Robust Radiotherapy Plan Generation via Scenario Overlap
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radiotherapy treatment planning methods, such as margin-based and probabilistic approaches, fail to generate robust treatment plans universally applicable across different clinical cases and treatment modalities, especially in regions of heterogeneous density, leading to suboptimal results.
Innovation Solution
A method that involves accessing scenarios representing uncertainties, determining mappings of a region of interest for each scenario, calculating weights based on the overlap of these mappings, and optimizing an optimization function using these weights to generate robust treatment plans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If margin-based treatment planning is used, then robustness to uncertainties is improved, but the Planning Target Volume becomes larger than necessary, increasing dose to healthy tissue
Solution Approach 1:
The patent changes the parameter representation from fixed geometric margins to probability-based scenario mappings. By representing uncertainties as probability distributions and mapping ROIs across multiple scenarios with associated probabilities, the method dynamically adjusts the effective treatment volume based on actual uncertainty realizations rather than applying uniform margins, thereby reducing unnecessary volume expansion while maintaining robustness.
2Reliability
If probabilistic optimization approaches are used, then robust treatment plans are generated for heterogeneous density regions, but the dose distribution becomes blurred with blunt fall-offs
Solution Approach 1:
The patent performs preliminary action by pre-defining multiple uncertainty scenarios and their probability distributions before optimization. By preparing scenario mappings in advance and using them to guide the optimization process, the method maintains sharp dose fall-offs while accounting for uncertainties, avoiding the blurred distributions that result from conventional probabilistic approaches that treat all scenarios equally during optimization.
3Reliability
If worst case scenario optimization is used, then robust treatment plans are generated, but the plan becomes too conservative in many situations
Solution Approach 1:
The patent changes the optimization parameter from worst-case scenario selection to probability-weighted scenario evaluation. By incorporating probability distributions that reflect the actual likelihood of different uncertainty realizations, the method adjusts the effective conservatism level dynamically - highly probable scenarios have greater influence while improbable worst cases have minimal impact, producing adaptable plans that are robust without being excessively conservative.
4Device complexity
If a single treatment plan is generated without scenario consideration, then the planning process is simple, but the plan is not robust to uncertainties in target position, cancer cell location, and patient density
Solution Approach 1:
The patent segments the treatment planning process by dividing uncertainties into distinct scenarios with specific probability distributions. Each scenario represents a particular realization of uncertainties (target position, cancer cell location, density variations), and the overall robust plan is constructed by optimizing across these segmented scenarios. This segmentation approach maintains manageable complexity while systematically addressing multiple uncertainty sources.
Data Source
AI summary
A method for generating a robust radiotherapy treatment plan, using scenario-based robust optimization, is provided. Weights which are dependent on the overlap of different scenario-specific mappings of a region of interest is used in the optimization.