Evaluation of daily patient positioning for radiotherapy with a commercial 3D surface-imaging system (Catalyst™)

F. Walter, P. Freislederer, C. Belka, C. Heinz, M. Söhn, F. Roeder

To report our initial clinical experience with the novel surface imaging system Catalyst™ (C-RAD AB, Sweden) in connection with an Elekta Synergy linear accelerator for daily patient positioning in patients undergoing radiation therapy. We retrospectively analyzed the patient positioning of 154 fractions in 25 patients applied to thoracic, abdominal, and pelvic body regions. Patients were routinely positioned based on skin marks, shifted to the calculated isocenter position and treated after correction via cone beam CT which served as gold standard. Prior to CBCT an additional surface scan by the Catalyst™ system was performed and compared to a reference surface image cropped from the planning CT to obtain shift vectors for an optimal surface match. These shift vectors were subtracted from the vectors obtained by CBCT correction to assess the theoretical setup error that would have occurred if the patients had been positioned using solely the Catalyst™ system. The mean theoretical set up-error and its standard deviation were calculated for all measured fractions and the results were compared to patient positioning based on skin marks only. Integration of the surface scan into the clinical workflow did not result in a significant time delay. Regarding the entire group, the mean setup error by using skin marks only was 0.0 ± 2.1 mm in lateral, -0.4 ± 2.4 mm in longitudinal, and 1.1 ± 2.6 mm vertical direction. The mean theoretical setup error that would have occurred using solely the Catalyst™ was -0.1 ± 2.1 mm laterally, -1.8 ± 5.4 mm longitudinally, and 1.4 ± 3.2 mm vertically. No significant difference was found in any direction. For thoracic targets the mean setup error based on the Catalyst™ was 0.6 ± 2.6 mm laterally, -5.0 ± 7.9 mm longitudinally, and 0.5 ± 3.2 mm vertically. For abdominal targets, the mean setup error was 0.3 ± 2.2 mm laterally, 2.6 ± 1.8 mm longitudinally, and 2.1 ± 5.5 mm vertically. For pelvic targets, the setup error was -0.9 ± 1.5 mm laterally, -1.7 ± 2.8 mm longitudinally, and 1.6 ± 2.2 mm vertically. A significant difference between Catalyst™ and skin mark based positioning was only observed in longitudinal direction of pelvic targets. Optical surface scanning using Catalyst™ seems potentially useful for daily positioning at least to complement usual imaging modalities in most patients with acceptable accuracy, although a significant improvement compared to skin mark based positioning could not be derived from the evaluated data. However, this effect seemed to be rather caused by the unexpected high accuracy of skin mark based positioning than by inaccuracy using the Catalyst™. Further on, surface registration in longitudinal axis seemed less reliable especially in pelvic localization. Therefore further prospective evaluation based on strictly predefined protocols is needed to determine the optimal scanning approaches and parameters.