Program Information
Commissioning of 2D/3D Matching On a Varian TrueBeam
J Dorand*, D Ellerbusch , L Fong de los Santos , J Ma , D Pafundi , Mayo Clinic, Rochester, MN
Presentations
SA-B-BRD-2 (Saturday, March 5, 2016) 10:30 AM - 12:30 PM Room: Grand Ballroom D
Purpose: To commission the Varian TrueBeam 2D/3D 6 degree of freedom (6DoF) registration algorithm. 2D/3D matching uses an orthogonal 2D image pair matched to projections calculated from a 3D data set.
Methods: MIMI, BrainLab Pelvis, and Rando C-Spine phantoms were used for commissioning. 6DoF CBCT registration was used as the gold standard, with absolute maximum differences computed between 6DoF 2D/3D and CBCT automatch results. Automatch algorithm configuration parameters were investigated and compared to Varian suggested parameters using the MIMI phantom. The Varian suggested parameters were used to test variations in reference CT slice thickness, acquisition angle, kV/kV vs. MV/kV, ROI selection, and the inclusion of an additional step in the matching algorithm process.
Results: Varian suggested matching algorithm configuration parameters, along with additional parameter set definitions, yielded maximum absolute differences of 0.5 mm/0.6° and 0.8 mm/0.8°, respectively. Reference CT slice thicknesses of 1 mm and 2 mm for all three phantom geometries had maximum absolute differences of 0.4 mm/0.4° and 0.5 mm/0.3°, respectively. AP/Lat and posterior oblique acquisition angles yielded maximum absolute differences of 0.5 mm/0.4° and 0.8 mm/0.7°, respectively. Utilizing the MIMI and BrainLab pelvis phantoms, kV/kV vs. MV/kV results were 0.5 mm/0.3° compared to 0.7 mm/1.7°, respectively. Rando C-Spine ROIs of one vertebral body vs. three vertebral bodies yielded results of 0.7 mm/1.3° and 0.2 mm/0.7°, respectively, and the inclusion of an additional step in the matching algorithm resulted in differences of 0.1 mm/0.2° compared to 0.2 mm/0.7°.
Conclusion: The 6DoF 2D/3D registration algorithm was determined to perform as well as the 6DoF CBCT algorithm for the rigid phantom geometries tested. ROI selection was found to have the greatest impact on differences between 2D/3D and CBCT matching results. A prospective clinical pilot of head and neck patients will test the efficacy of 6DoF 2D/3D registration on non-rigid anatomy.
Contact Email: