Abstract
BACKGROUND: Fractionated stereotactic radiosurgery (fSRS) using LINAC-based volumetric-modulated arc therapy (VMAT) has become widely adopted and achieved favorable outcomes. However, this technique involves extremely small field sizes, making pre-treatment quality assurance (QA) with high-resolution detectors essential. PURPOSE: This study aimed to evaluate the error detection sensitivity and utility of EPID using 3D in vivo dosimetry based on dose-volume histogram (DVH) analysis for brain fSRS pre-treatment QA. METHODS: VMAT plans were generated for spherical planning target volumes (PTVs) with diameters of 1-3 cm centered in SRS MapCHECK. A monitor unit (MU) output error was introduced in increments, and the leaf gap width was systematically opened and closed. Finally, multi-leaf collimator (MLC) shift errors were applied. The error detection sensitivity was evaluated with these intentional errors. We compared this with the SRS MapCHECK 2D gamma analysis conventionally used in pre-treatment QA. For in vivo dosimetry, 3D dose calculations were performed with PerFRACTION using cine-mode electronic portal imaging device images and log files. DVH parameters such as D98%, D95%, D2%, and average doses of the PTV were compared between the base plan and the error plans. The 2D gamma analysis was performed using global gamma of 3%/2 mm, 3%/1 mm, 2%/2 mm, 2%/1 mm, and 1%/1 mm. RESULTS: The 2D gamma analysis revealed no MU output errors for the 1 cm target across all criteria. The 2%/1 mm criterion effectively detected errors in leaf gap width and MLC shifts in some cases. Conversely, all error types were detected by 3D in vivo dosimetry using D95% of the target. CONCLUSION: In pre-treatment QA for brain fSRS, 3D in vivo dosimetry based on DVH analysis demonstrated superior error detection sensitivity compared to conventional 2D gamma analysis. Our results suggest that 3D in vivo dosimetry is a useful tool for pre-treatment QA in brain fSRS.