Optical Texture Analysis for Retinal Layer Abnormality Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting retinal nerve fiber layer (RNFL), ganglion cell layer (GCL), and inner plexiform layer (IPL) abnormalities, such as thickness deviation maps, are impaired in eyes with myopia and fail to discern different levels of optic nerve damage, leading to false positives and compromised diagnostic performance.
Innovation Solution
Optical texture analysis of RNFL/GCL/IPL using optical coherence tomography (OCT) images, which measures optical density and computes optical texture signature values to visualize and quantify abnormalities without normative databases, employing non-linear transformations and deep learning for pattern recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If RNFL/GCL/IPL thickness deviation maps are used for detection, then diagnostic standard is established, but sensitivity and specificity are impaired in eyes with myopia and advanced optic neuropathies
Solution Approach 1:
The patent transforms the detection parameter from RNFL/GCL/IPL thickness to optical density values extracted from OCT images. This parameter change enables detection in myopic eyes and advanced glaucoma cases where thickness-based methods fail, as optical density reflects tissue composition and microstructural changes independent of overall layer thickness
Solution Approach 2:
The patent introduces a new dimension of analysis by computing optical texture signature values through non-linear transformations of optical density measurements. This creates a topographic map that visualizes abnormalities in a different dimensional space, allowing differentiation of damage levels that appear uniform in traditional thickness maps
2Ease of operation
If RNFL/GCL/IPL thickness profiles are analyzed, then conventional detection method is established, but different levels of optic nerve damage cannot be discerned in advanced glaucoma
Solution Approach 1:
The patent applies non-linear transformations to optical density values to compute optical texture signature values. This transformation preserves subtle variations in tissue properties that are lost in linear thickness measurements, enabling differentiation of damage severity levels in advanced glaucoma while maintaining ease of automated analysis
Solution Approach 2:
The patent creates a topographic representation of optical texture signature values that adds spatial dimensionality to the analysis. This topographic map visualizes the distribution and severity of abnormalities across the retinal surface, providing information about damage progression and heterogeneity that single-value thickness profiles cannot convey
3Ease of operation
If red-free photography is used to visualize RNFL abnormalities, then standard visualization is established, but sensitivity for detection is limited
Solution Approach 1:
The patent replaces the optical microscopy-based red-free photography with quantitative analysis of OCT optical coherence tomography data. This substitution transitions from qualitative visual assessment to quantitative measurement of optical density and texture properties, significantly improving detection sensitivity while providing automated analysis capability
Solution Approach 2:
The patent extracts optical density values from OCT images and applies non-linear transformations to compute optical texture signature values. This parameter transformation converts the raw OCT signal into a enhanced visualization format that amplifies subtle abnormalities, providing both superior sensitivity and automated quantification compared to manual red-free photography assessment
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
Improves sensitivity and specificity for detecting RNFL/GCL/IPL abnormalities, allowing differentiation of various levels of optic nerve damage, particularly in advanced glaucoma, and provides accurate visualization and quantification of retinal layer abnormalities.
Implementation Method 1
using optical coherence tomography (OCT) images, which measures optical density
Data Source
AI summary
Optical texture analysis of the inner retina, including the retinal nerve fiber layer (RNFL), ganglion cell layer (GCL), or inner plexiform layer (IPL), or a combination of these layers, can be used to detect and quantify RNFL/GCL/IPL abnormalities. From a set of scans of a retina, anterior and posterior boundaries of an inner retinal layer of interest can be determined. Optical density measurements at specific locations on the retina and depths within the layer of interest can be extracted from the scans. From these measurements, a set of optical texture signature values corresponding to different locations can be computed, where the optical texture signature value for a given location provides information about a tissue composition of the inner retinal layer at that location. The texture signature values can provide a topographical map of a retinal layer, which can facilitate detection and quantification of abnormalities.