LVO Test Device Using Eyeball Deviation Biomarkers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing acute large vessel occlusion (LVO) using non-contrast computed tomography (CT) are limited by subjective radiologist interpretation, leading to variable symptom identification and delayed treatment, particularly in primary and secondary hospitals where advanced imaging is not readily available.
Innovation Solution
A large vessel occlusion test device and method utilizing non-contrast CT images, which includes image input, pre-processing, non-rigid registration, and AI-based classification to identify biomarkers such as dense MCA sign, early ischemic changes, and eyeball deviation, providing rapid LVO testing and notification to medical staff.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If non-contrast CT is used for LVO diagnosis, then the scanning time is reduced and accessibility is improved, but the diagnostic accuracy and objectivity deteriorate due to subjective radiologist interpretation
Solution Approach 1:
The patent replaces the mechanical/manual interpretation process of radiologists with an AI-based automated analysis system. The AI model objectively identifies LVO biomarkers (eyeball deviation, dense MCA sign, early ischemic changes) from non-contrast CT images, eliminating subjective variability while maintaining rapid scanning capabilities. This substitution resolves the contradiction by providing both speed and objective accuracy.
Solution Approach 2:
The patent introduces an AI-based image analysis system as an intermediary between the non-contrast CT scanning and final diagnosis. This intermediary automatically detects and quantifies LVO biomarkers, serving as a bridge that transforms subjective radiologist interpretation into objective, standardized measurements while preserving the advantages of rapid non-contrast CT scanning.
2Measurement precision
If advanced imaging methods (CTA, CTP, MR DWI) are used for LVO diagnosis, then the diagnostic accuracy is improved, but the scanning time increases and accessibility to primary/secondary hospitals is reduced
Solution Approach 1:
The patent extracts and isolates the most critical LVO diagnostic biomarkers (eyeball deviation, dense MCA sign, early ischemic changes) that can be reliably identified from non-contrast CT images. By focusing only on these key indicators rather than requiring full advanced imaging protocols, the system achieves adequate diagnostic accuracy with significantly reduced scanning time and without requiring specialized equipment unavailable in primary and secondary hospitals.
Solution Approach 2:
The patent changes the diagnostic parameters from requiring multiple advanced imaging modalities (CTA, CTP, MR DWI) to utilizing only non-contrast CT with AI-based analysis of specific biomarkers. This parameter change enables diagnosis in settings with limited resources while maintaining clinical utility by focusing on the most sensitive and specific indicators of LVO.
3Device complexity
If manual radiologist interpretation is used for non-contrast CT, then the device complexity is reduced, but the productivity and treatment speed deteriorate due to subjective analysis and time delays
Solution Approach 1:
The patent implements a self-service diagnostic system where the AI model automatically analyzes non-contrast CT images and generates LVO assessment results without requiring extensive manual radiologist interpretation. The system autonomously identifies biomarkers, calculates probabilities, and provides diagnostic recommendations, thereby increasing productivity and treatment speed while maintaining manageable system complexity through the use of existing CT infrastructure.
Data Source
AI summary
A large vessel occlusion (LVO) test device using eyeball deviation includes an image input unit configured to input non-contrast computed tomography (CT); a pre-processing unit; an image processing unit; and a determination unit. The objects include both eyes and the upper part of the brain corresponding to the entire cerebral cortex from the midbrain. The image processing unit configured to non-rigid register at least one atlas for the non-cis contrast CT images pre-processed from the pre-processing unit, normalize each cell of the non-contrast CT image to which the atlas is registered to have a value of 0 or 1, and extract a region of interest in a restored non-contrast CT image by combining each cell of the non-contrast CT image to which the atlas is registered through inverse transformation.