Endoscope Localization Using Branch Imaging and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical procedures like bronchoscopy face challenges in precisely aligning the endoscope within the patient's anatomy due to uncertainties caused by adverse events such as patient coughing or airway collapse, necessitating time-consuming manual initialization steps and potential backtracking, which prolongs the procedure and complicates navigation.
Innovation Solution
A system utilizing a machine learning model to determine the endoscope's position within the luminal network based on imaging data, eliminating the need for manual initialization steps and enabling real-time refinement of registration, even after adverse events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual initialization steps are performed to establish endoscope position, then initial position estimate is obtained, but procedure time is prolonged
Solution Approach 1:
The system performs preliminary actions by pre-processing medical imaging data (CT, MRI, or ultrasound scans) to create a 3D luminal network model before the procedure begins. This model includes pre-identified anatomical landmarks and branching structures that will be used during the procedure for rapid position estimation without requiring manual initialization steps during the actual procedure.
Solution Approach 2:
The patent replaces the manual mechanical initialization process (physically positioning the endoscope at specific anatomical landmarks) with an automated image-based system. The machine learning model automatically estimates endoscope position by comparing real-time imaging data with the pre-created 3D luminal network model, eliminating the need for manual mechanical positioning and landmark identification.
2Reliability
If manual initialization is required, then position tracking can begin, but navigation complexity increases due to potential backtracking
Solution Approach 1:
The system implements continuous feedback by repeatedly estimating the endoscope position throughout the procedure using the machine learning model and comparing it with the 3D luminal network model. This real-time feedback allows the system to track position continuously and automatically adjust for deviations caused by adverse events, eliminating the need for manual re-initialization and backtracking.
Solution Approach 2:
The 3D luminal network model with pre-identified anatomical landmarks and branching structures is created before the procedure, providing a ready-reference framework that simplifies real-time position tracking. This preliminary preparation eliminates the need for complex manual navigation decisions during the procedure.
3Measurement precision
If traditional navigation methods are used, then initial position can be established, but accuracy deteriorates after adverse events
Solution Approach 1:
The machine learning model continuously estimates the endoscope position throughout the entire procedure, providing uninterrupted position tracking. This continuous action ensures that position accuracy is maintained even after adverse events, as the system constantly updates the position estimate based on current imaging data rather than relying on periodic manual re-initialization.
Solution Approach 2:
The system creates a comprehensive 3D luminal network model before the procedure that includes detailed anatomical landmarks and branching structures. This pre-prepared reference framework enables rapid and accurate position re-estimation after adverse events without requiring the endoscope to be withdrawn or re-positioned manually.
Data Source
AI summary
Certain aspects relate to systems and techniques for localizing and/or navigating a medical instrument within a luminal network. A medical system can include an elongate body configured to be inserted into the luminal network, as well as an imaging device positioned on a distal portion of the elongate body. The system may include memory and processors configured to receive from the imaging device image data that includes an image captured when the elongate body is within the luminal network. The image can depict one or more branchings of the luminal network. The processor can be configured to access a machine learning model of one or more luminal networks and determine, based on the machine learning model and information regarding the one or more branchings, a location of the distal portion of the elongate body within the luminal network.