On-line image guidance for frameless stereotactic radiotherapy of lung malignancies by cone beam CT: Comparison between target localization and alignment on bony anatomy

• Published in: Acta oncologica 2008, Vol.47 (7), p.1422-1431
• Main Authors: Masi, Laura, Casamassima, Franco, Menichelli, Claudia, Pasciuti, Katia, Doro, Raffaela, Polli, Caterina, D'imporzano, Elena, Bonucci, Ivano
• Format: Article
• Language: English
• Subjects: Bone and Bones - diagnostic imaging; Cone-Beam Computed Tomography; Humans; Lung Neoplasms - diagnostic imaging; Lung Neoplasms - surgery; Movement; Observer Variation; Online Systems; Radiosurgery; Respiration
• ISSN: 0284-186X; 1651-226X
• DOI: 10.1080/02841860802251534

Introduction. Free-breathing stereotactic radiotherapy for lung malignancies requires reliable prediction of respiratory motion and accurate target localization. A protocol was adopted for reproducibility and reduction of respiratory motion and for target localization by CBCT image guidance. Tumor respiratory displacements and tumor positioning errors relative to bony anatomy alignment are analyzed. Materials and method. Image guided SRT was performed for 99 lung malignancies. Two groups of patients were considered: group A did not perform any breathing control; group B controlled visually their respiratory cycle and volumes on an Active Breathing Coordinator (ABC) monitor during the acquisition of simulation CT and CBCT, and treatment delivery. GTV on end inhale and exhale CT data sets were fused in an ITV and the extent of tumor motion evaluated between these 2 phases. A pre-treatment CBCT was acquired and aligned to the reference CT using bony anatomy; for tumor positioning the ITV contour on the reference CT was matched to the visible tumor on CBCT. Interobserver variability of tumor positioning was evaluated. ITV and CBCT tumor dimensions were compared. Results. 3D tumor breathing displacement (mean±SD) was significantly higher for group A (14.7±9.9 mm) than for group B (4.7±3.1 mm). The detected differences between tumor and bony structure alignment below 3 mm were 68% for group B and 45% for group A, reaching statistical significance. Interobserver variability was 1.7±1.1 mm (mean±SD). Dimensions of tumor image on CBCT were consistent with ITV dimensions for group B (max difference 14%). Conclusions. The adopted protocol seems effective in reducing respiratory internal movements and margin. Tumor positioning errors relative to bony anatomy are also reduced. However bony anatomy as a surrogate of the target may still lead to some relevant positioning errors. Target visualization on CBCT is essential for an accurate localization in lung SRT.