Machine-Learned Radiotherapy Planning for Faster Dose Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current radiotherapy treatment planning is time-consuming and complex due to the need for manual optimization of radiation beam parameters to balance dose delivery to tumors and minimize damage to surrounding healthy tissue, particularly when dealing with multiple organs at risk, and is hindered by anatomical variations between patients.

Innovation Solution

A machine learning model, specifically a deep neural network, is trained to estimate optimization variables for radiotherapy treatment plans, enabling faster solution of optimization problems using conventional optimization methods.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual optimization techniques are used to balance dose delivery to tumors and minimize damage to surrounding healthy tissue, then treatment plan quality can be optimized, but the time required to generate treatment plans increases significantly

Engineering Contradiction:
Improvetreatment plan qualityVSAvoidtreatment plan generation time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary segmentation of anatomical structures and pre-calculates dose distributions for multiple beam configurations before final optimization. This advance preparation reduces the time required for manual plan generation while maintaining quality by having critical data ready beforehand.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

An automated optimization algorithm acts as an intermediary between the planner's objectives and the final treatment plan. This intermediary automatically adjusts beam parameters, fluence maps, and other variables to satisfy dose constraints, reducing the need for time-consuming manual trial-and-error adjustments while preserving plan quality.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If the number of organs at risk (OARs) is increased for comprehensive coverage, then treatment plan completeness improves, but the complexity of the optimization process increases

Engineering Contradiction:
Improvetreatment plan completenessVSAvoidoptimization process complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The optimization problem is segmented into manageable components: critical structure identification, constraint generation, dose calculation, and parameter optimization. Each segment handles specific OARs and their associated constraints separately, making the overall complex process more tractable and automated.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system automatically adjusts optimization parameters such as weighting factors, constraint thresholds, and beam parameters based on the specific anatomy and OARs present. This dynamic parameter adaptation allows comprehensive OAR coverage without requiring manual reconfiguration of the optimization process for each case.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If conventional optimization methods are used to solve radiotherapy treatment plan problems, then accurate dose distribution can be achieved, but the computational time required increases

Engineering Contradiction:
Improvedose distribution accuracyVSAvoidcomputational time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The system pre-calculates dose kernels, beam geometry parameters, and anatomical segmentation data before the optimization process. This preliminary computation prepares all necessary data structures and matrices, reducing the computational burden during the actual optimization while maintaining dose distribution accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

Traditional iterative numerical optimization methods are replaced or augmented with machine learning-based prediction models. These models predict optimal beam parameters and fluence maps directly from anatomical data, bypassing time-consuming iterative calculations while preserving dosimetric accuracy through learned patterns from training data.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3999171B1Radiotherapy treatment plan optimization using machine learning
Publication Date: 2025.11.19 ELEKTA AB
  • EP3999171B1 patent drawingFigure 1
  • EP3999171B1 patent drawingFigure 2A
  • EP3999171B1 patent drawingFigure 2B

AI summary

Techniques for solving a radiotherapy treatment plan optimization problem are provided. The techniques include receiving a radiotherapy treatment plan optimization problem; processing the radiotherapy treatment plan optimization problem with a machine learning model to estimate one or more optimization variables of the radiotherapy treatment plan optimization problem, wherein the machine learning model is trained to establish a relationship between the one or more optimization variables and parameters of a plurality of training radiotherapy treatment plan optimization problems; and generating a solution to the radiotherapy treatment plan optimization problem based on the estimated one or more optimization variables of the radiotherapy treatment plan optimization problem.