MRI Device Tracking via 1D Projections and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current MRI imaging methods for device tracking during medical interventions are inefficient due to the need for manual tracking of 2D layers, which is time-consuming, and 3D imaging offers lower time resolution, making it challenging to accurately position interventional devices like needles and catheters in real-time.
Innovation Solution
A method using one-dimensional (1D) projections with off-resonance spin excitation pulses to localize MR signals along a single axis, combined with machine-learned algorithms to adjust projection parameters for precise positional information of interventional devices, enabling fast and accurate device tracking without requiring optimally positioned imaging planes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If 2D MRI imaging methods are used for online monitoring, then the time resolution is fast, but manual tracking of 2D layers is time-consuming and positioning accuracy deteriorates
Solution Approach 1:
The system performs automatic device tracking through computer implementation, where the MRI system itself automatically determines device position from projections without requiring manual intervention. The computer-implemented method automatically processes projection data to track device location, eliminating the need for operators to manually track 2D layers while maintaining fast time resolution.
Solution Approach 2:
The patent replaces manual mechanical tracking operations with automated computational processing. Instead of manually analyzing 2D image layers, the system uses computer-implemented algorithms to automatically determine device position from 1D projections, substituting human-operated mechanical tracking with automated digital processing.
2Measurement precision
If 3D MRI imaging is used to avoid positioning problems, then device position is accurately captured, but time resolution deteriorates significantly
Solution Approach 1:
The patent segments the 3D imaging problem into multiple 1D projection measurements taken at different angles. Instead of acquiring a complete 3D volume, the system collects sequential 1D projections along different directions and reconstructs device position from these segmented measurements, achieving accurate positioning with much faster acquisition than full 3D imaging.
Solution Approach 2:
The system performs partial imaging by acquiring only the necessary 1D projection data required for device localization rather than complete 3D volumetric data. This partial action approach collects sufficient information for positioning purposes without the excessive time cost of full 3D imaging, achieving the minimum necessary measurement for the tracking function.
3Loss of information
If conventional MRI imaging is used, then the device itself is visualized, but signal from the device is absent and only extinction artifacts are visible
Solution Approach 1:
The patent introduces an intermediary approach by detecting the device indirectly through its effect on surrounding MR signals rather than attempting to detect the device directly. The method uses the device's magnetic susceptibility effects to modulate signals from adjacent tissues, creating visible projections that indicate device position without requiring the device itself to generate a signal.
Solution Approach 2:
The patent converts the harmful magnetic susceptibility artifacts and signal extinction caused by the device into beneficial localization information. Instead of treating the signal loss and distortion as problems to be eliminated, the method exploits these effects to create positive contrast projections that clearly indicate device position, turning the previously harmful artifacts into useful tracking signals.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for rapid and accurate localization of interventional devices, reducing intervention time, increasing position accuracy, and eliminating the need for manual tracking, thereby enhancing the robustness and reproducibility of device positioning during medical procedures.
Implementation Method 1
an off resonance spin excitation pulse, which excites spins in the immediate vicinity of the interventional device based on local B0 magnetic field distortions by the interventional device
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for localization of an interventional device introduced into an examination object in an examination volume of a MRI system is provided. A least one white marker contrast projection of the examination volume is acquired based on at least one projection parameter. Positional information of the interventional device in the examination volume is determined using the projection. A MRI image of a slice of the examination volume including at least part of the interventional device is acquired using the determined positional and/or orientational information of the interventional device. The at least one projection parameter is adjusted, wherein adjusting the at least one projection parameter comprises applying trained functions to the at least one acquired projection, wherein the trained functions were trained with training projections associated with known positional information of an interventional device in each training projection.