Purpose: The constraint on a minimum segment area in VMAT optimization is utilized to restrain small MLC apertures due to limited accuracy on TPS’s small-field output factor modeling. Since this constraint is applied to a total MLC aperture area per control point (CP), it does not effectively prevent small island apertures, e.g., island aperture area<1cm². We quantify the amount of these apertures in clinical plans as the first step for investigating their dosimetric and delivery impact.
Methods: 21 head-and-neck, 21 pelvis-node, and 21 prostate-bed clinical VMAT plans are selected, which are optimized with Elekta Agility MLC in Pinnacle³. Each DICOM RT-plan is exported from TPS for analysis. A CP may have multiple unconnected, island apertures. A table is created for every CP to record the MLC parameters of each island aperture: index of leaf pair(s) forming the aperture, leaf positions, and aperture area. The process is performed for all CPs. Finally, Modulation Area Index, MAI(r), is defined as a cumulative histogram of these island-aperture areas, r, calculated from all CPs and normalized by the total MLC opening area. The MAI(r) quantifies as a percentage of the island aperture for r<0.1cm², r<0.8cm², and r<0.6cm².
Results: Average MAI(r<1cm²) is 3.4%, 1.3%, and 0.3% with maximal 8.22%, 2.67% and 1.6% for the head-and-neck, prostate-node, and prostate-bed plans; for r<0.8cm², average MAI is reduced to 2.3%, 0.8%, and 0.2%; when r is further decreased to <0.6cm², MAI is still non-zero as 0.6%, 0.1%, and 0.1%. Statistical significance, p<0.05, is observed with the three disease sites.
Conclusion: We propose a plan-specific metric quantifying VMAT modulation complexity: optimization for head-and-neck produces the most island apertures, while prostate-bed produces the least. Future study is to investigate the impact of these apertures on the passing rates of patient-specific QA and TPS’s dose calculation accuracy on target coverage.