Intracardiac Echocardiography Volume Stitching via Preoperative Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Percutaneous cardiac interventions using intracardiac echocardiography (ICE) face limitations due to a small spatial angle of view, making it difficult for clinicians to recognize and guide anatomical structures within the heart, as existing technologies rely heavily on ICE intensities for volume registration, which can be inaccurate.
Innovation Solution
The method involves spatially aligning and combining ICE volumes with a preoperative physiological model of the patient's cardiac system, using machine-trained classifiers to register and fuse ultrasound data, thereby increasing the field of view and improving accuracy by using a patient-specific model for guidance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ICE volumes are registered using ICE intensities, then the registration process is simple, but the accuracy is poor
Solution Approach 1:
A preoperative physiological model serves as an intermediary between multiple ICE volumes and the final stitched volume. The model provides a common reference coordinate system that mediates the registration process, replacing direct ICE intensity-based alignment with model-based feature and surface matching, thereby improving accuracy while maintaining manageable complexity through automated classification algorithms
Solution Approach 2:
A preoperative model (copy of the patient's cardiac anatomy from preoperative imaging) is created and used as a reference for registering ICE volumes. This copy contains accurate anatomical features and surfaces that serve as reliable matching targets, eliminating the need for direct ICE intensity comparison and significantly improving registration precision
2Area of stationary object
If a single ICE volume is used, then the device complexity is low, but the field of view is limited
Solution Approach 1:
Multiple ICE volumes are merged into a single stitched volume by registering them to a common preoperative model and combining their data in a unified coordinate system. This merging process expands the effective field of view while using automated feature matching and machine-trained classifiers to manage the complexity of integrating multiple data sets
Solution Approach 2:
The approach transitions from viewing individual ICE volumes in isolation to integrating them in a unified multi-dimensional coordinate system based on the preoperative model. This dimensional framework allows volumes from different spatial perspectives to be combined, effectively expanding the field of view while providing a structured method to handle the complexity
3Productivity
If real-time stitching is implemented, then the productivity is improved, but the processing complexity increases
Solution Approach 1:
The preoperative physiological model is created and prepared before the ICE procedure, with all anatomical features and surfaces pre-processed and stored in a ready-to-use format. This preliminary preparation eliminates the need for complex real-time model construction, enabling fast real-time stitching by simply matching ICE volumes to the pre-existing model using machine-trained classifiers
Solution Approach 2:
Traditional manual or intensity-based registration methods are replaced with automated feature and surface matching algorithms enhanced by machine-trained classifiers. This substitution of mechanical/manual processes with intelligent automated systems accelerates the stitching process to real-time speeds while managing processing complexity through efficient classification techniques
Data Source
AI summary
Different intracardiac echocardiography volumes are stitched together. Different volumes of a patient are scanned with ICE. To stitch the volumes together, creating a larger volume, the volumes are spatially aligned. The alignment is based on feature, surface, or both feature and surface matching of the ICE volumes with a preoperative model of the same patient. The matching with the model indicates a relative position of the ICE volumes with each other. Using machine-trained classifiers may speed performance, allowing for real-time assembling of a volume from ICE data as the catheter is moved within the patient.