Cloud-Based Monte Carlo Patient-Specific Quality Assurance (QA) Method for Volumetric-Modulated Arc Therapy (VMAT): Clinical and Educational Impacts
X Chen*, L Xing, E Mok, G Luxton, K Bush, Stanford University Cancer Center, Stanford, CAMO-D-108-1 Monday 2:00PM - 3:50PM Room: 108
Purpose:Patient-specific QA for VMAT is labor-intensive,especially in cases which tend to be associated with severe heterogeneities or small aperture beams. We developed a robust cloud-based Monte Carlo method as an independent QA tool or backup for commercial dosimeters. It is also educationally innovative because it helps physicists to comprehend the underlying physics of complex VMAT procedures.
Methods:The workflow of the cloud-based voxel Monte Carlo method(cVMC++) developed here is the following: (1) After a VMAT plan is approved by the physician, a dose verification plan is created and delivered to the phantom using the Varian Trilogy™ or TrueBeam™ system. (2) Actual delivery parameters (i.e., dose fraction, gantry angle, and MLC leaf position) are extracted from dynamic log or trajectory files. (3) Based on the delivery parameters, a new phase space is constructed for each control point. The 3D dose distribution in the phantom containing the detector array (e.g., Delta 4 from ScandiDos) are then recomputed using cVMC++. (4) Comparison and Gamma index analyses are then conducted to evaluate the agreement between measured, recomputed, and planned dose distributions. To enhance the educational function, we are developing visualization and animation modules to display the intermediate processes such as leaf movement, dose distribution and DVH accumulation with each control point. The software has been wrapped as a web-based service so that it can readily be used by physicists via a browser. To test the robustness of this system, we examined several representative VMAT treatments.
Results:For all cases, agreement between measured, recomputed and planned doses was examined with the gamma criterion of >90% of the points satisfying <3% of relative dose difference and <3 mm of distance to agreement.
Conclusion:The proposed method offers a robust patient specific QA tool for VMAT treatments and provides a valuable educational tool for physicists.
Funding Support, Disclosures, and Conflict of Interest: NIH (1R01 CA133474) and Department of Radiation Oncology of Stanford University Seed Grant