Technical performance of a commercial laser surface scanning system for patient setup correction in radiotherapy

In conformal radiotherapy, careful setup of the patient and setup verification prior to irradiation is essential. The technical performance of a commercial 3D-surface imaging system (Galaxy, LAP Laser, Lüneburg, Germany) for patient setup correction was evaluated. The system reconstructs a 3D-surface model by scanning the patient with a laser line while a camera records its reflections. This surface model is then compared with a reference model and a setup correction with 6 degrees of freedom is derived. The calibration stability of the system was investigated using the daily check phantom of the manufacturer. The accuracy and reproducibility of the system were investigated with an anthropomorphic phantom by performing 1D- and 3D-shifts with and without breathing of the phantom, respectively. In addition, measurements in a healthy volunteer were performed. With a few exceptions, the day-by-day variations of the calibration were <0.5 mm in LAT and LNG direction and <1.5 mm in VRT direction, respectively. Besides day-by-day variations, also baseline-shifts of up to 3 mm were observed. The lowest observed accuracy of the system in detecting pre-defined shifts of the rigid phantom was found in lateral direction. Here, mean deviations of −0.15 ± 0.46 mm for 1D-shifts and −0.12 ± 0.26 mm for 3D-shifts were found. For the ventilated phantom, the lowest observed accuracy was found in vertical direction with mean deviations of 1.16 ± 0.6 mm for 1D-shifts and −0.45 ± 0.57 mm for 3D-shifts. In a healthy volunteer, the accuracy was lowest in longitudinal direction with 1.7 ± 1.5 mm. The overall technical accuracy of the surface imaging system can be considered to be acceptable for application in fractionated radiotherapy. For special radiotherapy techniques such as SBRT, an increased accuracy might be necessary. To define the clinical role of the system, patient studies for different target locations are required.