skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: SU-D-202-04: Validation of Deformable Image Registration Algorithms for Head and Neck Adaptive Radiotherapy in Routine Clinical Setting

Abstract

Purpose: To evaluate the ROI contours and accumulated dose difference using different deformable image registration (DIR) algorithms for head and neck (H&N) adaptive radiotherapy. Methods: Eight H&N cancer patients were randomly selected from the affiliated hospital. During the treatment, patients were rescanned every week with ROIs well delineated by radiation oncologist on each weekly CT. New weekly treatment plans were also re-designed with consistent dose prescription on the rescanned CT and executed for one week on Siemens CT-on-rails accelerator. At the end, we got six weekly CT scans from CT1 to CT6 including six weekly treatment plans for each patient. The primary CT1 was set as the reference CT for DIR proceeding with the left five weekly CTs using ANACONDA and MORFEUS algorithms separately in RayStation and the external skin ROI was set to be the controlling ROI both. The entire calculated weekly dose were deformed and accumulated on corresponding reference CT1 according to the deformation vector field (DVFs) generated by the two different DIR algorithms respectively. Thus we got both the ANACONDA-based and MORFEUS-based accumulated total dose on CT1 for each patient. At the same time, we mapped the ROIs on CT1 to generate the corresponding ROIs on CT6more » using ANACONDA and MORFEUS DIR algorithms. DICE coefficients between the DIR deformed and radiation oncologist delineated ROIs on CT6 were calculated. Results: For DIR accumulated dose, PTV D95 and Left-Eyeball Dmax show significant differences with 67.13 cGy and 109.29 cGy respectively (Table1). For DIR mapped ROIs, PTV, Spinal cord and Left-Optic nerve show difference with −0.025, −0.127 and −0.124 (Table2). Conclusion: Even two excellent DIR algorithms can give divergent results for ROI deformation and dose accumulation. As more and more TPS get DIR module integrated, there is an urgent need to realize the potential risk using DIR in clinical.« less

Authors:
; ; ;  [1];  [1];  [2];  [3]; ; ;  [4]
  1. University of Science and Technology of China, Hefei, Anhui (China)
  2. (China)
  3. Saint Vincent Medical Center, Bridgeport, CT (United States)
  4. The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui (China)
Publication Date:
OSTI Identifier:
22624391
Resource Type:
Journal Article
Resource Relation:
Journal Name: Medical Physics; Journal Volume: 43; Journal Issue: 6; Other Information: (c) 2016 American Association of Physicists in Medicine; Country of input: International Atomic Energy Agency (IAEA)
Country of Publication:
United States
Language:
English
Subject:
60 APPLIED LIFE SCIENCES; 61 RADIATION PROTECTION AND DOSIMETRY; ACCELERATORS; ALGORITHMS; BIOMEDICAL RADIOGRAPHY; COMPUTERIZED TOMOGRAPHY; HEAD; HOSPITALS; IMAGE PROCESSING; IMAGES; MEDICAL PERSONNEL; NECK; NEOPLASMS; PATIENTS; RADIATION DOSES; RADIOTHERAPY; SKIN; SPINAL CORD; VALIDATION

Citation Formats

Zhang, L, Pi, Y, Chen, Z, Xu, X, Wang, Z, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, Shi, C, Long, T, Luo, W, and Wang, F. SU-D-202-04: Validation of Deformable Image Registration Algorithms for Head and Neck Adaptive Radiotherapy in Routine Clinical Setting. United States: N. p., 2016. Web. doi:10.1118/1.4955644.
Zhang, L, Pi, Y, Chen, Z, Xu, X, Wang, Z, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, Shi, C, Long, T, Luo, W, & Wang, F. SU-D-202-04: Validation of Deformable Image Registration Algorithms for Head and Neck Adaptive Radiotherapy in Routine Clinical Setting. United States. doi:10.1118/1.4955644.
Zhang, L, Pi, Y, Chen, Z, Xu, X, Wang, Z, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, Shi, C, Long, T, Luo, W, and Wang, F. Wed . "SU-D-202-04: Validation of Deformable Image Registration Algorithms for Head and Neck Adaptive Radiotherapy in Routine Clinical Setting". United States. doi:10.1118/1.4955644.
@article{osti_22624391,
title = {SU-D-202-04: Validation of Deformable Image Registration Algorithms for Head and Neck Adaptive Radiotherapy in Routine Clinical Setting},
author = {Zhang, L and Pi, Y and Chen, Z and Xu, X and Wang, Z and The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui and Shi, C and Long, T and Luo, W and Wang, F},
abstractNote = {Purpose: To evaluate the ROI contours and accumulated dose difference using different deformable image registration (DIR) algorithms for head and neck (H&N) adaptive radiotherapy. Methods: Eight H&N cancer patients were randomly selected from the affiliated hospital. During the treatment, patients were rescanned every week with ROIs well delineated by radiation oncologist on each weekly CT. New weekly treatment plans were also re-designed with consistent dose prescription on the rescanned CT and executed for one week on Siemens CT-on-rails accelerator. At the end, we got six weekly CT scans from CT1 to CT6 including six weekly treatment plans for each patient. The primary CT1 was set as the reference CT for DIR proceeding with the left five weekly CTs using ANACONDA and MORFEUS algorithms separately in RayStation and the external skin ROI was set to be the controlling ROI both. The entire calculated weekly dose were deformed and accumulated on corresponding reference CT1 according to the deformation vector field (DVFs) generated by the two different DIR algorithms respectively. Thus we got both the ANACONDA-based and MORFEUS-based accumulated total dose on CT1 for each patient. At the same time, we mapped the ROIs on CT1 to generate the corresponding ROIs on CT6 using ANACONDA and MORFEUS DIR algorithms. DICE coefficients between the DIR deformed and radiation oncologist delineated ROIs on CT6 were calculated. Results: For DIR accumulated dose, PTV D95 and Left-Eyeball Dmax show significant differences with 67.13 cGy and 109.29 cGy respectively (Table1). For DIR mapped ROIs, PTV, Spinal cord and Left-Optic nerve show difference with −0.025, −0.127 and −0.124 (Table2). Conclusion: Even two excellent DIR algorithms can give divergent results for ROI deformation and dose accumulation. As more and more TPS get DIR module integrated, there is an urgent need to realize the potential risk using DIR in clinical.},
doi = {10.1118/1.4955644},
journal = {Medical Physics},
number = 6,
volume = 43,
place = {United States},
year = {Wed Jun 15 00:00:00 EDT 2016},
month = {Wed Jun 15 00:00:00 EDT 2016}
}