Abstract
We compared the effect of set-up error and uncertainty on two radiation therapy treatment plans for head and neck cancer: one using intensity modulated radiation therapy (IMRT) and one using conventional three-dimensional conformal radiation therapy (3D-CRT). We used a Pinnacle3 (Philips Medical Systems, Markham, Ontario) system to create the two treatment plans (7-beam IMRT and 5-beam 3D-CRT) for the same volumetric data set, based on the objectives and constraints defined in the Radiation Therapy Oncology Group H-0022 protocol. In both plans, the dose–volume constraints for the targets and the organs at risk (OARS) were met as closely as the beam geometries would allow. Monte Carlo–based simulations of set-up error and uncertainty were performed in three orthogonal directions for 840 simulated “courses of treatment” for each plan. A systematic error (chosen from distributions characterized by standard deviations ranging from 0 mm to 6 mm) and random uncertainties (2 mm standard deviation) were incorporated. We used a probability approach to compare the sensitivities of the IMRT and the 3D-CRT plans to set-up error and uncertainty in terms of equivalent uniform dose (EUD) to targets and OARS. Based on the EUD analysis, the targets and OARS showed considerably greater sensitivity to set-up error with the IMRT plan than with the 3D-CRT plan. For the IMRT plan, target EUDS were reduced by 4%, 7.5%, and 10% for 2-mm, 4-mm, and 6-mm set-up errors respectively. However, even with set-up error, the mandible, spinal cord, and parotid EUDS always remained lower with the IMRT plan than with the 3D-CRT plan. We conclude that, when quantified by EUD, IMRT plan doses to OARS and targets are more sensitive to set-up error than are 3D-CRT-plan doses. However, as judged by the differences between target and OAR doses, IMRT retains its superiority over 3D-CRT, even in the presence of set-up error.