Tomographic Data Analysis for Ischemic Stroke Severity Assessment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for diagnosing ischemic strokes, such as the ASPECTS scoring system, lack precision in assessing the severity and functional impact of ischemic damage in brain tissue, which can hinder timely and effective treatment decisions.

Innovation Solution

A computer-implemented method that processes three-dimensional tomographic data to align brain images with reference images, classifies voxels based on attenuation differences, and assigns voxel scores indicating damage likelihood, with weights for functional relevance, to provide a cumulative score for assessing ischemic stroke severity and location.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual assessment methods like ASPECTS scoring system are used, then diagnostic process is simple and quick, but measurement precision of ischemic damage severity is insufficient

Engineering Contradiction:
Improveprecision of ischemic damage assessmentVSAvoidcomplexity of analysis method
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The brain is divided into multiple three-dimensional regions of interest (ROIs) including infarct core, penumbra, and healthy tissue. Each ROI is independently analyzed with specific attenuation thresholds (e.g., -10 to -30 HU for penumbra, -30 to -50 HU for infarct core), enabling precise localization and characterization of ischemic damage at different stages.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

A computer-implemented analysis system acts as an intermediary between raw CT data and clinical decision-making. The system automatically processes attenuation values, performs three-dimensional segmentation, calculates volumes, and generates diagnostic reports, thereby achieving high measurement precision without requiring complex manual interpretation by clinicians.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If detailed three-dimensional voxel analysis is performed, then diagnostic accuracy improves, but processing time increases

Engineering Contradiction:
Improveaccuracy of stroke severity determinationVSAvoidtime for image processing
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary segmentation of the brain into anatomical regions and pre-defines attenuation value ranges for different tissue types before detailed analysis. Three-dimensional ROIs are pre-established based on standard brain atlases, allowing rapid extraction and classification of voxels within each region without requiring exhaustive whole-brain processing.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

Different analysis strategies are applied to different brain regions based on their functional importance and expected pathology. Critical regions such as the penumbra and infarct core receive detailed voxel-level analysis with specific attenuation thresholds, while less critical areas use coarser segmentation, optimizing the balance between diagnostic accuracy and processing time.

Inventive Principle:
Principle #3Local quality

3Reliability

If attenuation thresholds are set to detect early ischemic changes, then detection sensitivity increases, but false positive rate increases

Engineering Contradiction:
Improvedetection sensitivity of early ischemic strokeVSAvoidfalse positive detections
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The system uses multiple attenuation thresholds to differentiate between various stages and types of ischemic tissue. For example, penumbra is identified with attenuation values of -10 to -30 HU, while infarct core uses -30 to -50 HU. This multi-threshold approach allows the system to detect subtle early changes while maintaining specificity through progressive classification.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The analysis system incorporates feedback mechanisms where detection results from one region inform the analysis of adjacent regions. Voxels classified as abnormal in one ROI are used as reference for determining normal ranges in neighboring ROIs, allowing the system to adapt to individual patient anatomy and reduce false positives while maintaining high sensitivity for true pathology.

Inventive Principle:
Principle #23Feedback

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

This method enables rapid and accurate determination of ischemic stroke severity and potential for recovery, guiding appropriate treatment decisions by highlighting damaged brain regions and their functional significance.

Implementation Method 1

measuring x-ray attenuation along multiple paths through the cross-section by scanning a source and an opposed sensor around the object and deducing the cross-sectional image by computation

Methodology Applied
Scientific EffectX-ray attenuation: X-Ray

Implementation Method 2

information about the three-dimensional structure of an object can be obtained by combining information from multiple two-dimensional images in closely spaced planes, or by a performing a scan along a helical path around the object

Methodology Applied
Scientific EffectTomography: Tomography

Data Source

PatentEP3701495B1Tomographic data analysis
Publication Date: 2023.12.06 BRAINOMIX LTD
  • EP3701495B1 patent drawingFigure 1
  • EP3701495B1 patent drawingFigure 2
  • EP3701495B1 patent drawingFigure 3

AI summary

Data from a tomographic scan (14) that provides three-dimensional information about a patient's brain comprises the steps of: filtering and re-sampling (21) the data to produce a three- dimensional image; performing registration (23) to align the three-dimensional image with a reference image (16), using 3-D rigid and/or non-rigid transformations; identifying (25) image features in the aligned image, to identify which voxels or regions of adjacent voxels correspond to image features that represent structures within the brain that are expected to be evident; classifying (26) each voxel within an identified image feature by a voxel score that corresponds to the difference between the attenuation of that voxel and the expected attenuation at that region of the brain; and deducing a cumulative score that combines the voxel scores from all the voxels of at least a region of the brain. This method can provide a medical professional with a rapid indication of the status of the brain tissue, which can be used to guide the selection of treatment to best improve the prospects for a patient, particularly a patient who has had an ischaemic stroke.