Quantification of Treatment Plan Deliverability in Breast Volumetric-modulated Arc Therapy With Agility Multi-leaf Collimator

The aim was to assess the complexity of breast volumetric-modulated arc therapy (VMAT) plans using various indices and to evaluate their performance through gamma analysis in predicting plan deliverability. A total of 285 VMAT plans for 260 patients were created using the VersaHD™ linear accelerator with a Monaco treatment planning system. Corresponding verification plans were generated using the ArcCHECK detector, and gamma analysis was conducted employing various criteria. Twenty-eight plan complexity metrics were computed, and Pearson's correlation coefficients were determined between the gamma passing rate (GPR) and these metrics. The average GPR values for all plans were 97.7%, 89.9%, and 78.0% for the 2 mm/2%, 1 mm/2%, and 1 mm/1% criteria, respectively. While most complexity metrics exhibited weak correlations with GPRs under the 2 mm/2% criterion, leaf sequence variability (LSV), plan-averaged beam area (PA), converted area metric (CAM), and edge area metric (EAM) demonstrated the most robust performance, with Pearson's correlation coefficients of 0.57, 0.50, -0.70, and -0.56, respectively. Metrics related to beam aperture size and irregularity, such as LSV, PA, CAM and EAM, proved to be reasonable predictors of plan deliverability in breast VMAT.