Endoscopic Vessel Tree Registration via Furcation Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Minimally invasive coronary artery bypass grafting (CABG) procedures face challenges in registering intra-operative endoscopic images with pre-operative 3D vessel tree images due to spatial constraints, lack of tactile feedback, and coordinate system handling errors, which hinder precise localization and guidance of the endoscope.
Innovation Solution
A robotic guiding system that employs an endoscope and endoscope controller for image registration, matching graphical representations of vessel tree furcations between pre-operative 3D images and intra-operative endoscopic images, using a robot unit with motorized joints and an image processing module to align and overlay the images, facilitating precise endoscope positioning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an optically tracked pointer is used to digitalize artery positions and register to pre-operative 3D tree using ICP algorithm, then the artery position data can be obtained, but the method becomes impractical for minimally invasive CABG due to spatial constraints of small port access
Solution Approach 1:
The patent replaces the mechanical optically tracked pointer system with an endoscope-based imaging system. Instead of using physical pointers that require manual manipulation and optical tracking, the system uses an endoscope to capture images of the vessel tree, which are then processed through image registration algorithms to achieve the same positioning and measurement objectives without the spatial constraints of manual pointer manipulation
Solution Approach 2:
The patent creates a digital copy of the pre-operative 3D vessel tree image and registers it with the intra-operative endoscopic images. This allows the surgical team to overlay and compare the pre-operative plan with the actual intra-operative anatomy without needing physical access to manipulate pointers or tactile feedback, solving the spatial constraint problem while maintaining measurement precision
2Measurement precision
If surface based matching is used to reconstruct heart surface from endoscope view and match to pre-operative CT data, then surface registration can be achieved, but the method fails when endoscope view is too small and operates in suboptimal local maximum due to smooth heart surface without specific features
Solution Approach 1:
The patent changes the registration parameters from surface-based matching to vessel tree-based matching. Instead of relying on the smooth heart surface geometry which lacks distinctive features, the system registers based on the branching patterns and topological structure of the vessel tree, which provides more distinctive and reliable features for accurate matching even with limited endoscope views
Solution Approach 2:
The patent transitions from 2D surface matching to 3D volumetric vessel tree matching. By utilizing the three-dimensional branching structure and topological relationships of the vessel tree rather than just surface geometry, the system achieves more reliable registration that is less sensitive to limitations in endoscope field of view and surface feature availability
3Ease of operation
If an assistant holds the endoscope while the surgeon holds two instruments and issues commands, then the surgeon can perform surgical tasks, but hand-eye coordination is hindered due to multiple coordinate systems and translation errors between surgeon's frame of reference and assistant's frame of reference
Solution Approach 1:
The patent implements a robotic system that allows the surgeon to directly control the endoscope without an assistant. The surgeon's movements and commands are directly translated into endoscope positioning through the robotic system, eliminating the intermediate translation step between surgeon's intent and endoscope action. This self-service control mechanism removes the coordinate system translation errors that occur when an assistant interprets and executes surgeon commands
Solution Approach 2:
The patent integrates real-time feedback by overlaying the registered pre-operative 3D vessel tree image with the live endoscopic view. This provides the surgeon with immediate visual feedback about the endoscope position relative to the target anatomy, enabling precise hand-eye coordination and localization without relying on verbal commands or intermediate translations by an assistant
Data Source
AI summary
An image registration system employs an endoscope and an endoscope controller. In operation, the endoscope generates an intra-operative endoscopic image of a vessel tree (e.g., an arterial tree or a venous tree) within an anatomical region, and the endoscope controller image registers the intra-operative endoscopic image of the vessel tree to a pre-operative three-dimensional image of the vessel tree within the anatomical region. The image registration includes an image matching of a graphical representation of each furcation of the vessel tree within the intra-operative endoscopic image of the vessel tree to a graphical representation of each furcation of the vessel tree within the pre-operative three-dimensional image of the vessel tree.