Integrated segmentation and nonrigid registration for application in prostate image-guided radiotherapy
Chao Lu1, Sudhakar Chelikani, Zhe Chen
1Department of Electrical Engineering, Yale University, New Haven, CT, USA. chao.lu@yale.edu
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|October 1, 2010
Summary
This study presents an automated method for segmenting and registering images in image-guided radiotherapy (IGRT), improving prostate cancer treatment accuracy. The new approach matches manual segmentation quality while enhancing image registration performance.
Area of Science:
- Medical Physics
- Radiotherapy Technology
- Medical Imaging
Background:
- Image-guided radiotherapy (IGRT) systems use in-room cone-beam CT (CBCT) for treatment verification.
- Accurate segmentation of anatomical structures and image registration are crucial for radiotherapy treatment parameters.
- Current methods rely on manual segmentation of CBCT images, which is time-consuming and may introduce variability.
Purpose of the Study:
- To develop and evaluate an integrated automatic segmentation and constrained nonrigid registration method for 3D CBCT images.
- To simultaneously achieve accurate segmentation of organs at risk and robust image registration.
- To improve the efficiency and accuracy of image-guided radiotherapy workflows.
Main Methods:
- An integrated automatic segmentation and constrained nonrigid registration algorithm was developed.
- The method was tested on 24 sets of real patient data.
- Quantitative analysis compared the automated method against manual segmentation and traditional registration techniques.
Main Results:
- The automated segmentation achieved accuracy comparable to manual segmentation.
- The constrained nonrigid registration significantly outperformed both rigid and non-rigid registration methods.
- Clinical application demonstrated promising results for treatment parameter determination.
Conclusions:
- The integrated automatic segmentation and registration method offers a viable alternative to manual processes in IGRT.
- This approach enhances accuracy and efficiency in radiotherapy image analysis.
- The method shows potential for improving clinical outcomes in image-guided radiotherapy.
