Vascular Disease Diagnosis via CFD and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing vascular diseases, such as fractional flow reserve (FFR) measurement, are invasive, time-consuming, and costly, and machine learning approaches face challenges with data quality and quantity, limiting their clinical applicability.
Innovation Solution
A method and apparatus that utilize biometric authentication information and computational fluid dynamics to calculate fractional flow reserve and flow feature information, applying these to machine learning models to determine vascular disease diagnosis and treatment decisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If invasive FFR measurement method is used, then diagnostic accuracy is improved, but patient harm and procedural complexity increase
Solution Approach 1:
The patent creates a virtual copy of the patient's blood vessel geometry from medical images and performs CFD simulations on this digital twin, eliminating the need for invasive physical measurements while maintaining diagnostic accuracy through computational modeling of blood flow dynamics
Solution Approach 2:
The patent replaces the mechanical invasive pressure measurement system with a computational fluid dynamics simulation system that uses numerical methods to calculate FFR values, thereby eliminating the need for physical catheter insertion and pressure sensors
2Measurement precision
If invasive FFR measurement method is used, then diagnostic accuracy is improved, but diagnosis time increases
Solution Approach 1:
The patent performs preliminary extraction of geometric features from medical images and pre-processes the vascular geometry data, so that when FFR calculation is needed, the computational simulation can proceed immediately with prepared data, significantly reducing the time required for actual diagnosis
Solution Approach 2:
By creating a virtual model of the blood vessel and performing repeated CFD simulations on this digital copy, the system can quickly obtain FFR values without the time-consuming invasive procedural steps required by traditional methods
3Extent of automation
If general machine learning technique is used, then automation is improved, but data quality and quantity requirements become insufficient for clinical applicability
Solution Approach 1:
The patent transforms the machine learning approach by changing the input parameters from raw medical images to extracted geometric feature parameters, and by incorporating physiological constraints and boundary conditions into the learning process, thereby improving reliability with available data quality and quantity
Solution Approach 2:
The patent introduces CFD simulation results as an intermediary between traditional medical imaging and machine learning diagnosis, providing physically accurate FFR values that serve as training labels and validation metrics, thereby enabling reliable automated clinical decision-making
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
This approach enables accurate and efficient vascular disease diagnosis, reducing the need for invasive procedures and minimizing user intervention, while quickly determining the necessity of surgery, thereby reducing diagnosis costs and improving diagnostic accuracy.
Implementation Method 1
a CFD processing step of applying the geometric feature parameter information to computational fluid dynamics (CFD) to calculate flow feature information
Data Source
AI summary
Disclosed are a method for diagnosing a vascular disease and an apparatus therefor. A vascular disease diagnosing method according to an exemplary embodiment of the present disclosure includes: an information acquiring step of acquiring patient information for a diagnosis subject; an FFR processing step of applying a geometric feature parameter information generated based on the patient information to a first learning model to calculate fractional flow reserve (FFR) information; a CFD processing step of applying the geometric feature parameter information to computational fluid dynamics (CFD) to calculate flow feature information; and a diagnosing step of determining a vascular disease based on the fractional flow reserve information and the flow feature information and determine whether to perform a surgery on the vascular disease.