Online Adaptive Radiation Therapy Replanning With ML Contour Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The lengthy and clinically impractical time required for segmenting anatomy and evaluating radiation treatment plans in online adaptive radiation therapy (ART) hinders efficient delivery, necessitating faster and more robust methods for segmentation and plan evaluation.
Innovation Solution
Implementing systems and methods that utilize textural analysis and machine learning algorithms to rapidly identify the need for online adaptive replanning (OLAR), automatically correct inaccurate contours, and generate synthetic computed tomography (CT) images from magnetic resonance images, thereby improving the efficiency and accuracy of radiation therapy delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional manual segmentation and plan evaluation methods are used in online adaptive radiation therapy, then accuracy of anatomy segmentation and plan evaluation is maintained, but treatment delivery time becomes excessively long (10-30 minutes)
Solution Approach 1:
The patent uses auto-segmentation algorithms to create automated copies of anatomy contours based on imaging data, replacing manual contouring. This allows rapid generation of segmentation results that closely approximate manual segmentation accuracy, reducing treatment time while maintaining sufficient precision for clinical decision-making in online adaptive radiation therapy
Solution Approach 2:
The patent replaces the mechanical manual process of contour drawing and plan evaluation with automated computer-based algorithms. Machine learning models and automated planning systems perform segmentation and evaluation tasks that previously required manual intervention, dramatically reducing treatment delivery time while maintaining or improving consistency and accuracy
2Productivity
If automated auto-segmentation algorithms are used to reduce treatment time, then treatment delivery speed increases (to 5-10 minutes), but accuracy and reliability of contour segmentation may deteriorate
Solution Approach 1:
The patent incorporates feedback mechanisms where auto-segmentation results are automatically evaluated against quality criteria, and controversial cases are flagged for manual review. This feedback loop ensures that automated segmentation maintains high accuracy by catching and correcting potential errors, while still achieving rapid processing for the majority of cases
Solution Approach 2:
The patent performs preliminary automated segmentation and evaluation before final treatment delivery, allowing time for quality verification and manual correction if needed. This preliminary action ensures that even though automation is used, there is a buffer to verify accuracy before the treatment is executed
Data Source
AI summary
Systems and methods for accelerated online adaptive radiation therapy (“ART”) are described. The improvements to online ART are generally provided based on the use of textural analysis and machine learning algorithms implemented with a hardware processor and a memory. The described systems and methods enable more efficient and accurate online adaptive replanning (“OLAR”), which can also be implemented in clinically acceptable timeframes. For example, OLAR can be reduced from taking 10-30 minutes down to 5-10 minutes.