CAD Intracranial Aneurysm Detection 3D Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting intracranial aneurysms in 3D medical image data, such as MRA, are time-consuming and prone to missing small aneurysms due to overlapping vessels and low sensitivity, especially in invasive and non-invasive imaging techniques.
Innovation Solution
A system that assigns points of interest in 3D image datasets based on defined characteristics associated with aneurysms, calculates features like distance, radius, planeness, and shape index, and identifies aneurysm suspects by applying sieving rules and probability scoring to reduce false positives and enhance detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If radiologists review MRA datasets using current methods, then they can detect aneurysms, but the process is very time-consuming
Solution Approach 1:
The patent segments the complex task of aneurysm detection into distinct processing stages: automatic POI assignment based on vessel characteristics, feature calculation for each POI, sieving through multiple filtering rules, and probability scoring. This segmentation automates the review process while maintaining detection accuracy, reducing radiologist time investment from reviewing entire datasets to evaluating only high-probability suspects.
Solution Approach 2:
The patent introduces a computer-aided detection system as an intermediary between the MRA dataset and the radiologist. This intermediary automatically performs initial screening, feature extraction, and suspect ranking, presenting only the most likely aneurysm candidates to the radiologist for final verification, thereby dramatically reducing review time while preserving detection reliability.
2Reliability
If radiologists use MIP images to review datasets, then they can visualize blood vessels, but small aneurysms are often missed due to overlapping vessels
Solution Approach 1:
The patent applies local quality analysis by calculating multiple geometric features (planeness, cylinder surfaceness, Gaussian curvature, mean curvature, shape index) specifically at each point of interest on the vessel surface. This localized multi-feature analysis enables detection of subtle local deviations indicative of small aneurysms that would be invisible in global MIP views where vessels overlap.
Solution Approach 2:
The patent transitions from 2D MIP image review to 3D surface analysis by assigning points of interest on the 3D reconstructed vessel surface and calculating geometric features in three-dimensional space. This dimensional transformation allows detection of small aneurysmal protrusions that are obscured in 2D projections, significantly improving sensitivity to small aneurysms.
3Measurement precision
If the system assigns POIs and calculates multiple features for each POI, then detection accuracy improves, but system complexity increases
Solution Approach 1:
The patent performs preliminary action by automatically assigning points of interest to specific locations on blood vessels based on predefined geometric characteristics before feature calculation. This preliminary assignment focuses subsequent complex feature analysis only on relevant locations, improving detection precision while managing system complexity through targeted rather than exhaustive analysis.
Solution Approach 2:
The patent replaces manual radiologist analysis with an automated computer-based system that systematically assigns POIs and calculates geometric features. This mechanical substitution of human expertise with algorithmic processing maintains high measurement precision through consistent application of multiple features while managing complexity through automated workflows and sieving rules.
4Measurement precision
If the system generates a comprehensive list of aneurysm suspects with probability scores, then identification accuracy improves, but false positives may increase
Solution Approach 1:
The patent implements feedback through iterative sieving rules that progressively filter POIs based on multiple geometric features, with each filtering stage providing feedback that refines the suspect list. Probability scores are assigned based on cumulative evidence from multiple feature evaluations, allowing the system to balance identification accuracy against false positive rates by adjusting scoring thresholds.
Solution Approach 2:
The patent employs parameter changes by varying the stringency of sieving rules and probability score thresholds to optimize the balance between detection accuracy and false positive rate. By adjusting these parameters, the system can generate comprehensive suspect lists with high identification accuracy while maintaining acceptable false positive rates through configurable filtering criteria.
Data Source
AI summary
A computer-aided system identifies aneurysm suspects in 3D image datasets. The system takes the raw image dataset as input and assigns one or more points of interest (POIs) in the image data. The system determines one or more features for each POI and identifies one or more aneurysm suspects from among the assigned POIs based on the determined features.