Glaucoma Diagnosis Data Fusion via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current glaucoma diagnosis and progression analysis rely on isolated structural or functional measurements, which are subjective, variable, and lack adequate diagnostic accuracy across patient populations and disease dynamic ranges, necessitating improved systems and methods for integrating structural and functional data using machine learning classifiers.
Innovation Solution
The integration of visual field test data and optical coherence tomography data through a knowledge-based approach, incorporating anatomical relationships and data distribution knowledge to fuse measurements at the basic and decision levels, enhancing the accuracy of diagnostic measurements and classifier decisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If isolated structural or functional measurements are used for glaucoma diagnosis, then the diagnostic process is simple and quick, but the diagnostic accuracy is insufficient and subjective variability is high
Solution Approach 1:
The patent combines multiple diagnostic measurements (OCT structural measurements and visual field functional measurements) into a unified diagnostic system. The machine learning classifier integrates data from both measurement types to produce a combined diagnostic output, thereby improving diagnostic accuracy while managing the complexity through automated processing.
Solution Approach 2:
The patent introduces a machine learning classifier as an intermediary component that processes and integrates data from multiple diagnostic measurements. This intermediary automatically correlates structural and functional measurements, reducing subjective variability and improving diagnostic accuracy without requiring manual integration by clinicians.
2Measurement precision
If multiple diagnostic tests are used in conjunction, then the diagnostic accuracy improves, but the interpretation becomes difficult and highly variable across observers
Solution Approach 1:
The machine learning classifier serves as an intermediary that automatically interprets and integrates results from multiple diagnostic tests. It processes OCT and visual field data together and produces a unified diagnostic output, eliminating the need for clinicians to manually correlate multiple tests and reducing observer variability.
Solution Approach 2:
The system provides automated feedback by processing multiple test results through the machine learning classifier and returning a integrated diagnostic assessment. This feedback mechanism guides clinicians by presenting a synthesized interpretation rather than requiring them to manually integrate multiple complex test results.
3Loss of time
If subjective visual review of multiple reports is performed, then the process is quick, but the variability across observers is high and accuracy is limited
Solution Approach 1:
The machine learning classifier acts as an intermediary that automatically processes and integrates multiple diagnostic measurements, replacing subjective visual review. It maintains the efficiency of automated processing while significantly improving diagnostic consistency by eliminating observer variability through standardized algorithmic interpretation.
Data Source
AI summary
Systems and methods for improving the reliability of glaucoma diagnosis and progression analysis are described. The measurements made from one type of diagnostic device are adjusted based on another measurement using a priori knowledge of the relationship between the two measurements including the relationship between structure and function, knowledge of disease progression, and knowledge of instrument performance at specific locations in the eye. The adjusted or fused measurement values can be displayed to the clinician, compared to normative data, or used as input in a machine learning classifier to enhance the diagnostic and progression analysis of the disease.