Related Experiment Videos
Large deformation three-dimensional image registration in image-guided radiation therapy
Mark Foskey1, Brad Davis, Lav Goyal
1Department of Radiation Oncology, University of North Carolina, USA. mark_foskey@unc.edu
Physics in Medicine and Biology
|December 8, 2005
Summary
This study introduces a new deformable image registration framework to accurately process 3D CT scans for image-guided radiation therapy, even with challenging bowel gas artifacts. The method improves organ segmentation and assesses radiation dose effects from tissue motion.
Area of Science:
- Medical Physics
- Image Analysis
- Radiation Oncology
Background:
- Deformable image registration (DIR) is crucial for image-guided radiation therapy (IGRT).
- Standard DIR algorithms assume point-wise correspondence between images, which is often invalidated by intra-treatment anatomical variations like bowel gas.
- Prostate cancer radiation therapy is particularly affected by these variations, impacting treatment accuracy.
Purpose of the Study:
- To present and validate a novel DIR framework capable of handling regions with no image correspondence, specifically addressing bowel gas in prostate IGRT.
- To demonstrate the utility of this enhanced DIR framework for automated organ segmentation and dosimetric analysis.
- To evaluate the impact of non-rigid tissue motion on radiation dose delivery and biological effect.
Main Methods:
- Developed a deformable image registration framework extending existing algorithms to accommodate regions lacking correspondence.
- Applied the registration technique for automated organ segmentation and performed statistical validation against human raters.
- Utilized the framework to assess the dosimetric and biological effects (using a linear-quadratic model) of non-rigid tissue motion during treatment.
Main Results:
- The proposed DIR framework successfully processes serial 3D CT images, accounting for regions without point correspondence.
- The organ segmentation method derived from the registration technique demonstrated high accuracy when compared to multiple human raters.
- The study quantified the dosimetric and biological impact of tissue motion, providing insights into treatment plan robustness.
Conclusions:
- The developed DIR framework offers a robust solution for processing IGRT images affected by anatomical variations like bowel gas.
- This approach enhances organ segmentation accuracy and provides a method to evaluate the clinical impact of tissue motion in radiation therapy.
- The framework is applicable beyond prostate cancer treatment, offering broad utility in image-guided interventions.