Room: AAPM ePoster Library
Purpose: To measure GTV motion for having accurate targeting in SBRT of NSCLC and for predicting the target motion in future real-time MRI or optical image-guided and respiration gated SBRT.
Methods: Fourteen lung GTVs from 0.5 to 111 cc were precisely delineated at the first phase in 4DCT scans. All GTVs were then transferred to other phases using local deformable image registration (DIR). DIR artifacts were corrected using threshold-based contouring. GTVs and the ipsilateral diaphragm displacements (DD), and GTV relative locations to the Apex, mediastinum and posterior chest wall (CW) are recorded. GTV relative locations of rZ, rY and rX are defined as ratios of (the stable reference surface to GTV center)/lung dimension, in Z from apex to diaphragm, Y from posterior CW to anterior CW, and X from mediastinum to lateral CW, respectively. GTV displacements were then linearly regressed with individual variables. Daily 4D CBCT were acquired for checking GTV motion and prediction.
Results: Measured GTV movements in the lateral, anterior and longitudinal directions are 0.3 to 9 mm, 1 to 5 mm and 0.3 to 17 mm, respectively. Measured DD are 1 to 24 mm with an average of 13 mm for 8 cases using abdominal compression and the same for 6 cases without abdominal compression. GTV lateral motion is small (< 3 mm excepting for a 9 mm movement) and unpredictable. GTV longitudinal and anterior displacements are predicted as rZ×DD - 0.1 with R2 = 0.85 and 0.3rY×DD+0.1 with R2 = 0.31. GTV trajectories were consistent with daily CBCT scans.
Conclusion: GTV motion is complicated with anatomic and other patient conditions but mostly measurable with errors of < 3 mm using 4D CT scans. A linear correlation of large GTV displacements with DD and relative locations might be useful in future real-time IGSBRT.
Not Applicable / None Entered.