Robust Dose Prediction Models for Radiation Therapy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radiation treatment planning methods lack robustness in accounting for perturbations during treatment, such as patient movement, which can lead to suboptimal dose distribution and increased exposure to healthy tissues.
Innovation Solution
Generating robust dose prediction models that account for perturbations by determining field-specific planning target volumes, delineating organ-at-risk sub-volumes, and training multiple dose prediction models using nominal and perturbed dose distributions to predict dose-volume histograms and optimize treatment plans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional dose prediction models are used, then the treatment planning process is simpler, but the models fail to account for patient movement and perturbations during treatment
Solution Approach 1:
The patent applies preliminary action by generating multiple perturbed treatment plans before the actual treatment to account for potential patient movements and setup errors. By pre-calculating dose distributions for various perturbation scenarios (e.g., patient position shifts, organ motion), the model prepares robustness against future uncertainties without adding complexity to the actual treatment delivery process
Solution Approach 2:
The patent implements dynamics by creating a flexible dose prediction model that can adapt to different perturbation scenarios. The system dynamically adjusts dose predictions based on various hypothetical patient positions and anatomical variations, allowing the treatment plan to remain robust even when actual treatment conditions deviate from the nominal plan
2Manufacturing precision
If the treatment plan optimizes for one clinical goal, then that specific goal is improved, but other conflicting clinical goals may deteriorate
Solution Approach 1:
The patent applies segmentation by dividing the treatment planning process into multiple independent optimization scenarios, each addressing a specific perturbation case. By segmenting the dose prediction into multiple perturbed plans (e.g., different patient positions, organ locations), the system can evaluate and optimize for multiple conflicting clinical goals simultaneously, then integrate results to achieve a robust overall treatment plan
3Reliability
If multiple perturbed treatment plans are generated, then robustness against patient movement is improved, but the treatment planning time increases
Solution Approach 1:
The patent applies partial action by selecting a representative subset of perturbation scenarios rather than exhaustively evaluating all possible patient movements and anatomical variations. By focusing on the most clinically relevant perturbations (e.g., typical setup errors, expected organ motion ranges), the system achieves robustness without requiring excessive computational time for planning
Data Source
AI summary
Nominal values of parameters, and perturbations of the nominal values, that are associated with previously defined radiation treatment plans are accessed. For each treatment field of the treatment plans, a field-specific planning target volume (fsPTV) is determined based on those perturbations. At least one clinical target volume (CTV) and at least one organ-at-risk (OAR) volume are also delineated. Each OAR includes at least one sub-volume that is delineated based on spatial relationships between each OAR and the CTV and the fsPTV for each treatment field. Dose distributions for the sub-volumes are determined based on the nominal values and the perturbations. One or more dose prediction models are generated for each sub-volume. The dose prediction model(s) are trained using the dose distributions.