Patient-Specific Esophageal Flow Analysis With Reduced-Order CFD
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical imaging techniques for diagnosing esophageal disorders, such as barium swallow esophagram and video fluoroscopy swallowing exam, provide only qualitative information about esophageal transport, requiring significant manual effort for analysis and limiting their clinical application.
Innovation Solution
A method using deep learning and computational fluid dynamics to analyze medical image data, including segmentation with a neural network and reduced-order modeling, to provide quantitative data on flow rate, pressure, and esophageal wall properties, enhancing diagnostic capabilities with minimal user input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual segmentation and analysis of fluoroscopy images is performed, then detailed bolus geometry can be obtained, but significant time and effort are required
Solution Approach 1:
The patent replaces manual mechanical segmentation processes with an automated deep learning-based image processing system. The neural network automatically segments bolus geometry from fluoroscopy images without requiring manual tracing or measurement, thereby maintaining measurement precision while eliminating the time-consuming manual analysis step.
Solution Approach 2:
The system enables self-service automation where the computational model automatically performs bolus segmentation and flow analysis without human intervention. The deep learning model processes images autonomously, extracting geometric information and computing flow rates, pressures, and wall properties without requiring manual input or expert analysis time.
2Object-affected harmful factors
If qualitative imaging data is used, then minimally invasive diagnosis is maintained, but quantitative flow information is not obtained
Solution Approach 1:
The patent transforms qualitative imaging data into quantitative flow information by changing the parameter representation from visual/qualitative descriptors to numerical/quantitative metrics. The system computes flow rates, pressures, and wall mechanical properties from standard fluoroscopy images, thereby obtaining quantitative data without requiring additional invasive measurements or specialized imaging modalities.
Solution Approach 2:
The patent introduces a computational model as an intermediary that bridges qualitative imaging data and quantitative flow information. This virtual intermediary processes the visual data from minimally invasive fluoroscopy and generates quantitative flow metrics, effectively mediating between the non-invasive imaging modality and the quantitative diagnostic information needed for clinical decision-making.
3Loss of information
If complex analysis methods are applied to medical images, then comprehensive flow data can be obtained, but device and process complexity increases
Solution Approach 1:
The patent creates a universal computational framework that performs multiple analysis functions simultaneously. The same deep learning-based segmentation and reduced-order model infrastructure extracts bolus geometry, computes flow rates, determines pressure fields, and estimates wall mechanical properties from a single integrated system, thereby obtaining comprehensive flow data without requiring separate specialized tools for each measurement type.
Data Source
AI summary
Flow through tubular organs (e.g., the esophagus) is analyzed based on fluid mechanics analysis of medical images. Using computational fluid dynamics, a reduced-order model is constructed and implemented to predict flow rate and fluid pressure developed inside flexible tubular organs inside the body. As one non-limiting example, the constructed model can be applied to analyze esophageal transport using fluoroscopy image sequences to predict flow rate, pressure, esophagus wall stiffness, and active relaxation.