Hepatic Vessel Branch Labeling in Liver Disease Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current diagnostic systems for liver cancer detection and diagnosis involve multiple image modalities and phases, leading to a large amount of data that is challenging to effectively utilize, requiring improved methods for data management and visualization to enhance physician throughput and accuracy.
Innovation Solution
A computer-assisted liver disease diagnosis system with a graphical user interface that facilitates the retrieval, manipulation, and visualization of visual and non-visual data, including lesion detection, segmentation, and treatment planning, using mechanisms like marking enforcement, hierarchical representation, and interactive exploration to streamline the diagnostic process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple image modalities and phases are acquired for liver cancer detection, then detection accuracy is improved, but data complexity and difficulty of effective utilization increase
Solution Approach 1:
The patent segments the complex vascular structure into individual vessels and branches, allowing separate labeling and analysis. The system divides the hepatic vasculature into portal veins and hepatic veins, and further into branches, enabling manageable processing of complex data while maintaining accurate detection capabilities
Solution Approach 2:
The patent transforms 3D volumetric imaging data into a 2D graphical user interface representation with hierarchical labeling. The index map displays vessel relationships in a simplified 2D format while preserving 3D spatial information, making complex multi-phase imaging data easier to interpret
2Measurement precision
If multiple image modalities and phases are acquired for liver cancer detection, then detection accuracy is improved, but physician throughput decreases
Solution Approach 1:
The system performs preliminary automatic labeling of vascular structures before physician review. The computer automatically identifies and labels vessels and branches in the imaging data, preparing the data in advance so physicians can focus on verification and diagnosis rather than manual labeling, thereby improving throughput while maintaining accuracy
Solution Approach 2:
The system enables self-service by allowing physicians to interactively correct and refine automatic labeling results directly in the graphical interface. The interactive labeling mechanism lets physicians make adjustments without requiring complex manual analysis of raw imaging data, streamlining the workflow
3Loss of information
If interactive labeling of vessel branches is implemented, then diagnostic information completeness is improved, but time consumption increases
Solution Approach 1:
The system implements partial action by allowing physicians to label only the specific vessel branches that are relevant to the diagnosis rather than requiring complete labeling of the entire vascular tree. The hierarchical structure enables selective labeling of parent vessels or specific branches based on diagnostic needs, reducing time consumption while maintaining information completeness
Data Source
AI summary
A method for labeling vessel branches forming first and second vessel systems. A 3D image is segmented to obtain vessel branches. First and second root points are then determined. Starting from the first root point, the vessel branches are traced to derive a tracing path and a break point is determined, that separates the tracing path into two portions. A region of interest in the 3D image is determined with respect to the break point, in which center lines are assigned to the first or second vessel system. A 3D cutting structure is generated based on distances measured from points on the center lines to the 3D cutting structure. Graph representations are constructed for the first and second vessel systems. The vessel branches are labeled with different labels based on the graph representations.