2631 - End-to-End Credentialing for Simulation-Omitted MR-guided Adaptive Radiotherapy: A Streamlined and Practical Approach

Purpose/Objective(s): Simulation-omitted MR-guided radiotherapy (MRgRT) can shorten time-to-treatment, reduce reliance on simulation imaging resources, and streamline patient scheduling. It leverages diagnostic MR imaging for reference planning and MR-guided adaptive radiotherapy (ART) platform for online re-optimization to account for the patient’s actual setup and anatomy variation at the time of treatment. However, no standardized end-to-end credentialing framework currently established to verify the workflow safety, systematic consistency, accuracy between dose calculation and delivered dose. To address this gap, we developed a streamlined and practical end-to-end credentialing test using a commercially available phantom with interchangeable inserts, enabling institutions to independently validate simulation-omitted MRgRT workflows and support safe clinical implementation.

Materials/Methods: The credentialing approach was implemented on a 1.5T MR-Linac system using a MR-compatible phantom, which supports point dose measurement at multiple anatomical locations (target, spinal cord, and liver) and planar film dosimetry. A calibrated PinPoint MR ionization chamber was used for point-dose measurements, and Gafchromic film was used for planar dose verification. Pseudo-diagnostic MR image sets and corresponding digital structures were created and intentionally deformed to mimic anatomical variability expected in simulation-omitted workflows. The electron density of structures was captured from CT images. A reference treatment plan was created using an institutional lung prescription 60Gy with an IMRT technique and Monte Carlo dose calculation. Standard simulation-omitted MRgRT online adaptive workflows were followed. Adaptive plans generated from daily MRIs were delivered over multiple fractions. Three point-dose measurements at PTV, spinal cord and liver were compared against TPS-reported values, and sagittal and coronal planar film results were evaluated using global gamma analysis.

Results: The point-dose differences were < 1.8% of PTV and < 2cGy in liver and spinal cord for the adaptive plans, despite the chamber positioning uncertainty and potential electron returned effects surrounding the chamber. All planar film measurements exceeded a 93% gamma pass rate using 3%/3 mm.

Conclusion: This work establishes a streamlined, practical, cost-effective end-to-end credentialing methodology for simulation-omitted adaptive MRgRT workflows. The proposed approach verifies the accuracy of online re-optimization and dose delivery, supporting safe clinical implementation and cross-institutional standardization.