Related Experiment Video
Updated: Jun 18, 2026
06:09
Measuring 3D In-vivo Shoulder Kinematics using Biplanar Videoradiography
Published on: March 12, 2021
Automatic marker detection and 3D position reconstruction using cine EPID images for SBRT verification
Sang-June Park1, Dan Ionascu, Fred Hacker
1Department of Radiation Oncology, Dana-Farber/Brigham and Women's Cancer Center, Harvard Medical School, Boston, Massachusetts 02115, USA. spark@lroc.harvard.edu
Medical Physics
|November 26, 2009
Summary
An automated algorithm accurately extracts markers from electronic portal imaging device (EPID) images, improving stereotactic body radiation therapy (SBRT) validation. This method enhances accuracy and efficiency in tracking tumor motion and patient setup errors.
Area of Science:
- Medical Physics
- Radiation Oncology
- Image Processing
Background:
- Manual marker tracking for validating respiratory gating and SBRT using electronic portal imaging devices (EPIDs) is time-consuming.
- Developing automated methods is crucial for improving the efficiency and utility of EPID-based validation.
Purpose of the Study:
- To develop and evaluate an automatic algorithm for extracting fiducial markers from EPID images.
- To reconstruct the 3D positions of markers for enhanced accuracy in SBRT validation.
- To assess the algorithm's performance in phantom and patient studies.
Main Methods:
- An image processing algorithm based on the Laplacian of Gaussian function was used for marker detection.
- A marker registration technique, utilizing image intensity and spatial transformations from CT data, was applied to reduce false positives.
- 3D marker positions were reconstructed by backprojecting 2D positions from EPID images.
Main Results:
- Spatial accuracies of <1 mm were achieved in 2D and 3D marker locations in static phantom studies.
- The marker detection success rate increased from 88.8% to 100% with the addition of the marker registration technique in dynamic phantom studies.
- Intrafractional tumor motion (3.1-11.3 mm) and interfractional patient setup errors (0.1-12.7 mm) were measured in SBRT patients.
Conclusions:
- The developed automatic algorithm efficiently and accurately extracts marker locations from MV images.
- This automated approach provides valuable data for off-line retrospective verification of SBRT by measuring intrafractional tumor motion and interfractional setup errors.