PCCT-Based CMD Assessment Using Virtual FFR and CFR Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional diagnostic methods for coronary microvascular disease (CMD) are invasive, costly, and risky, failing to detect subtle abnormalities in small blood vessels, leading to inadequate diagnosis and increased patient risk.
Innovation Solution
A machine learning-based assessment model using PCCT images to non-invasively determine a microvascular coronary resistance index (MCRIML) through a trained model calibrated with FFR and CFR values, enabling precise diagnosis of CMD without invasive procedures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional invasive diagnostic tests are used to assess coronary microvascular disease, then diagnostic accuracy for CMD is improved, but patient risk and procedure complexity increase
Solution Approach 1:
The patent creates a virtual copy of the invasive diagnostic process through machine learning models that simulate fractional flow reserve (FFR) and coronary flow reserve (CFR) measurements. The system trains neural networks on data from patients who underwent invasive testing, then applies these trained models to non-invasive imaging data to predict FFR and CFR values without actually performing invasive procedures on new patients.
Solution Approach 2:
The patent replaces the mechanical/invasive catheter-based measurement system with a computational imaging system. Instead of physically inserting catheters into coronary arteries to measure pressure and flow, the system uses machine learning algorithms processed from non-invasive imaging data to derive the same diagnostic parameters (FFR and CFR) that traditionally required invasive mechanical measurement.
2Reliability
If conventional invasive testing procedures are performed to diagnose CMD, then diagnostic reliability is improved, but procedure duration and cost increase
Solution Approach 1:
The system performs preliminary training of machine learning models on datasets from patients who underwent invasive testing. This pre-trained knowledge is then applied to new patients without requiring them to undergo the time-consuming invasive procedures. The model has already learned the relationships between imaging features and FFR/CFR values from the training data, enabling rapid assessment of new patients.
Solution Approach 2:
The patent creates virtual replicas of the invasive diagnostic outcomes through computational modeling. The machine learning models generate predicted FFR and CFR values that copy the diagnostic information obtained from invasive testing, but without requiring the actual invasive procedure to be performed on each patient, thereby eliminating procedure duration while maintaining diagnostic reliability.
3Adaptability or versatility
If traditional diagnostic methods are used for CMD, then diagnostic capability is limited, but the methods are simpler and less costly
Solution Approach 1:
The patent develops a multi-functional machine learning assessment model that can evaluate multiple diagnostic parameters (FFR, CFR, and CMD diagnosis) from a single set of non-invasive imaging data. The system is trained to simultaneously predict multiple outcomes, making it a universal diagnostic tool that replaces multiple separate testing procedures with one integrated assessment.
Solution Approach 2:
The patent introduces machine learning algorithms as an intermediary between non-invasive imaging data and diagnostic conclusions. The neural networks serve as computational mediators that translate raw imaging features into clinically meaningful diagnostic parameters (FFR, CFR values), bridging the gap between simple imaging and complex diagnostic interpretation without requiring invasive intermediate steps.
Data Source
AI summary
Systems and methods for determining an assessment of coronary microvascular disease of the patient are provided. Patient data of a patient is received. The patient data comprises patient characteristics and vessel characteristics. An assessment of coronary microvascular disease of the patient is determined based on the patient data using a machine learning based assessment model. Results of the assessment of coronary microvascular disease of the patient are output.