CT Imaging Assessment for Coronary Test Selection Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for assessing coronary artery disease using whole-heart Agatston scores are not accurate in determining which patients require additional diagnostic tests, leading to unnecessary testing and resource drain, as only a small percentage of patients identified by these scores are actually ischemia-positive.
Innovation Solution
A cardiovascular assessment method using machine learning on assessment features extracted from digitized imaging data, such as CT images, to predict the need for additional diagnostic tests, providing a personalized assessment of patients with obstructive coronary artery disease.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If whole-heart Agatston scores are used to assess coronary artery disease, then patient screening coverage is improved, but measurement precision deteriorates leading to unnecessary testing
Solution Approach 1:
The patent segments the assessment process into multiple stages: first using Agatston scoring for broad screening, then applying machine learning analysis to extract multiple features (calcification patterns, vessel morphology, plaque characteristics) from the same imaging data to refine patient selection for additional diagnostic tests. This segmentation allows the system to maintain wide screening coverage while improving precision in identifying high-risk patients.
Solution Approach 2:
The patent transforms the single-parameter Agatston score into a multi-parameter assessment system using machine learning. The system extracts numerous features from the CT imaging data including calcification distribution patterns, vessel wall characteristics, and anatomical relationships, then integrates these parameters through trained models to generate a refined risk assessment that accurately identifies patients needing further diagnostic evaluation.
2Reliability
If additional diagnostic tests are performed on all patients with elevated Agatston scores, then reliability of diagnosis is improved, but loss of resources increases
Solution Approach 1:
The patent applies machine learning analysis to CT imaging data as a preliminary filtering step before ordering additional diagnostic tests. The system pre-processes the available imaging data to identify patients with features highly predictive of obstructive coronary artery disease, thereby preparing a refined list of candidates for further testing. This preliminary action prevents unnecessary resource expenditure on low-risk patients while ensuring high-risk patients receive appropriate follow-up.
Solution Approach 2:
The patent introduces machine learning algorithms as an intermediary between the initial Agatston scoring and additional diagnostic testing. This intermediary layer analyzes multiple imaging features and clinical parameters to generate a risk stratification that guides decision-making about further testing. The intermediary prevents direct escalation to expensive follow-up tests for all patients, instead selectively referring only those with high predicted probability of obstructive disease.
3Measurement precision
If machine learning analysis is applied to extract features from digitized imaging data, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent implements a multi-functional machine learning system that performs several assessment tasks using the same underlying model infrastructure. The system simultaneously evaluates calcification patterns, vessel morphology, plaque characteristics, and anatomical relationships to generate comprehensive risk assessment. This universal approach improves measurement precision across multiple diagnostic dimensions while avoiding the complexity of separate specialized tools for each assessment type.
Solution Approach 2:
The patent employs machine learning models trained on large datasets to automatically extract and interpret imaging features without requiring manual analysis by radiologists or cardiologists. The system self-performs the complex task of identifying subtle patterns in CT images that correlate with obstructive coronary artery disease, thereby achieving high measurement precision while reducing the complexity burden on clinical workflows by automating the analytical process.
Data Source
AI summary
In some embodiments, the present disclosure relates to a method that includes accessing data stored in an electronic memory. The data includes digitized imaging data from a segmented non-contrast computerized tomography (CT) image of an obstructive coronary artery disease (OCAD) patient. A plurality of assessment features are extracted from the data. The plurality of assessment features include image based features that characterize one or more of calcifications, fat tissue, heart structures, bone density, muscle, a lung, and breast tissue. The plurality of assessment features and a plurality of clinical factors are provided to a machine learning stage that is configured to generate a medical assessment corresponding to whether or not the OCAD patient would benefit from additional diagnostic tests to identify a presence and extent of the obstructive coronary artery disease.