assessment of coronary disease by optimizing acquisition settings during imaging, reducing radiation and contrast agent doses, and improving diagnostic image quality.
Implementation Method 1
classifying the diagnostic image data to extract at least one quantitative feature of the vessel of interest from at least one acquisition image of the plurality of acquisition images
Data Source
AI summary
A method and apparatus for analyzing diagnostic image data are provided in which a plurality of acquisition images of a vessel of interest having been acquired with a pre-defined acquisition method is received at a trained classifying device and classified, by the classifying device, to extract at least one quantitative feature of the vessel of interest from at least one acquisition image of the plurality of acquisition images. The at least one quantitative feature is then output associated with the at least one acquisition image while the acquisition of the diagnostic image data is still in progress and one or more adjustable image acquisition settings are adjusted based on the at least one quantitative feature to optimize the acquisition of the diagnostic image data.