SU‐E‐J‐32: Dosimetric Evaluation Based On Pre‐Treatment Cone Beam CT for Spine Stereotactic Body Radiotherapy: Does Region of Interest Focus Matter?

Abstract

Purpose:Spine stereotactic body radiotherapy requires very conformal dose distributions and precise delivery. Prior to treatment, a KV cone‐beam CT (KV‐CBCT) is registered to the planning CT to provide image‐guided positional corrections, which depend on selection of the region of interest (ROI) because of imperfect patient positioning and anatomical deformation. Our objective is to determine the dosimetric impact of ROI selections.Methods:Twelve patients were selected for this study with the treatment regions varied from C‐spine to T‐spine. For each patient, the KV‐CBCT was registered to the planning CT three times using distinct ROIs: one encompassing the entire patient, a large ROI containing large bony anatomy, and a small target‐focused ROI. Each registered CBCT volume, saved as an aligned dataset, was then sent to the planning system. The treated plan was applied to each dataset and dose was recalculated. The tumor dose coverage (percentage of target volume receiving prescription dose), maximum point dose to 0.03 cc of the spinal cord, and dose to 10% of the spinal cord volume (V10) for each alignment were compared to the original plan.Results:The average magnitude of tumor coverage deviation was 3.9%±5.8% with external contour, 1.5%±1.1% with large ROI, 1.3%±1.1% with small ROI. Spinal cord V10 deviation from plan was 6.6%±6.6% with external contour, 3.5%±3.1% with large ROI, and 1.2%±1.0% with small ROI. Spinal cord max point dose deviation from plan was: 12.2%±13.3% with external contour, 8.5%±8.4% with large ROI, and 3.7%±2.8% with small ROI.Conclusion:A small ROI focused on the target results in the smallest deviation from planned dose to target and cord although rotations at large distances from the targets were observed. It is recommended that image fusion during CBCT focus narrowly on the target volume to minimize dosimetric error. Improvement in patient setups may further reduce residual errors.