Robust Radiotherapy Planning via Dose Distribution Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radiotherapy treatment planning methods are uncertain due to anatomical changes during treatment, as they fail to accurately account for dose mapping uncertainties caused by tumor shifts and deformations, leading to potentially conservative or inaccurate dose delivery.
Innovation Solution
A robust optimization approach that considers the uncertainty in deformation vector fields by generating a distribution of mapped doses using multiple image registrations and error estimates, allowing for more reliable dose mapping and treatment planning that accounts for anatomical changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If deformable image registration is used to map doses between images, then anatomical changes can be accommodated, but uncertainty in the deformation map increases
Solution Approach 1:
The patent applies dynamics by making the dose mapping process adaptive rather than static. Multiple deformation vector fields are generated to represent different plausible anatomical configurations, and the optimization dynamically adjusts the treatment plan to be robust across all these scenarios. This resolves the contradiction by accepting the uncertainty (worsening reliability) but compensating through dynamic adaptation (improving versatility) to ensure adequate dose delivery.
Solution Approach 2:
The patent changes the parameter representation from a single deterministic deformation map to a distribution of multiple plausible deformation vector fields. By representing the deformation as a probability distribution rather than a fixed value, the system accounts for the inherent uncertainty in DIR while maintaining adaptability to anatomical changes. This parameter transformation resolves the contradiction by explicitly modeling uncertainty rather than ignoring it.
2Ease of manufacture
If a single deformation vector field is trusted for dose mapping, then the process is simple, but the dose delivery may be inaccurate in regions of anatomical change
Solution Approach 1:
The patent applies preliminary action by generating multiple deformation vector fields and evaluating dose distributions before final treatment plan optimization. By preparing multiple plausible scenarios in advance and incorporating them into the optimization process, the system ensures accurate dose delivery without requiring complex manual corrections during treatment. This resolves the contradiction by performing the computationally intensive work beforehand, maintaining simplicity during actual treatment while ensuring accuracy.
3Reliability
If multiple image registrations are performed to account for uncertainty, then dose mapping robustness improves, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the treatment planning process into distinct stages: generating multiple deformation vector fields, mapping doses for each scenario, and performing robust optimization. By segmenting the complex task of handling uncertainty into manageable components, the system achieves robust dose delivery while keeping computational complexity tractable through structured processing of multiple scenarios.
4Reliability
If conservative dose estimates are applied to account for uncertainty, then safety is improved, but the treatment may be overly conservative and reduce productivity
Solution Approach 1:
The patent applies feedback by using the results from multiple deformation scenario evaluations to inform and adjust the treatment plan optimization. The optimization process receives feedback from all plausible deformation scenarios and adjusts beam parameters to achieve robust dose delivery. This resolves the contradiction by using systematic feedback from multiple scenarios to find the optimal balance between safety and efficiency, rather than applying blanket conservative estimates that reduce productivity.
Data Source
AI summary
Generating a robust radiotherapy treatment plan for a treatment volume, defined using a plurality of voxels, of a subject. A first and at least one second image of the treatment volume are received. A distribution of mapped doses in the first image is generated by mapping a dose defined in the at least one second image to the first image using image registration. An optimization problem is defined using at least one optimization function for a total dose, related to the radiotherapy treatment. At least one optimization function value is calculated based on at least two mapped doses in the distribution of mapped doses in the first image. A radiotherapy treatment plan is generated by optimizing the at least one optimization function value evaluated by taking into account the at least two mapped doses in the distribution of mapped doses.