Sensitivity analysis of physics and planning SmartArc parameters for single and partial arc VMAT planning

We investigate the sensitivity of various physics and planning SmartArc parameters to generate single and partial arc VMAT plans with equivalent or better plan quality as IMRT. Patients previously treated with step-and-shoot IMRT for several treatment sites were replanned using SmartArc. These treatment sites included head and neck, prostate, lung, and spine. Effect of various physics and planning SmartArc parameters, such as continuous vs. binned dose rate, dynamic leaf gap, leaf speed, maximum delivery time, number of arcs, and control point spacing, were investigated for Elekta Axesse and Synergy linacs. Absolute dose distribution was measured by using the ArcCHECK 3D cylindrical diode array. For all cases investigated, plan metrics such as conformity indices and dose homogeneity indices increased, while plan QA decreased with increasing leaf speed. Leaf speed had a significant impact on the segment size for low dose per fractionation cases. Constraining leaf motion to a lower speed not only avoids tiny large leaf travel and low-dose rate value, but also achieves better PTV coverage (defined as the volume receiving prescription dose) with less total MUs. Maximum delivery time, the number of arcs, and the spacing of control points all had similar effects as the leaf motion constraint on dose rate and segment size. The maximum delivery time had a significant effect on the optimization, acting as a hard constraint. Increasing the control point spacing from 2 to 6 degrees increased the PTV coverage, but reduced the absolute dose gamma passing rate. Plans generated using continuous and binned dose rate modes did not show any difference in the quality and the delivery for the Elekta machines. Dosimetric analysis with a 3D cylindrical QA phantom resulted in 93.6%-99.3% of detectors with a gamma index (3%/2 mm) < 1 for all cases.