Program Information
Effect of Dose Grid Position in VMAT Optimization Using Varian Photon Optimizer (PO) Versus Varian Progressive Resolution Optimizer (PRO)
H Norris*, A Morrow , Baylor Scott & White Health, Temple, TX
Presentations
PO-BPC-Exhibit Hall-13 (Saturday, March 18, 2017) Room: Exhibit Hall
Purpose: Varian has introduced a new IMRT/VMAT optimization algorithm, Photon Optimizer (PO), which utilizes a fixed dose grid (1.25, 2.5, or 5mm grid spacing) for fast multi-resolution dose calculations during optimization; versus the previous Progressive Resolution Optimizer (PRO) algorithm for VMAT, which utilizes a point cloud determined by the resolution and shape of user-defined segments. Calculation point cloud density is increased near structure surfaces and boundaries. Herein we seek to quantify the effect of dose grid boundaries as well as the effect of dose grid resolutions in PO.
Methods: We created a computational phantom dataset which included outer body, target, and avoid contours. Using identical optimization objectives and prescriptions, three plans were created on the phantom: PO with 2.5mm grid spacing; PO with 5.0mm grid spacing; and PRO with default resolution. Then the structures, body, and isocenter were all shifted relative to the dose grid, and each plan was reoptimized and recalculated. Plans were normalized uniformly, and DVH metrics were compared among the plans.
Results: For the 2.5mm grid and for six shifts relative to the dose grid, the minimum target dose changed by 22±8%; the maximum target dose changed by 5±2%; and the maximum avoid dose changed by 8±3%. For the 5.0mm grid, min target change=12±4%; max target change=4±2%, and max avoid change=10±3%. For the PRO, min target change=7±2%, max target change=1.1±0.4%; max avoid change=5±2%. Overall the deviations in the PO are roughly twice that of the PRO.
Conclusion: The effect of the dose grid boundary in the PO algorithm creates inconsistencies amongst plans due to volume sampling errors. This leads to reduced target coverage and reduced dose homogeneity over the target volume and increased doses to critical structures. The effect of dose grid size is under further investigation.
Contact Email: