Evaluation of setup errors for head-and-neck cancer localized with final isocenter marking method via cone beam CT

Objective: To evaluate the setup errors of image guided radiation therapy (IGRT) for head-and-neck cancer using kilovoltage cone beam CT (kV CBCT). Methods: 256 patients with head-and-neck cancer were treated with intensity modulated radiation therapy (IMRT) from March 2009 to October 2011. All patients were immobilized with head-and-neck mask and localized with final isocenter marking method using the Philips PQS CT or Philips Brilliance CT Big Bore scanners,which were equipped with LAP movable laser systems. The CT images were transferred to a Varian Eclipse V8.6 workstation for contouring and planning. A kV cone-beam CT scans was acquired, and registered before the treatment for every patient on a Varian iX linear accelerator via OBI system. The setup errors in the right-left (RL), superior-inferior (SI), and anterior-posterior (AP) directions were recorded. Results: The setup errors for the 473 datasets followed a Gaussian distribution. The systematic errors ± random errors in the RL,SI and AP were (-0.6 ± 1.3), (0.5 ± 1.6) and (0.9 ± 1.7) mm,respectively. The planning target volume (PTV) margins were calculated respectively as 2.4, 2.4 and 3.4 mm according to the formula of M=2.5Σ +0.7δ The margins of 288 sets of data using the Big Bore CT scanner were calculated as 2.0, 2.1 and 1.7 mm,respectively. Conclusions: The setup errors using final isocenter marking method are smaller than those using reference point marking method. The result derived from this retrospective study could be used to set the margin between CTV and PTV. (authors)