Toward Robust Patient-Specific Quality Assurance for Magnetic Resonance-Guided Online Adaptive Radiation Therapy
Abstract
Purpose
To develop and evaluate a log-file–based PSQA method for the ViewRay MRIdian MR-Linac that incorporates actual machine delivery information to improve the robustness of adaptive treatment plan verification.
Methods
RT plan and delivery log files were parsed to extract monitor units (MU), gantry angle, and multileaf collimator (MLC) positions. Segment apertures were reconstructed by interleaving the upper and lower stacks of the dual-focused, dual-stack ViewRay MLC as binary masks shifted by half a leaf width, yielding an effective resolution of 0.10 mm in the leaf motion direction and 4.15 mm in the leaf width direction at isocenter. MU-weighted beam’s-eye-view fluence maps were generated per gantry angle by summing segment apertures weighted by MU. Planned and delivered fluence maps were compared using gamma analysis with a 1%/1 mm criterion. All software solutions were developed in-house.
Results
The method was evaluated on an adaptive intensity-modulated radiation therapy (IMRT) plan comprising 28 fields. Gamma analysis (1%/1 mm) of the reconstructed fluence maps yielded passing rates ≥98% for 27 of 28 gantry angles, with one gantry angle exhibiting a 93% passing rate.
Conclusion
A log-file–based PSQA method was developed for adaptive treatment plan verification on the MRIdian MR-Linac. The proposed approach provides delivery-informed plan verification and complements existing independent dose calculation methods toward more robust PSQA in MR-guided radiotherapy.