130. Inter-fraction variability of respiratory-induced tumour motion in lung SBRT: 4D-CBCT

Abstract Purpose The aim of this study was to estimate the inter-fraction variability of respiratory-induced motion of lower lobes target volumes in lung SBRT treatments. Methods 72 4D-CBCT scans from 17 patients were studied. For SBRT planning, patients were scanned on a GE Bright-Speed 4D-CT scanner and multi-phases images were reconstructed. For breath control a compression abdominal belt was used. GTV was contoured on each phase and the ITV was generated. A PTV was created from the ITV with a 5 mm margin. In-room 4D image-guidance was performed by the Symmetry XVI Elekta system. 4D-CBCT data sets were matched with reference CT images using two image registration steps. Firstly a clipbox containing patient anatomy for set-up correction was defined and then a mask of 5 mm around PTV was created for the second registration. Only the volume of the reference planning CT within the mask is used for soft-tissue registration. Matching results were used for table shifts to correct daily patient setup and base line shifts. Baseline shift was measured subtracting clipbox from mask shift. Tumour motion was quantified for all the treatment sessions as the absolute range of the respiratory cycle. The difference between the max and min range for each patient has been evaluated as an index of the inter-fraction target motion variability. Results A large inter-fraction baseline shifts of lung tumours has been observed (see table). Inter-fraction variability of respiratory-induced tumour motion resulted within 2 mm in LL and AP directions and higher than 3 mm in 50% with max values up to 5.5 mm. Conclusions Respiration correlated 4D-CBCT allows accurate target localization in the presence of breathing induced target motion and validation of PTV margins. Inter-fraction variability of the tumour motion observed in this study suggested to increase the ITV margin to at least 7 mm in SI direction.