Differentiable Dose Functions for Faster Radiotherapy Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radiotherapy treatment planning is time-consuming and complex, often requiring a trial-and-error process due to the presence of multiple organs at risk (OARs), and existing optimization methods fail to accurately consider dose calculations, leading to inefficient and inaccurate treatment plans.
Innovation Solution
A computer-implemented method using a machine learning model that estimates radiotherapy treatment plan parameters based on a derivative of dose calculations, incorporating a loss function to optimize treatment plans more efficiently and accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional optimization techniques are used to create treatment plans, then treatment plans can be generated considering clinical and dosimetric objectives, but the process becomes time-consuming and complex
Solution Approach 1:
The patent replaces traditional manual optimization processes with a machine learning model that uses neural networks to predict treatment plan parameters. The system substitutes the mechanical trial-and-error optimization process with an AI-based predictive model that can generate treatment plans much faster while maintaining accuracy through training on clinical outcomes.
Solution Approach 2:
The patent changes the approach from traditional optimization parameters to machine learning parameters. Instead of iteratively adjusting treatment parameters to minimize dose to OARs, the system uses a neural network trained on historical data to directly predict optimal treatment parameters, transforming the optimization problem into a prediction problem.
2Ease of operation
If manual trial-and-error optimization is used, then treatment plans can be adjusted to meet clinical objectives, but the process becomes complicated by the presence of multiple OARs
Solution Approach 1:
The patent replaces the complex manual optimization process with an automated machine learning system. The neural network model automatically handles the complexity of multiple OARs by learning from historical treatment data, eliminating the need for planners to manually navigate through complex optimization scenarios.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between the treatment planning problem and the solution. This intermediary absorbs the complexity of multiple OARs and optimization constraints, translating them into a prediction problem that can be solved efficiently without requiring manual intervention in the complex optimization process.
3Measurement precision
If traditional optimization methods are used, then treatment plans can be generated, but they fail to accurately consider dose calculations
Solution Approach 1:
The patent substitutes traditional dose calculation methods with a machine learning-based approach. The neural network model is trained on accurate dose calculations from historical data, allowing it to predict treatment parameters that directly account for dose distribution accuracy without requiring time-consuming iterative optimization.
Solution Approach 2:
The patent performs preliminary training of the machine learning model on accurately calculated dose data from historical treatments. This preliminary action allows the model to learn the relationship between treatment parameters and dose distributions in advance, enabling fast and accurate treatment plan generation without requiring time-consuming dose calculations during the planning process.
Data Source
AI summary
Techniques for generating a radiotherapy treatment plan parameter are provided. The techniques include receiving radiotherapy treatment plan information; processing the radiotherapy treatment plan information to estimate one or more radiotherapy treatment plan parameters based on a process that depends on the output of a subprocess that estimates a derivative of a dose calculation; and generating a radiotherapy treatment plan using the estimated one or more radiotherapy treatment plan parameters.