ESTRO 2020 Abstract book

S921 ESTRO 2020

directly while the shape Y movement took the leaf thickness into account by either using a threshold or interpolation algorithm. The MLC control system was running in 1ms cycle mode handling displacement information and commanding new leaf positions. Detectors were fixed to the moving phantom (cos 2 amplitude of 2cm) with breathing cycles of 12/min up to 60/min. Films were evaluated for penumbra, gamma index (3%, 2mm) and dose difference for static, non-tracked and tracked scenarios. Ionization chamber measurements were used to show dynamic processes.

rotation offsets obtained by a Winston-Lutz (WL) test using the MV panel. Results Mean translational and mean rotational isocenter shifts of the non-coplanar treatment for the population of 7 volunteers are (0.24±0.09) mm and (0.15±0.07) degrees, respectively, averaged over the couch angles. Isocenter shift values larger than 0.5mm were found for couch angles 45˚ and 90˚ (fig.1.). These results can be correlated with the couch rotation offsets obtained by regularly repeated WL tests: mean vector deviations between the couch and radiation isocenters at gantry 0° are 0.36±0.15, 0.24±0.09, 0.30±0.14, 0.59±0.11 and 0.60±0.13 (mean±SD in mm) for couch rotations 0°, 315°, 270°, 90° and 45°, respectively, revealing isocenter shift values larger than 0.5 mm for couch rotations 90° and 45°.

Conclusion This work shows that it is feasible to monitor patient positioning for a non-coplanar single isocenter SRS treatment using a commercial OST with submillimeter and subdegree accuracy without compromising the desired 1mm GTV-PTV treatment margin. Further, a regular WL test should be part of the SRS-specific QA program of a linac. PO-1600 Evaluation of a High Dynamic Multileaf Collimator for Real-Time tumor tracking C. Murillo 1 , S. Seeber 1 , P. Haering 2 , C. Lang 2 , M. Splinter 2 1 DKFZ - Deutsches Krebsforschungszentrum, Electronic & Embedded Systems Development Lab, Heidelberg, Germany ; 2 DKFZ - Deutsches Krebsforschungszentrum, Research Group e040, Heidelberg, Germany Purpose or Objective Image guided radiotherapy (IGRT) and adaptive RT are used to minimize dose deposition in normal tissue due to inter-fractional motion. Besides this, tumor tracking provides an efficient way to deposit dose to a moving target. Therefore, a Multileaf Collimator (MLC) needs to be accurate, precise and fast. Known experiments, achieving tracking with 3cm/s (isocentric) MLCs, show low dosimetric accuracy for faster organ/tumor motion. A new in-house developed High Dynamic MLC (HDMLC) has therefore been evaluated. This system uses a real-time control unit, fulfills industrial automation standards and achieves leaf speeds up to 20cm/s. Quantification of the dosimetric accuracy and dose conformity are the main goals of this study. Material and Methods The HDMLC evaluation was done using a Siemens Artiste (X-ray 6 MV) and 1D motion phantom together with EBT films (Gafchromic) and a PTW 729 detector. Experiments were made with an X1 side only HDMLC (39 x 3mm leaves) while X2 was linac formed. A magnetic encoder was placed on the phantom and acquired by the MLC controller. By placement of the phantom 45° relative to the leaf motion, X and Y movements were achieved (Fig 1). Circle and square shapes were used for the experiment. The MLC shape X movement followed the encoder signal

Results In general, experiments performed with the HDMLC presented significant improvements in all aspects compared to the results without tracking. Higher cycle rates lead to a higher penumbra (1 mm for 12/min and 6mm for 60/min), which is up to 15 times better than untracked. The gamma distribution shows corresponding results (Fig 2). Interpolation for the y axis was most effective on more rounded target shapes. The main drawback observed in all experiments was the mechanical backslash of the leaves on the HDMLC, which could be resolved by a direct coupling of the position sensor into the leaf.

Conclusion