Single X-ray Imager 3D Position Estimation via Spatial Probability Density
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radiotherapy methods struggle to accurately estimate three-dimensional target positions from two-dimensional projections, especially during prostate cancer treatment, due to the ambiguity in reconstructing actual tumor trajectories from CBCT projections and the limitations of single gantry-mounted kV x-ray imagers.
Innovation Solution
A method using a single x-ray imager to estimate retrospective and real-time 3D target positions by determining 2D position components and resolving position components along imager axes with a spatial probability density, specifically employing a Gaussian distribution centered at the isocenter or patient coordinate system, and utilizing maximum likelihood estimation to determine the target position and trajectory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single x-ray imager is used to capture 2D projections, then device complexity is reduced, but measurement precision of 3D target position deteriorates due to ambiguity in trajectory reconstruction
Solution Approach 1:
The patent introduces radiopaque markers as intermediary objects attached to or near the target. These markers serve as visible surrogates that can be reliably detected in 2D projections, enabling indirect measurement of 3D target position through mathematical reconstruction algorithms that use the markers' known geometric relationships
Solution Approach 2:
The patent acquires multiple 2D projections from different angular positions during a gantry rotation and uses tomographic reconstruction algorithms to recover 3D target position information. This transforms the problem from direct 3D measurement to indirect 3D reconstruction through multiple 2D views, resolving the ambiguity inherent in single-view 2D imaging
2Measurement precision
If CBCT projections are used for 3D target position estimation, then measurement capability is improved, but reliability of trajectory reconstruction deteriorates due to infinite possible 3D trajectories producing identical 2D projections
Solution Approach 1:
The patent attaches radiopaque markers with specific geometric configurations (e.g., different shapes, sizes, or orientations) to different locations on or near the target. These markers have distinct local visual characteristics in the projections, enabling the reconstruction algorithm to distinguish between different 3D configurations and resolve the ambiguity of infinite possible trajectories
Solution Approach 2:
The patent pre-attaches radiopaque markers to the target or target surrogate before treatment begins. This preliminary placement establishes a known geometric reference frame that constrains the possible 3D trajectories, allowing reliable reconstruction by comparing observed marker positions in projections against the known geometric relationships established beforehand
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 achieves accurate 3D target position estimation with errors less than 1 mm, enabling real-time tracking and motion compensation during radiotherapy, improving the precision of tumor trajectory estimation and reducing uncertainties in prostate and tumor motion monitoring.
Implementation Method 1
imaging a target to determine a two-dimensional (2D) position of the target
Data Source
AI summary
The present invention provides a method for estimation of retrospective and real-time 3D target position by a single imager. The invention includes imaging a target on at least one 2D plane to determine 2D position and/or position components of the target, and resolving a position and/or position component along at least one imager axis of the target using a spatial probability density. The present invention provides a probability-based method for accurate estimation of the mean position, motion magnitude, motion correlation, and trajectory of a tumor from CBCT projections. The applicability of the method for tumors with periodic respiratory motion and for prostate are provided. Clinical feasibility is demonstrated for a pancreas tumor. The method includes monoscopic tracking of the 3D prostate position utilizing the spatial probability density to estimate the unresolved motion from the resolved motion. The method is applicable to prostate tracking even with a population-based probability density.