3D Dataset Sub-Volume Generation Using Anatomic Landmarks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radiologists face challenges in efficiently and consistently visualizing complex anatomic structures within three-dimensional datasets for accurate diagnosis, as existing methods lack repeatability and reproducibility, especially when dealing with closely spaced and intricately structured data like blood vessels in the brain.
Innovation Solution
The method involves creating a sub-volume within a 3D dataset by placing geometric objects at precise anatomic landmarks, allowing for division of the dataset into manageable parts, with pre-selected landmarks and objects enabling repeatable and reproducible analysis. This includes using a 3D cursor for feature extraction, filtering, and visualization techniques such as transparency and artificial intelligence for naming and measuring structures, and employing extended reality for optimized viewing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional visualization methods are used for complex anatomic structures, then radiologists can view the entire 3D dataset, but it becomes difficult to efficiently and consistently visualize specific structures due to lack of repeatability and reproducibility
Solution Approach 1:
The patent divides the complex 3D anatomic dataset into multiple sub-volumes based on pre-selected anatomic landmarks. Each sub-volume contains a specific structure of interest (e.g., a particular blood vessel segment), allowing radiologists to systematically examine each component separately rather than searching through the entire dataset, thereby improving visualization precision and consistency.
Solution Approach 2:
The patent employs pre-selected anatomic landmarks and pre-defined geometric objects that are prepared in advance according to standard radiologic checklists. This preliminary setup enables repeatable and reproducible sub-volume extraction across different patients and radiologists, eliminating the need for manual, time-consuming boundary definition for each case.
2Reliability
If manual boundary definition is used for each case, then flexibility in viewing is maintained, but repeatability and reproducibility across multiple patients are lost
Solution Approach 1:
The system pre-defines anatomic landmarks and geometric objects based on standardized radiologic checklists before patient data is processed. This preliminary configuration ensures that the same structures are consistently identified and extracted across different patients and operators, achieving high repeatability without requiring complex manual adjustments during operation.
Solution Approach 2:
The patent creates a universal framework where pre-selected landmarks and geometric objects can be applied across multiple patients and different anatomic structures. The same methodological approach works for various vascular territories and pathologies, making the system both repeatable and easy to operate without requiring case-specific customization.
3Measurement precision
If the entire 3D dataset is displayed, then complete anatomical context is provided, but closely spaced and intricately structured features become difficult to distinguish
Solution Approach 1:
The patent segments the large 3D dataset into smaller, manageable sub-volumes, each containing a specific anatomic structure or region of interest. This segmentation reduces the amount of data that needs to be processed and displayed at once, allowing radiologists to closely examine intricate features without the visual clutter of the entire dataset, thereby improving feature detection accuracy.
Solution Approach 2:
The system extracts only the relevant sub-volume containing the structure of interest from the complete 3D dataset, discarding or setting aside the rest. This extraction process isolates closely spaced and intricately structured features from surrounding anatomy, enabling detailed visualization and measurement without interference from other structures.
Data Source
AI summary
A method, apparatus and computer program for generating a sub-volume within a 3D dataset in a consistent, repeatable fashion. To accomplish this, geometric object(s) (e.g., 2D planes) are placed at precise anatomic landmarks with precise sizes and orientations. This serves to divide the 3D object into multiple parts (e.g., a first portion of the 3D volume has a first set of voxels and a second portion of the 3D volume has a second set of voxels). This process continues as multiple additional geometric objects are placed so that certain features of the 3D dataset can be extracted (i.e., shown with the best viewing settings). This process when used in conjunction with a radiologist's checklist enables efficient volume-by-volume viewing.