Study of deformable image registration parameters and their effect on the determination of an adaptive radiotherapy protocol in VMAT patients

• Κύριος συγγραφέας: Αλαμάνη, Ευγενία
• Άλλοι συγγραφείς: Alamani, Evgenia
• Γλώσσα: English
• Έκδοση: 2022
• Θέματα: Deformable image registration; Head and neck cancer; Adaptive radiotherapy; Dice similarity coefficient; Anatomic changes; Καταχώρηση εικόνων με εύκαμπτο μετασχηματισμό; Καρκίνος κεφαλής-τραχήλου; Προσαρμοσμένη ακτινοθεραπεία; Ανατομικές διακυμάνσεις

Purpose: The aim of this study is to attempt to determine the optimal time or a time-frame for the creation of an adaptive radiation therapy (ART) replan protocol by monitoring and evaluating anatomical changes of the clinical target volumes (CTV) and parotid glands during the course of Volumetric modulated arc therapy (VMAT) for head and neck cancer patients. Materials and methods: 31 anonymized patients diagnosed with head and neck cancer were classified into three groups based on the anatomical location of the tumor. Among the 31 patients, 13 patients emerged with malignant neoplasm of oropharynx, 11 with malignant neoplasm of larynx and 7 with malignant neoplasm of nasopharynx. SmartAdapt® Deformable image registration (DIR) software was utilized for the image registration process. The data set of each patient included the planning CT image (pCT) and a Cone Beam Computed Tomography (CBCT) image acquired prior to fraction. Each pair of CT-CBCT images was rigidly registered based on bony anatomy. Rigid registration process was followed by an Automatic Deformable image registration process between the selected CT-CBCT image pairs. The deformed structures generated by the DIR process were analyzed relative to the corresponding structure volumes on the planning CT image to assist in monitoring anatomical changes. Results: The most demanding structures to be reproduced by the DIR software were the parotid glands, exhibiting mean DSC score values lower than those of the CTV. In addition, both parotid glands underwent a significant reduction in their volumes in comparison to the CTV for all three head and neck categories. The range of the mean percentage of volume change for the CTV was between -0.87% to -8.12%, 0.88% to -21.41% and -0.03% to -4.78% for the oropharynx, larynx and nasopharynx patients respectively. Accordingly, the range of the mean percentage of volume change for the left parotid gland was between -0.88% to -21.41%, -0.71% to -15.95% and -2.19% to -31.55% for the oropharynx, larynx and nasopharynx patients respectively. The range of the mean percentage of volume change for the left parotid gland was between -0.51% to -21.52%, -0.57% to -20.71% and -3.49% to -35.71% for the oropharynx, larynx and nasopharynx patients respectively. For oropharynx patients the optimal time for re-planning based on the CTV volume change was the 17th fraction and the 6th and 5th fractions based on the left and right parotid gland volume change. Accordingly, for larynx patients it was the 10th and 6th based upon the left and right parotid glands volume change and for nasopharynx patients fractions 4 and 5 based upon the left and right parotid volume change . Conclusions: The study of deformable image registration parameters can assist in monitoring anatomical variations. Significant volume changes of the parotid gland over the course of fractions can trigger the need for re-planning in order to avoid severe toxicities. The determination of the optimal time for re-planning for the three head and neck cancer categories based exclusively on anatomical changes can lead to the development and implementation of an adaptive radiotherapy clinical protocol that aims in automatization and optimization of the quality of treatment.