Deep Learning Plaque Segmentation in Intravascular OCT Images

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current imaging modalities like x-ray angiography and intravascular ultrasound struggle to accurately characterize coronary calcification, particularly in terms of thickness and composition, which can hinder stent deployment and increase the risk of thrombosis and in-stent restenosis during percutaneous coronary interventions.

Innovation Solution

The development of a deep learning-based method for automated segmentation and quantification of vascular plaques in intravascular optical coherence tomography (IVOCT) images, using techniques such as data augmentation, transfer learning with SegNet architecture, and conditional random fields to refine segmentation results, enabling the computation of a calcification score that predicts stent deployment success.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If x-ray angiography is used to describe vessel lumen, then the treatment planning information is provided, but specific information regarding vascular wall composition including calcification thickness is not available

Engineering Contradiction:

Improvevascular wall composition informationVSAvoidimaging modality complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the vascular wall into distinct tissue types (fibrous, lipid, calcified plaque) through automated image analysis. The system segments IVOCT images to identify and characterize different plaque components, providing detailed composition information without requiring complex multi-modality imaging.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent uses an automated image analysis system as an intermediary between IVOCT imaging and clinical decision-making. This intermediary processes the high-resolution IVOCT images to extract quantitative plaque characteristics, translating complex imaging data into actionable clinical information about vascular wall composition.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If intravascular ultrasound (IVUS) is used to identify calcification location, then the location is detected, but thickness assessment is impossible due to acoustic shadowing

Engineering Contradiction:

Improvecalcification thickness measurementVSAvoidcalcification thickness measurement difficulty

Core Design Contradiction:

Measurement precisionVSDifficulty of detecting and measuring

Solution Approach 1:

The patent replaces the mechanical ultrasound-based measurement system with an optical imaging system (IVOCT). By substituting acoustic waves with light waves, the system eliminates the acoustic shadowing problem that prevents thickness measurement in IVUS, enabling precise calcification thickness assessment through optical coherence tomography.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If manual analysis of IVOCT images is performed to characterize plaques, then detailed plaque characterization is achieved, but time consumption and subjectivity increase

Engineering Contradiction:

Improveplaque analysis efficiencyVSAvoidplaque characterization consistency

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent implements self-service through automated image analysis that performs plaque characterization without human intervention. The system automatically segments images, identifies plaque types, and quantifies characteristics, making the analysis process independent of operator skill and experience while maintaining high accuracy and consistency.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent employs feedback mechanisms where the automated analysis system continuously refines its classifications based on learned patterns from training data. The system uses feedback from image features and previously classified examples to improve the consistency and accuracy of plaque characterization across different cases.

Inventive Principle:

Principle #23Feedback

4Ease of operation

If high balloon pressures are applied to fracture calcification, then stent deployment is enabled, but risk of thrombosis and in-stent restenosis increases

Engineering Contradiction:

Improvestent deployment feasibilityVSAvoidthrombosis and in-stent restenosis risk

Core Design Contradiction:

Ease of operationVSObject-affected harmful factors

Solution Approach 1:

The patent applies preliminary action by performing detailed plaque characterization before stent deployment. The automated analysis system identifies calcified plaques and assesses their thickness and composition in advance, allowing clinicians to plan appropriate pre-dilation strategies and select suitable stents, thereby reducing the need for high-pressure balloon inflation and associated complications.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11710238B2Plaque segmentation in intravascular optical coherence tomography (OCT) images using deep learning

Publication Date: 2023.07.25 CASE WESTERN RESERVE UNIV

AI summary

Embodiments discussed herein facilitate segmentation of vascular plaque, training a deep learning model to segment vascular plaque, and/or informing clinical decision-making based on segmented vascular plaque. One example embodiment accessing vascular imaging data for a patient, wherein the vascular imaging data comprises a volume of interest; pre-process the vascular imaging data to generate pre-processed vascular imaging data; provide the pre-processed vascular imaging data to a deep learning model trained to segment a lumen and a vascular plaque; and obtain segmented vascular imaging data from the deep learning model, wherein the segmented vascular imaging data comprises a segmented lumen and a segmented vascular plaque in the volume of interest.