|Table of Contents|

Setup errors of kilo-voltage CBCT image-guided radiotherapy for nasopharyngeal carcinoma under different registration conditions(PDF)

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

Issue:
2020年第3期
Page:
282-288
Research Field:
医学放射物理
Publishing date:

Info

Title:
Setup errors of kilo-voltage CBCT image-guided radiotherapy for nasopharyngeal carcinoma under different registration conditions
Author(s):
ZHANG Guoqian ZHANG Shuxu ZENG Qingxing LUO Songgui LI Ping YANG LuWANG Linjing ZHOU Lu WANG Ruihao LIAO Yuliang
Department of Radiotherapy, Cancer Hospital of Guangzhou Medical University, Guangzhou 510095, China
Keywords:
Keywords: nasopharyngeal carcinoma kilo-voltage cone-beam computed tomography bone registration grey registration setup error
PACS:
R811.1;R730.55
DOI:
DOI:10.3969/j.issn.1005-202X.2020.03.005
Abstract:
Abstract: Objective To compare the results obtained by different registration methods in the kilo-voltage cone-beam computed tomography (CBCT) image-guided radiotherapy for nasopharyngeal carcinoma (NPC) for providing a reference for setup correction. Methods The CBCT images of 100 NPC patients who were treated in the Cancer Hospital of Guangzhou Medical University between May 2018 and August 2019 were analyzed retrospectively. Two registration areas, namely large area (the whole scanning area) and small area (planning target volume), and two registration algorithms, namely bone and grey registration algorithms, were adopted in the study. According to different combinations of the registration area and different registration algorithms, 4 registration methods, namely large area and bone registration (LB), large area and grey registration (LG), small area and bone registration (SB), and small area and grey registration (SG), were used for kilo-voltage CBCT image registration. The translation errors and rotation errors were analyzed. Results The average translation error range of the isocenter after the registration with 4 different methods was -0.29~1.07 mm, and the average rotation error ranged from -0.1° to 0.61°. No significant differences were found between LB and SB in the translation and rotation errors (P=0.05-0.82), except for the translation error in Y direction (P=0.00); and there was no statistical difference between LG and SG in all translation and rotation errors (P=0.14-0.64). Moreover, significant differences were found between LG and LB in the translation and rotation errors (P=0.00-0.02), except for the translation and rotation errors in Y direction (P=0.67, 0.57); and the comparison between SB and SG showed that there were significant differences in translation and rotation errors (P=0.00-0.02), except for the translation errors in X direction and the rotation error Y direction (P=0.36, 0.72). The result of Pearson correlation analysis revealed that there were positive correlations between any two registration methods in translation and rotation errors in X, Y and Z directions (R=0.48-0.98, P<0.01). With the registration of LG and SG, the 95% limits of agreement (LoA) of translation errors in X, Y and Z directions was [1.19, -1.25], [0.95, -1.01], [1.13, -1.31], and that of the rotation errors was [0.80, -0.85], [0.69, -0.60], [1.02, -1.13]. The setup error was consistent, according to the 2 mm and 2° boundary standards. Conclusion When kilo-voltage CBCT is used for automatic image registration in image-guided radiotherapy for NPC, the registration results in three-dimensional directions may be different due to different registration modes (bone or grey registration); and with the full scanning area or planning target volume as registration area, the registration results are similar.

References:

Memo

Memo:
-
Last Update: 2020-04-02