Room: Karl Dean Ballroom A1
Purpose: DEAR is a novel dynamic radiation technique that can achieve highly conformal dose distributions through simultaneous couch and gantry motion during delivery. Recently we developed DEAR dose calculation algorithm using Geant4 based VirtuaLinac, and established framework of integrating it to Eclipse. To ensure safe delivery of the complex DEAR plan, comprehensive quality assurance must be performed. The purpose of this study is to develop a feasible QA method for DEAR.
Methods: In clinical IMRT and VMAT QA, the treatment plan is delivered on phantom/imager with minimum modification to ensure plan deliverability and dose accuracy. QA for DEAR follows the same logic. Verification plan is generated and applied to flat water phantom. It keeps dose rate, MUs, distances between control points, cutout, SSD unchanged. The gantry is reset to 0Â° and couch motion is limited in the lateral and longitudinal direction. Measurement is performed using EBT3 films in solid water phantom. DEAR doses are calculated using full VirtuaLinac simulation or hybrid method which uses pre-calculated VirtuaLinac small field dose as kernel and convolves with DEAR plan trajectory. 2D Î³-analysis is used to quantify the differences between calculated and measured doses. Two DEAR plans with uniform and non-uniformed beam fluence for 6(9) MeV electrons are used to demonstrate the technique.
Results: Comparing full simulation with measurement, the 2D Î³ passing rates (2%/2mm) are 94%(90%) for uniform and 94%(95%) for non-uniform plans at 6(9) MeV, demonstrating excellent agreement between VirtuaLinac calculation and measurement. Dose from hybrid method is almost identical to full simulation (100% pass rate at 1%/1mm), while full simulation uses ~350cpu hours for DEAR plan(10cm) to achieve 2.5% precision. Therefore, the hybrid methed is suitable for QA purpose.
Conclusion: We have developed and validated a feasible QA method for DEAR. Future study includes the gantry rotation and 3D couch motion.
Funding Support, Disclosures, and Conflict of Interest: This work is partially supported by a Varian's research grant.