Memetic Optimization Algorithm for Radiation Therapy Planning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current radiation therapy treatment planning methods are computationally intensive and take hours to develop, often resulting in suboptimal dose homogeneity and conformality, with increased exposure to surrounding tissues and organs.

Innovation Solution

A method using memetic optimization algorithms, including population-based heuristics and meta-heuristics, to determine beam geometry and dosimetric parameters, calculate dose deposition coefficients, and generate solutions that minimize radiation exposure to healthy tissues while ensuring adequate tumor dose, employing techniques such as evolutionary computation and swarm intelligence to optimize beam weights and positions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If known inverse planning algorithms are used to optimize treatment plans, then dose homogeneity and conformality are improved, but computational time increases significantly taking hours to develop

Engineering Contradiction:

Improvedose homogeneityVSAvoidcomputational time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent replaces traditional mechanical/mathematical optimization algorithms with a neural network-based intelligent system. The neural network is trained offline to learn the complex relationship between treatment parameters and dose distribution, enabling it to rapidly generate optimized treatment plans without requiring intensive real-time computational resources.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The neural network performs preliminary learning and pattern recognition during an offline training phase, storing optimized solutions and dose distribution patterns in its weight structures. During actual treatment planning, the pre-trained network can rapidly retrieve and adapt solutions without repeating the intensive optimization calculations from scratch.

Inventive Principle:

Principle #10Preliminary action

2Manufacturing precision

If known inverse planning algorithms are used to optimize treatment plans, then dose conformality to target volume is improved, but computational resources and processing time are excessively consumed

Engineering Contradiction:

Improvedose conformalityVSAvoidtreatment plan development speed

Core Design Contradiction:

Manufacturing precisionVSProductivity

Solution Approach 1:

The patent substitutes traditional iterative mathematical optimization algorithms with a neural network system that has learned optimal dose conformality patterns during training. The neural network processes treatment planning data through its layered structure, rapidly producing conformal treatment plans without the computational burden of traditional algorithms.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The neural network creates simplified representations (copies) of complex dose distribution patterns and optimization relationships during training. These learned patterns are stored in the network's weight structures and can be rapidly replicated and adapted for new treatment cases, avoiding repeated intensive calculations.

Inventive Principle:

Principle #26Copying

3Device complexity

If traditional treatment planning methods are used, then computational simplicity is maintained, but radiation exposure to surrounding healthy tissues and organs increases

Engineering Contradiction:

Improvecomputational complexityVSAvoidradiation exposure to healthy tissue

Core Design Contradiction:

Device complexityVSObject-affected harmful factors

Solution Approach 1:

The patent replaces simple forward planning methods with an intelligent neural network system that has learned to minimize radiation exposure to healthy tissues. The neural network optimizes beam parameters by processing complex relationships between beam geometry, tissue density, and dose distribution, automatically reducing harmful radiation exposure without requiring complex manual adjustments.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The neural network performs self-optimization of treatment plans by automatically adjusting beam parameters to minimize radiation exposure to healthy tissues. The system uses its learned knowledge to autonomously identify and avoid critical structures, eliminating the need for extensive manual optimization while reducing harmful radiation exposure.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS8792614B2System and method for radiation therapy treatment planning using a memetic optimization algorithm

Publication Date: 2014.07.29 WITTEN MATTHEW R

AI summary

A method for the optimization of radiation therapy treatment plans is disclosed. The disclosed method is equally-applicable to robotic radiosurgery as well as other types of radiosurgical delivery, intensity-modulated radiotherapy (IMRT), volumetric modulated arc therapy (VMAT), and three-dimensional conformal radiotherapy (3DCRT). A population-based heuristic approximation is used to perform a global search, and subsequently, a deterministic local trajectory search is employed to further refine the initial solution.