Room: Stars at Night Ballroom 2-3
Purpose: Online adaptive replanning (OARP) in magnetic resonance guided radiotherapy offers unprecedented ability to improve beam targeting precision. The OARP workflow cannot accommodate measurement-based patient-specific quality assurance (QA). Instead, the re-optimized plans are verified against an independent Monte Carlo-based dose calculations. Due to the complexity of the re-planning process, a comprehensive end-to-end verification of software QA is pivotal. The purpose was to assess performance of volumetric scintillation imaging system for end-to-end OARP QA as a part of real-time scintillation-based 3D dosimetry solution.
Methods: A multi-target plan was delivered by ViewRay MR-linac (0.34T, 6MV FFF) to a water-based scintillator phantom (1g/L quinine sulphate, Ã¸127x300mm). Intensified, gated CMOS camera recorded projected scintillation images through the base of the phantom at 20Hz framerate. Cumulative scintillation image served as surrogate to the projected dose distribution. First, gamma pass-rate was evaluated between the measured scintillation image and CT-based dose plan. Second, MR image of the phantom was recorded by MR-Linac and used for dose calculation with an artificially modified structure set. The adapted MR-based plan was delivered to the phantom and gamma pass-rate between scintillation image and MR-based dose image was evaluated. Finally, 3D geometrical beam volume was reconstructed from scintillation images and multi-leaf collimator position logs in order to allow 3-dimensional gamma and dose volume histogram (DVH) analysis.
Results: 2D gamma pass-rate of the measured dose distribution compared against the calculated CT-based dose was 96% at 3%/3mm criteria. The pass-rate of MR-based optimized plan dropped to 94%. The majority of non-passing pixels were found in the sub-surface area, most likely caused by an electron-return effect.
Conclusion: Scintillation images of both original and re-planned deliveries agreed with treatment planning system calculations. Scintillation surrogated dose real dose distribution and enabled high-resolution, rapid QA of online adaptive replanning, as well as a 3D DVH comparisons.
Funding Support, Disclosures, and Conflict of Interest: This work has been sponsored by NIH research grant R01EB023909. B Pogue is the president and co-founder of DoseOptics LLC. P Bruza is the principal investigator in SBIR subaward B02463 (prime award NCI R44CA199681, DoseOptics LLC). This work was not financially supported by DoseOptics.