Deep Learning OCT System for High-Resolution TCFA Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurate identification of vulnerable atheromatous plaques, such as thin cap fibroatheroma (TCFA), in coronary arteries is time-consuming and requires extensive expertise, limiting quick and accurate diagnosis during cardiovascular procedures.
Innovation Solution
A deep learning-based diagnostic method using optical coherence tomography (OCT) images, which involves image acquisition, feature extraction (including fibrous cap thickness and necrotic core features), region-of-interest setting, and high-risk lesion determination, with optional lesion display using Grad-CAM or guided Grad-CAM for visualization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If expert reading of hundreds of cross-sectional images is performed one by one, then accurate identification of TCFA is achieved, but it takes a lot of time
Solution Approach 1:
The patent segments the complex task of TCFA identification into multiple processing stages: image acquisition, image processing to generate pullback images, automated analysis by the processor, and result generation. This segmentation allows parallel processing and eliminates the sequential manual review of hundreds of cross-sectional images, thereby reducing diagnosis time while maintaining accuracy through systematic multi-step analysis
Solution Approach 2:
The patent introduces an intermediary automated processing system between the raw OCT images and the final diagnosis. The processor acts as an intermediary that automatically performs image processing, generates pullback images, and analyzes them to identify TCFA, replacing the need for expert manual review and significantly reducing the time required while preserving diagnostic accuracy
2Loss of information
If manual reading of OCT images is performed, then quantitative information on TCFA can be obtained, but it is slow and cannot be done quickly during cardiovascular procedures
Solution Approach 1:
The system implements self-service through automated processing where the processor automatically performs image processing, generates pullback images, and extracts quantitative information about TCFA without requiring manual intervention. This self-service capability enables rapid acquisition of complete quantitative information during cardiovascular procedures, simultaneously improving both information completeness and diagnosis speed
3Productivity
If deep learning-based automated diagnosis is used, then diagnosis speed is improved, but system complexity increases
Solution Approach 1:
The patent applies universality by designing a multi-functional processor that performs multiple tasks: image processing, pullback image generation, TCFA identification, and quantitative analysis. This single multi-functional component handles all diagnostic operations, improving diagnosis speed while managing system complexity through functional integration rather than requiring separate specialized systems for each task
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 rapid and accurate evaluation of coronary artery atherosclerotic plaques, effectively distinguishing high-risk lesions like TCFA with high resolution, thereby aiding in the prediction of myocardial infarction and improving vascular response assessment post-interventional therapy.
Implementation Method 1
optical coherence tomography (OCT) image of a coronary artery lesion
Data Source
AI summary
The disclosure purposes to provide an optical coherence tomography (OCT)-based system for diagnosing a high risk lesion such as a vulnerable atheromatous plaque by using an artificial intelligence model through deep learning. A deep learning-based diagnostic method of diagnosing a high risk lesion of a coronary artery includes: acquiring an OCT image of a coronary artery lesion of a patient; extracting a first feature of a thin cap from the OCT image; setting a region of interest included in the OCT image on a basis of the first feature; and determining whether the region of interest includes a high risk lesion.