Sci-Thurs PM: Planning-09: Simulation of RapidArc Delivery Errors for Head and Neck Cancer

Abstract

RapidArc is a novel commercial delivery system that simultaneously varies the multi-leaf collimator (MLC), dynamic dose rate and gantry positions. Single arc 360° RapidArc plans were generated on Eclipse v8.6 with a prescription dose of 60 Gy (2.4Gy/fraction) for 3 head and neck cancer patients. Systematic and random errors were applied to the MLC (±0.5 mm, ±1 mm, ±2 mm), gantry positions (±0.25°, ±1°, ±2°) and MU per control point (±1.25%, ±2.5%, ±5%) by editing the DICOM plan file using an in-house MATLAB program. Erroneous DICOM plans were re-imported for dose calculation and compared to the original plan based on dose differences and 3D gamma analysis. For random errors, the maximum dose deviations for the three plans were 0.95±0.20Gy, 2.02±0.58Gy, 3.93±0.78Gy for 0.5, 1 and 2 mm errors in the MLC, 1.60±0.39Gy, 1.84±0.30Gy and 2.40±0.42Gy for the 0.25, 0.5 and 1° errors in gantry positions and 0.14±0.03Gy, 0.28±0.02Gy, 0.55±0.08Gy for 1.25, 2.5 and 5% errors in the MU per control point. For the systematic errors, the maximum dose deviations for the three plans were 3.72±0.87Gy, 7.31±1.67Gy, 14.42±3.09Gy for the 0.5, 1 and 2 mm errors in the MLC, 2.59±1.03Gy, 2.82±0.91Gy, 3.58±1.03Gy for the 0.25, 0.5 and 1° errors in the gantry and 0.97±0.16Gy, 2.10±0.49Gy, 3.96±0.57Gy for the 1.25, 2.5 and 5% errors in the MU per control point. For the delivery parameters and ranges of errors that were studied, systematic errors in the MLC positions produced the largest dose deviations to the final dose distribution as compared to the original plan.