AI Material Property Tracking for Radiation Therapy Targets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional radiation therapy methods face challenges in accurately tracking moving target structures due to patient motion, particularly when relying on implanted fiducial markers, which can be unreliable and invasive.
Innovation Solution
A template-based, markerless approach for target structure tracking is implemented using material properties such as density and effective atomic number, generating templates from planning images and matching them with real-time treatment images using AI engines to improve positional verification and dose accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If implanted fiducial markers are used for target structure tracking, then tracking can be performed, but additional implantation risks arise and markers may migrate becoming unreliable
Solution Approach 1:
The patent extracts the tracking function from implanted markers and relocates it to naturally occurring anatomical structures. By using AI-based segmentation to identify and track organs-at-risk and tumors directly in treatment images, the system eliminates the need for foreign body implants while maintaining tracking capability. This resolves the contradiction by removing the harmful implantation aspect while preserving the beneficial tracking function.
Solution Approach 2:
The patent creates a digital copy or model of the target structure's material properties and geometry from planning images. This virtual model is then used for tracking by comparing it with treatment images through AI-based segmentation and registration. This copying approach replaces physical markers with a digital representation, eliminating implantation risks while maintaining reliable tracking through material property analysis.
2Measurement precision
If conventional template matching is used without material property analysis, then processing is simpler, but tracking accuracy decreases due to patient motion
Solution Approach 1:
The patent changes the parameters used for template matching from simple geometric or intensity-based comparisons to material property-based parameters. By analyzing effective atomic number, electron density, and other material characteristics of the target structure, the system achieves more accurate tracking that is insensitive to motion and deformation. This parameter transformation resolves the contradiction by improving accuracy through physics-based properties while the AI automation keeps processing complexity manageable.
Solution Approach 2:
The patent replaces conventional mechanical or geometric template matching methods with an AI-based system that uses material property analysis. Instead of relying on rigid geometric transformations, the system uses neural networks to segment and identify target structures based on their physical properties in treatment images. This substitution improves tracking accuracy under motion conditions while the automated AI processing handles the increased complexity.
3Measurement precision
If material property-based template matching is implemented, then tracking accuracy improves, but computational requirements and processing time increase
Solution Approach 1:
The patent performs preliminary action by pre-calculating and storing material property parameters of the target structure during the planning phase. These pre-computed material property profiles are then used as reference templates during treatment. By preparing the material property data in advance, the system reduces the computational burden during real-time treatment, allowing accurate tracking without excessive processing time delays.
Solution Approach 2:
The patent implements self-service by using the target structure's own material properties as the tracking signature. The AI system automatically segments the target in treatment images and extracts its material characteristics without requiring external markers or complex reference systems. This self-based approach streamlines the process and reduces computational overhead compared to marker-based or cross-modality registration methods.
Data Source
AI summary
Example methods and systems for image data processing for target structure tracking are described. In one example, a computer system may obtain treatment image data associated with a target structure of a patient requiring radiation therapy. The treatment image data may be acquired using an imaging system during a treatment phase of the radiation therapy. The computer system may process the treatment image data using an artificial intelligence (AI) engine to generate material property data representing a particular material property associated with the target structure. The material property data may be generated to be matchable against a template that also represents the particular material property for tracking the target structure based on the particular material property during the treatment phase.