Stereotactic body radiation therapy (SBRT) delivers relatively high doses to a target, typically in a hypofractionated fashion. With large fraction sizes, ideal target margins must minimize normal tissue exposure while ensuring tumor coverage. Daily On-Board Imaging (OBI) for target localization facilitates reductions in margins. In this study, we qualitatively compared 2D orthogonal kV vs. 3D cone beam CT (CBCT) OBI techniques and their abilities to assess the degrees of inter- and intra-fractional variation in patients undergoing SBRT.
Materials/Methods
32 patients with 40 lesions received 123 fractions of SBRT. Customized alpha cradles were used for immobilization, and all patients had CT-based simulation. Patients with lung, liver, or adrenal masses underwent 4D CTs to evaluate respiratory effects on tumor motion. Based on tumor- and patient-specific factors, either gated or free breathing techniques were used. Patients were initially aligned with lasers to external marks drawn on skin and cradles. Then, 2D orthogonal images were taken and aligned to the planning DRR using bony landmarks. The treatment table was shifted based on 2D matching, and changes to the isocenter in the anterior-posterior (AP), cranial-caudal (CC), and medial-lateral (ML) planes were recorded. Pre-treatment CBCT images were acquired and matched to the planning CT based on soft tissue anatomy, clips, fiducial markers, and bone. Additional changes in isocenter position due to CBCT were recorded. Finally, post-treatment 2D kV or 3D CBCT images were taken and compared to pre-CBCT images to reflect any intra-fractional changes. Absolute averages, standard deviations, and root mean squares (RMS) were calculated and displayed.
Results
Isocenter shifts recorded after 2D, pre-treatment CBCT, and post-treatment OBI are shown below. After initial set up to external marks with laser guidance, 2D kV images revealed set up deviations of 0.60 cm (RMS). CBCT detected additional isocenter shifts of 0.39 cm (RMS). Post-treatment OBI demonstrated intra-fractional variations of 0.14 cm (RMS). The individual shifts seen within the standard 3 orthogonal planes (AP, CC, ML) showed no obvious directional error bias. With regard to site-specific shifts, the deviations for liver lesions appeared to exceed those for lung and spine masses.
Conclusions
For patients undergoing SBRT, CBCT allows further refinement of treatment set up and target localization beyond that provided by 2D bony registration. These advances in target tracking/immobilization during treatment allow for improved confidence in the accuracy and precision of treatment delivery in SBRT.
Author Disclosure: D.S. Yoo, None; Z. Wang, None; J. Nelson, None; J. Kirkpatrick, None; S. Yoo, None; J. Wu, None; J. Meyer, None; N. Larrier, None; L. Marks, None; F. Yin, None.
FULL TEXT
