[P262] Verification of Deformable Image Registration (DIR) program for clinical application in head and neck cancer. (August 2018)

• Record Type: Journal Article
• Authors: Suphaphong, Sithiphong; Tangboonduangjit, Puangpen; Worawongsakul, Rasin; Sakulsingharoj, Siwaporn
• Is Part Of: Physica medica. Volume 52(2018)Supplement 1
• Page Start: 175
• Publication Date: 2018-08
• Subjects: Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2018.06.541
• Languages: English
• ISSNs: 1120-1797

Abstract

Abstract : Purpose: Deformable image registration (DIR) program has played an importance role in radiotherapy, for example, image registration from multi-modality and adaptive radiotherapy (ART). Velocity™ v3.2.1 is a commercial software for multipurpose such as DIR, automatic segmentation and PET contouring. The reliability of DIR should be evaluated before clinical used. Nevertheless, most of verification program for DIR are commercial software. This study aims to verify the DIR program in case of there is no commercial verification software. Methods: I'mRT phantom™ with 3 inserted fiducial markers and Anthropomorphic Rando phantom™ (head and neck part) with 12 inserted fiducial markers were used for deformable registration evaluation. For evaluation of DIR, the translation (lateral, longitudinal and vertical directions), rotation, pitch and yaw rotation of I'mRT™ and Rando phantom™ were performed for calculation of target registration error (TRE). Eclipse™ treatment planning v13.6 was used for TRE by measuring the center of line profile of fiducial markers among translated phantoms. Nine cases of head and neck cancer CT images were selected for evaluation the TRE and DVH changes of tumor and critical organs. Results: Target registration error of rigid phantom was 0.39 ± 0.40 mm, maximum value was 1.0 mm. For anatomical phantom, TRE was 0.93 ± 1.01 mm with maximum value of 5.5 mm. The highest value was found at the border bone region. For clinical CT images, TRE was 2.4 ± 1.76 mm and maximum difference were found at uvula and epiglottis region. For inverse consistency method, the result showed good agreement (percentage differences were less than 3%) of DVH report between initial CT and deformed CT images. Conclusions: Velocity™ can be used for image registration in head and neck region. For more reliability, further investigation will be performed in clinical data to achieve consistency evaluation.