Dose Prediction Model Selection for Radiation Therapy Planning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current radiation treatment planning methods face challenges in navigating the Pareto surface effectively, especially with a large number of quality metrics, leading to complex navigation and interpretation difficulties, and the need for more versatile approaches to balance clinical goals.

Innovation Solution

Selecting a dose prediction model based on clinical goals allows for the generation of balanced plans using knowledge-based planning, expanding the evaluation region and incorporating Pareto surface navigation with a reduced number of sliders focused on significant quality metrics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple quality metrics are included in MCO to comprehensively evaluate treatment plans, then the evaluation becomes more thorough and clinically meaningful, but the navigation of the Pareto surface becomes more complex and difficult to interpret

Engineering Contradiction:
Improveevaluation thoroughnessVSAvoidnavigation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the large number of quality metrics into hierarchical groups (e.g., target coverage metrics, organ-at-risk metrics, dose distribution metrics). This segmentation allows planners to navigate the Pareto surface in a structured manner, evaluating metrics within each group separately while maintaining comprehensive evaluation across all groups.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension to the Pareto surface navigation by adding clinical goal priorities and weighting schemes. Instead of navigating only in the space of quality metric values, the system allows navigation in an extended space that includes goal importance weights, transforming the navigation problem from purely geometric to include clinical decision-making dimensions.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If planners manually navigate the Pareto surface using sliders for each quality metric to find optimal plans, then customization to clinical goals is possible, but the process becomes time-consuming and challenging with many metrics

Engineering Contradiction:
Improveclinical goal customizationVSAvoidplanning time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent implements preliminary action by pre-calculating and storing Pareto optimal plans for various combinations of quality metric weights and clinical goals. When a planner needs to evaluate treatment plans, the system retrieves pre-computed Pareto surfaces rather than calculating them in real-time, significantly reducing planning time while maintaining customization capability.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system enables self-service by automatically generating and updating Pareto surfaces based on input clinical goals and quality metrics without requiring manual intervention for each navigation step. The system autonomously performs the optimization calculations and presents results to the planner.

Inventive Principle:
Principle #25Self-service

3Productivity

If the Pareto surface is used to represent tradeoffs between clinical goals, then comprehensive optimization is achieved, but it becomes difficult to define clinically meaningful points in the domain

Engineering Contradiction:
Improveoptimization effectivenessVSAvoidclinical meaningfulness identification
Core Design Contradiction:
ProductivityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent introduces clinical expert rules and guidelines as intermediaries between the mathematical Pareto surface and clinical decision-making. These rules translate the abstract tradeoffs represented on the Pareto surface into clinically meaningful recommendations by incorporating established medical knowledge and practice patterns.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system allows dynamic adjustment of parameter interpretations on the Pareto surface based on clinical context. By changing how metrics are weighted and prioritized according to specific clinical scenarios, the same Pareto surface can reveal different clinically meaningful points for different patient situations.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3342461B1Selecting a dose prediction model based on clinical goals
Publication Date: 2023.09.27 SIEMENS HEALTHINEERS INTERNATIONAL AG
  • EP3342461B1 patent drawingFigure 1~2
  • EP3342461B1 patent drawingFigure 3
  • EP3342461B1 patent drawingFigure 4

AI summary

A clinical goal for radiation treatment of a patient is set. A dose prediction model 150 is selected from a number of dose prediction models based on the clinical goal. A radiation treatment plan is then generated for the patient using the dose prediction model that was selected based on the clinical goal.