ESTRO 36 Abstract Book

S896 ESTRO 36 _______________________________________________________________________________________________

Purpose or Objective This pilot study aims to prospectively assess the setup uncertainty in head and neck (HN) using thermoplastic mask immobilization on a dedicated 1.5T MR-sim for radiotherapy based on a cohort of healthy volunteers. Material and Methods 11 healthy volunteers received a total of 142 scans (each scanned 4-40 times) on a 1.5T MR-sim in a treatment position immobilized with customized 5-point thermoplastic mask with shoulder fixation to simulate HN- RT fractions. Volunteers were carefully aligned and positioned by RT therapists using a 3D external laser system. The imaging protocol consisted of a T1w SPACE sequence (TR/TE = 420/7.2ms, FOV = 470mm, 256 coronal slices, isotropic voxel size of 1.05mm, acquisition time ~5min, geometric distortion correction ON). Six Degree- of-Freedom rigid body registration based on normalized mutual information (sampling factor = 5%) were used to pair-wisely register images with respect to the first session for each subject. Systematic and random errors of translation and rotation in positional setup were calculated. One sample t-test and box-plot were used to evaluate positional variation of translation and rotation with a significance level of 0.05. Results The overall positional setup repeatability results were presented in Table I. The group mean translation was <1mm and mean rotation was <1º, respectively. The systematic error in LR, AP and SI translation was 0.48mm, 0.23mm and 0.53mm, respectively. The systematic error in roll, pitch and yaw rotation was 0.07°, 0.002° and 0.27°, respectively. The random error in corresponding direction was 1.83mm, 0.62mm and 1.88mm for translation, and 0.36°, 0.01° and 1.09° for rotation, respectively. 20 (~14.0%), 0 (~0%), and 25 (~18.6%) out of all 142 scan sessions had displacements >1mm in the LR, AP, and SI directions, respectively. Displacements >2 mm were seen only in the SI direction in one session (~0.7%). One-sample t-test showed significance of group mean error in AP translation ( p< 0.001) and all rotations ( p = 0.02 in roll and p <0.001 in pitch and yaw). The box-plot of inter-session translation and rotation was shown in Fig.1. No significant differences were observed in translation in LR and SI direction, while AP translation was significantly smaller. This could be partially explained by the fixed AP level of the MR-sim couch. The yaw rotation was significantly larger than roll and pitch rotation, which may result from the relatively inadequate immobilization performance in this direction, but yet to be further investigated.

Conclusion Small head and neck setup uncertainty and thus high positional repeatability could be achieved on a 1.5T MR- sim, suggesting the potentials of MRI for precise HN-RT. EP-1651 Dosimetric impact of rotations correction in Stereotactic RT. How much a 6DoF couch is useful? S. Chiesa 1 , S. Menna 2 , A.R. Alitto 1 , L. Azario 2 , G.C. Mattiucci 1 , M. Ferro 3 , M. Massaccesi 1 , A. Re 1 , A. Piermattei 2 , V. Valentini 1 , M. Balducci 1 1 Università Cattolica del Sacro Cuore -Fondazione Policlinico Universitario A. Gemelli, Radiation Oncology Department- Gemelli-ART, Rome, Italy 2 Università Cattolica del Sacro Cuore - Fondazione Policlinico Universitario A. Gemelli, Medical Physics Department, Rome, Italy 3 Università Cattolica del Sacro Cuore -Fondazione di Ricerca e Cura “Giovanni Paolo II”, Radiation Oncology Department, Campobasso, Italy Purpose or Objective Setup accuracy and organ motion control are essential in stereotactic radiation therapy (SRT) due to the use of sharp dose gradients and tight margins around the target volume. PRO-ART Project was designed to evaluate set up errors and dosimetric impact of rotational patient positioning correction using a 6-Degree of Freedom (6DoF) robotic couch. Material and Methods Patients with lung, brain and abdominal lesions were enrolled and immobilized with Uni-frame or trUpoint Arch for brain, Breast board or Body Pro-Lok for chest and abdomen (CIVCO support system) lesions. Eclipse™ Treatment Planning Systems (Varian Medical System®,Palo Alto,CA) was used for dose calculations of VMAT plans. A daily KV-Cone Beam Computed Tomography (CBCT) was performed before each treatment fraction and translational and rotational shifts were identified, recorded and applied on the Protura TM Robotic couch to correct the position. Using MIM 5.5.2 software, the simulation CT was rigidly registered with CBCT, considered as primary CT, for each fraction. After registration, translational errors were applied to correct the CT position, obtaining a tCT, i.e. CT with only translational errors correction. Then, rotational errors were corrected too, obtaining roto-translated CT (rtCT). Reference treatment plan was copied to translated tTP and to roto-transalted rtTP and dose distributions were re- calculated, obtaining two treatment plans for each fraction. DVHs dosimetric parameters were compared. Results In this study 179 CBCT were performed on 47 patients (14 with primary tumours and 33 with metastatic lesions) and 358 treatment plans were calculated (179 tTP and 179 rtTP). Geometric and dosimetric analysis are reported in Table 1. There was no correlation between translational and rotational shifts. Rotational shifts resulted greater than 1° in 40% of cases. Average variations in PTV and CTV V95% were negligible, but we observed variations of PTV V95% >2% in 30%, 8% and 12% of cases (CBCTs) respectively for brain, thorax and abdomen patients. OARs proximity to the target caused variations >2% and rotations around each axis could determine important changes depending